Vitreous Microsurgery

FIFTH EDITION
Vitreous Microsurgery
FIFTH EDITION

Steve Charles, MD
Vitreoretinal Surgeon and Founder
Charles Retina Institute
Clinical Professor of Ophthalmology
University of Tennessee
Memphis, Tennessee
Adjunct Professor of Ophthalmology
Columbia College of Physicians and Surgeons
New York City, New York

Jorge Calzada, MD
Vitreoretinal Surgeon
Charles Retina Institute
Clinical Instructor in Ophthalmology
University of Tennessee
Memphis, Tennessee

Byron Wood, MS
Chief Photographer and Director of Graphics
Charles Retina Institute
Memphis, Tennessee
Senior Executive Editor: Jonathan W. Pine, Jr.
Senior Product Manager: Emilie Moyer
Senior Manufacturing Manager: Benjamin Rivera
Marketing Manager: Lisa Lawrence
Design Coordinator: Joan Wendt
Production Service: SPi Technologies

© 2011 by LIPPINCOTT WILLIAMS & WILKINS, a WOLTERS KLUWER business

Two Commerce Square
2001 Market Street
Philadelphia, PA 19103 USA
LWW.com

All rights reserved. This book is protected by copyright. No part of this book may be reproduced in any form by any means, including
photocopying, or utilized by any information storage and retrieval system without written permission from the copyright owner, except for
brief quotations embodied in critical articles and reviews. Materials appearing in this book prepared by individuals as part of their official
duties as U.S. government employees are not covered by the above-mentioned copyright.

Printed in China

Library of Congress Cataloging-in-Publication Data

Charles, Steve.
Vitreous microsurgery / Steve Charles, Jorge Calzada, Byron Wood. —5th ed.
p. ; cm.
Includes bibliographical references and index.
Summary: “Written by a pioneering leader in the development of vitreoretinal surgical techniques and instruments, Vitreous Microsur-
gery is a comprehensive how-to guide to all vitreoretinal procedures. This thoroughly updated Fifth Edition describes many new techniques
and refinements of established procedures and includes new chapters and new illustrations. More than 170 three-dimensional full-color
illustrations—many by the Charles Retina Institute’s resident medical artist, Byron Wood—enable surgeons to clearly visualize the tech-
niques. The focus of the text is on the decision making process a surgeon goes through in evaluating the best course of treatment for his/her
patient undergoing vitreous surgery. The book describes in detail clinically proven methods of managing the anterior and posterior segment
vitreous surgery patient in a systematic manner. The text is organized in a building block approach with general methodology preceding its
application to specific disease states. The book stresses algorithms for intra-operative decision making, relying on knowledge of physical
principles and performed in the order of ascending risk. A companion website included with purchase offers the fully searchable text and
an online image bank”—Provided by publisher.
ISBN 978-1-60831-503-1 (hardback)
1. Vitreous body—Surgery. 2. Microsurgery. I. Calzada, Jorge. II. Wood, Byron. III. Title.
[DNLM: 1. Vitreous Body—surgery. 2. Eye Diseases—surgery. 3. Microsurgery—methods. WW 250 C477v 2011]
RE501.C48 2011
617.7’46059—dc22
2010018076

Care has been taken to confirm the accuracy of the information presented and to describe generally accepted practices. However, the
authors, editors, and publisher are not responsible for errors or omissions or for any consequences from application of the information in
this book and make no warranty, expressed or implied, with respect to the currency, completeness, or accuracy of the contents of the pub-
lication. Application of the information in a particular situation remains the professional responsibility of the practitioner.
The authors, editors, and publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accor-
dance with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government
regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert
for each drug for any change in indications and dosage and for added warnings and precautions. This is particularly important when the
recommended agent is a new or infrequently employed drug.
Some drugs and medical devices presented in the publication have Food and Drug Administration (FDA) clearance for limited use in
restricted research settings. It is the responsibility of the health care provider to ascertain the FDA status of each drug or device planned for
use in their clinical practice.

To purchase additional copies of this book, call our customer service department at (800) 638-3030 or fax orders to (301) 223-2320. Inter-
national customers should call (301) 223-2300.

Visit Lippincott Williams & Wilkins on the Internet: at LWW.com. Lippincott Williams & Wilkins customer service representatives are
available from 8:30 am to 6 pm, EST.
10 9 8 7 6 5 4 3 2 1
The Fifth Edition is dedicated to the quest to
better serve humankind.

In keeping with the approach taken with
previous editions, the fifth edition is designed
as a technology book rather than an academic
analysis of the literature, a statistical review of
outcomes, or a series of clinical cases and images. This edi-
tion is intended to be fine grained with emphasis on every
nuance of surgical technique and tool technology.
Surgical and medical retina strategies have under-
gone significant paradigm shifts since the fourth edition
was written. The fifth edition includes new chapters on
the Constellation vitrectomy system, uveitis, retinal com-
plications of permanent keratoprosthesis, and anti-VEGF
therapy.
Many techniques and parameters unique to 25-gauge
sutureless vitrectomy are discussed in detail. The authors
use 25-gauge sutureless technique for all cases, and micro-
incisional 23-, 25-, or 27-gauge vitrectomy is growing in
popularity on a worldwide basis.
The section on anesthesia for vitreoretinal surgery
has been completely rewritten from an anesthesiologist’s
Preface
perspective in collaboration with Gary Fanning, MD, and
Jay Mattingly, MD, leading experts on this subject.
Kenneth Fung, MD, MBA made a substantial contribu-
tion to the chapter on retinal vein occlusions.
The retinopathy of prematurity chapter has been com-
pletely rewritten to address new examination, laser guide-
lines, anti-VEGF therapy, and changing, more conservative
indications for surgery.
All chapters have been updated, and new references
have been included. All illustrations have been updated; the
majority of illustrations have been updated to a 25-gauge
approach instead of a 20-gauge approach, and many new
illustrations have been included.
It is planned that this text will go forward with an update
approximately every 5 years with my able coauthors, associates,
and great friends Byron Wood, MS and Jorge Calzada, MD.
Steve Charles, MD
vii

This book is made possible because Lippincott
Williams & Wilkins invited us to write a fifth
edition based, in part, on extensive sales of the
fourth edition as well as Chinese, Portuguese,
Turkish, and planned Spanish editions. I would like to thank
the many surgeons, libraries, and universities worldwide
that purchased the previous edition. We are very fortunate
to have worked with Jonathan Pine and Emily Moyer of
Lippincott Williams & Wilkins to produce this larger and
improved textbook.
I would like to especially thank my incredible coau-
thors, Byron Wood, MS and Jorge Calzada, MD. Byron Wood
has created the illustrations, as he has for four previous edi-
tions; these are even more detailed, anatomically and surgi-
cally accurate, and beautiful to view as art. Jordi Calzada has
done a superb job of restructuring all references as well as
adding new ones. In addition, many concepts were refined
viii
Acknowledgm ents
because of his input. It is a great pleasure to work with great
friends who are dedicated to producing the highest quality
work.
I would also like to thank my wonderful, loyal, and
hardworking office and OR staff who provided clinical sup-
port, allowing me to perform over 700 vitrectomies per year
for the last 30 years, which ultimately drove the concepts
described in this book.
I would like to thank Alcon Laboratories for working
with me over the years to develop most of the wonderful
tools and systems described in this text. The engineers, man-
agers, and craftsmen have enabled surgeons to provide the
best possible care to the patients.
I would like to thank my three amazing daughters—
Kelli Ross, MD, Kerri Charles, MD, JD, and Marci Charles—
for their love and their patience with their often preoccupied
father.
 Contents

Preface vii Chapter 8

Acknowledgm ents viii Combined Phacoemulsification Vitrectomy
Procedures 98
SECTION I Chapter 9
Pre surg ical Evaluat io n an d 25-Gauge Vitrectomy 103
Im ag in g 1 Chapter 10
Chapter 1 20-Gauge Sutured Wound Construction
Presurgical Evaluation 2 and Closure 112

SECTION II SECTION III

Surg ical Te ch n o lo g y an d In -Offi ce Pro ce d ure s 117

Te ch n iq ue s 17 Chapter 11
Laser Photocoagulation 118
Chapter 2
Surgical Systems, Tools, and Operating Chapter 12
Room 18 Pneumatic Retinopexy and Office
Fluid-Gas Exchange 121
Chapter 3
Constellation Architecture 32
SECTION IV
Chapter 4
Visualization and Illumination 37 Dise ase -Sp e cifi c
Man ag e m e n t 125
Chapter 5
Anesthesia and Operating Room Chapter 13
Efficiency 40 Prophylaxis of Retinal Breaks and Scleral
Buckling 126
Chapter 6
General Posterior Segment Techniques 45 Chapter 14
Vitrectomy for Retinal Detachment 131
Chapter 7
Vitrectomy Techniques and Technology Chapter 15
for Anterior Segment Problems 76 Giant Breaks 138

ix
x Contents

Chapter 16 Chapter 26
Proliferative Vitreoretinopathy 144 Retinopathy of Prematurity and Pediatric
25-Gauge Vitrectomy 225
Chapter 17
Epimacular Membranes and Vitreomacular Chapter 27
Traction Syndrome 157 Pediatric Traction Retinal Detachments 231

Chapter 18 Chapter 28
Macular Holes 164 Management of Vitreoretinal
Complications Associated with
Chapter 19 Keratoprosthesis 236
Diabetic Retinopathy 171
Chapter 29
Chapter 20 Inadvertent Penetration of the Eye 239
Venous Occlusive Diseases 189
Chapter 30
Chapter 21 Management of Suprachoroidal
Treatment of Choroidal Neovascular Hemorrhage 241
Membranes 193
Chapter 31
Chapter 22 Complications of Vitreoretinal Surgery 244
Intravitreal Injection Technique 204
Chapter 32
Chapter 23 Surgical Self-Education 249
Trauma 205
Chapter 24 Index 253
Endophthalmitis 216

Chapter 25
Surgical Management of the Uveitis
Patient 221
 SEC TIO N I
Pre surg ical Evaluat io n an d Im ag in g

1
C H AP TER
1
PRESURGICAL EVALUATION
Vitreous surgery has been applied successfully
to a wide spectrum of ocular diseases. The com-
plex set of biologic, systemic risk, technique,
and technology considerations confronting the
surgeon creates a difficult decision-making environment.
Rather than emphasizing a search for good prognosis or easy
cases, the goal must be to avoid bilateral or irreversible visual
disability. Many potential vitreous surgery patients have pro-
found socioeconomic problems as a result of blindness and/
or systemic disease. A significant attempt to understand the
patient’s medical and socioeconomic situation is essential to
effective and compassionate health care delivery.
Since the introduction of the first edition of this book,
the scope of vitreous microsurgery has expanded greatly.
Whereas vitrectomy surgery was initially considered only
on the most advanced and complicated clinical situations,
we can safely perform 25-gauge vitrectomy with minimal
discomfort to the patient and, most importantly, with low
surgical complication rates. This allows the use of vitrectomy
for epimacular membranes (EMMs) with relatively good
preoperative visual acuity, relieving patients’ symptoms and
improving functional outcomes in ways that would not have
been possible in decades past.
The surgeon must examine both eyes personally to
reach the best clinical decision. The surgeon must develop
an open, direct, and warm relationship with the patient
before surgery can be considered. It is probably better to dis-
cuss specific statistical results of surgery on similar patients
rather than to discuss the details of a specific methodology.
Rather than using qualitative words such as “excellent” or
“poor,” using a percentage success rate is a more meaning-
ful way to communicate. The frequency of the most impor-
tant complications should be mentioned as well as the range
and mean of postoperative vision in functional and numeric
terms. All available family members and the surgeon’s sec-
retary, technician, assistant, or nurse should witness this
conversation. A complex, legalistic permission document is
far less important than meaningful, compassionate, under-
standable, personal communication with the patient. The
patient should be informed about postoperative positioning,
pain, medications, operating time, type of anesthesia to be
2
used, the requirement for postoperative examinations, and
limitations on activity.
The surgeon must take personal responsibility for obtain-
ing financial support for the economically disadvantaged
patient and act as a consumer advocate concerning facility
fees. It is unconscionable to turn away a patient or permit the
facility to do so because of the patient’s financial status.
It is the surgeon’s responsibility to organize the schedul-
ing process so that the preoperative assessment or logistical/
bureaucratic delay does not affect the prognosis. Waiting
lists are highly inappropriate for patients with significant
visual problems. If surgery is performed within a few days of
the original examination, the surgeon will be more acutely
aware of significant aspects of the patient’s history, findings,
and imaging during surgery.
SYSTEMIC FACTORS
Ag e
Age is a consideration at both extremes of the timeline. While
age cannot be used as an arbitrary criterion, it is clear that
both the neonate with severe pulmonary problems and the
elderly patient with multisystem systemic disease have higher
anesthesia risks. Surgery on the premature infant can be
delayed until pulmonary function and body weight increase,
allowing safer anesthesia. Most vitreoretinal surgery today is
performed under local anesthesia with anesthesia monitoring,
but children and certain adults require general anesthesia.
Sensory deprivation from poor vision may contribute to
disorientation and agitation in certain elderly patients. Uni-
lateral visual loss is well tolerated in many elderly patients
because of less demanding occupational and recreational
needs. While it is the surgeon’s obligation to determine
ocular status and needs, medical consultants and anesthesia
team can help determine and manage systemic risk.
Diab e t e s
As diabetes is common in vitrectomy patients, the vitreous
surgeon must be quite familiar with the complications of this

multisystem disease. An individualized approach is superior
to arbitrary surgical criteria. The presence of a systemic infec-
tion is a contraindication to vitrectomy. All ocular conditions
except endophthalmitis and severe glaucoma can wait until a
systemic infection has been treated.
End-stage renal disease decreases life span and creates
difficult socioeconomic as well as medical problems. The
stable transplant patient can undergo successful vitrectomy,
although immunosuppressant drugs create increased infection
risk, especially with general anesthesia. The dialysis patient
has increased risk with general anesthesia risk and may have
increased bleeding with vitrectomy. While surgical conser-
vatism is indicated for patients with end-stage renal disease,
many carefully selected patients can benefit from vitrectomy.
Card io p ulm o n ary Dise ase
Specific guidelines in the area of cardiopulmonary disease
again must be replaced by an individualized approach with
reliance upon medical and anesthesia consultations. Although
local anesthesia is safer than general, there remains some
risk associated with local anesthesia, pointing to the need
for careful medical assessment and a monitored approach by
an anesthesiologist or a certified registered nurse anesthetist
(CRNA) with anesthesiology supervision. Pulse oximetry,
EKG, continuous blood pressure monitoring, and frequent
blood glucose assessments are mandatory.
The recent trend of vitrectomy surgery moving from
the hospital setting, where subspecialized consultants are
readily available, to the ambulatory surgery center requires
the surgeon to be aware of the anesthesia risk of each
patient. We feel that high-risk patients require hospital-
based surgeries.
Ob e sit y an d Ob st ruct ive Sle e p Ap n e a
Syn d ro m e
A major practical problem that has become more prevalent
in the recent years is morbid obesity and associated obstruc-
tive sleep apnea. These patients require surgical tables that
can withstand extreme weights. In addition, surgical bed
designs that have an off-centered pedestal to allow room for
the surgeon’s legs can tip over with morbidly obese patients.
These patients may not be able to lie flat during surgery due
to orthopnea, requiring either delaying elective surgery until
the patient loses enough weight to safely undergo the pro-
cedure or operating on patients on uncomfortable angles for
the surgeon if the bed is positioned with the head elevated
enough for adequate ventilation.
Pickwickian, morbidly obese, patients can become
hypoxic postoperatively. If an inexperienced certified nurse
assistant increases the concentration of inspiratory oxygen
to improve orthopnea, the oxygen-driven respiratory drive
(since these patients have chronic hypercapnia that desensi-
tizes the CO2 chemoreceptors) can be suppressed and lead
to immediate postoperative hypoxia.
CHAPTER 1 ■ Presurgical Evaluation 3
OCULAR FACTORS
Examination of the vitrectomy patient embodies the same
steps, but with different emphasis, that required for other
ophthalmic specialties. A problem-oriented approach with
emphasis on specific factors that have a high impact on deci-
sion making is more effective than a nonstructured, “routine”
examination.
Visual Fun ct io n Te st in g
Visual function testing can be the most difficult component
of previtrectomy evaluation unless approached properly.
Careful, precise measurement of visual acuity is more rel-
evant than an array of laboratory tests.
Visual Acuit y
The inherent subjectivity of visual acuity testing demands
random presentation of stimuli, multiple repetitions, and
sophisticated examiners. Total coverage of the other eye with
the patient’s palm, not fingers, pressed against the bridge of
the nose is essential to determine light perception. Specifi-
cally, the examiner cannot ask, “Can you see the light?” but
rather should say, “Tell me when the light goes on and off.”
The patient must describe the direction of randomly pre-
sented motion when testing for “hand motion.” Multiple
repetitions increase reliability because of problems stem-
ming from patient guessing. The patient’s neurological/psy-
chological status and educational level may interfere with
accurate testing.
The light perception determination is of particular sig-
nificance because patients without light perception should
never have vitrectomy unless it is clearly understood that the
procedure is to prevent the spread of infection or, in rare
instances, for cosmetic preservation of the globe. Dense
opaque media requires a bright testing light such as that
from an indirect ophthalmoscope at 7.5 volts.
The finger-counting method is fraught with error
because of the great variance in finger size, position, color,
and illumination. It is preferable to use a printed, 20/200-
sized “E” on a white card. This can be presented in random
orientation with multiple repetitions, and the distance can
be measured to determine acuity as a fraction (e.g., 1/200).
A patient with 1/200 vision typically cannot see well enough
to walk, while a 5/200 patient typically is ambulatory, and an
individual with 9/200 vision can read large print with mag-
nification. Placing all these patients into the finger-counting
category is misleading.
Flat screen display, computer-driven vision testing
devices are accurate and rapid ways of testing visions of
20/400 or better and can replace more traditional projector
charts (1). The printed charts with special illumination used
by the Early Treatment of Diabetic Retinopathy Study (2) are
very accurate but cumbersome.
4 SECTION I ■ Presurg ical Evaluation and Im aging
Current refraction must be in place for visual acuity
testing at the 20/400 or better level. Pinholes are difficult for
retinal patients to use and may reduce contrast sensitivity
in the presence of retinal disease, resulting in underestimat-
ing visual function. Near vision testing is essential because
of hereditary myopia and induced myopia from nuclear
sclerosis and encircling buckles.
Co n t rast Se n sit ivit y
The ability to detect changes in contrast is a significant
measure of macular function. Not infrequently a patient’s
perception of visual changes and visual recovery is more
dependent on contrast sensitivity than on visual acuity, since
most real world scenarios involve reduced contrast. We do
not routinely perform formal contrast sensitivity testing on
our patients. On the other hand, an adequate history from
the clinician can gather information regarding visual func-
tion in reduced light that can help orient surgical decision
making in patients with macular disease, even when formal
contrast sensitivity testing is not performed.
Me t am o rp h o p sia
Amsler grid evaluation can help explain a patient’s visual
complaints and can assist in the recommendation of sur-
gery in patients with EMMs and macular striae. No objective
evaluation has been developed for metamorphopsia analysis,
but adequate history and proper Amsler grid testing can
demonstrate the presence of visual distortion that may be
visually disabling and that can be improved with surgery.
We do not recommend using only visual acuity testing for
determination of need for macular surgery, since many high-
functioning patients with relatively good visual acuity can
have disabling, correctable metamorphopsias.
Pro je ct io n an d Fie ld Te st in g
Marked opaque media creates light scatter and might cause
a false impression of being able to perceive the direction of
projection, while recent retinal detachments can have nor-
mal projection. Occasionally, an advanced glaucoma patient
with opaque media will respond only in the temporal field,
but frequently, the patient indicates a larger field due to light
scattering.
En t o p t ic Ph e n o m e n o n
The patient should be able to see the shadows of the reti-
nal vessels if a focal light source (transilluminator) is rubbed
against the closed lids. Care must be taken because these
lights generate heat. If the light source is turned on after
skin contact is made and the skin contact is never broken,
heat dissipation is usually adequate. Many patients report
the shadows as appearing like leaves or cracks, and there
appears to be excellent correlation with attachment of the
posterior 30 degrees of retina. However, many patients with
normal retinas cannot respond in a positive manner to this
test, greatly reducing its value.
Co lo r Discrim in at io n
Patients with recent total retinal detachments and many
long-term, low detachments have excellent color discrimi-
nation with large targets. Because 90% of the cones are extra-
macular, patients with large macular scars have normal color
function with large targets. Standardized color testing meth-
ods are highly dependent on 20/200 or better visual acuity,
negating their value in opaque media cases. Hemorrhagic
vitreous acts as a red filter, altering color discrimination.
Mad d o x Ro d
Detection of the orientation of a Maddox rod is said to cor-
relate with good postoperative function in opaque media
cases. Patients with recent total detachments and advanced
glaucoma can provide accurate answers to this test, thereby
eliminating its value. Certain inoperable patients respond
positively to this test, further decreasing its value.
Two -Po in t Discrim in at io n
Two-point sources of light (transilluminators) can be
moved laterally, initially touching and then progressing to
1-m distance. Reporting of “two lights” at close distance
(< 2.5 cm) corresponds with better retinal function. Patients
with recent or partial retinal detachments can respond nor-
mally to this test, and light scatter can cause false negatives,
dramatically reducing the value of this test.
Lase r In t e rfe ro m e t ry
Dense cataracts and vitreous hemorrhage prevent visualiza-
tion of laser interference fringes. Although laser interferom-
etry can be used with minimally opaque media, it measures
vernier-type acuity, with only moderate correlation with
Snellen acuity.
VALUE OF PSYCHOPHYSICAL TESTING
In general, patients with a positive response to all of the above
psychophysical tests show better postoperative results and,
conversely, those with a negative response typically achieve
poor visual function (3). The subjective nature of these tests
and the many instances of false positives and false negatives
greatly decrease their value. Examiners may use these meth-
ods to seek out good prognosis cases, some of which do not
really require surgery, but a negative response to these tests
frequently indicates an operable retinal detachment demand-
ing immediate vitrectomy. At present, the authors use only
visual acuity testing to assess function.

PUPIL TESTING
An afferent pupillary defect discovered with the swinging
flashlight test can be of great value in preoperative evaluation.
Glaucoma, retinal detachment, ischemic optic neuropathy, and
optic nerve trauma are the most frequent causes of an abnor-
mal response. If one afferent system (retina and optic nerve)
and one efferent system (third nerve, ciliary ganglion, and
iris sphincter) are intact, the test can have value. Bilateral iris
disease eliminates the utility of this test. As in psychophysical
methods, a negative response may indicate an operable retinal
detachment rather than an inoperable eye. Large macular dis-
ciform scars can cause a moderate afferent pupillary defect.
SLIT LAMP EXAMINATION
A problem-oriented examination is far more effective than
a “routine” examination. Specific, subtle details can have a
major impact on presurgical decision making.
Co rn e a
Epithelial erosions are common in diabetics and some
patients who have had recent surgery. Corneal edema may
preclude the excellent visualization required for vitrectomy.
Unless vitreous-corneal touch, lens-corneal touch, or glau-
coma to be treated by vitrectomy is present, the net effect of
vitrectomy will be to further decrease endothelial function.
Guttata and low endothelial cell counts indicate the possibil-
ity of postoperative corneal edema and suggest more rigid
vitrectomy indications. Blood staining, scarring, or edema, if
very extensive, is an indication for penetrating keratoplasty,
Descemets stripping endothelial keratoplasty (DSEK), or
endoscopic surgery. Vitrectomies can often be performed by
working around central corneal or lenticular opacities.
Iris-An t e rio r Ch am b e r
The extent of pupillary dilation is important if the lens is to
be retained. On occasion, a clear lens lensectomy will be nec-
essary in an emergency vitrectomy only to allow the removal
of anterior chamber or perilenticular blood or fibrin.
Iris and trabecular neovascularization is subtle in its
early stages but of extreme importance. The iris surface over-
lying the sphincter should be examined using a magnifica-
tion of 24× to 40× prior to dilation. Large vessels, ectropion
uveae, and peripheral anterior synechia are late changes and
are not required for a diagnosis of iris neovascularization.
Capillaries on the iris or trabecular surface indicate activity,
while larger vessels will persist after regression of the capil-
lary activity. Examiners with red-green color discrimination
problems find it difficult to recognize iris neovascularization
and make this condition a negative factor in the decision to
become a vitreoretinal surgeon. Dilated stromal vessels or
CHAPTER 1 ■ Presurgical Evaluation 5
exposed iris vessels due to prior cataract surgery or certain
lens implants are usually quite easy to differentiate from iris
neovascularization. Topical phenylephrine decreases the vis-
ibility of iris neovascularization due to vasoconstriction.
While significant trabecular meshwork neovasculariza-
tion can be present with minimal iris neovascularization, this
is somewhat unusual. Because gonioscopy is time-consuming
and may damage the corneal epithelium, the iris is usually
used as an indicator of anterior segment neovascularization,
but gonioscopy detection of trabecular meshwork neovascu-
larization is the most sensitive and relevant indicator of vas-
cular endothelial growth factor (VEGF) levels in the anterior
chamber. Cells and flare are present in the anterior chamber
of many patients requiring vitrectomy and are at most a rela-
tive contraindication. Fibrin syndrome usually contraindi-
cates all but emergency vitrectomies for endophthalmitis.
Le n s Op acit ie s
Blood or pigment on the anterior or posterior surface of the
lens or lens implant may necessitate lens or implant removal
for adequate visualization. Many vitreous hemorrhage
cases have posterior subcapsular cataracts. The lens should
be removed if significant posterior subcapsular changes
are present to provide excellent surgical visualization and
because these changes virtually always rapidly progress post-
operatively. Cortical changes, if moderate, can be tolerated.
Moderate nuclear sclerosis will permit adequate visualiza-
tion and progresses more slowly than posterior subcapsular
opacities following vitrectomy.
In t rao cular Pre ssure
Indentation (Schiotz) tonometry is generally inadequate
and outdated for the measurement of intraocular pressure
(IOP). Applanation tonometry is adequate but is inaccurate
with irregular corneas. The Tonopen is more accurate with
irregular corneas and ocular rigidity problems, causes less
corneal trauma, and does not result in fluorescein in the tear
film and anterior chamber. Tonopen tonometry is easier than
applanation tonometry for patients with marked blepharos-
pasm and can be used with a contact lens in place.
Low pressure (hypotony) has an undeserved bad repu-
tation. Most low pressures are due to reparable wound leaks
or reparable retinal detachments with increased uveal-scleral
outflow. Low pressure does not cause phthisis; rather, phthisis
may cause hypotony. Phthisis might best be defined as ocular
collapse from a cyclitic membrane. In no instance should
hypotony delay an otherwise indicated vitrectomy. Ciliary
body shutdown is an overrated, theoretical cause of hypot-
ony, which occurs only after massive destruction of the ciliary
epithelium by cryopexy, ultrasound, laser, or infection.
Elevated IOP has many causes, and a complete dis-
cussion is beyond the scope of this book. Pupillary block,
hemolytic (erythroclastic), and phacolytic glaucoma are
treated by vitrectomy. Neovascular glaucoma (NVG) can
be made worse by vitrectomy unless anti-VEGF agents,
6 SECTION I ■ Presurg ical Evaluation and Im aging
panretinal photocoagulation (PRP) or retinal reattachment
can be accomplished during the procedure, thereby decreas-
ing VEGF levels.
SLIT LAMP BIOMICROSCOPY OF
THE RETINA
An understanding of vitreoretinal anatomy and the forces
acting on the retina is essential in the evaluation of potential
vitrectomy cases. In large part, vitreous drawing is a use-
less exercise; the traction forces are evident from retinal con-
tours and visibility of the vitreous is far better at the time
of surgery. Small retinal breaks, macular edema, macular
holes, EMMs, subretinal neovascular membranes, and subtle
neovascularization are best observed with slit lamp biomi-
croscopy. Cloudy media decreases the value of slit lamp bio-
microscopy. Plano-concave (flat, Goldman) contact lenses
with antireflective coatings provide much better lateral
resolution than 78- to 90-diopter (D) lenses and eliminate
corneal asphericity. In addition, flat or three-mirror contact
lenses provide better axial resolution, better depth of field,
and improved stereopsis than 78- to 90-D lenses and wide-
field contacts, although the noncontact lenses are easier to
use and somewhat more comfortable for the patient.
INDIRECT OPHTHALMOSCOPY
An understanding of the vitreoretinal anatomy and the forces
acting on the retina is a necessity in clinical decision mak-
ing. The examiner cannot be a passive observer of vitreous
“bands” but must grasp the concepts of the continuum of the
anterior and posterior vitreous cortex (PVC) and the forces
causing retinal contour changes.
The indirect ophthalmoscope is essential to vitreoreti-
nal examination because of the stereopsis, wide field, and
contrast obtained even with nonoptimal media. It should be
used on its brightest setting (7.5 volts) in most instances. The
20-D lenses have greater light-gathering ability and increased
contrast compared to the 28-D lenses, which are better for
small pupils and gas bubbles. If possible, the pupil should be
widely dilated, the room dark, and the examiner well dark-
adapted. Continued dark adaptation, by keeping the entire
examining unit dimly illuminated, is far more efficient than
waiting for dark adaptation to occur during each examina-
tion. Noncognitive repositioning of the hand holding the con-
densing lens allows prism effect to enable peripheral viewing,
reduced astigmatism, and viewing around opacities.
VITREOUS CONFIGURATION
The examiner must make a specific effort to look at rather
than through the vitreous with the indirect ophthalmoscope.
Much can be learned from the geometric configuration of
opaque or semiopaque vitreous. A common tendency is for
the clinician to give up and describe the opaque media case
as “no view,” “no retinal detail,” or “red reflex.”
Vitreous consists of a three-dimensional (3D) matrix of
collagen fibers suspended in hyaluronic acid. The normal
shape of the vitreous is roughly spherical with a small ante-
rior concavity for the lens. The continuity of the multilay-
ered outer surface (cortex) is the most essential concept in
developing an understanding of surgical anatomy. The vitre-
ous base is not a space-occupying structure but a zone of
adherence between retina and vitreous. The vitreous base
represents a zone connecting the anterior vitreous cortex
(AVC) and the PVC.
Posterior vitreous detachment (PVD or posterior
vitreous separation) occurs in approximately 70% of the
population and can be thought of as a normal process
although it is a factor in many vitreoretinal problems. Cata-
ract surgery and other anterior segment procedures such as
refractive lens exchange, ocular trauma, vitreous, myopia,
aging, and hereditary processes increase the likelihood of
PVD. Syneresis is traditionally described as the development
of cavities within the vitreous and vitreous “collapse,” which
then causes separation from the internal limiting membrane
(ILM) of the retina. A better way to think of PVD is that vit-
reous collagen becomes less adherent to certain parts of the
retina, central vitreous collagen fibers cross-link and clump,
and saccadic motion facilitates a separation from the retina.
Cavities do not passively form in the vitreous. Collagen
clumping creates this appearance. Collapse exerts minimal
force because the vitreous collagen is nearly neutrally buoy-
ant in hyaluronan and the high water content of vitreous.
Degenerative changes (syneresis), hemorrhage, inflam-
mation, thermal effects, and cellular migration and prolif-
eration can cause the shortening of vitreous surfaces, which
is best termed “hypocellular vitreous contraction.” Dense
hemorrhagic vitreous has been described as “organized” by
many, while it should more correctly be termed “opacified.”
“Organization” is a specific term applied to the proliferation
of fibroblasts. It is usually observed only with severe ocular
trauma or retinal necrosis.
Clear vitreous is fully capable of creating sufficient trac-
tion force to cause a total traction retinal detachment (TRD).
Many examiners refer to “vitreous bands” in their analysis of
vitreoretinal pathology. Discrete vitreous bands are markedly
overemphasized and typically represent only a more opaque
section of the continuous PVC. The clear contiguous por-
tions of PVC often create a tremendous force on the retina,
demonstrating that “band”-oriented vitreous description or
surgical approach is useless.
As interaction between a few cells and the vitreous col-
lagen fibers causes contraction, the PVC pulls away from the
posterior retinal surface but remains adherent to the vitreous
base. The vitreous is normally more adherent to the optic
nerve head, macula, and retinal vessels. Abnormal adher-
ence is common in areas of neovascularization, chorioretinal
scarring, trauma sites, and prior photocoagulation. Further

vitreous contraction will create a conical configuration of
the PVC because of the posterior adherence areas. Epiretinal
membranes (ERMs) may occur at the attachment sites but
also occur independently of the vitreous in the form of pro-
liferative vitreoretinopathy (PVR) star folds, EMMs, or the
ridge or sheetlike ERMs seen in proliferative diabetic retin-
opathy and retinopathy of prematurity (ROP). Frequently,
the vitreous will separate from the retinal surface in some
areas, leaving scattered zones of adherence connected by
the PVC. A large fenestration in the PVC may occur at the
macula. Posterior vitreous separation is often accompanied
by a Weiss ring, which is an elliptical condensation of PVC
formerly adherent to the optic nerve head. Posterior vitreous
separation is somewhat illusory in that an additional layer or
layers of vitreous often remain adherent to the retina in the
presence of an apparent PVD.
Saccadic motion of the eye and a change in head posi-
tion can help differentiate between flaccid PVC with retinal
adherence and taut PVC causing traction on the retina. Sac-
cadically induced motion can create dynamic traction capa-
ble of creating a retinal break but incapable of causing TRD.
The traction referred to in this book might more correctly be
termed “static traction.”
The conical surface of the PVC in diabetic vitreous
hemorrhages, TRDs, or penetrating trauma may have one
or more apices created by zones of adherence. The most
typical apex of the single point cone is the optic nerve.
Two-point conical PVC surfaces are usually attached to the
optic nerve and a vascular arcade (most frequently super-
otemporal). It is essential to understand that the PVC must
bridge between these vitreoretinal attachment points. Nasal
location of the apex of the conical surface usually precludes
traction detachment of the macula. The apices can be acute
angles or broad truncations of the conical PVC, indicating
small or larger zones of adherence, respectively. If the vitre-
ous is taut, the shape of the PVC apex can be thought of
as a mirror image of the shape of the traction detachment.
Single acute angle apices result in simple conical traction
detachments. Ringlike apices along the arcades and disk
create ringlike traction detachments. A broad area of adher-
ence of the contracted PVC to the retina causes “tabletop”
traction detachments. The continuity of the PVC between
each apex must always be kept in mind. If the retina can be
seen in some areas but not others, the shape of the retina
in the visible areas can be used to extrapolate the contour
as well as attached or detached configuration of adjacent
nonvisible retina.
VITREOUS CLARITY
Some assessment of the age of a vitreous hemorrhage should
be made. Discrete blood clots must go through thrombolysis
and dispersion phases before clearing can occur. Subposte-
rior vitreous detachment (sub-PVD) and preretinal hemor-
rhage clear much more rapidly than hemorrhage within the
CHAPTER 1 ■ Presurgical Evaluation 7
vitreous and should be described and managed accordingly.
The retina can frequently be seen through semiopaque vit-
reous. It is useful to grade vitreous hemorrhage from 1+ to
4+ or clear, semiopaque, and opaque, to permit subsequent
examiners to assess the rate of clearing. Fresh blood is bright
red; the color of older hemorrhages becomes less saturated
red and finally becomes yellow or gray.
RETINAL CONFIGURATION
The transition from retinal to vitreoretinal surgery necessi-
tates a change in examination goals. A compulsive, ritualis-
tic quest for retinal breaks must be replaced in large part by
an approach that emphasizes understanding of the forces
acting on the retina. Retinal breaks eliminate the 0.8 mm Hg
pressure gradient normally existing across the retina, result-
ing in a convex contoured rhegmatogenous retinal detach-
ment. Damage to the retinal pigment epithelium (RPE)
pump mechanism or increased vascular permeability also
results in loss of the transretinal pressure gradient, which
produces a convex configuration of the retina termed exuda-
tive retinal detachment. In contrast, the transretinal pressure
gradient causes the concave shape of the retinal surface in
TRD cases. A concave configuration with the presence of
one or more retinal breaks means the normal transretinal
pressure gradient is still present and proves the presence of
traction. In contrast, convex retinal detachment with no vis-
ible retinal break must have a rhegmatogenous or exudative
component.
Perpendicular or oblique traction on the retina is best
recognized by alterations in retinal contour. Steep changes
in retinal contour result only from PVC adherence and trac-
tion at the apex of the elevated retina. Broader zones of PVC
adherence create truncated cones (mesa-like) of retinal ele-
vation. Ridge-like changes in the retinal surface result from
perpendicular (vitreous sheet) or ERM traction. Star folds
and EMMs result from traction directed inward. Macular
holes result from forces due to ILM elasticity, fenestrated
ERMs, and/or residual vitreous and are directed outward.
Retinal incarceration in a trauma or drainage site results in
retinal foreshortening and must be relieved by retinal resec-
tion so that the retina can conform to the RPE.
RETINAL DETACHMENT
MECHANISM CONFIGURATION
Exudative Convex
Rhegmatogenous Convex
Traction Concave
DIRECTION OF RETINAL SURFACE
TANGENTIAL TRACTION FINDING
Inward (centripetal) Star folds, EMM
Outward (centrifugal) Macular hole
8 SECTION I ■ Presurg ical Evaluation and Im aging
RETINAL BREAKS
The detection of retinal breaks assists in surgical planning
but can be more difficult in the patient with decreased vitre-
ous clarity. In contrast to scleral buckling surgery, the view
is always far better in the operating room than in the office.
The reward for prolonged office examination of the patient
can be examiner fatigue and patient discomfort.
Retinal breaks are frequently located adjacent to abrupt
changes in retinal contour associated with vitreous traction.
If the retina is concave, it is certain that these breaks have
played a minor role in the retinal detachment. As any break
may contribute to postvitrectomy detachments, all breaks
must be recognized at the time of vitrectomy and treated.
Retinal breaks are frequently located adjacent to retinal-
RPE adherence areas or previous retinopexy sites acted
on by vitreous traction. A drawing and description of the
position of all retinal breaks should be made preoperatively
and reviewed moments before or during surgery to avoid
overlooking areas needing retinopexy. Too much emphasis
is placed on the poor prognosis of so-called “combined”
traction and rhegmatogenous retinal detachments associated
with diabetic retinopathy.
RETINAL NEOVASCULARIZATION
Active neovascularization of the disk (NVD) or neovascular-
ization “elsewhere” (NVE) can be thought of as a marker for
the presence of VEGF in the vitreous cavity. Assessment of
neovascularization is important not because of intraopera-
tive bleeding, which is usually easy to control, but because
it correlates with postoperative NVG and fibrovascular pro-
liferation on the AVC and is an absolute indication for initial
or supplemental PRP and anti-VEGF therapy. Large vessels
in an ERM without active capillary budding remain pres-
ent even after PRP and anti-VEGF agents markedly decrease
VEGF levels. These larger vessels are more impressive but
are not as significant as capillary activity.
EPIRETINAL MEMBRANE
Since “fibrosis” in the vitreous cavity is only present in the
context of severe trauma or inflammation, this term should
rarely be used. It should be recognized that an ERM is usu-
ally contiguous with the PVC in areas not having a posterior
vitreous separation in the context of diabetic TRD. ERMs
should be examined for color: a brownish pigment may
indicate RPE etiology, while white coloration points more
toward a glial mechanism although RPE cells dedifferentiate
and lose pigment over time. It is essential to assess the rela-
tionship of the ERM to any retinal elevation. Broad areas of
ERM, when contracted, cause larger areas of retinal elevation
than do very small epicenters of ERM.
The ERMs in PVR are typically less opaque than the
membranes associated with diabetic TRD or ROP. For this
reason, they are frequently overlooked, and such misnomers
as “retinal stiffening” are applied. Retinal stiffening in most
disease processes is from periretinal membrane proliferation,
even if the ERMs or subretinal membranes are transparent
and difficult to see. Every fixed fold or star fold must have
an ERM, subretinal membrane, or incarceration as its cause
unless the problem is developmental. The surgical approach
can only be planned by determination of the location of these
membranes. The exact indications for removal are discussed
in the chapters on appropriate disease states.
FOREIGN BODY LOCALIZATION
Localization of an intraocular foreign body (IOFB) has
become somewhat of an end unto itself in the workup of
trauma cases. Vitreous surgery has radically changed the
approach to IOFB patients. For this reason, there is little
need for extensive localization studies in most instances.
If the fundus cannot be seen, in most cases, vitrectomy will
be performed and accurate visualization of the foreign body
will become possible intraoperatively. The relative inaccu-
racy of preoperative localization of foreign bodies near the
ocular wall can contribute to mismanagement.
The nature of the injury causing the foreign body assists
in localization. Most steel foreign bodies are created by ham-
mering and seldom have the velocity to double penetrate the
globe. For this reason, they may be assumed in most cases
to be intraocular and managed with vitrectomy and forceps
removal. In contrast, most shotgun injuries are double per-
forating because of the high velocity, and even if a lead pel-
let is intraocular, it need not be removed for 10 to 14 days.
Computerized axial tomography is expensive and only mod-
erately accurate, while magnetic resonance imaging (MRI)
is contraindicated because a ferrous foreign body will be
moved by the magnetic field. Foreign body localization tech-
niques not utilizing a contact lens are extremely inaccurate
and can give false information. Foreign body localization
methods using a contact lens have the risks of infection and
prolapse of ocular contents from the pressure of application
of the contact lens. Ultrasonic methods, although better, are
difficult to interpret when the foreign body is near the ocular
wall. This will be discussed further in the section on diag-
nostic ultrasound. Anteroposterior and lateral x-rays of the
orbit are important, however, to determine the presence and
number of IOFBs but are not as useful in localization.
DIAGNOSTIC ULTRASOUND
Diagnostic ultrasound has greatly improved the management
of the opaque media patient. This method, however, is best
utilized in the hands of the potential surgeon rather than in
a diagnostic laboratory setting. The surgeon has the entire

clinical picture in mind at the time of ultrasonic examination,
making possible better integration of the ultrasonic diagno-
sis into the thought process. If the ultrasound equipment
is located in the surgeon’s primary examining room, it will
be utilized much more frequently and can be done without
charge if needed. Technicians should not be used for what
should be thought of as acoustic ophthalmoscopy.
Clin ical Aco ust ic Ph ysics
An understanding of the physical principles involved in the
interaction between ultrasonic energy and biologic materi-
als is essential for accurate ultrasonic diagnoses. Ophthalmic
ultrasonography utilizes pulse-echo reflection ultrasound.
Brief pulses of ultrasonic energy having a 10-MHz, or
greater, center frequency are repeated at a rate of 1 to 5 kHz,
allowing time for the same transducer to receive the reflected
echoes. Knowledge of the average speed of ultrasonic energy
propagation through tissue (~1,540 m/s) permits the real-
time calculation and display of the distance between the
transducer and the echo-producing structure as a dimen-
sion on the two-dimensional (2D) display. Acoustic energy
is reflected and refracted at interfaces between materials of
different acoustic densities (Fig. 1.1).
If the face of the piezoelectric transducer crystal has a
small radius of curvature, the result is a point focus with
poor depth of field. The length of the eye (~25 mm) requires
a weaker focusing to obtain an appropriate depth of field.
This wide beam width (3 mm at the 6-dB point) creates
poor lateral resolution (Fig. 1.2): Targets in the near field are
Fig ure 1 .1 ■ Materials of different acoustic densities reflect and refract acoustic energy at their
interfaces.
CHAPTER 1 ■ Presurgical Evaluation 9
displayed as double, while far-field targets are displayed as
spread out laterally. This is inescapable without using com-
puted sonography, which is currently not available in oph-
thalmic ultrasound systems.
Axial resolution is a function of frequency, with higher
frequencies resulting in better axial resolution. Higher fre-
quencies are absorbed more by biologic structures, mandat-
ing higher powers to obtain the same small echo sensitivity.
The risk of cataract formation determines the maximum
power limit that can safely be utilized. In practice, this com-
promise results in the use of 10- to 20-MHz ultrasound with
about 0.15-mm axial resolution, which is more than an order
of magnitude better than lateral resolution (Fig. 1.3). Axial
resolution is degraded when the wide beam is reflected from
curved surfaces such as the crest of a TRD.
The greatest reflection of ultrasound is obtained when
the beam strikes the interface perpendicularly. Beams reflect-
ing tangentially from the equatorial ocular wall result in little
reflection back to the transducer. With accurate representa-
tion of echo amplitudes, an entire circular eye cross section
cannot be displayed.
Because the speed of sound is greater in denser tissues
such as the lens, structures behind the lens are displayed as
being closer, and refraction occurs at the edge of the lens
(Fig. 1.4). Dense structures such as lens, intraocular lens
(IOL), IOFB, and scleral buckles have multiple internal
reflections and display evenly spaced false echoes of decreas-
ing amplitude behind the structure (Fig. 1.5). The echoes
produce paradoxical movement with transducer reposition-
ing, which aids in their recognition. Dense structures such as
10 SECTION I ■ Presurgical Evaluation and Im aging

Fig ure 1 .2 ■ Poor lateral resolution results from the inherently wide ultrasound beam .

Fig ure 1 .3 ■ 10-MHz ultrasound produces axial resolution ten tim es greater than the lateral
resolution.
 CHAPTER 1 ■ Presurgical Evaluation 11

Fig ure 1 .4 ■ Multiple, evenly spaced false echoes result from internal reflection with an IOL.

Fig u re 1 .5 ■ Structures p osterior to d ense tissue ap p ear closer than their actual location
because of increased speed of sound in denser m aterials.
12 SECTION I ■ Presurgical Evaluation and Im aging
calcified cyclitic membranes, IOLs, and foreign bodies create
significant shadowing behind them because of the absorp-
tion of acoustic energy.
The absorption of ultrasonic energy as it passes twice
through the tissue results in distant echoes being displayed
with relatively less amplitude. Electronically increasing the
gain for distant targets can compensate for this absorption.
This is called time-varied gain.
Electronics that automatically display interfaces such
as cornea, lens capsule, retina, and sclera have resulted in
diagnostic inaccuracies. Increasing the amplitude and clip-
ping the peaks to display interfaces means that all echoes
are displayed as equal amplitude. With this approach, the
vitreous and retina can easily be confused. Similarly, elec-
tronic differentiation to detect interfaces eliminates the lower
amplitude echoes within the lens, vitreous, subretinal fluid,
suprachoroidal space, and tumors.
A-Scan Ve rsus B-Scan
Time amplitude ultrasonography (A-scan) was the origi-
nal method of ultrasonography but has virtually no util-
ity in the opaque media workup. A-scan information is
one-dimensional (1D) and is analogous to a “needle in
the haystack” approach. Extremely experienced A-scan
ultrasonographers can spatially integrate the 1D informa-
tion and obtain some limited value. The typical clinician,
however, has far more difficulty with this approach. Quan-
titative A-scan echography has been overemphasized as
being capable of a tissue diagnosis. A-scan echo amplitude
is highly dependent on the angle at which the ultrasonic
beam strikes the reflecting structure. Oblique angles cause
marked attenuation of the reflected echo. Highly convo-
luted retinal detachments will have areas of high and low
reflections. For this reason, there is a large sampling error
in this 1D approach.
B-Scan
Sector-type B-scan ultrasonography is a 2D approach in
which a knifelike slice or plane of acoustic section is made
through the tissue, as opposed to the 1D needle approach
of A-scan. Echoes are displayed as intensity-modulated pix-
els. As with A-scan ultrasonography, the best reflections are
from those structures roughly perpendicular to the beam.
For this reason, the cornea, anterior and posterior lens, and
retinal and scleral surfaces are the best-displayed structures.
The equatorial sclera and lens are seen less well, except
when the eye is rotated or approached from different angles,
emphasizing the need to move the eye and probe during the
examination.
Th re e -Dim e n sio n al
Slow rotation of a sector scan will produce a conical volume
of information that can be displayed as a conical 3D image
or 3D slices using perspective, shading, motion parallax,
and various other digital graphics technologies. Because the
image data set is acquired from a single vantage point, struc-
tures with surfaces that are more tangential than perpendic-
ular to the scan beam will be invisible or depicted with far
less reflectivity. Current 3D ultrasonic imaging techniques
are of minimal value in vitreoretinal disease and best used
for tumor volume determination.
Re al-Tim e
Real-time is defined as appearing virtually continuous and
instantaneous to the human observer. It is accomplished by
using a scanning rate of 10 to 30 scans per second. Real-
time ultrasound has many advantages in presurgery decision
making. As opposed to a static scan captured with digital
frame store, continuous tomography is possible by moving
the transducer. As the transducer is moved with respect to
the eye, the examiner can observe the screen and the relation-
ship of the probe to the eye and integrate multiple 2D scans
into a 3D mental picture of the intraocular contents utilizing
proprioceptive awareness of the transducer position. Any
attempt to reconstruct this picture from a protocol-driven
approach using individual images taken at discrete positions
is much more difficult and inaccurate.
Real-time ultrasound is of great value in detecting
rigidity or, conversely, mobility of ocular structures. The
PVC, when detached from the retinal surface in a PVD con-
figuration, is highly mobile. Loose vitreous apices attached
at a single point to the retina are quite mobile but become
immobile if they are taut with resultant TRD. Rhegmatog-
enous retinal detachments are somewhat less mobile than
the detached vitreous but have definite undulation motion
during saccadic motion of the eye. Retinal detachments with
PVR have much less mobility than a typical rhegmatogenous
retinal detachment (4). Giant breaks have marked retinal
mobility (5).
With real-time capability, moving patients and even
those with nystagmus can easily be examined. In fact,
it is advantageous to elicit repeated ocular motion in
all directions to decrease the sampling error intrinsic to
2D scanning and to test the rigidity and mobility of ocular
structures.
Preretinal hemorrhage can be seen to flow from one
position to the other, and dislocated lenses, implants, and
foreign bodies can be seen to move with saccadic motion,
changing gaze, or patient repositioning (supine or seated).
Real-time adds the fourth dimension (time) to the 3D image
reconstructed in the examiner’s cerebral cortex. When the
posterior vitreous has detached from the retina, it will lie in
contact with the retina when the patient is in the supine
position. Saccadic motion can then demonstrate with
real-time ultrasound that the apparent adherence is only
gravity-induced contact of the vitreous with the retina. The
saccadic-induced motion of preretinal hemorrhages can help
differentiate them from disciform scars.
 CHAPTER 1 ■ Presurgical Evaluation 13

Gray Scale are measured in ultrasound by the use of an assured

speed of sound, increased tissue density causes more
Gray scale B-scan relates echo amplitude (acoustic reflectivity
distant echoes to be displayed as being closer to the
or absorption) to the brightness of specific pixels on the
transducer.
display. The vector positional data from the sector scan are
5. Displays have only a 20- to 30-dB dynamic range,
displayed as a sector on the screen; very reflective echoes
while ultrasound interpretation of ocular tissue
have white dots on the screen, less reflective echoes have a
produces 60- to 100-dB dynamic range. For this
dimmer shade of white, and absence of echo is displayed as
reason, high gain must be used for minimal vitreous
black. Simply stated, the more sound reflected from the ocu-
hemorrhage or normal vitreous surfaces, while the
lar structure, the more light emitted from the corresponding
gain must be turned down to avoid saturating high-
area of the screen. Earlier ultrasound units had static scan-
amplitude echoes such as IOLs.
ning with bistable storage cathode ray tubes (CRTs), and the
echo amplitude information was therefore not displayed two-
dimensionally. Combined A-scan was utilized to obtain a lim- Ve ct o r Scan n in g
ited, 1D impression of echo amplitude. If a combined A-scan
The process of electronically removing a line from a real-time
approach was used and the A-scan intersected the retina
B-scan and displaying it in the typical time amplitude, A-scan
obliquely, it might be called vitreous. The pattern recognition
manner is called vector scanning or simply simultaneous
capabilities of the examiner, however, can, in effect, join the
A-scan and B-scan. It makes possible the combined A-scan
dotted lines between the highly reflective areas using a gray
approach with its wide dynamic range combined with the
scale B-scan. With low noise and high dynamic range signal
real-time, gray scale, 2D topographic information. It is of
processing, even diffuse red blood cells and normal vitreous
more value in looking at internal tumor architecture than it
can be seen. Long-term retinal detachment causes atrophy,
is in the typical opaque media previtrectomy workup.
decreasing acoustic reflectivity just as cellular proliferation in
the vitreous can increase reflection, making amplitude alone
an insufficient criterion to distinguish retina from vitreous. Im ag e Arch ivin g an d Vie win g
When gray scale is coupled with real-time (6), it permits
Hard copy images can be used to establish to the reimburse-
a further appreciation of retinal structures, such that diagno-
ment or legal community that an ultrasound examination
sis is made in a pattern recognition manner rather than by
was performed. Rarely, prior digital or hard copy images can
itemization of individual diagnostic criteria. The authors con-
be compared with the current on-screen image to determine
sider a real-time, gray scale B-scan to be an absolute necessity
if a change has occurred. The authors again emphasize that
for the workup of the opaque media patient (7,8).
ultrasound is a key component of the clinician’s examina-
tion, not a test, archival, or photo opportunity.
Te ch n ical Lim it at io n s
Diagnostic ultrasound has certain limitations: Exam in at io n Me t h o d s
1. The typical semifocused B-scan transducer creates a Contact scanning on the eyelid is used in the majority of
beam that is 2 to 3 mm in width, and as a result, the opaque media examinations before vitrectomy. A monolithic
target is displayed widened in the lateral orientation. blob of highly viscous acoustic coupling material is applied
A target is imaged before and after the beam center to the probe before application to the upper lid for contact
is lined up with the target during the sector scan. scans. The patient is asked to look multiple times in every
Highly focused beams improve lateral resolution but direction, and the scan is made in the anteroposterior direc-
markedly decrease the depth of focus. tion both through the lens for the classic picture and then
2. Axial resolution is a function of the frequency of the outside the lens for better resolution and to aid in 3D think-
ultrasound. Higher frequency ultrasound, however, ing. It is best that this examination be performed in the stan-
is absorbed more by tissue, thus requiring dangerous dard examining room and that an ultrasound unit is always
power levels to achieve the same sensitivity or imag- available. Putting the unit on a wheeled cart can also increase
ing limited to the anterior segment. Practical 20-MHz utilization. Ultrasound should be used at every visit in which
retinal B-scan technology is now available. the patient has opaque media, because the retinal situation
3. Beam inclination to a reflecting “surface” markedly can change very rapidly from one visit to the next (9,10).
decreases the amplitude of the reflected echo. The
complex topography of vitreoretinal disease pre-
vents quantitative echography because of this beam OPTICAL COHERENCE TOMOGRAPHY
inclination–induced decrease in echo amplitude.
4. Denser structures such as a nuclear sclerotic cataract, Optical coherence tomography (OCT) (11) is an imaging
IOL, or calcified tissue absorb sound but increase the technique that produces high-resolution cross-sectional
speed of sound within the tissue. Because distances images of optical reflectivity. It is based on the principle of
14 SECTION I ■ Presurgical Evaluation and Im aging
Fig ure 1 .6 ■ OCT im ag e of EMM.
low-coherence interferometry where distance information
concerning various ocular structures is extracted from time
delays of reflected signals. OCT systems use light waves
emitted by a superluminescent diode operating at 840 to
870 nm and between 200 microwatts and 1 milliwatt to
construct the images. This is analogous to the application
of sound waves in B-scan ultrasonography. The use of light
waves enables OCT to achieve an optical axial resolution of
7 mm and a digital depth resolution of 3.5 mm using the
Heidelberg Spectralis HRA-OCT spectral domain OCT. The
initial systems were time domain (Zeiss Stratus) and used a
moving mirror in the reference arm which required much
longer acquisition times, resulting in motion blur and less
scan lines. Less scan lines degrade the image quality and pro-
duce more artifacts. Spectral domain OCT systems (Fourier
domain) use a spectrometer analyzed by Fast Fourier Trans-
formation instead of a moving mirror in the reference arm.
OCT images can be presented as either cross-sectional
images or topographic maps. Although autosegmentation
algorithms can produce topographic maps of isolated retina
or RPE (retinal pigment epithelium), this has not been dem-
onstrated to be of clinical value except for following RPE
loss or drusen volume in clinical trials. Cross-sectional or
B-mode imaging is accomplished by acquiring a sequence
of 40,000 A-scans (Heidelberg Spectralis HRA-OCT) across
a section of retina.
Retinal thickness can be assessed longitudinally using
serial OCT images. The presence of cystic spaces can be
detected by the presence of focal areas of very low reflectivity
Fig ure 1 .7 ■ OCT im age of m acular hole.
within thickened neurosensory retina. OCT images that
topographically display retinal thickness can be produced.
Topographic maps and cross-sectional OCT images can be
displayed using a false color scheme thought by some to
facilitate interpretation; however, this approach produces
false interfaces; gray scale images allow better visualization
of detail and facilitate correct interpretation.
OCT has become an invaluable tool in evaluating
EMMs (12) (Fig. 1.6), partial- and full-thickness macular
holes (13) (Fig. 1.7), diffuse and cystoid macular edema,
vitreomacular traction, submacular fluid, macular schisis,
macular cysts, and choroidal neovascular membranes (14).
Spectral domain OCT often demonstrates clinically relevant
pathology not visible with contact lens slit lamp biomicros-
copy or angiography. It has become an essential part of mac-
ular assessment when macular surgery is being considered.
The authors prefer the Heidelberg Spectralis HRA-OCT
to other available systems for many reasons. The Heidelberg
Spectralis HRA-OCT produces 40,000 A-scans per second
and the best OCT resolution available. This device also has
TruTrack, which tracks eye movements and sets the OCT
scan line on the same position on the reference images or
scanning laser ophthalmoscope (SLO). The combined SLO
produces 15 dB better signal-to-noise ratio than fundus
camera–based angiography systems. In addition, the Heidel-
berg Spectralis HRA-OCT produces very high-quality auto-
fluorescence images using tracking and signal averaging (Fig.
1.8). The 3D reconstruction enables overlapping of SLO and
OCT images. OCT image XY location is simultaneously dis-
played on infrared, autofluroscence (AF), or SLO angiogram
images by using a clickable cursor.
We recommend that retina surgeons learn to analyze
spectral domain OCT the way a radiologist reads a brain
MRI: All scans should be seen by the physician to avoid
missing pathology that a single line scan printout may not
show. The best place to examine OCT data is at the com-
puter terminal where all image raw data are stored rather
than from an image database where only a few scans per ses-
sion are stored, based on the opinion of the technician, or,
even less, from a single scan printout.

Fig ure 1 .8 ■ Tracking and im age averaging on the Spectralis
produce high-quality autofluorescence im ages.
BRIGHT FLASH
ELECTRORETINOGRAPHY
Bright flash electroretinography was originally developed in
the context of recording the human early receptor poten-
tial (ERP) (15). It was discovered that bright strobes could
overcome the light-absorbing property of dense vitreous
hemorrhage and allow the elicitation of an electroretinogram
(ERG) in these densely opaque media patients. There are
many problems with this method, however (16,17). Many
ophthalmologists incorrectly think that a retinal detachment
will still have a recordable ERG that can be used to deter-
mine retinal viability. However, because of conduction prob-
lems, a total retinal detachment of any age does not produce
a recordable ERG. Similarly, some examiners believe that
a nonrecordable bright flash ERG means that a vitrectomy
should not be done. In most cases, the remaining part of
the clinical and ultrasound examinations will reveal a total
detachment, which requires immediate vitrectomy. In con-
trast, many surgeons wish to do cases that have a recordable
bright flash ERG. While this is certainly easier for the sur-
geon, if these are unilateral moderate vitreous hemorrhages,
these are just the cases that may not require vitrectomies. Use
of this test, therefore, as a prognostic means must be thought
of in terms of the complete clinical picture.
Although it is true that a decreased B-wave correlates
with ischemic inner retinal layers, this is also not a valu-
able criterion because the small portion of the retina around
the macula can be well perfused and capable of good vision
postvitrectomy. Further, false negatives occur in extremely
dense vitreous hemorrhage cases because even the bright
strobes used cannot penetrate the ocular media sufficiently.
Extensive PRP may cause a nonrecordable ERG similar to
CHAPTER 1 ■ Presurgical Evaluation 15
that in retinitis pigmentosa. Therefore, a patient who had
a vitreous hemorrhage after a successful PRP would have a
nonrecordable bright flash ERG, causing some examiners to
think that surgery was not indicated. If a protocol mentality
allows the bright flash ERG to be utilized in a patient with
clear media just because his or her initial ERG testing was
not recordable, permanent damage to the retina can result
from the bright strobe.
Because of the many false positives and false negatives
described above and the possibility of technical laboratory
error, bright flash ERG has very little value in the preopera-
tive evaluation. Certainly, if the ERG is recordable and other
clinical factors suggest surgery, the patient has a better prog-
nosis. Having a better prognosis, however, in many clinical
situations is not the reason to have a vitrectomy.
VISUAL EVOKED POTENTIAL
The visual evoked potential (VEP) also has limited value
because of difficult interpretation, equipment complexity,
false negatives, and false positives. As with ERG testing, a
patient with severely decreased preoperative VEP might not
be considered for vitrectomy, when in reality, he or she could
have useful vision after vitrectomy if a retinal detachment is
repaired.
References
1. Timberlake GT, Mainster MA, Schepens CL. Automated visual
acuity testing. Am J Ophthalmol 1980;90:369.
2. NAS-NRC Committee on Vision and Recommended Standard
Procedures for the Clinical Measurement and Specification of
Visual Acuity: Report of Working Group 39. Adv Ophthalmol.
3. Michels RG, Ryan SJ. Preoperative evaluation of patients for
vitreous surgery. In Gitter KA, ed. Current concepts of the vitreous
including vitrectomy. St. Louis: C.V. Mosby, 1976:121–128.
4. Han DP, Lewandowski M, Mieler WF. Echographic diagnosis
of anterior hyaloid fibrovascular proliferation. Arch Ophthalmol
1991;109:842–846.
5. Genovesi-Ebert F, Rizzo S, Chiellini S, et al. Echographic study
of the vitreoretinal interface in giant retinal tears. Ophthalmo-
logica 1998;212(Suppl 1):89–90.
6. Charles S, Griffith J, Lindgren W. Unpublished paper, 1974.
7. Capeans C, Santos L, Touriño R, et al. Ocular echography in
the prognosis of vitreous hemorrhage in type II diabetes mel-
litus. Int Ophthalmol 1997–1998;21:269–275.
8. Kumar A, Verma L, Jha SN, et al. Ultrasonic errors in analysis of
vitreous hemorrhage. Indian J Ophthalmol 1990;38:162–163.
9. Jack RL. Ultrasonographic ocular evaluation prior to vitrectomy.
In Irvine AR, O’Malley C, eds. Advances in vitreous surgery.
Springfield: Charles C. Thomas Company, 1976:100–112.
10. Jack RL, Hutton WL, Machemer R. Ultrasonography and
vitrectomy. Am J Ophthalmol 1978;78:265.
11. Huang D, Swanson EA, Lin CP, et al. Optical coherence tomog-
raphy. Science 1991;254:1178–1181.
12. Wilkins JR, Puliafito CA, Hee MR, et al. Characterization of
epiretinal membranes using optical coherence tomography.
Ophthalmology 1996;103:2142–2151.
13. Hee MR, Puliafito CA, Wong C, et al. Optical coherence tomog-
raphy of macular holes. Ophthalmology 1995;102:748–756.
16 SECTION I ■ Presurgical Evaluation and Im aging
14. Michels S, Rosenfeld PJ, Puliafito CA, et al. Systemic bevaci-
zumab (Avastin) therapy for neovascular age-related macular
degeneration twelve-week results of an uncontrolled open-label
clinical study. Ophthalmology 2005;112:1035–1047.
15. Galloway NR. Early receptor potential in the human eye.
Br J Ophthalmol 1967;51:21.
16. Fuller DG, Knighton RW, Machemer R. Bright flash
electroretinography for the evaluation of eye with opaque vit-
reous. Am J Ophthalmol 1975;80:214.
17. Fuller D, Knighton R, Machemer R. Bright flash ERG. In: Irvine
AR, O’Malley C, eds. Advances in vitreous surgery. Springfield:
Charles C. Thomas Company, 1976:97–99.
 SEC TIO N II
Surg ical Te ch n o lo g y an d Te ch n iq ue s

17
C H AP TER
2
SURGICAL SYSTEMS, TOOLS,
AND OPERATING ROOM
High-technology vitreous surgery requires
intensive and continuous attention to the equip-
ment and operating environment. Equipment,
supplies, training, staffing, maintenance, and stor-
age of the equipment are ultimately the surgeon’s responsibil-
ity. The requirements for effective high-consequence surgery
should outweigh logistical considerations. Blaming problems
on equipment, companies, nurses, technicians, administra-
tors, or managed care is irresponsible and ineffective.
Vitreoretinal surgery has undergone many engineer-
ing advances in the four decades since it was first intro-
duced. Most of the important developments resulted from
an intuitive rather than a physics-based design approach. By
contrast, medical school curriculums emphasize a biology-
based approach to diagnosis and therapy. The intent of this
chapter is to provide a physics and engineering framework
for understanding the essential techniques and technologies
used in vitreoretinal surgery today.
There are many similarities between the task domain
of the pilot and of the surgeon. Pilots are required to under-
stand the engineering principles behind the propulsion,
hydraulic, pneumatic, avionics, and electrical systems,
while surgeons often use tools and machines with little or
no understanding of how they work. The physics of flight is
very complex but is relatively well understood by both engi-
neers and pilots. Tissue cutting, turbulence, friction, state
change in materials, and fracture propagation are so complex
that a precise mathematical description of the underlying
physics is impossible. Nevertheless, valid assumptions about
the physical principles involved and estimation of the math-
ematical parameters can and should be made. Vitreoretinal
surgery requires a wider variety of techniques and technolo-
gies than cataract surgery, because it is used to treat a broader
array of disease entities and anatomic configurations using a
greater array of technologies. Knowledge of the engineering
concepts involved in vitreoretinal surgery can facilitate better
selection of tools and techniques as well as infusion, suc-
tion, duty cycle, and cutting rate parameters. Surgeons who
understand vitreoretinal surgical systems can problem solve,
18
troubleshoot, improvise, and perform better and faster in the
operating room (OR). A surgeon should be able to set up all
machines, set all parameters and modes, and connect, prime,
and test all tools without the help of a nurse or technician.
This knowledge reduces dependency on OR personnel and
resultant anxiety if key people are not available.
OPERATING ROOM ENVIRONMENT
Sch e d ulin g
An attempt should be made to schedule vitrectomies when
the surgeon and team are not too rushed or heavily bur-
dened mentally. If this means early, late, day-off, or week-
end surgery, it must be done. Some surgeons become tense if
they are heavily committed for time following their surgery
time. As some vitrectomies can be lengthy, especially during
the learning phase, time must be available to facilitate con-
centration on the surgery rather than a subsequent event.
A designated OR is preferable to ad hoc room utilization as it
facilitates equipment accessibility, maintenance, and stock-
ing and storage of disposables.
In st rum e n t St o rag e
The approach in which instruments are assembled from
various sources just prior to the case is inadequate for high-
technology vitreous surgery. It is not optimal to share these
specialized instruments with other surgeons and proce-
dures. A tray that is equipped for all basic vitreous, retinal,
and microsurgical methods should be prepared and stored
together. The removal of dense lenses, intraocular lenses
(IOLs), and large intraocular foreign bodies through the
limbus requires that the appropriate instruments be present
on the basic vitrectomy tray. Similarly, scleral buckling tech-
niques may be required and probably should not dictate the
use of a second tray. An identical sterile backup tray should
be available at all times in the OR or adjacent storeroom.
 CHAPTER 2
All fragile instruments such as intraocular scissors,
forceps, contact lenses, and the like must be kept sterile at
all times in transparent peel-packs stored in well-labeled
trays, cabinets, or boxes. These instruments are best stored
in a lockable cabinet in the OR or a movable cabinet that is
taken to the OR before use. They should be stored in spe-
cific places in the cabinet to permit daily inventory control
and maintenance. If not stored in a single cabinet, the instru-
ments tend to become lost and unavailable at a critical time
in vitrectomy.
All disposable materials such as packs, patient drapes,
microscope drapes, sutures, needles, IOLs, gases, and tub-
ing sets should be stored in sufficient numbers in a spe-
cific place near the OR. In this way, faulty or inadvertently
contaminated materials and instruments can be instantly
replaced without having to be ordered from a central supply
area. The many steps in a complex procedure should not
be delayed while waiting for instruments or materials to be
delivered.
A fast para-acetic acid (Steris) and/or gas plasma (Sterad)
as well as a fast cycle steam autoclave should be immedi-
ately available to the OR personnel to reduce case turnover
time as well as resterilization of inadvertently contaminated
equipment if sufficient backups are not available.
Pre surg ical Eq uip m e n t Te st in g
All equipment should be set up and tested prior to the
administration of anesthesia. If needed vitrectomy, lensec-
tomy, phacoemulsification, endophotocoagulation, operat-
ing microscope, or scissors equipment is not functional, the
case must be postponed. Infusion fluid should be run copi-
ously through all tubing to remove any bubbles or particulate
material. Only after all equipment is tested and the surgeon
is present can the team begin local or general anesthesia.
Op e rat in g Ro o m Pe rso n n e l
Preferably, a single surgical technician or nurse should be
responsible along with the surgeon for all instruments as
well as for the ordering, maintenance, and inventory of dis-
posable materials. Preferably, this individual should also
function in the office environment to enhance information
transfer, efficiency, and patient confidence. This same indi-
vidual can assist in the recording and compilation of preop-
erative, intraoperative, and postoperative data for outcome
testing, medical records, and billing. Participation in the
follow-up care provides personal evidence of the impact of
surgical success and failure. This individual should then be
in command of the remainder of the team and should be
responsible for backup personnel in their absence. A friendly,
cooperative atmosphere with a sense of humor is conducive
to the team-play attitude, which is as necessary in the OR as
it is for a successful sports team. This is preferable to a tense,
angry, chauvinistic, find-someone-to-blame attitude unfor-
tunately so prevalent in ORs.
■ Surgical System s, Tools, and Op erating Room 19
Vid e o Re co rd in g
While certain cases may have teaching value, video recording
should not become a time- and concentration-consuming
consideration in the OR. Centering the microscope and
producing the best video opportunity should not consume
costly labor dollars or a surgeon’s mental focus. At best, it
is a moderately efficient way to teach and is often used for
promotional purposes. 3-CCD cameras provide better color
purity and image quality than 1-CCD cameras but require
over three times as much light. Firewire (IEEE 1394) or
DVCAM is currently the optimal connection scheme between
the camera and the digital recording systems. Recording can
be on hard drives or DVDs. Currently, the best image quality
is obtained with AVI files, although MPEG is a reasonable
alternative for DVDs as it requires less storage space. Direct-
to-edit (DTE) systems are far more efficient than systems
requiring capturing the video at a time subsequent to the
surgical procedure. DTE permits the first review of the video
using fast forward rather than normal speeds.
Tissue Cut t in g
An external force placed on an object will cause it to move,
unless the object is constrained. A force placed on a con-
strained object will cause an internal reaction called stress
that resists the external force. Tensile (pulling) and compres-
sive stresses occur normal (perpendicular) to the cross-sec-
tional plane of an object. Guy wires on a suspension bridge
are subject to tensile stress while the road surface is subject
to compressive stress by a traveling car. Forces that act in
the plane as opposed to normal to the plane are called shear
stress and can be identified as transverse or torsional (twist-
ing). A rivet between two sliding metal plates may be subject
to transverse shear. The drive train of a car has a number of
torsional stresses when the throttle is rapidly depressed.
Force placed on a constrained object induces defor-
mation in relation to the amount, duration, and speed of
onset of the external force. The proportion of the amount of
deformation per unit length is known as strain. Tensile and
compressive stresses will cause a lengthening or shortening
of an object. Transverse shear stress causes bending, and tor-
sional shear stress causes a twisting deformation (Fig. 2.1).
Relatively small amounts of stress cause small amounts of
strain. After the stress is removed, the object will return to its
original size and shape in a process known as elastic strain.
Increasing amounts of stress will eventually cause a perma-
nent deformation in the object that persists after the stress
is removed. In material such as steel, the permanent defor-
mation is caused by changes in the crystalline structure of
the metal. Eventually, small cracks appear that interconnect,
leading to rupture and failure of the metal.
Biological tissues, particularly in the eye, display more
complicated characteristics. An externally applied force can
cause deformation, with the possibility of elastic strain. With
continuing force, a “flow” may occur, such as that seen in
20 SECTION II ■ Surgical Technology and Techniques
Fig ure 2 .1 ■ Force p laced on a constrained object ind uces
deform ation in relation to the am ount, duration, and speed of
onset of the external force.
a viscous fluid. This combination of elasticity and viscous
flow is termed viscoelasticity. For example, it is common for
the scleral buckling effect from a sponge exoplant to appear
larger 1 or 2 days later. This is caused by viscoelastic changes
in the sclera.
As noted above, the remarkable complexity of tissue
cutting prevents an exact physics-based understanding and
mathematical representation. The author has found it useful
to define several types of tissue cutting in order to develop
and utilize more effective techniques and devices: elongation,
shear, sharpness, inertial, vaporization, and enzyme-assisted.
Ideal cutting can be defined as separating tissue into two
pieces without producing any remote forces or physical or
chemical effects.
Elongation simply means to apply force to tissue or a
tissue interface until the tissue fails, tears, or breaks. This
is analogous to the methodology used in tensile strength
testing. Membrane peeling is an example of this mode of
action. Epiretinal membranes (ERMs) are thought to be
approximately 100 times stronger than retina, which limits
the application of this method to cases with low adherence
between the retina and the membrane.
Shear can be defined as moving two square-edged cut-
ting edges past each other with the target tissue interposed.
Scissors implement shear type cutting with nonparallel
blades that push the tissue forward as the blades close. This
concept can be referred to as an “exclusive shear.” Scissors
cut at a point that advances toward the tips with blade clo-
sure. Vitreous cutters also work primarily by shear. How-
ever, in contrast to scissors, vitreous cutters cut on a line
rather than a point. Vitreous cutters utilize parallel cutting
edges and, therefore, do not produce the “squeeze out”
force produced by scissors. Vitreous cutters may be termed
“inclusive” shears. Vitreous cutters require a pressure gradi-
ent across the port to cause fluid or tissue to move into the
port. The actual cutting edges are recessed from the outside
diameter of the cutter by the thickness of the outer “needle”
wall. Tissue larger than the diameter of the port must be
elastically deformed to pass through the port.
Sharpness can be defined as high pressure per unit area
created by a blade with a thin cross section. Knives work
by sharpness, while scissors are never “sharp” or “dull.”
It is not known if microscopic surface roughness facili-
tates cutting or increases undesirable tissue displacement.
The tip of the microvitreoretinal (MVR) blade, although
originally developed for making the sclerotomies (Devel-
oped by Steve Charles, September 1976), is occasionally
used to incise intraocular tissue including ERM. Angled
knives have been developed to delaminate ERMs but have
not been widely adopted because of the risk of iatrogenic
retinal breaks.
Inertial cutting utilizes a rapidly moving cutting element
impacting against tissue at rest and is the mode of action
of the phacoemulsifier. The term “emulsification,” however,
is inaccurate as this refers to overcoming surface (interfa-
cial) tension to produce small droplets of a liquid otherwise
immiscible in water. It is probable that vitreous cutters do
not achieve velocities sufficient for significant inertial cut-
ting effects.
Lasers and the peak electron avalanche knife (PEAK)
work by producing vaporization of tissue, which inherently
produces bubbles and acoustic (remote, mechanical) effects.
Colloquial terms such as “ablation” and “disruption” are mis-
leading as they are not physics-based terms. Lasers produce
power densities sufficient to vaporize tissue or water by utiliz-
ing spatial or temporal coherence. Temporal coherence means
that power is delivered in a very short interval, whereas spatial
coherence means the energy is delivered to a very small vol-
ume. Spatial coherence inherently produces a small volume
of tissue vaporization. Therefore, tissue cutting is dependent
upon a contiguous line of spots or propagation of a tear that
occurs with Yittrium Aluminium, garnet (YAG) laser capsu-
lotomy. ERMs and vitreous are very elastic as well as resis-
tant to fracture propagation. As a result of these properties,
it requires very high-power densities to produce significant
tissue cutting. High-power densities produce many bubbles
as well as a propagating acoustic pressure wave, which can
damage tissue remote to the intended site. The erbium YAG
laser requires shielding from the retina, produces bubbles,
requires vitrectomy system–like fluidics, and removes tissue
at a very slow rate (1). Femtosecond lasers used for LASIK
flap creation and other anterior segment applications utilize
very high temporal coherence and inherently broadband
(white) light but are not applicable to ERM removal.
Enzyme-assisted approaches to vitreoretinal surgery
such as microplasmin-assisted vitrectomy have been stud-
ied for over three decades but have not been shown to be
safe and effective at this time. Enzymatic methods have
been directed at liquefying the vitreous or for separation of
the posterior vitreous cortex from the retina (2). Problems
include the following: (a) potential for damage to the retina,
lens, or zonules; (b) inflammation; (c) potential for prion,
 CHAPTER 2
viral, or bacterial contamination from biological materials;
(d) the need to inject the enzyme at least 30 minutes before
surgery; and (e) complicated, costly preparation processes.
Autologous plasmin enzymes are obtained from the patient’s
own plasma and, thus, will not have prion or viral contami-
nation and appear to be safe with respect to damage to ocu-
lar tissues.
Surgica l Fluidics
A pressure difference across the suction port of a vitrec-
tomy instrument is necessary to cause substances to enter
the port. The term “passive egress” is incorrect because
there is no physical difference between transorifice pressure
(TOP, delta P) achieved by high infusion pressure and that
achieved by negative pressures (vacuum) on the inside of
the port. Semirigid materials such as ERM and lens nucleus
must be deformed to pass through a suction port in response
to a TOP gradient. Markedly low TOP will prolong the pro-
cess of vitreous removal; large gradients create vitreoretinal
traction and retinal breaks.
In general, a maximum suction setting of 150 mm
Hg should be used for 20-gauge systems and 650 mm Hg
for 25-gauge systems when removing vitreous. The lowest
suction force that enables vitreous removal in a reasonable
time should be used to reduce vitreoretinal traction. The
safest way to clear air lock and plugged tubing is by flush-
ing the system outside the eye and not by increasing the
suction force. After the vitreous has been removed, higher
suction force can be used to carefully remove adherent or
previously delaminated ERMs. Suction-induced vitreoretinal
traction is a key factor in the production of intraoperative
Fig ure 2 .2 ■ Intraoperative IOP
is d eterm ined b y infusion p res-
sure that is reduced by resistance
th roug h in fusion system wh ile
flow is occurring m inus vacuum
level at console reduced by resis-
tance in suction system including
cutter port interm ittent closing.
■ Surgical System s, Tools, and Op erating Room 21
retinal breaks. A baseline incidence of aphakic-like retinal
detachments will exist as long as vitreous is removed using
suction force. Factors other than suction force such as lattice
degeneration, preexisting retinal breaks, and vitreous incar-
ceration in the sclerotomies also cause postvitrectomy retinal
detachments.
Infusion Fluidics
Pars plana vitrectomy requires both infusion and aspiration;
there are many parallels between infusion and aspiration flu-
idics because they are influenced by the same physical prin-
ciples. Resistance to fluid flow is determined by the internal
diameter of a lumen, length of tubing, and flow restrictions
in the cannula. Fluidic resistance is proportional to the
fourth power of the diameter (Hagen-Poiseuille equation)
and linearly related to the length. The impact of diameter
is significant because of the fourth-power relationship and
clinically relevant because of the transition from 20-gauge
(0.89 mm) to 23-gauge (0.75 mm) or 25-gauge (0.5 mm)
technology in recent years.
The resistance of the infusion cannula is more than that
produced by 84 inches of connected tubing. Ohm’s law (volt-
age = current × resistance [E = IR]) is mathematically equiva-
lent to Ohm’s law for fluid flow (pressure gradient = flow ×
resistance). Fluidic resistance is an advantage in the aspira-
tion system. By contrast, fluidic resistance is a disadvantage
in the infusion system because it produces a pressure drop
between the infusion pressure and the intraocular pressure
(IOP), which must be compensated for. The pressure drop
occurs only during flow and increases with higher flow rates
(Fig. 2.2). It is typically 20 mm Hg or more during core
22 SECTION II ■ Surgical Technology and Techniques
vitrectomy. Conversely, the infusion pressure is equal to the
IOP when there is no flow, that is, during membrane peel-
ing, scissors segmentation or delamination, and endophoto-
coagulation. The highest flow rate and subsequent pressure
drop occurs with the 20-gauge fragmenter, excessively low
cutting rates, and nontapered extrusion cannulas.
Effects of Low IOP
Excessively low IOP during vitrectomy is common in part
because of undue concern about the effects of higher pres-
sure. Low IOP causes pupillary constriction for unknown
reasons; miosis often results in a need to perform potentially
complication-producing techniques such as using iris retrac-
tors and injecting intraocular epinephrine. Low IOP permits
deformation of the corneal dome from the force of a handheld
irrigating contact lens; the author prefers an IOP of 45 mm Hg
unless there is low arterial perfusion pressure as occurs in
children or in some instances with general anesthesia. The
most significant complication of low IOP is bleeding from
resected vessels, most commonly in diabetic traction retinal
detachment cases or those with elevated neovascularization.
Suprachoroidal hemorrhage during vitrectomy is virtually
always iatrogenic, caused by inadvertent suprachoroidal infu-
sion with secondary shearing of the short posterior ciliary ves-
sels. Hypotony is 100% preventable during vitrectomy, which
is not the case with cataract surgery because of the need to
open the eye for IOL implantation. Suprachoroidal infusion is
preventable by careful attention to surgical technique during
infusion cannula placement and rapid recognition of displace-
ment of the infusion cannula into the suprachoroidal space.
Effects of High IOP
Excessively high IOP during vitrectomy can lead to the
occlusion of the central retinal artery and corneal edema.
Corneal edema is more likely to occur when the endothelial
cell count is low as a result of previous intraocular surgery,
trauma, or Fuchs’ dystrophy. Corneal edema occurs almost
instantaneously when the IOP is excessive but takes substan-
tially longer to resolve when the IOP is normalized. Fortu-
nately, corneal edema is an early warning sign of excessive
IOP before the retinal vessels come into view.
Once the retinal vessels are in view, the retinal vessels
can be monitored for interrupted flow.
Intentionally elevated IOP is a useful technique to con-
trol bleeding when segmenting or delaminating vascularized
ERMs, resecting scar tissue, or removing a subretinal chor-
oidal neovascular membrane. Rapid foot pedal control is
essential when using transient high IOP to control bleeding.
Sa fety Preca utions for Infusion Ca nnula
Pla cement
Wound construction is discussed extensively in another tuto-
rial but several safety points are important to emphasize. The
infusion cannula must be inserted in a relatively firm eye so
that it penetrates completely through the choroid and nonpig-
mented pars plana epithelium, as opposed to just the sclera.
Care must be taken to choose a location away from a preexist-
ing choroidal or suprachoroidal hemorrhage or known abnor-
mality of the pars plana (trauma scar, persistent hyperplastic
primary vitreous (PHPV)). The cannula should be visualized
with the operating microscope or an indirect ophthalmo-
scope, not an unaided eye prior to use to avoid subretinal or
suprachoroidal infusion from failure to visualize transparent
tissue over the cannula tip. The infusion cannula tubing must
be adhesively fastened to the drape with a significant service
loop to prevent cannula dislocation from ocular rotation to
see the periphery. The cannula must be placed close to the
3 o’clock or 9 o’clock position so that it will not be displaced
by the lower lid or speculum during ocular rotation.
Infusion System Technologies
All systems available prior to the development of the Accu-
rus (Alcon, Inc., Ft. Worth, Texas) used gravity-based infu-
sion systems. Gravity-based systems have three significant
disadvantages: There is no digital readout of infusion pres-
sure, and surgeons cannot mentally convert inches (or
centimeters) of water to millimeters of mercury; IV poles
cannot be controlled by the surgeon; and motorized IV
poles are slower than the VGFI (vented gas-forced infu-
sion) and the more advanced Alcon Constellation Vision
System pressurized infusion/aspiration system. The older
gas-forced infusion is better than gravity-based systems
because gas-forced infusion readily produces a direct digital
readout. The VGFI implemented on the Accurus was even
better because it allowed rapid decrease as well as increase
of infusion pressure via surgeon foot pedal command. The
Constellation Vision System has servo-controlled IOP com-
pensation (not infusion pressure). The Constellation Vision
System calibrates the resistance of the infusion system dur-
ing push-priming, measures actual flow using a proprietary
sensing system, calculates fluid Ohm’s law in real time, and
adjusts infusion pressure to produce the selected IOP with
±2 mm Hg.
Asp irat io n Fluid ics an d Vit re o us Cut t in g
Pars plana vitrectomy requires both infusion and aspiration;
there are many parallels between infusion and aspiration flu-
idics because they are influenced by the same physical prin-
ciples. Resistance to fluid flow is determined by the internal
diameter of a lumen or port, length of tubing, cannula or
tool, and the cutter port opening and closing, cyclically
obstructing the port. Fluidic resistance is proportional to the
fourth power of the diameter (Hagen-Poiseuille equation)
and linearly related to the length. The impact of diameter
is very significant because of the fourth-power relation-
ship and clinically relevant because of the transition from
20-gauge (0.89 mm) to 23-gauge (0.75 mm) or 25-gauge
 CHAPTER 2
Fig ure 2 .3 ■ High cutting rates
cause p ort-b ased flow lim iting ,
which increases fluidic stability.
(0.5 mm) technology in recent years. The resistance of the
cutter inner needle and the infusion port limit the flow far
more than the 84 inches of connected tubing. Ohm’s law,
voltage = current × resistance (E = IR), is mathematically
equivalent to Ohm’s law for fluid flow; pressure gradient =
flow × resistance.
Port-based flow limiting is a term coined by the prin-
cipal author that encompasses both the flow limiting result-
ing from smaller diameter cutters and that caused by higher
cutting rates and biased closed duty cycles. Higher cutting
rates cyclically interrupt the flow through the port, thereby
increasing fluidic resistance (Fig. 2.3). High cutting rates,
and therefore higher fluidic resistance at the port, are ben-
eficial for all cases and all tasks because they increase flu-
idic stability, which, in turn, decreases pulsatile vitreoretinal
traction on both detached (retinal motion) and attached
retina and therefore iatrogenic retinal breaks. The principal
author refers to the amount of fluid that passes through the
port during an open-close cycle as “pulse flow.” High cut-
ting rates produce many small-volume pulses with much
less remote (far field) effects, that is, pulsatile vitreoretinal
traction, than those that occur using lower cutting rates.
Small pulse flow means that the vitreous does not have time
to produce remote effects because of the Force = Mass ×
Acceleration (F = MA) relationship. Higher cutting rates
do not cut collagen fibers better; this is because the veloc-
ity of the cutter does not increase with higher cutting rates
on pneumatic cutters. In addition, port-based flow limit-
ing decreases surge, and therefore iatrogenic retinal breaks,
after sudden elastic deformation of the dense ERM or scar
tissue through the port. High cutting rates in addition to
■ Surgical System s, Tools, and Op erating Room 23
producing desirable port-based flow limiting reduce the
travel of uncut vitreous collagen fibers through the port.
25 gauge provides more resistance than 23 gauge because
fluidic resistance is proportional to the fourth power of the
diameter. Surgeons often incorrectly believe that 25-gauge
vitrectomy is “inefficient” or produces insufficient flow rates
when, in fact, it is safer because of less pulsatile vitreoretinal
traction.
Port-based flow limiting relies on the same physical
principles as high-vacuum, low-flow phaco as first imple-
mented on Alcon MicroFlare ABS and MicroTaper ABS
phaco systems, now the standard of care. High-vacuum,
low-flow phaco produces better anterior chamber stability
and decreased fluid surge after occlusion break, which are
directly analogous to the advantages of port-based flow lim-
iting for posterior vitrectomy.
Th e Ro le o f Vit re ct o m y Te ch n iq ue
Phaco technique is largely based on using aspiration to move
lens material away from the lens capsule (to prevent capsu-
lar defects and vitreous loss). In marked contrast, the vitre-
ous cutter port should be moved to the vitreous rather than
the vitreous pulled to the port using excessive flow rates
(Fig. 2.4). Phaco surgeons must consciously focus on mov-
ing the port to the vitreous because their phaco experience
teaches them the opposite approach.
Higher flow rates from larger diameter cutters are not
more (or less) efficient; efficiency is defined as the volume of
vitreous removed per volume of infusion fluid. Similarly, effi-
ciency is not a function of cutting rate; efficiency is entirely
24 SECTION II ■ Surgical Technology and Techniques
Fig ure 2 .4 ■ Cutting while pulling away increases vitreoretinal
traction from vacuum force at cutter port.
driven by technique. Keeping the port constantly immersed
in vitreous produces efficiency.
The author refers to the optimal technique as continu-
ous engage and advance vitrectomy. Current emphasis on
efficiency and faster operating times can result in the unin-
tended consequence of pulling the cutter back while aspirat-
ing, greatly increasing vitreoretinal traction (Fig. 2.5).
Ph ysics o f Vit re ct o m y
Vitreous is a very complex tissue with low homogeneity; the
physical properties vary widely from patient to patient and
disease to disease and change dramatically as the vitrectomy
progresses. Vitreous hyaluronan acts as a Non-Newtonian,
pseudoplastic fluid similar to viscoelastic agents in the ante-
rior chamber resisting deformation into the cutter port. Early
in the vitrectomy, surgeons often believe that “nothing is
happening” when in fact hyaluronan is being removed, and
they may react by unsafely increasing flow rates, usually by
decreasing the cutting rate. Hyaluronan acts as a dampening
agent reducing vitreoretinal traction from pulsatile flow
through the port. Hyaluronan is diluted as the vitrectomy
progresses, decreasing the dampening effect, clearly an issue
because vitreous cutting close to the retina is typically per-
formed after core vitrectomy. Furthermore, infusion fluid
changes the electrochemical properties of the vitreous, dra-
matically decreasing its viscosity. It is of interest that vitre-
ous viscosity is reduced by a factor of five in minutes after
removing it from the eye or enucleating an animal eye.
Cut t e r Te ch n o lo g y
Pneumatic cutters are much lighter and more compact than
electric cutters, thereby improving dexterity (Weber-Fechner
law) and decreasing hand fatigue. Although this is commonly
misunderstood, disposable tools actually reduce per-case
costs because they eliminate cleaning, rinsing, drying, wrap-
ping, sterilization, storage, replacement, and spare parts
Fig u re 2 .5 ■ Cutting wh ile ad van cing shears tissue in its
orig inal position, which m inim izes vitreoretinal traction.
costs. Cleanup of any tool with a lumen that includes cutters,
scissors, forceps, and cannulas has the potential of creating a
TASS (toxic anterior segment syndrome)-like inflammation
from biological materials from previous patients, enzymes
used in ultrasonic cleaning, autoclave water impurities, and
the like. In addition, vitreous cutters, scissors, and forceps,
especially in smaller form factors (23 and 25 gauge), have
fragile cutting and gripping surfaces that are damaged by
cleaning and sterilization process.
The author developed the InnoVit dual actuation
scheme to eliminate the spring used to open the port after
the pressure pulse on the diaphragm closed the port. Elimi-
nation of the spring increased cutting rates and cutter veloc-
ity at the time of closure. The InnoVit utilized a limited angle
rotary cutting scheme rather than an axial (guillotine) cut-
ting action. The UltraVit developed for the new Alcon Con-
stellation system uses a diaphragm-based, dual actuation,
axial cutting design (Fig. 2.6). Duty cycle is defined as the
percentage of port open time versus total time. Lower duty
cycle results in more port-based flow limiting, and there-
fore fluidic stability, and less pulsatile vitreoretinal traction.
Fig u re 2 .6 ■ Cutaway view of Alcon UltraVit 5,000 cp m
cutter.
 CHAPTER 2
Higher duty cycle produces greater flow and more pulsatile
vitreoretinal traction, suitable only for core vitrectomy. The
Alcon Constellation UltraVit currently cuts at 5,000 cuts per
minute and has variable duty cycle control, enabling the
control of port-based flow limiting independent of cutting
rates.
Asp irat io n Te ch n o lo g y
The VISC (vitreous infusion suction cutter), RotoExtractor
Vitreophage, and other early vitrectomy machines used a
syringe operated by the assistant to produce aspiration. This
method was highly imprecise and even dangerous because
of stiction in the syringe and control by someone other than
the surgeon. Conor O’Malley and Ralph Heinz developed
the Ocutome 800 that utilized on-off foot pedal control of
preset vacuum, which was a huge improvement. The author
developed linear (proportional) vacuum for the Ocutome
8000 working with engineers at CooperVision. The Ocular
Connection Machine (OCM) system that the senior author
developed with InnoVision had extremely fast fluidics but
was never commercialized; fortunately, much of the technol-
ogy was incorporated into the Alcon Accurus system, and
the entire package is implemented and improved upon on
Fig ure 2 .7 ■ Alcon Constellation Vision System .
■ Surgical System s, Tools, and Op erating Room 25
the Alcon Constellation (Fig. 2.7). Vacuum response to a foot
pedal command to decrease vacuum is far more important
from a safety perspective than a command to increase vac-
uum. Response time is driven by many factors: size of vac-
uum chamber in the cassette, proportional valve(s) response
time, embedded controller, and use of a real-time operating
system. The Alcon Accurus had the first real-time operating
system and distributed embedded processing resulted in a
25-millisecond response time compared to ten times lon-
ger for the closest competitor. The Constellation Vision Sys-
tem has flow control using flow sensing and servo feedback
control and is twice as fast as the Accurus. This technology
produces rapid, nonpulsatile control unlike the pulsatile,
slower flow control produced by a peristaltic pump system.
In addition to flow control, there is a flow-limiting mode;
these systems will increase safety near the retinal surface,
especially with mobile retina. Port-based flow limiting pro-
duced by high cutting rates (5,000), smaller lumens (23,
25, and 27 gauge), and now variable duty cycle control
is instantaneous, while console-based flow control must
interact through two-way pass of the fluidic signal through
84 inches of compliant tubing.
It has long been incorrectly taught that cutting rates
should be reduced when removing dense ERMs, scar tis-
sue, and lens material or performing core vitrectomy. The
surgeon reaction time from seeing a visual target to mov-
ing a foot pedal is approximately 400 milliseconds because
of visual and cognitive processing, generation of a motor
response, propagation through the spinal cord and leg, and
contraction of the lower leg muscles. The surgeon reaction
time is more than an order of magnitude longer than the
25 milliseconds it takes a modern venturi-based system to
respond to a foot pedal command. The senior author has
calculated that approximately thirty times the volume of the
fluid contained between the tip and the port goes through
the port in the time that occurs between when the surgeon
decides to lift the pedal and when the vacuum is actually
reduced at the cutter port.
Cutter Movement
Many surgeons habitually pull back the cutter while suc-
tion is applied, which greatly increases vitreoretinal traction.
Some surgeons have recommended pulling the cutter away
from vitreous while cutting, but this technique is dangerous
because the force created by probe movement is added to
the force created by suction. This is particularly dangerous
when slow cutter speeds, relatively high suction levels, or
malfunctioning cutters are used. In cataract or penetrating
keratoplasty, vitrectomy cutting while pulling the probe out
of the vitreous has the added disadvantage of creating air-
lock as air is introduced into the system. If material becomes
caught in the port, the surgeon should not jerk the probe
back (“burned-hand” reflex) but should wait and have an
assistant squeeze the suction tubing or, better yet, use the
foot pedal reflux mode.
26 SECTION II ■ Surgical Technology and Techniques
With the Accurus or especially the Constellation,
accurately controlled low suction force allows the port
to be turned toward the retina or 90 degrees rather than
180 degrees away. The probe should be advanced toward
the tissue to cut, just as one would use any other power
tool. Excessive suction must be applied if the port is turned
anteriorly, away from the vitreous to be cut. The combina-
tion of the high-performance cutters, advanced fluidics,
and especially high-speed cutting achieves the ultimate
goal of cutting tissue in its original position. Surgeons
accustomed to the vitreous being drawn to the port at
first may think that these systems are not cutting, but later
they learn to move the cutter to the tissue to be removed.
The authors refer to the optimal approach as continuous
engage and advance, emphasizing the need to take the port
to the vitreous, never pull back while cutting, and avoid a
start-stop approach.
Port Confi gura tion
Large ports decrease the force per unit area but allow the
entry of larger tissue pieces without cutting or deformation
of the tissue. Ports closer to the tip are advantageous only
when high cutting rates and high-performance cutter/flu-
idic systems permit safe operation near the retina. Although
most ERMs are removed by delamination or forceps mem-
brane peeling, cutter operation near the retina is especially
advantageous in removing ERM in diabetic traction detach-
ment cases using conformal cutter delamination or foldback
delamination.
Self-Sha rpening
Higher force-to-weight ratio pneumatic actuators are pre-
ferred over electric actuators to drive self-sharpening cutters.
Self-sharpening cutters remain sharp throughout the proce-
dure. Any self-sharpening cutter must be disposable because
self-sharpening results in loss of metal and, ultimately, fail-
ure of the probe.
There has been an evolution of systems utilized to con-
trol suction force for vitreous surgery. Manually operated
syringes controlled by the assistant provided excessive fluc-
tuation of suction force because of syringe friction and a vol-
ume/flow rather than pressure control. Manually operated
syringes caused large swings in TOP and excessive vitreo-
retinal traction. Use of the mechanical syringe drive offered
better control because of its mechanical advantage. Peristal-
tic pump systems primarily control flow rather than pres-
sure and typically undergo transient peaks of high TOP as
materials are impacted in the port. As the material deforms
and moves rapidly through the port, this excessive pulse of
“residual” suction force is transmitted to the surrounding vit-
reous, creating undue vitreoretinal traction. Venturi pumps
are safer than peristaltic or scroll pump–based vitrectomy
systems and reduce pulsatile (i.e., peristaltic) vitreoretinal
traction.
Vacuum systems controlled by the surgeon’s foot offer
a major advantage over surgeon’s hand or assistant control.
Suction force is analogous to the pressure one places on a
knife or scissors to cut and should be directly controlled by
the surgeon. The maximum vitreoretinal traction is under
constant control with a controlled vacuum system. Control-
ling the vacuum with a button on the cutter causes inadver-
tent hand movements and operator fatigue. Foot-controlled
solenoid valves interrupt the connection of the suction port
to a collection bottle at a preset vacuum level but create an
abrupt onset of suction and inability to decrease suction
without stopping the process. Better control is made possible
by having the foot pedal proportional to the suction force.
Simply stated, pedal position correlates with suction force.
Proportional suction control, frequently called linear suc-
tion (3,4), permits continuous optimization of the suction
force rather than using the circulating nurse for resetting the
level of suction. A preset maximum should be utilized, with
a constant use of just enough suction to remove vitreous as
judged visually. Readout of this vacuum-produced force is
available on the console but seldom observed because the
correct force is judged by viewing vitreous flowing into the
port and unwanted retinal movement.
A distinct advantage of linear suction is the control over
the rate of increase, and especially decrease, of the suction
force. The suction force is gradually increased until vitre-
ous movement into the port is just achieved. Fast response
time is essential for proportional (linear) suction systems.
The Constellation and Accurus systems have been opti-
mized to decrease the vacuum rapidly (25 ms) in response
to a foot pedal command to reduce suction, without the
need to stop cutting. This precise control requires match-
ing dynamic port resistance, cutting rate, tubing resistance,
and console fluidics. A real-time operating system imple-
mented on a dedicated fluidics processor provides guaran-
teed response time. The Accurus system and especially the
Constellation system coupled with the high-speed cutters
enable safe dissection with the probe virtually touching the
retinal surface.
Vitreous Cutter Design Issues
The initial electric motor–powered, nondisposable vitreous
cutters such as the Machemer-Parel VISC and the Douvas
RotoExtractor utilized continuous rotary motion and non-
disposable cutters. This combination often produced marked
winding and vitreoretinal traction as the cutters began to have
wear-induced poor apposition of the shearing surfaces. Non-
disposable axial cutters such as the O’Malley-Heinz Berkley
Bioengineering Ocutome eliminated the winding problem
noted above but did not solve the wear-induced poor cutting
problem. Disposable, pneumatic, axial cutters, also referred
to as guillotine cutters, such as the Alcon Accurus and Alcon
Constellation UltraVit are self-sharpening, which ensures
that all vitreous collagen fibers flowing through the port are
cut if cutting rates are sufficiently high.
 CHAPTER 2
The initial vitreous cutters utilized brush-type direct
current (DC) motors that are too costly to be disposable. The
nondisposable Ocutome utilized a bellows pneumatic actua-
tor, which dramatically reduced the weight and size of the
cutter system from hundreds of grams to 22 g. Diaphragm
(Accurus, Constellation UltraVit), dual-activated piston
(InnoVit), and Alcon Constellation UltraVit dual-actuated
diaphragm cutters weigh less than 10 g. Pneumatic actuators
produce roughly ten times the force per unit mass and force
per unit volume compared to electrodynamic actuators.
Fatigue is reduced and tactile sensing and dexterity increased
with lighter finger and hand loading. Shorter handles reduce
torque on the hand due to cable/tubing friction on the drape
as well as unintentional pulling by the assistant.
Pneumatic actuators have much lower moving mass than
direct electromagnetic actuators such as solenoids. Higher
moving mass reduces acceleration, assuming a constant force
(F = MA, where F is force, M is mass, and A is acceleration).
Swash plate cams driven by a continuous rotary motor have
velocities limited by the cam follower “floating” as well as
by spring return force limitations. Although some advocates
have stated that the motion of a cam-based cutter is a triangle
wave, this is impossible, because it would require infinite
acceleration when the cutter reverses direction. Electromag-
netic actuator forces are limited by the heat produced by I2R
(current squared times resistance) losses in the motor wind-
ings. Electrically driven cutters become noticeably hot when
driven at high cutting rates. The spring return on axial cutters
reduces the velocity as the port approaches closure (Hooke’s
law, F = −KY, where F is force, K is the spring constant, and
Y is the displacement). The Alcon InnoVit cutter (US Patent
5,176,628) utilizes dual pneumatic actuation to eliminate the
spring slowdown problem and a gear-driven cutter to pro-
duce limited rotary motion, which eliminates winding. Rotary
motion allows the port to be closer to the tip because an over-
shoot distance is not needed for the moving cutter element
to slow down and reverse direction. The Alcon Constellation
UltraVit cutter utilizes a dual-actuated diaphragm cutter that
eliminates the spring, eliminates the friction of the InnoVit
piston seals, and enables variable duty cycle control.
In fusio n Sit e Op t io n s
Infusion instruments, as well as other vitreous surgery
instruments, can enter the eye through the limbus, ciliary
body, or pars plana.
Full Function Probes
The generic term “vitreous infusion suction cutter” (VISC)
implies the full function probe concept. While having all
functions on one probe was an original goal in the design of
vitrectomy instruments (5–7), in almost all situations, sepa-
ration of functions is more flexible and permits smaller inci-
sions (8,9). Restated, three-port vitrectomy is preferred to
one-port vitrectomy for virtually all situations.
■ Surgical System s, Tools, and Op erating Room 27
Fig ure 2 .8 ■ 25-gauge infusion cannula.
Infusion Ca nnula s
Although an infusion sleeve can be slipped over a 20-gauge
cutter to restore the full function probe concept, this is only
rarely useful for selected anterior vitrectomy scenarios. Infu-
sion cannulas through the pars plana create far less turbu-
lence and decrease fluid throughput (Fig. 2.8). Separation
of the infusion function from the vitreous cutter decreases
its size and enhances flexibility. Proportional (linear) suction
connected to various extrusion cannulas (Chapter 4) is far
more efficient with separate infusion and does not require
the presence of the vitrectomy instrument in the eye to
provide infusion. The wide separation of the infusion and
egress ports decreases turbulence and operating time when
removing blood products from the eye. Internal fluid-air/
gas exchange and internal drainage of subretinal fluid (SRF)
(Chapter 4) are similarly more efficient with a separate
infusion system. If a retinal break is present or occurs dur-
ing vitrectomy, the SRF will increase when exposed to the
probe-located infusion of the full function probe, while the
separate infusion system approach results in decreased SRF
if the cutter is brought near the break. The infusion system
is placed without the vitrectomy probe, following the “first
in, last out” rule to provide infusion and pressure control
throughout the case.
Infusion into the suprachoroidal or subretinal space is
the only complication unique to the infusion cannula sys-
tem. These complications are manageable but, more impor-
tantly, preventable by careful technique (Chapter 4). Longer
(4 mm) cannulas than the 2-mm cannulas originally recom-
mended reduce the chance of inadvertent suprachoroidal
or subretinal infusion and do not strike the lens, unless the
cannula is pushed by excessive rotation of the eye, causing
contact with the lids, or if the surgical assistant pulls on the
tubing by mistake.
Alt e rn at ive In fusio n De vice s
A 30-degree bend in a blunt, end-opening, handheld can-
nula permits access to the anterior chamber via a pars plana
28 SECTION II ■ Surgical Technology and Techniques
entry site over the bridge of the patient’s nose (10). These
cannulas can be used for infusion when the pars plana can-
not be visualized, as in most ROP cases, preexisting supra-
choroidal hemorrhages, pars plana abnormalities, and some
cases of anterior vitrectomy.
Sharp infusion needles risk inadvertent puncture of
ocular structures and damage to the cutter. The bevel can
leak fluid when the tip is near the wound and can infuse into
the choroid or cornea. Scalp vein needles (butterfly needles,
winged infusion sets) and silicone tubing can inadvertently
contact the corneal endothelium and should be avoided.
In fusio n Fluid
Infusion fluids that include a bicarbonate buffer, dextrose,
and glutathione in addition to the usual components of
Ringer’s solution should be considered mandatory in all cases
(Alcon Laboratories BSS Plus) (11). The tubing and connec-
tors must be handled with meticulous, sterile technique.
Bubbles should be prevented by flushing, and care should
be taken to avoid negative pressure in infusion bottles. Cer-
tain surgeons use lactated Ringer’s solution or balanced salt
solution (BSS) rather than BSS Plus in a misguided attempt
to save money. High labor cost due to slow surgery, not infu-
sion fluid, is the most significant cost driver, yet it receives
less emphasis because it points back to the surgeon. Glucose
was initially added to the infusion fluid before the advent of
frequent blood glucose testing during surgery. Twenty years
ago, the procedure was to utilize intravenous dextrose 5%
in water to prevent hypoglycemia during local or general
anesthesia. The principal author has not added anything
to the infusion fluid for over 25 years and never observes
significant posterior subcapsular changes. The safety of the
cumulative dose of antibiotics or epinephrine in the infusion
bottle has not been established. The principal author has
performed over 26,000 vitrectomies, with only three cases
of postoperative endophthalmitis. The author (S.C.) had
two cases of endophthalmitis in 1975; one case had severe
oral sepsis, and multilayer drapes were used, permitting the
patient to exhale on the operative site. The second case had
an infected suprapubic cystotomy and, in retrospect, should
have been postponed until the systemic infection had been
successfully treated. The third case occurred when no sub-
conjunctival antibiotics were used at the end of the case, an
experiment that was immediately terminated. The authors
never add epinephrine or any other agent to the infusion
fluid because of the risk of contamination and toxicity and
the possibility of the incorrect dose or agent.
En d o illum in at io n
Full function probes utilize a coaxial fiberoptic illumination
system. While this is vastly superior to microscope illumina-
tion, it is inflexible and illuminates only the cutter tip and
surrounding vitreous. A separate endoilluminator (Fig. 2.9)
may be used bimanually with intraocular instruments other
Fig u re 2 .9 ■ En d oillum in ator elim inates lig h t scatter that
would occur with transcorneal illum ination and allows the light
source to be m oved around for optim al focal, d iffuse, specular,
and retro illum ination.
than the cutter and allows illumination of any area of interest.
Retro, focal, and specular illumination with the endoillumi-
nator allows better visualization of clear vitreous and pro-
duces much less scattered light (glare) than transcorneal slit,
chandelier, or torpedo illumination. Safety is increased by
the ability to illuminate peripheral vitreous during mem-
brane or vitreous removal. Diathermy can be added to the
disposable endoilluminator, which is ideal for diabetic cases
still using 20-gauge systems. Various divergence angles can
be used for a spot or flood light effect. Wide-angle illumina-
tion devices produce more uniform illumination for video
but make visualization of clear vitreous more difficult. Wide-
angle illumination using a standard 78-degree endoillumina-
tor is essential when using wide-angle visualization systems
for peripheral viewing in rhegmatogenous retinal detach-
ment, proliferative vitreoretinopathy (PVR), and giant break
cases.
Erg o n o m ic Issue s
All hand instruments should be held in three fingertips. Scis-
sors and forceps should rest on the web of tissue extend-
ing from the base of the thumb to the second digit, while
vitreous cutters should be shorter to permit greater angular
motion. Fingertips are softer and have better tactile sensa-
tion than the bony portions of the fingers. Triangulation of
instruments between three fingertips and this web is the
most stable grip and occupies approximately 35 mm of han-
dle length. Any handle or cable extending beyond this grip
moves the center of gravity behind the grip point and cre-
ates torque on the fingers. Extension beyond 35 mm gives
dangerous leverage to an assistant inadvertently striking the
handle or pulling on the cable, tubing, or fiber. All tools
should have a concavity or another means of axially con-
straining the grip. Axially constraining the grip allows the
 CHAPTER 2
surgeon to loosen the grip and reduce fatigue without fear of
the probe slipping. Consistent tip to grip distance facilitates
cerebellar learning (so-called muscle memory) for the sur-
geon and reduces bumping the retina. An hourglass-shaped
probe (Developed by Steve Charles, September, 1976)
wedges between the fingers, reducing the force required to
prevent slippage in either direction.
The lighter the probe the better; a heavier electrically
driven cutter creates excess loading of the proprioceptive
sense (Weber-Fechner law) and decreases sensitivity. Excess
weight coupled with excessive length is extremely muscle
fatiguing and tremor inducing. Anxiety and caffeine accen-
tuate natural tremor as well. Pneumatically driven, dispos-
able, self-sharpening, hourglass-shaped, small, lightweight
probes are the best design compromise. Nonslip surface of
the cutter allows easier grip by the surgeon and improves
control of the instrument with less firmness of grasp.
OPERATING MICROSCOPE
An operating microscope with power zoom, focus, and
two-axis power translational (XY) or rotational (pitch, yaw)
movement is necessary for vitrectomy (Fig. 2.10) (12).
Coaxial stereo assistant’s microscopes are essential for vit-
rectomy. The power XY system can be controlled by a
three-axis, six-direction chin switch (13), but this can cause
Fig ure 2 .1 0 ■ Zeiss operating m icro-
scop e with m otorized XY m ovem ent,
stereo observer scope, em bedded 3-CCD
television cam era, and sterile controls for
releasing brakes to enable m anual posi-
tioning of the m icroscope.
■ Surgical System s, Tools, and Op erating Room 29
temporomandibular joint discomfort. Zoom, microscope
light on-off, and XY reversal for wide-angle systems must
be controlled on the foot pedal. Speech recognition systems
have been used to control microscope position but are ill
suited to control any of the real-time tasks. An additional
observer tube can be utilized for teaching other surgeons,
technicians, or nurses. High-quality television in the OR is
very advantageous for team coordination and in-room teach-
ing, but recording is overutilized.
En d o p h o t o co ag ulat io n
Endophotocoagulation was initially developed for retin-
opexy and panretinal photocoagulation (PRP) (14). The
principal author developed a technique of coagulating
retinal surface bleeders in diabetic cases with the 20-gauge
Chang end-aspirating endophotocoagulation probe but
now uses 25-gauge laser probes for hemostasis without
aspiration. Although some surgeons prefer the laser indi-
rect ophthalmoscope (LIO) to endophotocoagulation, the
LIO can cause cornea, lens, and iris damage; longer operat-
ing time; and surgeon neck strain. The far periphery can be
visualized extremely well with wide-angle viewing systems;
using an LIO is unnecessary and has no advantages in the
majority of cases. A 532-nm, diode-pumped, frequency up-
converted laser is currently the best choice for endophoto-
coagulation.
30 SECTION II ■ Surgical Technology and Techniques
Diat h e rm y
Diathermy is the term used to describe the use of radio frequency
(RF) energy to produce heat and thereby coagulate proteins.
Although intraocular diathermy can be used for retinopexy,
the primary application is hemostasis. Bipolar means all the RF
energy must pass between two intraocular electrodes within
the eye rather than between one electrode and patient ground.
Bipolar is safer than monopolar because RF might propagate
through the optic nerve with monopolar diathermy. The Alcon
disposable bipolar endoilluminator and its predecessor, the tis-
sue manipulator, have a small antenna as one electrode and the
tool outer shank as the return path. This small antenna can be
thought of as “focused” diathermy as is preferred to the standard
intraocular bipolar with a symmetrical electrode configuration.
The principal author found the disposable bipolar endoillumi-
nator to be an ideal “second” instrument during the delamina-
tion of diabetic traction retinal detachments before switching
exclusively to 25-gauge surgery for all cases.
OPERATING ROOM SETUP
The operating table and head support must be very rigid
and stable and have minimal thickness between the patient’s
head and the surgeon’s knees. The surgeon must sit up
very straight, stretching to reach the microscope in order to
decrease all-too-common cervical and lumbar spine prob-
lems caused by leaning and bending forward. The ideal
surgeon chair or stool should have back support and no
armrests, and the shape and elasticity of the seat should exert
no pressure on the sciatic nerves or perineum. The chair or
stool must have a foot pedal or, even better, a full circle foot
control for height adjustment by the surgeon.
The surgeon’s hands rest on the patient’s face during
vitrectomy and not directly on the “wrist rest.” The wrist rest
is used primarily to create a trough in conjunction with the
drape, permitting continuous aspiration of fluid runoff from
the eye. The trough is a necessity to protect the foot pedals
from water damage, to contain contaminated fluids, and to
catch dropped instruments. A high position of the wrist rest
will result in retinal or lens damage if the patient turns his or
her head from side to side.
The microscope should usually be placed on the patient’s
right side to create a tent with the drape under which the
anesthesia personnel can visualize the airway and EKG leads.
An assistant can be located on the right or left side of the
patient, but the left side is preferable because of access to the
console and instruments. Hand instruments are best kept on
a Mayo stand between the surgeon and the assistant on the
left, with a back table placed behind for other surgical tools
and disposables. The vitrectomy system and 532-nm diode-
pumped laser are stacked on a cart at the patient’s left hand,
with a sterile shelf or Mayo stand in front for the associated
handpieces and tubing. This stand should be placed over
the patient’s chest with the bipolar handpiece, contact lens,
contact lens suction, and infusion tubing, after the patient is
prepped and draped. The Mayo stand over the patient’s chest
creates a tent that allows the anesthesiologist to visualize the
chest wall movement, abdomen, airway, and EKG leads. This
setup is used regardless of which eye is to be operated upon,
thus permitting more rapid and consistent setting up.
Placing the equipment on a rack over the patient pre-
vents the surgeon from viewing the console display. In addi-
tion, access to the patient is limited in the case of a medical
emergency or restless patient. A modern vitrectomy system
such as the Alcon Constellation Vision System (Fig. 2.11) is
Fig ure 2 .1 1 ■ Standardized OR
config uration enab les rap id and
efficient setup and takedown.
 CHAPTER 2
best placed at the patient’s left hip. Dual xenon light sources
are embedded in the Constellation system.
Microscope drapes are essential to prevent fibers, wires,
tubing, and instruments from being contaminated. The micro-
scope is handled by many people and is directly over the
patient, and it is therefore a potential source of contaminated
particles falling on the eye.
Powder-free gloves should be used on all cases. Spe-
cific care must be taken for patients with latex allergy. Care
should be taken to prevent cotton fibers, plastic particles,
and cellulose material from touching the instruments. The
authors almost never touch the eye with cellulose sponges or
cotton-tipped applicators to avoid particulate matter being
left in the eye. Surgeon masks with an integral plastic flap
to prevent fogging are easier on the skin than foam rubber
or adhesives. It has been reported that there is a higher inci-
dence of facial basal cell carcinoma in surgeons who taped
their mask for many years.
Use of a single-piece drape with an integral fluid collec-
tion trough is the most efficient method of draping. The drape
should be transparent over the airway and face for safety. The
drape should be thicker than typical drapes to avoid inadver-
tent perforation. Head drapes and other drapes are not needed,
saving time and money. Drapes without an opening for the eye
should be used and a cut made in the drape after the drape
is adhesively applied with the lids open. The flaps created by
this method are then folded over the lid margins and held in
place by the speculum to keep cilia and lid bacteria away from
the operative site. The skin must be degreased and dry, or
the drape will not adhere. The drape must first be applied to the
superior orbital ridge, then the nasal canthus, and finally the
zygoma and outer portion of the orbital ridge in order to have
a complete seal. A complete seal prevents water from running
into the patient’s hair and ears and, more importantly, prevents
exhaled air from reaching the eye. The patient’s exhaled air is a
sterility risk and causes fogging of the contact lens. Tincture of
benzoin can be used to increase drape adherence.
Pre p Te ch n iq ue
Povidone-iodine (Betadine 5%) prep should be used even if
the patient has an iodine allergy. Betadine drops should be
used for the cul-de-sac before and immediately after surgery.
Lash trimming is no longer performed because it has been
reported that it increases bacteria loading in the conjunctival
cul-de-sac, emphasizing the need to cover the lashes with the
drape. The corneal epithelium should be covered with the lids
or irrigated frequently with BSS to markedly reduce the need
for scraping. The authors almost never need to remove the
corneal epithelium. There is little, if any, evidence that vis-
coelastics preserve the corneal epithelium better than BSS.
■ Surgical System s, Tools, and Op erating Room 31
St e rilizat io n Syst e m s
The steam autoclave, when properly utilized, kills bacteria,
fungi, parasites, and prions (transmissible spongiform enceph-
alopathy, bovine spongiform encephalopathy). Vaporized
hydrogen peroxide (Sterad) and washer disinfecting, spori-
cidal (Steris) systems cause less damage to certain instruments
than the steam autoclave but probably do not kill prions.
Surg ical Effi cie n cy
Reduced operating, setup, cleanup, and turnover time
reduces labor costs. Standardization and stable staffing
reduce training time and associated costs, errors, and mate-
rials waste. Intuitive user interfaces allow the utilization of
less-skilled staff and reduce errors and surgeon frustration.
It is the surgeon’s responsibility to lead by example and
encourage a team approach rather than blaming, criticizing,
and becoming angry.
References
1. Banko A. Apparatus for removing blood clots, cataracts, and
other objects from the eye. United States Patent 3,732,858.
Filed February 14, 1969.
2. Machemer R, Buettner H, Norton EW, et al. Vitrectomy: a
pars plana approach. Trans Am Acad Ophthalmol Otolaryngol
1971;75:813.
3. Charles S, Wang C. Linear suction control system for the vitre-
ous cutter. Arch Ophthalmol 1981;99:1613.
4. Charles S, Wang C. Motorized gas injector for vitreous surgery.
Arch Ophthalmol 1981;99:1398.
5. Peyman GA, Dodick NA. Experimental vitrectomy. Arch Oph-
thalmol 1971;86:548.
6. Douvas NG. The cataract roto-extractor (a preliminary report).
Trans Am Acad Ophthalmol Otolaryngol 1973;77:792.
7. Federman JL, Cook K, Bross R, et al. Intraocular microsurgery 1:
new instrumentation (SITE). Ophthalmic Surg 1976;7:82.
8. O’Malley C, Heintz RM. Vitrectomy via the pars plana—a new
instrument system. Trans Pac Coast Otoophthalmol Soc Annu
Meet 1972;53:121–137.
9. O’Malley C, Heintz RM. Vitrectomy with an alternative instru-
ment system. Am Ophthalmol 1975;7:585–588.
10. May DR. Anterior chamber infusion with the 30 degree bent
needle. Ocutome Fragmatome Newsletter 1979;4:4.
11. Edelhauser HF, VanHorn DL, Schultz RO, et al. Compara-
tive toxicity of intraocular irrigating solutions on the corneal
endothelium. Am J Ophthalmol 1976;81:473.
12. Parel J-M, Machemer R, Aumayr WA. New concept for vitreous
surgery for automated operating microscope. Am J Ophthalmol
1974;77:161.
13. Charles S, McCarthy C, Eichenbaum DA. Chin operated switch
for motorized three axis microscope movement. Am J Ophthal-
mol 1975;80:150.
14. Charles S. Endophotocoagulation. Retina 1981;1(2):117–120.
C H AP TER
3
CONSTELLATION ARCHITECTURE
The Constellation is the culmination of over
three decades of evolutionary development of
vitreous cutters and fluidics, new and improved
tools, illumination sources, phaco technology,
systems integration, efficiency systems, and advanced user
interface design. The principal author is the architect of the
Alcon Constellation Vision System.
The vitreous infusion suction cutter (VISC) and Roto-
Extractor were single-port, large-incision, so-called full
function, slow-speed, rotary, electric cutters with aspiration
provided by a syringe operated by the assistant. The Berkley
Bioengineering Ocutome 800, developed by Conor O’Malley
and Ralph Heinz, was the first three-port, 20-gauge (0.89
mm) system and had the first lightweight, pneumatic probe
and surgeon foot pedal–controlled, on-off aspiration, a major
advance. Berkley Bioengineering subsequently was acquired
by Coopervision, and Coopervision was later acquired by
Alcon. The Coopervision Ocutome 8000, developed by
Carl Wang, his engineering team, and the author, had the
first linear suction (now used on all vitrectomy and phaco
machines), an integrated light source, and a connected
fragmenter. The MidLabs MVS system, developed by Carl
Wang and the author, had the first disposable pneumatic
cutter, a crucial improvement over reusable cutters with
low-performance cutting. After the original MidLabs sys-
tem was acquired by Alcon Laboratories, the author started
InnoVision and began the development of the Ocular Con-
nection Machine (OCM). The OCM had a dual actuation
InnoVit cutter with limited angle rotary cutting at 1,500
cuts per minute, linear diathermy, tool ID, an articulated
arm with integrated tubing management, servo-controlled
intraocular pressure (IOP), a graphical user interface with
soft keys, integrated xenon illuminator, integrated frag-
menter, auto–gas mixing, auto fluid-air exchange valving,
and power scissors. The InnoVision OCM technology was
never commercialized and was later acquired by Alcon Lab-
oratories, and I became a consultant for Alcon. Many of the
OCM concepts were improved upon and incorporated into
the highly successful Accurus system, which included an
advanced graphical user interface with soft keys and global
functions, VGFI (vented gas-forced infusion), integrated
32
fragmenter, silicone injector (VFC), and a halogen light
source. 25-gauge and later 23-gauge tools were developed
for the Accurus platform and are now in their third genera-
tion. A nonintegrated EyeLite 532-nm diode-pumped laser
and later a nonintegrated xenon source were developed to
use with the Accurus system.
The Constellation incorporates new, improved imple-
mentations of all the concepts incorporated in the OCM as
well as the Accurus and adds many new capabilities. The
vitreous cutter is the pivotal component of any vitreoretinal
surgery system; the Constellation UltraVit cutter utilizes dual
pneumatic actuation (no spring return axial cutter) such as
the InnoVit and currently operates at 5,000 cuts per minute
(Fig. 3.1). Because it is diaphragm driven, it eliminates the
friction of the InnoVit piston, provides the familiar axial cut-
ter format, and scales better to 23-, 25-, and even 27-gauge
form factors. The cutter is driven using a proprietary variable
duty cycle technology dependent on the dual actuation sys-
tem. Variable duty cycle control enables the use of a biased
closed approach to produce increased port-based flow limit-
ing when working in close proximity to the retina, so-called
vitreous shaving. Alternatively, port-based flow limiting can
be decreased by using a biased open approach enabling
greater flow rates when doing core vitrectomy. The aspi-
ration system has an extremely low latency response time
to a foot pedal command for vacuum decrease or increase
because of a new, triple proportional valve aspiration sys-
tem and cassette design. Sensor-based, fast-response, digital
flow control and flow limiting facilitate working safely near
the retinal surface. The aspiration system provides continu-
ous linear (proportional) reflux as well as microreflux for all
aspiration tools.
The Constellation utilizes a real-time operating system
and distributed processor architecture to ensure reliability
as well as 2× faster response time than the Accurus. It has
a switched Ethernet architecture with 42 printed circuit
boards, a Pentium, five microprocessors, and many Field
Programmable Gate Arrays and uses over 600,000 lines of
code. The electronics have power backup for situations such
as the cord being pulled out or power failure in the operating
room (OR) to facilitate resumption of surgery.
 CHAPTER 3 ■ Constellation Architecture 33

Fig u re 3 . 1 ■ Cu t away view

of UltraVit 5,000 cp m vitreous
cutter showing d ual d iap hrag m
actuation.

Integrated pressurized infusion using a two-chamber

system and servo control of the IOP is unique to the Con-
stellation and is especially valuable in high-flow scenarios
such as removal of dislocated dense lens fragments using
the fragmenter with 23- or 25-gauge infusion. IOP compen-
sation will reduce the sudden IOP decrease and resultant
bleeding after dense epiretinal membrane deforms through
the cutter port. It is likely that IOP control will enable the
use of lower average infusion pressures. The infusion system
has automatic bottle-out warning and enables changing the
bottle with no interruption in fluid flow or bubbles.
The system can be configured with one or two dual-port
xenon illumination sources to facilitate the use of illuminated
tools, chandeliers, and Torpedoes (Fig. 3.2). The new xenon
illuminator design is more efficient and produces longer, 400-
hour lamp life. The xenon optical system produces greater
than 25 lumens using 23- and 25-gauge fibers at 200 hours.
Radio frequency identification (RFID) connectors on the illu-
mination tools automatically adjust the initial xenon source
intensity depending on specific tool characteristics: light
throughput, typical working distance, and divergence angle.
This set point is 8 to 10 lumens, optimal light intensity for all
tools in 20, 23, and 25 gauges. The surgeon can increase illu-
mination, if needed, to the maximum FDA-allowed output.
RFID in the tool connectors activates parameter and
mode setup, decreasing the setup time as well as workload and
training requirements for the circulator. RFID also activates a
new time-saving, push-prime system for the vitreous cutter
and any connected extrusion tools. A sterile articulated arm
incorporating a tubing management system enables priming
and testing of all infusion and aspiration components before Fig ure 3 .2 ■ One or two dual-port xenon illum inations system s
the patient is anesthetized, prepped, and draped, significantly can supply m ultiple endoillum inated devices with over 25 lum ens
decreasing the setup time and eliminating the need for a Mayo of white light.
34 SECTION II ■ Surgical Technology and Techniques
Fig ure 3 .3 ■ Diagram of the em bedded PurePoint 532-nm laser with a thin-disk laser engine for
im proved control and reliability.
stand. Embedded wizards and video help facilitate faster
setup even when using less experienced personnel.
The Constellation can be configured with an embedded
PurePoint 532-nm laser. The PurePoint laser is a novel,
advanced, thin-disk 532-nm solid-state laser (Fig. 3.3). The
thin-disk laser engine reduces thermal lensing, which is the
change in the index of refraction of the Nd:Vanadate lasing
medium as it heats up. The thin-disk design produces a more
constant laser output, especially with higher powers and firing
rates because of thermal stability. The lithium triborate frequency
upconversion crystal (1,064–532 nm), thermal electric cooler,
and all optomechanical components are solder mounted in a
fixed, mechanically and thermally stable position. A separate
foot pedal controls the laser power and standby/ready/standby
control with voice verification, eliminating dependence on the
Fig ure 3 .4 ■ Auto gas fill calculates the d esired g as concentration and elim inates the need of
a circulator to operate the gas tank regulator.
circulator. The Constellation graphical user interface eliminates
the need for a separate display and controls.
An auto infusion valve replaces the stopcock used for
fluid-air exchange, eliminating delay while air travels through
84 inches of tubing and bubbles. The auto infusion valve can
be controlled by the surgeon’s foot pedal or the scrub tech
(Fig. 3.4) using the Constellation sterile user interface.
Auto gas syringe fill system enables filling the syringe from
attached tanks of SF6 and C3F8 reduce gas wastage, ensures
sterility, and eliminates the need for the circulator to assist the
scrub tech. A software applet calculates how much air to add
to the gas to produce the desired concentration of gas in air.
The VFC power silicone injector supports simultane-
ous aspiration and has RFID to automatically configure the
Constellation mode.
 CHAPTER 3 ■ Constellation Architecture 35

Fig ure 3 .5 ■ Power forceps offer linear control and disposable tip s.

Advanced phaco technology including Ozil torsional
phaco is embedded in the Constellation to support both
combined phacoemulsification-vitrectomy (phaco-vit) pro-
cedures and phaco only procedures.
The Constellation has proportional control of new,
higher frequency, 1.5-MHz, sinusoidal diathermy system
producing 10 Watts maximum. Higher frequency diathermy
produces a more focused lesion possibly reducing retinal
damage.
Power forceps with linear (proportional) control and
power scissors with single-cut and multicut modes sup-
port disposable Alcon Grieshaber disposable DSP forceps
and scissors tips (Figs. 3.5 and 3.6). Bimanual surgery is
supported by single-pedal linear control of power forceps
to grasp and stabilize epiretinal during the initial segment of
pedal travel followed by control of the power scissors with
further depression of the pedal.
The RFID system and a barcode reader wand (Fig. 3.7)
support the end-of-case transmission to a wireless printer
for inventory control, cost accounting, and billing. Turnover
time is greatly reduced because of RFID, barcode reader,
automated push-prime system, testing and setup, and the
sterile articulated arm system. Surgical parameters, laser log,
tool, and consumable use data are generated automatically
and printed on wireless printer for analysis or incorporation
into an operative note.

Fig ure 3 .6 ■ Power scissors have single-cut and m ulticut m odes and have d isposab le tips.
36 SECTION II ■ Surgical Technology and Techniques

Fig u re 3 .7 ■ Barcod e read er wand allows the inp ut of d isp osab les used for instrum en t
program m ing, inventory control, and inform ation m anag em ent.
C H AP TER
4
VISUALIZATION AND
ILLUMINATION
Optimal visualization is required for vitreoretinal
surgery. Many surgeons utilize wide-angle visu-
alization systems for all cases because of habit,
video considerations, or the perception that a
second surgeon is required to hold a contact lens. A plano
(flat) contact lens provides significantly greater lateral and
axial (depth) resolution than a noncontact (BIOM, Oculus
Optikgeräte GmbH, Wetzlar-Dutenhofen, Germany; EIBOS,
Möller-Wedel, Wedel, Germany) or contact-based (Volks
Reinverting Operating Lens System, Volk Optical, Inc.,
Mentor, Ohio; AVI Panoramic Wide-Angle Viewing System,
Advanced Visual Instruments, Inc., New York) wide-angle
visualization system. A plano contact lens should be used for
all macular and traction retinal detachment surgeries. Surgi-
cal technicians usually do an excellent job of holding the
contact lens if they are comfortably seated, view through a
stereo observer tube, and are treated with respect. Assistant
surgeons frequently mentally focus on the surgical procedure
instead of positioning the contact and require reminders to
center the lens. Sewed-on contact lenses were developed to
eliminate the need for an assistant to hold the lens but cre-
ate many problems. Bleeding and bubbles under the lens,
use of expensive viscoelastics, cost of sutures, suboptimal
centering, and damage to the conjunctiva are all potential
problems with sewed-on lenses. Sewed-on contact lenses
are inappropriate for 23/25-gauge, sutureless, transconjunc-
tival surgery because of inherent conjunctival damage and
bleeding. Self-stabilizing contact lenses are useful if a steady-
handed assistant is not available, but they still require fre-
quent recentering by the assistant. Prism lenses can be used
for viewing the periphery but slow down the surgery and
can lead to lens or retinal damage because of paradoxical
movement. Indirect ophthalmoscope viewing for vitrectomy
is grossly inferior to operating microscope-endoilluminator
viewing. The author does not agree with the use of micro-
scope-mounted slit lamp illumination during vitrectomy
because of light scattering by the cornea and lens as well as
other issues.
Wide-angle visualization should be used for all
rhegmatogenous retinal detachment cases, proliferative vit-
reoretinopathy, and giant breaks when viewing the periph-
ery. On occasion, wide-angle visualization is necessary to
visualize the peripheral portion of the posterior vitreous
cortex when operating large tabletop diabetic traction reti-
nal detachments. Wide-angle viewing is ideal to examine
the periphery for iatrogenic retinal breaks after the comple-
tion of vitrectomy and for removal of peripheral intraocular
foreign bodies or lens fragments, especially at the inferior
vitreous base. Contact-based wide-angle visualization (Volk,
AVI) provides 10 degrees greater field of view than noncon-
tact viewing (BIOM) and eliminates all corneal asphericity,
which is common after cataract surgery, LASIK, photorefrac-
tive keratectomy, limbal relaxing incisions, radial keratotomy,
astigmatic keratotomy, penetrating keratoplasty, pterygium
surgery, and corneal lacerations. Noncontact wide-angle
visualization systems require much greater ocular rota-
tion than contact-based systems for viewing the periphery,
thereby increasing the 25-gauge tool flex problem. The most
recent generation of articulated and flexible laser probes as
well as stiffer vitreous cutters, scissors, forceps, and endoil-
luminators virtually eliminate the 25-gauge tool flex issue.
ILLUMINATION
Wide-angle visualization requires some form of wide-angle
illumination. The principal author prefers the standard
25-gauge, 78-degree endoilluminator to chandelier or Tor-
pedo devices. Focal illumination, specular illumination,
and retroillumination are better for visualizing clear, color-
less vitreous, internal limiting membrane (ILM), and shiny
epiretinal membranes than diffuse illumination produced by
chandeliers and Tornambe Torpedoes (Insight Instruments,
Stuart, Florida and Alcon Laboratories, Fort Worth, Texas).
Slit lamp biomicroscopy was developed to provide focal illu-
mination, retroillumination, and specular illumination to
37
38 SECTION II ■ Surgical Technology and Techniques
better visualize transparent structures. Specular illumination
is used when looking for ILM or shiny epiretinal mem-
branes, similar to the way an oblique light source is used
to inspect a shiny surface after sanding, painting, cleaning,
or waxing. Retroillumination is produced by reflected light
from the white sclera transmitted through the retina, retinal
pigment epithelium, and choroid; it is therefore more use-
ful in lightly pigmented eyes. Another very effective way to
produce retroillumination is to reflect endoilluminator light
from the metal surface of the vitreous cutter; this technique
cannot be performed with a chandelier or Tornambe Tor-
pedo. The author uses this technique consciously, but it is
likely that many surgeons use this technique instinctively.
Focal illumination is utilized by positioning the endoillumi-
nator near the port of the vitreous cutter, scissors, or forceps;
the surgeon should be careful to use minimal light intensity
when near the macula.
Some surgeons use a chandelier or Torpedo to enable
bimanual surgery, typically using forceps in one hand and
scissors or a vitreous cutter in the other, although a pic or
microvitreoretinal (MVR) blade can be used as well. Scis-
sors produce a push-out force, which increases with each
use if they are reusable scissors and further increases with
the reuse of disposable scissors. Bimanual surgery is often
used to offset this push-out force using forceps in one hand
for epiretinal membrane stabilization and scissors or the vit-
reous cutter in the other. The author uses disposable curved
scissors for both segmentation and delamination and rarely
performs bimanual surgery.
The use of indocyanine green and other staining agents
for the ILM as well as triamcinolone particulate marking of
the vitreous is driven, in part, by the use of wide-angle dif-
fuse illumination sources that reduce a surgeon’s ability to
view transparent, colorless structures such as clear vitreous,
the ILM, and thin, transparent epiretinal membranes. The
dynamic range of charge-coupled-device (CCD) cameras
and displays for surgical video is 2 to 3 log units (f-stops),
while the surgeon’s eye is 7 log units (f-stops), which, in
part, drive the use of wide-angle illumination systems to
provide uniform illumination for video. The objective is safe,
effective, rapid surgery; making a video is far less important.
A large-screen display in the operating room does, however,
enhance technician and nurse interest, involvement, effi-
ciency, and education. Large flat-screen displays are ideal for
teaching visiting surgeons, residents, and fellows.
Vitreous without hemorrhage, ILM, epiretinal mem-
branes, and the retina are all colorless; the argument that
green or yellow light improves the ability to see these struc-
tures is without merit. Green light improves the contrast of
red structures (blood and blood vessels), but this is never
a problem. Yellow light makes visualization of the macular
xanthophyll nearly impossible, which is crucial in many
cases of macular surgery, diabetic traction retinal detach-
ment, and difficult trauma cases with displaced maculas.
Although there are clinical situations in which vitreo-
retinal surgery must be performed with a suboptimal view
because of corneal disease or cataract, the use of optimal
illumination and visualization technology is an essential goal
in all cases.
PHOTOTOXICITY
Phototoxicity has become a potential issue because of the
availability of high-brightness xenon and mercury vapor
sources. Xenon light sources do not burn out suddenly; the
light output decreases over time, creating potential risk of
excessive light levels if the power is not reduced after lamp
replacement. Chandeliers, illuminated infusion cannulas,
illuminated tools, and small-diameter tools produce less
light output for a given light source intensity because of
inherent losses or larger divergence angles. The light source
intensity must be reduced if the surgeon chooses to utilize a
20-gauge standard endoilluminator with high light through-
put for macular surgery immediately after doing a case using
low–light throughput tools or operating a dense vitreous
hemorrhage in a darkly pigmented fundus requiring higher
light source power settings. The best practice is to start every
case, particularly macular surgery, with low light intensity
and increase the intensity gradually until sufficient illumina-
tion is reached. The latest generation of 25-gauge tools have
high light throughput; the author typically sets his xenon
source at approximately15%.
ICG and other stains are chromophores and, there-
fore, dramatically increase the risk of phototoxicity, espe-
cially when using high-intensity xenon and mercury vapor
sources. Phototoxicity is additive to the well-known chemi-
cal toxicity of the dye molecule in addition to pH, osmolar-
ity, and diluent toxicity issues.
Video is another factor in phototoxicity; because CCD
cameras are not as sensitive as the surgeon’s eye, some sur-
geons use 30/70 beam splitters to produce more light for the
video channel, which produces the unintended consequence
of requiring higher illumination levels to optimize the sur-
geon’s view. Single-chip CCD cameras are more than three
times more sensitive than three-chip cameras but produce
inferior color quality. Most surgeons use three-chip cameras,
which require higher illumination levels. The optimal micro-
scope objective has a focal length of 175 mm. Although
150-mm objectives produce 1 log unit (f-stop) more light,
contamination of tools, tubing, wires, and fibers due to lim-
ited clearance between the microscope and the eye is likely.
A 200-mm objective requires 1 f-stop more light than a
175-mm objective and should not be used.
ENDOSCOPE SYSTEMS
Endoscope systems are rarely, if ever, needed in vitreous sur-
gery. They can theoretically be of value to dissect, drain sub-
retinal fluid (SRF), and laser peripheral retinal detachments
in phakic eyes with small pupils. If capsular fibrosis limits

the view, it can be dissected with the scissors to enlarge the
opening, eliminating the need for an endoscope. The authors
rarely use iris retractors because of iris damage, bleeding,
and postoperative inflammation. If the pupil is small, the
periphery can be seen by using wide-angle viewing, care-
ful positioning of the head and eye, scleral depression, and
sphincterotomies (in aphakic eyes). Endophotocoagulation
of the ciliary body can be performed using scleral depression
or with endoscopic viewing. GRIN rod endoscopes have 15×
more pixels than fused coherent fiber bundle–based systems
but have limited optical resolution because of low modu-
lar transfer function (MTF) microoptics. Combining white
light illumination, laser endophotocoagulation, and a fluid
channel can increase the utility of an endoscope. The fluid
channel plugs easily if used for aspiration and is seldom, if
ever, needed for infusion. The authors find little practical
value for endoscopic vitreous surgery.
Bibliogra phy
Andonegui Navarro J, Marcuerquiaga Arriaga J. Xenon light
induced phototoxic lesions [Article in Spanish]. Arch Soc Esp
Oftalmol 2000;75:117–120.
Azzolini C, Brancato R, Venturi G, et al. Updating on intraoperative
light-induced retinal injury. Int Ophthalmol 1994–1995;18:269–
276.
Birngruber R, Gabel VP. Thermal versus photochemical damage in
the retina—thermal calculations for exposure limits. Trans Oph-
thalmol Soc UK 1983;103:422–427.
Charles S. Illumination and phototoxicity issues in vitreoretinal
surgery. Retina 2008;28:1–4.
Charles S. Retinal pigment epithelial abnormalities after macular
hole surgery [Letter to the Editor]. Retina 1993;13:176.
Donovan M, Carmody RJ, Cotter TG. Light-induced photoreceptor
apoptosis in vivo requires neuronal nitric-oxide synthase and gua-
nylate cyclase activity and is caspase-3-independent. J Biol Chem
2001;276:23000–23008. [Epub ahead of print 2001 Mar 8.]
Fisher PL, Suh DW, Rapp LM. Evaluation of retinal susceptibility to
light damage in pigmented rats supplemented with beta-carotene.
Curr Eye Res 1996;15:219–223.
Fuller D, Machemer R, Knighton RW. Retinal damage produced by
intraocular fiber optic light. Am J Ophthalmol 1978;85:519–537.
Ham WT Jr, Mueller HA, Ruffolo JJ Jr, et al. Action spectrum for
retinal injury from near-ultraviolet radiation in the aphakic mon-
key. Am J Ophthalmol 1982;93:299–306.
Harwerth RS, Sperling HG. Prolonged color blindness induced
by intense spectral lights in rhesus monkeys. Science 1971;174:
520–523.
Keller C, Grimm C, Wenzel A, et al. Protective effect of halot-
hane anesthesia on retinal light damage: inhibition of meta-
bolic rhodopsin regeneration. Invest Ophthalmol Vis Sci 2001;42:
476–480.
CHAPTER 4 ■ Visualization and Illum ination 39
Kim SR, Nakanishi K, Itagaki Y, Sparrow JR. Photooxidation of
A2-PE, a photoreceptor outer segment fluorophore, and protec-
tion by lutein and zeaxanthin. Exp Eye Res 2006;82:828–839.
Koch FH, Schmidt HP, Mönks T, et al. The retinal irradiance and
spectral properties of the multiport illumination system for vitre-
ous surgery. Am J Ophthalmol 1993;116:489–496.
Kraushar MF, Harris MJ, Morse PH. Monochromatic endoillu-
mination for epimacular membrane surgery. Ophthalmic Surg
1989;20:508–510.
Kuhn F, Morris R, Massey M. Photic retinal injury from endoillumi-
nation during vitrectomy. Am J Ophthalmol 1991;111:42–46.
Lawwill T. Three major pathologic processes caused by light in the
primate retina: a search for mechanisms. Trans Am Ophthalmol
Soc 1982;80:517–579.
Maia M, Haller JA, Pieramici DJ, et al. Retinal pigment epithelial
abnormalities after internal limiting membrane peeling guided
by indocyanine green staining. Retina 2004;24:157–160.
McDonald HR, Irvine AR. Light-induced maculopathy from the
operating microscope in the extracapsular cataract extraction and
intraocular lens implantation. Ophthalmology 1983;90:945–951.
Meyers SM, Bonner RF. Retinal irradiance from vitrectomy endoil-
luminators. Am J Ophthalmol 1982;94:26–29.
Michels M, Lewis H, Abrams GW, et al. Macular phototoxicity
caused by fiberoptic endoillumination during pars plana vitrec-
tomy. Am J Ophthalmol 1992;114:287–296.
Miller SA, Landry RJ, Byrnes GA. Endoilluminators: evaluation of
potential retinal hazards. Appl Opt 2004;43:1648–1653.
Noell WK, Walker VS, Kang BS, Berman S. Retinal damage by light
in rats. Invest Ophthalmol Vis Sci 1966;5:450–473.
Poliner LS, Tornambe PE. Retinal pigment epitheliopathy after
macular hole surgery. Ophthalmology 1992;99:1671–1677.
[Comment in Ophthalmology 1993;100:1604–1605.]
Remé CE. The dark side of light: rhodopsin and the silent
death of vision. The Proctor lecture. Invest Ophthalmol Vis Sci
2005;46:2671–2682.
Sparrow JR, Zhou J, Ben-Shabat S, et al. Involvement of oxidative
mechanisms in blue-light-induced damage to A2E-laden RPE.
Invest Ophthalmol Vis Sci 2002;43:1222–1227.
van den Biesen PR, Berenschot T, Verdaasdonk RM, et al. Endoil-
lumination during vitrectomy and phototoxicity thresholds. Br J
Ophthalmol 2000;84:1372–1375.
Williams TP, Howell WL. Action spectrum of retinal light-damage
in albino rats. Invest Ophthalmol Vis Sci 1983;24:285–287.
Wu WC, Hu DN, Roberts JE. Phototoxicity of indocyanine green
on human retinal pigment epithelium in vitro and its reduction
by lutein. Photochem Photobiol [Epub ahead of print September
2004.]
Yanagi Y, Inoue Y, Jang WD, Kadonosono K. A2e mediated pho-
totoxic effects of endoilluminators. Br J Ophthalmol 2006;90:
229–232.
Yanagi Y, Iriyama A, Jang WD, Kadonosono K. Evaluation of the
safety of xenon/bandpass light in vitrectomy using the A2E-laden
RPE model. Graefes Arch Clin Exp Ophthalmol 2007;245:677–
681. [Epub ahead of print 2006 Oct 17.]
Zilis JD, Machemer R. Light damage in detached retina. Am J Oph-
thalmol 1991;111:47–50.
C H AP TER
5
ANESTHESIA AND OPERATING
ROOM EFFICIENCY
PREOPERATIVE EVALUATION
Regardless of the type of anesthesia contemplated for
vitreoretinal (VR) surgery, the patient should undergo a thor-
ough preoperative evaluation prior to the procedure. Under
most circumstances, this evaluation should occur well before
the day of surgery so that required treatment can be per-
formed in advance to help ensure that the patient is in opti-
mal condition prior to surgery. Specific investigations, such
as chest X-ray, ECG, and blood chemistries, should be per-
formed only when dictated by the findings of thorough his-
tory and physical examinations. So-called “screening labs”
are not indicated when the appropriate history and physical
examinations are negative.
GENERAL VERSUS LOCAL ANESTHESIA
Both general and local anesthetic techniques are accept-
able for VR surgery; however, the authors prefer to do the
vast majority of their cases using monitored local anesthesia
for a variety of reasons: (a) local anesthesia offers increased
safety for patients, especially those in high-risk categories,
(b) local anesthesia saves time and reduces cost, and (c) local
anesthesia provides rapid recovery and prolonged analgesia,
both of which are especially important in the outpatient
population.
Not all patients are appropriate candidates for VR
surgery under local anesthesia. Immature, mentally defi-
cient, claustrophobic, and uncooperative patients are best
managed with general anesthesia. Patients with language
barriers, however, can frequently be managed extremely well
with local anesthesia if a competent translator can be found.
Estimated surgical time is an additional consideration when
choosing general versus local anesthesia. Surgeons requiring
more than 90 minutes for a given VR procedure should con-
sider general anesthesia over local anesthesia, as patients may
become restless and uncomfortable when asked to lie com-
pletely still for such long periods. An additional indication
40
for general anesthesia is the patient who insists upon it,
although these patients will be rare if properly informed
and reassured by a sympathetic surgical team. Nitrous oxide
should be avoided not only because of the potential inter-
action with intraocular gas bubbles but also because it has
been shown in large, randomized multicenter trials to result
in a 30% higher incidence of nausea and vomiting.
MONITORING DURING SURGERY
Regardless of the type of anesthesia used, the patient must
be carefully monitored during surgery. Appropriate monitor-
ing begins with the continuous presence of an anesthesiolo-
gist or certified registered nurse anesthetist during the entire
procedure. If sedation is being given, it is not in the patient’s
best interest to have the surgeon or circulating nurse moni-
toring the patient, as may be the case in a brief procedure
performed under strictly local anesthesia without sedation.
Basic monitoring includes continuous ECG, noninvasive
blood pressure, and pulse oximetry. End-tidal CO2 monitor-
ing is additionally essential during general anesthesia and
can also be helpful during local anesthesia. Core tempera-
ture monitoring is indicated during longer procedures under
general anesthesia to help ensure that thermal preservation
procedures are successful and to help in monitoring for
the rare occurrence of malignant hyperthermia. In diabetic
patients, the ability to monitor blood glucose in the intraop-
erative and perioperative periods is also important in order
to recognize and treat extremes of both hyperglycemia and
hypoglycemia.
BLOOD PRESSURE CONSIDERATIONS
DURING GENERAL ANESTHESIA
It is common for VR surgeons to become angry if the patient
moves at all during surgery. An unintended consequence
of this tendency is for the anesthesia provider to maintain
 CHAPTER 5
deeper levels of anesthesia to prevent movements, which
may result in low enough systemic blood pressures to com-
promise cerebral, myocardial, and retinal perfusions. During
VR surgery, intraocular pressure (IOP) should be controlled
in the 35 to 45 mm Hg range. Ocular ischemia and central
retinal artery occlusion can occur if low systemic blood pres-
sures are allowed to persist during the procedure. To ensure
adequate levels of general anesthesia and immobility of the
patient, adequate, monitored muscular relaxation combined
with processed EEG (i.e., bispectral analysis) monitoring
should be considered so that excessively deep levels of gen-
eral anesthesia can be avoided.
SEDATION DURING LOCAL
ANESTHESIA
In general, patients having VR surgery under local
anesthesia should have minimal sedation, most of which
should be given at the time of the block. Patients should
not be sedated too deeply during VR surgery for a num-
ber of reasons. In the first place, airway obstruction may
occur, requiring manual support and interruption of the
procedure. This has been described as AWAC (anesthesia
without airway control). Secondly, respiratory movements
during sleep or near sleep often result in magnified move-
ments of the head, which greatly hinder the progress of the
surgeon who is seeing these movements magnified 20 to
40 times through the operating microscope. Thirdly, some
patients become quite talkative and social when overly
sedated. It may be impossible for them to quit talking and
moving despite the most vigorous admonitions to do so.
The only way to manage these patients is to stop all seda-
tion completely or to convert to general anesthesia. Finally,
patients who are asleep or nearly asleep are prone to awak-
ening suddenly and being totally disoriented, resulting in
movements, which can be devastating, even in the hands of
the finest surgeon. The goal of sedation should be control
of anxiety, rather than true sedation with decreased aware-
ness of surroundings.
Judicious amounts of sedatives and/or opioid agents
can be helpful during local anesthesia for VR surgery, espe-
cially in the patient who is very apprehensive or slightly
claustrophobic. Brevital, thiopental, midazolam, propofol,
alfentanil, remifentanil, ketamine, and others have been
promoted to provide good operating conditions and accept-
able patient sedation for a variety of procedures performed
under local anesthesia. Brevital, thiopental, and alfentanil
have been largely supplanted by propofol. Remifentanil
has a higher profile of nausea and vomiting than the other
narcotics. Ketamine will cause nystagmus and probably
increases IOP. Although the regional block should pre-
vent ocular movements, ketamine is a questionable choice
unless a mentally challenged or combative patient had to
be done under regional anesthesia. The cost of propofol has
significantly decreased since the drug became generic, but
■ Anesthesia and Operating Room Efficiency 41
this drug only remains sterile in a syringe for a maximum
of 12 hours, which may be a consideration in developing
countries. For VR surgery, the emphasis must be placed
on balancing patient comfort and satisfaction while provid-
ing the most stable conditions for surgery. In general, this
means using small doses of rapid-onset, short-acting drugs
given continuously with very careful monitoring of effect.
The goals are to assist the patient in lying perfectly still for
60 to 90 minutes without falling asleep, to enhance analge-
sia, and to provide a measure of amnesia. These are not eas-
ily achieved, but they can be accomplished in most patients
by an experienced and knowledgeable anesthesia team.
PSYCHOLOGICAL PREPARATION
FOR LOCAL ANESTHESIA
In preparing patients for VR surgery under some form of
local anesthesia, it is important to give them specific details
about the experience so that they will suffer no surprises.
They need to know about the drape and about not being
able to see during the procedure. They also need to know
that plenty of fresh air will be provided for them under the
drape and that breathing under the drape will not be a prob-
lem. This is the perfect opportunity to discuss the patient’s
fears, such as claustrophobia, positional dyspnea, positional
pain, and similar concerns. One may discover during these
discussions that a particular patient might be better man-
aged with general anesthesia.
The patient should also be given a realistic estimate of
the length of the procedure and the need for lying extremely
still. Almost anyone can lie still for 30 to 45 minutes, but for
longer procedures the patient must be reassured that short
“time-outs” can be arranged to allow for some movement.
Patients must also be aware that an anesthesia pro-
vider will be constantly present and dedicated to monitor-
ing their condition and to act as liaison with the rest of the
team. It is extremely important for the anesthesia provider
and surgeon to communicate freely during the procedure,
both with each other and with the patient. Simple means
for communication with minimal movement, such as hand-
holding or handheld signaling devices, give the patient a
feeling of comfort in knowing that it is possible to alert the
team to a problem while not jeopardizing the surgical field.
If the patient cannot speak English, it is imperative to have
a translator in the room who is fluent in the patient’s native
language.
CHOICE OF LOCAL ANESTHESIA
There are essentially four types of local anesthesia commonly
used in ophthalmic surgery: topical, retrobulbar, peribulbar,
and sub-Tenon’s. Topical anesthesia is useful in a variety of
operations, but it has limitations in VR surgery due to the
need for complete akinesia during many VR procedures,
42 SECTION II ■ Surgical Technology and Techniques
such as macular surgery and membrane peeling. The terms
“retrobulbar” and “peribulbar” are confusing and imprecise,
and they should perhaps be replaced by the terms “intra-
conal” and “extraconal,” which more accurately describe
the intended location of the needle in the orbit. These tech-
niques carry a risk, albeit small, of major complications,
such as ocular perforation, bleeding, and brainstem anes-
thesia, but both are very useful for VR surgery, providing
excellent akinesia, anesthesia, and prolonged postoperative
analgesia. Sub-Tenon’s anesthesia offers an increased level
of safety over intraconal and extraconal techniques. Sub-
Tenon’s anesthesia may not be appropriate for patients who
have had previous scleral buckling, as scleral perforation
with a sub-Tenon’s cannula has been reported in such a
patient.
The two main local anesthestics used for retrobulbar
anesthesia are lidocaine and bupivacaine. Although bupiva-
caine (Marcaine) has a longer duration of action, it has sig-
nificant cardiotoxicity (1). For the usual macular surgeries
and retinal detachment surgeries, the duration of action of
lidocaine is sufficient for painless surgery. On the other hand,
a longer and more complicated case, such as combined pha-
covitrectomy for a diabetic with dense cataract and complex
traction retinal detachment, may take some surgeons longer
than 90 minutes, including the time for setting up and prim-
ing the machine. These cases are best blocked with bupiva-
caine or done under general anesthesia.
TECHNIQUE FOR INTRACONAL
ANESTHESIA
A 27-gauge needle is preferred to larger needles and blunt
so-called “retrobulbar” needles that cause much more pain
when going through the collagenous lid septum. In addition,
retrobulbar needles often penetrate the septum abruptly
after considerable force is applied and may then perforate
the eye. The conventional 1.5-inch needle is too long for
many orbits and should be replaced by a 1.25-inch needle
to avoid impaling the optic nerve in the orbital apex. The
entry point should be at the outer “corner” of the orbital rim,
not at the outer 1/3, inner 2/3 junction in order to reduce
potential damage to the eye and inferior oblique muscle. The
needle should be directed on a straight, not curved, path
along a plane intersecting the visual axis. The authors use
2% plain lidocaine without epinephrine to reduce the risk
of arrhythmias and hypertension and avoid using bicarbon-
ate because of reports and personal experience with lateral
rectus paralysis for months after surgery. The authors recom-
mend applying pressure on the entire orbit with the palm of
the hand immediately after withdrawing the needle to reduce
bleeding and disperse the anesthetic agent. The authors do
not use Wydase because approximately half of the random-
ized trials show no effect while the others show minimally
faster onset and the agent is likely to cause severe damage if
injected into the eye.
REBLOCKING DURING THE
PROCEDURE
Sometimes local anesthesia must be supplemented during
surgery. This can occasionally be accomplished with topi-
cal anesthesia, but the authors most commonly supplement
intraoperatively by placing a flexible cannula into Tenon’s
space and injecting additional local anesthetic. An additional
intraconal injection can also be performed by placing the
needle between Tenon’s capsule and the sclera to enter the
intraconal space. Most often, reblocking is necessary when
the block has been inadequate, when the patient is undergo-
ing a reoperation, and when the procedure is prolonged.
FACIAL NERVE BLOCKS
Separate facial nerve blocks are rarely indicated, especially
if a well-performed extraconal or high-volume intraconal
block is used. Avoiding a facial nerve block spares the patient
a painful injection and prevents the bleeding, swelling,
and other complications that occasionally accompany these
blocks. If the patient is a marked “squeezer,” the orbicularis
occuli can be easily and effectively blocked by inserting a
0.5-inch 30-gauge needle transconjunctivally into the lower
lid just beneath the orbicularis and injecting about 1.5 mL
of local anesthetic.
SOURCES OF PAIN DURING
VITREORETINAL SURGERY
Local anesthesia needs to be quite complete if the experience
is to be pain free. Manipulation of the iris, ciliary body, and
sclera can all be painful, especially if blunt instruments are
being used. Thermal stimulation is also an important source
of discomfort. Cryopexy is very painful, more so than laser
or even radio-frequency cautery (bipolar diathermy). Lasers
in the near-infrared range are more painful than the argon
laser at 514 nm or the diode-pumped, frequency-doubled
CW YAG laser at 532 nm. Some patients also experience
pain or discomfort during extreme globe rotations dur-
ing vitrectomy. As one or more of these modalities may be
employed during VR surgery, it is important that the patient
receives adequate anesthesia.
CARBON DIOXIDE ISSUES
Patients lying awake under the drape frequently complain
that they “cannot get enough air.” As pulse oximetry rou-
tinely records normal oxygen saturation in these patients,
their complaints are frequently attributed to anxiety. In fact,
CO2 often builds up under the drape, resulting in hypercar-
bia and a feeling of air hunger. This may be noted by a rise in
 CHAPTER 5
the baseline if capnography is being used, even though the
peak expired CO2 may be normal or only slightly elevated.
An easy solution to this problem is to ensure adequate air or
oxygen supplementation near the patient’s nose and mouth
as well as active evacuation of the exhaled gases by way
of a large-bore vacuum line placed under the drapes. The
vacuum line also facilitates cooling, which can be an issue
as well.
AIR/ GAS AND GENERAL ANESTHESIA
If gas and/or air are introduced into the eye during VR
surgery, nitrous oxide should be turned off at least 10
minutes beforehand and fresh gas flow into the anesthesia
machine should be increased to ensure adequate wash-
ing out prior to introduction of the gas. Failure to do so
results in a smaller-than-desired gas bubble within the eye
and lower-than-desired IOP postoperatively when nitrous
oxide diffuses out of the bubble. Conversely, if a patient has
a bubble in the eye from a previous procedure, nitrous oxide
should be avoided from the beginning in order to prevent
the expansion of the bubble by diffusion of nitrous oxide
into it, thus raising IOP. In fact, patients must be warned to
alert physicians to the presence of the bubble should they
require emergency surgery for a nonophthalmic condition.
ANESTHETIC CONSIDERATIONS FOR
SPECIFIC PROCEDURES
En d o p h t h alm it is
Endophthalmitis is an acute situation in which cultures must
be taken and therapy instituted as quickly as possible. In
many situations, cultures and even core vitrectomy can be
performed under topical anesthesia. If general anesthesia is
required, surgery cannot be delayed to allow the stomach to
empty.
Th e Op e n Glo b e
Each patient must be thoroughly evaluated, as the choice
of anesthesia will depend on the extent of the injury and
the ability of the patient to cooperate. Often, initial wound
closure can be accomplished under topical and intracam-
eral anesthesia. In cooperative patients with limited damage,
orbital regional anesthesia can be safely used, provided that
the person performing the block has had sufficient experi-
ence, uses limited volumes of anesthetic, and injects very
slowly (i.e., 1 mL every 30–60 seconds) while closely watch-
ing the eye. When general anesthesia is required, the issue of
whether or not to use a depolarizing muscle relaxant arises.
As there are advocates on both sides of this issue, the choice
must be left to the anesthesia provider, who will make a
decision based on the total clinical picture.
■ Anesthesia and Operating Room Efficiency 43
Scle ral Buckle s
Many presenting for scleral buckling procedures will be
high myopes. These patients have long axial lengths, often
accompanied by posterior staphylomas and scleral thinning.
Sub-Tenon’s cannula techniques might be considered in
these patients to lessen the risk of perforation.
Regional anesthesia for scleral buckling procedures may
be complicated by the fact that the orbital retractor can cause
significant orbital rim pain even in the presence of complete
ocular anesthesia. Additionally, with traction of the extraoc-
ular muscles, the oculocardiac reflex may occur. Most com-
monly, the resulting bradycardia will return to normal when
traction is released, and the reflex will diminish over time.
Intravenous atropine is more effective than glycopyrrolate in
blocking the reflex, but its use is associated with the higher
incidence of subsequent tachyarrhythmias. Local anesthetic
injection may block the bradycardia, but the reflex is also
seen in the presence of a complete block.
Patients who have had previous scleral buckles and
present for another procedure may be difficult to block.
As the buckling may slightly elongate the eye, one must be
aware of an increased danger for perforation. As scarring
occurs, normally “safe” procedures may become less safe,
and ocular perforation has been reported with sub-Tenon’s
anesthesia in a patient with a previous scleral buckle.
An t ico ag ulat io n Issue s
In their practice, the authors virtually never stop anticoagu-
lation or antiplatelet therapy prior to VR surgery, although
it is wise to ensure that the patient taking warfarin com-
pounds has an international normalized ratio (INR) in the
therapeutic range (generally 2–3). Stopping anticoagulants
risks causing morbidity or mortality from a variety of causes,
including stroke, myocardial infarction, pulmonary embo-
lism, and deep venous thrombosis. In the authors’ opinion,
the dangers of intraoperative hemorrhage are grossly over-
emphasized when compared to the dangers of stopping
therapeutic anticoagulation. Properly performed VR surgery
is safe on patients with therapeutic INRs (2,3). Use of can-
nula techniques for local anesthesia greatly reduces the risk
of hemorrhage in these patients, as does the use of short
(1.25 inches) 27-gauge needles placed in the less vascu-
lar areas of the orbit (i.e., avoiding the superior half of the
orbit in general and especially the superonasal quadrant) for
orbital blocks.
Po st o p e rat ive Pain
One source of postoperative pain is the injection of antibi-
otics and steroids into the periocular tissues at the end of
the procedure. This pain can be reduced by injecting these
substances into the sub-Tenon’s space with a cannula if con-
junctival incisions have been made, which is not the case
with 25-gauge, sutureless surgery. In addition, injection of a
44 SECTION II ■ Surgical Technology and Techniques
long-acting local anesthetic, such as bupivacaine, at the end
of the procedure with a flexible cannula can greatly reduce
postoperative pain. This is especially important in the occa-
sional patient who requires general anesthesia for VR surgery
and those undergoing scleral buckles.
SUMMARY
The vast majority of VR procedures can be safely, comfort-
ably, and efficiently performed under local anesthesia with
minimal sedation. Compared with general anesthesia, prop-
erly performed and monitored local anesthesia offers the
patient an increased level of safety, reduced recovery times,
and prolonged postoperative pain relief. Nonetheless, the
choice of anesthesia technique must be based on the needs
of the patient, the requirements of the surgeon, and the
skills of the anesthesia provider, ever keeping in mind that
our ultimate goal is a satisfied patient with a good visual
outcome.
References
1. Marwick PC, Levin AI, Coetzee AR. Recurrence of cardiotoxic-
ity after lipid rescue from bupivacaine-induced cardiac arrest.
Anesth Analg 2009;108:1344–1346.
2. Dayani PN, Grand MG. Maintenance of warfarin anticoagulation
for patients undergoing vitreoretinal surgery. Arch Ophthalmol
2006;124:1558–1565.
3. Fu AD, McDonald HR, Williams DF, et al. Anticoagulation with
Warfarin in vitreoretinal surgery. Retina 2007;27:290–295.
C H AP TER
6
GENERAL POSTERIOR SEGMENT
TECHNIQUES
STRATEGIES
All surgery requires a strategy, game plan, or algorithm before
being undertaken. The complexity of high-technology vit-
reoretinal surgery requires extensive preoperative planning.
The surgical algorithm must include an expected series of
events but must contain branches to allow for the contin-
gencies and discoveries that occur in the course of surgery.
The main branches of the algorithm are determined by the
expected techniques, such as lens removal, endophotoco-
agulation, silicone oil, gas, scleral buckling, etc. Equipment,
disposables, gas, oil, buckles, etc. must be available for any
possible additional techniques driven by the individual case
and any intraoperative findings or complications.
SPECULUM
A heavy wire speculum is an effective means of retracting
the eyelids. Larger bladed specula that lift the lids may inter-
fere with the contact lens and infusion cannula. The drape
should not have a precut hole; it should be placed with the
lids open and cut so the flaps completely cover both lid mar-
gins using the speculum to hold them in place. Covering
the lashes and lid margins is done to reduce the chance of
endophthalmitis and to prevent cilia from being entrapped
in tissue. Aspirating speculums reduce fluid runoff under
the drape or onto the foot pedals, surgeon’s feet, and floor
but tend to be bulkier.
INFUSION FLUID
High-quality infusion fluid containing dextrose, glutathione,
and a bicarbonate buffer should be used for all cases. This
should be freshly prepared just prior to the operation. High-
quality commercial solutions (Balanced Salt Solution, BSS
Plus, Alcon Laboratories) should be utilized in all cases. Use
of BSS Plus reduces corneal endothelial cell loss as well as
trabecular meshwork and lens damage compared to other
solutions. The authors do not add glucose, epinephrine, or
antibiotics to the fluid to reduce the chances of using incor-
rect agents or concentrations as well as to reduce the risk
of corneal, lens, and retinal toxicities. Plastic infusion fluid
containers should be avoided because air diffuses through
the plastic during storage, causing bubbles. Two-component
infusate systems should be designed so that the substance in
the actual infusion container is safe if the second component
is inadvertently omitted, as is the case with BSS Plus.
IMPORTANCE OF VISUALIZATION
Optimal visualization is essential to safe vitrectomy. While it
is obviously desirable to preserve the corneal epithelium and
the lens, neither one is as important as the ability to perform
a safe vitrectomy. This simple principle is often overlooked
with adverse consequences.
Pup illary Dilat io n
The preoperative administration of topical adrenergic
(phenylephedrine 2.5%–10%) and anticholinergic (tropic-
amide 1%, cyclopentolate 1%, hyoscine 0.25%, or homa-
tropine 2.5%) dilating drops is critical to safe vitrectomy.
Minimization of hypotony and mechanical iris trauma usu-
ally allow the preoperative dilation to be sustained through-
out the case. If pupillary constriction occurs, a small pulse
of intraocular epinephrine can redilate the pupil. This epi-
nephrine should be labeled “for injection,” contain minimal
amounts of sodium bisulfite antioxidant, and be diluted to
1:10,000 (1). Intracardiac injection preparations, although
labeled as containing no preservative, contain too much
sodium bisulfite and are potentially toxic to the corneal
endothelium. Surgical pupillary sphincterotomies should be
avoided unless the above steps fail and the fundus cannot be
seen adequately for safe surgery. Iris surgery leads to postop-
erative glare, photophobia, and cosmetic problems, as well
45
46 SECTION II ■ Surgical Technology and Techniques
as increased postoperative inflammation. Iris retractors and
sutures cause iris trauma, inflammation, longer operating
time, and increased cost and should only be used if there
is no other option for surgical visualization. Contact wide-
angle visualization systems with a trained assistant obviate
the need for iris retractors, iris surgery, and intracameral epi-
nephrine in almost all cases.
Le n s Re m o val
If lens removal is necessary, phaco or lensectomy should be
performed before proceeding with vitrectomy, unless the
lens is subluxated or dislocated. Relative miosis may occur
as the vitrectomy progresses, and early lens removal permits
better access to the periphery and prevents unnecessary
iris surgery. The view is always better in aphakic or pseu-
dophakic eyes, and safe, effective vitrectomy should never
be compromised to preserve the lens, even if it is relatively
clear. Saving the anterior lens capsule until the end of the
operation in lensectomy cases hinders the view and offers no
significant advantages. Forceps removal of the lens capsule
in lensectomy cases facilitates total capsule removal without
iris touch and resultant miosis and inflammation.
Wound Construction Issues for Microincisiona l
Vitrectomy Surgery
A transconjunctival, sutureless vitrectomy technique was
introduced by Eugene de Juan, to reduce pain, inflamma-
tion, and operating time, although from the outset, concern
regarding hypotony and endophthalmitis was expressed
with this technique, as with sutureless, small-incision cata-
ract surgery (2). de Juan recommends leaving vitreous in
the wound to reduce the incidence of wound leaks, but the
authors disagree with this approach because leaving vitreous
in the sclerotomies increases the likelihood of vitreous wick
(and therefore increases the likelihood of endophthalmitis)
and increases the likelihood of subsequent hypocellular con-
traction of the collagen fibers and retinal breaks immediately
posterior to the sclerotomies (3). In the authors’ opinion, vit-
reous wicks are more likely to be a factor in endophthalmitis
than hypotony with retrograde flow of tear film into the eye.
Angula ted Wound Construction
Although the authors initially used straight-in wound con-
struction for 25-gauge vitrectomy, they changed to angulated
wound construction after the technique was introduced for
23-gauge vitrectomy. The purpose of angulated wound con-
struction is to create a scleral tunnel that functions as a flap
valve, similar to no-stitch cataract surgery incision. It is not,
however, considered a biplanar wound; the change in trocar
insertion angle is to avoid intraocular contact. The wound
already has been made at the point of trocar insertion. Some
surgeons call the uptilt in insertion angle supination, which
is incorrect. A single plane, 15 degree incision produces the
best scleral tunnel and is less likely to result in an inadvertent
suprachoroidal infusion than a flatter (5 degree) itrajectory.
Sta biliza tion of the Eye
Traditionally, surgeons have attempted to stabilize the eye to
offset translational and rotational forces acting on the eye
during trocar-cannula insertion. An alternative approach
is to allow the eye to move deeper into the orbit but pre-
vent eye rotation by rigidly enforcing the correct geometric
relationship between the trocar and the eye (4). Using the
forefinger on the surgeon’s hand to triangulate the trocar
support on the insertion hand facilitates better control of
hand and trocar attitude. Some surgeons have described this
technique as pressing down on the eye; while this is true,
the key point is to maintain the correct insertion angle and
direction by continuously modifying the handle angle while
allowing the eye to freely move. Initially, this hand holds the
cotton-tip applicator used to displace the conjunctiva, but
the cotton-tip applicator is dropped after the trocar impales
the conjunctiva against the sclera, maintaining the displaced
position. Incidentally, the authors use the same two-handed
approach when performing intravitreal injections and ret-
robulbar blocks.
The surgeon should allow a large service loop before
adhesively fastening the infusion line to the drape to avoid
traction and inadvertent suprachoroidal infusion or can-
nula pullout during ocular rotation. Unless a filtering bleb
or other conjunctival conditions prevent this approach, the
sclerotomies should be placed to facilitate the greatest pos-
sible angular intraocular access.
The infusion cannula port is traditionally placed infer-
otemporally near the lower border of the lateral rectus, but
it is better to move it closer to the 3 o’clock or 9 o’clock
position to avoid contact with the lower lid. The supero-
nasal sclerotomy should be placed on a virtual line from
the lowest point of the bridge of the nose to the center
of the pupil, facilitating better angular movement. Similarly,
the superotemporal incision should be placed on a virtual
line extending from the lowest point of the supraorbital rim
to the center of the pupil. The position of the 25-gauge can-
nulas may require modification in the presence of a filtering
bleb, an area of previous trauma with a sclera wound, or if
there is underlying choroidal effusion or hemorrhage.
The authors initially made the sclerotomies 3.5 mm
posterior to the limbus but have found that moving them to
4.0 mm posterior to the limbus reduces the chance of air in
the anterior chamber of phakic or pseudophakic eyes during
fluid-air exchange (FAX).
Conjunctiva l Displa cement
The concept of conjunctival displacement for sutureless,
25-gauge surgery was developed by de Juan; some sur-
geons using 23-gauge systems have abandoned conjunctival

displacement (in the authors’ opinion, unwisely) because
various techniques to stabilize the globe made displacement
difficult. Tornambe Torpedoes (Insight Instruments, Stuart,
Florida) and other chandelier systems are inserted without
conjunctival displacement, which removes vitreous at the site,
or without angulated wound construction; both a lack of con-
junctival displacement or failure to perform angulated wound
construction increases the likelihood of endophthalmitis (5).
Reduction of Wound Lea ks
The senior author first reported the use of FAX to reduce
wound leaks after 25-gauge straight-in wound construc-
tion, but with the adoption of scleral tunnel wound con-
struction, the author currently utilizes the FAX technique to
reduce wound leaks only in patients with a high degree of
myopia or Marfan syndrome or in other patients with thin
sclera.
Logically, wounds made during 20-gauge (0.89-mm
tools) and 23-gauge (0.75-mm tools) surgeries are signifi-
cantly larger than wounds made during 25-gauge (0.5-mm
tools) surgery, increasing the likelihood of wound leaks and
making wound construction more challenging. Tools used
in 23-gauge surgery were developed to address tool stiffness
and flow restriction issues associated with the first-genera-
tion 25-gauge tools (6). Subsequent generations of 25-gauge
tools have increased tool stiffness and reduced flow restric-
tion; therefore, the authors do not see an advantage of
23-gauge surgery.
When withdrawing the cannulas at the end of a case, the
surgeon should take care to avoid wound eversion. Cannulas
should be withdrawn slowly on the same trajectory as they
were inserted, because the sclera is relatively inelastic and
the goal is to close the scleral tunnel. Scleral tunnel closure
is also dependent on leaving the infusion set at least 25 mm
Hg, which will cause the intraocular pressure (IOP) to press
on the inside of the wound. The surgeon should use point
pressure with smooth forceps, not a cotton-tip applicator, on
the scleral tunnel to press on the outside of the tunnel, simi-
larly to how pressure is placed when withdrawing a needle
or IV line from the antecubital vein. The conjunctiva should
be repositioned while point pressure is applied on the tun-
nel. Some surgeons withdraw the cannulas over the endoil-
luminator. Withdrawing on the insertion trajectory, leaving
the infusion turned on, and point pressure on the tunnel are
far more important in achieving tunnel closure, though.
If the wounds leak, then they should be sutured with-
out hesitation with a single, transconjunctival absorbable
polyglycolic suture such as 8-0. PolySorb. PolySorb has
been associated with less irritation than polyglactin (Vicryl,
Novartis, Basel, Switzerland) (7). It is not necessary to open
the conjunctiva to suture the sclerotomy. Some surgeons
unwisely use sutured-on contact lens support rings, which
can damage the conjunctiva (8).
Transconjunctival 23-gauge and 25-gauge, sutureless
vitrectomies allow flexibility in cannula location. The surgeon
CHAPTER 6 ■ General Posterior Segm ent Techniques 47
should avoid filtering blebs, previous conjunctival wounds,
areas of pars plana scarring, high choroidals, or suprachoroidal
hemorrhage to ensure that the cannula extends into the vitre-
ous cavity and not in the suprachoroidal space. If required,
contiguous placement of all cannulas on the same hemisphere
or quadrant can be performed. Additionally, the infusion port
with some vitrectomy systems (Accurus and Constellation
Vision System, Alcon, Fort Worth, Texas) can be moved from
one cannula to another to facilitate tool access to pathology.
Several early studies demonstrate an apparent increase
in endophthalmitis rates after 23-gauge and 25-gauge surger-
ies. The Wills study has been widely quoted, but the surgeons
involved in the study now believe that the factors resulting
in the apparent increase in endophthalmitis are no longer
significant (9,10). Although hypotony is considered a factor
for endophthalmitis, the author believes that vitreous wicks
and omission of subconjunctival antibiotics are more impor-
tant issues and are bigger risk factors for the development
of endophthalmitis. Some vitreoretinal surgeons stopped
using subconjunctival antibiotics after cataract surgeons
abandoned the practice; however, topical antibiotics achieve
minimum inhibitory concentration antibiotic levels in the
anterior chamber but not in the vitreous cavity of phakic or
pseudophakic eyes. Harry W. Flynn, as well as the author,
recommend using both an antibiotic effective against com-
mon Gram-positive organisms and one effective against
virulent Gram-negative organisms. The authors prefer
administering 20 mg tobramycin subconjunctivally and cef-
tazidime 50 mg; vancomycin 25 mg is a good alternative to
ceftazidime if the patient is allergic to penicillin. Injection in
the inferior cul-de-sac may help the surgeon avoid reflux of
the aminoglycoside into the eye and, therefore, the possibil-
ity of retinal toxicity.
VITRECTOMY OBJECTIVES
Vitreous removal requires a thorough understanding of
surgical anatomy and must proceed in a systematic fashion.
The vitreous body should be thought of in terms of discrete
surfaces that are removed in a specific order. The goals should
not be band cutting, obtaining a view on the posterior pole, or
core vitrectomy only but should be directed at management
of the underlying vitreoretinal process. When vitrectomy
was performed using rotating cutters with syringe-controlled
suction, the surgeon was able to remove much of the vitreous
without moving the instrument from the center of the vitre-
ous cavity. This occurred because of excessive suction and
cutter winding, which drew the vitreous inward; however,
such traction is now recognized to be dangerous. From this
traction-induced central relocation of the vitreous arose the
erroneous concept of “core vitrectomy.” In fact, many eyes
with sufficient vitreoretinal disease to require vitrectomy do
not even have a vitreous “core.” Recent trauma, rarely recent
retinal detachments, and macular hole cases may have rela-
tively normal vitreous requiring core vitrectomy.
48 SECTION II ■ Surgical Technology and Techniques
Better cutters, high cutting rates, fast fluidic controllers,
and proportional (linear) suction control facilitate cutting
the vitreous without moving it from its original position. The
surgeon accustomed to low-performance systems may at first
be confused by the better systems’ minimal pulsatile traction
and may falsely conclude that they are not working.
If the eye is aphakic at the commencement of vitrec-
tomy, the anterior vitreous cortex (AVC) should be removed
first, starting centrally and progressing peripherally. Any
attachments to anterior segment wounds or the iris should
be removed before proceeding posteriorly. In phakic eyes,
the AVC is frequently adherent to the posterior lens cap-
sule, making removal difficult without lens damage. A clear
AVC not causing traction should be avoided in eyes with-
out retinal detachment to reduce lens damage. Similarly, the
AVC should be retained in pseudophakic eyes to reduce the
chances of fogging due to condensation on the intraocular
lens (IOL) in eyes that have had a YAG capsulotomy.
Pars plana infusion cannulas permit the surgeon to
change sides between the vitrectomy instrument and the
endoilluminator, thereby gaining access to the whole pos-
terior curve of the lens. Removal of the AVC in a phakic eye
requires direct microscope viewing and coaxial plus endoil-
lumination without a contact or wide-angle system to avoid
lens damage. A combined nasal and temporal approach is
best for AVC removal in phakic eyes. Eyes with fibrovascu-
lar proliferation on or near the pars plana or a significant
inflammatory component should have lensectomy with for-
ceps capsule removal as a precaution against the formation
of a cyclitic membrane at the lens-AVC interface.
After removal of the AVC, the second objective is usu-
ally the removal of the posterior vitreous cortex (PVC). Eyes
Fig ure 6 .1 ■ Penetration through
the opaque PVC should be m ade
nasally or into an area known to
have PVD and no d etachm ent if
p ossible as determ ined b y ultra-
sound or ophthalm oscopy.
requiring trans–pars plana vitrectomy usually have total
vitreoretinal contact, partial posterior vitreous detachment
(PVD) with a conical vitreous configuration, or total PVD
with a frontal plane configuration. Entry into the PVC should
be made in an area nasally (Fig. 6.1) or preferably known
to have a PVD by ophthalmoscopy or ultrasound. The por-
tions of PVC extending between areas of vitreoretinal adher-
ence are known as bridging and are areas where tangential
traction occurs. All portions of the PVC not in contact with
the retina, that is, both the conical surface and the bridging
portions, must be removed to relieve traction. However, the
vitreous “skirt,” which is the anterior edge of the truncated
cone, must be treated differently. Because retinal breaks can
be caused by surgically induced traction on the skirt, only
enough skirt should be removed to allow satisfactory surgi-
cal visualization and to prevent the superior skirt from cov-
ering any portion of the view when the patient is upright.
If blood products are incorporated in the layers of the skirt,
they should be carefully debulked to diminish postoperative
hemolytic and erythroclastic glaucoma.
EPIRETINAL MEMBRANES
The PVC may include new collagen and glial cells on its sur-
face at sites of vitreoretinal adherence in addition to hypocel-
lular contraction of preexisting vitreous collagen. This type
of pathoanatomy is usually referred to as an epiretinal mem-
brane (ERM). The treatment of ERM is a challenging and
essential component of management or vitreoretinal cases.
The approach to ERM varies with the disease; specific tech-
niques will be discussed with the specific diseases.

Forceps membrane peeling, scissors segmentation, and
scissors delamination are all useful approaches to ERM. The
goal of ERM surgery is to reattach the retina while mini-
mizing recurrences and complications. If the membrane is
loosely adherent and can be removed without iatrogenic
retinal breaks or bleeding, forceps membrane peeling is
the best approach. In diabetes, retinopathy of prematurity
(ROP), and some trauma and proliferative vitreoretinopathy
(PVR) cases, the membrane is usually quite adherent, requir-
ing scissors segmentation and/or delamination. Excellent
visualization is necessary for ERM dissection.
Sim p lifi e d Ep ire t in al Me m b ran e Disse ct io n
Current approaches to vitreoretinal surgery utilize a wide
variety of techniques and technologies. The techniques for
PVR, proliferative diabetic retinopathy (PDR), ROP, epimac-
ular membranes (EMMs), macular holes, and other causes of
vitreoretinal traction typically vary considerably with the dis-
ease process. Fibrin syndrome, corneal problems, cataract,
light toxicity, postanesthesia complications, postoperative
pain, lid and conjunctival hyperemia, and edema are more
frequent after longer procedures. Postoperative atelectasis,
infection, pulmonary embolism, and increased length of
stay have been associated with longer operating times in
the surgical literature. The principal author has observed a
relationship between number of tools used, longer operating
times, and worse surgical outcomes. A unified, all-disease
approach to tools, techniques, and algorithms could theo-
retically reduce costs associated with longer operating times,
staff training time, setup time, and instrument acquisition
and maintenance. Advancements in manufacturing technol-
ogy generally decrease the cost while increasing the quality
Fig ure 6 .2 ■ Mem brane peeling
with end-grasping forceps does not
require entering into the potential
space between the retina and the
m em b ran e or fin d in g an ed g e;
therefore, it is m uch less likely to
d am age the retina than a p ic or
an approach using a forceps blade
under the retina.
CHAPTER 6 ■ General Posterior Segm ent Techniques 49
of the product by reducing the number of steps or processes.
A central theme of this book is to present a simplified, faster,
less-step, less-tool approach to vitreoretinal surgery.
At present, the authors use end-grasping forceps for all
epiretinal and subretinal membrane peeling and fine curved
scissors for all segmentation and delamination. Pics and bent
needles are not needed for EMM dissection and increase
the potential for retinal trauma. End-grasping forceps and
curved scissors are the only dissection tools used by the
authors for virtually all vitreoretinal surgery. The following
section describes a unified approach to tool and technique
selection for vitreoretinal surgery that was developed for use
by novice and advanced surgeons alike.
Man ag e m e n t o f Ep ire t in al Me m b ran e s
Membra ne Peeling
Robert Machemer, developed ERM peeling only 2 years
after he developed pars plana vitrectomy. He used a bent
23-gauge needle to peel macular puckers, which we now call
EMMs. His technique, although effective, required a sharp
needle tip adjacent to the retina surface. Conor O’Malley,
soon thereafter introduced the pic—a rounded tip instru-
ment, safer for membrane peeling. The principal author
developed forceps membrane peeling using end-grasping
forceps and a pinch peeling, en face technique because
membranes slip off the needles and pics and because for-
ceps were still required to remove the membrane from the
eye. A single-step, end-grasping forceps membrane peeling
approach is safer and faster than initiating peeling with a
blade, pic, or membrane scraper and subsequently removing
the ERM with forceps (Fig. 6.2). End-grasping forceps are
50 SECTION II ■ Surgical Technology and Techniques
used with both blades on the anterior surface of the ERM,
a safer approach than attempting to place one blade under
the membrane risking retinal damage similar to peeling with
needles, microvitreoretinal (MVR) blades, or pics. Using pics
or MVR blades requires finding or constructing an “edge,”
which is a potentially dangerous maneuver. Finding the
outer perimeter of an ERM can be difficult with thin, so-
called “glassy,” membranes; retinal surface damage can result
from using this approach. Making an edge involves using an
MVR blade to slice through the ERM, a potentially danger-
ous maneuver if the patient moves or the visualization is not
optimal. The senior author developed the MVR blade and
was an early adopter of pics but has used only end-grasping
forceps membrane peeling for more than three decades. An
end-grasping forceps technique requires perfect alignment
of the forceps blades at the tip, which cannot be maintained
with blades with long axial gripping surfaces, particularly
with multiple cleaning and sterilization cycles for reusable
tools or, worse yet, reuse of disposable forceps. Smaller
23-gauge and 25-gauge forceps are more vulnerable to dam-
age during cleaning and sterilization; the authors use dispos-
able 25-gauge Grieshaber Revolution DSP ILM (Alcon, Ft.
Worth, Texas) forceps for all ERMs, PVR membranes, and
internal limiting membrane (ILM) peeling without pics or
MVR blades. The senior author developed conformal forceps
so that the blade tips would have the same radius of curva-
ture as the retinal surface in contrast to the Eckardt forceps
(Dutch Ophthalmic USA, Exeter, New Hampshire) that have
sharp corners that tend to grasp retina. The DSP ILM forceps
are small enough that the radius of curvature is not a signifi-
cant issue.
Yasuo Tano, MD, developed the diamond-dusted mem-
brane scraper, which some surgeons have adopted. The
authors prefer the end-grasping forceps technique because
it is less likely to damage the retinal surface, especially in
smaller 23-gauge and 25-gauge form factors that produce
greater force per unit area.
Interna l Limiting Membra ne Peeling
ILM peeling was introduced for macular hole surgery and
has been widely accepted; it is generally agreed that it
improves closure rates. A few papers suggested that ILM
produced worse visual outcome despite hole closure;
the authors believe that this is due to indocyanine green
(ICG) toxicity, not ILM peeling. The standard justification
for ILM peeling is to remove “tangential traction.” Successful
ILM peeling guarantees removal of vitreomacular traction,
residual PVC not contiguous with the detached vitreous,
and any ERM that might be present. The principal author
believes that an additional and critical role for ILM peeling
is increasing retinal compliance by approximately 50% so
that the inward-directed surface tension force from a gas
bubble can almost immediately reapproximate the inner
margins of the hole.
Anslem Kampnik, MD, has emphasized that ILM
peeling in EMM surgery produces a reduction in recur-
rence rates. The authors have been following his recom-
mendations by peeling ILM in all EMM cases for several
years; they agree with his observation and believe that
an additional benefit is the elimination of residual striae,
thereby producing better visual outcomes and faster
improvement.
Staining and Particulate Marking
Many surgeons still use ICG staining of the ILM, although its
use is decreasing. The authors have never used ICG because
of concerns about the toxicity of the agent and vehicle and
because it is simply unnecessary. One factor driving danger-
ous and unnecessary ICG use is using noncontact optical
systems such as the BIOM and EIBOS systems that decrease
both axial and lateral resolutions. An additional factor driv-
ing ICG use is using inadequate forceps that are incapable of
end grasping.
Triamcinolone particulate marking, often incorrectly
referred to as staining, is not specific to ILM or ERMs. The
optimal use for triamcinolone is for visualizing the vitre-
ous. Triamcinolone is more often required with a Tornambe
Torpedo (Alcon) or a chandelier illumination system that
produces diffuse illumination, making visualization of the
vitreous, ILM, and ERMs difficult. Triamcinolone is useful
when learning vitrectomy and for the identification of the
PVC anatomy in complex cases.
Scissors Segmenta tion a nd Dela mina tion
The principal author developed scissors segmentation of
ERMs more than three decades ago to address the issue
of iatrogenic retinal tears caused when attempting to peel
adherent ERMs, which are common in diabetic traction
retinal detachment cases. Segmentation is used to reduce
tangential force on the retina by separating an ERM into
“epicenters,” (Fig. 6.3) better termed vascularized adherence
sites. Segmentation is performed by placing one blade of the
scissors under the ERM, between the retina and the ERM,
and the other blade above the ERM. Although this was ini-
tially performed with “vertical scissors,” curved scissors are
far superior for segmentation. Because blade width is much
greater than blade thickness, vertical scissors require more
space between the retina and the ERM than curved scissors
(Fig. 6.4). Segmentation is now primarily used for access seg-
mentation to expose the delamination plane. Using curved
scissors expedites transitioning from access segmentation to
delamination without tool exchange.
The principal author developed scissors delamination
2 years after developing segmentation to address the issue of
residual ERM after segmentation. Residual ERMs in diabetic
traction retinal detachment cases resulted in small areas of
chronic retinal elevation that ultimately caused atrophic
retinal holes and late rhegmatogenous retinal detachments.
 CHAPTER 6 ■ General Posterior Segm ent Techniques 51

Fig u re 6 .3 ■ Scissors seg m en-

tation of ERMs reduces tangential
traction b ut leaves ep icen ters,
wh ich m ay b leed o n th e cu t
ed g es. The b lood acts as a sub -
strate for glial recurrence. Residual
ERM m ay cause residual elevation
and late retinal detachm ent from
atrop h ic holes. Seg m entation is
used when the m em brane is too
ad h erent to p eel an d the retina
too thin for delam ination.

Scissors delamination is performed by inserting both
scissors blades in the potential space between the ERM and the
retina and severing the attachment points (Fig. 6.5). Simplisti-
cally, segmentation means to cut up the ERM, and delamina-
tion means to cut it off. Initially, delamination was performed
with “horizontal” scissors, actually 135 degrees. For almost
two decades, the senior author has used only curved scissors
for conformal delamination. Curved scissors are preferred
to “horizontal” scissors because the curvature of the scissors
matches the curvature of the retina, reducing the likelihood of
impaling the retina with scissors tips. The same design concept
is used for skis—the tips curve up. If the scissors are intro-
duced under the ERM with the blades wide open and then
closed, attachment points will be ripped instead of sheared,

Fig ure 6 .4 ■ Fine curved scissors

are better than vertical scissors for
segm entation because blade width
is greater than blade thickness and
the curve is conform al with retinal
contour, reducing iatrogenic reti-
nal breaks.
52 SECTION II ■ Surgical Technology and Techniques
which causes retinal breaks (Fig. 6.6). When the scissors are
introduced under the ERM with the blades closed and then
opened similar to the technique used for blunt dissection in
general surgery, attachment points will be ripped instead of
sheared, creating retinal breaks. The optimal technique is to
open the scissors just enough to engage an attachment point,
Fig u re 6 .5 ■ Scissors d elam i-
n ation of ERMs is the p referred
m eth od of rem ovin g ad h eren t
m em b ranes as it p erm its com -
p lete rem oval of ERM.
cut, pull back slightly, move laterally to the next attachment
point, advance slightly, and cut again (Fig. 6.7). It is not nec-
essary to lift or stabilize the ERM when disposable scissors or
reasonably new reusable scissors are used. Because an ERM is
taut and there is a pressure gradient across the retina, the ERM
will scroll up and separate from the retina without lifting.
Fig ure 6 .6 ■ The shear point of
scissors m oves forward as the scis-
sors close, creating an exclusion
force that can tear the retina.

Fig u re 6 .7 ■ Th e scissors are
m oved back and forth while m ak-
ing sm all cuts with the tips barely
op en ing ; th e ERM will scroll up
b ecause of in h eren t elasticity,
rarely requiring lifting.
Inside-Out Versus Outside-In Technique
Robert Machemer, used an outside-in technique for bent
needle membrane peeling, and many surgeons emulate this
technique when using delamination. A better approach for
end-grasping forceps membrane peeling, scissors segmenta-
tion, and scissors delamination is to initiate the dissection
centrally and peel, segment, or delaminate outward. Inside-
out dissection is safer because the retina is thicker and
100 times stronger centrally than in the periphery, especially
in ischemic patients and postpanretinal photocoagulation;
the retina is redundant centrally in traction retinal detach-
ment cases, and the view is better centrally.
En Bloc Dissection
En bloc dissection, as initially described, utilized an out-
side-in technique and advocated using the PVC to lift the
ERM. Putting traction on the PVC results in peripheral retina
breaks while the surgeon is focusing on the ERM dissection.
En bloc is a poor term when applied to ERM dissection; it
is only meaningful in cancer surgery where the notion is to
minimize the dispersion of cancer cells. A better way to think
about this issue is to forget the outdated notion that the stan-
dard surgical algorithm is first performing core vitrectomy,
then creating a PVD, and finally dissecting ERMs. If a pos-
terior vitreous separation PVD is not present in a diabetic
traction retinal detachment (TRD) case, proceed directly to
an inside-out delamination starting at or near the optic nerve
head. It is not necessary to remove all ERM in one piece as
suggested by the term en bloc; it is preferable to segment
in several pieces, each of which will be contiguous with the
CHAPTER 6 ■ General Posterior Segm ent Techniques 53
PVC. When most or all the ERM has been delaminated using
an inside-out direction, creation of PVD will be accomplished
as well or it will be irrelevant because there are no longer any
posterior attachment points.
Bima nua l Surgery
Bimanual surgery is a poor term; surgeons have always used
both hands. The point of bimanual surgery is to accomplish
what the author calls “forceps stabilization of epiretinal
membranes,” using forceps to offset the scissors push-out
force (Fig. 6.8). Scissors cut at a point that starts near the
pivot point and moves toward the tip, creating undesirable
push-out force. By contrast, a vitreous cutter cuts on a curved
line, which makes manufacturing much more difficult and
requires vacuum for imbrications into the port but eliminates
push-out force. An unintended consequence of bimanual
surgery is to inadvertently lift the ERM, thereby creating iat-
rogenic retinal breaks. Making small delamination cuts with
scissors almost closed obviates the need to lift the ERM and
visualize the attachment points. Bimanual surgery requires
the use of either illuminated tools, which are not available in
23-gauge or 25-gauge systems, or a chandelier or Torpedo,
which produces diffuse illumination, making visualization of
vitreous, ILM, and transparent membranes more difficult.
Viscodissection
Viscodissection is performed by injecting a viscoelastic in
the potential space between the ERM and the retina. Viscodis-
section is seldom performed because it has not been shown
54 SECTION II ■ Surgical Technology and Techniques
to be advantageous and it increases cost and operating time.
Viscodissection pressurizes the potential space between the
ERM and the retina, which can result in a hydraulic retinal
break if the ERM is tightly adherent. If the ERM is adher-
ent with no gap between the ERM and the retina, the vis-
cocannula may create a retinal break during introduction.
Viscoelastic substances significantly decrease the interfacial
tension of silicone oil, increasing the likelihood of emulsifi-
cation. It is virtually impossible to remove the viscoelastic;
the viscous material potentially increases glial recurrence
rates by retaining cells, fibronectin, and cytokines such as
vascular endothelial growth factor (VEGF), basic fibroblast
growth factor, and transforming growth factor beta.
Power Scissors a nd Forceps
Power scissors and power forceps address the problem of
inadvertent motion of the tool tip during manual actuation.
A potentially offsetting issue is inadvertent hand motion due
to whole leg motion while moving the foot pedal. Older gen-
eration power scissors and power forceps were too heavy and
did not have high-fidelity proportional control; the Constel-
lation (Alcon) power scissors and forceps have solved these
design issues and also have a single foot pedal control mode
for bimanual surgery.
Hemosta sis
Endolaser is better than diathermy for treating the severed
vessel ends of attachment points after delaminating ERMs
in diabetic traction retinal detachment surgery. The authors
Fig ure 6 .8 ■ Illum inated forceps
allow stab ilization of the ERM to
offset p eelin g , seg m en tation , or
delam ination forces.
believe that using excessive diathermy on retinal surface
bleeders can result in subclinical retinal necrosis and late
atrophic holes. On occasion, larger vessels may require dia-
thermy before or after transection, especially in delayed
trauma surgery and when performing retinectomy.
St ab ilize d Surg e ry
Retinal movement in response to fluidic or mechanical dis-
section forces is often an indication that retinal tearing may be
imminent. Ideal vitreous and ERM removal would not cause
any force on the retina but is impossible as described above.
Many techniques have been developed that, although pro-
moted for a variety of reasons, primarily function by reduc-
ing surgically induced retinal motion: (a) flow limiting using
high cutting rates and 25-gauge fluidics, (b) forceps stabiliza-
tion of ERM (bimanual surgery), (c) perfluorocarbon liquids
used during ERM dissection, and (d) viscodissection.
Perfluorooctane has twice the specific gravity of bal-
anced salt solution. Therefore, it produces twice the gravita-
tional force on the retina (11).
In addition, inertial stabilization, because of the F =
MA relationship, offsets pulsating fluidic forces. Perfluo-
rocarbon liquids are immiscible in balanced salt solution,
blood, and subretinal fluid (SRF) and are therefore used
because of surface tension properties (interfacial tension) to
unfold giant breaks and remove SRF in all types of retinal
detachments (Fig. 6.9). The interfacial tension effect also
prevents transretinal fluid flow (uveal-scleral outflow) and
tends to reduce retinal edema while the perfluoron (PFO)
is present.
 CHAPTER 6 ■ General Posterior Segm ent Techniques 55

Fig u re 6 . 9 ■ PFO st ab ilizes

the retina to offset forces due to
m em brane peeling while floating
out subretinal fluid through ante-
rior retinal breaks.

Ep ire t in al Me m b ran e Disse ct io n that excessive retinal surface damage or bleeding will result
Flo wch art if delamination is used. This is sometimes the case with
atrophic, long-standing, diabetic traction retinal detach-
Removal of all apparent vitreoretinal traction should pre-
ments because of ischemia and prior extensive panretinal
cede ERM dissection unless there is limited separation of
photocoagulation (PRP).
the posterior cortical vitreous from the retina. The princi-
If the retina is pushed away by the scissors or if peel-
pal author developed the concept of inside-out membrane
ing is causing excessive retinal movement, forceps stabiliza-
delamination preceding removal of the posterior cortical
tion of the ERM can be used. If retinal mobility is excessive,
vitreous for cases without apparent PVD. Abrams and
PFO or air stabilization of the retina can be used; this is
Williams (12) subsequently recommended an outside-in
typically required in retinal detachment, giant break, and
approach for delamination before the cortex is removed,
PVR cases. Fast cutting and 25-gauge fluidics produce port-
which they termed “en bloc.” In their paper, they rec-
based flow limiting that greatly increases fluidic stability
ommended using traction on the vitreous to lift up the
by reducing pulse flow, thereby reducing pulsatile retinal
ERM. This approach may create peripheral retinal breaks
movement as well as surge after sudden deformation of
because of counter traction and is not recommended
ERM through the port.
by the authors. Many now call the inside-out delamina-
After all vitreoretinal traction and all apparent ERM-
tion before posterior cortical vitreous separation method
induced retinal foreshortening has been managed, inter-
“en bloc” even though that is not how it was originally
nal or direct transscleral needle drainage of SRF should
described.
be initiated in cases with a rhegmatogenous component.
The first stage of ERM dissection is to carefully test
After retinal movement in the direction of attachment
the adherence of the ERM to the retina. In general, EMM,
ceases, internal FAX should be started while continuing
macular hole or ILM, and PVR have less adherence than
to drain SRF. If the retina does not completely reattach
PDR or ROP. If the adherence is low, inside-out forceps
or subretinal air appears, the remaining traction should
membrane peeling is the best approach. If the adherence is
be managed by further vitrectomy, peeling, segmenta-
high, inside-out delamination is the next step. Access seg-
tion, delamination, subretinal traction removal, or, lastly,
mentation can be used to facilitate finding a tissue plane
retinectomy. The silicone soft-tip cannula allows the aspi-
for delamination. In general, almost all ERM is delami-
ration of SRF and preretinal hemorrhage while minimiz-
nated using an inside-out orientation in PDR cases. The
ing trauma to the retina and is the preferred tool for FAX
procedure is halted at the segmentation step if it is thought
by the authors.
56 SECTION II ■ Surgical Technology and Techniques
Cut t e r De lam in at io n
Modern high-speed microincisional vitrectomy cutters allow
safe, direct delamination of epiretinal tissue with the cutter.
Cutter delamination permits continuous aspiration of blood
from the dissection surface, improving surgical visualiza-
tion, decreases surgical time by minimizing instrument han-
dover, decreases surgical costs, and decreases intraoperative
hemorrhage from the hypotony that occurs during replacing
instruments on open cannulas.
Two techniques of cutter delamination can be performed:
conformal cutter delamination and foldback cutter delami-
nation. Conformal cutter delamination involves placing the
cutter opening directly against the edge of the epiretinal tissue
and modulating the angle of attack and suction until the cut-
ter removes the epiretinal tissue without removing the under-
lying retina (Fig. 6.10). Segmentation can also be performed
with the cutter by placing the cutter directly over the valleys of
retinal folds, between retinal attachment points, where clear-
ance from the retina is the highest, and removing the bridg-
ing tractional tissue. Foldback delamination involves placing
the cutter above the epiretinal tissue, with the cutter opening
facing in the direction of the leading edge of the dissection
and allowing the epiretinal tissue to “foldback” into the cutter
(Fig. 6.11). This is best performed when there is a free mobile
edge to the epiretinal tissue that can move toward the cutter.
Suct io n Me m b ran e Pe e lin g o r Po st e rio r
Vit re o us De t ach m e n t Cre at io n
If a loose edge of a flexible ERM or PVC requires peeling, it can
be held in the port of the vitrectomy cutter using the suction-
only mode of the foot pedal or a soft-tip extrusion cannula.
Fig ure 6 .1 0 ■ Conform al cutter
d elam in atio n rem o ves ep ireti-
nal tissue without rem oving the
underlying retina.
The tool can then be moved, lifting the membrane or vitreous
away from the retina (13). The suction force can, however,
cause retinal traction, and certain membranes are difficult to
imbricate into a port. Forceps peeling is preferable in most
instances except for PVD creation in macular hole cases.
“Ch o p st icks” Me m b ran e Pe e lin g
If a flexible ERM has been partially lifted from the retina
surface by peeling, removal can be completed by holding
the membrane between the endoilluminator and the vitrec-
tomy probe. However, if excessive pressure is utilized with
this bimanual method, the instruments may slip past one
another (14). Outward spreading of a pair of blunt tools
such as the probe and the endoilluminator can be used to
separate the PVC or an ERM from the retina by blunt dissec-
tion. The blunt dissection, spreading method is very useful
in PVR cases.
Vacuum Cle an in g –Ext rusio n
When an opening is first made in the PVC, nonclotted
blood, hemoglobin products, and erythroclasts tend to
flow out the opening into the vitreous cavity, especially if
suction is turned off and on repeatedly. This is often falsely
perceived as active bleeding and should be removed by
continuous vacuum cleaning (Fig. 6.12) before continu-
ing with PVC truncation. This approach ensures excellent
visualization and avoids the flow of blood products to the
lens, cornea, and trabecular meshwork.
The vacuum cleaning method utilizes a soft-tip cannula
and a nonpulsatile, controlled pressure gradient across
 CHAPTER 6 ■ General Posterior Segm ent Techniques 57

Fig ure 6 .1 1 ■ Foldback delam i-

n ation is b est p erform ed wh en
there is a free m obile edge to the
ep iretin al tissue th at can m ove
toward the cutter.

the needle port. Keeping the cannula away from the infu-
sion port minimizes turbulence. If console or foot pedal
controlled, this method has been referred to as “extrusion”
(38). The key to understanding extrusion is to remember
these features: (a) end opening, (b) low, precisely controlled
pressure gradient, and (c) nonpulsatile. The original and
now-obsolete form of vacuum cleaning utilized the flute (not
fluted) needle, in which egress is controlled by the surgeon’s
forefinger over the exit port on the side of the handle (15).
This has been incorrectly termed passive egress. Whether
these techniques are considered “active” or “passive” is
irrelevant; precise control of transorifice pressure is the key
concept. The extrusion method utilizes a 20- or, preferably,
25-gauge soft-tip cannula, vitreous cutter on suction-only

Fig u re 6 .1 2 ■ After m akin g

an op en in g in th e PVC, sub -
PVD b lood p rod ucts sh ould b e
rem o ved b y extrusion with a
soft-tip cannula b efore continu-
ing with PVC truncation.
58 SECTION II ■ Surgical Technology and Techniques
mode or an end-aspirating laser probe if the surgeon elects
to use 20-gauge technology connected to the console-based
suction system, with foot control by the surgeon (16). Pro-
portional (linear) suction control allows greater control over
the extrusion process and has supplanted the flute needle for
vacuum cleaning. Proportional suction applied to a cannula
is very effective for removing blood products, substances to
be exchanged, and SRF. Extrusion allows for reflux using the
foot pedal.
Vacuum cleaning–extrusion should be utilized only
when there is a region of nonclotted blood in the vitreous
cavity, the sub-PVD space, or on the retinal surface. The can-
nula should never touch the vitreous or retinal surface but
may be held very close to the retina if the transorifice pres-
sure is kept low. Intentionally lowered IOP can be used to
induce bleeding, so that the source of the hemorrhage can
be located and treated.
The Chang end-aspirating endolaser probe is ideal for
surgeons using 20-gauge technology for removal of prereti-
nal blood and active bleeding while treating surface bleeders
and vascular attachment points and performing PRP in dia-
betic cases. The authors only perform 25-gauge surgery.
EXCHANGES AND SUBRETINAL
FLUID DRAINAGE
Surface Te n sio n Man ag e m e n t an d Exch an g e s
Air, various inert gases (SF6, C3F8) (17–20), silicone oil,
and perfluorocarbon liquids are utilized primarily because
they are immiscible in water (balanced salt solution, retina,
vitreous). This immiscibility is best referred to as interfacial
surface tension.
Surface tension is a physical property of liquids that
depends on intermolecular attraction forces within the liquid
itself. Whereas deep in a liquid all molecules are attracted to
neighboring molecules in all directions, in the surface of a
liquid, molecules are only attracted to those in the plane of
the surface and underneath it, but not above the surface. This
causes surface liquid molecules to form a “film” of closely
packed molecules. The energy required to break this film is
termed surface tension. Different liquids will have different sur-
face tensions, which depend on the chemical characteristics
of their molecules and their intermolecular attraction forces.
Liquids with high electrostatic intermolecular attraction
forces, such as water (surface tension of 73 dynes/cm), have
higher surface tension than nonpolar liquids, such as silicone
oil (surface tension of 20 dynes/cm). Perfluorocarbon liquids
have intermediate surface tension properties (50 dynes/cm).
When two immiscible liquid phases interact, the particular
tension at the interphase is called interfacial tension. The sur-
face tension or interfacial tension values of clinical significance
for vitreous surgery are water/air (or gas) = 73 dynes/ cm, sili-
cone oil/water = 40 dynes/cm, and perfluorooctane/water =
50 dynes/cm. As mentioned above, vitreous, retina, BSS, and
aqueous humor behave as water.
Surface tension management of rhegmatogenous
retinal detachments is based on creating a “liquid surface”
interphase at the level of a retinal hole (or holes). This inter-
phase, by virtue of its surface tension, creates a film over the
hole that prevents the passage of fluid through the hole and
restores the transretinal pressure gradient (Fig. 6.13). This
enables the RPE cells to absorb and drain the SRF, resolving
Fig ure 6 .1 3 ■ Interfacial tension
allows restoration of the transreti-
nal pressure gradient.

the retinal detachment. The term “tamponade” is meaning-
less from a physics perspective. It is derived from French
and means “to plug” and not “to press.” If a gas or silicone
bubble surface is not in direct contact with a retinal hole,
it will not “seal” the hole, and the rhegmatogenous compo-
nent will not be adequately managed. The interfacial tension
of an air or gas bubble “seals” the hole even when it is not in
contact with the RPE.
Density (specific gravity) determines if a substance will
sink or float and the force it will produce at the retinal con-
tact surface. Air, gases, and silicone oil are less dense than
saline or vitreous and migrate opposite to the gravity vec-
tor, and they rise inside the eye. Perfluorocarbon liquids are
almost twice as dense as saline or vitreous and, therefore,
migrate to the lowest part of the eye.
Air, SF6, and C3F8 have identical intraoperative proper-
ties (21,22), although they are different with respect to bub-
ble expansion and duration in the eye. Gases such as C3F8
(23–29) produce greater duration in the eye because the gas-
eous diffusion rate of these larger molecules is much slower
than nitrogen in the blood. Abrams showed that an empiri-
cal isoexpansive mixture of each inert gas with air could be
determined so that complete intraocular filling could be used
without increased postoperative pressure. The isoexpansive
concentration is approximately 25% for SF6 and 18% for
C3F8. Nitrous oxide anesthesia should not be used with air
or any of the inert gases as it can cause rapid intraoperative
expansion of a preexisting bubble or marked postoperative
shrinkage of the bubble.
In t e rfacial Te n sio n Ag e n t s
Exchange techniques are utilized in vitreoretinal surgery to
inject or remove interfacial tension agents: air and longer
acting gases, perfluorocarbon liquids, and silicone oil after
the vitreous has been removed. Although these substances
are often called tamponade agents, this is a poor term.
Using correct terms from physics and chemistry promotes
the understanding of surgical concepts. For example, some
surgeons inject gas after EMM surgery thinking that it will
press on and eliminate retinal folds, which is simply not the
case. Interfacial tension is not related to viscosity or density.
Air and air-gas mixtures produce greater interfacial tension
(70 nM/M) than silicone oil (40–45 nM/M) or perfluo-
rooctane (50 nM/M). The silicone-PFO interface produces
only 5 to 7 nM/M interfacial tension, explaining why these
substances become mixtures over time. Density determines
where the bubble goes; interfacial tension determines what
it does. Many surgeons have said that PFO cannot be used
with 25-gauge surgery because it is too viscous, which is
incorrect; PFO is less viscous than BSS, although it has
almost twice the density demonstrating that density and
viscosity are not related. PFO has approximately the same
interfacial tension as silicone oil, although oil has over 1,000
times greater viscosity. Most surgeons use FAX coupled
with internal drainage of SRF to attach the retina, treat the
CHAPTER 6 ■ General Posterior Segm ent Techniques 59
retinal breaks with endophotocoagulation, and then perform
air-silicone exchange (ASX) for PVR cases. They are often
surprised when the retina is detached on the first postopera-
tive day; this is because the interfacial tension of air is greater
than that of silicone oil and the force of residual traction is
greater than the interfacial tension of oil. Emulsification is
related to many factors, but the most numerically significant
parameter is interfacial tension, which is not related to vis-
cosity explaining why the emulsification rate of 1,000 cs oil
is same as 5,000 cs oil.
There are many ways to think about interfacial tension
agents: Does the agent float (air, gas, and silicone oil) or sink
(PFO)? Does the agent absorb over time (air, gas) or remain in
place (PFO, oil)? Can the patient see through the bubble (sili-
cone oil) or not (air, gas, and PFO)? Can the substance be left
in the eye long term (oil) or only medium term (gas, PFO)?
Factors in silicone oil emulsification include steep bub-
ble radius due to an incomplete fill and the use of viscoelas-
tics or postoperative bleeding or inflammation, all of which
decrease interfacial tension. Nystagmus is another factor in
emulsification because it increases the shear rate, in essence
spinning off small bubbles.
Te rm in o lo g y
Logical terminology is essential in a procedural context as well.
English language convention is that word order is relevant;
FAX means replace fluid with air. Often, surgeons incorrectly
call this gas-fluid exchange, which actually means remove
gas and replace it with fluid. The preferred naming conven-
tion is to use IDS for internal drainage of SRF, FAX for fluid-
air exchange, AGX for air-gas exchange, ASX for air-silicone
exchange, PGX for perfluorooctane-gas exchange, PSX for
perfluorooctane-silicone exchange, etc.
Drain ag e o f Sub re t in al Fluid
Total FAX and near total IDS are utilized by most vitreo-
retinal surgeons when repairing retinal detachments using
vitrectomy. The principal author invented IDS over three
decades ago in order to drain SRF through a preexisting or
iatrogenic retinal break or drainage retinotomy rather than
through the sclera, choroid, and RPE, risking bleeding and
retinal incarceration. Unlike cryopexy during scleral buck-
ling, vitrectomy using internal drainage and FAX permits
postreattachment retinopexy, which in turn enables precise
location of the endolaser without overtreatment and treat-
ing both the retina and the RPE, which results in stronger
adherence. In addition, internal drainage, when coupled
with FAX (Fig. 6.14), enables the assessment of residual vit-
reoretinal traction, a maneuver the principal author refers to
as the reattachment experiment. An assessment of residual
vitreoretinal traction can determine the need for interface
vitrectomy, additional forceps peeling of ERMs, scissors seg-
mentation or delamination, or retinectomy, all performed
“under” air.
60 SECTION II ■ Surgical Technology and Techniques
The first IDS was performed using the vitreous cutter,
which is still a useful technique in some cases. Soon thereaf-
ter, the principal author developed the flute needle, a nonta-
pered, end-opening, blunt cannula with a port on the handle
to control fluid egress with the surgeon’s fingertip (hence
the term flute, like the musical instrument). This cannula
was also used to remove free blood products using a tech-
nique called “vacuum cleaning” similar to cleaning the bot-
tom of a swimming pool. Many surgeons called this “passive
egress,” which is incorrect; IOP can produce a substantial
pressure difference across the port relative to atmospheric
pressure, which is no safer than so-called active suction.
Conor O’Malley soon thereafter suggested that fluid egress
Fig ure 6 .1 5 ■ Straight soft-tip extrusion cannula.
Fig u re 6 .1 4 ■ FAX req uires a
soft-tip or altern ative extrusion
cannula to rem ove fluid while air is
injected through the infusion can-
nula powered by the low-p ressure
air source on the console.
would be better controlled by using the console and foot
pedal, a technique he called “extrusion.” The author imme-
diately adopted extrusion because it prevented inadvertent
motion of the tip when covering or uncovering the port on
the handle and enabled more precise vacuum control as well
as foot pedal–controlled reflux.
Sandy Grizzard and Harry Flynn independently devel-
oped various flexible or soft-tip cannulas that are preferable
to rigid cannulas for all applications (Figs. 6.15 and 6.16).
Soft-tip cannulas are less likely to damage the retina, optic
nerve, RPE, or choroid due to patient movement or posi-
tioning errors caused by suboptimal visualization. Soft-tip
cannulas should be inserted through the retina break to
Fig ure 6 .1 6 ■ Angulated soft-tip extrusion cannula.

Fig u re 6 .1 7 ■ In tern al d rain -
age can be perform ed through a
p reexisting retinal b reak or m id-
peripheral drainage retinotom y.
a position just above the RPE to reduce the likelihood of
retinal incarceration in the port. Small-diameter, especially
25 gauge, soft-tip cannulas are very flexible and will flex to
conform to the RPE surface when inserted through a reti-
nal break or retinotomy. Soft-tip, 25-gauge cannulas always
provide sufficient flow rates when used with proportional
(linear) vacuum with a preset maximum of 650 mm Hg.
IDS should be initiated before starting FAX to prevent SRF
from being displaced posteriorly by the buoyancy of the air.
Internal drainage can be performed through a preexisting reti-
nal break or midperipheral drainage retinotomy well away from
the macula and vascular arcades (Fig. 6.17). Internal drainage
should be continued during FAX to prevent the posterior dis-
placement of subretinal fluid; the view is often transiently lost
until bubbles coalesce, necessitating very stable positioning of
the soft-tip cannula. Removal of the flap from flap tears not
only eliminates traction but also reduces the likelihood of reti-
nal incarceration in the soft-tip cannula. It is best to perform
a slow yet continuous drainage of the SRF. This promotes the
migration of all SRF toward the area of drainage and maxi-
mizes drainage efficiency. Fast aspiration leads to reattachment
of the area of drainage with peripheral areas of persistent SRF
that may not be contiguous with the area of drainage.
Dra ina ge Retinotomy
Although the senior author developed IDS and independently
codeveloped “relaxing” retinotomy or retinectomy, drainage
retinotomies were infrequently used for many years. More
recently, small drainage retinotomies have been utilized safely
and effectively when the tear could not be adequately visualized
or adequately reached with the drainage cannula. A small
CHAPTER 6 ■ General Posterior Segm ent Techniques 61
spot of diathermy can be made using the disposable bipolar
diathermy before making a small, round retinotomy with the
single cut mode of the vitreous cutter. The diathermy needle
can itself be used to punch through the retina while cauter-
izing, leading to a small, round hole without hemorrhage. The
hole should be made rather posterior, away from retinal ves-
sels but outside the temporal arcades. The soft-tip cannula can
then be used to drain all the SRF through the retinotomy.
Th e Re at t ach m e n t Exp e rim e n t
Internal (simultaneous) FAX combined with IDS is an excel-
lent test for relief of traction on the retina. This can be
thought of as the reattachment experiment. If during this
procedure the retina does not completely attach, further
vitreous removal, forceps membrane peeling, scissors seg-
mentation, scissors delamination, subretinal surgery, retine-
ctomy, or scleral buckling is necessary. Surface tension
stabilizes the retina and allows visualization of mechanical
factors preventing reattachment. If all these techniques fail,
this test serves to indicate inoperability.
If subretinal air appears, it is indicative of residual trac-
tion that must be treated as outlined above. Direct needle
drainage of SRF posterior to a scleral buckle tied after the
subretinal air appears will remove subretinal air, but a poste-
rior drainage retinotomy is usually a better choice.
Total intraoperative reattachment eliminates postopera-
tive questions concerning leakage through the tear or the rate
of RPE pumping of SRF. In addition, retinopexy performed after
reattachment allows for more precise localization and less risk
of RPE migration and permits treatment of the retina as well as
the RPE, ensuring better adherence. The guesswork method
62 SECTION II ■ Surgical Technology and Techniques
of using a small expanding bubble without completion of SRF
drainage leaves many questions unanswered at the time of vit-
rectomy and, consequently, lowers the success rate.
In t rao cular Le n s Fo g g in g Durin g
Fluid -Air Exch an g e
Fogging is not specific to silicone IOLs, although they have a
greater thermal mass than polymethyl methacrylate (PMMA)
or acrylic IOLs and have a higher posterior capsular opaci-
fication rate than acrylic IOLs. Fogging occurs when a YAG
laser capsulotomy has been performed and the cataract or
vitreous surgeon removes the AVC, enabling the warm air
bubble to come into contact with the relatively cooler IOL
causing condensation. PFO is ideal when fogging of the
IOL occurs (Fig. 6.18). If fogging occurs, the air should
be removed and the retina attached with PFO, followed by
endophotocoagulation around all retinal breaks, and finally
by placing the soft-tip cannula in the optic nerve cup to per-
form a PFO-gas exchange. Often, a few drops of PFO will
be retained if fogging occurs during the exchange, but these
can be removed at a later time if necessary. Alternatively, ASX
could be performed; condensation will disappear when the
oil reaches the posterior surface of the IOL, and IDS fol-
lowed by endophotocoagulation to all breaks can then be
completed under oil.
Air-Gas Exch an g e
AGX is performed after IDS combined with FAX is used
to attach the retina and endophotocoagulation has been
Fig u re 6 .1 8 ■ Fog g ing occurs
when the warm air bubble com es
into contact with th e relatively
cooler IOL, causing condensation.
completed around all retinal breaks (Fig. 6.19). It is essential
to use an isoexpansive gas concentration as developed by
Gary Abrams to prevent high IOP or undersize bubbles.
Sulfur hexafluoride (SF6) should be mixed with air to
achieve a 25% concentration; 18% is appropriate for C3F8. It
is dangerous to estimate the volume of the air bubble in the
vitreous cavity and inject a higher gas concentration. This
inaccurate method solves no problems and can result in very
high IOP and central retinal artery occlusion or inadequate
postoperative bubble size. Some surgeons use partial FAX
and higher gas concentrations, which are inaccurate and
prone to high IOP or inadequate bubble errors. The air-gas
concentration should be carefully determined; many errors
have occurred by confusing cubic centimeters with percent,
a potential disaster if a different size syringe is used than that
used when the percent calculation was done.
Liq uid Pe rfl uo ro carb o n Te ch n iq ue s
Liquid perfluorocarbon (perfluorooctane, PFO) can be used
to remove SRF in cases without PVR or giant breaks. There
are several pros and cons of using PFO compared to IDS,
although the techniques can be combined. PFO should be
injected slowly, near the retinal surface, preferably into the
optic nerve cup or nasally to prevent damage to the macula.
The MedOne 25-gauge, dual-bore PFO injection cannula per-
mits fluid egress as the PFO is injected to maintain a normal
IOP (Fig. 6.20). The cannula must be retracted as the PFO
level rises to avoid loss of PFO through the fluid egress port.
PFO is best injected by placing a full 7.5-mL vial in
Alcon VFC MedOne dual-bore cannula.
 CHAPTER 6 ■ General Posterior Segm ent Techniques 63

Fig u re 6 .1 9 ■ AGX p rod uces

an isoexp ansive concentration of
g as, sig n ificantly red ucing p ost-
operative variation in bubble size
and elevated IOP.

The VFC system from the Accurus and Constellation
systems is ideal to control infusion of PFO into the eye. The
surgeon should be cautious to keep the maximum infusion
pressure set at 10 mm Hg and should continuously visual-
ize the fundus during the VFI injection of PFO. A similar
technique can be used to inject tissue plasminogen activa-
tor into the subretinal space in cases of massive subretinal
hemorrhage.
SRF and infusion fluid float anteriorly when PFO is
injected into the vitreous cavity, which often results in SRF
being displaced anterior to the most anterior retinal break
unless a giant retinal break is present. Anterior displacement
of SRF can be managed by extending a retinal break to the
ora with scissors or the vitreous cutter, making a very small
drainage retinotomy right at the ora, or by inserting a soft-tip
cannula through a retinal break and carefully removing the
SRF without removing PFO.

Fig u re 6 .2 0 ■ Th e d ual-b ore

cannula m ust be retracted as the
PFO level rises to avoid loss of PFO
through the fluid egress port.
64 SECTION II ■ Surgical Technology and Techniques
Removal of the PFO should be made carefully, under
air infusion, to prevent any layer of aqueous fluid (BSS, SRF)
that may be above the PFO to enter the subretinal space.
Sometimes, a small amount of SRF can become trapped in a
doughnut configuration anterior to the retinal tear and may
be difficult to perceive until the PFO is removed and the
surgeon notices the presence of posterior SRF. To achieve
complete removal of PFO and SRF, the removal of PFO
should begin anteriorly, trying to remove any amount of
BSS above the PFO meniscus. The tip of the aspiration can-
nula should slowly be moved to the location of the retinal
tear, and aspiration should be continued until the PFO-air
meniscus is clearly observed. If any SRF above the PFO is
present, the surgeon can remove it at this time immediately
above the PFO meniscus through the retinal tear. This is
critical in giant retinal tear cases to prevent slippage. Once
the PFO-air meniscus has been identified and the presence
of BSS or SRF has been ruled out or removed, the surgeon
can proceed to drain the rest of the PFO at the level of the
optic nerve head.
Me d ium -Te rm Pe rfl uo ro n
The principal author has used medium-term PFO to repair
selected inferior retinal detachments and giant retinal breaks
since PFO became available. The technique allows the patient
to stand, sit, fly, and even work and drive if there is good
vision in the other eye. The technique involves completing
the vitrectomy and removing all traction followed by attach-
ing the retina with PFO and then endophotocoagulation sur-
rounding all retinal breaks and suspicious areas. A total PFO
Fig ure 6 .2 1 ■ Triam cinolone is
injected into the air bubble, which
reduces the foreign body reaction
that can occur with PFO.
fill is utilized and air turned on after total PFO fill; a very
thin air bubble at the pars plana prevents PFO loss while
removing the cannulas. Triamcinolone 4 mgm (Alcon Tries-
ence, preservative free) is then injected into the air bubble
and the cannulas are removed (Fig. 6.21). Use of triamcino-
lone reduces the foreign body reaction that can occur with
PFO left in the eye for 14 to 16 days until the laser treatment
areas are pigmented. Triamcinolone is removed when the
PFO is removed at 14 to 16 days using the 25-gauge, three-
port technique; removal of triamcinolone prevents steroid
glaucoma and cataract.
Gian t Re t in al Bre aks
David Wong emphasizes that a full fill with PFO with the
infusion fluid turned off prevents posterior slippage of the
giant break. Exchange of PFO for air to enable AGX or ASX
must be done precisely to prevent slippage; the soft-tip can-
nula for PFO removal must be positioned above the PFO-
fluid interface just below the air-fluid interface at all times
so that all BSS (containing SRF) is removed before any PFO.
Obviously the cannula must be slowly advanced posteriorly
as the exchange is done.
Silico n e In fusio n
ASX is preferable to fluid-silicone exchange because air
helps keep silicone out of the anterior chamber. The prin-
cipal author uses 25-gauge, transconjunctival vitrec-
tomy in all cases and performs ASX by injecting 1,000
cs oil through one cannula with a Alcon cannula while
 CHAPTER 6 ■ General Posterior Segm ent Techniques 65

Fig ure 6 .2 2 ■ Following FAX and IDS and usually retinop exy, a silicone injector such as the
Alcon VFC is used to inject silicone oil through a short, thin-wall cannula while air through the
infusion cannula m aintains IOP and air egress occurs with an extrusion cannula behind the lens/
IOL or in the anterior cham ber (A/ C) of aphakic eyes. Injection is stop ped and the infusion line
clam ped with a hem ostat when silicone enters the infusion cannula.

allowing air to escape through the other, open cannula
(Fig. 6.22). The air infusion cannula line is clamped when
oil refluxes up the infusion tubing and the oil injection is
continued, carefully monitoring the IOP until all the air
comes out the open cannula in phakic or IOL eyes. Silicone
oil is filled up to the pupillary plane in aphakic eyes. Air
egress is made possible in phakic eyes and IOL eyes by posi-
tioning the eye and patient’s head so that the open cannula is
at the highest point, although an air bubble can be removed
by extrusion with the soft-tip cannula if care is taken to stay
out of the oil.
Silico n e Re m o val
Silicone oil should not be removed unless the retina is 100%
attached, all retinal breaks are surrounded by confluent laser,
and, preferably, three rows of 360-degree laser have been
applied. It is not necessary to remove oil in older patients,
especially if an PC is present and silicone does not enter
the anterior chamber. The authors use 25-gauge, three-port
technique to remove silicone oil by placing a 0.25-inch-
long segment of silicone tubing (fragmenter tubing) on the
Alcon Constellation VFC and placing it over the outside of
one of the cannula hubs (Fig. 6.23). The infusion maintains
the IOP, and the other cannula is left open to allow addi-
tional egress of small silicone bubbles. Complete silicone oil
removal is made possible by positioning the eye and patient’s
head so that the cannula used for extraction is at the highest
point. Repeated fluid-air-fluid exchanges may be needed to
displace any silicone oil that may be adherent to the zonules.
This technique allows complete removal of all silicone oil
through the 25-gauge cannulas.
Silico n e Re o p e rat io n s
The principal author has reoperated all silicone cases for
EMM or redetachment without removing the oil for over
25 years. Before 25-gauge vitrectomy was developed, the
author used a short, silicone infusion tubing–cannula device
attached to the Alcon Constellation VFC to infuse silicone oil
as SRF is removed. The principal author developed another
method that works well with 25-gauge silicone reopera-
tions, a two-port, sequential, inject-oil-and-remove-SRF
method using the Alcon 25G silicone injection cannula. The
techniques utilized for surgery without removing the oil will
be described in the next section.
In t e rface Vit re ct o m y
The principal author developed a concept known as inter-
face vitrectomy that is based on performing a variety of tech-
niques with air, PFO, or silicone oil in the eye. Vitrectomy,
forceps membrane peeling, scissors segmentation or delam-
ination, diathermy, and retinectomy all work well with
66 SECTION II ■ Surgical Technology and Techniques
these agents in the eye. The term “interface” was selected to
emphasize that the vitreous cutter must be placed outside
the air, PFO, or silicone oil bubble. Air and silicone oil float
in infusion fluid but PFO sinks; it is crucial to understand
that there is always a fluid layer at the retinal surface and
the retina is over 90% water as well as immiscible in air,
PFO, or oil. The principal author introduced the concept
Fig u re 6 .2 3 ■ Silicone oil can
be com pletely rem oved by attach-
ing a 0.25-inch-long segm ent of
silicone tubing to the Alcon Con-
stellation VFC and p lacing it over
the outside of one of the cannula
hubs.
of endophotocoagulation “under” air over three decades
ago, and Stanley Chang introduced membrane peeling
“under” PFO for PVR cases many years ago as well. Inter-
face vitrectomy prevents an increase of SRF during traction
removal, allows realistic assessment of remaining traction,
confines bleeding to the interface, and stabilizes the retina.
Air stabilizes the retina by spring dampening (Fig. 6.24),
Fig ure 6 .2 4 ■ Air stabilizes the
retina by spring dam pening.
 CHAPTER 6 ■ General Posterior Segm ent Techniques 67

Fig u re 6 . 2 5 ■ Silico n e o il
stab ilizes th e retin a b y viscous
dam pening.

silicone oil by viscous dampening (Fig. 6.25), and PFO by
inertial (F = MA) and gravitational (high specific gravity)
effects (Fig. 6.26). The author uses 25-gauge vitrectomy for
these cases and is very careful to avoid applying vacuum
while the port is in silicone oil. Moderate amounts of subreti-
nal silicone can be removed in conjunction with retinectomy
and ERM removal without plugging the vitreous cutter by
using 600–650mm Hg vacuum.
Prior to the development of FAX by the senior author
(30), fluid was withdrawn through a single needle, which was
then used to inject gas into the collapsed eye after turning a
stopcock. AGX is usually best performed at the end of the case,

Fig u re 6 .2 6 ■ PFO stab ilizes

th e retin a b y in ertial (F = MA)
an d g ravitation al (h ig h sp ecific
gravity) effects.
68 SECTION II ■ Surgical Technology and Techniques
after FAX and retinopexy. Similarly, most surgeons perform
FAX and retinopexy before ASX. Some surgeons recommend
fluid-silicone exchange and omit the FAX step. The fluid-
silicone exchange approach can make visualization as well as
complete reattachment for retinopexy more difficult.
Fluid-perfluorocarbon exchange is preferred to FAX
before perfluorocarbon placement in retinal detachment,
PVR, and giant break cases. PGX or PSX is usually performed
after laser retinopexy.
The authors never remove silicone oil when perform-
ing reoperations as forceps membrane peeling, scissors
segmentation or delamination, retinectomy, laser, etc. are
effective “under” silicone (Fig. 6.27). Silicone is infused to
replace silicone lost through the sclerotomies. The suction
(vacuum) is used at the highest setting (600–650mm Hg),
and scissors segmentation or delamination, forceps mem-
brane peeling, retinectomy, subretinal surgery, and endo-
photocoagulation work in the usual manner.
Vit re ct o m y Un d e r Air
Certain portions of the vitreous surgery procedure can
be completed after FAX (31). At times, continuous severe
bleeding cannot be controlled with combined extrusion and
bipolar diathermy or endophotocoagulation. If internal FAX
is performed in such an instance, the bubble will confine
the blood to a small space so that overall visualization is
improved and diathermy or endophotocoagulation of the
vessel can be completed. The air-vitreous interface is quite
visible, which facilitates the removal of residual vitreous by
placing the cutter port barely into the vitreous.
Fig u re 6 .2 7 ■ Forcep s m em -
brane peeling, scissors segm enta-
tion or delam ination, retinectom y,
laser, etc. can be perform ed with-
out rem oving silicone oil.
The air-vitreous interface facilitates visualization of
residual vitreous after an exchange is performed. The senior
author often places the vitreous cutter port just posterior to
the air-vitreous interface and performs additional vitrectomy
under air (Fig. 6.28). This can be done with PFO and sili-
cone oil as well, which is termed “interface vitrectomy.”
Forceps membrane peeling, scissors segmentation and
delamination, subretinal surgery, retinectomy, laser endopho-
tocoagulation, and foreign body removal can all be performed
under air if it is not an aphakic eye with striate keratopathy.
Fogging of IOLs if a YAG capsulotomy has been performed
is a serious limitation of this method. The authors used to
refrigerate the BSS Plus prior to all vitrectomies except those
with IOLs because hypothermia reduces inflammation, isch-
emic damage, endothelial damage, iris bleeding, and light
toxicity, but low fluid volumes associated with current vitrec-
tomy technology dramatically reduce the advantages. Silicone
lenses fog much more than acrylic or PMMA lenses because
they have a higher thermal inertia. Prior YAG capsulectomy
and intraoperative removal of the AVC create the fogging
problem by allowing air access to the IOL surface. IOLs fog
because they have significant thermal inertia and are cooled
by room temperature infusion fluid, and the air in the eye
is saturated with water vapor. Viscoelastics can be injected
against the posterior surface of the lens to reduce fogging, but
this creates a morphed image, increases cost, and increases
silicone emulsification. The authors use a soft-tip cannula as
a windshield wiper until the procedure is completed or the
fogging abates. PFC liquids such as PFO completely elimi-
nate this problem by facilitating removal of SRF and perform-
ing endophotocoagulation before PFO-gas exchange.

Fig ure 6 .2 8 ■ Th e air-vitreous
in terface facilitates visualization
and rem oval of resid ual vitreous
after an exchange is perform ed.
Po st o p e rat ive Po sit io n in g
Postoperative positioning requirements will depend on the
location of the tears being treated and lens status. A phakic
patient can develop a gas cataract from contact of the gas to
the posterior lens surface. While this cataract can be transi-
tory, once it persists for longer than 48 hours, it may become
permanent. The first requirement is that the location of the
treated tear is at the highest position in the eye to allow con-
tact with the gas. Once the laser retinopexy becomes pig-
mented, further head positioning is not required to keep the
retina attached. Once this occurs on a pseudophakic patient,
strict head positioning can be relieved. If the patient is pha-
kic, forward head tilt (depending on the remaining size of
the bubble) is required to keep the gas away from the poste-
rior lens surface.
Making an analogy to a “cast on a broken arm” can
enhance patient acceptance. Phrases such as “put the bub-
ble on the trouble” and “longer duration bubbles result in
higher success rates” help the patient gain understand-
ing. Numerous pillows facilitate comfortable positioning.
A comparison can be made to sleeping with crossed arms
on a school desk and a table at home used with a pil-
low for padding. Wingback chairs and sofas are similar
to the car, bus, or airplane seating position and can be
used to explain the required position. Some surgeons
recommend expensive special chairs or support systems,
which the authors believe to be unnecessary in most
cases. These devices require immobility, which increases
muscle cramping, psychological stress, and deep venous
thrombosis.
CHAPTER 6 ■ General Posterior Segm ent Techniques 69
The reduced atmospheric pressures intrinsic to air
travel or rapid ground travel to higher altitudes dictate
marked caution when a bubble is present. Bubble expansion
from air travel causes a severe increase in IOP and can cause
vascular occlusion (32,33). Air travel is best prohibited while
an air or gas bubble is present, although clinically it has been
observed that a bubble less than 10% of ocular volume can
be tolerated. Ground travel should be used to prevent this
severe problem. Patients should be cautioned to avoid rapid
ground travel from low to high altitudes.
Sub re t in al Fluid Re m o val
Indica tions for Dra ina ge of Subretina l Fluid
The debate over the necessity of drainage of SRF in sclera
buckling has raged for years and will not be settled by this
book. Drainage of SRF is essential in difficult retinal detach-
ment cases because it instantaneously determines the need
for subsequent surgical steps. It is not always possible for
even the most experienced surgeons to correctly determine
the exact steps to accomplish retinal reattachment from the
preoperative examination. The principal author refers to SRF
drainage with FAX as the “reattachment experiment” because
it frequently helps determine the need for further vitreous
removal, peeling, segmentation, delamination, retinectomy,
or scleral buckling.
Removal of virtually all SRF before fluid-air-gas or
fluid-air-silicone exchange increases the likelihood of injec-
tion of the correct volume of isoexpansive gas mixtures
or silicone oil. It is virtually impossible for the surgeon to
70 SECTION II ■ Surgical Technology and Techniques
accurately estimate the volume of the subretinal space or
vitreous cavity. Expanding gas mixtures should not be used
with total fluid-gas exchange because of the risk of central
retinal artery occlusion due to high pressure. Similarly, if the
volume of isoexpansive gas or silicone oil required is over-
estimated, central retinal artery occlusion can result. If the
volume of gas or oil is underestimated, the area of contact
with the surface tension management agent may be insuf-
ficient, causing the procedure to fail.
The principal author introduced the concept of internal
(transretinal) drainage of SRF and postreattachment retin-
opexy in 1973. Prior to that time, retinopexy was performed
before the vitrectomy and expanding gas bubbles were used
to slowly reattach the retina postoperatively. Postreattach-
ment retinopexy probably reduces the dispersion of retinal
and RPE cells, potentially reducing PVR recurrences. Postre-
attachment retinopexy allows accurately placed, controlled
intensity retinopexy to the retina as well as the RPE. Over-
treatment is a significant factor in PVR recurrences and fibrin
syndrome.
Internal drainage methods and exchanges allow the
benefits of postreattachment retinopexy; accurate air, gas,
and silicone volumes; and the reattachment experiment to
be used on a routine basis.
Externa l Versus Interna l Dra ina ge
External drainage of SRF requires an opening through the
sclera, choroid, and RPE. By contrast, internal or transretinal
drainage requires no eye wall incision. Internal drainage
enables complete drainage of the SRF, prevents incarceration
of retina in the sclera, prevents choroidal bleeding, reduces
RPE damage, and is the technique that should be used dur-
ing pars plana vitrectomy.
Direct Needle Tra nssclera l Dra ina ge of Subretina l
Fluid for Sclera l Buckling
The senior author developed direct (transscleral) needle
drainage of SRF to reduce the complications of the scleral
cutdown method during scleral buckling. Direct needle
drainage has been shown in a randomized trail (Cairns) to
eliminate the problem of retinal incarceration in the drain
site (>3% with the scleral cutdown method), allow more
complete drainage, and reduce hypotony, without adding
any complications. Direct needle drainage of SRF can be
performed through the conjunctiva, a buckle, or a choroi-
dal detachment (edema). This method is performed using a
25- to 27-gauge, 0.5-inch, disposable needle on a tubercu-
lin syringe with the plunger removed. A transparent needle
hub is preferred because it allows early visualization of fluid
egress (34). A site is selected that has the highest retinal
detachment by viewing through the operating microscope
using the endoilluminator and contact lens. Indirect oph-
thalmoscopy can be used if an operating microscope is not
being utilized. The bevel of the needle is always oriented
away from the retina to prevent retinal entrapment. The tip
of the needle is kept away from the sclera by intentional
“shanking” until the entry point is located. The needle is then
oriented in a more tangential manner and advanced slowly
through the sclera until a slight proprioceptive “pop” is felt
similar to venipuncture (Fig. 6.29). At this point, a brown
Fig u re 6 .2 9 ■ Direct n eed le
d rain ag e of SRF is in itiated b y
inserting 25- to 27-gauge, 0.5-inch
needle (attached to a 1-m L syringe
with p lung er rem oved ) throug h
the sclera slightly anterior to the
g reatest h eig h t of th e retin al
d etach m en t. A p ro p rio cep tive
“pop” will be felt when the needle
is through the sclera.
 CHAPTER 6 ■ General Posterior Segm ent Techniques 71

Fig u re 6 .3 0 ■ Th e n eed le is
ad van ced th roug h the ch oroid
an d RPE un til it can b e seen
under the retina. A brown “pencil
p oint” of RPE tissue can be seen
just before the RPE is penetrated.
Rap id , sm all-am p litu d e m o ve-
m ent of the retina and shifting of
SRF toward the n eed le ind icate
that flow is occurring.

pencil point–like structure may be seen in the subretinal
space, which is the RPE stretched over the needle tip. The
needle is then advanced slightly under direct visualization
through the RPE until the silvery sheen of the needle is seen
under the retina (Fig. 6.30). At this point, the retina begins to
tremble due to turbulence in the needle lumen. The needle
cannot be visualized if the SRF is turbid, but the trembling
of the retina will indicate fluid egress. The detachment may
increase near the drainage site as the pseudoplastic mass
of SRF shifts toward the egress lumen. The needle should
be angulated slightly until it is tangential to the retina after
the initial placement. The needle should be held immobile
until all the SRF is gone (Fig. 6.31). Pigment typically comes
through the needle hub as the last fluid drains.

Fig u re 6 . 3 1 ■ Th e n e e d le
sh ould b e h eld im m ob ile un til
all SRF is d rained and the retina
drapes over the needle; pigm ent
often ap p ears in SRF in the clear
n eed le h u b wh en d rain ag e is
com p leted.
72 SECTION II ■ Surgical Technology and Techniques
ENDODIATHERMY
Although neovascularization (NVE) is frequently encountered
in diabetic vitrectomy, intraoperative bleeding is seldom of
great consequence. Because endocoagulation can cause tis-
sue shrinkage, retinal breaks, nerve fiber layer damage, and
optic atrophy, it should be reserved for large vessels that
are to be transected or that actually bleed during surgery.
When extensive networks of NVE are associated with ERM,
it is best to control bleeding during the dissection by using
transient elevation of IOP. Vascular attachment points can
then be treated as needed with endocoagulation. Frequently,
vascularized ERM has only a few connections to the retinal
arterial system, which are only recognized and treated after
delamination of the ERM. Use of preoperative bevacizumab
in proliferative retinopathies has also decreased the problem
of intraoperative hemorrhage.
Every attempt should be made to prevent and control
bleeding; blood can form a substrate and stimulus for post-
operative cellular proliferation. As it is not always possible
to safely control bleeding, it is fortunate that the aphakic,
vitrectomized eye rapidly clears postoperative hemorrhage.
Elevated new vessels are best coagulated with diathermy.
Unipolar diathermy, which was used originally, has been
supplanted by bipolar diathermy because of the latter’s
greater safety (35,36). If unipolar diathermy is applied near
the optic nerve, radio frequency energy can travel through
the optic nerve, causing optic atrophy and permanent blind-
ness. Unimanual bipolar diathermy (UBD) functions well if
care is taken to maintain at least 1 mm distance from the
optic nerve. By controlling bleeding with transient increases
in IOP, UBD can be very effective in treating vascular attach-
ment points.
Bimanual bipolar diathermy (BBD) (37) can function
through any two intraocular instruments or one intraocular
and one extraocular instrument by the simple attachment of
unshielded wires and clips. Insulated microclips or connec-
tors on the instrument handles can be used for BBD (Charles).
BBD clips are ideal for 25-gauge surgery. This method has
been virtually supplanted by the disposable bipolar endoil-
luminator for 20-gauge surgery, which combines diathermy
with the endoilluminator. This device is ideal for coagulat-
ing vascular attachment points during delamination and seg-
mentation of ERM. The disposable bipolar endoilluminator
is typically utilized in conjunction with the vitreous cutter
or fine curved scissors. It is not necessary to interrupt the
procedure in order to place a specialized instrument in the
eye, and bleeding can be washed away continuously using
the cutter to permit excellent visualization of the bleeding
vessel during dissection. By intentionally allowing the IOP
to become very low, bleeding can be stimulated so that its
source can be identified and coagulated. Although bipolar
diathermy can be used for retinopexy, there is little effect
at the RPE level. Endophotocoagulation is better suited to
retinopexy than is endodiathermy.
Lase r En d o p h o t o co ag ulat io n
Xenon arc light energy can be delivered to the retinal surface
with the endophotocoagulator (37). The prototype was
an adapter to the Zeiss xenon photocoagulator. The first
commercial endophotocoagulator was an adapter to the
Clinitex Log III portable xenon arc photocoagulator (38).
A xenon endophotocoagulator has high beam divergence
and reliability problems and is no longer used by most
surgeons.
Laser endophotocoagulation was developed later
(Peyman, Fleischman, Charles, and Landers; all indepen-
dently and simultaneously) and is preferable to xenon
because of less beam divergence, better reliability, and more
rapid firing. Virtually all compact, laser photocoagulator sys-
tems can be utilized for endophotocoagulation.
Near IR (infrared) diode lasers (Tano) are 20% to 30%
efficient in converting electrical to light energy in contrast to
ion lasers (argon) at 0.0001% or less. For this reason, they
use standard, single-phase electrical power instead of high-
current, three-phase power. Because of their efficiency, they
are far more reliable, are more compact, and do not require
internal or external water cooling. The near IR wavelength
is the major disadvantage of these lasers because it is much
more difficult to judge the correct power threshold and the
lesion is deeper.
Diode-pumped frequency up-converted, continuous,
532-nm lasers are ideal for all operating room and most
office photocoagulation. They have all the above described
advantages of diode laser but use an ideal wavelength for
hemoglobin absorption and xanthophyll avoidance.
The endophotocoagulator is used primarily for the
treatment of bleeding from surface NVE, for retinopexy, and
for PRP. Endophotocoagulation should be used for surface
bleeding from specific sites, usually after scissors delamina-
tion of ERMs. When using the endophotocoagulator, the
endoilluminator is usually held in the opposite hand to pro-
vide diffuse illumination. 20-gauge illuminating laser probes
(Chang) were developed to solve this problem. When there
is acute bleeding, it is recommended that the suction can-
nula or vitreous cutter be alternated with the laser probe
to remove blood and facilitate precise coagulation. The
20-gauge aspirating endophotocoagulator probe (Chang)
can be used to aspirate SRF during endophotocoagulation as
reported by the principal author.
The endophotocoagulator is well suited for PRP con-
comitant with vitreous surgery for diabetic retinopathy,
venous occlusive disease, hemoglobinopathies, and telangi-
ectasia. As discussed in the chapter on diabetic retinopathy,
endo-PRP can bring about a decreased incidence of neovas-
cular glaucoma, AVC fibrovascular proliferation, and recur-
rent flat NVE leading to postoperative bleeding. The only
limitation of endophotocoagulation is that elevated retina
cannot be treated.
In contrast to endocryopexy, endophotocoagulation is
a noncontact method. Therefore, dispersion of RPE cells,

choroidal bleeding, retinal tears, and the increased wound
size associated with endocryopexy are avoided. Endophoto-
coagulation lesions are between 600 and 1,000 mm in diam-
eter, depending on the distance from the tip of the probe
to the retinal surface, the beam divergence, and the power
setting. If the retina is detached, endophotocoagulation
must be preceded by FAX and IDS, which bring the retina
and RPE into contact to permit energy absorption. For focal
treatment of retinal breaks, the continuous mode is utilized
to treat in a confluent manner (painting) around the breaks.
This technique minimizes the possibility of undertreat-
ment or overtreatment, which frequently occurs with the
placement of discrete photocoagulation spots in rows. Peri-
odically during treatment, small amounts of SRF will shift
posteriorly, making repeated internal drainage necessary to
permit retinopexy.
The endophotocoagulator should never be utilized in
air (gas) when there is blood on its tip, or damage to the
probe may result. PRP under air to areas of retina that had
been detached before surgery is a common cause of fibrin
syndrome because of overtreatment.
A microscope filter attenuates the laser energy in the
optic path of the operating microscope during endophotoco-
agulation and ensures continued dark adaptation and safety.
The fovea and optic nerve must be avoided, and all treat-
ment must be initiated at low energy with stepwise increases
until the desired effect is achieved.
The endophotocoagulator can be used to dilate the
pupil by treating the posterior surface or margin of the iris
sphincter. Endophotocorepexy is of special use when iris
NVE prohibits sphincterotomy with the vitrectomy
instrument.
The ciliary processes can be endophotocoagulated in
glaucoma cases, with subsequent decrease in pressure. The
senior author first reported endocyclophotocoagulation but
is concerned that it is significantly overutilized by cataract
surgeons in conjunction with phacoemulsification. Many
glaucoma experts share this concern based on the observa-
tion that the apparent benefits are short term and postopera-
tive CME increases.
RETINOPEXY
All retinal breaks, except macular holes, peripapillary
breaks, and retinotomies for submacular surgery, should be
treated with some form of retinopexy. This policy is neces-
sary because of the impossibility of predicting which retinal
break will result in detachment, and it is justified because of
the relative safety of retinopexy.
Although it may be contrary to the tendency of the
scleral buckling surgeon, retinopexy should be used only
after vitrectomy, surgical dissection, IDS, FAX, and comple-
tion of SRF drainage have reattached the retina. Performed
at the beginning of the operation, retinopexy may cause
CHAPTER 6 ■ General Posterior Segm ent Techniques 73
choroidal hemorrhage due to the combination of engorged
choroidal vessels. Postreattachment retinopexy will ensure
better visualization, so that all breaks can be identified and
iatrogenic breaks can be treated. Completion of internal
fluid-gas exchange and IDS will confine any RPE cells mobi-
lized by retinopexy to the area of the break and possibly
decrease the incidence of PVR.
Transscleral cryopexy in vitrectomy has long been vir-
tually replaced by laser endophotocoagulation and laser
indirect ophthalmoscope (LIO). Cryopexy disperses live
RPE and glial cells and causes more inflammation and PVR.
Endocryopexy is an unsafe, contact-based method that
requires an enlarged opening and is fortunately no longer
used by most surgeons. This method causes increased cel-
lular proliferation and choroidal hemorrhages. As discussed
previously, endophotocoagulation now performs the tasks
formerly handled by endocryopexy. Endophotocoagulation
is used far more commonly than transscleral retinopexy.
Transscleral retinopexy is used primarily for scleral buckling
without vitrectomy.
Tran sscle ral Diat h e rm y
Transscleral diathermy probably creates less proliferation
than cryopexy but causes more scleral damage. It can be used
with fiberoptic transillumination to find and treat any breaks,
except in the macular and peripapillary areas. The small
probe fits under buckles more easily than a cryoprobe.
Because of the thermal and electrical insulating proper-
ties of an air (gas) bubble, diathermy causes larger retinal
lesions and less scleral damage in an air (gas)-filled eye than
in a fluid-filled eye. Endophotocoagulation is preferred in
virtually all instances.
INTRAOPERATIVE FLUORESCEIN
ANGIOSCOPY
Intravenous sodium fluorescein dye (3.0 mL of 25%) can be
administered during vitreous surgery after the retina is visu-
alized (31). The dye can then be made to fluoresce by put-
ting a fluorescent interference-type exciter filter in the path
of the endoilluminator light source. The sources of bleeding,
especially in reoperation, can be readily identified.
In je ct io n o f An t ib io t ics an d St e ro id s
Subconjunctival antibodies should be injected in all patients
at the end of surgery. A 30- or 32-gauge needle can be used
to inject the antibiotics in the inferior cul-de-sac. This tech-
nique has low risk of sclera perforation. Antibiotics effective
against Gram-positive penicillinase producers and Gram-
negative organisms should be used (the authors currently use
a combination of cefazolin and tobramycin and replace the
cefazolin with vancomycin in penicillin-allergic patients).
74 SECTION II ■ Surgical Technology and Techniques
This use of antibiotics is necessary with the extensive
amount of tubing connections, instrumentation, and irrigat-
ing fluid utilized in vitreous surgery. It is thought, however,
that the myelinolytic effect of aminoglycoside antibiotics
may contribute to an occasional transient neuritis-like facial
pain after surgery. This lasts several weeks, can be managed
medically, and always disappears. Most “scleral buckle pain”
is probably caused by this mechanism.
A subconjunctival steroid (triamcinolone acetonide
20 mg in 0.5 mL or dexamethasone 4 mg) should be injected
in every case unless the patient is an infant, has a macular
hole, or is known to be a steroid glaucoma responder. Perio-
cular injections are more effective than systemic or topical
steroids and are medically safer than systemic steroids. Intra-
vitreal steroids can be used in conjunction with vitrectomy.
The duration of action is shorter (~1 month) than in the
nonvitrectomized eye (3 months). There is a 30% to 50%
risk of steroid glaucoma and a significant cataract risk, and
nonclosure of macular holes has been reported.
BUCKLING ANTICIPATED OR
PREVIOUS BUCKLE PRESENT
The operating microscope is useful for all scleral buckling
procedures, even if vitrectomy is not planned. The authors do
not currently perform vitrectomy buckles; primary vitrectomy
produces better outcomes with less complications and shorter
operating times. If, on the other hand, a decision is made to
perform combined vitrectomy-buckle, conjunctival incision
should be performed to expose the sclera as explained in the
sclera buckling chapter. Better visualization reduces inadver-
tent scleral perforation and damage to the muscles and ocular
blood supply and improves conjunctival closure and patient
comfort. Practice with the microscope improves surgical skills.
If an existing buckle is to be revised or removed, it should be
exposed under the microscope by dissection with a hockey
stick–type blade. A small suction cannula can be used to aspi-
rate routine bleeding, which usually subsides with minimal
diathermy. Bipolar diathermy coagulation should be used
only for major bleeding as these techniques increase conjunc-
tival and episcleral scarring if used excessively. Bipolar forceps
produce more focused diathermy than the “eraser.”
Excellent exposure can be obtained by using a microret-
inal retractor without a slot. The slot on the Schepens retrac-
tor allows prolapse of orbital tissues into the surgical field.
The assistant’s hand is kept out of the surgical field because of
the right angulated handle and can exert a lateral pull rather
than a twisting force, resulting in reduced hand fatigue. If
scleral buckling elements are to be removed, it is best to
place the infusion cannula, verify placement through the
microscope, and turn on the infusion before buckle removal
to prevent hypotony. Care must be taken when rotating the
eye temporally and inferiorly so as not to rotate the interior
portion of the infusion cannula posteriorly into the periph-
eral retina when the cannula contacts the lid.
SUMMARY
This chapter provides a discussion of the basic building
blocks of successful posterior vitrectomy. Each step must be
understood, and the proper equipment must be available
so that complications or unexpected findings can be man-
aged safely, rapidly, and effectively. Even the infrequent
vitreous surgeon who tries to limit himself or herself to
“easier” cases can find the unexpected and is obliged to use
these “advanced” techniques to safely complete the surgery.
Implicit in understanding the building blocks is using them
in the appropriate sequence.
References
1. Edelhauser HF, Hyndiuk RA, Zeeb A, Schultz RO. Cor-
neal edema and the use of epinephrine. Am J Ophthalmol
1982;93(3):327–333.
2. Kunimoto DY, Kaiser RS, Wills Eye Retina Service. Incidence
of endophthalmitis after 20- and 25-gauge vitrectomy. Oph-
thalmology 2007;114:2133–2137. [Epub 2007 Oct 4.]
3. Martidis A, Chang TS. Sutureless 25-gauge vitrectomy: risky or
rewarding? Ophthalmology 2007;114:2131–2132.
4. Rizzo S, Genovesi-Ebert F, Vento A, et al. Modified incision in
25-gauge vitrectomy in the creation of a tunneled airtight scle-
rotomy: an ultrabiomicroscopic study. Graefes Arch Clin Exp
Ophthalmol 2007;245:1281–1288. [Epub 2007 Feb 21.]
5. Mason JO III, Yunker JJ, Vail RS, et al. Incidence of endophthal-
mitis following 20-gauge and 25-gauge vitrectomy. Retina
2008;Jul 28.
6. López-Guajardo L, Vleming-Pinilla E, Pareja-Esteban J, Teus-
Guezala MA. Ultrasound biomicroscopy study of direct and
oblique 25-gauge vitrectomy sclerotomies. Am J Ophthalmol
2007;143:881–883.
7. Inoue M, Shinoda K, Shinoda H, et al. Two-step oblique inci-
sion during 25-gauge vitrectomy reduces incidence of postop-
erative hypotony. Clin Exp Ophthalmol 2007;35:693–696.
8. Asheesh T, Shah GK, Fang A. Visual outcomes with 23-gauge
transconjunctival sutureless vitrectomy. Retina 2008;28:258–262.
9. Gupta OP, Ho AC, Kaiser PK, et al. Short-term out-
comes of 23-gauge pars plana vitrectomy. Am J Ophthalmol
2008;146:193–197. [Epub 2008 Jun 10.]
10. Chen E. 25-gauge transconjunctival sutureless vitrectomy. Curr
Opin Ophthalmol 2007;18:188–193.
11. Landers MB. Removal of blood from the retinal surface in pars
plana vitrectomy. Am J Ophthalmol 1978;86:427.
12. Abrams GW, Williams GA. “En bloc” excision of diabetic mem-
branes. Am J Ophthalmol 1987;103(3):302–308.
13. Charles S (Developer). Suction forceps membrane peeling.
Presented at Wilmer Vitrectomy Course, The Johns Hopkins
School of Medicine, Baltimore, May 1976.
14. Charles S (Developer). Chopsticks membrane peeling. Pre-
sented at Wilmer Vitrectomy Course, The Johns Hopkins
School of Medicine, Baltimore, May 1976.
15. Wang CT, Charles S. Microsurgical instrumentation for vitrec-
tomy, Part 1. J Clin Eng 1983;8(4):321.
16. O’Malley C (Developer). Extrusion method. Ocutome Frag-
matome Newsletter 1978;3(1):3.
17. Fineberg E, Machemer R, Sullivan P. SF6 for retinal detach-
ment surgery, a preliminary report. Mod Probl Ophthalmol
1974;12:173.
18. Fineberg E, Machemer R, Sullivan P, et al. Sulfur hexafluoride
in owl monkey vitreous cavity. Am J Ophthalmol 1975;79:67.

19. Machemer R. Intravitreal injection of sulfur hexafluoride gas
(SF6). In: Freeman HM, Hirose T, Schepens CL, eds. Vitreous
surgery and advances in fundus diagnosis and treatment. New
York: Appleton-Century-Crofts, 1977:421–422.
20. Bourgeois JE, Machemer R. Results of sulfur hexafluoride gas
in vitreous surgery. Am J Ophthalmol 1983;96:405.
21. Killey F, Edelhauser H, Aaberg TM. The effects of intraocular
sulfur hexafluoride and freon gas on intraocular pressure and
vitreous volume. Arch Ophthalmol 1978;96:521.
22. Miller B, Lean JS, Miller H, Ryan SJ. Intravitreal expanding gas
bubble: a morphologic study in the rabbit eye. Arch Ophthalmol
1984;102:1708.
23. Chang S, Lincoff H, Coleman J, et al. Perfluorocarbon gases in
vitreous surgery. Ophthalmology 1985;92:651.
24. Lincoff A, Haft D, Liggett P, Reifer C. Intravitreal expansion of
perfluorocarbon bubbles. Arch Ophthalmol 1980;98(9):1646.
25. Lincoff A, Kreissig I. Intravitreal behavior of perfluorocarbons.
Surv Ophthalmol 1981;2:17.
26. Lincoff H, Coleman J, Kreissig J, et al. The perfluorocarbon
gases in the treatment of retinal detachment. Am Acad Ophthal-
mol 1983;90:546.
27. Lincoff H, Kressig J. Intravitreal behaviour of perfluorocarbons.
Dev Ophthalmol 1981;2:17.
28. Lincoff H, Mardirossian J, Lincoff A, et al. Intravitreal longevity of
three perfluorocarbon gases. Arch Ophthalmol 1980;98:1610.
CHAPTER 6 ■ General Posterior Segm ent Techniques 75
29. Lincoff A, Lincoff H, Iwamoto T, et al. Perfluoro-n-butane.
A gas for a maximum duration retinal tamponade. Arch Oph-
thalmol 1983;101:460.
30. Aaberg TM, Abrams GW, Edelhauser HF. Intraocular sulfur
hexafluoride: experimental and clinical correlation. In: Inter-
national Symposium on New and Controversial Aspects of
Vitreoretinal Surgery, Texas Medical Center, Houston, Texas,
St. Louis: C.V. Mosby, 1977:393–397.
31. S. Charles. unpublished data, 1974.
32. Fuller D. Flying and intraocular gas bubbles (letter). Am J
Ophthalmol 1981;91:276.
33. Dieckert JP, O’Connor PS, Schacklett DE, et al. The effects of
air travel on eyes with intraocular gas. Presented at the Annual
Meeting, American Academy of Ophthalmology, Atlanta, GA,
October 2, 1985.
34. Charles S. Controlled drainage of subretinal and choroidal
fluid. Retina 1985;5(4):233.
35. Machemer R. Letter to the editor. Am J Ophthalmol 1977;
83:282.
36. Schepens C. Letter to the editor. Am J Ophthalmol 1978;
85:574.
37. Charles S, White J, Dennison C, Eichenbaum D. Bipolar biman-
ual intraocular diathermy. Am J Ophthalmol 1976;81:101.
38. O’Malley P. Portable xenon arc light coagulator. Br J Ophthalmol
1973;57(12).
C H AP TER
7
VITRECTOMY TECHNIQUES AND
TECHNOLOGY FOR ANTERIOR
SEGMENT PROBLEMS
Although developed for posterior vitreoretinal
surgery, the vitrectomy instrument has wide-
spread application to anterior segment surgery
(1–8). Every anterior segment microsurgeon
must have vitrectomy techniques and equipment at his or her
fingertips to manage vitreous loss during cataract surgery and
other common problems. The anterior segment microsurgeon
with extensive knowledge of corneal physiology, aqueous
dynamics, intraocular lenses (IOLs), cataract complications,
and microsurgical techniques need not rely on a posterior
segment surgeon to manage all anterior segment vitreous
problems. Similarly, an anterior segment surgeon should not
delve into complex vitreoretinal problems just because the
technology is available. The demands are so complex that it is
a very unique individual who is an expert in both arenas. Per-
haps the most important issue to convey to anterior segment
surgeons is that proper management of vitreous is critical for
the prevention of postoperative retinal detachments.
VITRECTOMY INSTRUMENT CHOICE
The large size and fluid flow characteristics of full function
probes make them as inappropriate for anterior segment
surgery as they are for posterior segment surgery. Dispos-
able, small, lightweight, pneumatic, 23- or 25-gauge cutters,
high-speed cutting, and fast proportional (linear) suction
control are a necessity for anterior segment surgery, just as
they are for vitreoretinal procedures (Chapter 2).
INFUSION SYSTEMS FOR ANTERIOR
VITRECTOMY
Infusion can be provided by the infusion sleeve (9), which
slips over the vitrectomy instrument through a limbal
76
incision for a one-incision technique (Chapter 2). This
method is easy to use but results in a return to the turbulent
fluid flow observed with full function probes (Fig. 7.1). Use
of the infusion sleeve should be limited to vitrectomy in soft
eyes with no sideport, a rare situation.
A bimanual approach (10) with a blunt 30-degree bent
infusion cannula as the infusion device is used if the pars
plana cannot be seen, as well as for retinopathy of prema-
turity (ROP) (5). This technique allows the independently
movable infusion to hydrate and mobilize the lens material
for aspiration. By using 25-gauge infusion, the infusion nee-
dle can be interchanged with a cutter to provide better access
to the posterior chamber and periphery.
The authors encourage anterior segment surgeons to
develop bimanual surgery dexterity by using bimanual irri-
gation and aspiration techniques for lens cortex removal.
The surgeon can use the bimanual I/A irrigation tip during
anterior vitrectomy as well. If this instrument is used, the
decrease in caliber (and irrigation flow capacity) should be
compensated by decreasing the aspiration flow limit of the
vitrectomy system, otherwise collapse of the anterior cham-
ber can occur during surgery.
Unsupported butterfly needles or silicone tubing should
not be used for anterior chamber infusion because they may
damage the endothelium and offer no help in accessing the
posterior chamber for manipulation of cortex. Sew-on cor-
neal infusion cannulas do not permit changing the active
instrument from one hand to the other and can cause cor-
neal damage via mechanical trauma or fluid directed against
endothelium.
ULTRASONIC FRAGMENTATION
Aspirating ultrasonic fragmenters such as the Alcon Accurus
four-crystal device and the Constellation permit 20-gauge
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s
Fig u re 7 .1 ■ In fusion sleeves
in crease in fu sio n flu id tu rb u -
lence, resulting in endothelial cell
dam age.
incisions to be used for relatively dense cataracts (11,12).
The 20-gauge aspirating fragmenters require a separate
infusion cannula identical to modern vitrectomy systems.
The phacoemulsifier is analogous to the full function vitrec-
tomy approach and similarly requires a larger incision and
creates more turbulence.
Current 25- and 23-gauge vitrectomy systems lack a
fragmenter in both gauge diameters. The only available frag-
menter at the time of this writing is 20 gauge. On the other
hand, modern fluidics in both 25- and 23-gauge systems
allow removal of soft nuclei with the vitrectomy probe alone.
Only harder nuclei require the creation of a 20-gauge wound
to introduce the fragmenter.
In fusio n So urce s
Infusion can be accomplished with the same options used
with the vitrectomy instrument: bimanual, with 20-gauge,
30-degree bent, blunt cannula, infusion sleeve, sew-on infu-
sion cannula or 25-gauge trocar-cannula system.
The metal infusion sleeve, which has the potential of
dampening the sonic energy, is not intended for use with
the fragmenter. Bimanual infusion with the angulated, blunt
cannula can be kept away from the suction port, decreasing
turbulence, and may be used for selective mobilization and
hydration of lens material in certain cases.
Vacuum So urce s
Syringes, peristaltic pumps, or venturi-based systems can
provide suction for the vitrectomy probe or aspirating
fragmenter. For the reasons discussed in Chapter 2,
77
venturi-based vacuum systems are far superior to the other
approaches because they reduce vitreoretinal traction and
ocular collapse. A vitrectomy system with proportional
(linear) suction control of a venturi source offers the most
precise control over suction level currently available.
Leaving the fragmenter lumen open to the atmosphere
(incorrectly called “passive egress”) is inappropriate because
precision, foot control, and reflux capability are lost. The tub-
ing can be squeezed to create reflux if the iris or retina becomes
impacted in the port. Certain vitrectomy systems (Accurus
and Constellation Vision System) offer foot pedal–controlled
reflux. Foot pedal control is also available with the vitrectomy
mode of phaco systems such as the Alcon Legacy and Infinity.
Modern phaco technique(s) utilizing high suction and
limited flow rates are not used with current fragmenters
because of the 20-gauge needle configuration.
In t e rm it t e n t Ve rsus Co n t in uo us
So n ifi cat io n an d Asp irat io n
Bursts of intermittent sonification alternated with aspira-
tion have been recommended for fragmenter use, but they
prolong the procedure and cause unnecessary heating of the
sclera (13). Aspiration without sonification is ineffective and
time consuming. Luminal fluid flow is the most effective
means of dissipating frictional heat generated by the frag-
menter in the wound. Squirting fluid on the needle shank
externally does not cool the internal scleral wound. By com-
bining continuous sonification with continuous aspiration, a
safe and rapid lensectomy can be performed using the con-
tinual luminal fluid flow for scleral cooling.
78 SECTION II ■ Surgical Technology and Techniques
Lim b al Ve rsus Pars Plan a Ap p ro ach
fo r An t e rio r Vit re ct o m y
Elective anterior vitrectomy during routine cataract removal
should be reserved only for selected pediatric cataracts or
for those cases in which the anterior vitreous cortex and an
opaque posterior capsule resistant to polishing are firmly
adherent. Complete posterior capsulectomy from the lim-
bal approach almost always results in anterior vitrectomy
because of the approach angle and posterior displacement of
the posterior capsule from the infusion system.
While the limbal approach to anterior vitrectomy is more
comfortable for the anterior segment surgeon, it creates more
endothelial and iris trauma than pars plana surgery. Limbal
anterior vitrectomy should be reserved for intraoperative cata-
ract surgery complications if the surgeon is not comfortable
with sideport infusion combined with pars plana vitrectomy.
Peripheral anterior synechia frequently form at the limbal
wounds postoperatively. The pars plana approach allows more
complete removal of peripheral membranes and vitreous with-
out iris trauma (see “Pupillary Membranes”). If conversion to
posterior vitrectomy becomes necessary, the limbal approach
is inappropriate because the angulated instruments cause
striate keratopathy. Most pupillary or retro-IOL membranec-
tomies should be done through the pars plana. Translimbal
bimanual lensectomy should be used for radiation cataracts in
retinoblastoma patients and in some cases of severe iris-retinal
adherence when pars plana entry is unsafe.
PARS PLANA LENSECTOMY
In d icat io n s
Pars plana lensectomy (PPL) is not indicated for elective
cataract removal unless the lens is subluxated or dislocated.
The obligatory interruption of the anterior vitreous cor-
tex increases retinal complications (cystoid macular edema
[CME] and retinal detachment), and it is not compatible with
endocapsular IOL implantation. The principal role of PPL is
in combination with indicated pars plana vitrectomy (PPV).
Many phakic, proliferative vitreoretinopathy (PVR)
(Chapter 9) and giant break (Chapter 10) cases often require
PPL even if the lens is clear. Many trauma cases require
removal of the lens because of associated lens damage or
ciliary body damage.
Aphakic eyes allow better visualization, better approach
to the anterior vitreous cortex and peripheral vitreous, and eas-
ier fluid-air or fluid-gas exchange; eliminate the substrate for
cyclitic membrane formation; and allow faster egress of cells,
protein, fibrin, fibronectin, hemorrhage, and growth factors.
Clear lenses should not be removed in a vitrectomy performed
for proliferative diabetic retinopathy (PDR) because of the
increased risk of neovascular glaucoma unless there is fibrin
syndrome, there is a high probability of postoperative bleed-
ing, or silicone is used for large retinectomies. It is thought
that the anterior vitreous cortex and the lens act as a barrier
to diffusion of vascular endothelial growth factor (VEGF) to
the anterior segment; VEGF causes neovascularization of the
iris and trabecular meshwork. For this theoretical reason and
to reduce posterior subcapsular cataract rates, the anterior
vitreous cortex should be avoided in diabetic vitrectomies
in which the lens is left in place, unless it is opaque. The
indications for elective clear lens and anterior vitreous cortex
removal without IOL implantation in the diabetic patient will
be explained in a subsequent chapter.
De ve lo p m e n t o f Pars Plan a Le n se ct o m y
The development of vitrectomy and phacoemulsification has an
intertwined and interesting history. The first vitrectomy infusion,
suction cutter was patented by Anton Banko in 1969 in response
to vitreous complications of early phacoemulsification. Banko
had learned of the need for a vitreous cutter because he devel-
oped the fluidics for Kelman’s early phacoemulsifier. Machemer
developed the trans-PPV procedure in 1970 and shortly there-
after performed lensectomy with the vitreous suction cutter
(VISC). It was soon discovered that the vitreous cutter would
not handle significant nuclear sclerosis. Girard developed the
fragmenter with Sparta in 1972 as a phacoemulsifier alternative
without the coaxial infusion sleeve. One of the authors (S.C.)
was the first advocate of PPL using the Girard fragmenter with
aspiration. Girard advocated vitrectomy with his unit and later
recommended using the fragmenter for routine cataract sur-
gery. The author believes that ultrasonic vitrectomy and PPL for
routine cataract surgery are unsafe. Shock adapted a dental unit
to cataract surgery just as Kelman had done at an earlier date
but, like Girard, could not use a coaxial infusion sleeve, which
had been patented by Kelman. The Shock technique required
a large, intentionally leaky wound and was used with infusion
through the fragmenter needle rather than suction. Machemer
used the Shock system with a large, leaky pars plana incision
for cases with nuclear sclerosis too great for the VISC.
Co n ve n t io n al Frag m e n t e r Te ch n iq ue s
Current practice for PPL with the fragmenter begins after
standard 25-, 23-, or 20-gauge port placement. The sur-
geon may attempt lensectomy with the vitreous cutter if the
lens does not appear too dense. Once a decision to perform
lensectomy with a 20-gauge fragmenter is done, the con-
junctiva and sclera should be opened as described in the
20-gauge wound construction chapter.
The next step in conventional lensectomy is to incise
the equatorial lens capsule with the microvitreoretinal
(MVR) blade. The author used the fragmenter to incise the
lens capsule before endocapsular lensectomy was developed.
The fragmenter method was initiated to avoid the stress that
capsular incision with the MVR blade creates on the zonules.
Many surgeons advocate penetration of the nucleus with the
MVR blade. This step is unnecessary if the nucleus is soft and
creates excessive stress on the zonules if the nucleus is hard.
After incising the lens capsule, the fragmenter is used in
what phaco surgeons would call a sculpting mode to remove
lens material. The author has advocated sculpting in a plane
parallel to the iris, starting on the temporal side adjacent to
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s
the entry site. An initial thick sheet of cortex and nucleus
is removed while being careful to avoid the cortex near the
anterior and posterior lens capsules. The posterior capsule
is much weaker than the anterior capsule and is usually not
intact by the time the posterior cortex has been removed.
Some surgeons recommend alternating aspiration and
sonification. Unlike the phacoemulsifier, the fragmenter has no
infusion sleeve and must rely on fluid flow through the lumen
for cooling. The needle has approximately 0.003 inches of lon-
gitudinal movement and generates significant frictional heat.
By contrast, the author has always recommended continuous
and simultaneous sonification and aspiration. The constant
fluid flow cools the needle and therefore the sclera. If white,
particulate matter (lens milk) appears at the needle tip, the sur-
geon must release the foot pedal immediately to avoid scleral
burns. If the needle becomes clogged, it should be back flushed
with a syringe filled with balanced salt solution (BSS) with the
ultrasound activated after double-checking to determine that
the needle is outside the eye. Aspiration of saline from a cup
is never effective, nor is it necessary to clean the needle with a
stylet or replace the needle if it becomes plugged.
In fusio n Op t io n s
Many surgeons recommend placing a separate infusion into
the lens rather than using the pars plana infusion cannula. If
the lens is soft, infusion directly into the lens is not necessary.
If the lens is hard, infusion into the lens will not reach the
temporal side where the sculpting must begin. Infusion into
the lens tends to force lens material through ever-present
defects in the posterior capsule. For this reason, the authors
use only the pars plana infusion cannula for lensectomy
cases except during hydrodissection.
Le n s Cap sule Re m o val
In the presence of anterior PVR or severe intraocular inflam-
mation, placement of an IOL should not be performed and
Fig u re 7 .2 ■ Alco n Co n stel-
lat io n frag m en t er an d p h aco
probes.
79
complete capsulectomy is recommended to decrease post-
operative inflammation and remove the substrate for cyclitic
membrane formation. Most surgeons use the vitreous cutter to
remove the capsule. The senior author has recommended end-
opening forceps to remove the capsule since the late 1970s.
The author has used the diamond-coated forceps developed
with Grieshaber since they became available in the late 1980s.
Zonulorhexis is performed in a circular motion identical to
capsulorhexis. Care must be taken to avoid engaging the vitre-
ous so as to prevent undue force on the retina. Using the vitre-
ous cutter to remove lens capsule frequently damages the iris,
causes miosis, prolongs operating time, may cause bleeding,
and leaves residual lens material. Residual capsule and lens
material cause inflammation and may lead to cyclitic mem-
branes and closure of peripheral iridectomies, especially the
inferior peripheral iridectomy necessary in aphakic silicone
oil cases.
Frag m e n t e r Pe rfo rm an ce
The original fragmenter handpieces sold by Berkley Bioengi-
neering, Coopervision, Sparta, Fibrasonics, MidLabs, Storz,
and Alcon were all made by Fibrasonics. These fragmenters
were useful but had less power than the phacoemulsifier and
hence more difficulty with dense nuclear sclerosis. Alcon
introduced a four-crystal fragmenter in the early 1990s that
uses the same ultrasonics as the Legacy and Infinity pha-
coemulsifiers. The Alcon Accurus and Constellation frag-
menters (Fig. 7.2) will handle dense nuclear sclerosis just as
well as the phacoemulsifier.
Ph aco e m ulsifi cat io n wit h Pars Plan a
Vit re ct o m y
As both phacoemulsification and vitrectomy techniques have
become more sophisticated, performing safe combined pha-
co-vitrectomy technique is a possibility for select patients.
This issue is discussed further in Chapter 8.
80 SECTION II ■ Surgical Technology and Techniques
An t e rio r Le n s Cap sule Re t e n t io n an d Le n s
Im p lan t at io n in t h e Sulcus
The late Ron Michels was long an advocate of preserving the
anterior lens capsule until the end of the vitrectomy to reduce
damage to the endothelium and trabecular meshwork induced
by erythrocytes, infusion fluid, and turbulence. Kokame and
Blankenship reported retention of the anterior lens capsule to
permit the implantation of posterior chamber lens in the ciliary
sulcus at the end of the vitrectomy. They recommended per-
forming an anterior capsulotomy after lens implantation. Some
phaco surgeons have been critical of this procedure because
endocapsular implantation has better results than ciliary sulcus
placement in elective cataract surgery. The authors have had
very good results with this method with no cases of decen-
tration. Capsular fibrosis occurs rapidly if the capsule is in
contact with silicone oil. The authors have found that polish-
ing the capsule increases the chance of preserving a clear cap-
sule. The authors use Acrysof three-piece foldable acrylic lens
from Alcon through a 2.2-mm keratome, clear corneal inci-
sion with implantation in the sulcus, anterior to the retained
anterior capsule after the vitrectomy. One should avoid the use
of silicone lenses as they absorb silicone oil and have marked
condensation problems during fluid-air (gas) exchange. Phaco-
vitrectomy has virtually eliminated the need for this method.
En d o cap sular Le n se ct o m y Me t h o d
Conventional lensectomy, as described previously, has many
of the attributes of endocapsular phaco except that it starts
with a risky, equatorial capsular incision. This equatorial
defect frequently leads to capsular tears that extend into
the anterior capsule. Extension of capsular tears was a
frequent problem in phaco surgery until the continuous
capsulorhexis was developed. Cataract surgery has seen a
progression from intracapsular to extracapsular surgery and,
ultimately, phacoemulsification. Phaco has evolved from
anterior chamber phaco to iris plane phaco to endocapsular
surgery. Continuous capsulorhexis has dramatically reduced
problems with capsular tears for the cataract surgeon.
En d o cap sular Le n se ct o m y Surg ical Se q ue n ce
One of the authors (S.C.) has been using posterior capsu-
lorhexis, cortical cleaving hydrodissection, and sculpting
since 1994 and has found these techniques of great value
in lensectomy. This approach is referred to as endocapsular
lensectomy.
Endocapsular lensectomy begins with placement of the
ports as described previously for conventional lensectomy.
The endoilluminator is essential to stabilize the eye and
provide a controllable red reflex. The endoilluminator can
be touched to the lens capsule in various locations to aid
in visualization without the light toxicity associated with
retina-reflective red reflex methods. The sclerotomy for the
fragmenter is made superotemporally for the capsulorhexis,
hydrodissection, sculpting, and cortex-aspiration tools.
A central posterior capsulorhexis is made with vitre-
ous cutter (Fig. 7.3) after a limited anterior vitrectomy. The
vitrectomy is performed to prevent vitreoretinal traction
from the fragmenter. Cortical cleaving hydrodissection is
the next step. A blunt, 27-gauge cannula attached with a
Fig u re 7 .3 ■ Posterior cap su-
lorh exis is p erform ed with th e
vitreous cutter after anterior vit-
rectom y to p revent eng ag em ent
of vitreous in the fragm enter.
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s 81

Fig u re 7 .4 ■ Cortical cleaving

h yd ro d issect io n is p erfo rm ed
with a b lun t 27-g aug e can n ula
attached to a 3- to 5-m L syringe
via short length of tub ing.

short length of tubing to a 5-mL syringe operated by the
assistant is used for this step (Fig. 7.4). If endocapsular
lensectomy is attempted with the 25-gauge cutter without
the fragmenter, hydrodissection can be performed with a
27-gauge needle advanced through the 25-gauge cannula to
the edge of the posterior capsulectomy. Sculpting without
nuclear cleaving or rotation is then initiated at the temporal
margin of the nucleus to avoid damaging the capsule with
the fragmenter. The Alcon 20-gauge, four-crystal fragmenter
is used to sculpt the nucleus, epinucleus, and finally the
cortex (Fig. 7.5). This fragmenter has the same ultrason-
ics as the Legacy. The anterior cortex is avoided to prevent

Fig ure 7 .5 ■ The nucleus, then

ep in ucleus, an d then cortex are
sculp ted with the frag m enter.
82 SECTION II ■ Surgical Technology and Techniques
damaging the anterior capsule. The vitreous cutter is used
for aspirating the remaining cortex. The vitreous cutter is
better than classic irrigation/aspiration (I/A) tools because
of the cutting capability that acts like chopping. The ante-
rior capsule is polished after aspirating the cortex (Fig. 7.6).
Almost any posterior chamber lens could theoretically be
utilized, but foldable lenses offer the advantage of a smaller
Fig u re 7 .6 ■ I/ A is p erform ed
with th e vitreous cutter so th at
cortex can b e cut and asp irated
after u sin g g en tle su ctio n t o
rem ove from capsular bag.
incision. The hydrophobic foldable acrylic multipiece lens
(Alcon Acrysof) is currently being utilized because of the
low incidence of capsular opacification. Scleral tunnel or
clear cornea incisions can be used (Fig. 7.7), and the lens
is positioned in the ciliary sulcus, over the intact anterior
capsule. Silicone lenses, as previously stated, should not
be used because they have severe condensation problems
Fig ure 7 .7 ■ A foldable Acrysof
IOL is in serted th roug h a clear
corneal or scleral tunnel incision.
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s
Fig u re 7 .8 ■ Zo n ulorh exis is
p erform ed in a circular fash ion
with th e in tern al lim itin g m em -
b ran e (ILM) o r en d -g rasp in g
forceps.
during fluid-air exchange, have higher posterior capsular
opacification rates, and absorb silicone oil. Multipiece
(three-piece) foldable acrylic lenses are preferable in the
sulcus to single-piece lenses, because the thickness of the
haptics in the latter might rub against the posterior iris and
cause postoperative iritis and pigment dispersion.
Pars Plan a Cap sulo t o m y Aft e r Le n se ct o m y
Capsulotomy is used if the intent is complete capsule removal.
An MVR blade is placed through the sclerotomy and used to
incise the anterior and posterior capsules centrally to create
an “edge.” The fragmenter almost always ruptures the pos-
terior capsule, eliminating the need for posterior capsulo-
tomy. The forceps method of total capsule removal requires
an edge. If the lensectomy is accomplished with the vitre-
ous cutter without the need for enlargement to a 20-gauge
incision, a 25- or 27-gauge needle can be used through the
25-gauge cannula to incise the anterior lens capsule.
Pars Plan a Cap sule ct o m y Aft e r Le n se ct o m y
End-grasping forceps should be used to remove the anterior
and posterior capsules after capsulotomy. Iris contact should
be avoided. A circular, zonulorhexis approach is better than
pulling across the eye because it reduces traction on the
peripheral retina (Fig. 7.8).
Avo id an ce o f Vit re o us in t h e Frag m e n t e r
The fragmenter emulsifies formed vitreous gel but does
not emulsify collagen fibers, giving a false impression of
vitrectomy. The vitreous fibers remain intact, and suction
applied with the fragmenter creates dangerous vitreoretinal
83
traction (Fig. 7.9). When vitreous enters the fragmenter,
the vitrectomy probe, not the fragmenter, should be used to
remove the vitreous. The fragmenter should never be used
for emergency or elective vitrectomy.
Sub luxat e d Le n se s Durin g Vit re ct o m y
Subluxated lenses are best managed by completing the
vitrectomy through the pars plana before approaching the
lens. Posterior dislocation is frequently seen and presents
no problem with the pars plana approach. After the vitre-
ous surrounding the lens is removed, the 20-gauge aspirat-
ing ultrasonic fragmenter can safely be used to complete
the lensectomy as described above (Fig. 7.10). Rhexis and
hydrodissection are very difficult if the lens is subluxated or
dislocated and may be omitted.
Po st e rio r Dislo cat io n o f Le n s Mat e rial Durin g
Cat aract Surg e ry
Surgica l Psychodyna mics
Cataract surgery has been one of the most frequently per-
formed surgical procedures worldwide for over a century.
While inexperience is known to cause a higher complication
rate, high surgical volumes and phenomenal success rates
can also cause surgeons to become complacent and make
judgment errors when complications do occur. Busy sched-
ules, observers, and video may contribute to faulty decision
making when the capsule ruptures and lens material and the
vitreous exchange locations. High success rates, outpatient
surgery, no stitch, no patch, emmetropia, and topical
anesthesia elevate patient expectations unrealistically, which
can place more pressure on the surgeon.
84 SECTION II ■ Surgical Technology and Techniques
Ea rly Recognition a nd Ma na gement of Defects
in the Lens Ca psule
Modern microscopes enhance the red reflex, allowing early
recognition of capsular defects. Very commonly, the posterior
capsular defect is not recognized by direct visualization until
later in the surgery, and early recognition is important to
prevent further complication of the procedure. Therefore,
Fig u re 7 .9 ■ Rem oval of th e
p h aco p rob e can cause vitre-
ous p rolap se in to th e an terior
cham ber.
the surgeon needs to notice subtle indirect changes during
cataract surgery that can alert to the presence of a poste-
rior capsular tear and vitreous prolapse. These are (a) sud-
den, unexpected deepening of the lens or iris diaphragm,
(b) difficulty in rotating the lens inside the capsular bag,
(c) difficulty in manipulating and moving lens fragments,
and (d) minor decentration of the lens nucleus. Any of these
Fig ure 7 .1 0 ■ Sub luxated lens
m aterial can b e safely rem oved
with the fragm enter only after all
vitreous has been rem oved.
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s
signs should suggest either instability of the lens support
structures or presence of vitreous intermingled with the
remaining lens fragments. If a posterior capsular tear is sus-
pected, the surgeon should immediately decrease all the
phaco fluidic parameters (lower the height of the infusion
bottle, lower the vacuum, and lower the aspiration flow rate)
and proceed operating in “slow motion.”
Most surgeons feel the urge to quickly withdraw the
phaco tip from the eye when they suspect that a posterior
capsular tear has occurred, but doing so acts as a suction
device that would guarantee immediate vitreous prolapse
into the anterior chamber. The surgeon must admit that the
defect has occurred rather than rationalize because of the
psychological factors described previously. When a capsular
defect is recognized, the first actions should be to leave the
phaco tip inside the eye (with the pedal on position zero)
and inject a viscoelastic into the capsular defect using the
sideport incision, creating a barrier between the posterior
capsule and the anterior vitreous cortex. High-viscosity,
cohesive viscoelastics injected into the defect can serve as a
temporary barrier, enabling removal of remaining lens mate-
rial and removal of the phaco tip from the eye without caus-
ing further vitreous loss. If the lens implant can be placed
over the defect and posterior to the lens material, it can itself
function as a barrier.
Many surgeons focus exclusively on the prevention or
management of posterior dislocation of lens material rather
than the more serious matter of reducing vitreoretinal traction
and subsequent retinal detachment. Any maneuver designed
to prevent posterior dislocation that increases vitreoretinal trac-
tion should not be employed. Kelman described the use of a
needle inserted through the pars plana to prevent lens material
from falling posteriorly. This method ignores the pressure that
must be placed on the eye to place the needle and the anterior
movement of the vitreous that occurs without a barrier. The
next section discusses the management of vitreous that pro-
lapses through the capsular defect. This discussion intention-
ally precedes the discussion of the management of dislocated
lens material because reduction of vitreoretinal traction and
retinal detachment prevention are the most important issues.
Vitreous Loss
As previously stated, use of the phacoemulsifier to remove
vitreous is a dangerous step that should never be under-
taken. Phaco probes liquefy hyaluronic acid but do not cut
collagen fibers. Use of a large-bore needle to aspirate “liquid”
vitreous should be avoided because of the obligate vitreoreti-
nal traction. The theoretical “pockets” of liquid vitreous are
more difficult to locate than the fountain of youth.
Cellulose sponge vitrectomy as developed by Kasner
has been an obsolete and dangerous method for two decades
in spite of the important role it played before machine vit-
rectomy. A cellulose sponge causes significant traction on the
retina as the sponge is lifted to transect the adherent vitreous
(Fig. 7.11) and because of wicking. Removal of all vitreous
85
by a vitreous cutter causes virtually no inflammation, while
marked inflammation is the rule after sponge vitrectomy.
Mechanical damage to the iris caused by contact with the
sponge as it swells and is lifted appears to be the cause of
this inflammation, as could the tugging of vitreous fibers on
a miotic pupil, and the unnecessary and prolonged intraocu-
lar manipulation required to finish a surgery using inappro-
priate equipment and techniques. One of the authors (S.C.)
has also observed cellulose material on the anterior vitreous
cortex after sponge vitrectomy has been performed. One can
speculate that this retained material causes inflammation in
addition to that caused by iris trauma. Testing for vitreous
can be accomplished by injecting air into the anterior cham-
ber via the sideport incision and looking for fragmentation
of the bubble. Alternatively, a single drop of sterile fluores-
cein from a newly opened ampule can be used to stain the
vitreous. Injection of small quantities of triamcinolone sus-
pension (Kenalog) in the anterior chamber through one of
the wounds can also aid in highlighting the prolapsed vitre-
ous, as the steroid particles are adsorbed by the vitreous.
This technique permits the visualization of very thin vitre-
ous strands that might be missed otherwise and that cause
vitreoretinal traction. Although theoretically this technique
might decrease postoperative inflammation, this is not the
intended consequence and is only a potential side benefit.
Vitrectomy with a high-quality vitreous cutter is the
preferred method of managing vitreous that presents in the
anterior chamber. Alcon builds high-quality cutters for use
with their phaco systems such as the Legacy and Infinity
which now has a 23-guage Ultravit, 2500 cut/minute cutter.
These cutters should be operated at the highest possible cut-
ting rate and low vacuum/flow (Fig. 7.12). Posterior vitreous
surgeons use vacuum control in addition to flow control to
reduce vitreoretinal traction. The anterior segment machines
frequently utilize peristaltic pumps, which do not directly
control the vacuum. The best procedure is to use a very
low flow rate and vacuum settings to reduce traction on the
retina. The cutter should be advanced or held stationary
during vitrectomy, never retracted. Pulling the cutter back
while vacuum is applied dramatically increases vitreoretinal
traction (Fig. 7.13). The introduction of the cutter through
the original cataract incision will cause excessive fluid egress
due to the mismatch of size and diameter of the cutter to
the phaco wound. It is preferable to perform bimanual vit-
rectomy with two limbal paracenteses separated 90 degrees
apart (one at each side of the larger phaco wound): one for
the sleeveless vitreous cutter and the second for an infusion
cannula. It might be necessary to place temporary nylon
sutures on the original phaco wound to stop any fluid and
vitreous egress through that wound while the vitrectomy is
performed. Sideport infusion is preferable to “dry” vitrec-
tomy because it prevents hypotony and therefore reduces
the chance of choroidal hemorrhage. Air should be used
instead of infusion fluid to keep the vitreous from hydrat-
ing and coming forward. The air helps to delineate the sur-
face of the vitreous and keep it confined by surface tension.
86 SECTION II ■ Surgical Technology and Techniques
Sweeping the wound for vitreous is dangerous because of
the vitreoretinal traction it causes.
Disloca ted Lens Ma teria l
Phacoemulsifiers, lens loops, and irrigation should never
be utilized in an attempt to extract lens material from
Fig ure 7 .1 1 ■ Cellulose sponge
vitrectom y or testing for vitreous
causes excessive vitreoretinal trac-
tion because of lifting to cut and
wicking and should never be used.
The sponge also causes m echani-
cal traum a to the iris, resulting in
postoperative inflam m ation.
the vitreous cavity. If lens material falls posteriorly, there
is a natural tendency for the surgeon to chase it with the
phaco probe. The phaco probe gives the appearance of vit-
reous emulsification but does not sever the collagen fibers
(Fig. 7.14). The surgeon must consciously stop, relax, and
plan before performing any further maneuvers. The best
plan is usually to let the material fall posteriorly and focus
Fig u re 7 .1 2 ■ Vitreous in the
an t e rio r ch am b er sh o u ld b e
rem oved with a hig h-quality vit-
reous cutter using the m axim um
cuttin g rate availab le an d low
vacuum / flow rate.
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s 87

Fig u re 7 .1 3 ■ Pulling b ack d uring low-sp eed cuttin g with suction ap p lied causes severe
vitreoretinal traction.

Fig ure 7 .1 4 ■ The phacoem ulsifier does not cut collagen fibers; it only disperses the gel, giving
the illusion of vitrectom y. The phaco p robe should never be introduced into the vitreous cavity
in order to prevent dangerous vitreoretinal traction.
88 SECTION II ■ Surgical Technology and Techniques
on vitreous cleanup and IOL implantation. Lens loops can
put significant traction on the retina and cause retinal breaks
and detachments (Fig. 7.15). Foulds, and subsequently
Machemer, used a saline stream directed at the retina to cre-
ate experimental retinal detachments. There is a significant
risk of retinal breaks if saline irrigation is used in an attempt
to move the lens material anteriorly (Fig. 7.16).
Fig ure 7 .1 5 ■ Lens loops cause
sig nificant vitreoretinal traction if
used in the vitreous cavity.
If the pupil is large, the cornea is clear, and the
surgeon and available staff are optimum for posterior vitrec-
tomy, immediate intervention may be undertaken. In most
instances, though, it is preferable to perform posterior vit-
rectomy and removal of lens material at a second procedure.
This procedure should be performed when the cornea is clear,
the wound is sealed, and the pupil is well dilated. The timing
Fig ure 7 .1 6 ■ BSS irrig ation to
m ove the lens m aterial anteriorly
can cause retinal breaks.
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s
can be from several days to weeks later. If there is a moderate
amount of cortex, no inflammation, no glaucoma, and no
lens-corneal touch, a vitrectomy may not be necessary.
Posterior vitrectomy requires a surgeon specifically
trained in posterior vitrectomy techniques and a sophis-
ticated vitrectomy system. An infusion cannula placed
through the pars plana is essential. An assistant-supported,
irrigating, corneal contact lens is easier and faster to use than
a sutured-on contact lens. Wide-angle visualization systems
increase cost, complexity, and the learning curve, although
they provide an excellent view of the periphery (unneces-
sary in anterior vitrectomy). A fiberoptic endoilluminator is
essential for all cases. Light reflexes from the cornea prevent
the surgeon from having an optimal view if coaxial illumina-
tion is used. Iris retractors increase inflammation and may
cause a distorted pupil after surgery. Some surgeons have
advocated indirect ophthalmoscopic visualization, but an
inverted view and the need to support the lens with one
hand make this approach dangerous.
All vitreous should be removed before removing any
dislocated lens material. Many surgeons have the miscon-
ception that lens material can damage the retina if it falls
posteriorly. Inappropriate techniques, not the lens, damage
the retina. It is dangerous and unnecessary to leave a layer
of vitreous under the lens material until it is removed. Some
cortex may be removed with the vitreous cutter, but dense
nuclear material requires the phacofragmenter. The Alcon
four-crystal fragmenter utilizes the same drive electronics
and piezo driver as the Infinity phaco probe and, similarly, is
able to handle the majority of nuclear sclerosis cases.
After removal of the vitreous, the fragmenter is intro-
duced and moved to the surface of the lens material. Linear
Fig ure 7 .1 7 ■ PFC liquids such
as PFO can be used to float a hard
nucleus into the anterior cham -
ber enabling rem oval with a lens
loop .
89
(proportional) suction is increased gradually until the lens
material is picked up. The lens material is then moved away
from the retina and the foot pedal is used to activate soni-
fication in midvitreous. Proportional fragmentation power
is used to adjust the power until sufficient sculpting occurs
without bouncing of the lens material. If the fragmenter
drills into the lens, the endoilluminator is used to push the
lens material off the tip. Alternatively, the endoilluminator
can be used to crush or “chop” the fragment that is speared
on the fragmenter tip. This process is continued until all lens
fragments are removed.
Perfluorocarbon (PFC) liquids (Chang) were intro-
duced to vitreoretinal surgery for unfolding giant breaks,
draining subretinal fluid, and stabilizing the retina during
the dissection of epiretinal membranes. PFC liquids can also
be used to float the lens material away from the retina, allow-
ing aspiration-fragmentation or phacoemulsification to be
performed anteriorly (Fig. 7.17). This method increases cost
and may require a subsequent procedure to remove residual
PFC liquids. The PFC liquid method is safe but unneces-
sary, unless there is extremely dense nuclear sclerosis that
should not have been managed with phaco in the first place.
It may be useful for surgeons with limited posterior vitrec-
tomy experience.
It is imperative to evaluate the retinal periphery at the end
of removal of dropped lens fragments. The retinal periphery
may be impossible to fully evaluate for retinal tears preopera-
tively due to the presence of the lens fragments and dispersed
cortex. Despite what a referring cataract surgeon tells the con-
sulting vitreoretinal specialist, if vitreous loss occurred dur-
ing the cataract surgery, we should assume the presence of
peripheral retinal tears until otherwise demonstrated.
90 SECTION II ■ Surgical Technology and Techniques
Ha rd Disloca ted Lens Ma teria l
Hard dislocated nuclei can usually be crushed between the
endoilluminator and the aspirating 20-gauge ultrasonic frag-
menter (14,15). After the pieces are made smaller, they can
be removed with the fragmenter. If the fragmenter becomes
plugged, it can be removed from the eye and back flushed
with sonification as often as required.
If the bimanual crush technique is not effective because
of a dense, black nucleus, internal simultaneous fluid-gas
exchange can be performed if PFC liquids are not available.
The endoilluminator should then be removed and the sclero-
tomy plugged with a scleral plug. The surgeon’s other hand
can then make a razor blade-scissors limbal incision of 90
to 110 degrees to remove the nucleus. The iris will prolapse
unless the infusion is turned off after the section is made. An
assistant can elevate the cornea and the lens can be removed
with a 1-mm cryoprobe without touching the endothelium.
Perfluoron (PFO) can be used to eliminate the need for a
cryoprobe. The wound is then closed with a running shoe-
lace or X-type 8-0 or 9-0 monofilament nylon suture. This
method has not been needed for decades.
Intra ocula r Lens Impla nta tion
Some retinal surgeons are opposed to lens implantation in
cases of posterior dislocation of lens material for unclear rea-
sons. The authors recommend lens implantation if certain
conditions are met. If the posterior capsule can support an
IOL, it can be placed in the bag with the haptics rotated
away from the capsular defect. If the posterior capsule will
not support an IOL, the IOL can be implanted in the ciliary
sulcus anterior to the anterior lens capsule.
If the capsule is not sufficient to support the IOL, an
anterior chamber lens can be used. Anterior chamber lenses
are contraindicated if there is a low endothelial cell count,
significant open angle glaucoma, or uveitis. Judgment is
required to set the level of cell counts and severity of glau-
coma that represent contraindications.
If there is insufficient capsular support for a posterior
chamber lens and low corneal endothelial cell counts or signif-
icant glaucoma prevents the use of an anterior chamber lens, a
sutured lens can be used. This technique requires practice and
can result in significant intraoperative and postoperative com-
plications, including suprachoroidal hemorrhage as well as late
complications such as endophthalmitis and suture breakage.
CONGENITAL AND PEDIATRIC
CATARACTS
PPL is suitable for cataracts occurring in infants and young
children. Extracapsular cataract extraction technique in this
age group uniformly results in capsular clouding requiring
discission or YAG capsulotomy (16–19). Limbal anterior vit-
rectomy (17) reduces late vitreoretinal traction and detach-
ment. A pars plana approach has less corneal problems and
permits removal of all peripheral cortex and capsule. One of
the authors (S.C.) has used pars plana lensectomy with total
capsule romoval many pediatric cataracts since 1975 and
has never had a postoperative retinal detachment. The con-
ventional I/A approach results in a fibrous ring often includ-
ing iris, neocortex, ciliary body, pars plana, and peripheral
vitreoretinal traction. This configuration can lead to retinal
detachment with even minor trauma. Phaco and IOL implan-
tation often with primary posterior capsulorhexis and ante-
rior vitrectomy can be used in children of appropriate age, if
there is a normal-sized eye and no glaucoma or vitreoretinal
problem (20–23).
TRAUMATIC CATARACTS
Most traumatic cataracts are either subluxated or have had
interruption of the anterior vitreous cortex. Vitrectomy
instruments should be used to remove all vitreous and soft
lens material using standard pars plana techniques. After vit-
rectomy, the fragmenter can be utilized if denser lens material
is present. The retinal periphery should be always examined
at the end of lensectomies for subluxated lenses, since the
original contusive force that created zonular damage could
have created a retinal dialysis or giant retinal tear.
PUPILLARY MEMBRANES
Multiple tissues contribute to the formation of pupillary
membranes. There is a wide range in density from an opaque
hyaloid to extremely dense calcified membranes. These
membranes may be postsurgical or posttraumatic or may
occur spontaneously. A range of approaches to these prob-
lems, depending on density, is required (24). Neodymium
YAG laser capsulectomy can be utilized for thin or mod-
erately dense pupillary membranes. The extravagant term
“anterior segment reconstruction” is inappropriate. As dis-
cussed earlier, the pars plana approach is preferable to the
limbal approach in managing most pupillary membranes.
Me m b ran o t o m y (Discissio n )
Simple membranotomy (discission) or neodymium YAG
laser techniques will suffice for a thin membrane such as
lens capsule. The MVR blade can be utilized translimbally
or through the pars plana for the task. If membranectomy
is required, pars plana membranotomy serves as the initial
stage to provide an edge on which the vitrectomy instrument
can imbricate and cut. As an alternative, 20- to 25-gauge
scissors can be used to segment the membrane and free it up
from the zonules and any scarring that is present.
Me m b ran e ct o m y
After an edge is made with the MVR blade, the vitrectomy
instrument can be used to remove a moderately dense lens,
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s
iris, or fibrous material. The infusion cannula should be
placed only if the pars plana is visible through the membrane
to verify infusion at the correct location. The best approach
if the infusion site cannot be visualized is bimanual, using
a 20-gauge, 30-degree angulated, blunt infusion cannula,
through the superonasal pars plana.
De n se Me m b ran e ct o m y
If the vitrectomy cutter using moderate suction levels can-
not cut a dense membrane, scissors should be utilized. The
membrane is cut into many small triangular pieces that can
be removed with the vitrectomy instrument. Bipolar biman-
ual diathermy (Chapter 4) can be used to control bleed-
ing from these membranes. Care should be taken to avoid
excessive tissue removal; a 4- to 5-mm opening is usually
sufficient if the iris is incorporated into the membrane. Off-
center opening can be used to avoid areas of corneal astig-
matism or opacity. All tissues adherent to the ciliary body
can be removed to reduce hypotony and phthisis.
While an 8-mm opening resembles a dilated pupil, it
causes glare, discomfort, reduced vision, and cosmetic prob-
lems postoperatively. Occasionally, a calcified membrane will
be resectable outside the calcified zone or crushable with a
sclerectomy punch. Limited anterior vitrectomy should then
follow to prevent subsequent pupillary block or corneal
touch.
ANTERIOR VITRECTOMY
Anterior vitrectomy is useful for a wide variety of anterior
segment problems: aphakic pupillary block, vitreous touch,
CME with vitreous to the wound, vitreous loss at cataract
surgery, aphakic keratoplasty, secondary IOL implantation,
and aphakic trabeculectomy. Each of these will be discussed
individually, but the methodology is similar and will be dis-
cussed in the following sections.
Tran slim b al An t e rio r Vit re ct o m y
The simplest and most frequent form of anterior vitrec-
tomy is that required for vitreous loss at cataract surgery or
aphakic keratoplasty. The vitrectomy probe is placed in the
central anterior vitreous space and used with minimal move-
ment. The iris, endothelium, and especially the peripheral
vitreous are avoided. Slight increased suction is required,
compared to closed vitrectomy, because there is no infu-
sion to contribute to transorifice pressure. Care should be
taken to avoid air into the cutter port as it increases vacuum
requirements, thereby decreasing safety. A rapid cutting rate
is to be used to reduce vitreoretinal traction and improve
fluidic stability. The approach in penetrating keratoplasty is
similar, but more extensive vitrectomy is usually required to
prevent postoperative corneal vitreous touch. Closure on an
air bubble is best to avoid residual strands to the wound.
91
In fusio n So urce s fo r Clo se d
An t e rio r Vit re ct o m y
Unless the presence of dense hyphema, cataract, or pupillary
membranes dictates the use of the bimanual, angulated can-
nula approach, the pars plana infusion cannula is the best
method for elective anterior vitrectomy. If more extensive
vitrectomy is required, the pars plana infusion cannula
should be placed after the initial bimanual infusion tech-
nique and removal of anterior opacities has been used to
obtain a view of the pars plana. If preoperative ultrasound
examination demonstrates attached retina with normal pars
plana anatomy, the 25-gauge infusion cannula can be placed
in a standard fashion prior to anterior vitrectomy.
Vit re o co rn e al To uch
Anterior vitrectomy can cause complete clearing of corneal
edema due to vitreous contact, if rigid case selection is uti-
lized (25). It is essential to obtain a history showing that
the cornea was once clear after lens removal, subluxation,
or dislocation (26) and that endothelial cell counts are
reasonable. In this way, surgery for corneal edema primar-
ily due to endothelial trauma, which will not respond to
subsequent vitrectomy, can be avoided. Long-term touch
with retrocorneal membrane formation is equally unrespon-
sive to anterior vitrectomy. The pars plana approach prevents
further endothelial trauma, in contrast to a limbal approach.
Vitrectomy should be completed behind the iris and using
the 25-guage infusion cannula to minimize anterior cham-
ber turbulence and reduce endothelial damage.
Ap h akic Pup illary Blo ck
While simple slashing of the anterior vitreous cortex or YAG
laser vitreolysis will suffice in some cases of pupillary block,
more frequently an anterior vitrectomy is required (27). The
pars plana approach decreases iris and endothelial damage
in these shallow chamber cases.
Ciliary Blo ck Glauco m a
Trans-PPL and trans-PPV will alleviate ciliary block (malig-
nant, aqueous misdirection) glaucoma in every case. The
creation of an unobstructed passage of fluid from the vit-
reous cavity through the anterior vitreous, zonule or lens
capsule diaphragm, and iris is curative for this condition.
A stepwise approach can be utilized to guarantee this fluid
passage. Core vitrectomy can be first performed, followed by
anterior vitrectomy. In the setting of pseudophakia, a periph-
eral iridectomy from the posterior pars plana approach can
be performed. Deepening of the anterior chamber and ref-
ormation of the angle may be required if a flat chamber has
been present for over 24 hours. It is advisable to deepen the
chamber with BSS instead of a viscoelastic, since reflatten-
ing of the chamber implies that the aqueous misdirection
92 SECTION II ■ Surgical Technology and Techniques
has not been relieved. Viscoelastics may artificially maintain
anterior chamber depth despite persistence of the misdirec-
tion syndrome. If the iridectomy does not repair the malig-
nant glaucoma, explantation of the IOL may be required.
Aqueous misdirection in a phakic patient may require vit-
rectomy-lensectomy, since it is impossible to create anterior
vitrectomy and iridectomy without violating the lens capsule
and creating a cataract.
Ap h akic Filt e rin g Pro ce d ure s
Most aphakic filtering procedures fail because of closure of
the filtration site by episcleral tissue proliferation. Occa-
sionally, vitreous can plug a filtration site internally. On
occasion, this can be recognized at the time of aphakic
trabeculectomy and managed with the vitrectomy instru-
ment. The sleeve is quite useful in these cases because
the incision is already large and the eye too soft to place
the infusion cannula or make a second incision. Vitrec-
tomy plays a role in the management of other glaucoma
problems such as phacolytic glaucoma, as discussed
elsewhere.
Vit re o us Lo ss at Cat aract Surg e ry
If vitreous presents in the anterior chamber during cata-
ract surgery, the vitrectomy instrument can be used with
sideport air infusion (Fig. 7.18) for anterior vitrectomy. So-
called dry vitrectomy is not advised because of unavoid-
able hypotony and risk of suprachoroidal hemorrhage.
Hypotony also causes striate keratopathy and miosis. Air
should be used instead of BSS to eliminate hydration of
the vitreous as well as prevent vitreous incarceration in the
wound. If a sideport is present, infusion should be used
through this site to reduce turbulence. If “one-handed”
technique was used without a sideport, a sideport for infu-
sion is shoule be placed.
Cyst o id Macular Ed e m a
There is no need to perform anterior vitrectomy if vitreous is
in the cataract wound and there is no CME. If only a strand
of vitreous is seen in contact with the wound, laser YAG
vitreolysis might be sufficient to sever a strand rubbing on
the iris and help relieve CME. If CME is present without vit-
reous in the wound, vitrectomy is unlikely to improve the
edema (28,29). Topical Nevanac should be tried for several
months when postoperative CME is present. Topical predni-
solone or Durezol should be used as well unless the patient
is a steroid responder. If topical treatment is not effective
and the patient is not a steroid responder, retrobulbar, juxta-
scleral triamcinolone or intravitreal triamcinolone should be
tried before PPV is considered. Retrobulbar steroids prob-
ably diffuse through the sclera, but posterior juxta-scleral
delivery produces a higher drug level at the macula. Topi-
cal and subconjunctival treatments target the source of the
inflammation in postoperative CME (Irvine Gass syndrome)
but produce 1/100 the anterior chamber drug level in phakic
or pseudophakic eyes (two-compartment eyes). Leakage of
the perifoveal capillary bed persists after vitrectomy, and the
Fig ure 7 .1 8 ■ Bim anual vitrec-
tom y utilizing a 23-gauge vitreous
cutter and BSS infusion throug h
a 21-g aug e can n ula is th e p re-
ferred m ethod of anterior vitreous
rem oval.
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s
visual improvement is usually modest. It is probable that
the visual improvement results primarily from clearing the
media. It appears that vitrectomy rarely causes CME, and
rarely a successful case will develop recurrent CME in spite
of having the vitreous removed. Because of the modest differ-
ences between vitrectomized and nonvitrectomized eyes in
the national collaborative study (30) and the risk of aphakic
retinal detachment and endothelial damage, caution must be
applied in the utilization of vitrectomy for CME (31–33).
If surgery is undertaken, the infusion cannula can be
used superotemporally and the vitrectomy instrument inf-
erotemporally. In this way, the fibers from iris border to a
superior cataract wound can be removed without dangerous
sweeping techniques. All vitreous adherence to the iris should
be eliminated because this may play a role in inflammation
and CME. Low suction force should be used to reduce post-
operative aphakic retinal detachment. Vitreous scissors can be
used to sever iris-to-wound vitreous fibers. Subconjunctival
steroids are essential at the end of surgery unless the patient
is a steroid glaucoma responder. Posterior juxta-scleral triam-
cinolone injected at the end of the case maximizes the dose to
the posterior segment while reducing the risk of steroid glau-
coma compared to subconjunctival or peribulbar steroids.
Ap h akic Ke rat o p last y
Successful aphakic keratoplasty can be threatened by vitre-
ous touch, pupillary block, and iris touch to the host-donor
interface. Anterior, open sky vitrectomy can be used without
infusion in these cases. Care should be taken to avoid pull-
ing the probe back during cutting to avoid vitreoretinal trac-
tion and air ingestion. Chamber deepening by pressing on
the midiris with a small iris spatula through the pars plana
wound can deepen the chamber without placing an instru-
ment in the angle.
PHACOLYTIC GLAUCOMA
Macrophages laden with ingested lens material clogging the
trabecular meshwork are best managed with vitrectomy.
Gentle suction with a small, blunt cannula near the trabe-
cular meshwork can enhance the removal process. Careful
irrigation of the angle can be utilized as well. Frequently,
cortex is entrapped in the peripheral cortical vitreous, usu-
ally inferiorly. Care should be taken to remove this material
without vitreoretinal traction.
UVEITIS
Cataracts secondary to severe uveitis are usually best
removed by a trans-PPL and trans-PPV approach. Vitrec-
tomy with lens removal for uveitis usually results in the
deposition of inflammation on the posterior lens capsule
and little, if any, improvement in CME and therefore should
93
be avoided. Cataract surgery without vitrectomy may result
in pupillary or cyclitic membranes and pupillary block. It is
not necessary to “quiet” eyes with steroids before vitrectomy
and lensectomy. Vitrectomy reduces, rather than increases,
inflammation and the risk of phthisis.
Iris contact and retractors should be avoided to reduce
inflammation. Scleral depression can be used to remove all
peripheral lens material with the forceps method without
iridectomy. Subconjunctival repository steroids are essen-
tial in these cases unless the patient is a steroid responder.
The authors rarely use systemic steroids because of the well-
known medical risk factors.
The authors strongly urge surgeons to avoid pha-
coemulsification in severe uveitis cases. Capsule retention
and the IOL result in a two-compartment eye with increased
retention of prostaglandins, complement, cells, and other
components of the inflammatory process in the vitreous cav-
ity, thereby increasing CME.
The lens capsule should be removed after lensectomy
using end-grasping forceps to reduce lens-related inflam-
mation and prevent retro-iris membranes, which increase
hypotony, phthisis, and vitreoretinal traction.
Juxta-scleral triamcinolone injected at the end of the
case maximizes the dose to the posterior segment while
reducing the risk of steroid glaucoma compared to subcon-
junctival or peribulbar steroids.
INTRAOCULAR LENSES
Many IOL problems can be managed effectively with a vitrec-
tomy. At times, a normal-functioning IOL must be removed
in the context of posterior vitreoretinal surgery. It is essential
that the implant surgeon be familiar with the vitrectomy-
based management of postoperative problems.
Re t ro le n t al In t rao cular Le n s Me m b ran e s
The YAG laser can manage most retro-IOL membranes. Pars
plana or limbal membranotomy (discission) is infrequently
indicated. Denser membranes require trans–pars plana
membranectomy with the vitrectomy instrument. While the
infusion sleeve can be utilized, it increases probe diameter
and reduces access to the membrane. Standard infusion can-
nula placement is preferred to maintain adequate intraocular
pressure during the surgery. Membranotomy with the scis-
sors, needle, or MVR blade must precede membranectomy
to provide an edge. Occasionally, dense membranes will
require scissors to radially segment and circumferentially
dissect the membrane away from the ciliary body and iris.
Dislo cat e d In t rao cular Le n s Re p o sit io n in g
Posterior chamber lens implants occasionally dislocate into
the vitreous cavity. On very rare occasion, positioning and
miotics can return the implant to proper position without
94 SECTION II ■ Surgical Technology and Techniques
surgery. If this approach is unsuccessful, vitrectomy is
required. Surgical manipulation of a dislocated IOL without
prior vitrectomy causes vitreoretinal traction and should be
avoided.
Complete vitrectomy with the infusion cannula,
vitrectomy probe, contact lens or wide-angle viewing sys-
tem, and endoilluminator should precede IOL reposition-
ing to avoid vitreoretinal traction (Fig. 7.19). End-grasping
Fig ure 7 .1 9 ■ Com plete vitrec-
tom y sh ould p reced e IOL rep o-
sitioning to p revent vitreoretinal
traction.
forceps (34) should be used to grasp the IOL optic with
the endoilluminator providing illumination and additional
support for the lens. The implant can be placed in the intact
portion of the capsule, the ciliary sulcus, or in the ante-
rior chamber (Fig. 7.20). Rotation of the lens in the cap-
sular bag away from the defect that caused the dislocation
can be effective in certain cases (Fig. 7.21). Ciliary sulcus
placement can be used if the anterior and posterior capsules
Fig ure 7 .2 0 ■ The IOL can b e
placed in the capsule if sufficiently
intact, in th e sulcus if sufficient
cap sule is p resent, or in the ante-
rior cham ber if no Fuchs d ystro-
phy or significant glaucom a.
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s 95

Fig u re 7 .2 1 ■ Rotation of the

IOL in to the cap sular b ag with
the hap tics away from the cap -
sular d efect can occasion ally b e
effective.

are fused together and there is sufficient capsular support
(Fig. 7.22). Anterior chamber relocation of certain IOLs can
be utilized if there is no glaucoma and the endothelial cell
counts are good. Vaulting of plate IOLs and silicone IOLs
limits the value of this method. A peripheral iridectomy with
the vitreous cutter should be used in all cases to prevent
pupillary block.
Ciliary sulcus suturing through positioning holes was
reported by the principal author but is seldom possible or
indicated today. Sutures can be passed around the haptics for
ciliary sulcus suturing. This method is complex and requires
experience and careful planning and results in a significant
incidence of late complications such as suture breakage and
endophthalmitis.

Fig u re 7 . 2 2 ■ Ciliary su lcu s

placem ent can be used if there is
sufficient cap sular sup p ort.
96 SECTION II ■ Surgical Technology and Techniques
Suturing of the IOL to the iris with McCannel sutures
may be performed in certain situations. To perform this
technique, the IOL should be grasped with the forceps and
the optic passed through the pupil into the anterior cham-
ber, while keeping the haptics below the iris. Pharmaco-
logic constriction of the pupil with carbachol in the anterior
chamber will keep the IOL in the pupil capture position.
The absence of a functional pupil sphincter makes the iris
suturing of IOLs significantly more difficult. Once the IOL is
stable, a viscoelastic can be injected into the anterior cham-
ber to push the iris posteriorly and delineate the contour of
the haptics. Suturing with a long, straight or curved spatu-
lated needle with 10-0 Prolene can be then performed by
passing the needle through a limbal paracentesis, through
the iris, underneath the haptic, through the iris, and out
of the eye through the limbus. An exit paracentesis is not
required. Mobilization of the loops of Prolene with a Kuglen
hook toward a central paracentesis directly above the haptic
position and out through the limbus allows the surgeon to
tie a knot that can be then reintroduced into the eye, secur-
ing the position of the haptic. This technique can be per-
formed in both haptics, if there is total luxation of the IOL,
or can be performed on a single haptic, if the contralateral
haptic is stable.
Dislo cat e d In t rao cular Le n s Re m o val
On occasion, marked inflammation or retinal problems will
necessitate removal of a dislocated IOL. A total vitrectomy
should be performed as described previously and the lens
optic grasped with forceps. The IOL should be brought
anteriorly and a scleral plug placed in the endoilluminator
opening. A limbal or clear corneal section of appropriate
size should be made with the surgeon’s other hand using
the diamond knife-scissors technique also used for very
dense nuclei and large intraocular foreign bodies. The IOL
can be passed from the 25-gauge forceps to a second forceps
such as 20-gauge diamond-coated forceps through the lim-
bal incision for explantation. After the IOL is removed, the
limbal wound is closed with a running shoelace or X-type
10-0 monofilament nylon suture. Smaller diameter sutures
cause less astigmatism but require care to prevent wound
leaks during vitrectomy.
In t rao cular Le n s Re m o val Be fo re
Tran s–Pars Plan a Vit re ct o m y
Occasionally, IOLs should be removed because of fibro-
vascular membranes from diabetic anterior vitreous cortex
fibrovascular proliferation, trauma, or uveitis. The hap-
tics should be cut with scissors with viscoelastics used to
maintain the anterior chamber. The haptics can be left in
the eye if they are surrounded by a fibrous capsule and
cannot be rotated out with forceps. Multiple X-type 9-0
or 10-0 monofilament nylon sutures facilitate tight wound
closure.
Ep it h e lial In g ro wt h
Vitrectomy instruments can permit removal of the fibrous
tissue, anterior vitreous cortex, and lens remnants in the
treatment of epithelial ingrowth (35). Pretreatment of the iris
with laser photocoagulation aids in identification of the tis-
sue and its removal. It does not kill all abnormal cells but can
be advantageous in defining the extent of involvement. The
wound should be explored and the edges excised. An iridec-
tomy with a 1- to 2-mm margin should then be performed
with Vannas scissors. The wound should be closed tightly
with a running shoelace, 8-0 monofilament nylon suture.
Internal fluid-air exchange (Chapter 4) should be performed
by injecting air through the infusion system while allowing
fluid egress with the vitrectomy instrument. Transcorneal
and transscleral cryopexies are then applied over affected
areas with a 2-mm margin. The thermal insulating effect of
the air bubble causes more uniform destruction of abnormal
tissue with less scleral damage.
The cytotoxin, 5-fluorouracil, has been used in con-
junction with the surgical approach by some investigators in
the hope of inhibiting regrowth of epithelial cells. Its long-
term efficacy is uncertain. Fortunately, ingrowth through
cataract wounds is extremely rare because of small incision
techniques and advanced wound construction.
References
1. Michels RG. Anterior segment and vitreoretinal surgery through
the pars plana, Part I. Ann Ophthalmol 1976;8:1353.
2. Michels RG. Anterior segment and vitreoretinal surgery through
the pars plana, Part V. Ann Ophthalmol 1976;8:1497.
3. Michels RG, Stark WJ. Vitrectomy technique in anterior segment
surgery. Trans Am Acad Ophthalmol Otolaryngol 1976;81:382.
4. Taylor HR, Michels RG, Stark WJ. Vitrectomy methods in ante-
rior segment surgery. Ophthalmic Surg 1979;10(10):25.
5. Michels RG. Anterior segment applications of vitrectomy tech-
niques. Trans Ophthalmol Soc UK 1978;98(4):458.
6. Stark WJ, Michels RG. Anterior segment surgery using instru-
ments designed for pars plana vitrectomy. Trans Penn Acad
Ophthalmol Otolaryngol 1981;34:27.
7. Michels RG, Paton D. Results of radical anterior vitrectomy: a
preliminary report of 26 cases. Ophthalmic Surg 1970;1:33.
8. Charles S. Anterior segment vitrectomy. In: Carroll D, ed. Sur-
gery of the eye. New York: Churchill Livingstone, 1988.
9. Charles S. The Charles anterior segment infusion sleeve. Ocu-
tome Fragmatome Newsl 1978;3(2):6.
10. May D. Closed vitrectomy for vitreous prolapse during cataract
extraction. Ocutome Fragmatome Newsl 1979;4(2):2.
11. Girard LJ, Hawkins RS. Cataract extraction by ultrasonic aspi-
ration vitrectomy by ultrasonic aspiration. Trans Am Acad Oph-
thalmol Otolaryngol 1974;78:50.
12. Charles S. Trans-pars plana and posterior chamber lensectomy
with the Girard phacofragmenter and automated suction. In:
Emery J, ed. Current concepts in cataract surgery, selected pro-
ceedings of the Fifth Biennial Cataract Surgical Congress. St. Louis:
C.V. Mosby; 1978.
13. Charles S. Trans-pars plana lensectomy update. Ocutome Frag-
matome Newsl 1980;5(3).
14. Michels RG, Shockett DE. Vitrectomy techniques for removal
of cataract lens material. Arch Ophthalmol 1977;95(10):1767.
 CHAPTER 7 ■ Vitrectom y Techniques and Technology for Anterior Segm ent Problem s
15. Morse PH. Intracapsular cataract extraction at the time of
vitrectomy. Ophthalmic Surg 1979;10(3):65.
16. Parks MM, Hiles DA. Management of infantile cataracts. Am J
Ophthalmol 1967;63:10.
17. Parks MM. Posterior lens capsulectomy during primary cata-
ract surgery in children. Ophthalmology 1983;90:344.
18. Calhoun JH. Cutting-aspiration instruments. Int Ophthalmol
Clin 1977;17(4):103.
19. Chrousos GA, Parks MM, O’Neill JF. Incidence of chronic glau-
coma, retinal detachment and secondary membrane surgery in
pediatric aphakic patients. Ophthalmology 1984;91:1238.
20. Hamill MB, et al. Pediatric cataracts. Curr Opin Ophthalmol
1999;10(1):4–9.
21. Malukiewicz-Wisniewska G, et al. Intraocular lens implanta-
tion in children and youth. J Pediatr Ophthalmol Strabismus
1999;36(3):129–133.
22. Simons BD, et al. Surgical technique, visual outcome, and com-
plications of pediatric intraocular lens implantation. J Pediatr
Ophthalmol Strabismus 1999;36(3):118–124.
23. Zwaan J, et al. Pediatric intraocular lens implantation. Surgical
results and complications in more than 300 patients. Ophthal-
mology 1998;105(1):112–118.
24. Treister G, Machemer R. Pars plana approach for pupillary
membranes. Arch Ophthalmol 1978;96(6):1014.
25. Wilkinson CP, Ramsey JJ. Closed vitrectomy for the vitreous
touch syndrome. Am J Ophthalmol 1980;90(3):304.
97
26. Snip RC, Kenyon KR, Green WR. Retrocorneal fibrous
membrane in the vitreous touch syndrome. Am J Ophthalmol
1975;79(2):233.
27. Irvine A. Pars plana vitrectomy for malignant and aphakic
pupillary block glaucoma. Trans Pac Coast Otoophthalmol Soc
1977;58:189.
28. Orth DH, Henry MD. Management of Irvine Gass Syndrome
using argon laser photocoagulation and pars plana vitrectomy.
Presented at the Bicentennial Cataract Surgical Congress,
Miami Beach, Florida, February 1977.
29. Rice TA, Michels RG. Vitreous wick syndrome-current surgical
management. Am J Ophthalmol. 1978;85:656.
30. Fung WE. Vitrectomy for chronic aphakic cystoid macular
edema. Ophthalmology 1985;92:1102.
31. Pendergast SD, et al. Vitrectomy for chronic cystoid macular
edema. Am J Ophthalmol 1999;128(3):317–323.
32. Ikeda T, et al. Vitrectomy for cystoid macular edema with
attached posterior hyaloid membrane in patients with diabe-
tes. Br J Ophthalmol 1999;83(1):12–14.
33. Holekamp NM. Treatment of pseudophakic CME. Ocul Immu-
nol Inflamm 1998;6(2):121–123.
34. Wilson DL. A new intraocular foreign body retriever. Ophthal-
mic Surg 1975;6(4):64.
35. Stark WJ, Michels RG, Maumenee AE, et al. Surgical manage-
ment of epithelial ingrowth. Am J Ophthalmol 1978;85(6):772.
C H AP TER
8
COMBINED
PHACOEMULSIFICATION
VITRECTOMY PROCEDURES
Many surgeons now combine phacoemulsifica-
tion with vitrectomy for a variety of reasons. This
approach adds significant complexity but is indi-
cated in certain situations. There is a widespread
but incorrect notion that vitrectomy inevitably leads to cata-
ract. It is widely appreciated that vitrectomy leads to progres-
sion of preexisting nuclear sclerosis, likely due to ascorbic
acid depletion and resultant permanent increase in the partial
pressure of oxygen in the vitreous cavity by 7 to 12 mm Hg
(Holekamp, Chang, Steffanson). BSS Plus (Alcon) has been
available for three decades (Edelhauser) and has eliminated
the development of posterior subcapsular cataract during
the procedure, yet many surgeons use BSS or, even worse,
lactated Ringer’s solution. Other surgeons choose inappro-
priately to add a variety of compounds including bicarbon-
ate, dextrose, antibiotics, and epinephrine, all of which can
lead to posterior subcapsular cataract, especially if infrequent
mixing errors do not occur. Bicarbonate is unnecessary and
inappropriate to use with BSS Plus because it is correctly buff-
ered without additives. Dextrose was added three decades
ago when diabetic patients were often markedly hyperglyce-
mic during surgery, but the widespread availability of serum
glucose monitoring during surgery has eliminated this issue.
Contact of the posterior lens with a gas bubble over a period
of several days leads to posterior subcapsular cataract; this
occurs only if the patient does not maintain the correct posi-
tion and when there is discontinuity in the anterior vitreous
cortex. Patient education is very important. Some of what is
described as poor compliance by the patient should be attrib-
uted to poor patient education. The vast majority of younger
patients with a clear lens undergoing vitrectomy will retain
a clear lens for decades if BSS Plus without additives is used
and gas bubble contact with the lens is avoided. In short,
cataract is not inevitable after vitrectomy.
Optimal visualization is essential for vitrectomy, espe-
cially if epiretinal or internal limiting membrane (ILM)
98
dissection, drainage of subretinal fluid through retina breaks,
or retinopexy is required. Posterior subcapsular cataracts
interfere with visualization more than nuclear sclerotic cata-
racts. If the surgery is elective, as is typically the case with
macular surgery, cataract surgery can be performed 1 month
before vitrectomy if the cataract is likely to interfere with
visualization during vitrectomy. A potential problem with
pre–pars plana vitrectomy (pre-PPV) cataract surgery is the
accurate determination of axial length. The Zeiss IOL Master
uses the retinal pigment epithelium (RPE) for measurement
in contrast to A-scan ultrasound that uses the ILM. Mea-
surement from the RPE is not a problem with epimacular
membranes (EMMs) or vitreomacular traction syndrome;
however, the A-scan ultrasound axial length may be reduced
when these conditions are present or increased when mac-
ular holes are present. Fixation is an additional issue with
macular disease; it may be difficult to determine if the axial
length is measured in the fovea or an extramacular region
with either technology.
Cataract surgery performed as a separate procedure
after PPV permits more accurate axial length measurement.
The procedure can be performed before or shortly after PPV
if cataract interferes with examination of the retina and a
retinal detachment (RD) is present. For this scenario, the
options are phaco-vit or pars plana lensectomy with com-
plete removal of the posterior capsule. Although endocapsu-
lar lensectomy with retention of the anterior capsule followed
by intraocular lens (IOL) implantation in the ciliary sulcus
is possible, this typically results in severe capsular opacifica-
tion. Therefore, the anterior capsule could not be utilized as
a barrier for silicone oil posterior retention. The lensectomy,
complete removal of the capsule with forceps and inferior
iridectomy, approach has been advocated for these cases, but
phaco-vit approach is now favored because it enables the
preservation of a silicone oil barrier and optimal correction
of aphakia.
 CHAPTER 8 ■ Com bined Phacoem ulsification Vitrectom y Procedures 99

COMPARISON AMONG SEQUENTIAL PHACO AND VITRECTOMY, COMBINED

PHACO-VITRECTOMY, AND PARS PLANA LENSECTOMY/ CAPSULECTOMY
SEQUENTIAL PHACO
AND PPV
Refractive Outcome Good if axial length can be
determined
Difficulty Level Easiest
Best Suitable for Isolated macular pathology
and elective cases
IOL Considerations Usually not an issue except
cannot use silicone lenses
Surgical Equipment Usually not an issue
Cost Usually not an issue
COMBINED PHACO-PPV
Difficult to predict, especially in
silicone oil and EMM, MH cases
Difficult (successful vitrectomy requires
near perfect phaco outcomes to maintain
corneal clarity and pupillary dilation)
Nonelective cases with significant
cataracts (e.g., RD with proliferative
vitreoretinopathy [PVR] stage C or less)
IOLs can be subluxated forward when
gases are used, leading to IOL optic
capture in pupil
Must use a vitrectomy machine with
combined phaco and vitrectomy
capabilities, otherwise needs two
separate machines
Significantly higher costs (viscoelastic,
IOL, time, surgical packs). May be cost
prohibitive in the ambulatory surgery
center setting
PARS PLANA LENSECTOMY/
CAPSULECTOMY
Worst; effective lens position
difficult to determine except for ACL
Moderately difficult
RDs with anterior PVR (stage D),
markedly inflamed eyes
Best to leave eye aphakic and only
consider secondary IOL at a later
date
May require enlargement of one
25-gauge sclerotomy to a 20-gauge
size for fragmenter in hard nuclei
Only additional surgical cost is
marginal (the use of the fragmenter)

ACL, anterior chamber lens; EMM, epimacular membrane; MH, macular hole.

[IOP] maintenance, nonleaking wounds, rigid aspiration

SILICONE OIL ISSUES
system), which are equally essential for vitrectomy.
Phaco and vitrectomy are both safer and more effective
Silicone oil has a different index of refraction than vitreous.
when a closed system without leaks is utilized. The principal
Because the posterior surface of most IOLs is convex, instead
author has long recommended using 45 mm Hg, unless per-
of plano, the usual IOL power calculations cannot be used.
fusion pressure is compromised, just as Mackool advocates
Although many surgeons think silicone oil must be removed
elevated infusion pressure for phaco. Two-handed methods
after a certain number of months, this is simply not true. The
are best for both procedures both for optimal access and
incidence of silicone oil–related glaucoma is approximately
for surgical manipulation. A spatula, chopper, or nucleus
11% (1), silicone oil is not toxic to the retina, and corneal com-
manipulator is used to hold back the capsule and manipu-
plications are even less common. A not uncommon scenario is
late and position lens material similar to using the endoil-
PPV plus silicone oil to accomplish reattachment, removal in
luminator to hold back retina or peel vitreous with bimanual
several months followed immediately by redetachment, and
spreading technique.
reoperation with replacement of the silicone oil. There is sim-
High vacuum, a flared tip, and low flow (for sculpting
ply no need to remove oil contained behind a posterior cham-
the nucleus) promote efficiency without surge. This concept
ber lens; the refractive effect is approximately four diopters (D),
is the same as the principal author’s port-based flow-limiting
and contrary to popular belief, oil does not cause decreased
concept for PPV. The author developed linear (proportional)
vision. If there is a full fill behind a posterior chamber intraoc-
control of vacuum for vitrectomy, and this was subsequently
ular lens (PCL), change in focus with changing head position
applied to all phaco machines for aspiration control of both
as well as emulsification is minimized. IOL calculations must
flow and vacuum. Unfortunately, many PPV surgeons make
anticipate the ultimate retention or removal of oil. There is a
little use of linear control, fully depressing the pedal and
greater impetus to removal of oil in younger patients.
changing vacuum settings on the console. Mackool empha-
sizes avoiding position 2 (aspiration) or 3 (aspiration and
GENERAL PHACOEMULSIFICATION ultrasound) unless actually intending to remove material at
CONCEPTS that time. The surgeon may elect to engage the “continuous
irrigation” mode in the Constellation and Infinity machines,
Many key phaco concepts are similar to vitrectomy princi- which maintains constant irrigation despite foot pedal posi-
ples. Richard Mackool and others emphasize the importance tion and avoids unintended position 0 of the foot position.
of maintaining optimal visualization, frequent adjustment of The continuous irrigation mode can be toggled on and off
microscope focal plane, and stability (intraocular pressure when entering and exiting the eye to save irrigation fluid.
100 SECTION II ■ Surgical Technology and Techniques
There are also essential differences between phaco and
PPV; the phaco tip should be positioned near the center of
the chamber and lens fragments mobilized to the phaco tip
by the use of an appropriate aspiration flow rate to protect
the capsule and iris. In marked contrast, the vitreous cutter
should be brought to the vitreous; vitreous should not be
drawn to the port by high vacuum.
GENERAL PHACOEMULSIFICATION
PRINCIPLES
A small, tapered (reduced incision width), nonleaking
sideport incision should be constructed. This can be done
with the Alcon I-Knife. The entry should be created at a
location that will be 2 to 2.5 clock hours from the phaco
incision prior to making the latter incision. The separation
between the sideport and the keratome incision should be
refined to allow for surgeons “comfortable hand position.”
The globe can be stabilized with an index finger placed
against the nasal conjunctiva during the creation of these
incisions.
For microincisional surgery, the temporal clear corneal
incision should be made with a 2.2-mm Alcon metallic ker-
atome with a superior bevel (diamond knives can result in
inaccurate incision size if not inserted and removed at the
exact same meridian). The initial entry should be made with
the blade at approximately 10-degree angle to the cornea.
After entering the cornea for 0.25 to 0.5 mm, the angle of
the blade is lowered so that it is parallel to the cornea. When
the blade has penetrated the cornea so that the 2-mm mark
on its anterior surface reaches the external incision, the tip of
the blade is directed slightly downward (parallel to the iris)
and the anterior chamber is entered.
The anterior capsule should be stained with trypan blue,
and this can be done efficiently with one injection from a
syringe that sequentially delivers air, trypan blue, and then
BSS to remove as much of the trypan blue as possible from the
chamber prior to viscoelastic injection. Sequential injections
of Viscoat (anterior) and Provisc or other hyaluronic acid–
based ophthalmic viscoelastic device (OVD) posterior to the
Viscoat is done to replace the air with clear viscous material.
The continuous curvilinear capsulorhexis (ccc) should
be initiated centrally with a bent needle, forceps, or cystotome.
Capsulorhexis forceps (Mackool Microincisional Capsulor-
hexis Forceps) are then used to create the ccc. It is important
to regrasp the elevated flap at a position that is relatively close
to the margin of the advancing tear during the creation of the
capsulorhexis.
Hydrodissection using a flat 25-gauge cannula to ele-
vate the anterior capsule prior to injection of BSS results in
the separation of lens material from the capsule (capsular-
cortical cleavage). Hydrodissection is performed in at least
two locations, usually 180 degrees apart. After each injec-
tion, the nucleus should shift forward and should be repo-
sitioned by gentle depression with the cannula prior to
repetition of the hydrodissection maneuver. Viscodissection
is then accomplished by injecting Viscoat beneath the
anterior capsule for 1,800 opposite the phaco incision. In
eyes with evidence of zonular laxity, viscodissection should
be performed for 360 degrees.
The nucleus can now be rotated in order to verify that
the hydrodissection or viscodissection has been successful;
however, this maneuver is not mandatory if the surgeon is
confident that separation has been successful.
A Mackool chopper or other chopper or spatula is
introduced through the sideport incision, and the phaco
tip is inserted through the primary incision without the
need to fixate the latter with forceps. Viscoelastic material
is aspirated from the surface of the lens in order to prevent
obstruction of the phaco tip by OVD. The nucleus is then
sculpted and divided by either chopping or other nuclear
division methods. Mackool prefers to sculpt a central bowl
of approximately two phaco tips (2 mm) in width prior
to performing phaco chop, and he stabilizes the nucleus
during sculpting by placing the Mackool chopper over
the edge of the equatorial nucleus opposite the phaco
incision.
Note that a nonleaking phaco incision is extremely
important to reduce intraocular turbulence and excessive
fluid flow through the eye. Rotation of instruments utilizing
the incisions as if they were an oarlock prevents distortion
of the cornea and compression of the phaco sleeve against
the vibrating tip (this reduces the risk of incision burn). The
authors prefer the Alcon 450 mini-flared tip with enhanced
Ultrasleeve and torsional phacoemulsification.
Chopping of the nucleus is an excellent, efficient
technique with the exception of nuclei that are elastic or
extremely dense (red-black nuclei). In the former case, high-
vacuum levels can be used to remove the peripheral lens
material until only a small central nuclear plate remains. The
latter can then be elevated with a spatula and removed.
Red-black nuclei are best divided into two heminuclei
by sculpting a long, very deep central trough prior to crack-
ing with a spatula and phaco tip.
Whenever possible, use both instruments to rotate the
nucleus and/or quadrants in order to reduce the amount of
stress placed on the lens zonules. Try to avoid the placement
of the phaco tip in the peripheral regions of the posterior
chamber, and after one or two nuclear quadrants have been
removed, a spatula should be placed posterior to the phaco
tip during subsequent nucleus removal in order to prevent
anterior movement of the lens capsule (infusion misdirec-
tion syndrome) and possible capsule aspiration.
A 900 I/A tip with a small (0.25 mm) port permits
access to cortex for 360 degrees. In this manner, cortex can
be removed with the port of the I/A tip facing anteriorly.
Note that the PPV can be performed prior to inser-
tion of the IOL. This is highly desirable as a more uniform
view of the peripheral vitreous cavity is obtained when not
looking through the edges of the optic after completion of
the vitrectomy. The only risk of delaying the introduction
 CHAPTER 8
of the IOL is the possibility of inadvertent capsular tear
during anterior vitrectomy, which is prevented by the
presence of the IOL creating tension on the capsule and
diminishing its mobility. The capsular sac is then filled
with Provisc prior to injecting the single-piece acrylic IOL
via a Monarch injector and D cartridge. Do not remove
the viscoelastic until the haptics have released from the
optic. If necessary, their release can be facilitated by using
a chopper to place traction on a haptic while the optic is
stabilized with the I/A tip. After the haptics have opened,
the viscoelastic can be aspirated with the 900 I/A tip, fol-
lowed by stromal hydration of both the primary and the
sideport incisions.
PHACO-VITRECTOMY TECHNIQUES
Combined phacoemulsification with intraocular lens (IOL)
implantation and vitrectomy is a safe technique when used
in the appropriate patients (2). The threshold for the use of
iris expander is lower than for standard phaco. For example,
a case with a pupil size of 4 mm or less may be done without
problems as standard phaco since anterior chamber infu-
sion during phaco maintains pupillary size but may develop
marked miosis when the anterior chamber infusion is ter-
minated at the end of the phaco. Early use of iris expanders
prevents severe miosis and a very difficult vitrectomy. If iris
expanders are used, do not remove until the completion of
the vitrectomy.
Iris prolapse through the cataract incision leads to mio-
sis during vitrectomy. Keratome wounds should be made 2 ×
2 mm, which is longer than some phaco surgeons use, in
order to prevent iris prolapse.
Staining of the anterior lens capsule with trypan blue is
usually required. The red reflex can often be poor in cases
that need combined phaco-vitrectomy approach, usually
from coexisting vitreous hemorrhage or RD. The need for
capsular staining may be underestimated if the only criterion
used is the appearance of the lens. Since adequate continu-
ous curvilinear capsulorhexis (CCC) is critical in combined
phaco-vitrectomy approach, it is best to use trypan blue
staining in the majority of cases.
Viscoelastic can be left in after phaco to maintain cham-
ber depth and pupil size. In the setting of combined vit-
rectomy, IOP elevation in the early postoperative period is
rare. If silicone oil is to be used at the end of the vitrectomy,
the surgeon should remove the viscoelastic from the ante-
rior chamber immediately before the oil fill, and it should be
replaced with air. This prevents oil emulsification from the
viscoelastic interaction.
The cataract incision should be sutured with 10-0 nylon
in all phaco-vitrectomy cases. This prevents failure of the
self-sealing clear corneal incision mechanism if hypotony
occurs during the vitrectomy or postoperatively.
If the posterior lens capsule is not clear at the end of
the phaco and this interferes with visualization during the
■ Com bined Phacoem ulsification Vitrectom y Procedures 101
vitrectomy procedure, a small posterior capsulectomy can be
performed with the vitreous cutter to improve visualization
of the retina after the IOL is well positioned. The vitreous
cutter is ideal for the construction of a posterior continuous
curvilinear capsulorhexis. This rarely leads to problems other
than potential fogging of the IOL if air is used as vitreous
infusion or allowing silicone oil access to the AC. YAG laser
capsulotomy is ineffective in the presence of silicone oil.
Higher levels of postoperative inflammation are typical
in phaco-vitrectomy cases. Very frequent dosing of topical
steroids in the early postoperative period (e.g., every hour)
with a rapid taper to qid after 1 or 2 weeks postoperatively
often prevents inflammatory complications.
Even if the visual potential of an eye is poor, there are
still benefits of placing an IOL in an eye that will require
long-term silicone oil retention is planned. The IOL can
form a barrier to the forward movement of the silicone and
prevent emulsification glaucoma and keratopathy. In addi-
tion, aphakic eyes with long-term silicone oil fill contain-
ing an inferior RD may develop neovascular glaucoma and
require Ahmed valve placement. In these eyes, if the eye is
aphakic, the silicone oil can exit the vitreous cavity through
the valve into the subconjunctival space, causing an inad-
equate intraocular oil fill that, in turn, causes progression of
the previously contained RD.
PHACOEMULSIFICATION PRINCIPLES
FOR SILICONE OIL–FILLED EYES
Because silicone oil “floats” over aqueous, oil will frequently
push forward on the posterior lens capsule causing something
similar to “positive pressure” during phaco. This is likely the
main cause of complications during the procedure. A tight seal
of the cataract incision around the phaco probe significantly
decreases turbulence and intraoperative incision leaks and sta-
bilizes capsular position and chamber depth. The height of the
infusion bottle with gravity-fed infusion should be elevated to
the maximum to maintain lens capsule inflation during the
case. If a small bubble of silicone oil enters the AC through
intact zonules, it is best to ignore it. All cortical fragments must
be removed without damaging the capsule. In standard phaco,
one may leave small cortical fragments without problems. If
a cortical fragment is left in a silicone oil case, it may not be
reabsorbed as expected and can cause inflammation and oil
emulsification. It is best to use bimanual I/A or angulated, that
is, 90-degree, I/A tip to remove subincisional cortex. If a small
posterior capsular tear occurs during the last phase of the I/A,
immediate introduction of an OVD or IOL can seal the hole and
prevent further oil bubbles from coming into the AC. If a large
capsular tear is encountered, insertion of a multipiece acrylic
IOL into the ciliary sulcus and capture of the lens optic through
the capsulorhexis can be done to isolate the AC from the oil; if
this cannot be done, it may be best to avoid IOL placement and
proceed with total capsular removal to prevent leaks of oil into
the AC that can lead to silicone oil glaucoma and keratopathy.
102 SECTION II ■ Surgical Technology and Techniques
ISSUES REGARDING INTRAOCULAR
LENS CALCULATIONS FOR SILICONE
OIL–FILLED EYES
Silicone oil impedes ultrasound A-scan measurements by
slowing of ultrasound waves and decreasing the intensity of
the posterior spike. If media clarity allows measurements to
be made with the IOL Master (which utilizes laser optical
measurements instead of ultrasound), this is the best option.
Otherwise, calculations based on the contralateral eye mea-
surements, prior refractions, etc. must be used. Therefore,
do biometry with IOL Master soon after initial surgery with
silicone prior to the onset of cataract and other possible reti-
nal complications such as pucker or recurrent RD, which
may affect measurement.
Because the refractive index of silicone oil is higher than
water, the refractive power of the IOL/silicone interface is
lower than the refractive power of the IOL/vitreous interface
(the refractive power is proportional to the ratio between
refractive indexes on both sides of the refractive interface,
based on Snell’s law). In effect, silicone oil diminishes the
power of the IOL, creating a hyperopic outcome. The higher
the lens power of the IOL placed, the higher the total loss of
its power in a silicone oil–filled eye, since the difference is
based on a percentage loss of power. For this reason, highly
myopic eyes tend to be closer to emmetropia than hyperopic
eyes when IOLs are placed in silicone oil–filled eyes, because
they require lower power IOLs. Theoretically, the IOL design
that provides the most reliable refractive results in silicone
oil–filled eyes is a planoconvex lens (the plano side faces the
Fig u re 8 .1 ■ A layer of aq ue-
ous wedged between the IOL and
the silicone oil can create m ultiple
refractive surfaces and p rism atic
effects that can b e im p ossib le to
predict or correct with an IOL.
silicone oil interface). This neutralizes the refractive interface
between IOL and silicone and places all the refractive power
on the front surface of the IOL. This design may be difficult
to obtain and is usually made of polymethyl methacrylate
(PMMA) and is not available with modern, foldable acrylic
material. Another source of error in refractive outcomes in
silicone oil eyes relates to the completeness of the oil fill. If
there is any degree of underfill, a layer of aqueous can wedge
between the IOL and the silicone oil (Fig. 8.1). This creates
multiple refractive surfaces and prismatic effects that can be
impossible to predict or correct with an IOL. If the goal is
long-term fill with silicone oil, the IOL can be calculated for
an emmetropic or mildly myopic result. If the goal is even-
tual removal of the silicone oil, placement of an IOL that is
appropriate for the eye after oil removal should be done.
The majority of the cases that require silicone oil and pha-
coemulsification with IOL placement are incapable of achiev-
ing excellent acuity due to the underlying retinal pathology;
there is normally no need to achieve a precise refractive result.
In the event of major refractive surprises or if the visual acu-
ity happens to be very good, the IOL can always be replaced
within the first month or two after implantation.
References
1. Al-Jazzaf AM, Netland PA, Charles SM. Incidence and manage-
ment of elevated intraocular pressure after silicone oil injection.
J Glaucoma 2005;14:40–46.
2. Hwang JU, Yoon YH, Kim DS, et al. Combined phacoemulsi-
fication, foldable intraocular lens implantation, and 25-gauge
transconjunctival sutureless vitrectomy. J Cataract Refract Surg
2006;32:727–731.
C H AP TER
9
25-GAUGE VITRECTOMY
Sutureless, transconjunctival 25-gauge vitrectomy
was introduced by DeJuan and colleagues (1,2)
and has rapidly become an accepted tool in the
vitreoretinal surgery armamentarium. Suture-
less vitrectomy is dependent on two concepts: (a) smaller
diameter instruments and therefore smaller sclerotomies and
(b) conjunctival displacement before making the transcon-
junctival sclerotomies so that the conjunctiva covers the scle-
rotomy after surgery is completed. Trocars are used to place
flexible, thin-wall cannulas, resulting in 23.5-gauge scleroto-
mies. The cannulas are used to maintain alignment between
the offset conjunctival incision and the sclerotomy and facili-
tate finding the small incisions when inserting tools.
CASE SELECTION
Initially it was thought that 25-gauge, sutureless vitrectomy
was only indicated for epimacular membranes, macular holes,
vitreomacular traction, and surgery for retinal venous occlu-
sion. Early on, many surgeons thought 25-gauge surgery was
inappropriate for vitreous hemorrhages, rhegmatogenous
retinal detachments, proliferative vitreoretinopathy (PVR),
diabetic traction retinal detachments, or giant breaks. Experi-
ence has shown, however, that 25-gauge surgery is ideal for
vitreous hemorrhages, rhegmatogenous retinal detachments,
PVR, and giant breaks and is applicable for diabetic traction
retinal detachments as well. Few surgeons now believe that
branch vein decompression (sheathotomy) and radial optic
neurotomy (RON) are effective procedures, making 25-gauge
surgery for these procedures a moot point except in the con-
text of vitrectomy alone for the reduction of macular edema
via vascular endothelial growth factor (VEGF) reduction in the
macula and increased oxygen tension in the vitreous cavity.
Dislocated lens material (dropped nucleus) and removal
of intraocular foreign bodies (IOFBs) require one 20-gauge
sclerotomy because dense membranes have too much flow
resistance for small-lumen instruments and IOFBs are too
large. The technique of combining two 25-gauge suture-
less incisions with one 20-gauge sutured incision has been
described as 20/25 vitrectomy by the authors.
Patients with prior or anticipated glaucoma-filtering
procedures are ideal candidates for 25-gauge surgery.
Patients with severe dry eyes, ocular surface disorders, and
scarred conjunctiva are excellent candidates for 25-gauge
surgery as well.
TROCAR-CANNULA SYSTEM
The purpose of the trocar is to make a 25-gauge sclerotomy
and enable simultaneous insertion of flexible 23.5-gauge self-
retaining cannula that fits over the trocar (Fig. 9.1). The con-
junctival incision is intentionally displaced from the scleral
incision so that the two incisions will not be aligned after
cannula withdrawal and the conjunctiva will cover the scle-
rotomy (Fig. 9.2). Repeated insertion and withdrawal of tools
is accomplished through the cannula, which maintains the
alignment of the conjunctival and scleral incisions and pro-
tects the wounds. The conjunctiva should be displaced using a
cotton-tip applicator by gently moving it anteriorly or circum-
ferentially or some combination thereof (Fig. 9.3). An attempt
should be made to avoid conjunctival and scleral vessels
to reduce postoperative subconjunctival hemorrhages. The
Alcon noncoring trocar is based on a modified microvitreoret-
inal (MVR) blade and requires much less insertion force than
competitive hypodermic needle-based, coring-type designs.
A prospective, consecutive clinical series by the author shows
that the Alcon system using fluid-air exchange demonstrates
significantly less hypotony than published results with the
coring-type trocar. The Alcon trocar cannula usually requires
no rotation when inserted. If some resistance is encountered,
small-amplitude back and forth rotation will facilitate inser-
tion. The cotton-tip applicator should be handed back to the
scrub technician or dropped so that the forefinger of this hand
can be used to guide the trocar cannula during insertion. The
trocar should be aimed toward a virtual point about 2 mm
anterior to the center of the eye or about 2 mm posterior to
the lens. The incisions should be made 4 mm posterior to the
limbus (Fig. 9.4) unless there is a pars plana abnormality from
retinopathy of prematurity (ROP), trauma, or pars planitis; a
large choroidal detachment; suprachoroidal hemorrhage; or
103
104 SECTION II ■ Surgical Technology and Techniques

Fig ure 9 .1 ■ A 25-gauge trocar cannula from Alcon.

high retinal detachment that would necessitate making the

FLUIDICS
incisions through the pars plicata. The 25-gauge incisions
may be made 3 mm posterior to the limbus in aphakic eyes.
Ohm’s law for fluids is directly analogous to Ohm’s law for
The inferotemporal incision should be made just below the
electricity and teaches that pressure (gradient) is equal to the
3 or 9 o’clock position to reduce bleeding and pain but as far
resistance times the flow. Resistance to flow is proportional
as possible from the lower lid so that it will not be displaced if
to the fourth power of the inner diameter of the lumen. The
the eye is rotated inferiorly. The superonasal incision is usually
much higher resistance of 25-gauge cutters and infusion
made on a virtual line from the lowest point of the bridge of
cannulas was initially thought by many to be a major dis-
the nose extending through the center of the pupil and then
advantage. It turns out that increasing infusion pressure to
plugged with a special plug made for 25-gauge cannulas. This
50 to 60 mm Hg while flow is occurring (dynamic state) and
location reduces tool flexion issues and facilitates peripheral
lowering it to 35 to 45 mm Hg (static state) when using for-
and anterior access. The purpose of the plug is to prevent vit-
ceps, pics, scissors, or the endophotocoagulator (static state)
reous prolapse or fluid loss through the port while making the
solves the problem. The Alcon Constellation pressurized
third incision. The superotemporal incision is usually made on
infusion system and the Accurus Vented Gas Forced Infusion
a virtual line extending from the lowest point of the supraor-
(VGFI) allow rapid and accurate switching between static
bital rim through the center of the pupil again to reduce tool
and dynamic infusion pressures, and the Constellation pres-
flexion and facilitate anterior and peripheral access.
sure compensation system directly addresses this problem.

Fig u re 9 .2 ■ Wound leaks are p revented b y conjunctival disp lacem ent with resp ect to the
0.5-m m scleral incision.
 CHAPTER 9 ■ 25-Gauge Vitrectom y 105

Fig u re 9 .3 ■ Circu m feren tial

or anterior conjunctival disp lace-
m ent with a cotton-tip actuator
is done before entry of the trocar
cannula.

The 23/25/27-gauge cutters have more resistance than
20-gauge cutters because of the smaller bore coaxial inner
needle as well as the cutter intermittently closing the port
during the open-close cycle, thus requiring greater com-
pensation with respect to the vacuum settings used with
20-gauge vitrectomy. The authors use 650 mm Hg maxi-
mum setting and proportional (linear) suction and the Con-
stellation system. Fast cutting with 20-gauge cutters is also
advantageous because of port-based flow limiting, which
results in decreased pulsatile vitreoretinal traction. The
smaller bore of the 23/25/27-gauge cutters also produces
port-based flow limiting (Fig. 9.5). The 25-gauge cutter pro-
duces one half the pulse flow of the 2,500-cpm 20-gauge cut-
ters and therefore has 2× greater fluidic stability. In addition,
port-based flow limiting reduces surge when dense epireti-
nal membrane (ERM) suddenly deforms through the port

Fig u re 9 . 4 ■ In fero t em p o ral

trocar-cannula entry should be 3.5
m m p osterior to the lim b us just
inferior to the 3 or 9 o’clock p osi-
tion. Sup eronasal trocar-cannula
entry should b e m ad e 3.5 m m
p osterior to the lim bus on a vir-
tual line extending from the low-
est point of the bridge of the nose
through the center of the p up il.
Sup erotem p oral trocar-can n ula
entry should b e m ad e 3.5 m m
p osterior to the lim bus on a vir-
tual line extending from the low-
est p oint of the sup raorb ital rim
through the center of the pupil.
106 SECTION II ■ Surgical Technology and Techniques

Fig ure 9 .5 ■ The sm aller lum en

of a 25-g aug e cutter p rod uces
p ort-b ased flow lim iting like fast
cuttin g , wh ich in creases fluid ic
stab ility an d red uces surg e an d
iatrog en ic b reaks after sud d en
elastic d eform ation of the ERM
through the port.

analogous to occlusion break surge in phaco procedures. Fast Multifunction tools such as the end-aspirating laser
cutting has the additional advantage of reducing vitreoretinal probe (Chang) and the disposable bipolar endoilluminator
traction by minimizing collagen fiber travel through the port. are not available in a 25-gauge diameter.
The authors recommend using a cutting rate of 5,000 cpm
To o l Fle xio n
for all 25-gauge tasks and cases and have found that core
vitrectomy times are very reasonable when using 650 mm Tool flexion is the most significant complaint about 25-gauge
Hg vacuum. systems, but this problem can be overcome using certain tech-
niques. The second-generation Alcon endoilluminators and
endolaser probes are 58% stiffer than the first-generation tools.
CUTTER DESIGN OPTIONS Careful positioning of the sclerotomies reduces tool flexion.
The superonasal incision should be placed aligned with the
Pneumatic disposable cutters weigh about 1/14 as much as lowest point on the bridge of the nose. The superotemporal
available electric cutters, thereby increasing dexterity, reduc-
ing fatigue, and reducing flexion of the shaft. Pneumatic
cutters vary (optimize) the duty cycle as a function of cut-
ting rate, which is an advantage compared to the fixed duty
cycle of electric cutters, because of the wide range of materi-
als properties encountered during vitrectomy (from air to
perfluorooctane to vitreous to dense ERM).

TOOL ISSUES

Disposable forceps and scissors are advantageous for 25-gauge

surgery (3). The advantages of disposable 25-gauge instru-
ments include elimination of sterilization and cleaning issues
(Toxic Anterior Segment Syndrome (TASS)-like inflammation
from retained biological material and enzymes from ultrasonic
instrument cleaning) and the inevitable damage that occurs Fig ure 9 .6 ■ Disposable 25-gauge Internal Lim iting Mem brane
to fragile reusable instruments. The authors use Alcon DSP forceps from Alcon are used to p eel ERMs, p osterior vitreous cor-
disposable forceps (Fig. 9.6) and scissors (Fig. 9.7). tex, and internal lim iting m em brane (ILM).

Fig ure 9 .7 ■ Disposable 25-gauge curved scissors from Alcon
are used for segm entation and delam ination.
incision should be placed at a low point on the brow, typically
just above the 3 or 9 o’clock meridian. The inferotemporal
incision should be placed just below the 3 or 9 o’clock merid-
ian. The patient’s head can be rotated toward the target pathol-
ogy during surgery to better view the periphery. The patient
can cooperate when using local anesthesia and the anesthesi-
ologist can help when general anesthesia is used.
A significant part of the tool flexion experienced by
surgeons is technique dependent. Proper technique should
Fig ure 9 .8 ■ A 20/ 25 endocapsular lensectom y requires higher infusion pressure and judicious use of
linear suction to prevent sudden hypotony.
CHAPTER 9 ■ 25-Gauge Vitrectom y 107
maintain the eye on the primary position or close to the
primary position with proper rotation of the instruments
around the pivot point of the sclerotomy and cannula. Contact
wide-angle lenses decrease the need for extreme eye rotation.
Scleral depression by the scrub technician or assistant can
provide peripheral access without probe-induced ocular rota-
tion. Contact-based wide-angle visualization systems (Volk,
AVI) require the surgeon to maintain the eye near the primary
position and rotating the instruments around the sclerotomies
instead of using translational forces to rotate the eye.
Using a 23-gauge approach reduces tool flexion but raises
significant questions about wound leaks, which demands
more challenging wound construction and leak issues.
COMBINED 25-GAUGE AND
20-GAUGE VITRECTOMY
Certain limitations of 25-gauge vitrectomy can be over-
come by enlarging one port to 20 gauge for certain tasks
(4). One sclerotomy can be enlarged for fragmenter use for
lensectomy (Fig. 9.8) or removal of dislocated lens material
(Fig. 9.9), or use of the Machemer-Parel diamond-coated for-
ceps for intraocular foreign body removal. The technique var-
ies depending on the indications previously described, but in
general, an 8-0 nylon sclerotomy suture cut on the knot was
used for the 20-gauge sclerotomy and an 8-0 Biosorb conjunc-
tival suture cut on the knot was used at the inflection point
108 SECTION II ■ Surgical Technology and Techniques
of an L-shaped mini-conjunctival flap. Utilizing one 20-gauge
sclerotomy with two 25-gauge sclerotomies to address the
mechanical properties of dense lens material or the physical
size of IOFBs (Fig. 9.10) was found to be safe and effective. The
20/25 technique can also be used for 20-gauge conformal for-
ceps, end-gripping diamond-coated forceps, the Chang end-
aspirating laser probe, the disposable bipolar endoilluminator,
Fig ure 9 .9 ■ A 20/ 25 rem oval
of d islocated len s m aterial m ust
be done after perform ing a com -
p lete vitrecto m y. A 2 5 -g au g e
infusion port requires higher infu-
sion pressure and judicious use of
linear suction to p revent sudden
hyp otony with a 20-g auge frag -
m enter.
or certain illuminated instruments for the so-called bimanual
surgery, although the authors do not use these methods. Use
of 20-gauge fragmenter with 25-gauge infusion requires eleva-
tion of the infusion pressure and proper technique with the
fragmenter, avoiding aspiration when there is no lens frag-
ment occluding the tip of the instrument. This prevents ocular
collapse from the mismatch of inflow and outflow of fluid.
Fig ure 9 .1 0 ■ A 20/ 25 rem oval
of IOFB with Machem er-Parel dia-
m ond-coated forceps from Alcon
Grieshaber should be done after
com p lete vitrectom y to red uce
vitreoretinal traction.

VISUALIZATION
Smaller diameter fibers can increase light losses, which can
be compensated for by using one of the new xenon sources,
such as the Accurus Xenon Illuminator. The Accurus Xenon
Illuminator has ISO standard filtering to produce a safe
white-appearing light with low aphakic hazard function
(AHF). The second-generation endoilluminators have twice
as much light throughput, have much larger divergence
angle, and are twice as stiff. The Alcon Constellation Vision
System incorporates dual two-port xenon illuminators that
are more efficient than first-generation xenon systems and
have longer lamp life.
WOUND LEAK ISSUES
Postoperative wound leaks have always created the
greatest concern about sutureless, 23/25-gauge technol-
ogy. Oblique wound construction produces a scleral tun-
nel that is self-closing when the intraocular pressure is
normal (Fig. 9.11); turning the infusion off at the end of
the case makes no sense. Repositioning the conjunctiva
with a cotton-tip applicator after cannula removal is essen-
tial but not sufficient; firm pressure should be applied with
the closed forceps over the scleral tunnel created by oblique
wound construction to close the outer wall of the tunnel.
Fig ure 9 .1 1 ■ Obliq ue wound construction p roduces a scleral tunnel, which is self-closing when the
intraocular pressure is norm al.
CHAPTER 9 ■ 25-Gauge Vitrectom y 109
Wound leaks can lead to hypotony, resulting in choroidal
detachments, bleeding, vitreous volume enhancement, or
reoperations to suture the wounds. Postoperative vitreous
wicks can occur with wound leaks and increase the risk
of endophthalmitis. Intraoperative wound leaks through
the cannula can result in bleeding, vitreous prolapse with
secondary vitreoretinal traction, miosis due to hypotony, or
even retinal prolapse. The author initially raised the issue
of conjunctival antibiotics gaining access to the vitreous
cavity through unsutured sclerotomies, resulting in retinal
toxicity, and anecdotal reports have substantiated this con-
cern. De Juan has advocated allowing vitreous to plug the
wounds, while the author introduced the idea of using flu-
id-air exchange to reduce wound leaks. Initially the author
recommended partial fluid-air exchange, but this left the
inferotemporal sclerotomy unprotected when the patient
was seated or standing; total fluid-air exchange (Fig. 9.12)
was later used for all cases not requiring gas or silicone oil
if there was a concern about wound leaks. Proper oblique
sclera wound construction has decreased the need for fluid-
air exchanges, and the authors currently do this technique
only rarely.
Wound leaks are more common in high myopes,
Marfan’s syndrome, and other patients with connective tis-
sue disorders and thin sclera. Many surgeons have noted
the role of vitreous incarceration in 20-gauge wounds as
a factor in postvitrectomy retinal detachment. The senior
author has emphasized that the surface tension effect of
110 SECTION II ■ Surgical Technology and Techniques

Fig ure 9 .1 2 ■ Fluid-air exchange can be utilized at the end of cases not requiring g as or silicone oil,
if there is a concern about wound leaks.

Fig ure 9 .1 3 ■ The conjunctiva m ust b e rep ositioned after the cannulas are rem oved so that the
conjunctival incision is not overlying the sclerotom y. The sclera should be sup ported while slowly with-
drawing the cannulas to prevent air, gas, or silicone loss and vitreous prolapse.

an air bubble can reduce or eliminate wound leaks and
prevent vitreous wicks until fibrin seals the wound just as
gas bubbles are used to “seal” retinal breaks and macular
holes. The author has long suggested that an air bubble in
the anterior chamber is the best way to keep vitreous out
of the anterior chamber and away from the wounds after
rupture of the posterior capsule in cataract surgery. Air is
also an excellent way to intraoperatively seal wound leaks
in trauma cases. The authors recommend removing as
much peripheral vitreous as possible in 20-gauge as well as
23/25-gauge cases.
POSTOPERATIVE ENDOPHTHALMITIS
AND RETINAL DETACHMENTS
Several surgeons have been concerned about a possible
increased risk of postoperative endophthalmitis and retinal
detachments (5) after 25-gauge vitrectomy. The rationale for
postoperative endophthalmitis would be postoperative entry
of bacteria into the globe through unsutured wounds. Since
the authors have started performing 25-gauge vitrectomy as
the standard surgery in their practice, they have only had
one case of postoperative endophthalmitis in over 700 cases.
This anecdotal case is not sufficient for them to presume an
increased risk of endophthalmitis in their practice. In terms
of prophylaxis, all patients undergoing 25-gauge vitrectomy
have the standard meticulous prepping with povidone-iodine
as well as postoperative injection of subconjunctival antibi-
otics. In addition, patients are instructed to remove the eye
patch upon returning home from surgery and start topical
antibiotics every hour while awake immediately postopera-
tively, attempting to maintain significant antibiotic concen-
trations in the ocular surface to decrease the conjunctival
bacterial load that may migrate through unsutured wounds.
The authors have not noticed an increase of postopera-
tive retinal detachments after 25-gauge vitrectomies in their
practice. Previous mentions of post–25-gauge vitrectomy
retinal detachments may be related to the surgical learning
curve associated with a new procedure rather than to a flaw
intrinsic to the equipment and techniques.
CHAPTER 9 ■ 25-Gauge Vitrectom y 111
CANNULA WITHDRAWAL
AND WOUND CLOSURE
Slow withdrawal of tools from cannulas reduces acute
vitreoretinal traction; think of the tool in the cannula as
a syringe with plunger being pulled back. Supporting the
sclera while removing the cannulas reduces wound eversion,
pressure on the globe, and the likelihood of vitreous pro-
lapse. The wounds should not be massaged after cannula
removal; the conjunctiva should be gently moved back (Fig.
9.13) into its original position. If a wound leak is detected, a
single 8-0 Biosorb suture can be placed through conjunctiva
and sclera to close these small wounds.
SUMMARY
25-gauge, sutureless vitrectomy requires modification of tech-
niques and settings but offers the advantage of better fluidic sta-
bility, faster visual improvement, much better patient comfort,
much less conjunctival damage, and, less importantly, tedium.
The time saved by not having to suture an infusion cannula in
place, make or close conjunctival incisions, and suture sclero-
tomies is largely offset by taking more time for the vitrectomy.
References
1. Fujii GY, De Juan E Jr, Humayun MS, et al. Initial experience
using the transconjunctival sutureless vitrectomy system for vit-
reoretinal surgery. Ophthalmology 2002;109(10):1814–1820.
2. Fujii GY, De Juan E Jr, Humayun MS, et al. A new 25-gauge
instrument system for transconjunctival sutureless vitrectomy
surgery. Ophthalmology 2002;109(10):1807–1812.
3. Capone A. 25 gauge pars plana vitrectomy: advances in instru-
mentation and broader indications. Program and abstracts of the
American Society of Retina Specialists 23rd Annual Meeting; July
16–20, 2005, Montreal, Canada.
4. Charles ST. 20/25 vitrectomy: combined 25 and 20-gauge vitrec-
tomy. Program and abstracts of the American Society of Retina Spe-
cialists 23rd Annual Meeting, July 16–20, 2005, Montreal, Canada.
5. Williams DF. The incidence of rhegmatogenous retinal detach-
ment (RRD) following 20 g versus 25 g vitrectomy. Program
and abstracts of the American Society of Retina Specialists 23rd
Annual Meeting, July 16–20, 2005, Montreal, Canada.
C H AP TER
10
20-GAUGE SUTURED WOUND
CONSTRUCTION AND CLOSURE
Whereas 25-gauge vitrectomy is usually
performed transconjunctivally with trocar-
cannula systems, 20-gauge vitrectomy usually
requires conjunctival incision, exposure of
the anterior sclera, and sclera closure and conjunctival
reapproximation with sutures. Despite performing nearly all
of their vitrectomies with 25-gauge techniques, the authors
still find the need for creation of 20-gauge incisions. These
may be required, for example, to introduce a 20-gauge
fragmenter handpiece to remove pieces of dropped dense
nucleus in the vitreous or to introduce diamond-coated 20-
gauge forceps to remove intraocular foreign bodies. This
chapter will focus on techniques of creation and closure of
20-gauge incisions.
CONJUNCTIVAL INCISIONS
Care should be taken to avoid highly scarred or vascular
areas, filtering blebs, setons, and flaps for sutured intraoc-
ular lenses (IOLs). Two temporal incisions are made, one
superotemporal for the cutter, scissors, forceps, fragmenter,
laser, etc. and another inferotemporal for the infusion can-
nula. They should be limbus-parallel, 1 mm posterior to the
limbus, and at the superior and inferior border of the lateral
rectus. The superonasal incision is usually for the endoil-
luminator and should be limbus-parallel, 1 mm posterior
to the limbus, and centered at the superior border of the
medial rectus. A small radial extension of the limbus-parallel
incisions at one end resulting in an L-shaped flap facilitates
better exposure and single suture closure.
For any vitreous surgery in which scleral buckling is
anticipated, it is best to make a continuous incision 1 mm
posterior to the limbus including each quadrant requiring
buckling. A limbal peritomy tends to bleed under the con-
tact lens, causes conjunctival retraction and redundancy
problems in the postoperative course, and makes postopera-
tive contact lens fitting difficult and dellen more common.
Incisions more than 1 mm behind the limbus increase the
112
risk of exposure of scleral sutures and make visualization of
the sclerotomies difficult because of the anterior flap.
TRACTION SUTURES
Traction sutures under the muscles are of no value in vitrec-
tomy because the instruments move the eye. Only if scleral
buckling is anticipated should traction sutures be used, and
then they should be placed beneath the muscles directly,
using a short-handled fenestrated muscle hook with cham-
fered hole. Large sutures (0 to 2-0 silk or cotton) do not cut
the muscle tendons and provide a good grip for the assis-
tant. Only the muscles absolutely required for access to the
sclera should be trapped to reduce postoperative pain, pto-
sis, and strabismus. The superior rectus should be avoided
if possible to avoid or reduce vertical strabismus as well as
ptosis secondary to levator and Mueller’s muscle damage. It
is never necessary to remove muscles for scleral buckling or
trauma repair in the author’s experience.
20-GAUGE SCLERAL INCISIONS
Sclerotomies for the endosurgical instruments will be focal
points for the remainder of the operation (1). If they are
incorrectly placed or constructed, the entire course of the
operation can be adversely influenced. The incisions should
be placed in the anterior pars plana to reduce the risk of
damage to the peripheral retina. Contrary to customary
teaching, bleeding of the ciliary body is infrequent and rarely
serious when it does occur. Except in children, patients with
an abnormal pars plana, and microphthalmic patients, the
sclerotomies should be placed 3 mm posterior to the limbus
if the lens is absent or to be removed and 4 mm posterior
if the lens is to be retained (Fig. 10.1). Some surgeons use
3.5 mm for all cases as the authors do for 25-gauge cases.
The authors use 25-gauge, sutureless, transconjunctival sur-
gery for almost all cases; the following techniques apply only
to sutured 20-gauge surgery.
 CHAPTER 10
Fig ure 1 0 .1 ■ All sclerotom ies are m ad e with MVR b lad e, 3 m m p osterior to the lim b us if
aphakic or lens is to be rem oved and 4 m m p osterior to the lim bus if lens is to be retained.
The disposable microvitreoretinal (MVR) blade is ideal
for making 20-gauge sclerotomies. The blade is lancet shaped
to ensure that the incision is centered at the initial entry point
(Fig. 10.2). The blade is 1.4 mm in width, which makes a lin-
ear incision that becomes 0.89 mm (20 gauge) in diameter
when deformed to a circular shape by the instruments. The
nontapered, 20-gauge shank and 25-mm length correspond
to the dimensions of the other endosurgical instruments.
Stilettos, myringotomy blades, illuminated stab needles,
20-gauge needles, and transilluminators are unnecessary
with this method. Prefirming prior to use of the MVR blade
Fig ure 1 0 .2 ■ MVR blade has a lancet tip that m akes a sym -
m etrical incision around the initial entry p oint, 1.4 m m b lad e
width to m ake a 0.89-m m incision for 20-g aug e tools; shaft is
20 gauge to m ake a round opening from the linear incision.
■ 20-Gaug e Sutured Wound Construction and Closure 113
is unnecessary because the tip of a high-quality MVR blade is
typically as sharp as a 30-gauge needle and provides a correct
size incision if a subsequent fluid injection is required.
The sharpness of a new MVR blade usually prevents
the nonpigmented pars plana epithelium and choroid from
being pushed into the eye without penetration. The pos-
sibility of penetrating the sclera but not the choroid or pars
plana epithelium should be a reminder to use these blades
for one case only. To verify proper penetration, the blade
should be seen through the operating microscope. Under
no circumstances should the naked eye be used to check
the cannula nor should the 20-gauge cannula be used for
infusion without inspection (although this is commonly
done with 25-gauge infusion cannulas). If the eye is soft, a
20-gauge needle should then be placed through the initial
MVR incision and infusion fluid injected with the needle
tip visible. This step will decrease the chances of supra-
choroidal and subretinal infusion. If excessive bleeding
through the sclerotomy occurs, it can be treated by exter-
nal-internal bipolar diathermy using a scleral plug as the
internal contact or by placing one blade of the bipolar in
the sclerotomy and the other on the scleral surface adjacent
to the bleeder.
20-GAUGE INFUSION CANNULA
PLACEMENT
Unless there is severe fibrovascular proliferation or some
other obstruction discussed above at this region of the pars
plana, the infusion cannula is placed near the inferior border
114 SECTION II ■ Surgical Technology and Techniques
of the lateral rectus. In this position, it does not hit the nose,
lids, or speculum or interfere with the surgical instruments.
Because it decreases the incidence of choroidal infusion
and directs infusion fluid further away from the lens than
does the 2-mm cannula, the 4-mm cannula should be used
in all cases. Unless it is carelessly directed anteriorly, the
4-mm cannula will not damage the lens. A 6-mm cannula
can be used for thick choroid cases but can bump the lens.
The authors make the sclerotomy for the infusion cannula
before placing the cannula retention suture. By placing the inci-
sion before the suture, each edge of the wound can be grasped
with 0.12 forceps and stabilized during suture passage. The
tension applied to the sclera makes possible a deep suture bite
and wound eversion without opening the deepest part of the
incision can be used to judge scleral thickness. The suture bites
should be at least three-fourth scleral depth and 1 mm long so
that they will not fall off the tabs of the infusion cannula. They
should be widely spaced to ensure imbrication of the cannula
into the sclera, similar to a buckle. A braided, flexible suture
(6-0 silk) is easier to place over the cannula footplates than is a
springy nylon or polypropylene suture. Some surgeons use the
cannula retention suture for wound closure. The authors are
opposed to this method because it compromises imbrication of
the cannula into the sclera as well as tight wound closure. Vicryl
sutures fray when used for infusion cannula retention and cause
significant and long-lasting inflammation (Vicryl scleritis).
After insertion, the cannula should always be inspected
with the operating microscope before the infusion is turned
on. If the pupil is very small, the indirect ophthalmoscope and
Fig ure 1 0 .3 ■ If tissue is seen over infusion cannula, it should be incised with MVR blade. The
blade is inserted from the opposite side of the eye if the eye is aphakic or lens is to be rem oved
and from the sam e sid e if the lens is to be retained.
condensing lens can be used for visualization. To see the cannula,
it is usually necessary to push it gently toward the center of the
eye, being careful to avoid the lens, if it is a phakic eye. If the tip
appears shiny and clear of overlying tissue, then infusion may
be turned on. Observing the cannula with the unaided eye and
endoilluminator has insufficient magnification and can result in
inadvertent suprachoroidal or subretinal infusion.
If tissue is seen over the cannula, the corrective action
depends on the status of the lens (Fig. 10.3). If the eye is
aphakic or will be at the end of the operation, the MVR blade
is inserted from the opposite side of the eye to incise the tis-
sue over the cannula internally, so that it will retract down
over the entire interior portion of the cannula. This tech-
nique is also effective if a choroidal infusion is present from
previous problems or occurs during surgery from disloca-
tion of the cannula. If the eye is to remain phakic, the can-
nula should be removed and a 20-gauge needle inserted to
repressurize the globe and compress the choroid against the
sclera before reinserting the cannula.
If the pars plana is not visible preoperatively, the pars
plana infusion cannula cannot be turned on at this point.
Infusion should begin using an angulated 30-degree,
20-gauge blunt infusion cannula placed through the supra-
nasal sclerotomy (Fig. 10.4). If the lens is absent or is to
be removed, the cannula can be visualized by indenting the
cannula toward the pupil. When the anterior segment opac-
ity has been removed facilitating a view of the cannula, the
surgeon can then turn on the infusion cannula before pro-
ceeding with the rest of the vitrectomy.
 CHAPTER 10
Fig ure 1 0 .4 ■ If the infusion cannula cannot be seen with the operating m icroscope, infusion
is initiated with the blunt, 20-gauge, 30-degree angulated infusion cannula. Infusion is changed
to the sew-in cannula after the anterior opacities are rem oved and the cannula becom es visible.
Basics o f In fusio n Can n ula Use
The infusion cannula should be placed as soon as the sclera
is exposed. The infusion should remain until just prior to
conjunctival closure. Any infusion system should be visual-
ized before use. This first-in, last-out approach is essential to
prevent hypotony leading to intraocular or suprachoroidal
choroidal hemorrhage as well as miosis.
INSTRUMENT SCLEROTOMIES
After the infusion cannula is inspected, the infusion is
turned on with the infusion pressure at 45 mm Hg in adults
or 35 mm Hg in children or patients with extremely low per-
fusion (blood) pressure. Use of the Alcon vented gas forced
infusion (VGFI) system is better than placing the bottle a
certain distance above the eye because it gives a direct digital
readout of the infusion pressure. Starting the infusion before
the second and third incisions are made inflates the uvea
against the sclera, permitting the other sclerotomies to be
made through the choroid, and nonpigmented ciliary epi-
thelium because they are pushed against the sclera like an
inner tube in a tire. This approach also prevents the miosis
and bleeding that result from hypotony.
The second scleral incision, usually for the endoillumi-
nator, is made superonasally. It is best made near along an
imaginary line extending from the lowest part of the bridge
of the nose through the center of the pupil, 4 mm from the
■ 20-Gaug e Sutured Wound Construction and Closure 115
limbus in eyes to remain phakic and 3 mm for aphakic eyes
or 3.5 mm with 25-gauge systems. This incision is plugged
with the endoilluminator held backhand as the MVR blade is
removed to maintain intraocular pressure (IOP) rather than
the unnecessary step of placing a scleral plug.
The third scleral incision is primarily for the vitreous
cutter, fragmenter, scissors, forceps, endophotocoagulator,
etc. It is placed along an imaginary line extending from the
lowest part of the bridge of the lateral aspect of the supraor-
bital rim through the center of the pupil, the usual distance
from the limbus. Vitreoretinal instruments bump the brow
or nose and have difficulty reaching the anterior chamber if
entered superiorly or nasally. Using the active instruments in
the hand corresponding to the eye being operated on (right
eye—right hand; left eye—left hand) and having the instru-
ments enter the eye at a point corresponding to the lowest
point of the nose and temporal orbital rim allow better access
to all areas in the globe. Many surgeons limit their capabili-
ties by always using the active instruments in their dominant
hand. There are certain maneuvers that involve exchanging
the active instrument to the opposite hand for better access
to the pathology. This is frequently the case with endophoto-
coagulation to peripheral retinal breaks, scissors delamina-
tion, submacular surgery, and drainage of subretinal fluid
through peripheral breaks.
In all cases, the placement of sclerotomies should avoid
previous incisions, areas of known fibrovascular proliferation
on the pars plana, and dense vascularity. Most attempts to use
previous sclerotomies result in wound leaks, tearing of the
116 SECTION II ■ Surgical Technology and Techniques

Fig ure 1 0 .5 ■ Running shoelace 8-0 or 9-0 nylon sutures with ends cut flush with knot elim i-
nate blebs and scleritis associated with absorbable sutures and p roduce tight closure.

sclera, and excessive bleeding and may lead to fibrovascular
proliferation. This is not the case with 25-gauge surgery.
WOUND CLOSURE AND
POSTOPERATIVE MEDICATION
Wo un d Clo sure
Cryopexy, diathermy, or thermal cautery should not be
applied to pars plana incisions. These modalities damage
the sclera and predispose it to poor healing, poor closure,
inflammation, vascularization, and tissue ingrowth. Any form
of retinopexy is inappropriate at pars plana sites because the
scleral incisions are anterior to the retina. If there is exces-
sive bleeding, minimal bipolar diathermy can be applied, but
an instrument through the wound or closure by the suture
will stop most bleeding. Absorbable sutures are not used in
the sclera by the authors because their inelasticity facilitates
wound leakage during the operation, postoperatively, and
during reoperations. The author (S.C.) has observed many
filtering blebs from the use of absorbable sutures by other
surgeons. Wound leaks may cause hypotony and predispose
to tissue ingrowth. Vicryl scleritis causes considerable post-
operative inflammation for many weeks. The best closure is
achieved with running shoelace 8-0 to 10-0 monofilament
nylon sutures (Fig. 10.5). A running or X-type suture can
be placed more rapidly than interrupted sutures, equalizes
the tensions between loops, and has fewer ends to protrude
through the conjunctiva. The ends should be cut right on
the knot with a sharp blade under operating microscope
visualization by moving the suture toward the blade. The
ends can be buried in the wound to prevent conjunctival
erosion, but this makes a watertight wound more difficult
to attain.
CONJUNCTIVA AND TENON’S
CAPSULE
The conjunctiva and Tenon’s capsule are closed with inter-
rupted 6-0 plain gut, 8-0 Biosorb, or equivalent sutures
trimmed on the knot in a single layer. The Tenon’s capsule
should not be sutured to the muscle insertions because this
decreases the lid fissure, limits ocular motility postopera-
tively, and makes reoperation more difficult. Longer lasting
sutures are very irritating to the patient and are unneces-
sary. Conjunctival closure with bipolar diathermy (coapta-
tion) can cause conjunctival shrinkage and fistulas and is
less secure than suturing.
Reference
1. O’Malley C, Heintz RM. Vitrectomy via the pars plana, a
new instrument system. Trans Pac Coast Otoopthalmol Soc
1972;53:121.
 SEC TIO N III
In -Offi ce Pro ce d ure s

117
C H AP TER
11
LASER PHOTOCOAGULATION
PASCAL
The Pascal method of photocoagulation allows the delivery
of a predetermined pattern by scanning the placement of the
laser spots and controlling the emission of the laser light to
high degrees of accuracy and precision. This laser system
(PASCAL; OptiMedica Inc.) provides 532-nm light through
a multimode step index optical fiber to an integrated gal-
vanometer-based optical scanner housed within a slit lamp
biomicroscope. The exit surface of the fiber is telecentrically
imaged through the scanning system onto the retina, provid-
ing a variety of spot sizes with nominally top-hat intensity
profiles. At the aerial image plane of the slit lamp micro-
scope, the laser spots measure 60, 100, 200, and 500 µm in
diameter, all at the same numerical aperture. Different core
diameter fibers are used to produce the different spots. Pulse
durations from 10 to −1,000 ms are available. These optical
pulses have 10 µs rise and fall times and a temporal power
stability of greater than 90%. A touch-screen graphic user
interface is used to control treatment parameters, includ-
ing the spot size, laser power, pulse duration, and pattern
geometry. Once the treatment parameters are appropriately
selected, a foot pedal is used to activate the laser. The Pascal
Photocoagulator enables the physician to deliver multiple
laser lesions with a single footswitch depression by automat-
ing the emission of laser light with as much as 56 pulses
within half a second.
The PASCAL laser utilizes short-duration pulses
(20–30 ms) to limit thermal diffusion (1). Anterior thermal
diffusion can create damage to the nerve fiber layer. Poste-
rior thermal diffusion heats the choroid, causing pain and
even choroidal edema (effusion). Lateral thermal diffu-
sion is the most observable and clinically significant issue
because it results in “RPE creep,” enlargement of lesions
over time causing loss of central visual function. The PAS-
CAL laser utilizes a three-galvanometer system to rapidly
produce a precise pattern with programmable interlesion
spacing. Lighter intensity, smaller, more numerous spots
with precise spacing produce the best outcomes and can
be produced in a shorter time because of the scanning
system.
118
LASER PROPHYLAXIS FOR RETINAL
BREAKS
Though roughly 70% of the population develops a posterior
vitreous detachment (2), only about 4% of the population
has retinal breaks (3). Of these patients, only about 6 in
10,000 develop retinal detachments. Patients go on to
develop retinal detachment after retinopexy in about 2% of
cases, with the complication rate of retinopexy remaining
exceedingly low and difficult to measure (4). Determin-
ing the need for treatment is multifactorial and complex
at best.
The most conservative position is that only symp-
tomatic flap tears should be treated (5). However, many
large horseshoe tears, which all surgeons would agree
need treatment, are asymptomatic as many surgeons
have discovered in examinations carried out before laser-
assisted in situ keratomileusis (LASIK) and even routine
examinations (6).
Clinical characteristics in favor of treatment include
larger breaks, flap tears instead of round holes, breaks outside
lattice, superior location, and evidence of vitreous traction (7).
Larger breaks allow greater transhole flow, potentially exceed-
ing the capacity of the retinal pigment endothelium (RPE)
pump to stabilize a subclinical retinal detachment. Round
tears are thought to be less likely to have traction, but they are
often noted to have traction at the time of vitrectomy. Opercu-
lated breaks are least likely to have vitreous traction (8). Rolled
edges are said to be indicative of tangential traction, but the
internal limiting membrane (ILM) is elastic, and surgically
resected normal retina immediately rolls inward. Superior
breaks are probably slightly more significant because grav-
ity will decrease the likelihood of a stable subclinical retinal
detachment.
Pigmentation indicates chronicity, not adherence; there-
fore, pigmentation is only a relative contraindication (9).
Other factors favoring treatment include a history of reti-
nal detachment in the other eye, a family history of retinal
detachment, and physically active careers and/or sports (10).
Surgeons cannot predict which patients will be struck by an
air bag, experience a serious fall, or suffer another type of

trauma suggesting that trying to guess who needs laser by
evaluating lifestyle is problematic.
Socioeconomic situations suggesting that the patient is
less likely to return for follow-up should also be taken into
account. Most surgeons believe that anticipated LASIK, cata-
ract removal, or vitrectomy surgery is a reason to be aggres-
sive about treating asymptomatic retinal holes and similar
low-risk breaks. A high percentage of the population will
ultimately have cataract surgery, suggesting the need to treat
lower risk breaks in most patients (11).
COMPARISON WITH CRYOPEXY
Unlike laser, cryopexy disperses living RPE cells, possibly
increasing the risk of proliferative vitreoretinopathy (PVR)
and epimacular membranes. The PVR and epimacular mem-
branes that are said to be complications of treating retinal
breaks could actually be also directly related to the retinal
break causing the retinal glial cells and RPE cells to have
loss of contact inhibition (12). Cryopexy produces inflam-
mation and exudation but no immediate adherence, while
laser produces moderate, immediate adherence and no exu-
dation. In addition, cryopexy is more painful and produces
conjunctival damage as well (13).
TECHNIQUE
Spot spacing is a judgment issue in determining the optimal
number of rows. Wide spacing (underlap) raises the issue of
subretinal fluid leakage between the spots, while overlapping
results in areas of overtreatment. Most surgeons use approxi-
mately three rows of confluent spots. Many surgeons do not
treat round holes within lattice degeneration unless breaks
outside lattice are present that require laser. While the use of
discrete circular spots is the standard of care, movement of the
laser using a painting technique increases uniformity of the
thermal effect, although painting potentially produces more
pain due to heat diffusion.
Selecting optimal power is also a judgment issue.
Undertreatment may not produce enough pigmentation
to subsequently validate treatment adequacy, while the
so-called heavy treatment may produce excessive inflamma-
tion, possibly leading to PVR and epimacular membranes.
It is crucial to completely surround the retinal break. Many
patients receive insufficient treatment anterior to the break.
Three-mirror contact lenses (150 degrees field of view) are
the contact lenses most widely used to treat retinal breaks,
but various wide-angle (>130 degrees) lenses can be used for
all but the most peripheral breaks if care is taken to ensure
treatment anterior to the break. The Eisner (Crystal Lake,
Illinois) scleral depressor fits over the 3-mirror lens and
works well, although it is somewhat more uncomfortable
and generally underutilized. Ensuring maximal dilation of
the pupil is important to enable adequate treatment anterior
CHAPTER 11 ■ Laser Photocoag ulation 119
to the break. If posterior capsular opacification or cortical
cataract makes visualization anterior to the break impossible
even with scleral depression, the anterior treatment can be
continued to the ora at both ends of the break or laser retin-
opexy can be combined with cryopexy.
The presence of subretinal fluid around a tear (subclini-
cal retinal detachment) requires enclosure of the total area
of the subretinal fluid with at least three rows of laser. These
patients often require laser treatment from ora to ora for
complete enclosure, and this is best accomplished with laser
indirect ophthalmoscopy with sclera depression.
The PASCAL laser (OptiMedica, Santa Clara, California)
produces a precision pattern of shorter duration spots.
The pattern results in greater spacing uniformity. Using 20- to
30-ms–duration burns results in significantly less pain from
thermal diffusion to the choroid while not increasing treat-
ment time. The arc pattern works well for retinal breaks.
The laser indirect ophthalmoscope (LIO) is useful for
wheelchair patients and patients with spinal deformities,
such as severe osteoporosis or scoliosis. The LIO is also ideal
for treating through gas bubbles using trial-and-error head/
bubble positioning to optimize focus and access to the breaks.
The LIO is ideal for operating room use on children or for
treatment of the contralateral eye during retinal detachment
surgery often under general anesthesia. LIO treatment is eas-
ily utilized with scleral depression.
Retrobulbar blocks are seldom needed with laser retin-
opexy. There is an increased risk of globe penetration in
myopic patients with thin sclera, who, of course, are the very
patients with a higher incidence of retinal breaks.
If a patient has difficulty with pain during laser treat-
ment, the authors prefer subconjunctival or peribulbar
anesthesia with lidocaine rather than retrobulbar block.
This avoids the risks of retrobulbar injection and provides
adequate analgesia, although total anesthesia is not often
accomplished.
Patients can return to full activity 14 days after laser treat-
ment because tensile strength reaches the maximum level at
this time. Often, patients are counseled unnecessarily to avoid
work, housekeeping, and exercise for extended periods.
The longevity of the population, high expectations of
medical treatment, and the impossibility of trauma predic-
tion make treatment of most retinal tears advisable. The
advances in surgical equipment and techniques reduce
the likelihood of complications. The short recovery period
makes treatment acceptable to active and working patients.
References
1. Jain AM, Blumenkranz MS, Paulus Y, et al. Effect of pulse
duration on size and character of the lesion in retinal photoco-
agulation. Arch Ophthalmol 2008;126(1):78–85.
2. Snead MP, Snead DR, Mahmood AD, et al. Vitreous detachment
and the posterior hyaloid membrane: a clinicopathological
study. Eye (Lond) 1994;8(Pt 2):204–209.
3. Murakami K, Jalkh AE, Avila MP, et al. Vitreous floaters.
Ophthalmology 1983;90:1271–1276.
120 SECTION III ■ In-Office Procedures
4. Pollak A, Oliver M. Argon laser photocoagulation of
symptomatic flap tears and retinal breaks of fellow eyes. Br J
Ophthalmol 1981;65:469–472.
5. Kazahaya M. Prophylaxis of retinal detachment. Semin Oph-
thalmol 1995;10(1):79–86.
6. Wilkinson CP. Evidence-based analysis of prophylactic treat-
ment of asymptomatic retinal breaks and lattice degeneration.
Ophthalmology 2000;107(1):12–15.
7. Combs JL, Welch RB. Retinal breaks without detachment: nat-
ural history, management, and long-term follow-up. Trans Am
Ophthalmol Soc 1982;80:64–97.
8. Byer NE. What happens to untreated aymptomatic retinal
breaks, and are they affected by posterior vitreous detach-
ments? Ophthalmology 1998;105(6):1045–1049.
9. Morse PH, Eagle RC Jr. Pigmentation and retinal breaks. Am J
Ophthalmol 1975;79(2):190–193.
10. Davis MD. Natural history of retinal breaks without detach-
ment. Arch Ophthalmol 1974;92:183–194.
11. Robertson DM, Buettner H. Pigmented preretinal membranes.
Am J Ophthalmol 1977;83:824–829.
12. Glaser BM, Vidaurri-Leal J, Michels RG, et al. Cryotherapy
during surgery for giant retinal tears and intravitreal disper-
sion of viable retinal pigment epithelial cells. Ophthalmology
1993;100(4):466–470.
13. Kramer SG, Benson WE. Prophylactic therapy of retinal breaks.
Surv Ophthalmol 1977;22(1):41–47.
C H AP TER
12
PNEUMATIC RETINOPEXY
AND OFFICE FLUID-GAS
EXCHANGE
Air was used by Ohm and Rosengren over a
half century ago for the repair of retinal detach-
ments. Dominguez (1), and subsequently Hilton
and Grizzard (2), reported the use of in-office
injection of expanding gas for the repair of primary, rheg-
matogenous retinal detachment. Initially, cryoretinopexy was
applied before gas injection, while others used laser retin-
opexy after reattachment, analogous to the postreattachment
retinopexy concept developed by the principal author for
vitreoretinal surgery (3–9).
Retinal reattachment surgery outcome data are very dif-
ficult to interpret because of the large number of variables, vast
array of surgical options, and combinations of techniques, as
well as the relatively low number of cases per surgeon. Some
surgeons state that pneumatic retinopexy causes proliferative
vitreoretinopathy (PVR), and in spite of this, they inject gas after
scleral buckling. While it is clear that pneumatic retinopexy
should not be used for cases with PVR or vitreous traction,
it is not known with statistical accuracy what the indications
should be. It is clear that single, superior retinal breaks are the
ideal cases for pneumatic retinopexy, but these cases can also
be repaired with very high success rates by scleral buckling or
vitrectomy, gas, and laser. Pneumatic retinopexy costs less than
scleral buckling or vitrectomy-based repair. Some advocate
the use of 360-degree laser retinopexy in these cases, which
raises the issue of potential stimulation of PVR (10). The prin-
cipal author has utilized pneumatic retinopexy in a wide vari-
ety of cases since 1988 but uses the technique less frequently
in recent years because of failures due to progressive contrac-
tion of the vitreous, new retinal breaks, unpredictability, and
PVR. Patients with medical problems combined with simple,
superior retinal detachments are the best candidates for pneu-
matic retinopexy. The advent of transconjunctival, sutureless,
25-gauge vitrectomy allows many of the advantages of pneu-
matic retinopexy to be retained but in the context of the better
outcomes afforded by vitrectomy, gas, and laser.
The authors’ current approach to pneumatic retinopexy
involves avoiding cryotherapy unless an individual small tear
is seen that can be treated with a single cryo spot. Avoidance
of cryotherapy decreases the stimulus for PVR. If a larger tear
is seen and pneumatic retinopexy is performed, injection of
C3F8 gas without cryotherapy is the authors’ preferred course
of action. If reattachment is successful, secondary laser retin-
opexy to the retinal tear can be performed on the attached
retina. If the gas injection is unsuccessful at retinal reattach-
ment, the authors proceed to final surgical repair with vitrec-
tomy, and the added insult of cryo is avoided.
SURGICAL SEQUENCE
The authors use topical viscous lidocaine anesthesia applied
first as a drop and then with a sterile applicator at the intended
pars plana injection site. Retrobulbar or peribulbar anesthesia is
used in most cases, followed by cryopexy to the retinal breaks.
The patient should be prepped with Betadine 5% after
anesthesia is achieved. A sterile speculum is required to pre-
vent the needle from contacting the nonsterile lid margins
and lashes. The surgeon should use sterile gloves, sterile
technique, and a filter when drawing up the gas.
The authors routinely use C3F8 gas rather than SF6
because it expands three to four times (its original size)
compared to two times for SF6 (11–18). Greater expansion
means that less gas can be injected and therefore less para-
centesis will be required.
Injection of gas is best performed with the patient lying
on his or her side (Fig. 12.1), not supine, seated, leaning over
the examining table, or at the slit lamp. Multiple bubbles (fish
eggs) are completely preventable if the injection is performed
at the highest point of the eye, with the needle advanced just
past the pars plana epithelium. By using this method, all
gas is injected into the original bubble, preventing multiple
121
122 SECTION III ■ In-Office Procedures
Fig ure 1 2 .1 ■ Gas injection is perform ed at the highest point
of the eye with the patient lying on his or her side.
bubbles (Fig. 12.2). A 32-gauge needle should be used to
reduce leakage of gas after the injection. The smaller gauge of
the needle favors a slower rate of injection due to increased
resistance. Slower gas injection creates less fish egg bubbles.
A tuberculin or 3-mL syringe is used for the gas injection.
Repetitive in and out movement of the needle into the eye
is avoided if the patient is on his or her side, decreasing the
risk of endophthalmitis. By placing the needle only 3 mm
into the eye, inadvertent contact with the lens and retina is
reduced. It is difficult for a patient on his or her side to raise
the head and bring the needle into contact with the retina.
Approximately 0.3 to 0.6 mL of gas is injected, although the
most appropriate volume is controversial. It is the authors’
opinion that higher volumes give higher success rates but
demand paracentesis. The authors use paracentesis in all
cases. Immediately after the injection, the reclining patient
chair is tilted up to the seated position at the slit lamp with
the speculum remaining in place. Sitting the patient up has
the added advantage of rolling the bubble away from the
injection site, eliminating the need for an applicator stick to
prevent leakage. A 30-gauge or 27-gauge needle is advanced
parallel and adjacent to the limbus through the inferior clear
cornea into the anterior chamber. Oblique entry ensures a
self-sealing wound. Needle placement over the iris instead
of the pupil reduces the risk of lens damage. The patient
usually reports no light perception immediately after the
injection, but light perception returns as soon as the para-
centesis reduces the intraocular pressure.
If the breaks could not be treated before gas injection
because of a bullous detachment, retinopexy can be per-
formed 1 or 2 days after the gas injection when the subretinal
fluid has pumped out. The laser indirect ophthalmoscope
is preferred to the 3-mirror contact for postreattachment
retinopexy (Fig. 12.3). If there are concerns about the abil-
ity to visualize the breaks after reattachment, preattachment
retinopexy with transscleral cryopexy can be utilized.
IN-OFFICE FLUID-GAS EXCHANGE
Postvitrectomy retinal detachment without traction or PVR
or hemorrhage patients can be managed by in-office fluid-gas
exchange using isoexpansive concentrations of C3F8 (16%)
or SF6 (25%) (Fig. 12.4). The injection needle is positioned
Fig u re 1 2 . 2 ■ In sertin g t h e
30-g auge needle just b arely into
the eye ensures having a sing le
bubble.
 CHAPTER 12 ■ Pneum atic Retinopexy and Office Fluid -Gas Exchange 123

Fig u re 1 2 .3 ■ Retin o p exy is

perform ed with transconjunctival
cryop exy b efore gas injection or
using the laser in direct op hthal-
m oscope as shown.

3 to 4 mm posterior to the limbus at the highest part of the

eye as in pneumatic retinopexy. The needle is advanced only
3 mm into the eye to avoid multiple bubbles. This 32-gauge
needle is connected via a short piece of tubing to a 30- to
60-mL air- or gas-filled syringe operated by the assistant.
A second 25- to 30-gauge needle, depending on the viscosity
of the intraocular fluid, is positioned 4 mm posterior to the
limbus, at the lowest portion of the eye. This needle is con-
nected to a tuberculin syringe, with the plunger removed,
to act as a handle. The open end of this needle is positioned
over a waste can to collect the fluid. The exchange is contin-
ued until gas appears in the egress needle. The egress needle
is withdrawn and the intraocular pressure adjusted using the
gas syringe and tactile assessment of the IOP with a sterile
applicator stick.
Phakic patients postvitrectomy can develop posterior
subcapsular cataracts after fluid-gas exchange if strict face-
down positioning is not followed. This can make second-
ary laser retinopexy very difficult. Cryotherapy to the retinal
tear prior to the fluid-gas exchange should be considered in
phakic postvitrectomy patients.

References
1. Dominguez A. Cirugia precoz y ambulatoria del desprendimento
de retina. Arch Soc Esp Oftamol 1985;48:47–54.
2. Hilton GF, Grizzard WS. Pneumatic retinopexy: a two step out-
Fig u re 1 2 .4 ■ In -office two -n eed le (30-g aug e in jectio n , patient operation without conjuctival incision. Ophthalmology
25-g aug e eg ress) fluid -g as exchang e is p erform ed in p atients 1986;93:626.
wh o d evelop retinal d etachm en t after vitrectom y. Th is tech - 3. McDonald HR, Abrams GW, Irvine AR, et al. Management of
niq ue ensures a com plete fill with an isoexp ansive concentration subretinal gas following attempted pneumatic retinal reattach-
of long-acting gas. ment. Ophthalmology 1987;94:319 –326.
124 SECTION III ■ In-Office Procedures
4. Hilton GF, Kelly NE, Salzano TC, et al. Pneumatic retinopexy.
A Collaborative report of the first 100 cases. Ophthalmology
1987;94:307–314.
5. Roy FH. Master techniques in ophthalmic surgery. Williams and
Wilkins, 1995: 1118–1119.
6. Wilkinson CP, Rice TA. Michels retinal detachment. Mosby,
1997:596–612.
7. Brinton DA, Hilton GF. Pneumatic retinopexy. Ophthalmol Clin
North Am 1994;7:1.
8. Vygantas CM, Peyman GA, Daily MJ, Ericson ES. Octafluo-
rocyclobutane and other gases for vitreous replacement. Arch
Ophthalmol 1973;90:235.
9. Norton EWD. Intraocular gas in the management of selected retinal
detachments. Trans Acad Ophthalmol Otolaryngol 1973;77:OP-85.
10. Hilton GF, Tornambe PE. Pneumatic retinopexy: an analysis
of intraoperative and postoperative complications. Retina
1991;11:285—294.
11. Lincoff H, Mardirossian J, Lincoff A, et al. Intravitreal longevity of
three perfluorocarbon gases. Arch Ophthalmol 1980;98:1610.
12. Constable IJ, Swann DA. Vitreous substitution with gases. Arch
Ophthalmol 1975;93:416.
13. DeJuan E Jr, McCuen B, Tiedeman J. Intraocular tamponade
and surface tension. Surv Ophthalmol 1985;30:47.
14. Stinson TW III, Donion JV Jr. Interaction of intraocular air and
SF6 with nitrous oxide: a computer simulation. Anesthesiology
1982;56:385.
15. Lincoff H, Coleman J, Kreissig I, et al. The perfluorocarbon
gases in the treatment of retinal detachment. Ophthalmology
1983;90:546.
16. Fineberg E, Machemer R, Sullivan P. SF6 for retinal detach-
ment surgery. A preliminary report. Mod Probl Ophthalmol
1974;12:173.
17. Lincoff HA, Maisel JM, Licoff A. Intravitreal disappear-
ance rates of four perfluorocarbon gases. Arch Ophthalmol
1984;102:928.
18. Lowe MA, McDonald HR, Campo RV, et al. Pneumatic retin-
opexy. Surgical results. Arch Ophthalmol 1988;106:1672.
 SEC TIO N IV
Dise ase -Sp e cifi c Man ag e m e n t

125
C H AP TER
13
PROPHYLAXIS OF RETINAL BREAKS
AND SCLERAL BUCKLING
PREVENTION OF RETINAL
DETACHMENT
It is probable that the per capita incidence of retinal detach-
ment has decreased over the past three decades. The replace-
ment of intracapsular cataract surgery, first by extracapsular
surgery and then by phacoemulsification with endocapsular
intraocular lenses, has certainly contributed to the decline.
The widespread use of indirect ophthalmoscopy and retin-
opexy for prophylaxis of retinal breaks has probably had a
major impact as well. It is also likely that the increased use
of protective eyewear has made a contribution.
IS EVIDENCE-BASED TREATMENT OF
RETINAL DETACHMENT POSSIBLE?
The increasing number of retinal specialists per capita coupled
with fewer detachments per patient-year, variable pathology,
and significantly more treatment options has virtually guaran-
teed that the treatment of retinal detachment will never move
to an evidence-based paradigm because of the statistical com-
plexity. Therapeutic options include scleral buckling, vitrec-
tomy, and pneumatic retinopexy. Gas choices include air, C3F8,
and SF6. Buckle options include sponges versus “hard” silicone,
drainage versus nondrainage, encircling versus segmental,
radial versus circumferential, etc. Retinopexy choices include
cryotherapy versus laser (transscleral or laser indirect ophthal-
moscope) versus diathermy. Many of these therapies are used
in combination, making the analysis even more complex.
PROPHYLACTIC RETINOPEXY
Laser is preferred to cryotherapy for prophylactic retinopexy
because there is less pain and potentially less proliferative
vitreoretinopathy (PVR). Low to moderate intensity con-
fluent lesions with fewer rows are preferred by the authors
over the very common method of using many rows of heavy,
126
spaced-out lesions. Many authors recommend treatment
only if the retinal breaks are symptomatic (1–10). Relying
on the patient can be effective in a population of highly edu-
cated people but is less reliable in patients having less edu-
cation and more socioeconomic problems. Large breaks are
typically more significant than small breaks. Retinal detach-
ment in the fellow eye or another location in the same eye
or a family history of detachment is a relative indication for
treatment. Patients for whom cataract surgery is planned,
athletes, or certain careers with risk of high G-forces or non-
availability of care may indicate the need for prophylactic
treatment as well. Retinal breaks outside lattice are more sig-
nificant than breaks inside lattice. Superior breaks are prob-
ably more significant than inferior breaks. Pigment around
breaks does not indicate adherence to the retinal pigment
endothelium but does indicate chronicity.
Laser can be used to “wall-off,” better termed “laser
delimiting,” a retinal detachment which is small enough that a
triple row of laser will not significantly impair the visual field.
On occasion, laser-delimited detachments will spontaneously
reattach. Laser confinement of relatively small retinal detach-
ments has remarkably good long-term results and is almost
certainly underutilized because of habit and economics.
SCLERAL BUCKLING
Case Se le ct io n
Given the success of modern microincisional vitrectomy for
repair of simple and complex retinal detachments, the ques-
tion of when to recommend either procedure is still heavily
debated (11). Even in this era of evidence-based medicine, it
is exceedingly difficult to create a clinical trial that accounts
for the myriad of variables present in surgical techniques
by individual surgeons. The authors do not perform sclera
buckling in combination with vitrectomy, since there is no
additive benefit in their view of combining both procedures,
although the risks and complications of both techniques are
nonoverlapping and therefore additive.
 CHAPTER 13
The authors currently only recommend sclera buckling
in young phakic patients with simple retinal detachments,
anterior tears, and no PVR. The presence of PVR requires
vitrectomy and obviates sclera buckling.
Sp o n g e s Ve rsus Hard Silico n e
The smooth surface and relative incompressibility of “hard”
silicone make it superior to sponges for scleral buckling in
the author’s opinion. Higher extrusion and infection rates
make sponges less desirable (12). Sponges create a higher
buckle immediately under the sutures and a lower buckle
between the sutures. This potentially creates a higher inci-
dence of radial folds. The area of lesser buckling effect
between sutures corresponds with regions of bulging under
the conjunctiva, which creates a dellen-like effect leading to
exposure of the buckle (Fig. 13.1).
Im p lan t s Ve rsus Exp lan t s
While scleral dissection is an acceptable method of scleral
buckling, it is rarely performed currently because of its
inflexible, time-consuming nature. The original justifica-
tion for scleral dissection was the ability to avoid scleral
damage from full-thickness diathermy. Although burying
the buckle under a flp reduces extrusion, it increases intru-
sion, operating time, inflexibility, and the risk of scleral
perforation.
Fig ure 1 3 .1 ■ Silicone sponges bulg e between m attress sutures at the end of buckle, which
causes conjunctival erosion (dellen form ation), which lead s to exposure of buckle and extrusion.
In addition, b uckle is flexible, which results in a lower buckle between m attress sutures, leading
to radial folds and unsupported retinal breaks.
■ Prophylaxis of Retinal Breaks and Scleral Buckling 127
Bucklin g wit h Vit re o us Surg e ry
Scleral buckling is used with many variations for retinal
reattachment (13,14). A rapid, simplified form of scleral
buckling is preferred by the authors for all scleral buckling
(Fig. 13.2). Scleral buckling is significantly overutilized in
conjunction with vitreous surgery. Scleral buckling causes
a very significant incidence of strabismus, ptosis, pain, con-
junctival damage, and refractive error. Scleral buckling is not
indicated for giant break surgery, PVR, or routine rhegmatog-
enous retinal detachments. The elimination of vitreous trac-
tion coupled with complete intraoperative reattachment by
vitrectomy has eliminated, in the authors’ opinion, the need
for combined buckle-vitrectomy.
En circlin g Ban d s
Prophylactic scleral buckling with an encircling band can
then be thought of as making a new ora serrata to treat
peripheral vitreoretinal traction preceding retinal detach-
ment. Encircling bands for prophylactic buckling were used
frequently in the early days of vitrectomy (15). This approach
is not used today because of better cutters, techniques, flu-
idics, and dissection methods. Local anesthesia, outpatient
surgery, cost containment as well as avoidable complications
such as pain, strabismus, refractive error, conjunctival dam-
age, and ptosis all contribute to the virtual elimination of
prophylactic buckling.
128 SECTION IV ■ Disease-Sp ecific Managem ent

Fig u re 1 3 . 2 ■ A sim p lified

approach saves time, reduces com-
plexity, and reduces complications.

Circum fe re n t ial Exp lan t s The explant is trimmed from a larger piece of silicone
for a custom fit if a standard width implant will not work.
Because of their narrow configuration, bands alone are usu-
In every case, the explant width is made so that the outer
ally not utilized to treat specific retinal breaks. If a circum-
surface of the explant conforms to the contour of the globe
ferential explant can cover the posterior extent of a retinal
after tying up the sutures. Off-shelf explant with sutures
break, it is utilized in preference to a radial explant. Circum-
placed 1 to 3 mm wider than the explant will not provide
ferential explants require less exacting localization, do not
the correct effect. If a large ratio of implant width to suture
distort the macula, and cover a broader extent of vitreous
spread is utilized, the explant will be only partially imbri-
base pathology. Posterior breaks are managed with vitrec-
cated into the sclera. This external protuberance wears
tomy techniques. The principal author has not used radial
a hole in the conjunctiva by creating a dellen-like effect,
buckles or sponges for over 25 years in the buckling alone
is similar to that observed with sponges, and provides a
or vitrectomy setting.
less substantial buckling effect. A slight chamfering of the
Monofilament (5-0) nylon sutures are utilized with a
cut edges of the silicone seems to provide a lesser chance
single circumferential posterior scleral bite. In contrast to
of scleral or conjunctival erosion. Two or three mattress
radial suture bites, the circumferential bite can be quite long
sutures per quadrant provide the most consistent buck-
without reducing the posterior extent of the buckle. The sin-
ling effect and reduce the chance of extrusion or sutures
gle circumferential posterior scleral bite reduces by one half
pulling through the sclera. Explants without grooves are
the chances of perforating the retina, as compared to paired
preferred because they do not fold along the groove and
bites. This posterior bite is always placed 3 mm and prefer-
bands are almost never combined with buckles using the
ably 5 mm posterior to the most posterior aspect of the most
authors’ technique.
posterior break. A too-anterior positioning of the buckle
The circumferential buckle method described above
causes many reoperations after scleral buckling procedures.
is used by the authors for all scleral buckling. All detached
All anterior scleral bites are placed circumferentially in
quadrants are buckled, direct needle drainage of SRF is used
the scleral condensation, conforming to the rectus muscle
exclusively, and bands, sponges, and radial buckles are never
insertions. This provides an area of thicker sclera for greater
used. An attempt is made to avoid the superior rectus as
permanence. This muscle ring also conforms to the ora ser-
mentioned earlier and only muscles absolutely necessary to
rata; therefore, a scleral suture bite placed here in a circum-
exposure are engaged with traction sutures. Minimal retin-
ferential orientation cannot perforate the retina (Fig. 13.3).
opexy is used. Transscleral diode laser retinopexy is alter-
Extending all circumferential buckles to the ora serrata pre-
native to cryopexy; however, it cannot be used in highly
vents the anterior leakage of subretinal fluid (SRF) associated
pigmented patients and it is more difficult to judge treat-
with narrow bands or buckles placed more posteriorly.
ment intensity.
 CHAPTER 13 ■ Prophylaxis of Retinal Breaks and Scleral Buckling 129

Fig ure 1 3 .3 ■ Anterior b ites of m attress suture should b e m ad e in thicker sclera associated
with the rectus m uscle insertions. The thicker sclera red uces p ostop erative p ull-through, and the
location corresponds to the ora so that p erforation will not dam age the retina. Placing all sutures
this anterior prevents SRF from leaking anteriorly.

Fig ure 1 3 .4 ■ Scleral buckling with encircling tire is used for PVR. A band is not necessary, and
the ends of the buckle are joined by a butt joint. All im brication is accom plished by using two to
three m attress sutures per quadrant.
130 SECTION IV ■ Disease-Sp ecific Managem ent
Scle ral Bucklin g wit h a Bro ad
En circlin g Buckle
The authors no longer utilize scleral buckling for PVR. For
those surgeons who believe in using a buckle, they recom-
mend a 360-degree moderate height, broad buckle with the
suturing techniques described for circumferential buckles.
Encircling sponges can result in anterior or posterior leak-
age. As PVR is a diffuse process, no attempt is made to cus-
tomize the buckle contour; instead, a uniformly high, broad
encircling buckle is utilized (Fig. 13.4).
The posterior circumferential bites are made as posteri-
orly as possible without compression of vortex veins, and the
circumferential anterior bite is placed in the scleral conden-
sation at the muscle ring. The usual suture spread is 10 to 12
mm, with the usual tire being 6 to 9 mm in width. Imbrica-
tion will then cause the buckle to be flush with the surface
of the globe. The ends are joined with two interrupted 5-0
nylon sutures, with long bites through the implant mate-
rial, and the knot is then buried. No band, encircling suture,
or overlapping is utilized so as to provide a consistently
smooth internal buckle contour. The direct needle drainage
technique is used for the intraoperative elimination of SRF
before the encircling tire is anchored in place.
Scle ral De fe ct s
Thin sclera can be covered in most cases with a circumferen-
tial scleral buckling or encircling tire previously described.
Preserved sclera and fascia lata add a degree of complexity,
infection risk, and risk of dehiscence and are rarely utilized.
Full-thickness defects, unless extremely large, can usually be
repaired by suturing good tissue edges together rather than by
oversewing. This approach creates a scleral shortening effect
that may be beneficial with PVR or wound-related cellular
proliferation. A small leaking area can be handled with vari-
ous tissue adhesives. Sponges appear to be more erosive to
the sclera than hard silicone and are therefore not utilized.
References
1. Foos RY. Posterior vitreous detachment. Trans Am Acad
Ophthalmol Otolaryngol 1972;76:480.
2. Linder B. Acute posterior vitreous detachment and its retinal
complications. Acta Ophthalmol (supp)(KBH) 1966;87:1.
3. Foos RY. Anatomic and pathologic aspects of the vitreous body.
Trans Am Acad Ophthalmol Otolaryngol 1973;77:171.
4. Foos RY. Tears of the peripheral retina: pathogenesis, incidence,
and classification in autopsy eyes. Mod Probl Ophthalmol
1975;15:68–81.
5. Byer NE. Clinical study of retinal breaks. Trans Am Acad
Ophthalmol Otolaryngol 1967;71:461–473.
6. Byer NE. The natural history of asymptomatic retinal breaks.
Ophthalmology 1982;89:1033–1039.
7. Byer NE. Long term natural history of lattice degeneration of
the retina. Ophthalmology 1989;96:1369–1401.
8. Byer NE. Cystic retinal tufts and their relationship to retinal
detachment. Arch Ophthalmol 1981;99:1788–1790.
9. Byer NE. Natural history of posterior vitreous detachment with
early management as the premier line of defense against retinal
detachment. Ophthalmology 1994;101:1503–1513.
10. McHugh DA, Schwartz S, Dowler JG, et al. Diode laser contact
transscleral retinal photocoagulation: a clinical study. Br J
Ophthalmology 1995;79:1083–1087.
11. Schwartz SG, Flynn HW. Primary retinal detachment: scleral
buckle or pars plana vitrectomy? Curr Opin Ophthalmol 2006
Jun;17(3):245–250.
12. Russo CE, Ruiz RS. Silicone sponge rejection; early and late
complications in retinal detachment surgery. Arch Ophthalmol
1971;85:647.
13. Hahn YS, Lincoff A, Lincoff H, Kreissig I. Infection
after sponge implantation for scleral buckling. Am J
Ophthalmol1979;87:180.
14. Devenyi RG, de Carvalho Nakamura H. Combined scleral
buckle and pare plana vitrectomy as a primary procedure for
pseudophakic retinal detachments. Ophthalmic Surg Lasers
1999;30(8):615–618.
15. Desai UR, Strassman IB. Combined pars plana vitrectomy
and scleral buckling for pseudophakic and aphakic retinal
detachments in which a break is not seen preoperatively.
Ophthalmic Surg Lasers 1997;28(9):718–722.
C H AP TER
14
VITRECTOMY FOR RETINAL
DETACHMENT

visualization is essential for peripheral vitreous removal and

INDICATIONS
endolaser and frequently obviates the need for sclera depres-
sion vitrectomy. Contact-type wide-angle visualization (Volk or
The use of vitrectomy for retinal detachment repair is
AVI) provides a 10-degree wider field of view than noncontact
increasing (1) because of a number of factors: more surgeons
systems and eliminates all corneal asphericity due to cataract
are trained in vitrectomy, better cutters and fluidics, wide-
surgery, Limbal Relaxing Incision (LRI), Radial Keratotomy
angle visualization, better methods for drainage of subretinal
(RK), laser-assisted in situ keratomileusis, Penetrating Kerato-
fluid (SRF) and retinopexy are available, and fellows are less
plasty (PK), trauma, keratoconus, pterygium, and pterygium
likely to be taught scleral buckling. The decision between
removal. In addition, contact-based wide-angle visualization
vitrectomy, scleral buckling, and pneumatic retinopexy is a
markedly reduces the need for ocular rotation compared to
very complex issue with few, if any, scientific answers. In gen-
noncontact visualization (BIOM, EIBOS), thereby reducing
eral, vitrectomy, gas, and laser should be used when vitreous
tool flex with 25- and 27-gauge pars plana vitrectomy.
opacities or tractions make it unlikely that scleral buckling
All traction should be removed from each retinal break.
alone will succeed. Relative advantages of scleral buckling
It is crucial to remove traction on the anterior margin as
include less cataract progression and 80% to 90% success
well as from the more apparent flap. Amputation of the flap
rates in selected detachments. Advantages of vitrectomy
ensures removal of this component of the traction.
include no refractive change; no damage to the extraocular
If the breaks or tears can be easily seen after fluid-air
muscles or levator, therefore no strabismus or ptosis; and
exchange, they can be used for internal drainage of SRF.
much less postoperative pain and conjunctival damage.
If access to the breaks is difficult, a posterior drainage reti-
Many surgeons, including the authors, use vitrectomy, gas,
notomy or perfluorocarbon liquids are required. Drainage
and laser in most pseudophakic detachments (2). Many sur-
retinotomy can be initiated by using a disposable bipolar
geons are inclined to use vitrectomy, gas, and laser for supe-
endoilluminator or 25-gauge bipolar diathermy to coagu-
rior detachments and buckling for inferior detachments.
late, weaken, and mark the site (Fig. 14.1). Outside the
Evidence of traction on breaks, especially large horseshoe
arcades, at the most posterior extent of the SRF is often a
tears, suggests the need for vitrectomy as does the pres-
good location; nasal is better than temporal and superior is
ence of avulsed vessels. Vitreous blood or significant opaci-
better than inferior. Single-cut mode with the vitreous cutter
ties require vitrectomy to visualize the retinal breaks. The
will produce an ideal small, round retinotomy (Fig. 14.2).
authors rarely use scleral buckling and never combine vit-
The soft-tip cannula is preferred for internal drainage of SRF
rectomy with scleral buckling. Their current indications for
(Fig. 14.3). Internal drainage of SRF should precede fluid-air
sclera buckling are young phakic patients with clear lenses,
exchange, especially with the use of preexisting peripheral
retinal detachments with anterior tears, and the absence of
retinal breaks for drainage; this sequence prevents posterior
proliferative vitreoretinopathy.
shift of the SRF. If intraocular lens (IOL) fogging occurs dur-
ing fluid-air exchange, the air should be removed and per-
fluorooctane (PFO) used to displace the SRF anteriorly out
SURGICAL SEQUENCE through the retinal breaks (Fig. 14.4). Endolaser retinopexy is
then performed under PFO followed by PFO-gas exchange or
It is essential to remove as much peripheral vitreous as possi- medium-term PFO surface tension management. Fogging can
ble without damaging the lens or creating new retinal breaks. be anticipated if a YAG capsulotomy has been performed and
Scleral depression is very useful for this purpose. Wide-angle the anterior vitreous cortex has been disrupted or removed.
131
132 SECTION IV ■ Disease-Sp ecific Managem ent

Fig ure 1 4 .1 ■ Bip olar d iatherm y is used to m ark the site p rior to m aking d rainag e retino-
tom y to facilitate finding the site after retina reattached by internal d rainage of SRF and fluid-air
exchange.

Fig ure 1 4 .2 ■ Sing le-cut m od e on a 20-g aug e or 25-g aug e cutter results in a sm all, round
drainage retinotom y.
 CHAPTER 14 ■ Vitrectom y for Retinal Detachm ent 133

Fig ure 1 4 .3 ■ The soft-tip extrusion cannula can be used for drainage of SRF.

Fig ure 1 4 .4 ■ Liquid perfluorocarbon such as Alcon PFO displaces SRF anteriorly out through
retinal breaks into the anterior vitreous cavity. PFO should be used if fluid-air exchange results in
IOL fogg ing .
134 SECTION IV ■ Disease-Sp ecific Managem ent
Complete removal of SRF can be accomplished by
sequential drainage. After the initial drainage of SRF, there
may be a thin layer of diffuse SRF that may be imperceptible.
This SRF often recollects underneath the drainage retinotomy
after approximately 30 seconds. At this time, redrainage can
remove most of the remaining SRF (Fig. 14.5). This tech-
nique is best repeated until no SRF recollects. Laser treatment
Fig ure 1 4 .5 ■ Redrainage facili-
tates com plete rem oval of SRF.
around the retinal tear or retinotomy can then be performed
safely with sharp demarcation of laser uptake (Fig. 14.6).
The authors use 25-gauge, sutureless, transconjunctival
technique for all vitrectomy for retinal detachment. Advan-
tages of 25-gauge technique over 20-gauge technique include
(a) port-based flow limiting that produces greater fluidic
stability and therefore less retinal motion, (b) less pain, (c) less
Fig ure 1 4 .6 ■ Complete removal
of SRF im p roves the up take of
endolaser retinopexy around drain-
age retinotom y.

inflammation, and (d) less conjunctival damage (important for
patients with glaucoma filtering procedures and surface disor-
ders). A 25-gauge, soft-tip cannula is used for simultaneous
drainage of SRF and fluid-air exchange unless PFO is used.
The 25-gauge, soft-tip cannula is excellent for air-gas exchange
or PFO-gas exchange after endolaser to all retinal breaks.
Although many surgeons perform 360-degree laser in vit-
rectomy for retinal detachment and even macular hole cases,
it is not known whether this reduces postoperative retinal
detachment. Some surgeons use laser indirect ophthalmos-
copy (LIO) instead of endolaser. The LIO approach could the-
oretically increase nuclear sclerosis as well as iris and corneal
damage from laser energy absorption. It is possible that light
scatter from the cornea and lens could cause macular damage
as well. The only advantage to the LIO approach is less chance
of bumping the lens during peripheral endolaser treatment.
Lens bump can be eliminated by scleral depression and flex-
ible or articulated laser probes as well as by using the laser
probe in both the left and the right hand for better access.
It has not been shown that scleral buckling is needed
when using vitrectomy for retinal detachment repair, and the
authors believe that it is widely overutilized. The authors
do not use scleral buckling in vitrectomy cases in order to
reduce induced refractive error, strabismus, ptosis, pain,
conjunctival swelling, and conjunctival damage.
SURFACE TENSION MANAGEMENT
All rhegmatogenous retinal detachments and traction
retinal detachments with one or more retinal breaks require
gas, medium-term PFO, or silicone oil surface tension
Fig ure 1 4 .7 ■ Medium-term PFO
requires com pulsive, 360-degree
rem oval of vitreous traction using
wide-angle visualization.
CHAPTER 14 ■ Vitrectom y for Retinal Detachm ent 135
management. Removal of the hyaluronan gel dramatically
reduces the viscosity as aqueous humor fills the former vit-
reous cavity. This reduces the pressure gradient the retinal
pigment epithelium pump can achieve across the retina. Sub-
clinical retinal detachments rarely remain stable after vitrec-
tomy; even very small holes usually result in a total retinal
detachment in hours to days. Laser and cryo retinopexy sites
reach maximum tensile strength at 10 to 14 days; SF6 bubbles
often last 7 to 14 days, which means a full fill or isoexpansive
SF6 is required to avoid the use of the longer duration C3F8.
MEDIUM-TERM PERFLUORON
The authors have been using medium-term PFO for infe-
rior retinal detachments and giant breaks since 2002 in a
prospective, consecutive series. The advantages are as fol-
lows: (a) the patients can be seated, reclining, or supine,
not face down; (b) compared to scleral buckles, there is no
induced refractive error, no strabismus, no ptosis, no pain,
and no inflammation; and (c) the patients can fly on an air-
plane. Disadvantages are as follows: (a) two procedures are
required, the second one to remove the PFO (this is often
done with silicone oil as well) and (b) residual PFO droplets
in the anterior chamber and former vitreous cavity. Drop-
lets of PFO adhere to the zonules and pars plana and are
common even after very compulsive surgical removal. Some
investigators believe that PFO is toxic, but the authors have
seen no evidence of toxicity and the patients have excellent
visual function. Some patients develop mild inflammatory
deposits on the retina and the lens, but these disappear
in a few weeks after PFO removal. Intraocular pressure
136 SECTION IV ■ Disease-Sp ecific Managem ent
elevation has been seen by the authors in patients who have
PFO migration into the anterior chamber. This technique,
similar to using vitrectomy plus gas and laser, requires com-
pulsive, 360-degree removal of vitreous traction with empha-
sis on traction to all retinal breaks (Fig. 14.7). Adequate
peripheral vitrectomy requires wide-angle visualization and
may require scleral depression. After vitrectomy, PFO is
injected over the optic nerve with a MedOne double-bore
cannula that enables infusion fluid to leak through the exter-
nal lumen of the cannula to equilibrate the intraocular pres-
sure (Fig. 14.8). Confluent laser is applied with the Alcon
25-gauge articulating laser probe to each retinal break and
suspicious area. If SRF does not completely disappear, a very
small drainage retinotomy can be made and SRF aspirated,
being careful to not remove PFO. If vitreous traction is seen
after PFO injection, it can be removed “under” PFO by care-
fully keeping the cutter port outside the bubble of PFO. This
technique is called “interface” vitrectomy by the authors and
is similar to vitrectomy “under” air or “under” silicone oil.
SILICONE OIL
Silicone oil is essential for most cases of PVR and is often indi-
cated for giant retinal breaks and large retinal defects. Silicone
oil has less intrafacial tension (surface tension) than air or gas
but does not absorb, and therefore it remains in the eye indefi-
nitely. Many physicians incorrectly believe (a) silicone oil is toxic
and must be removed in a few months, (b) silicone oil reduces
vision, (c) glaucoma is common with oil and most glaucoma
in patients with silicone oil is oil related, (d) positioning is
not necessary with silicone oil, (e) silicone oil is effective for
Fig u re 1 4 .8 ■ PFO in jectio n
o ver t h e o p t ic n erve wit h a
25-gauge dual-bore cannula.
macular holes, (f) 5,000 cs has less emulsification than 1,000
cs oil, and (g) inferior peripheral iridectomy is needed in eyes
with in-the-bag posterior chamber lenses and intact capsules.
Unlike gas bubbles, silicone oil does not expand and therefore
can be used if the patient must fly. Positioning is required with
breaks below the horizontal meridian if retinopexy has been
performed to inferior breaks. The patient can lie on his or her
side for temporal and nasal retinal breaks. Retinopexy can be
omitted for macular holes causing retina detachment; large
breaks and retinectomies likely to cause PVR if treated; and
breaks not well seen during vitrectomy that can be treated
later when there is no SRF, retinal edema, or inflammation.
This technique is referred to as using “rhegmatogenous con-
finement” for “retinopexy avoidance.”
Air-silicone exchange rather than fluid-silicone
exchange is used by the authors. It is done as described in
Chapters 4 and 5 using 25-gauge technique.
SUMMARY
Vitrectomy as first surgery monotherapy for rhegmatog-
enous retinal detachment has significant advantages over
scleral buckling (3) with respect to strabismus, refractive
effects, ptosis, pain, conjunctival hyperemia, and chemosis.
Vitrectomy requires advanced tools such as wide-angle
visualization, PFO, and fast cutting. Vitrectomy approaches
require strict attention to the peripheral vitreous and espe-
cially traction on flap tears. A 25-gauge vitrectomy offers
better fluidic stability than 20- or 23-gauge vitrectomy and
less patient discomfort, less conjunctival damage, less sub-
conjunctival hemorrhage, and less chemosis.
 CHAPTER 14 ■ Vitrectom y for Retinal Detachm ent 137

References 3. Brazitikos PD, Androudi S, Christen WG, Stangos NT. Primary

pars plana vitrectomy versus scleral buckle surgery for the treat-
1. Kapran Z, Acar N, Altan T, et al. 25-Gauge sutureless vitrectomy ment of pseudophakic retinal detachment: a randomized clinical
with oblique sclerotomies for the management of retinal detach- trial. Retina 2005 Dec;25(8):957–964.
ment in pseudophakic and phakic eyes. Eur J Ophthalmol 2009
Sep-Oct;19(5):853–860.
2. Arya AV, Emerson JW, Engelbert M, et al. Surgical management
of pseudophakic retinal detachments: a meta-analysis. Ophthal-
mology 2006 Oct;113(10):1724–1733.
C H AP TER
15
GIANT BREAKS
Giant retinal breaks are traditionally defined as
breaks extending for greater than 90 degrees.
Because of the propensity of the retina to fold
over, surgical management of these detachments
was very difficult until the advent of vitrectomy and intraoc-
ular gas. The next major advance occurred when Chang ini-
tiated the application of perfluorocarbon (PFO) liquids to
giant break surgery (1). It is now relatively straightforward
to achieve surgical success in most cases. The long-term suc-
cess rate is determined by the incidence of proliferative vit-
reoretinopathy (PVR). Although vitreous dynamics, trauma,
and hereditary peripheral retinal changes play a role, the
exact pathogenesis of giant breaks remains unclear. The high
incidence of associated PVR presumably relates to the expo-
sure of vast areas of retinal pigment epithelium (RPE) and
retinal glial cells to the vitreous matrix. RPE cells migrate
along the vitreous and retinal surfaces. The exposed edge of
retina presumably enables glial cells to migrate to the retinal
surface as well. Vitreous is virtually never observed on the
photoreceptor side of the retina in these cases. Inward fold-
ing of the retina is caused by hypocellular contraction of the
vitreous, the inherent elasticity of the internal limiting mem-
brane (ILM), and PVR. Although one can occasionally dis-
tinguish between giant breaks and giant disinsertions, this
differentiation does not seem to play a role in management
strategies or prognosis.
CASE SELECTION
Quadrantic giant breaks can rarely be managed effectively
with scleral buckling alone if there is minimal folding of the
retina. Mild degrees of folding can be managed by adding
expanding C3F8 gas to the scleral buckling modality. Most
giant breaks with retinal folding are best managed with vit-
rectomy and PFO liquids. Giant breaks combined with PVR,
vitreous to a wound, dislocated lenses, or vitreous hemor-
rhage are absolute indications for vitrectomy, regardless of
the size of the break.
138
SURGICAL SEQUENCE AND
TECHNIQUES
Pat ie n t Ed ucat io n
Patients should be psychologically prepared for 1 to 3 weeks
in the face-down position after surgery, unless medium-term
PFO liquids (2–3 weeks) are used instead of gas. Patients
treated with silicone oil must be positioned as well if the
break extends below the horizontal meridian.
In cisio n s
The authors use 25-gauge vitrectomy, laser, internal drain-
age of subretinal fluid (SRF), and air-silicone exchange
(ASX) or air-gas exchange (AGX) in all giant break cases.
Intraoperative use of perfluoron (PFO) allows adequate
repositioning of the folded retina in large retinal tears.
Medium-term PFO prevents slippage of the giant break but
requires two procedures and cannot be used for superior
giant breaks. The patient must lie on his or her side for
nasal or temporal breaks and be seated, standing, or semi-
reclining for inferior breaks with the medium-term PFO
technique.
Man ag e m e n t o f t h e Le n s
Phaco-vit and IOL insertion should be used if there is
significant cataract preventing sufficient visualization. In
the past, giant retinal tears were considered indications
for pars plana lensectomy, but the authors do not find a
need for lensectomy at this time with modern vitrectomy
techniques.
Vit re ct o m y
Vitreous removal should be performed using very low suc-
tion force with proportional suction control and high cut-
ting rates (5,000 cuts per minute) because of marked retinal

mobility. The vitreous is very rarely attached to the posterior
edge of the giant break. The authors use 25-gauge technique
for all giant break cases because the greater fluidic stabil-
ity makes removal of vitreous near the highly mobile retina
much safer.
The vitreous must be trimmed to the anterior retinal
surface and away from the anterior flap to prevent vitreous
entrapment in the SRF drainage cannula and to prevent late
vitreous traction from causing redetachment (Fig. 15.1). It is
better to remove the detached nonpigmented pars plana epi-
thelium anterior to the break during removal of the periph-
eral vitreous.
Fluid -Air-Gas Exch an g e
Internal drainage of SRF and fluid-air exchange can be per-
formed without PFO liquids in selected cases with smaller
breaks and little or no folding. Internal drainage of SRF
must precede fluid-air exchange and continue during the
exchange. A soft-tip extrusion cannula should be positioned
just anterior to the edge of the giant break. The retina will
be gently pulled into position as the SRF is drained. Rotating
the eye toward the giant break enables drainage of the vitre-
ous fluid simultaneous with the SRF and facilitates complete
drainage of SRF and better repositioning of the retina.
Pe rfl uo ro carb o n Liq uid s
PFO liquids permit removal of all SRF and position the ret-
ina in a nonfolded state near its original position, permitting
CHAPTER 15 ■ Giant Breaks 139
Fig ure 1 5 .1 ■ Peripheral vitreous
must be removed for 360 degrees
as close as possible to the retinal
surface. The anterior flap of the
giant break should be rem oved.
A 25-gauge vitrectom y is ideal for
giant breaks because of greater
fluidic stability.
laser retinopexy to both the retina and the RPE. PFO liquids
are surface tension management agents similar to silicone
oil, air, and gas. Silicone and gas float because their density is
less than infusion fluid. PFO liquids are denser than infusion
fluid and retina and migrate to the lowest part of the eye.
Retina is less dense than PFO liquids and therefore floats
to its original position, limited only by vitreoretinal traction
and its inherent stiffness. Similarly, SRF is less dense than
PFO liquids and floats up through the giant break and into
the anterior vitreous cavity. The authors use PFO from Alcon
for all cases.
PFO liquid should be injected on the anterior side of the
retina (Fig. 15.2), near the optic nerve with a MedOne dual-
bore 25-gauge viscous fluid injection cannula. Because the
endoilluminator is plugging one sclerotomy and both PFO
and balanced salt solution (BSS) are incompressible, BSS Plus
must be allowed to egress through the dual-port cannula to
prevent increasing the intraocular pressure. The PFO should
be injected until the retina is unfolded and all SRF is gone.
PFO must reach the pars plana when injection is complete and
all SRF is gone (Fig. 15.3). David Wong has taught us that a
full fill with PFO with the infusion fluid turned off prevents
posterior slippage of the giant break. Exchange of PFO for
air to enable AGX or ASX must be performed precisely to
prevent slippage; the soft-tip cannula for PFO removal must
be positioned above the PFO-fluid interface at all times so
that all saline solution (containing SRF) is removed before any
PFO. The cannula must be slowly advanced posteriorly as the
exchange is done (Fig. 15.4). The BSS-PFO interface is easier
to see than with perfluorodecalin.
140 SECTION IV ■ Disease-Sp ecific Managem ent

Fig ure 1 5 .2 ■ Perfluoron (PFO) is slowly injected near the optic nerve on the anterior surface
of the retina, which displaces SRF anteriorly and unfolds the flap.

Fig ure 1 5 .3 ■ PFO injection is continued until the entire vial has been injected and the PFO is
anterior to the giant break.

If PFO liquids are used when giant breaks and PVR
coexist, care must be taken to avoid PFO liquids from
entering the subretinal space by removing all epiretinal
membrane by inside-out forceps membrane peeling with
the Alcon 25-gauge DSP forceps (Fig. 15.5). PFO liquids
will enter the subretinal space when the difference between
forces due to specific gravity and intrafacial tension effects
CHAPTER 15 ■ Giant Breaks 141
Fig u re 1 5 . 4 ■ Th e so ft -t ip
cannula for PFO rem oval m ust be
p osition ed ab ove th e PFO-fluid
interface at all tim es so that all
saline solution (containing SRF) is
rem oved before any PFO.
is less than the combined stiffness of the retina and PVR
membranes (2–4).
Re t in o p e xy
Confluent laser retinopexy should be used at the posterior
margin of the retinal break (Fig. 15.6). Multiple spots with
Fig ure 1 5 .5 ■ Ep iretinal m em -
b ranes should b e rem oved with
inside-out forcep s p eeling , using
conform al, d iam on d -coated , or
DSP forcep s.
142 SECTION IV ■ Disease-Sp ecific Managem ent
Fig u re 1 5 .6 ■ Confluent laser end op hotocoag ulation should b e ap p lied in rows on the
posterior edge of the break and extended to the ora at the two ends of the break.
interspersed untreated spaces necessitate a larger treatment area
and create the problem of leakage of SRF between the spots. Care
should be taken to extend the laser (treatment) around the ends
of the break to the pars plana to avoid SRF leakage. A 532-nm,
diode-pumped, frequency up-converted laser is the authors’
preference (5–7). Treatment intensity is more difficult to judge
with red or near infrared laser energy. Cryo is thought to cause
increased PVR (8–13) and retinal slippage compared to laser.
Transscleral diathermy can damage the sclera but can be very
effective in an air-filled eye, although it is no longer utilized.
Transscleral laser is less predictable, is not feasible in darkly
pigmented patients, and, like cryo and diathermy, requires
exposure of the scleral surface (14). Endophotocoagulation is
the preferred mode for all vitrectomy cases. Laser indirect oph-
thalmoscope may cause damage to the cornea, iris, and lens.
Po st o p e rat ive Surface Te n sio n Man ag e m e n t
The options for postoperative surface tension management
include medium-term perfluorocarbon liquids, silicone, and
gas. The advantage of medium-term PFO is that the slippage
that occurs at the time of PFO-gas or PFO-silicone exchange
is avoided. The minor disadvantage is that a second surgi-
cal procedure is required to remove the PFO. PFO is ideal
for medium-term tamponade but must be used off-label in
the United States. Gas will absorb, eliminating the need for
removal. Gas expands with air travel, bubble size decreases
with time, and the view is problematic for the patient and the
physician. Silicone is the best option for PVR combined with
giant breaks. The exchange from PFO to gas or silicone can
cause posterior slippage. Laser retinopexy must be applied
before this slippage occurs in order to position the laser cor-
rectly. The PFO-air exchange prior to AGX or ASX must be
performed by placing the extrusion cannula right behind the
lens or IOL and staying just below the air-BSS + SRF inter-
face as it moves back. In this way, the PFO is removed last
and posterior slippage minimized. The eye should be rotated
toward the break as the last few drops of PFO are removed
to ensure removal of all SRF. The patient must be positioned
prone immediately after surgery and continuously in order
for the retina to move back to a nonredundant position. On
some occasions, permanent folds will occur. Retinal folds do
not represent a serious problem unless the macula is involved
or very large areas of RPE are exposed. Damage to the retina
and/or RPE occurring in a prolonged attempt to eliminate
folds may unnecessarily stimulate PVR.
It is tempting to allow the patient to position part-
time or position other than prone if gas or silicone oil is
used or to shorten the prone period. These compromises, in
the authors’ opinion, reduce success rates with gas and oil.
Medium-term PFO allows all patients to be supine for sleep-
ing, and they can be semireclining with an inferior break and
on their side for nasal or temporal breaks.
RESULTS
With the previously described methodology, about 95% of
cases can be successfully reattached in the operating room
and remain reattached for the first 3 weeks. However, the

long-term results range from 50% to 90%, depending
on PVR, surgical techniques, and case selection (15,16).
PVR can be managed by reoperation with the approach
described earlier. Epimacular membranes (macular pucker)
occur in a significant number of cases and can be managed
effectively with substantial visual recovery after vitrectomy
revision and end-opening, inside-out forceps membrane
peeling.
MANAGEMENT OF THE FELLOW EYE
The high incidence of bilaterality in the nontraumatic,
genetically determined cases of giant breaks is of great
concern (17). The potential for giant break in the other
eye raises the question of prophylaxis to the other eye.
It is known that retinal breaks and even giant breaks fre-
quently occur at the posterior margin of previous retin-
opexy marks. It is well known that hypocellular vitreous
contraction can pull the retina away from a high, broad
scleral buckle. The authors currently laser retinal breaks
in the other eye but do not perform 360-degree retinopexy
or scleral buckling.
References
1. Chang S, Lincoff H, Zimmerman NJ, Fuchs W. Giant retinal
tears: surigical techniques and results using perfluorocarbon
liquids. Arch Ophthalmol 1989;107:761.
2. Glaser BM, Carter JB, Kupperman BD, Michels RG. Perfluo-
octane in the treatment of giant retinal tears with PVR. Oph-
thalmology 1991;98:1613.
3. Darmakusma IE, Glaser BM, Sjaarda RN, et al. The use of per-
fluoro-octane in the management of giant retinal tears without
PVR. Retina 1994;14:323.
CHAPTER 15 ■ Giant Breaks 143
4. Verstraeten T, Williams GA, Chang S, et al. Lens-sparing
vitrectomy with perfluorocarbon liquid for the primary treat-
ment of giant retinal tears. Ophthalmology 1995;102:17.
5. Charles S. Endophotocoagulation. Retina 1981;1:117.
6. Yoon YH, Marmour MF. Rapid enhancement of retinal adhesion
by laser photocoagulation. Ophthalmology 1988;95:1385.
7. Powell JO, Bresnick GH, Yanoff M, et al. Ocular effects of argon
laser radiation. II. Histopathology of chorioretinal lesions.
Am J Ophthalmol 1971;71:1267.
8. Campochiaro PA, Kaden IH, Vidaurri-Leal JS, Glaser BM. Cry-
otherapy enhances viable intravitreal dispersion of retinal pig-
ment epithelial cells. Arch Ophthalmol 1984;103:434.
9. Kreissig I, Lincoff H. Mechanism of retinal attachment after
cryosurgery. Trans Ophthalmol Soc UK 1975;95:148.
10. Johnson RN, Irvine AR, Wood IS. Endolaser, cryopexy, and
retinal reattachment in the air-filled eye: a clinicopathologic
correlation. Arch Ophthalmol 1987;105:231.
11. Jaccoma EH, Conway BP, Campochiaro PA. Cryotherapy
causes extensive breakdown of the blood retina barrier. A com-
parison with argon laser photocoagulation. Arch Ophthalmol
1985;103:1728.
12. Campochiaro PA, Kaden IH, Vidaurri-leal J, Glaser BM. Cryo-
therapy enhances intravitreal dispersion of viable retinal pig-
ment epithelial cells. Arch Ophthalmol 1985;103:434.
13. Glaser BM, Vidaurri-leal J, Michels RG, Campochiaro PA. Cry-
otherapy during surgery for giant retinal tears and intravitreal
dispersion of viable pigment epithelial cells. Ophthalmology
1993;100:466.
14. Haller JA, Lim J, Goldberg MF. Pilot trial of transscleral diode
laser retinopexy in retinal detachment surgery. Arch Ophthalmol
1993;111:952.
15. Freeman HM, Schepens CL, Couvillion GC. Current manage-
ment of giant retinal breaks 11. Trans Am Acad Ophthalmol
Otolaryngol 1970;74(1):59.
16. Machemer R, Allen AW. Retinal tears 180 degrees and greater.
Management with vitrectomy and vitreoretinal gas. Arch Oph-
thalmol 1976;94(8):1340.
17. Freeman HM. Fellow eyes of giant retinal breaks. Mod Probl
Ophthalmol 1979;20:267.
C H AP TER
16
PROLIFERATIVE
VITREORETINOPATHY
Proliferative vitreoretinopathy (PVR) can be
thought of as a reparative or scarring mechanism.
Contact inhibition is lost because of mechanical
disruption of a tissue such as the retinal pigment
epithelium (RPE), and the response is migration, contraction,
and moderate proliferation of cells. A defect in the internal
limiting membrane (ILM) can be created by vitreous trac-
tion, a retinal break, or membrane dissection. Glial cells then
migrate, proliferate to some extent, and contract the retinal
surface as if to repair the retina-ILM defect. Retinal breaks
“heal” as a result of this mechanism as has been shown in
macular hole surgery and retinotomies for submacular sur-
gery. Similarly, if the retina is separated from the RPE, the
RPE cells migrate to the retinal surface and the posterior
surface of the posterior vitreous cortex (PVC). Monocytes
can gain access to the retinal surface from the iris and cili-
ary body capillaries as a result of ocular inflammation. It is
known that these cells can migrate, proliferate, and contract
on the retinal surface (1–13). Glial, RPE, and monocytic cells
share in common the migration along an existing substrate
or to a new tissue surface. As cellular migration occurs, intra-
cellular contraction of the cells occurs, creating tangential
traction on the retina. All healing mechanisms studied have
demonstrated a microtubule, smooth muscle-like contrac-
tion process occurring with many cells acting in concert. The
myofibroblasts have coated pits, which have specific recep-
tor sites for collagen, fibrin, and elastin. These pits contain
fibronectin, which allows the cells to attach to collagen fibers
(14,15). Growth factors (TGF-b, etc.), metalloproteinases,
fibronectin, and receptors for collagen, elastin, and fibrin are
components of the migration and contraction phenomenon.
Proliferation (mitotic activity) is of much lesser importance.
This process can be thought of as hypocellular periretinal
scarring, analogous to hypocellular vitreous contraction. PVR
is somewhat of a misnomer; there is minimal mitotic activity
in PVR, and antimetabolites have been shown to be ineffec-
tive; in fact, the disease is more like a keloid. To emphasize the
hypocellular nature of this process, the authors suggest taking
144
the “P” out of PVR. Like a keloid, PVR is excessive healing or
scarring. Scarring is biologically identical to healing; scarring
is basically a functional description. PVR should be thought
of as a hypocellular interaction between various cell types
and the extracellular matrix (ECM). RPE cells, glial cells, and
inflammatory cells have been shown to be involved in PVR.
Better understanding of the disease process leads to better
treatment strategies; excessive retinopexy leads to PVR, and
repeated surgical interventions with brief intervals between
procedures are common causes of PVR. Lengthy procedures,
extensive retinopexy, iris trauma (retractors), combined
phaco and intraocular lens (IOL) implantation, and exces-
sive peripheral “membrane peeling” instead of retinectomy
are all factors in iatrogenic PVR. Unlike cancer surgery or
draining an abscess, the goal is minimal tissue destruction,
not removal of all pathologic tissue. The PVR surgeon should
think like a plastic surgeon, not a cancer surgeon.
Tangential shortening of the cells occurs prior to col-
lagen production, which may be thought of as a late stabi-
lization phase. The generation of basement membrane and
collagen can be thought of as a recreation of Bruch’s mem-
brane or the ILM, as if to duplicate the process that occurs in
embryonic development.
PVR can be localized and create isolated star folds, fixed
folds, subretinal changes, or epimacular membranes (16– 18).
It can be more widespread, which gave rise to the no longer
used term, “massive periretinal proliferation.” If the prolif-
eration of glial or RPE cells extended onto the contiguous
posterior surface of the PVC, the older term “massive vitre-
ous retraction” would have been used. At this time, there is
no proof of the role of hyalocytes in the pathogenesis.
Understanding surgical anatomy is a crucial component
of surgical intervention. Core vitrectomy is actually removal
of confluent PVC and anterior vitreous cortex (AVC) in the
frontal plane (FP); there is no spherical core vitreous in these
cases. It is important to recall that the vitreous base has no
volume and cannot be “shaved” as surgeons often state; it is
a zone of adherence between the peripheral vitreous and the

retina. Often surgeons state that they perform vitreous base
dissection; this is simply not the case, although so-called
shaving of the peripheral vitreous close to the retina surface
is essential. Circumferential vitreous collagen fibers shortened
by hypocellular contraction pull the equator inward; these
fibers are contiguous with the vitreous base and rarely remov-
able by peeling. Anterior-posterior–oriented vitreous col-
lagen formerly contiguous with the vitreous base shortened
by hypocellular contraction causes what the principal author
initially identified as anterior loop traction; resection of this
component of anterior PVR is essential to reduce the anterior
pull on equatorial retina. Epiretinal membranes (ERMs) are
always behind to the posterior margin of the vitreous base;
peeling peripheral PVR membranes is a misnomer as well.
Scarring in the subretinal space can occur in many
physical configurations, as discussed later (see “Subreti-
nal Proliferation”). A placoid configuration can create an
inverted star fold configuration. Subretinal bands (strands)
occur presumably as tubes of RPE cells proliferate along a
fibrin strand; these then contract, causing an extended den-
dritic configuration. If the band is circumferential at the
midretinal level, it can create an annular configuration with
a resultant closed cone retinal detachment.
CASE SELECTION
Vit re ct o m y Ve rsus Scle ral Bucklin g
Vitrectomy should be thought of as a mechanical approach
to a mechanical problem caused by retinal traction. It has no
known prophylactic benefit in the prevention of PVR nor is
it an approach to be utilized only after several scleral buckles
are tried. The authors utilize 25-gauge vitrectomy techniques
without scleral buckling for practically all cases of retinal
detachment secondary to PVR.
Op e rab ilit y
Extensive star folds, even in a closed cone (funnel) configuration,
are usually operable. If extensive membrane peeling has been
performed, the recurrent proliferation may be more adherent to
the retinal surface. Excessive retinopexy probably causes RPE
and glial reproliferation, while retinal surface dissection proba-
bly causes glial proliferation. If extensive subretinal placoid pro-
liferation is present, this is often inoperable. Redundant retina
can simulate diffuse subretinal placoid proliferation, making
clinical assessment difficult. The presence of large retinal breaks
and extensive surface proliferation was thought of as inoperable
before the advent of epiretinal and subretinal dissection, inter-
nal drainage of subretinal fluid (SRF), retinectomy, and silicone
oil. In addition to the mechanically inoperable configurations,
there are many patient-based and biologic decisions related to
operability. If the patient has an apparently operable mechani-
cal configuration but has had many recurrences after vitrec-
tomy, surgery might be contraindicated because of biologic
CHAPTER 16 ■ Proliferative Vitreoretinopathy 145
behavior, especially if the patient has another functional eye,
has medical problems, or is very elderly. Iris neovasculariza-
tion, glaucomatous cupping, and severe uveitis can influence
the decision of operability as well.
Cases with extremely recent proliferative activity and an
inflammatory component have a worse prognosis than inac-
tive cases (19). In such cases, a period of observation and
subconjunctival steroids should precede surgery (19; R.G.
Michels, personal communication). A definite life cycle as
described by S. J. Ryan characterizes this reparative or scar-
ring process. Recurrent proliferation is reduced if reoperation
is delayed until the proliferation stabilizes and inflammation
is minimal. Proliferative inactivity can be judged by the pres-
ence of minimal cells and flare, waning conjunctival hyper-
emia, comfort, and lack of further progression of ERMs.
SURGICAL SEQUENCE AND TECHNIQUES
Man ag e m e n t o f t h e Le n s o r In t rao cular Le n s
The crystalline lens may be retained in most cases of poste-
rior PVR but may require removal in the presence of anterior
PVR to allow adequate dissection of the anterior vitreous and
prevent cyclitic membrane formation. Endocapsular lensec-
tomy should be utilized if the eye is extremely inflamed;
otherwise, phaco-vit should be used to facilitate posterior
containment of silicone oil.
If a posterior chamber IOL is present, it should be
retained. Selected anterior chamber and iris plane lenses must
be removed. Implant removal can occasionally result in an
intraoperative hemorrhage or corneal and iris damage and
compromise the surgical result. Corneal incisions, viscoelas-
tics, and haptic cutting minimize the incidence of intraop-
erative hemorrhage. Cut haptics can be left if bleeding is
anticipated because of fibrous scarring around the haptic.
The hypotony that occurs during lens removal can theoreti-
cally cause choroidal hemorrhage.
COMPARTMENTALIZATION
Serum components and inflammation play a role in the cau-
sation and acceleration of PVR. A source tissue produces
growth factors, which then target substrate tissues. The tar-
get tissue (retina) is in the same compartment as the source
in PVR, but on remote surfaces in proliferative diabetic retin-
opathy (vascular endothelial growth factor [VEGF], iris and
trabecular meshwork). Viscoelastics, fibrin, inflammation,
and blood can be thought of as culture media enhancing PVR.
The lens or IOL creates a two-compartment eye, prolonging
the cell culture–like environment and providing a migration
substrate. Removal of the lens in PVR cases increases the
egress rate of cytokines, cells, and serum components such
as fibronectin through the trabecular meshwork. Similarly,
gases and silicone produce sequestration of cells and factors
at the retinal-bubble interface, enhancing PVR.
146 SECTION IV ■ Disease-Sp ecific Managem ent
VITRECTOMY
The AVC is removed first, along with all vitreous attachments to
cataract and trauma wounds, and the iris. Very low suction force
using the proportional suction control should be utilized to pre-
vent retinal breaks from suction-induced vitreoretinal traction.
The highest cutting rates should be used to increase fluidic
stability. The PVC is frequently in contact with the AVC and
removable in a single step. A posterior vitreous detachment
is almost always present in PVR cases. Hypocellular vitreous
contraction causes the PVC and AVC to come together in a
“frontal plane” configuration (Fig. 16.1). “Core vitrectomy” is a
misnomer; a core configuration does not exist in PVR eyes.
An t e rio r Pro life rat ive Vit re o re t in o p at h y
Anteroposterior vitreous fibers normally extend from the
retina at the posterior edge of the vitreous base to the pars
plana, ciliary body, and iris. Frequently, these fibers undergo
hypocellular vitreous contraction and pull the equatorial ret-
ina anteriorly into a circumferential fold. The author (S.C.)
first described this phenomenon as anterior loop contraction
in 1975. A circumferential equatorial ring of ERM or con-
tracted equatorial vitreous is typically present in PVR. Hypo-
cellular contraction of the AVC and PVC is also common
in PVR cases. Collectively, these structures are known
as anterior PVR. They are frequently incorrectly termed
“vitreous membranes” or the “vitreous base.” Anterior PVR
must be relieved to allow retinal reattachment.
Scleral depression by the assistant facilitates anterior PVR
dissection. The radial component of anterior PVR dissection
Fig u re 1 6 .1 ■ Th e AVC an d
PVC un d erg o h yp ocellular con -
traction of th e collag en fib ers
resulting in confluence and an FP
sheet. The FP com ponent should
be rem oved before forceps m em -
brane peeling of star folds.
can be done with the vitreous cutter, if broad, or delamination
scissors, if narrow (Fig. 16.2). The 25-gauge ILM forceps can
be used to peel posterior ERMs in an anterior direction to the
posterior edge of the vitreous base. Blunt dissection with the
vitreous cutter and endoilluminator can be used to separate
vitreous and ERM from peripheral retina. Delamination scis-
sors can be used to delaminate or segment the circumferen-
tial component (Fig. 16.3).
Ep ire t in al Me m b ran e s
Membrane peeling can cause retinal breaks and hemorrhage
implicated in the recurrence of cellular migration, contrac-
tion, and proliferation. Often, the ERMs peel easily with none
of these problems and allow complete release of tangential
traction (Fig. 16.4). End-grasping forceps membrane
peeling is preferable to using pics or forceps with one blade
under the ERM because the end-grasping forceps contact
the ERM, not the retina. Use of forceps without pics reduces
the incidence of iatrogenic retinal breaks. The principal author
refers to this approach as inside-out forceps membrane peel-
ing; it is the same direction of dissection used for epimacular
membranes, ILM, as well as scissors segmentation and delami-
nation. The senior author initially developed diamond-coated,
20-gauge forceps for membrane peeling, then developed
conformal forceps, and more recently transitioned to Alcon
25-gauge DSP ILM forceps with conformal tips. The senior
author has not used pics since 1978. Alcon 25-gauge
DSP ILM forceps are ideal for this technique. The end-
grasping forceps should be applied to the apparent epicenter
of a star fold, or the outer margin only if clearly visible. It is
 CHAPTER 16 ■ Proliferative Vitreoretinopathy 147

Fig ure 1 6 .2 ■ The radial (anter-

op osterior) com p onen t of ante-
rior PVR is also called anterior loop
traction. It can b e resected with
curved scissors if it is not b road
enough to perm it transaction with
the vitreous cutter. The 25-gauge
cutter is p referred b ecause the
com p act tip facilitates access.

usually best to use the forceps to peel toward the periphery
because central retina is much stronger than peripheral
retina. Peripheral ERMs may extend to the vitreous base; once
lifted from the retina surface, they can be removed with the
vitreous cutter often while removing anterior loop traction.
If the membranes are tightly adherent, they should be seg-
mented and/or delaminated with 25-gauge curved scissors.
Retinectomy may be required if forceps membrane peeling,
segmentation, or delamination of the ERM causing a star
fold, fixed fold, or zone of retinal-retinal adherence cannot
be completed. A specific attempt must be made to release
the ERM causing circumferential traction at the equator by
using scissors segmentation or delamination. If this cannot
be safely accomplished because of tight adherence to the

Fig u re 1 6 .3 ■ Curved scissors

can be used to segm ent or delam -
inate ERM and condensed circum -
ferential fib ers on the eq uatorial
surface of the retina.
148 SECTION IV ■ Disease-Sp ecific Managem ent
retinal surface, this portion of ERM can be left intact and
supported by the very broad and high buckle or retinectomy
performed. Frequently, dense ERM can be delaminated from
the retinal surface with curved delamination scissors with
the blades parallel to the retinal surface. Segmentation of the
denser portions of an ERM using the curved scissors is very
effective in releasing traction and tends to be underutilized
(Fig. 16.5). When the entire surface of the retina can be
Fig ure 1 6 .4 ■ Inside-out forceps
m em b ran e p eelin g with Alcon
25-gauge DSP ILM forceps is used
to rem ove ERMs. Pics, viscodissec-
tion, and forceps with one b lad e
under the ERM are m ore likely to
dam age the retinal surface.
visualized and the sharp angulations have been converted to
rounded contours, dissection is complete.
Sub re t in al Pro life rat io n
Many anatomic forms of subretinal proliferation occur, just
as there is variability in the configuration of ERMs. The term
“strands” is a misnomer because of the frequent occurrence
of placoid elements.
Fig u re 1 6 .5 ■ Insid e-out seg -
m en tation of d en se ERM with
curved scissors is very effective
and tends to be underutilized.
 CHAPTER 16 ■ Proliferative Vitreoretinopathy 149

Fig ure 1 6 .6 ■ Dendritic subreti-

nal p roliferation can b e rem oved
with forceps through a preexisting
retinal defect.

Placoid subretinal proliferation, if moderate in extent,
is managed by attempting internal drainage of SRF, fluid-
air exchange, and completion of SRF drainage. If the retina
becomes attached, subretinal surgery, and/or scleral buckling
are not required. If there is extensive proliferation, subretinal
forceps removal is required.
Dendritic proliferation can usually be managed by
forceps removal (Fig. 16.6) or segmentation, followed by
internal drainage of SRF, fluid-air exchange, and completion
of SRF drainage. The diamond-coated, end-opening forceps,
or 25-gauge DSP forceps are pushed through the retina over
the densest part of the subretinal proliferation (Fig. 16.7)
and used to grasp and remove the tissue (Fig. 16.8). Care
must be taken to avoid entrapping the retina in the forceps
or damaging the RPE or choroid. Frequently, the dendrite
will break, releasing the traction.

Fig ure 1 6 .7 ■ Dendritic subret-

inal p roliferation can be rem oved
with forceps after punch-through
retin o to m y an d sp re ad in g t o
enlarg e the op ening and g rasp
the “strand.”
150 SECTION IV ■ Disease-Sp ecific Managem ent
Fig ure 1 6 .8 ■ Sequential regrasping and use of the endoillum inator as a fulcrum enable long
subretinal dend rites to be com pletely rem oved.
Extension of dendritic proliferation in a rather posterior,
circumferential configuration gives rise to an annular ring con-
figuration. After completion of the vitrectomy and inside-
out forceps membrane peeling, the scissors are placed in
the subretinal space through an existing retinal break or a
punch-through retinotomy is made with the forceps in the
retina over the subretinal annulus. Scissors allow the sur-
geon to transect the annulus at a considerable distance from
the retinal defect used to gain entry to the subretinal space.
The endoilluminator can be used to illuminate the subretinal
space and to contact and palpate the retina to determine if
the traction has been alleviated. At times, it is necessary to
transect the annulus in several locations. If extensive mem-
brane is present in the subretinal space, a 90-degree or greater
circumferential retinectomy can be performed to allow visu-
alization and access to the subretinal space, although this
is rarely required. Direct puncture (punch-through) retin-
otomies can allow access to very posterior subretinal pro-
liferation. Internal drainage of SRF, fluid-air exchange, and
completion of SRF drainage followed by endophotocoagula-
tion and air-silicone exchange are required in most of these
severe PVR cases.
RETINECTOMY
For three decades, the principal author has advocated cir-
cumferential retinectomy rather than Machemer’s technique
of relaxing retinotomy. Retinectomy removes all tissue
anterior to the circumferential cut, reducing postoperative
hypotony and peripheral PVR recurrences. Unless retina
is incarcerated in a trauma wound or transscleral drain-
age site from prior scleral buckling, retinectomy should be
performed with air, perfluoron (PFO), or silicone oil in the
eye. The author uses the term “incremental retinectomy” to
make the point that it is usually better to do just enough
retinectomy to reattach the retina, which is best accom-
plished with a surface tension agent in the eye. If the retinec-
tomy reaches 270 degrees, it is usually better to extend for a
full 360 degrees because the remaining unresected quadrant
usually contracts postoperatively.
Th e Re at t ach m e n t Exp e rim e n t
The reattachment experiment is defined as the combination of
internal drainage of SRF, fluid-air exchange, and completion
of internal drainage of SRF in order to determine the need
for further dissection. In contrast to the normally low tran-
sretinal pressure gradient, a somewhat higher gradient is cre-
ated, forcing a moderately stiffened retina against the RPE.
Air infusion through the infusion cannula and SRF aspiration
through the soft-tip cannula should be utilized for this step.
The authors use the Alcon 25-gauge soft-tip extrusion can-
nula for drainage of SRF and all exchanges except silicone
oil and PFO injection and silicone oil removal. On occa-
sion, this step will uncover posterior areas of ERM not seen
before the retinal cone was opened up. Forceps membrane
peeling, delamination, or segmentation of these now visible

Fig ure 1 6 .9 ■ Interface vitrectom y is defined as positioning the cutter port just p osterior to air,
perfluoron (PFO), or silicone oil so that residual vitreoretinal traction can be rem oved.
membranes can be performed under air, further releasing
tangential traction.
On occasion, further vitreoretinal traction will be
uncovered by the reattachment experiment. This is an ideal
situation for interface vitrectomy (vitrectomy under air). The
port is positioned in the remaining vitreous, outside the air
bubble with the port midway between the retina and the vit-
reous (Fig. 16.9). Moderate to small amounts of subretinal
air appearing during the reattachment experiment indicate
that the traction has not been completely eliminated. Until
the remaining traction on the retina is relieved, it is impos-
sible to remove subretinal air unless the eye is refilled with
infusion fluid. It is better to leave the subretinal air in place
and proceed with further peeling, segmentation or delami-
nation, subretinal surgery, or retinectomy. The reattachment
experiment, if successful, should be followed by laser treat-
ment of all retinal breaks and then air-gas exchange or air-
silicone exchange.
At times, retinal defects will enlarge greatly or sponta-
neous defects will occur as the reattachment experiment is
performed. This will allow the release of tangential traction
on foreshortened retinas and permit retinal conformation
with the pigment epithelium. These defects occasionally are
extensive, leading to inoperability, but in many cases, they
are helpful in allowing retinal reattachment.
The basic mechanical problem in PVR retina is retinal
foreshortening combined with the requirement of the retina to
conform to the inner surface of the rigid eye wall. The tensile
CHAPTER 16 ■ Proliferative Vitreoretinopathy 151
strength and contraction forces of periretinal membranes
exceed the strength of the retina and the normal transretinal
pressure gradient by a factor of approximately 100.
Pe rfl uo ro carb o n Liq uid s
PFO can be used to stabilize the retina during ERM dis-
section. The term “tamponade” is inappropriate; PFO is
immiscible in fluid and therefore immiscible in the retina
and SRF. It provides 1.67× inertial and gravitational stabi-
lization. The MedOne dual-bore 25-gauge cannula is ideal
for injecting PFO. The optimal method for injecting PFO
is to attach a short length of tubing between the dual-bore
cannula and a syringe filled with half PFO and half BSS.
The BSS eliminates dead space, enabling use of all the PFO.
If a giant break is present, PFO will enter the subretinal space
if insufficient ERM has been removed. PFO-air exchange fol-
lowed by air-silicone exchange is usually a better approach
than PFO-silicone exchange. PFO-air exchange should be
performed by placing the tip of the soft-tip cannula just
below the air-liquid interface and slowly advancing the
cannula posteriorly following the interface down to the
retina. Removing all liquid vitreous, SRF, and infusion fluid
before PFO is removed prevents posterior slippage of giant
breaks and reaccumulation of SRF as the PFO-air exchange
is performed.
Steep folds can result from retinal incarceration in the
sclera at the prior sclerotomies, trauma sites, or previous
152 SECTION IV ■ Disease-Sp ecific Managem ent

Fig ure 1 6 .1 0 ■ If the reattachm ent experim ent using internal drainage of SRF and fluid-air
exchange results in subretinal air, additional vitrectom y, forceps m em brane peeling, segm entation,
or delam ination should be perform ed to reattach the retina. If the subretinal air persists, increm en-
tal retinectom y should be used. All tissue anterior to the circum ferential cut should be rem oved to
reduce hypotony and reproliferation, which distinguishes this technique from relaxing retinotom y.

drainage sites. If a single fold occurs, a retinal cut made
perpendicular to the long axis of the fold will release the
traction analogous to a plastic surgeon releasing a scar. Mul-
tiple steep folds radiating from an epicenter of periretinal
membrane may require retinectomy including or surround-
ing the epicenter. In general, radial folds require circumfer-
ential cuts just as the typical equatorial, circumferential fold
requires a radial cut.
Diffuse retinal foreshortening that cannot be managed by
membrane peeling, segmentation, or delamination requires
large circumferential retinectomies (Fig. 16.10). Other sur-
geons use the term “relaxing retinotomies,” indicating that
no retina is removed with their approach. The retinectomy
approach was developed by the author to remove all tissue
anterior to the circumferential “relaxing” retinotomy. The
advantages of complete removal are less hypotony from epi-
ciliary tissue, less iris neovasculization from ischemic tissue,
and less anterior movement of a silicone oil bubble due to
contraction of the circumferential ring of pars plana tissue.
Cases requiring large circumferential retinectomies are man-
aged similar to giant breaks with PVR using long-term sili-
cone oil tamponade instead of scleral buckling.
The need for large retinectomies frequently cannot be
anticipated in the office. Excessive retinectomies can result
from underestimation of the effectiveness of periretinal
dissection methods. Large retinectomies are needed only
after failure of the reattachment experiment. In all cases,
the retinal reattachment experiment should be attempted
after the completion of periretinal membrane dissection and
stopped immediately if subretinal air appears. Incremental
retinectomy with endodiathermy applied to retinal vessels
to be transected should be alternated with incremental addi-
tional drainage of SRF. This process should cease only when
the retina is so rigid that it cannot be mechanically unfolded
or it becomes completely reattached (20).
The contributing factors in achieving intraoperative reat-
tachment include surface tension of the fluid-air interface, size
of the retinal breaks, retinal stiffness, and the contour of the
eye wall. When periretinal membrane dissection is completed
and the reattachment experiment applied, further mechani-
cal forces on the retina may become apparent. Mechanically
holding the retina in position as reattachment is attempted
can be useful at times (21). This can be accomplished with
the endoilluminator, vitreous cutter, or a soft-tip cannula.
Scle ral Bucklin g
The principal author has not used buckles for PVR cases
for almost two decades and never revises buckles if pres-
ent. Buckles and buckle revision increase pain, inflamma-
tion, refractive error, hypotony, strabismus, ptosis, operating
time, and labor cost due to longer procedures. High scleral
buckles coupled with marked retinal foreshortening may
push the posterior retina away from the RPE.

VISCOELASTICS AS VITREOUS
SUBSTITUTES
Viscoelastics are non-Newtonian fluids and have exceedingly
low surface tension (tamponade) effects. They are pseudo-
plastic and will slowly deform and pass through a small retinal
break very easily. In addition, viscoelastics act as a sustained
release medium for cells and cytokines. For these reasons,
viscoelastics have no place in the PVR management.
SILICONE OIL
Silicone oil was first utilized without vitreous surgery
techniques as a surface tension and dissection device, pro-
ducing ERM peeling by forcing the retina back with con-
current drainage of SRF (22–27). This method will not
work with more extensive ERM or subretinal proliferation.
Recent studies with improved silicone oil (28–37) have not
shown the retinal toxicity noted in previous years. Infe-
rior peripheral iridectomy as developed by Ando allows
aqueous humor to pass from the ciliary processes into
the anterior chamber to supply metabolites to the corneal
endothelium (38).
Aqueous access to the cornea reduces silicone keratop-
athy. Keratopathy occurs in less than 10% of cases with long-
term oil (S. Charles, data presented at the Vitreous Society,
1996, using Adatomed 5,000 cs). Reduction of lower molec-
ular weight components, by using higher average molecu-
lar weight and therefore higher viscosity, and purification to
remove metal ions dramatically reduce silicone emulsifica-
tion, corneal changes, and glaucoma.
The physical effects of silicone are based on its inter-
facial (surface) tension resulting from its immiscibility with
water (aqueous). The ability to restore a transretinal pressure
gradient is a function of the size and shape of the retinal
break and the tangential forces (stiffness) on the retina. Vis-
cosity is not a factor in the transretinal pressure gradient,
and the silicone-aqueous interfacial tension is 25 dyne/cm,
which is much less than that of the air (gas) fluid interface,
which is 70 dyne/cm.
Long-term rhegmatogenous confinement (tamponade)
may obviate the need for retinopexy, therefore reducing repro-
liferation in response to tissue destruction (39). Long-term
tamponade limits the rhegmatogenous component from
contraction-created breaks and those missed at the time of
surgery. Silicone may act to prevent the wetting of the retinal
surface by cytokines implicated in reproliferation (J. Lean,
personal communication). Silicone can reduce reproliferation
by retinopexy avoidance but appears to have no direct role in
suppressing or stimulating reproliferation. Silicone may also
increase reproliferation because of sequestration of cells and
cytokines at the silicone-retinal interface and prevention of
access of pharmacologic agents to the retina surface.
Retinectomy and multiple previous procedures using
gas are relative indications for using silicone in PVR.
CHAPTER 16 ■ Proliferative Vitreoretinopathy 153
The patient must be informed about possible corneal problems,
glaucoma, and the possibility of silicone removal.
Silicone oil is used for rhegmatogenous confinement
for breaks not identified during surgery and new breaks due
to recurrent PVR. The authors prefer fluid-air exchange fol-
lowed by air-silicone exchange to fluid-silicone exchange
unless IOL fogging occurs. IOL fogging is due to contact
between the air and the IOL; therefore, YAG capsulotomy
or surgical capsule rupture and disruption of the AVC is
required. PFO is an excellent alternative to fluid-air exchange
if IOL fogging occurs. Some surgeons believe incorrectly that
IOL fogging is due to the use of silicone IOLs; it occurs with
all IOLs but is more common with silicone lenses because of
higher Posterior Capsular Opacification (PCO) and therefore
higher YAG rates and greater thermal mass.
Silicone oil should be used for rhegmatogenous confine-
ment often without retinopexy when retinectomy is required
for areas of highly adherent ERMs. The authors often utilize
retinectomy without retinopexy when reoperating cases ini-
tially operated by other surgeons that created a large number
of retinal breaks attempting to peel highly adherent mem-
branes. Long-term surface tension management is a crucial
tool in these difficult cases. Excessive retinopexy causes PVR
recurrences and fibrin syndrome; permanent silicone oil elim-
inates the need for retinopexy, hence the terms “rhegmatog-
enous confinement” and “retinopexy avoidance.” There is an
undue emphasis on silicone oil removal, which often results
in redetachment. Silicone oil complications are minimal when
the oil is confined behind an IOL. The emulsification and
complication rates for 1,000 cs and 5,000 cs oil are equal but
injection and removal are 5× faster with 1,000 cs oil.
Many surgeons incorrectly believe that removal of sili-
cone oil is required to reoperate for PVR or epimacular mem-
branes. The authors find it advantageous to operate “under oil”
just as they operate under air and PFO. This technique is called
interface vitrectomy. Vitreous removal, forceps membrane peel-
ing, retinectomy, and drainage of SRF are all effective under oil,
air, or PFO. Advantages include less surgical trauma, much less
operating time, and realistic assessment of residual forces on
the retina. It is crucial to recall that silicone oil has one third
the surface tension of gas or air; often retinas attached under
air in the operating room are noted to be partially detached
under oil on the first postoperative day. Two-port, 25-gauge
technique is ideal for reoperation “under” silicone oil. The
algorithm is to sequentially remove SRF and top up the oil
with the MedOne 25-gauge viscous fluid injection cannula
when the eye gets soft; then repeat the process until the ret-
ina is attached and the IOP is normal.
RETINAL SUTURES, TACKS,
AND INTENTIONAL INCARCERATION
Permanent fixation of the retina to the sclera has been utilized
to prevent inward curling or posterior sliding of large retinal
flaps or the edges of retinectomies. These methods have the
154 SECTION IV ■ Disease-Sp ecific Managem ent

Fig ure 1 6 .1 1 ■ After all of the SRF has been d rained, all retinal defects are surrounded by three
rows of confluent endophotocoagulation.

risk of retinal or choroidal hemorrhage, tissue damage, and Re t in o p last y

the creation of retinal breaks. Retinal tacks are more safely
Retinoplasty using a synthetic adhesive such as cyanoacrylate
and easily utilized than sutures or intentional retinal incar-
(46,47) could theoretically replace all vitreous substitutes
ceration. They can be repositioned as well as removed later.
and retinopexy. An ideal retinal patch should be very flex-
Different materials have been used for the construction of
ible, more elastic than the retina, less permeable than the
retinal tacks: polyacetal (40), steel (41), titanium (42,43),
retina, and not adhere to the RPE. Aqueous humor (the ideal
and certain alloys (H.G.). Sutures and incarceration are time
vitreous substitute) could be used in place of gas or silicone
consuming and frequently place additional traction on the
if a retinal patch were available.
retina. No permanent fixation method can prevent surface
contraction and detachment. Tacks cause bleeding, reprolif-
eration, distortion, and secondary breaks and have not been
utilized by the principal author for 30 years. RESULTS

Results are highly dependent on case selection, surgical meth-

RETINOPEXY
After the reattachment experiment has been successfully com-
pleted, endophotocoagulation can be applied to retinal breaks.
Postreattachment retinopexy (S.C.) permits treatment of both
the retina and the RPE for stronger adherence. It may limit
the migration of RPE cells implicated in PVR. Because retin-
opexy is implicated in PVR formation (44,45) and inflamma-
tion can cause retinal-retinal adherence, retinopexy should be
limited to the retinal breaks or suspected break areas. Panreti-
nal photocoagulation has never been used by the authors for
PVR because it can cause reproliferation and fibrin syndrome.
Similarly, the authors never use many rows of treatment but
rather surround retinal defects with a moderately wide row of
confluent laser (Fig. 16.11) unless silicone is used for retin-
opexy avoidance in highly active or inflamed cases.
ods, and surgical experience. Using conventional techniques
of vitreous surgery, about 50% to 85% of cases achieved
long-term retinal reattachment (48–50). With the previously
described methods and case selection criteria, 90% of cases
can be repaired surgically, with over 74% remaining attached
over the long term (51,52). About 50% of the cases sustain
improved vision better than 5/200, but many cases require
more than one procedure (53).
COMPLICATIONS
Pro life rat ive Vit re o re t in o p at h y Re curre n ce
A recurrence of PVR with secondary retinal separation is
the most frequent complication. In approximately one half
of the recurring cases, further forceps membrane peeling

air-silicone exchange, and endolaser photocoagulation can
cause successful reattachment.
Although the use of intraocular steroids and 5-fluo-
rouracil and its derivatives (54–61) for PVR cases has been
described, therapeutic efficacy has not been proven. It appears
that the migration phase of the RPE cells is more important
than the proliferation phase in the development of PVR. As
these drugs have a short half-life, better delivery systems are
required than those presently available. Cytotoxic agents have
low therapeutic ratios, which is a significant problem, espe-
cially when coupled with the highly variable drug elution rate
of sustained-release delivery systems. Other systemic medica-
tions, such as colchicine, have no proven use in the treatment
or prevention of PVR and should be avoided. Further under-
standing of the molecular processes of cellular and pericellular
contraction might eventually lead to directed pharmacologic
therapy in the future (akin to anti-VEGF therapy in choroidal
neovascularization). Toxic medications should not be utilized
in human patients on a “what if” approach.
In fl am m at io n
Because retinal breaks are exclusively treated with laser
endocoagulation, the usual exudative detachment and fibrin
syndrome associated with two or three rows of cryopexy are
eliminated. A very rare patient will experience severe inflam-
mation and transient iris neovascularization and go on to
(develop) phthisis bulbi. This rare complication could be from
anterior segment necrosis syndrome, sympathetic uveitis, or
some yet unexplained etiology but is usually due to exces-
sive retinopexy or operating on inflamed eyes. Subconjunc-
tival repository steroids without systemic steroids are used in
every PVR case unless the patient is a steroid responder.
References
1. Machemer R, Van Horn D, Aaberg TM. Pigment epithelial
proliferation. Am J Ophthalmol 1978;85:181.
2. Van Horn DL, Aaberg TM, Modene R, et al. Glial cell prolif-
eration in human retinal detachment with massive periretinal
proliferation. Am J Ophthalmol 1977;84(3):383.
3. Machemer R. Role of the pigment epithelium in vitreous
pathology. Trans Ophthalmol Soc UK 1975;95:402.
4. Machemer R, Laqua H. Pigment epithelium proliferation in
retinal detachment (massive penretinal proliferation). Am
J Ophthalmol 1975;80:1.
5. Machemer R. Massive periretinal proliferation (MPP). 1. Pigment
epithelial proliferation. Mod Prob Ophthalmol 1975;15:227.
6. Mandelhorn M, Machemer R, Fineberg E, et al. Proliferation
and metaplasia of intravitreal retinal pigment epithelium cell
autotransplants. Am J Ophthalmol 1975;80:227.
7. Laqua H, Machemer R. Clinical-pathological correlation in mas-
sive periretinal proliferation. Am J Ophthalmol 1975;80:913.
8. Laqua H, Machemer R. Glial cell proliferation in retinal
detachment (massive penretinal proliferation). Am J Ophthal-
mol 1975;80:602.
9. Campochiaro PA, Jerdan JA, Cardin A, et al. Vitreous aspi-
rates from patients with proliferative vitreoretinopathy stimu-
late retinal pigment epithelial cell migration. Arch Ophthalmol
1985;103(9):1403–1405.
CHAPTER 16 ■ Proliferative Vitreoretinopathy 155
10. Campochiaro PA, Jerdan JA, Glaser BM. Serum contains
chemoattractants for human retinal pigment epithelial cells.
Arch Ophthalmol 1984;102:1830.
11. Campochiaro PA, Glaser BM. Platelet-derived growth factor is
chemotactic for human retinal pigment epithelial cells. Arch
Ophthalmol 1985;103:576.
12. Pastor JC. Proliferative vitreretinopathy: an overview. Surv
Ophthalmol 1998;43(1):3–18.
13. Campochiaro PA. Pathogenis mechanisms in proliferative vit-
reoretinopathy. Arch Ophthalmol 1997;115:237–241.
14. Kohno T, Sorgente N, Ryan SJ. Fibronectin distribution at
the vitreoretinal interface. Invest Ophthalmol Vis Sci 1983;
24(Suppl.):240.
15. Kohno T, Sorgente N, Patterson R, et al. Fibronectin and
Laminin distribution in bovine eye. Jpn J Ophthalmol 1983;
27:496.
16. Anderson DH, Stern WH, Fisher SK, et al. The onset of pigment
epithelial proliferation following retinal detachment. Invest Oph-
thalmol 1981;21:10.
17. Machemer R. Pathogenesis and classification of massive pen-
retinal proliferation. Br J Ophthalmol 1978;62:737.
18. Machemer R, Aaberg TM, Freeman HM, et al. An updated clas-
sification of retinal detachment with proliferative vitreoretin-
opathy. Am J Ophthalmol 1991;112:159–165.
19. Charles S. Presentation at 11th Annual Estelle Doheny Eye
Foundation Conference. Los Anvitreouses, CA, September
1979.
20. Han DP. Relaxing rentinotomies and retinectomies. Surgical
results and predictors of visual outcome. Arch Ophthalmol 1990;
108(5):694–697.
21. Glaser BM. A new method of treating giant tears without
patient rotation with the use of intraocular gas and silicone
oil. Presentation at Annual Meeting of American Academy of
Ophthalmology. Atlanta, GA, October 2, 1985.
22. Cibis PA, Becker B, Okun E, et al. The use of liquid silicone in
retinal detachment surgery. Arch Ophthalmol 1962;68:590.
23. Cibis PA. Vitreous transfer and silicone injections. Trans Am
Acad Ophthalmol Otolaryngol 1964;68:983.
24. Okun E. Intravitreal surgery utilizing liquid silicone: a long-
term followup. Trans Pac Coast Otolaryngol Ophthalmol Soc
1968;49:141.
25. Okun E, Arribas NP. Therapy of retinal detachment compli-
cated by massive preretinal fibroplasia (long-term followup
of patients treated with intravitreal liquid silicone). In: New
Orleans Academy of Ophthalmology Symposium on Retina and
Reena Surgery. St. Louis: Mosby, 1969:278–293.
26. Scott JD. The treatment of massive vitreous retraction by the
separation of preretinal membranes using liquid silicone. Mod
Probl Ophthalmol 1975;15:285.
27. Scott JD. A rationale for the use of liquid silicone. Trans
Ophthalmol Soc UK 1977;97:235.
28. Labelle P, Okun E. Ocular tolerance to liquid silicone: an
experimental study. Can J Ophthalmol 1972;7:199.
29. Ober RR, Blanks JC, Ogden TE. Experimental retinal tolerance
to liquid silicone. Retina 1983;3:77.
30. Ober RR, Ryan SJ, Minckler DS, et al. Ocular tolerance to liq-
uid silicone: an experimental study. Invest Ophthalmol Vis Sci
1980;19(Suppl.):47.
31. Meredith TA, Lindsey DT, Edelhauser HF, et al. Electroretino-
graphic studies following vitrectomy and intraocular oil injec-
tion. Br J Ophthalmol 1985;69:254.
32. Momirov D, Van Lith GHM, Zivojnovic R. Electroretinogram
and electro-oculograms of eyes with intravitreously injected
silicone oil. Ophthalmologica 1983;186:183.
33. Freeman KD, Gregor ZJ. Electrophysiological responses after
vitrectomy and intraocular tamponade. Trans Ophthalmol Soc
UK 1985;104:129.
156 SECTION IV ■ Disease-Sp ecific Managem ent
34. Foerster M, Esser J, Laqua H. Silicone oil and its influence on
electrophysiologic findings. Am J Ophthalmol 1985;99:201.
35. Abrams GW, Azen SP, McCuen BW II et al. Vitrectomy with
silicone oil or long acting gas in eyes with severe PVR: results
of additional and long-term follow-up. Silicone Study report
11. Arch Ophthalmol 1997;115(3):335–344.
36. Diddie KR, Azen SP, Freeman HM, et al. Anterior proliferative
vitreoretinopathy in the silicone study. Silicone Study Report
Number 10. Ophthalmology 1996;103(7):1092–1099.
37. Vitrectomy with silicone oil or pleurofluorocarbon gas in eyes
wit severe PVR: results of a randomized clinical trial. Silicone
report 2. Arch Ophthal 1992;110(6):780–792.
38. Ando F. Intraocular hypertension resulting from pupillary
block by silicone oil. Am J Ophthalmol 1985;99:87.
39. Charles S. Vitrectomy for retinal detachment. Trans Ophthalmol
Soc UK 1980;100(4):542.
40. Ando F, Kondo J. A plastic tack for the treatment of reti-
nal detachment with giant tear (letter). Am J Ophthalmol
1983;95:260.
41. De Juan E, Hickingbotham D, Machemer R. Retinal tacks. Am
J Ophthalmol 1985;99:272.
42. Aaberg TM. Presentation at Duke Advanced Vitrectomy Course.
Durham, NC, April 1985.
43. Aaberg TM. Presentation at American Retina Society Meeting.
Cleveland, OH, November 1985.
44. Campochiaro PA, Koden IH, Vidaurri-Leal JS, et al. Cryother-
apy enhances intravitreal dispersion of viable retinal pigment
epithelial cells. Arch Ophthalmol 1984;103:434.
45. Glaser BM, Vidaurri-Leal J, Michels RG, et al. Cryotherapy
during surgery for giant retinal tears enhances dispersion of
viable retinal pigment epithelial cells. Ophthalmology 1993
Apr;100(4):466–470.
46. Hida T, Sheta SM, Proia AD, et al. Experimental transvitreal
cyanoacrylate retinopexy in a primate model. Am J Ophthalmol
1987 Jun 15;103(6):782–789.
47. McCuen BW II, Hida T, Sheta SM. Transvitreal cyanoacrylate
retinopexy in the management of complicated retinal detach-
ment. Am J Ophthalmol 1987 Aug 15;104(2):127–132.
48. Machemer R, Laqua H. Alogical approach to the treatment of mas-
sive periretinal proliferation. Ophthalmology 1978;85(6):584.
49. Machemer R. Massive periretinal proliferation: a logical approach
to therapy. Trans Am Ophthalmol Soc 1977;75:556.
50. Sternberg P, Machemer R. Results of conventional vitreous
surgery for proliferative vitreoretinopathy. Am J Ophthalmol
1985;100:141.
51. Charles S. Methodology and research on proliferative vitreo-
retinopathy. Presented at Retina Sodety Meeting. Cleveland,
OH, November 1985.
52. Charles S. Vitreous surgery for proliferative vitreoretinopathy.
Presented at Vitreous Society Meeting. Orlando, FL, October
1985.
53. Lewis H, Aaberg TM. Causes of failure after repeat vitrectomy
for recurrent proliferative vitreoretinopathy. Am J Ophthalmol
1991;15;111(1):15–19.
54. Machemer R, Sugita G, Tano Y. Treatment of intraocular pro-
liferations with intravitreal steroids. Trans Am Ophthalmol Soc
1979;77:171.
55. Tano Y, Chandler D, Machemer R. Treatment of intraocular
proliferation with intravitreal injection of triamcinolone ace-
tonide. Am J Ophthalmol 1980;90:810.
56. Chandler DB, Rozakis G, De Juan E, et al. The effect of triam-
cinolone acetonide on a refined experimental model of prolif-
erative vitreoretinopathy. Am J Ophthalmol 1985;99:686.
57. Stern WH, Lewis GP, Erickson PA, et al. Fluorouracil therapy
of proliferative vitreoretinopathy after vitrectomy. Am J Oph-
thalmol 1983;96:32.
58. Blumenkranz MS, Ophir A, Claflin AL, et al. Fluorouracil for
the treatment of massive periretinal proliferation. Am J Oph-
thalmol 1982;94:458.
59. Santana M, Wiedemann P, Kinmani M, et al. Daunomycin in
the treatment of experimental proliferative vitreoretinopathy-
retinal toxicity of intravitreal daunomycin in the rabbit. Graefes
Arch Clin Exp Ophthalmol 1984;221:210.
60. Wiedemann P, Sorgente N, Bekhor C, et al. Daunomycin in
the treatment of experimental proliferative vitreoretinopathy-
effective doses in vitro and in vivo. Invest Ophthalmol Vis Sci
1985;26:719.
61. Wiedemann P, Sorgente N, Kirmani M, et al. Daunorubicin in
the treatment of experimental MPP-effective doses in vitro and
in vivo. Invest Ophthalmol Vis Sci 1983;24(Suppl.):241.
C H AP TER
17
EPIMACULAR MEMBRANES
AND VITREOMACULAR
TRACTION SYNDROME
Membranes on the macular surface can result
from several pathogenic mechanisms with the
common theme of tissue damage and subse-
quent repair (1–5). Epimacular membranes
(EMMs) are hypocellular, largely collagen structures. EMMs
are also called macular puckers, cellophane maculopathy,
surface wrinkling retinopathy, and premacular fibrosis. Each
of these names has certain deficiencies, hence the currently
most widely accepted name, EMMs.
PATHOGENESIS
The so-called idiopathic type of EMM is caused by glial
migration and proliferation from a defect in the internal
limiting membrane (ILM) created by a posterior vitreous
separation (6). Retinal breaks, retinopexy, photocoagulation,
inflammation, and vascular disease (7) can lead to glial pro-
liferation (8–12) on the retinal surface. Retinal pigment epi-
thelial cells (13,14) can migrate through a retinal break and
proliferate on the retinal surface just as they do in proliferative
vitreoretinopathy (PVR). EMMs can be thought of as localized
glial or retinal pigment epithelium (RPE)–induced PVR.
ETIOLOGY OF VISUAL LOSS
Hypocellular contraction of the EMM causes nonrhegmatog-
enous elevation of the macula thought by the authors to be
responsible for a major fraction of the associated visual loss.
Fluid under the macula is universally seen on optical coher-
ence tomography (OCT). Reversible macular edema secondary
to macular separation from the fluid pumping mechanism of
the RPE contributes to visual loss as well. Although it is widely
stated that traction on the ILM can produce macular edema, it is
unclear what the mechanism would be and the concept remains
unproven. Although the terms “macular pucker” and “surface
wrinkling retinopathy” emphasize retinal distortion, some
patients have marked improvement in postoperative vision in
spite of persistent retinal distortion and metamorphopsia.
HISTORY
The typical EMM patient experiences a relatively rapid loss
of vision accompanied by metamorphopsia over a period
of several weeks, followed by relative stabilization of visual
function. In spite of this typical history, it is common prac-
tice for doctors to advise a patient with a recent history of
visual loss to, for example, the 20/50 level that he or she
should wait until the vision is reduced to 20/80 or worse
before considering surgery. In fact, the vision will usually
stabilize at a visual level at or near that noted on initial pre-
sentation. Because visual results are better with better pre-
operative vision and shorter duration, it is better practice to
make a decision on surgical intervention on the first visit.
CASE SELECTION
As with all surgical procedures, the decision to operate is a
multifactorial process based on symptoms, extent of visual
loss, visual needs, status of the other eye, age, duration,
medical status, and the presence of other ocular diseases.
There is no substitute for ethical, sound clinical judgment in
making the decision to operate.
The principal author’s visual acuity threshold for sur-
gery has moved from 20/200 to 20/40 in selected cases, as
the methodology has improved. A patient with preoperative
vision of 20/40 can and should be operated if the patient
is significantly symptomatic, is in good health, is relatively
young, and understands the issues. Specific visual acuity
157
158 SECTION IV ■ Disease-Sp ecific Managem ent
levels are less important than symptomatology and impact
on activities of daily living for recommendation of surgery
for EMMs. Duration is a relative rather than absolute crite-
rion because cases of 10 years’ duration have had significant
visual improvement following surgery. The visual improve-
ment in long-duration cases is presumably because the mini-
mal amount of subretinal fluid present in these cases leads
to minimal irreversible photoreceptor degeneration, just as
is the case in central serous retinopathy. Macular edema,
except in the vascular disease subgroup, is probably sec-
ondary to macular elevation, typically reversible and not a
contraindication to vitreoretinal surgery. Knowledge that the
patient had poor vision before the membrane occurred is an
absolute contraindication to surgery. The slow recovery of
vision after retinal reattachment surgery coupled with the
typical 1-month onset of EMM makes it difficult to make a
surgical decision in this situation. Patients with severe hered-
itary photoreceptor degeneration or a previous central reti-
nal artery occlusion frequently have wrinkling of the retinal
surface without an epiretinal membrane because of marked
decrease in retinal thickness. Surgery is contraindicated in
these situations (15–27).
SURGICAL SEQUENCE
AND TECHNIQUES
Vit re o m acular Tract io n Syn d ro m e
The posterior vitreous cortex (PVC) is rarely adherent
to typical EMMs. Some patients have macular elevation
secondary to hypocellular contraction of the PVC combined
with marked adherence of the vitreous to the macula (28–30).
This entity is known as vitreomacular traction syndrome.
Spectral domain OCT invariably demonstrates vitreomacu-
lar traction in these cases. When operating on these cases,
care must be taken to avoid tearing the fovea by imbrica-
tion of the vitreous into the port of the vitreous cutter. Fine
curved scissors can be used to delaminate the PVC from the
fovea prior to any removal of the vitreous (Fig. 17.1).
No n rh e g m at o g e n o us Pro life rat ive
Vit re o re t in o p at h y
Some patients have multiple star folds from PVR in addition
to an epimacular component. Removing these additional
epiretinal membranes is a stimulus for recurrent PVR and
is unnecessary unless they are causing macular elevation or
distortion.
Ne e d fo r Vit re ct o m y at t h e Tim e
o f Me m b ran e Pe e lin g
The principal author initially suggested the concept of
membrane peeling without vitreous removal but stopped
advocating this approach after several hundred cases
because the patients complain bitterly of floaters and
in-office fluid-gas exchange cannot be performed if the
patient develops a retinal detachment. If the vitreous has
been removed, a postoperative retinal detachment can be
managed by in-office, two-needle, fluid-gas exchange and
laser retinopexy. Anterior vitreous cortex removal is prob-
ably correlated with an increased incidence of posterior
subcapsular cataract probably related to fluid turbulence.
Fig u re 1 7 .1 ■ Curved scissors
are used to resect the PVC in vitre-
om acular traction syndrom e cases
to p revent avulsion of the fovea
caused by im brication of the taut
PVC into the cutter.
 CHAPTER 17 ■ Epim acular Mem branes and Vitreom acular Traction Syndrom e 159

Fig ure 1 7 .2 ■ Inside-out forceps

peeling is initiated at the epicenter,
which is the densest portion of the
EMM and the focal point of striae.
Pics and looking for an ed g e can
result in retinal d am ag e an d are
not required with this m ethod.

Avoiding the anterior vitreous cortex may reduce post- for surface grasping and must have precise alignment of the
operative posterior subcapsular cataract. blade tips. Eckardt developed effective end-opening forceps
that are preferred to diamond-coated forceps because the
diamonds are recessed from the tips of the forceps blades.
Ep ire t in al Me m b ran e Re m o val
Conformal forceps were then developed by the princi-
Machemer developed the concept of membrane peeling in pal author because Eckardt-type forceps have square cor-
1972 soon after his introduction of vitrectomy. Originally, ners that can tear the nerve fiber layer or cause bleeding.
peeling was performed with a bent needle. O’Malley sub- The tips of conformal forceps have a radius of curvature
sequently developed the concept of using a rounded, angu- which matches that of the retina. The authors use 25-gauge
lated instrument he called a pic to perform the peeling. The Alcon disposable, DSP ILM forceps (Fig. 17.3) for all cases.
principal author and the late Ron Michels popularized the The Alcon 25-gauge MVR pic is used by some surgeons if the
pic method. Bent needles and pics require the presence of membrane is very smooth and taut (so-called glassy), but
a visible outer margin of the EMM, frequently called an the authors rarely find this step necessary when using DSP
“edge,” unless a slit is made in the membrane. Searching
for an edge creates the risk of making a retinal break. The
principal author developed the concept of inside-out for-
ceps membrane peeling initially because of the difficulty of
finding an edge in certain cases. In contrast to Machemer’s
outside-in membrane peeling method, inside-out peeling
is initiated by surface grasping the EMM with end-opening
forceps (Fig. 17.2). Although the senior author formerly rec-
ommended making a slit in the apparent center of the EMM
using the microvitreoretinal (MVR) blade or a sharp-tipped
pic, this is not necessary with the Alcon 25-gauge ILM DSP
forceps. The center of the membrane can be identified by
noting the orientation of radial striae, the most elevated
retinal region, the most opaque region of the membrane,
and the relative movement of the membrane with respect
to the retina induced by lateral movement of the forceps or
MVR blade tip. End-opening forceps with diamond coating Fig u re 1 7 .3 ■ Alcon 25-g aug e ILM forcep s are id eal for
were developed by the senior author and Hans Grieshaber rem oving EMMs.
160 SECTION IV ■ Disease-Sp ecific Managem ent
forceps. Membrane peeling should be accomplished by
moving the forceps tangentially along the surface of the
retina in a circular fashion (Fig. 17.4). If the membrane
tears, it can be regrasped without removing any membrane
from the forceps because microteeth will penetrate several
layers of membrane and facilitate removal of the membrane
through the pars plana as well. The surgeon should always
observe the fovea during the peeling process rather than
focusing on the forceps in order to prevent tearing the fovea.
Areas of stronger adherence to the ILM can be detected by
noting fine fibers being lifted from the retinal surface dur-
ing the peeling process. If prominent striae are still noted
after peeling the epiretinal membrane, ILM peeling with the
25-gauge DSP forceps must be performed. Kampik intro-
duced the idea that ILM peeling reduces the incidence of
recurrent membranes, and the authors have validated this
observation. Scissors delamination with fine curved scis-
sors rather than peeling is utilized if strong adherence to
the fovea, vessels, or any region of the retina is noted during
the peeling process (Fig. 17.5). If there are marked folds,
the blunt, polished end of the vitrectomy instrument can be
used to gently push the retinal folds into better position, a
method referred to by the principal author as “burnishing.”
Moderate-sized peeled or delaminated membrane pieces
should be removed through the pars plana with the forceps.
If the membrane is very large or dense, it should be removed
with the vitrectomy probe.
Man ag e m e n t o f Re t in al Bre aks
No posterior retinal breaks occurred in the principal
author’s prospective, consecutive, single-surgeon series of
Fig ure 1 7 .4 ■ Inside-out forceps
m em b ran e p eeling is p erform ed
tan g ential to the retinal surface,
u sin g a circular m o tion wh ile
observing the fovea so that foveal
avulsion will not occur.
over 1,600 consecutive cases; all other series report a 5%
to 6% incidence of retinal breaks. The peripheral retina
should be inspected at the end of the case, and all retinal
breaks with elevated edges should be managed with fluid-
air exchange, air-gas exchange with 25% SF6, and laser
endophotocoagulation around the break(s). Laser alone can
be used if the edges are not elevated, but there is an increased
risk of immediate postoperative retinal detachment. Scleral
buckling is not required in these cases.
Man ag e m e n t o f Co e xist e n t Cat aract
Cataract surgery should be performed prior to vitreoretinal
surgery if a visually significant cataract is present. Although
this approach increases cost and subjects the patient to an
additional procedure, it enables a superior view at vitrec-
tomy and optimal refractive status. EMM surgery requires
excellent visualization and should not be attempted if
3+ or greater nuclear sclerosis or a significant posterior
subcapsular cataract is present. The fact that vitreous sur-
gery causes progression of preexisting nuclear sclerotic
cataract in a large number of cases should be taken into
account while considering when to remove the lens. Opti-
mal refractive outcomes are obtained when cataract sur-
gery incorporating all of the latest technology developed to
enable emmetropia is utilized. Although phaco-vit is now
recommended by the authors and many other surgeons for
many clinical scenarios for the management of vitreoretinal
disease combined with cataract or in anticipation of nuclear
sclerosis progression, it is difficult to obtain optimal refrac-
tive outcomes. It is often a better approach to perform vitre-
oretinal surgery first if the view is adequate and to have an
 CHAPTER 17 ■ Epim acular Mem branes and Vitreom acular Traction Syndrom e 161

Fig u re 1 7 . 5 ■ If t h e EMM
is extrem ely ad h eren t, scisso rs
delam ination with curved scissors
should b e used to avoid tearin g
the retina.

experienced cataract surgeon perform refractive phaco later Rh e g m at o g e n o us Re t in al De t ach m e n t

when nuclear sclerosis progression becomes significant.
The author’s series had a similar incidence of postoperative
retinal detachment to that reported by other authors (5%).
This complication is related to inadvertent force on the vitre-
VISUAL RESULTS ous base during vitreous removal, instrument introduction
forces, and incarceration of vitreous in the sclerotomies. This
About 40% of the patients in the senior author’s series
complication can be decreased by using high-quality cut-
had improved vision to 20/40 or better. Approximately
ters, low suction force (<80 mm Hg), highest possible cut-
56% improved to 20/80 or better. Over 85% of the patients
ting rates, and avoiding pulling the cutter away when cutting
improved two lines or greater in visual acuity. It is noted that
is activated. The retinal periphery should be inspected with
those with greater visual loss (1/200 to 5/200) preoperatively
the indirect ophthalmoscope at the end of the case and any
improve to approximately the 20/200 level postoperatively.
retinal breaks treated with endolaser, laser indirect ophthal-
Those in the 20/200 region preoperatively typically return to
moscope (LIO), or cryopexy and fluid-gas exchange with
the 20/40 to 20/50 level postoperatively. Those with 20/50
air or 25% SF6, if the edges are elevated. Alternatively, the
vision preoperatively usually return to the 20/20 to 20/25
contact wide-angle lens can be utilized to perform limited
level. Visual results have improved significantly since ILM
peripheral vitrectomy with simultaneous observation of the
peeling was combined with epiretinal membrane peeling in
retinal periphery for presence of retinal tears. If found, endo-
all cases.
photocoagulation can be performed.

Pro life rat ive Vit re o re t in o p at h y Re curre n ce

COMPLICATIONS
Re t in al Bre aks
No posterior retinal breaks were caused in the author’s series
of over 1,100 consecutive vitrectomies for EMM. All other
authors have reported an incidence of 5% to 7%. Although
the author’s series was not randomized to outside-in versus
inside-out peeling, this marked difference probably indicates
that inside-out forceps membrane peeling is a safer method
than using pics and outside-in peeling.
If PVR was successfully repaired prior to the development of
a membrane, it can be stimulated to recur by EMM surgery.
Usually, the PVR-induced redetachment can be successfully
repaired with repeat vitrectomy and silicone oil, but little
improvement in pre-EMM surgery vision is then obtained.
Cyst o id Macular Ed e m a
The author’s series noted a preoperative cystoid macular
edema (CME) incidence of 3.2%. This incidence was based
162 SECTION IV ■ Disease-Sp ecific Managem ent
on clinical observation, as routine fluorescein angiography
was not performed. The McDonald et al. series reports a 3%
incidence of CME.
En d o p h t h alm it is
The author’s series of over 1,400 cases had no cases of post-
operative endophthalmitis, while the literature reports an
average incidence of 0.7%. The author uses a one-piece sur-
gical drape that is folded over the lid margins and a micro-
scope drape in all cases. Subconjunctival Tobramycin and
Ancef are injected at the end of the case. High-quality infu-
sion fluid (Alcon BSS Plus) is used in all cases. Absorbable
sclerotomy sutures are not used.
Re curre n ce o f Ep im acular Pro life rat io n
Approximately 1.6% of treated EMM cases resulted in clini-
cally significant recurrent EMMs (31–34) in the author’s
series of over 1,100 cases. Successful reoperation can usu-
ally be performed with sustained visual improvement. The
recurrence rate emphasizes the reparative nature of the pro-
cess and the damage to the retinal surface associated with
membrane peeling. All patients probably have some repair of
the retinal surface after membrane peeling. The criterion for
defining a recurrence has not been established in the litera-
ture. In some patients, the folds disappear completely and a
recurrence can easily be determined if folds recur. In other
patients, the recurrent membrane causes macular elevation
with minimal striae and the membrane is more difficult to
visualize. Decreased visual function is usually the best indi-
cator of clinically significant recurrences if the patient had
initially experienced gradual visual improvement. Visually
significant cataract must be ruled out before the visual loss
can be attributed to the macula.
Re t in al Wh it e n in g
Immediate postoperative retinal whitening occurs at the
removal site in a significant percentage of cases. This disap-
pears spontaneously in several days and does not seem to
affect the visual outcome. It is probable that this phenom-
enon represents ganglion cell axoplasmic flow disruption.
Michels noted that retinal whitening is present preopera-
tively in a significant number of cases presumably because of
traction on the nerve fiber layer exerted through the ILM.
Cat aract
Posterior subcapsular cataracts after vitrectomy are largely
avoidable. Posterior subcapsular cataracts can be caused by
using low-quality infusion fluids such as lactated Ringer’s
solution rather than glutathione bicarbonate Ringer’s solu-
tion (Alcon BSS Plus).
The observed increase in the incidence of nuclear
sclerosis associated with EMM surgery is thought by the
authors to result from increased oxygen tension in the former
vitreous cavity after vitrectomy. The incidence of progression
of nuclear sclerosis has been reported to range from 10% to
100%. This wide variation could be accounted for by many
factors, including follow-up period, definition of clinical sig-
nificance, postoperative refraction, operating time, intraop-
erative cataract surgery, light source UV and IR content, and
unknown factors.
References
1. Kampik A, Kenyon KR, Michels RG, et al. Epiretinal and
vitreous membranes: comparative study of 56 cases. Arch Oph-
thalmol 1981;99:1445–1454.
2. Pearlstone AD. The incidence of idiopathic preretinal macular
gliosis. Ann Ophthalmol 1985;17:378.
3. Scudder MJ, Eifrig DE. Spontaneous surface wrinkling retin-
opathy. Ann Ophthalmol 1982;94:44.
4. Sidd RJ, Fine SL, Owens SL, et al. Idiopathic preretinal gliosis.
Am J Ophthalmol 1982;94:44.
5. Wise GN. Clinical features of idiopathic preretinal macular
fibrosis. Am J Ophthalmol 1975;79:349.
6. Roth AM, Foos RY. Surface wrinkling retinopathy in eyes
enucleated at autopsy. Trans Am Acad Ophthalmol Otolaryngol
1971;75:1047.
7. Wise GN. Clinical features of idiopathic preretinal macular
fibrosis. Am J Ophthalmol 1975;79:349.
8. Kenyon KR, Michels RG. Ultrastructure of epiretinal membrane
removed by pars plana vitreoretinal surgery. Am J Ophthalmol
1977;83(6):815.
9. Clarkson SG, Green WR, Massof D. A histopathologic review of
168 cases of preretinal membrane. Am J Ophthalmol 1977;84:1.
10. Green WR, Kenyon KR, Michels RG, et al. Ultrastructure of
epiretinal membranes causing macular pucker following reti-
nal reattachment. Trans Ophthalmol Soc UK 1979;99:63.
11. Kampik A, Green WR, Michels RG, et al. Ultrastructural fea-
tures of idiopathic progressive epiretinal membrane removed
by vitreous surgery. Am J Ophthalmol 1981;90:797.
12. Michels RG. A clinical and histopathological study of epiretinal
membranes affecting the macula and removed by vitreous sur-
gery. Trans Am Ophthalmol Soc 1982;80:580.
13. Laqua H. Pigmented macular pucker. Am J Ophthalmol 1978;
86(1):56.
14. Lindsey PS, Michels RG, Luckenbach M, et al. Ultrastructure
of epiretinal membrane causing retinal starfold. Ophthalmology
1983;90:578.
15. Michels RG. Surgical treatment of macular pucker. In: Stirpe
M, Convers M, Blankenship G, et al., eds. Advances in vitreoreti-
nal surgery. Filia, Rome; 1991.
16. Shea M. The surgical management of macular pucker. Can J
Ophthalmol 1979;2:110.
17. Wilkinson CP. Recurrent macular pucker. Am J Ophthalmol
1979;88(6):1029.
18. Machemer R. A new concept for vitreous surgery: two instru-
ment techniques in pars plana vitrectomy. Arch Ophthalmol
1974;92:407–441.
19. Trese MT, Chandler DB, Machemer R. Macular Pucker: prognos-
tic criteria. Graefes Arch Clin Exp Ophthalmol 1983;221:12–15.
20. Margherio RR, Cox MS, Trese MT, et al. Removal of epimacular
membranes. Ophthalmology 1985;92:1075–1083.
 CHAPTER 17 ■ Epim acular Mem branes and Vitreom acular Traction Syndrom e
21. McDonald HR, Verre WP, Aaberg TM. Surgical management
of idiopathic epiretinal membranes Ophthalmology 1986;93:
978–983.
22. de Bustros S, Thompson JT, Michels RG, et al. Vitrectomy for
idiopathic epiretinal membranes causing macular pucker. Br J
Ophtalmol 1988;72(9):692–695.
23. Rice TA, de Bustros S, Michels RG, et al. Prognostic factors in
vitrectomy for epiretinal membranes of the macula. Ophthal-
mology 1986;93:602–610.
24. Charles S. General posterior segment techniques. In: Charles S,
ed. Vitreous microsurgery. Baltimore: Williams & Wilkins;
1987:98.
25. Margherio RR. Discussion of Michels RG: vitrectomy for macu-
lar pucker. Ophthalmology 1984;91:1387–1388.
26. Gass JDM. Steroscopic atlas of macular diseases, 4th ed., St Louis:
Mosby; 1997:288–299, 938, 940, 944.
27. Barr CC, Michels RG. Idiopathic nonvascularized epiretinal
membranes in young patients:report of six cases. Ann Ophthal-
mol 1982;14:335–341.
163
28. Smiddy WE, Michels RG, Glaser BM, et al. Vitrectomy for
macular traction caused by incomplete vitreous separation.
Arch Ophthalmol 1988;106:624–628.
29. Melberg NS, Williams DF, Balles MW, et al. Vitrectomy for
vitreomacular traction syndrome with macular detachment.
Retina 1995;15:192–197.
30. Bellhorn MB, Friedman AH, Wise GN, et al. Ultrastructural
and clinicopathologic correlation of idiopathic preretinal mac-
ular fibrosis. Am J Ophthalmol 1975;79:366–373.
31. Michels RG. Vitreous surgery for macular pucker. Am J Oph-
thalmol 1981;92:628.
32. Michels RG. Surgery of macular pucker. In: Fine SL, Owens
SL, eds. Management of retinal vascular and macular disorders.
Baltimore: Williams & Wilkins; 1983:120–130.
33. Michels RG. Vitrectomy for macular pucker. Ophthalmology
1984;91:1384.
34. Michels RG. Macular pucker. In: Fraunfelder FT, Roy FH, eds.
Current ocular therapy, 2nd ed. Philadelphia: W.B. Saunders;
1984:440–442.
C H AP TER
18
MACULAR HOLES
Until 1991, it was thought that macular holes
were untreatable and their pathogenesis was
unknown. In that year, Kelly and Wendel devel-
oped the concept of using vitrectomy and fluid-
gas exchange to treat these patients (1). Initially there was
great skepticism about the treatment, but the facts prevailed.
The initial goal was to “seal” the hole much as is done for
rhegmatogenous retinal detachment and eliminate the cuff
of subretinal fluid that surrounds the hole. When it was
noted that many holes actually disappeared after surgery, and
patients obtained near-normal vision, the skeptics again did
not believe it. Fortunately, it is now accepted that complete
disappearance of the hole is the usual outcome after surgery.
PATHOGENESIS
It was widely believed that a posterior vitreous detachment
(PVD) pulled a full-thickness piece of tissue out of the mac-
ula, probably as a result of or during a saccade. Electron
microscopy of the so-called operculum, which occurs in
these cases, has shown that few photoreceptors are present
(2,3). Gass published a hypothesis for the pathogenesis (4,5)
of macular holes that has remained viable with minimal
modifications since the introduction of high-resolution opti-
cal coherence tomography (OCT). The concept is that radial
vitreous fibers remaining on the perimacular surface after
apparent posterior vitreous separation contract and slowly
tear the macula in a circumferential fashion. The Mueller
cells may play a role as well according to Gass. Many observ-
ers have noted that vitreous is attached to the optic nerve
after an apparent posterior vitreous separation characterized
by the presence of a prepapillary (Weiss) ring. Because of
these observations, the event might better be termed “delam-
ination of the posterior vitreous cortex.” The senior author
terms the cuff of fluid “the traction cuff.” Sjaarda has shown
using scanning laser ophthalmoscope (SLO) microperimetry
that the actual extent of scotoma extends far beyond the mar-
gin of the cuff (6). The vitreous that is attached to the inner
margin of the macular hole is rarely contiguous with the
vitreous attached to the midperipheral retina. Terminology
164
represents a serious problem in discussing these cases with
vitreoretinal surgeons.
Surgical success should be defined as clinical disap-
pearance of the hole, reconstitution of the foveal anatomy
on OCT, and marked improvement in vision.
INDICATIONS FOR MACULAR
HOLE SURGERY
Freeman et al. have shown that smaller, more recent holes
have the best prognosis. Size is much more important than
duration with respect to closure rates; duration plays no
role if size is controlled for. Duration probably plays a role
in visual success in successful cases. Most surgeons do
not suggest repairing holes secondary to chronic macular
edema from diabetic retinopathy, venous occlusive disease,
cystoid macular edema after cataract surgery, or secondary
to uveitis. Traumatic macular holes are a complex decision-
making process because of the high likelihood of associ-
ated photoreceptor, retinal pigment epithelium (RPE), and
optic nerve damage and because many will spontaneously
close within 1 to 2 months. If there is good evidence that
the macular hole is the only significant damage, these cases
can be considered for surgery after a reasonable period of
observation.
OCULAR COHERENCE TOMOGRAPHY
OCT is invaluable in the evaluation of macular hole patients
both preoperatively and after surgery. It can be difficult to
distinguish partial-thickness holes from full-thickness holes
clinically. Some holes are very small at the internal limiting
membrane (ILM) level but much larger at the level of the outer
retina; others have the opposite configuration. Vitreomacular
traction and epimacular membranes can be seen with care-
ful OCT examination. Some holes have no cuff, are ellipti-
cal, and are caused by typical epimacular membranes rather
than vitreous. These holes may be thought of as secondary
holes and round holes with a cuff above as primary or

classic holes. The Gass classification system using Stage I, etc.
is no longer used by the authors now that OCT is available.
Macular holes are now classified as partial thickness or full
thickness. Diameter is a major factor in driving outcomes
and can be accurately measured with OCT. Whether there
is a PVD or not is irrelevant clinically as an attempt must be
made to create a PVD in all cases and ILM peeling removes
any adherent posterior vitreous cortex. Some patients have
apparent closure after surgery but modest visual recovery;
some of these cases have edema or subretinal fluid as shown
by Kaiser, while others have a persistent defect in the outer
retina or disruption on the outer segment layer of the fovea
on spectral domain OCT.
MACULAR HOLE SURGERY
Macular holes were untreatable until Neil Kelly devel-
oped the concept of core vitrectomy followed by fluid-gas
exchange using an isoexpansive mixture of air and SF6 or
C3F8 gas. His goal was to reattach the cuff of subretinal fluid
around a full-thickness macular hole and therefore eliminate
the relative scotoma that surrounds the absolute scotoma.
Serendipitously he soon discovered that the macular hole
often closed with remarkable improvement in vision. Many
leading surgeons initially discounted his discovery, but for-
tunately it has become the standard of care. The prevailing
thought at the time was that substantial neural tissue was
avulsed by a PVD; an operculum was often seen, validating
this notion. Examination of surgically removed operculums
using the electron microscope revealed very little neural
tissue; most of the operculum proved to be glial tissue,
explaining why substantial visual improvement was pos-
sible. Subsequently, OCT demonstrated restoration of near-
normal or normal foveal anatomy in successfully operated
cases. Clearly, “closure” of macular hole is quite different
than the term “closure” when used in the context of retinal
detachment repair.
Much emphasis is placed on the substantial work of
Don Gass, which involved clinical observation of the evolu-
tion of macular holes, theoretical considerations concerning
the pathogenesis, and a classification system. Although all
surgeons agree that the posterior vitreous cortex is somehow
involved and the elasticity of the ILM plays a role, the patho-
genesis of macular holes remains unknown. Presurgical clas-
sification, even using spectral domain OCT, is incapable of
reliably determining if residual vitreous cortex is adherent to
the retinal surface rendering the classification system virtu-
ally useless. Macular holes are three times as common in
females as they are in males, but there is no explanation for
this interesting observation.
OCT is essential to determine if the hole is partial thick-
ness or full thickness; not infrequently clinical examination
is inadequate to detect a very small, full-thickness hole at
the base of a large-diameter partial-thickness hole. Occa-
sionally, a very small, inner layer, full-thickness hole will
CHAPTER 18 ■ Macular Holes 165
overlie a large, full-thickness outer layer hole producing an
appearance suggesting a macular cyst. Spectral domain OCT
is essential in the evaluation of macular disease. Time domain
OCT is no longer adequate. Charteris has shown that 30%
to 50% of partial-thickness outer macular holes (Stage I on
Gass’ Classification) spontaneously close if observed over
the long term. A multicenter clinical trial done before the
availability of OCT or ILM peeling did not demonstrate a
benefit of operating on a partial-thickness hole to prevent
progression to a full-thickness hole. Neither ultrasound nor
OCT can predict whether a partial-thickness hole will prog-
ress to become a full-thickness hole, and status of the other
eye is not helpful because holes are bilateral less than 10%
of the time.
Size of the macular hole is the only preoperative factor
that has been shown to drive surgical closure rates; duration
is not a determining factor if size is controlled for. Although
duration has an influence on visual outcomes, assessment of
the subhole RPE using spectral domain OCT and confocal
autofluorescence will probably prove to be more effective in
predicting visual outcomes in longer duration holes. Some
studies have shown reasonable visual outcomes after operat-
ing macular holes of relatively long duration.
Macular holes originating from macular cysts due to
chronic macular edema typically have relatively poor visual
outcomes because of macular ischemia related to underly-
ing diabetic retinopathy or retinal vein occlusion. Similarly,
macular holes arising from chronic inflammatory macular
edema have a relatively poor visual prognosis.
Traumatic macular holes can spontaneously close in the
first 4 to 6 weeks suggesting that a period of observation
should precede the determination of operability. An affer-
ent pupillary defect (APD) should be considered a relative
contraindication to surgical repair because it indicates asso-
ciated optic nerve damage. As mentioned above, assessment
of the subhole RPE using spectral domain OCT and confocal
autofluorescence will probably prove to be more effective in
predicting visual outcomes. Presence of a choroidal rupture
in the papillomacular bundle is a relative contraindication to
repair of a traumatic macular hole; often these patients will
have an APD.
The authors’ management of partial-thickness holes
has changed over the years. The senior author began
operating lamellar holes after Arthur Willis introduced
the concept of macular hole prevention surgery. Macular
hole prevention surgery was no longer performed after
the results of the Macular Hole Study and the Chart-
eris paper were published. The authors’ current practice
is to determine the need for surgery on partial-thickness
holes based on symptoms, particularly metamorphopsias
from the epiretinal membranes (ERMs) that commonly
surround lamellar macular holes. The goal is to improve
visual function, not prevent a full-thickness hole because
progression to a full-thickness hole cannot be predicted.
The authors have determined in recent years that surface
tension management using air or SF6 combined with ILM
166 SECTION IV ■ Disease-Sp ecific Managem ent
peeling is required to restore normal or near-normal foveal
anatomy and improve or eliminate symptoms when operat-
ing on symptomatic partial-thickness holes.
HOLE CLOSURE MECHANISMS
The conventional explanation of the role of surgical steps in
surgical hole closure is incomplete and, to an extent, prob-
ably incorrect. It is often stated that the role of PVD creation
is to eliminate vitreous traction on the macula, but, in fact,
the posterior vitreous is usually attached to the optic nerve
head, nasal, and midperipheral retina but not attached to
the macula or retina within the temporal arcades. Care-
ful PVD creation reduces the likelihood of inferior retinal
breaks caused by interaction of the bubble with residual
vitreous similar to what may occur with pneumatic retin-
opexy. The role of core vitrectomy is to enable the exchange
of vitreous for a large gas bubble and to ensure that a thin
layer of vitreous does not prevent contact of the gas bubble
with the hole.
Virtually all surgeons agree that ILM peeling improves
closure rates, but why this is true has a complex answer. ILM
peeling ensures removal of tangential traction due to resid-
ual vitreous on the retinal surface, which, although rare, can
occur even with an apparent PVD as evidenced by a Weiss
ring. In addition, ILM peeling guarantees successful removal
of ERMs that are occasionally present. ILM peeling increases
retinal elasticity by over 50%, which enables lateral surface
tension forces to close the hole as soon as the bubble comes
into contact with the hole. It is also likely that ILM peel-
ing also initiates mechanical signaling to the astrocytes to
heal the hole margins days after it is closed by lateral surface
tension.
Most surgeons use the term “tamponade” when describ-
ing the mechanism of action of air, gas, and silicone oil bub-
bles. To some, tamponade means to “press,” but tamponade
is from the French and means to “seal.” It is more accurate
and descriptive to use the term “surface tension manage-
ment” or “interfacial tension management” to describe the
function of the bubble. It is obvious that a bubble elimi-
nates transhole flow, which is the reason it is used in retinal
detachment repair. An additional and crucial function of the
bubble as Reppucci has pointed out is the lateral surface ten-
sion effect. Surface tension forces act along an interface of
two immiscible substances effectively pulling the surface
inward. Surface tension makes a droplet nearly spherical as
it falls away from a faucet and a soap bubble spherical. Rep-
pucci describes the bubble as bridging the hole, but the key
concept is that the bubble pulls the hole margins together
immediately after surgery as soon as the patient is positioned
with bubble in contact with the hole. Yet another function
of the bubble, as the senior author has pointed out, is the
prevention of transretinal flow (uveal-scleral outflow). Tor-
nambe has introduced the hydration hypothesis making
a note of the marked edema surrounding the hole as seen
on OCT; edema-mediated retinal thickening coupled with
the elasticity of the ILM causes eversion of the hole mar-
gins. Elimination of the edema facilitates restoration of near-
normal foveal anatomy, resolution of the edge eversion, and
approximation of the hole margins. The bubble dries out the
retinal surface, which probably signals the astrocytes to heal
the hole days after approximation of hole margins via the
lateral surface tension effect.
POSTERIOR VITREOUS DETACHMENT
CREATION METHODS
Many surgeons use a lateral (tangential) motion of the extru-
sion cannula or vitreous cutter to create a PVD. This method
creates shear force at the vitreous base, potentially leading to
iatrogenic retinal breaks. A better method is to position the
vitreous cutter at the nasal, superior and inferior disk margins
with the port oriented away from the center of the disk and
pull anteriorly (toward the cornea) using the vacuum-only
mode (Fig. 18.1). The anterior-pull, disk-margin method
safely, reliably, and quickly produces a PVD using 25-gauge
cutters. It is a misconception that higher flow rates help pro-
duce PVDs or that 25-gauge cutters cannot produce a PVD;
it is all about technique.
INTERNAL LIMITING MEMBRANE
PEELING
Unfortunately, many surgeons equate ILM peeling with
indocyanine green (ICG) staining, staining with alterna-
tive dyes, or particulate marking with triamcinolone. ICG
has been shown to be toxic to the retina and RPE and may
also be associated with increased intraoperative phototox-
icity (7). Many papers using ICG for ILM peeling report
the typical 90% closure rate but unacceptable visual out-
comes. One factor driving dangerous and unnecessary
ICG use is using noncontact optical systems such as the
BIOM and EIBOS, which decrease both axial resolution
and lateral resolution, making it more difficult to visual-
ize the ILM. There is anecdotal evidence for triamcino-
lone reducing closure rates and particles being trapped
within the hole or in the subretinal space. The authors use
Alcon 25-gauge DSP ILM end-grasping forceps (Fig. 18.2)
but never use pics, MVR blades, or membrane scrapers.
Pics, MVR blades, and membrane scrapers often result in
unnecessary retinal surface damage as attempts are made
to find or construct an “edge.” Forceps are still required to
actually peel the ILM when pics, scrapers, or MVR blades
are used to construct an edge, making this a two-step
approach. An additional factor driving ICG use is using
inadequate forceps incapable of true end grasping. For-
ceps used for ILM peeling should be end grasping and
25 or 23 gauge and should not be asymmetric or built
with a gripping surface extending more than 120 mm
 CHAPTER 18 ■ Macular Holes 167

Fig ure 1 8 .1 ■ After core vitrectom y, a PVD is created using suction-only m ode on the vitreous
cutter, p ulling anteriorly (not tang entially) over the op tic nerve. Peeling sh ould extend to
the m id p erip hery, and the p erip heral retina should b e exam ined for p eeling-ind uced retinal
breaks.

Fig ure 1 8 .2 ■ Conform al forcep s, 20-gauge DSP forceps, or 25-gauge DSP forceps (preferred)
are used to peel residual vitreous, epim acular m em brane, and ILM. ILM is peeled 360 degrees;
pics, m em b rane scrapers, and ICG staining are not required with this technique.
168 SECTION IV ■ Disease-Sp ecific Managem ent
Fig ure 1 8 .3 ■ Com plete fluid-air exchange is followed b y air-gas exchang e.
along the blades. Reusable forceps rapidly lose the ability
to grip at the leading edge of the gripping surface as the
surface is eroded and the blades are warped outward by
use, cleaning, and sterilization. Reuse of disposable for-
ceps rapidly degrades the ability to grip at the leading
edge as well.
SURFACE TENSION MANAGEMENT
The principal author used C3F8 for many years thinking
incorrectly that glial cells pulled the hole together and longer
term surface tension management would improve outcomes
but converted to SF6 in 2006 after learning of Reppucci’s
surface tension concept and observing that holes that were
not closed on the first postoperative day rarely, if ever, closed
(Fig. 18.3). Proof that glial cells do not pull the hole together
follows from the observations that normal or near-normal
foveal anatomy without a glial scar is the typical postopera-
tive OCT appearance (Fig. 18.4).
Positioning is controversial, in part because of seman-
tics and also from inappropriate pandering to the patients
suggesting that positioning is not necessary with “their”
technique. If patients remained supine continuously after
surgery, the hole would never close because surface ten-
sion from bubble contact is required to pull the hole mar-
gins together. It is unknown how long it takes for enough
adherence to occur that exposure to fluid would not cause
the hole to reopen, but it is probably 1 to 3 days. Fur-
ther positioning is necessary in phakic eyes to prevent gas
cataract.
Some surgeons advocate silicone oil for patients who
cannot or will not position after surgery. This is a seriously
flawed concept because silicone oil has roughly one third the
surface tension of a gas-fluid (actually gas-retina) interface.
A silicone oil bubble to a significant extent conforms to the
hole and RPE in the base of the hole, reducing the likelihood
of closure.
Vitrectomy produces accelerated nuclear sclerosis
progression rates in approximately 90% of cases, although
it does not produce cataract in cases with clear lenses
preoperatively, rarely the case with macular hole patients.
As Nancy Holekamp has shown, nuclear sclerosis progres-
sion is most likely due to a permanent increase in oxy-
gen tension of 7 to 12 mm Hg secondary to removal of
ascorbate. Ascorbate, a powerful antioxidant, is actively
secreted into the vitreous cavity, producing nine times the
serum level.
Combining cataract surgery with vitrectomy for macu-
lar holes or even removing lenses with minimal or no cata-
ract is advocated by some surgeons, since, in their opinion,
Fig ure 1 8 .4 ■ OCT after m acular hole surgery dem onstrates
near-norm al foveal anatom y with no glial scar.

cataracts will progress after vitrectomy. The authors do not
recommend phaco-vitrectomy technique for macular holes
because it produces less precise refractive outcomes and
increases the likelihood of anterior intraocular lens (IOL)
subluxation postoperatively. Gas bubbles also increase
the likelihood of iris capsule adhesions (synechia) during
the postoperative course. As described in Chapter 8, elective
macular surgery is best performed independently of phaco
surgery. If a cataract precludes adequate visualization for
ILM peeling, an experienced cataract surgeon can per-
form phacoemulsification with IOL implantation weeks
prior to vitrectomy. IOL calculations in macular disease
are best done with the IOL Master instead of A-scan ultra-
sonography, since the former calculates axial length to the
RPE and is unaffected by macular holes or ERMs as the
latter is.
ARCUATE RETINOTOMY
Aneesh Nikhra suggested that a relaxing incision in the
retina could be used to repair large, failed macular holes,
and the principal author developed the arcuate retinotomy
procedure. This technique is never used in primary surgery
cases. Alcon disposable 25-gauge curved DSP scissors were
used to make an arcuate retinotomy temporal to the macula.
The cut was made by separating the nerve fibers with special
care to avoid transecting nerve fibers. The arc length of the
retinotomy was 100 degrees centered around the temporal
horizontal meridian (Fig. 18.5). The arcuate retinotomy was
made approximately 250 mm from the edge of the macu-
lar hole. Internal drainage of subretinal fluid was performed
repeatedly to dry the macular hole. Drying the submacu-
lar RPE was necessary to allow surface tension to retain the
bridge of temporal retina in the appropriate position after
it was moved inward with the 25-gauge soft-tip cannula.
Repositioning of the temporal retinal bridge resulted in a
vertically oval macular hole and widening of the retinotomy
at the end of surgery. The horizontal diameter of the hole
markedly decreases at this point. Air-gas exchange with C3F8
or SF6 was the final step. In the authors’ series, which has
been submitted for publication, they achieved an 83% mac-
ular hole closure rate in these patients who otherwise would
be deemed inoperable.
Fig ure 1 8 .5 ■ OCT after arcuate retinotom y for large m acular
hole showing retinotom y site.
CHAPTER 18 ■ Macular Holes 169
POSTOPERATIVE POSITIONING
Most surgeons recommend full-time or near full-time face-
down positioning for the duration of a reasonably sized
bubble (8). The authors believe in full-time, face-down
positioning for 1 week. The patients are allowed to walk,
work, read, watch TVs placed on the floor, ride in cars (not
drive), and use the treadmill or stationary bicycle. Although
the patients often ask, a head-down position does not inter-
fere with eating, showering, reading, or using the restroom.
Face-down driving has consequences that extend beyond
vitreoretinal surgery.
BIOLOGIC MODIFIERS
Fibrin, serum, autologous platelet concentrate, thrombin,
whole blood, cryo precipitates, and transforming growth
factor beta (bovine, recombinant, and autologous) have
been used to accelerate glial proliferation in the hole (9–11).
Sterile endophthalmitis, uveitis, inflammatory prolifera-
tive vitreoretinopathy, and bacterial endophthalmitis have
been reported after use of these agents. The authors and
many other surgeons no longer use these agents because
of unproven efficacy and the risk factors noted previously.
Freeman et al. showed no effect of serum in a controlled
clinical trial.
RESULTS
Results have been reported as high as 100% for small series
with cases selected for small, short-duration holes. The
authors believe that 90% is a reproducible success rate with
SF6; 1 week of face-down positioning; and cortex, ERM, and
ILM peeling (12).
References
1. Kelly NE, Wendel RT. Vitreous surgery for idiopathic macu-
lar holes. Results of a pilot study. Arch Ophthalmol 1991;109:
654–659.
2. Madreperla SA, McCuen BW II, Hickinbotham D, et al. Clini-
copathologic correlation of surgically removed macular hole
opercula. Am J Ophthalmol 1995;120:197–207.
3. Ezra E, Munro PM, Charteris DG, et al. Macular hole opercula.
Ultrastructural features and clinicopathological correlation.
Arch Ophthalmol 1997;115(11):1381–1387.
4. Gass JD. Idiopathic senile macular hole. Its early stages and
pathogenesis. Arch Ophthalmol 1991;106:654–659.
5. Gass JD. Reappraisal of biomicroscopic classification of
stages of development of a macular hole. Am J Ophthalmol
1995;119:752–759.
6. Sjaarda RN, Frank DA, Glaser BM, et al. Assessment of vision
in idiopathic macular holes with macular microperimetry using
the scanning laser ophthalmoscope. Ophthalmology 1993;
100(10):1513–1518.
7. Gandorfer A, Aritoglou C, Kampik A. Toxicity of Indocyanine
Green in vitreoretinal surgery. In: Meyer CH, ed. Vital dyes
170 SECTION IV ■ Disease-Sp ecific Managem ent
in vitreoretinal surgery. Developments in ophthalmology. Karger:
Basel, 2008:69–81.
8. Tornambe PE, Poliner LS, Grote K. Macular hole surgery
without face down positioning. A pilot study. Retina 1997;17:
179–185.
9. Glaser BM, Michels RG, Kupperman BD, et al. Transforming
growth factor-beta 2 for the treatment of full thickness mac-
ular holes. A prospective randomized study. Ophthalmology
1992;99:1162–1173.
10. Ligget PE, Skolik SA, Horio B, et al. Human autologous serum
for the treatment of full thickness macular holes: preliminary
study. Ophthalmology 1995;102:1071–1076.
11. Thompson JT, Smiddy WE, Williams GA, et al. Comparison of
recombinant transforming growth factor beta-2 and placebo
as an adjuvant agent for macular hole surgery. Ophthalmology
1998;105:700–706.
12. Thompson JT, Sjaarda RN, Lansing MB. The results of vitreous
surgery for chronic macular hole. Retina 1997;17:493–501.
C H AP TER
19
DIABETIC RETINOPATHY
MEDICAL ISSUES
Diabetes is increasing in prevalence in the American population
for many reasons (1,2). Diabetics are living longer because of
dialysis, kidney and pancreas transplants, improved cardio-
vascular management, and pharmacologic advances. The
availability of high-calorie, high-carbohydrate, high-fat,
and high-sugar foods; fast food restaurants; cultural atti-
tudes toward eating; poor role models; large serving sizes;
and many other factors contribute to the problem (3). Joslin
observed that in 1912, there was no diabetes in the Pima
Indian population; the incidence is now almost 70%. Fried-
man and others have attributed this to the “thrifty gene,”
which enabled survival of man with “occasional” eating
rather the three-square meals that many believe we should
eat. It is of interest that beef, pork, egg, fast food, and dairy
business interests support the nutrition community and
make very large political donations. If dieticians, teachers,
parents, doctors, nurses, ministers, and coaches eat improp-
erly, there are no positive role models. The diabetic, antiath-
erosclerosis, anticancer, weight-loss, longevity, feel-good,
look-good, and fitness diet are virtually the same. If the doc-
tor eats properly, exercises, and educates the family and staff
to do the same, a positive model is created for all. Everyone
who comes into contact with the patient and family should
support better health habits.
IMAGING
Fluorescein angiography was instrumental in understand-
ing the stages in the pathogenesis of diabetic retinopathy
and remains a valuable diagnostic tool to determine the
extent and location of capillary nonperfusion, presence of
neovascularization, and management of diabetic macular
edema. The confocal, laser scanning, digital imager pro-
vides 15 dB better signal/background (contrast and reso-
lution), eliminates the bright flash that increases patient
comfort, and has a faster frame rate when compared to
digital or film-based angiography systems. Better image
quality is achieved because confocal imaging increases the
image to background ratio via rejection of light scattered by
the cornea, lens, and vitreous. In addition, lasers are more
efficient in stimulating the fluorescein fluorophore than
filtered noncoherent light. The use of contact wide-angle
angiography lenses permits the diagnosis of preequatorial
retinal neovascularization that was previously difficult to
diagnose (Fig. 19.1).
Optical coherence tomography has not only allowed
more precise diagnosis of diabetic macular edema but
also permitted detailed observation of the vitreomacu-
lar interface and the presence of traction on the macular
area that can be potentially relieved by vitrectomy surgery
(Fig. 19.2).
SCREENING
Up to 70% of the known diabetic population in America
does not get annual eye exams. Many studies have utilized
various methods to screen this population as well as monitor
patients with no retinopathy or preclinically significant mac-
ular edema (pre-CSME) nonproliferative diabetic retinopa-
thy (NPDR) at their last exam. It appears that color images
are somewhat more sensitive than monochrome. Digital
cameras have decreased in cost while increasing in quality
and have almost replaced standard film and Polaroid images.
Nonmydriatic fundus cameras have proven to be very effec-
tive for diabetic screening. Screeners, optometrists, general
ophthalmologists, and retinal specialists can evaluate the
images, if properly trained.
RETINAL THICKNESS MEASUREMENT
Spectral domain ocular coherence tomography (OCT)
(Heidelberg Spectralis or equivalent) is used to evaluate
macular thickness in patients with NPDR. Although the
studies demonstrated excellent sensitivity to macular edema,
the real issue is whether it is beneficial to perform focal laser
before CSME is present on clinical biomicroscopic examina-
tion. This issue will be addressed in the next section.
171
172 SECTION IV ■ Disease-Sp ecific Managem ent
Fig ure 1 9 .1 ■ The Staurenghi 230 SLO lens perm its angiogra-
phy of preequatorial neovascularization.
NONPROLIFERATIVE RETINOPATHY
The Early Treatment Diabetic Retinopathy Study (ETDRS)
has contributed greatly to our management of patients
with NPDR (4,5). The ETDRS defined CSME as (a) thick-
ening of the retina at or within 500 µm of the center of
the macula, (b) hard exudates at or within 500 µm of the
center of the macula if associated with thickening of adja-
cent retina, or (c) an area of retinal thickening one disk
diameter or larger, part of which is within one disk diam-
eter of the center of the macula. It should be emphasized
that the ETDRS was based on the clinical perception of
macular edema. The majority of the physicians, including
the authors, do not treat pre-CSME patients. Pre-CSME
patients should probably be treated based on upcom-
ing cataract surgery, poor result in the other eye, marked
peripheral capillary nonperfusion, upcoming panretinal
photocoagulation (PRP), or an estimation that the patient
was unlikely to return for follow-up visits. Most surgeons
use the Goldmann (flat) contact lens or macular lenses
and start with a 60- to 100-µm, 100-milliwatts, and 20-
to 30-ms spot. Many surgeons now use lighter treatment
compared to the Diabetic Retinopathy Study (DRS) and
ETDRS. Some surgeons primarily treat microaneurysms
and microangiopathy noted to leak on angiography, oth-
ers primarily use a light grid in the thickened, leaking
area, and most treat both. Preferred lasers include argon
green (514 nm), diode or flash tube–pumped diode lasers
(532 nm), and krypton yellow (577 nm). Infrared (IR)
diode (>800 nm) lasers are not recommended because of
absorption in the choroid and pain. The authors prefer
a 532-nm, diode-pumped, frequency up-converted con-
tinuous wave YAG laser on the PASCAL platform.
Fig u re 1 9 .2 ■ OCT im ag e of vitreom acular traction in a
diab etic eye.
Ph arm aco t h e rap y o f Diab e t ic Macular Ed e m a:
Co rt ico st e ro id s
Since the introduction of intravitreal triamcinolone for
the treatment of uveitic cystoid macular edema (6), it has
become clear that steroids have a direct effect on the reti-
nal vascular endothelium, decreasing leakage and restoring
the blood retinal barrier. This discovery has led to the wide-
spread use of intravitreal steroids for macular edema due to
multiple causes, including diabetic retinopathy and retinal
vein occlusions. Recent reports from the Intravitreal Steroid
Injection Studies–Diabetic Macular Edema (7) (ISIS-DME)
have shown that intravitreal injections of triamcinolone
(Kenalog) result in significant visual improvement (three or
more lines) in 38% of patients. Subgroup analysis revealed
that 62% of patients with cystoid macular edema improved
at least three lines of vision, whereas only 9% of those with
noncystoid macular edema had a similar response. The
macular edema recurred in approximately 50% of patients 6
months after the injection. Complications noted were steroid
glaucoma in approximately a third of patients. Intravitreal
steroids have other important complications in addition to
steroid glaucoma. Posterior subcapsular cataract formation
is a well-recognized complication of steroids and should be
considered whenever injections are to be performed in pha-
kic individuals. The authors do not use steroids in the man-
agement of diabetic macular edema because of the untenable
steroid glaucoma and cataract rates.
The widespread use of intravitreal injections of ste-
roids has been followed by various drug delivery systems
that promise to prolong the release of steroids with slow-
release devices implanted in the vitreous. Currently, there
are two types of drug delivery steroid implants in clinical tri-
als: biodegradable implants (e.g., Osurdex dexamethasone-
polyacticglycolic acid) and nonbiodegradable (e.g., Retisert
fluocinolone implant). Both types of implants have potential
advantages and disadvantages. Biodegradable implants have
the advantage of requiring one implantation (that can be
performed in the office with a 22-gauge injection applicator)
but may have nonlinear release kinetics, whereas nonbiode-
gradable implants, although having linear release kinetics,
require more complex surgical implantation and subsequent
removal unless the implant is left in place. The Retisert clinical

trials reported a 90% incidence of steroid glaucoma and a 34%
incidence of glaucoma filtering procedures. The risk/benefit
and cost/benefit analyses of these implants in comparison with
repeated intravitreal injections of triamcinolone suggest that
the implants are too costly, in addition to causing an unaccept-
ably high rate of steroid glaucoma and cataracts.
Because of steroid glaucoma and steroid-induced cata-
racts, the authors do not use intravitreal steroids for diabetic
macular edema and use a combination of anti-VEGF therapy
(Avastin), PASCAL laser, and topical nonsteroidal agents
(Nepafenac, Alcon).
Ph arm aco t h e rap y o f Diab e t ic Macular
Ed e m a: Vascular En d o t h e lial Gro wt h
Fact o r An t ag o n ist s
Macular edema is caused by vascular endothelial growth fac-
tor (VEGF) (8–10), the same agent that causes retinal neo-
vascularization in diabetes and choroidal neovascularization
in age-related macular degeneration. VEGF downregulates
the tight junctions of the endothelium of the retinal ves-
sels, causing breakdown of the blood-retinal barrier, and
therefore leakage of fluid and macromolecules into the reti-
nal intercellular space. Ischemia leading to VEGF produc-
tion may be a factor in certain macular edema cases. For
this reason, patients with macular edema not responding to
direct focal treatment of leaking microaneurysms, or areas of
leakage on fluorescein angiography, may respond to treating
areas of ischemia. Heavy grid photocoagulation probably has
little effect in reducing neovascularization but significantly
reduces central visual fields and therefore reading speeds
and often causes patients with excellent Snellen acuity to
complain that they “cannot see.”
The introduction of anti-VEGF therapy has led to a new
mode of therapy for macular edema secondary to NPDR as
well as for venous occlusive disorders. Bevacizumab (Avas-
tin) is an anti-VEGF antibody currently approved by the FDA
for systemic therapy of cancer. Intravitreal Avastin is widely
used for the treatment of choroidal neovascularization in age-
related macular degeneration. Avastin is also being used for
macular edema secondary to Branch Retinal Vein Occlusion
(BRVOs) (11) and Central Retinal Vein Occlusion (CRVOs)
as well as for retinal neovascularization in proliferative dia-
betic retinopathy (PDR). The initial published results as well
as the authors’ results demonstrate that anti-VEGF therapy
combined with laser should become standard therapy in the
management of diabetic macular edema and PDR.
Man ag e m e n t o f Diab e t ic Macular Ed e m a
The recent advances in pharmacotherapy and surgical ther-
apy for diabetic macular edema, added to the well-known
and time-proven approaches with laser photocoagulation,
provide the clinician multiple therapeutic possibilities.
Although there are no trials that clearly indicate which com-
bination and sequence of therapies should be employed, the
CHAPTER 19 ■ Diabetic Retinopathy 173
authors present their current perspective of management of
diabetic macular edema, with the understanding that this
protocol may change in the near future as clinical research is
presented. Combination therapy is an appealing concept but
is appropriate in some instances and not in others. Combina-
tion chemotherapy in oncology is utilized because the agents
have narrow windows between effective and toxic drug lev-
els as well as to provide multiple barriers to the evolution of
cancer cells. Combination therapy for infectious disease is
utilized in severe infections when the infectious agent has
not been identified and delayed treatment would produce
bad outcomes. Combination therapy for infectious disease
creates multiple barriers for evolution of the infectious agent
but unfortunately leads to higher incidence of resistance.
Focal and/or PRP laser plus anti-VEGF therapy is very effec-
tive in diabetic retinopathy and can be broadly defined as
combination therapy. Topical nonsteroidal (Nepafenac) ther-
apy in combination with laser and anti-VEGF compounds
is effective for diabetic macular edema because of multiple
mechanisms, VEGF, and inflammation.
PROLIFERATIVE RETINOPATHY
The DRS randomized, multicenter clinical trial proved that
PRP is effective for patients with PDR (12–15). Many sur-
geons now treat before the DRS treatment criteria are reached
(16), in part because of the subsequent ETDRS study data.
The DRS criteria are based on high-risk characteristics as
compared to standard photographs. The authors use light
treatment with small spot size (100 to 200 µm), using 20- to
30-ms duration to reduce thermal diffusion, with the spots
placed one spot size apart. Many surgeons treat using a very
large number of intense, large, 100-ms duration spots and
see the patient in follow-up in 3 months. Some of this behav-
ior is unfortunately driven by the so-called global period for
Medicare reimbursement. It is probably better practice to use
a moderate number of spots and reexamine the patients in
1 month. Some surgeons have a protocol mindset with respect
to PRP treatment. This approach may result in inappropriate
vitrectomy or peripheral cryo when the “protocol has been
completed” and active neovascularization is present.
It is better to perform PRP fill-in, combined with anti-
VEGF therapy, if there is any neovascularization and the
patient is greater than 1-month posttreatment. The authors
believe that this combined, incremental approach to therapy
reduces treatment-related complications and discomfort. If
the patients present with florid neovascularization, it is better
to use anti-VEGF therapy at the initial visit combined with
PASCAL PRP. PRP with PASCAL using 30-ms, less intense
burns does not destroy ischemic inner retina raising ques-
tions about efficacy, but studies have shown identical out-
comes with less pain. The principal author points out that
the destruction of RPE and photoreceptors results in trans-
synaptic degeneration of the bipolar, horizontal, amacrine,
and ultimately ganglion cells similar to what occurs in
174 SECTION IV ■ Disease-Sp ecific Managem ent
retinitis pigmentosa. Combining PASCAL PRP with Avastin
combines the benefit of rapid response to Avastin with
the durable response of PRP without the disadvantage of
complications secondary to longer duration, more intense
burns.
The authors virtually never use a retrobulbar block
for laser treatment. Red and IR lasers cause more pain than
532-nm lasers. Durations greater than 30 ms cause more
pain because of thermal diffusion; a duration of 30 ms on
PASCAL is a better choice for PRP. Larger spot sizes and
higher powers cause more pain, light scattering, and poten-
tially more nuclear sclerosis and occult macular photic
damage.
VITRECTOMY
Better medical management and laser photocoagulation
should significantly reduce the need for vitrectomy for the
complications of PDR. Vitrectomy, when indicated, has an
excellent prognosis when appropriate patient selection and
techniques are utilized (17).
CASE SELECTION
It is useful to divide blindness from diabetic retinopathy
into two subgroups: those cases requiring immediate sur-
gery and those in which surgery is elective. Traction retinal
detachment (TRD) involving the macula (MTRD), neovas-
cular glaucoma (NVG), and anterior vitreous cortex fibro-
vascular proliferation (AVCFVP) are permanently blinding
if left untreated. By contrast, the visual potential probably
does not change in the treatment of vitreous or preretinal
hemorrhage if there is substantial delay before surgery is
performed.
Vit re o us He m o rrh ag e
Early experimental work incorrectly concluded that vitreous
hemorrhage caused neovascularization via organization of
the blood clot. Vitreous hemorrhage is a result of neovas-
cularization rather than the cause. Although long-standing
vitreous hemorrhage can deposit iron on many intraocular
structures, there is usually no retinal damage from a vitre-
ous hemorrhage. Retinal detachment, macular damage, isch-
emia, and optic nerve function will determine the ultimate
visual outcome when long-standing vitreous hemorrhages
are removed, not the hemorrhage per se.
If the other eye has good vision, a unilateral hemorrhage
can be followed indefinitely with ultrasound, unless retinal
detachment, anterior vitreous cortex (AVC) neovasculariza-
tion, or iris neovascularization occurs. An eye without prior
PRP is at greater risk for the development of neovascular
complications and must be watched more closely. B-scan
ultrasonography should be repeated at each visit, preferably
at 1-month intervals until the blood clears or surgery is
performed. Ultrasonic evidence of posterior pole detach-
ment requires immediate vitrectomy. The usual question of
duration of a vitreous hemorrhage plays a less important role
in the surgical decision-making process than other factors. If
it does not appear that near-term clearing will occur, bilat-
eral hemorrhage requires surgery on the eye with the best
visual prognosis. Vitreous hemorrhage in a patient with only
one eye as well as the better eye of bilateral cases should be
operated on to improve visual function. Those patients with
shortened lifespan and multisystem disease need immediate
visual rehabilitation for emotional and social reasons. Sub-
posterior vitreous detachment and preretinal hemorrhages
clear more rapidly than does hemorrhage in the vitreous cor-
tex. For this reason, patients with bilateral or only-eye sub-
PVD or preretinal hemorrhage can be followed up for as long
as the patient’s emotional and social needs permit. If one
eye has macular ischemia and the other, better, eye devel-
ops a vitreous hemorrhage, vitrectomy may be indicated to
improve the patient’s overall visual function.
Tract io n Re t in al De t ach m e n t
TRD can be diagnosed by ophthalmoscopic or ultrasonic
examination. If the macula is detached, vitreous surgery
should be performed within 3 weeks, unless there are medi-
cal contraindications. If there is active neovascularization, it
is better to perform PRP before vitrectomy if possible. Because
of extensive exudation and fibrous proliferation, panretinal
cryopexy should not be utilized. If vitrectomy indications
are present, endo-PRP can be combined with vitrectomy. If
vitrectomy is postponed until PRP-induced or spontaneous
involution of neovascularization occurs, the incidence of
postoperative NVG and AVCFVP is dramatically reduced.
Recent studies demonstrate that intravitreal anti-VEGF
therapy with bevacizumab can precipitate TRDs in patients
with severe neovascularization (18). These patients should
be followed closely after anti-VEGF therapy and the surgeon
should be ready to proceed to vitrectomy if progression to
TRD is evidenced.
Because of the relatively high rate of biologic complica-
tions and medical risk factors, vitrectomy is not indicated in
extramacular TRD. This is true even if progression toward
the macula or a similar condition in the other eye seems to
“threaten” the macula. It is safer to operate on actual, rather
than predicted, visual loss. The rate of progression of extra-
macular TRD to include the macula is about 15% per year
(19,19a,20). After several years, progression to MTRD stabi-
lizes at a cumulative rate of about 30%, and there are many
patients with 5 to 10 years duration extramacular TRDs sur-
rounding the macula with good vision that never required
surgery.
Cataract surgery can result in anterior movement of the
vitreous with progression of extramacular TRD to macular
involvement. Once again, vitreous surgery should only be
performed if the macula actually becomes elevated (21).

MACULAR EDEMA, CYSTS,
AND SUBMACULAR EXUDATES
Hilel Lewis et al. first reported vitrectomy with peeling of
taut posterior vitreous cortex (PVC) from diffuse macula
edema (DME) eyes not responsive to focal laser photocoagu-
lation. SD OCT is very effective at determining the presence
of vitreoretinal traction and measuring macular thickness.
Many surgeons now believe that vitrectomy in DME cases
without vitreomacular traction can reduce macular edema
similar to the observations in vitrectomy without branch
vein decompression for BRVO and vitrectomy without radial
optic neurotomy in CRVO. The principal author has postu-
lated that vitrectomy or separation of the PVC could allow
greater egress of VEGF, known to be causative of macular
edema. Steffanson and later Holekamp have shown that vit-
rectomy produces higher oxygenation in the vitreous cavity
(22), which may improve macular edema as well. Peeling of
the internal limiting membrane can improve diabetic mac-
ular edema and should be considered in patients who are
resistant to laser and pharmacologic treatment (23). Some
surgeons have reported removal of submacular exudates or
draining macular cysts, but there is scant evidence of efficacy
for either procedure at this time and some concern about
retinal damage caused by the procedures.
CONTRAINDICATIONS
The absence of light perception indicates glaucomatous
optic atrophy, ischemic optic neuropathy, or extensive retinal
vascular occlusive disease and contraindicates vitreous sur-
gery. Corneal opacity; corneal, lid, or conjunctival infection;
and the inability to withstand local anesthesia are relative
contraindications.
Iris neovascularization can be an indication for surgery
in diabetic TRDs (24–26). Intravitreal bevacizumab has been
shown by the authors to cause regression of anterior seg-
ment neovascularization and revert the progression toward
NVG (27). Anti-VEGF therapy for NVG requires permanent
control of VEGF production through retinal reattachment
and PRP. Vitrectomy in an aphakic eye with active iris neo-
vascularization will result in rapid progression of postop-
erative NVG unless intravitreal bevacizumab and PRP are
performed preoperatively and intraoperatively.
Cases of several years’ duration may exhibit extensive
white vessels and retinal atrophy. If the temporal arcades are
not perfusing the macula, there is no need for vitreous sur-
gery because visual improvement will not occur. If the retina
is extremely atrophic but PRP has not previously been per-
formed, this is an indication that limited visual improvement
will occur from reattachment. There are, however, cases of 3
to 4 years’ duration that have improved to ambulatory vision
levels after vitrectomy. Late cases with atrophic retina produc-
ing little VEGF have a lesser incidence of NVG and AVCFVP.
CHAPTER 19 ■ Diabetic Retinopathy 175
SURGICAL SEQUENCE
AND TECHNIQUES
Vitreous surgery for hemorrhage or TRD requires a planned
sequence of surgical steps, with multiple branches depend-
ing on different intraoperative scenarios. As in all vitreous
surgery, a full complement of sterile tools and materials must
be immediately available.
An e st h e sia
The frequency of cardiovascular and renal disease in the dia-
betic patient requires careful preoperative evaluation by the
primary care physician or internist and utilization of cardi-
ologists, endocrinologists, and other consultants as needed.
The anesthesiologist should review the preoperative medi-
cal evaluation. Diabetic vitrectomies can be performed in the
ambulatory surgery center setting if systemic disease is stable,
an anesthesiologist is available and the patients systemic spe-
cialists (nephrologists, cardiologists) give adequate consent
for surgery. Patients that are not controlled systemically and
require vitrectomy surgery that cannot be delayed should
be operated in a hospital outpatient setting. It is essential
to have MD anesthesia function in an immediate availability
and supervisory role if CRNA anesthesia is utilized. Proximity
to the hospital ensures availability of cardiologists, endocri-
nologists, pulmonary specialists, and intensive care facilities
as well as providing higher facility fees than free-standing
ambulatory surgery centers, thereby facilitating access to best
technology. An intravenous line, EKG, pulse oximetry, blood
pressure monitoring, and oxygen mask with suction hose to
prevent hypercapnia must be utilized in all cases. The anes-
thesiologist or nurse anesthetist should make liberal use of
intraoperative serum glucose monitoring. Operating times
are always less than 1 hour and usually about 30 minutes,
which is compatible with local anesthesia and the associated
reduction in nausea, vomiting, and medical complications
of general anesthesia. Minimal, if any, sedation is used after
the block, which is performed with a 27-gauge, 1.25-inch
needle at the outer “corner” of the orbit.
In cisio n s
The authors use 25-gauge techniques for all cases, although
23-gauge methods have similar advantages and results.
Man ag e m e n t o f t h e Le n s
Lens removal correlates with an increased incidence of post-
operative NVG but prevents AVCFVP (28–30). The AVC and
lens apparently act as barriers to the anterior diffusion of
VEGF, and therefore their presence reduces the incidence of
NVG. A specific attempt should be made to leave the AVC, if
the lens is retained during vitrectomy in an effort to reduce
the development of posterior subcapsular cataract.
176 SECTION IV ■ Disease-Sp ecific Managem ent
Contact lenses for aphakic correction should be uti-
lized judiciously because of decreased corneal sensitivity
and infection risk but are effective in a surprising number of
patients. Spectacle correction is surprisingly well tolerated
in these patients. Implantation of posterior chamber lenses
after endocapsular lensectomy (ECL) or phacoemulsifica-
tion tends to keep the AVC and posterior capsule intact, may
decrease NVG, offers the best optical outcome, and can be
used in all but the most complicated cases.
ECL with a 20-gauge ultrasonic fragmenter is preferable
to phacoemulsification combined with vitrectomy if there is
significant fibrin syndrome or neovascularization. Preopera-
tive anti-VEGF therapy as well as injection of anti-VEGF
agents at the end of the case reduce the complications and
enable more frequent use of phaco-vit. Phaco combined with
vitrectomy reduces the number of procedures. If the phaco
is performed at the beginning of the procedure, pupillary
dilation and corneal clarity may suffer, thereby reducing sur-
gical visualization. If phaco and intraocular lens (IOL) inser-
tion are performed after the vitrectomy, unavoidable periods
of hypotony will increase bleeding. If significant cataract is
noted at the preoperative evaluation, best practice is usually
to send the patient to a cataract surgeon for phaco and IOL.
If the preoperative view is judged to be adequate to per-
form vitrectomy, phaco and IOL insertion can be performed
approximately 1 month after vitrectomy. If the preoperative
view was judged to be adequate but it was determined dur-
ing vitrectomy that the view was inadequate, endocapsular
fragmenter lensectomy with retention of the anterior cap-
sule is currently the author’s preference. If sufficient capsule
was retained and bleeding is unlikely, a foldable, multipiece
acrylic sulcus lens can be implanted anterior to the anterior
capsule. An anterior chamber lens can be implanted if there
is minimal likelihood of bleeding and the capsule is not suit-
able for a posterior chamber lens.
Vit re ct o m y
The continuity of the PVC is a critical concept in the
understanding and planning for the vitrectomy process.
The PVC will be completely adherent to the retina, par-
tially detached, or completely detached from the retina. The
“core” vitrectomy misconception stems from earlier days
when high suction and slow cutting pulled vitreous into the
central portion of the eye. Complete sectioning or truncation
of the PVC (Fig. 19.3), rather than “band cutting” or “core
vitrectomy,” must be understood before surgical success can
be obtained. These concepts apply whether the vitreous is
opaque, semiopaque, or clear.
Pro ce d ure if Po st e rio r Vit re o us
De t ach m e n t Is Pre se n t
If the PVC is detached from the retina, a central opening in
the vitreous should be created and linear extrusion with the
vitreous cutter or a 25-gauge, soft-tip cannula performed if
any subvitreous erythroclastic or hemolytic material is pres-
ent. This step is also known as vacuum cleaning or extru-
sion. When a clear effluent is obtained from this fenestration,
the opening should be enlarged with the cutter until only a
small “skirt” at the confluence of the AVC and PVC remains.
Particular care must be taken to trim the superior “skirt”
Fig u re 1 9 .3 ■ Trun cation o f
PVC relieves so-called anterop os-
terior traction on diabetic traction
detachm ents.

Fig u re 1 9 .4 ■ Lin ear suction
with a straight 25-gauge, soft-tip
cannula should be used to rem ove
all p reretin al b lo od p rod u cts,
p erm it b etter visualization an d
less p ostop erative erythroclastic
g laucom a, and enab le p hotoco-
ag ulation without d am ag ing the
retina.
if opaque, because it can hang down postoperatively and
obscure the seated patient’s view. Linear suction with a straight
25-gauge, soft-tip cannula should be performed to remove
all preretinal blood products, permit better visualization
and less postoperative erythroclastic glaucoma, and enable
photocoagulation without damaging the retina (Fig. 19.4). If
a complete posterior vitreous detachment (PVD) is present,
there is no perpendicular or oblique force on the retina, but
tangential force from epiretinal membranes (ERMs) can cause
TRD. Vascular attachment points of the ERM to the retina
should be treated with the bipolar diathermy or endophoto-
coagulator only if they bleed intraoperatively or appear active.
The Alcon 20-gauge disposable bipolar endoilluminator and
25-gauge bipolar diathermy are ideal for this technique.
Pro ce d ure if Part ial Po st e rio r Vit re o us
De t ach m e n t Is Pre se n t
If only a partial PVD has occurred, the vitreous will be
adherent to the retina at one or more epicenters. Typically,
the optic nerve and vessels serve as attachment points
because of glial proliferation. As the vitreous contracts,
these attachment points become the apex (apices) of the
now-conical PVC. This is the most common vitreous
configuration encountered in PDR. It is critical in these
patients to section or truncate the PVC to completely
eliminate any anteroposterior traction (Fig. 19.5). Vitre-
ous bands do not exist as such but are actually more vis-
ible portions of the PVC continuum. The PVC penetration
is initiated in an area indicated by preoperative ultrasound
or indirect ophthalmoscopy to have attached retina. In the
CHAPTER 19 ■ Diabetic Retinopathy 177
absence of this information, the first opening should be
made nasally in the midperiphery to avoid the macula and
in an area easily treatable if a retinal break occurs. After
an opening is made, extrusion (vacuum cleaning) through
the opening must be continued until a clear effluent is
obtained. It is at this time that the novice vitreous surgeon
could mistake voluminous old blood trapped behind the
vitreous for active bleeding. When the retina is visualized
through the opening, a safe circumferential truncation of
the PVC can proceed from this point and extend for 360
degrees. It is not necessary to make multiple openings in
the PVC or to dissect the layers of the PVC in “onion skin”
fashion. The “skirt” must be trimmed as described above
and the portion connected to the retina trimmed down
to near the retinal surface. Any areas of PVC connected
to two or more retinal points should be sectioned with
the vitrectomy instrument, if there is sufficient clearance
for the tip, or with the 25-gauge, curved segmentation/
delamination scissors. If these bridging areas of PVC are
vascularized, they can be precoagulated with the dispos-
able bipolar cautery.
Pro ce d ure if No Po st e rio r Vit re o us
De t ach m e n t Is Pre se n t
In some diabetics, the entire PVC is adherent to the retina. In
the unlikely event that the PVC is minimally adherent, mem-
brane peeling can be performed. Caution must be exercised
in attempting to peel the PVC in a single sheet because reti-
nal breaks can occur in sites remote from the forceps grasp
site. Areas of photocoagulation can cause adherence of the
178 SECTION IV ■ Disease-Sp ecific Managem ent
PVC to the retina as well as the choroid and sclera. If strongly
adherent zones are present, the PVC should be allowed
to remain in these locations and the scissors or vitreous
cutter used to sever all tangential traction (Fig. 19.6). PVC
truncation must be completed in every case but can be per-
formed with scissors in the case of shallow PVDs or following
inside-out delamination if a TRD is present (Fig. 19.7). This
Fig u re 1 9 .5 ■ Im b rication of
taut PVC into the port of the cut-
ter can cause posterior or periph-
eral retinal b reaks. It is b etter to
sever the PVC with scissors.
latter method, developed by the principal author, was later
termed “en-bloc” when described by others using a less safe,
outside-in direction of dissection. Even worse, the advocates
of en-bloc dissection, as originally described, suggest that
intentional traction on the vitreous can facilitate dissection.
Using the vitreous to lift the ERM has substantial risk of
causing peripheral retinal breaks from counter traction.
Fig u re 1 9 .6 ■ In sid e-out d el-
am in atio n o f ERM co n tig uo us
with PVC.

Fig ure 1 9 .7 ■ Rem oval of PVC
and ERM after inside-out delam i-
nation (en bloc).
DIABETIC TRACTION RETINAL
DETACHMENT
Understanding surgical anatomy is crucial to a success-
ful surgical approach; the central element is the continuity
of the PVC and the role of vitreoretinal (so-called AP trac-
tion) and tangential, ERM-driven traction. Three often used
phrases indicate a poor grasp of surgical anatomy and surgi-
cal principles: (a) core vitrectomy, (b) resection of traction
“bands,” and (c) “membrane peeling or stripping.” Let us
start with the incorrect concepts of core vitrectomy and trac-
tion bands. There is no “core” vitreous and there are no trac-
tion bands; the entire transparent PVC is the cause of TRD,
not a region of vitreous rendered opaque by chronic vitreous
hemorrhage or a so-called traction band. It follows that the
first and most important surgical goal is 360-degree trun-
cation of the conical (or truncated cone if table-top TRD)
PVC midway between the outer margin of ERM attachment
areas and the periphery. Secondly, the incorrect notion of
membrane peeling (stripping) must be addressed. ERMs in
the vast majority of diabetic TRD cases are highly adherent;
attempts to peel or strip the membranes using pics and simi-
lar tools usually result in iatrogenic retinal breaks. The prin-
cipal author developed scissors segmentation of ERMs and
subsequently scissors delamination of ERMs to address the
strong adherence problem.
Truncation of the PVC should be performed by using
the highest possible cutting rates and lowest vacuum while
applying the port to the surface of the PVC. It is dangerous
and inefficient to attempt to do core vitrectomy and pull the
vitreous to the port using excessive vacuum and flow rates.
CHAPTER 19 ■ Diabetic Retinopathy 179
Occasionally the PVC is so taut that it is better to truncate
some regions with scissors rather than the vitreous cutter.
Smaller cutters (25 or 23 gauge) usually obviate the need
for this step. Wide-angle illumination systems such as chan-
deliers and Tornambe Torpedos reduce the ability of the
surgeon to visualize transparent PVC, necessitating the use
of triamcinolone particulate marking. Similarly, wide-angle
visualization systems (BIOM, EIBOS, Volk, AVI) reduce lat-
eral and axial resolution and, just as chandeliers and Torpe-
does, hinder visualization of the transparent PVC.
Truncation of the PVC should be followed by manage-
ment of ERMs unless no PVD is present. If a PVD has not
occurred, inside-out delamination of ERM starting at or near
the disk extending out along the arcades should be initi-
ated rather than vigorously and dangerously trying to create
a PVD. “En bloc dissection,” as initially described, utilized
outside-in technique and advocated using the PVC to lift the
ERM. Putting traction on the PVC results in peripheral reti-
nal breaks while the surgeon is focusing on the ERM dissec-
tion. En bloc is a poor term when applied to ERM dissection;
it is only meaningful in cancer surgery when the notion is to
minimize the dispersion of cancer cells. A far better way to
think about this issue is to forget the outdated notion that
the standard surgical algorithm is first performing core vit-
rectomy, then creating a PVD, and finally peeling ERMs. If a
PVD is not present in a diabetic TRD case, it is better to go
directly to inside-out delamination starting at or near the
optic nerve head. It is not necessary to remove all ERM in
one piece as suggested by the term en bloc; it is preferable to
use access segmentation, cutting ERM into several segments,
each of which will be contiguous with the PVC. When most
180 SECTION IV ■ Disease-Sp ecific Managem ent

Fig ure 1 9 .8 ■ Access segmenta-

tion develops a plane for inside-out
delam ination. Forceps m em brane
p eeling , scissors seg m entation,
and scissors delamination are usu-
ally best performed in an inside-out
direction because central retina
is stronger than m ore peripheral
retina, the retina is redundant cen-
trally, and edges are hard to find,
creating the risk of iatrogenic retinal
breaks.

or all the ERM has been delaminated using an inside-out and delamination means to cut it off. Initially, delamination
direction, PVD creation will be accomplished as well or it was performed with so-called horizontal scissors, actually
will be irrelevant because there are no longer any posterior 135 degrees. For almost two decades, the principal author used
attachment points. only curved scissors for what is termed “conformal delamina-
tion.” The 25-gauge Alcon DSP curved scissors have been used
for all cases since they became available (Fig. 19.9). Curved
Scisso rs Se g m e n t at io n an d De lam in at io n
scissors are preferred to “horizontal” scissors because the
Scissors segmentation of ERMs is used to reduce tangential curvature of the scissors matches the curvature of the retina,
force on the retina by separating ERM into so-called epicen- reducing the likelihood of impaling the retina with scissors tips.
ters, better termed “vascularized adherence sites.” Segmenta- The same design concept is used for skis, the tips curved up.
tion is performed by placing one blade of the scissors under If the scissors are introduced under the ERM with the blades
the ERM, between the retina and the ERM, and the other blade wide open and then closed, attachment points will be ripped
anterior to the ERM. Although this was initially performed with
so-called vertical scissors, curved scissors are far superior for
segmentation because blade width is much greater than blade
thickness; “vertical” scissors require more space between the
retina and the ERM than curved scissors. The principal author
largely abandoned segmentation after developing delamina-
tion; segmentation is now primarily used for access segmen-
tation to expose the delamination plane (Fig. 19.8). Using
curved scissors expedites transitioning from access segmenta-
tion to delamination without tool exchange.
The senior author developed scissors delamination
2 years after developing segmentation to address the issue of
residual ERM after segmentation. Residual ERMs in diabetic
TRD cases resulted in small areas of chronic retinal elevation,
which ultimately caused atrophic retinal holes and late rheg-
matogenous retinal detachments. Scissors delamination is per-
formed by inserting both scissors blades in the potential space
between the ERM and the retina and severing the attachment Fig ure 1 9 .9 ■ Curved scissors are the best tool for both seg-
points. Simplistically, segmentation means to cut up the ERM m entation and inside-out conform al delam ination.
 CHAPTER 19 ■ Diabetic Retinopathy 181

Fig ure 1 9 .1 0 ■ Conform al cut-

ter delam ination is p erform ed by
rotating the cutter around the lon-
gitudinal axis to control the angle
of attack; ERM is fed into a p ort
that is p ositioned ap p roxim ately
9 0 d eg rees fro m th e tan g en t
plane to local retinal contour.

instead of sheared, which causes retinal breaks. If the scissors
are introduced under the ERM with the blades closed and then
opened similar to the technique used for blunt dissection in
general surgery, attachment points will be ripped instead of
sheared, creating retinal breaks. The optimal technique is to
open the scissors just enough to engage an attachment point,
cut, pull back slightly, move laterally to the next attachment
point, advance slightly, and cut again. It is not necessary to lift
or stabilize the ERM if disposable scissors or reasonably new
reusable scissors are used. Because ERM is elastic and there is a
pressure gradient across the retina, the ERM will scroll up and
separate from the retina without lifting.
Cut t e r De lam in at io n o f Ep ire t in al Me m b ran e s
Vitreous cutters are increasingly being used for removal
and segmentation of ERMs in diabetic TRD cases, but they
cannot replace scissors delamination for all cases. There
are three distinct techniques for using scissors in diabetic
TRD surgery: (a) conformal cutter delamination, (b) fold-
back delamination, and (c) segmentation of PVC segments
between adherence sites. Conformal cutter delamination is
performed by placing the port at the outer margin of the
ERM and feeding, not sucking or pushing the ERM into
the port (Fig. 19.10). It is conformal because the angle of
attack is continuously adjusted so that the port is rotated
away from the retinal surface as the cutter is moved along
the concavities and convexities of the TRD. This technique is
used for relatively rigid, thick ERMs not candidates for fold-
back delamination. Foldback delamination involves plac-
ing the cutter port on the anterior surface of the ERM just
behind the leading edge, allowing vacuum to cause the ERM
to fold back into the port (Fig. 19.11). With this technique,
the ERM protects the retina. Smaller cutters, 23 gauge or
preferably 25 gauge, are ideal for these techniques as well as
segmentation of PVC between vitreoretinal adherence sites
because the port is close to the tip; the tip has smaller diam-
eter, reducing axial and lateral setback from the target ERM.
In sid e -Out Ve rsus Out sid e -In Te ch n iq ue
Machemer used outside-in technique for bent needle mem-
brane peeling, and many surgeons unfortunately emulate this
technique when using delamination. A far better approach
for end-grasping forceps membrane peeling, scissors seg-
mentation, and scissors delamination is to initiate the dis-
section centrally and peel, segment, or delaminate outward.
Outside-in dissection is safer because the retina is thicker
and 100× stronger centrally than in the periphery, especially
in ischemic and post-PRP patients, the retina is redundant
centrally in TRD cases, and the view is better centrally.
Visco d isse ct io n
Viscodissection is performed by injecting a viscoelastic in the
potential space between the ERM and the retina. Viscodissec-
tion is seldom performed today because it has not been shown
to be advantageous and it increases cost and operating time.
Viscodissection pressurizes the potential space between the
ERM and the retina, which can result in a hydraulic retinal
break if the ERM is tightly adherent. If the ERM is very
182 SECTION IV ■ Disease-Sp ecific Managem ent

Fig u re 1 9 .1 1 ■ Fold b ack d el-

am in ation involves p lacin g th e
cutter p ort on the anterior sur-
face of the ERM just b ehind the
leading edge allowing vacuum to
cause the ERM to fold b ack into
the port.

adherent with no gap between the ERM and the retina, the He m o st asis
visco cannula may create a retinal break during introduction.
Endolaser is better than diathermy for treatment of the
Viscoelastic substances significantly decrease the interfacial
severed vessel ends of vascular attachment points after
tension of silicone oil, increasing the likelihood of emulsifi-
delaminating ERMs in diabetic TRD surgery (Fig. 19.12).
cation. It is virtually impossible to remove the viscoelastic;
The authors believe that using diathermy on retinal sur-
the viscous material potentially increases glial recurrence
face bleeders can result in subclinical retinal necrosis and
rates by retaining cells, fibronectin, and cytokines such as
late atrophic holes. On occasion, larger vessels may require
VEGF, basic fibroblast growth factor, and TGF-β.

Fig ure 1 9 .1 2 ■ Focal 25-g auge

laser is id eal for coag ulating vas-
cular attachm ents after delam ina-
tion because of sm all spot size.

diathermy before or after transsection, especially when
performing retinectomy.
Pan re t in al Ph o t o co ag ulat io n
PRP should be performed on all attached areas when operat-
ing diabetic TRD cases. However, PRP should not be per-
formed on areas that were detached preoperatively because
the ever-present thin layer of subretinal fluid (SRF) after
fluid-air exchange and internal drainage of SRF coupled with
retinal edema results in overtreatment and fibrin syndrome.
An t i–Vascular En d o t h e lial Gro wt h
Fact o r Th e rap y
Preoperative bevacizumab (off-label in the United States,
Avastin) rapidly causes regression of neovascularization in vir-
tually all cases, dramatically reducing intraoperative bleeding.
Caution is required, however, because rapid cicatrization can
occur, resulting in increased TRD. This effect can be avoided
in almost all instances by operating between 4 and 7 days after
injection. As Avery has pointed out, it is crucial to reinject
Avastin at the end of the case to avoid rebound. The principal
author recommends injecting Avastin at the end of all diabetic
vitrectomies recalling that the anti-VEGF effect of PRP may
take longer than Avastin to take effect and PRP and Avastin are
synergistic. Endo-PRP decreases NVG and AVCFVP in high-
risk cases but is probably not required in cases with com-
pletely regressed neovascularization. TRD cases are at higher
risk for neovascularization than are hemorrhage cases, but
unfortunately endophotocoagulation cannot be safely utilized
on elevated retina. PRP should not be performed on formerly
detached retina because a thin layer of SRF always remains,
the retina is often edematous, and overtreatment is inevitable.
Overtreatment results in fibrin syndrome.
Silico n e Oil
Silicone oil should be used for rhegmatogenous confinement
usually without retinopexy when retinectomy is required
for areas of highly adherent ERMs affixed to atrophic retina.
This is more often the case with long-standing extramacular
TRDs that finally developed a retinal break and progressed to
include the macula necessitating surgery. Midperipheral and
nasal areas are more likely to require retinectomy. The authors
often utilize retinectomy without retinopexy when reoperat-
ing cases initially operated by other surgeons who created
a large number of retinal breaks attempting to peel highly
adherent membranes. Long-term surface tension management
is a crucial tool in these difficult cases. Excessive retinopexy
causes glial recurrences and fibrin syndrome; permanent sili-
cone oil eliminates the need for retinopexy, hence the terms
“rhegmatogenous confinement” and “retinopexy avoidance.”
There is an undue emphasis on silicone oil removal, which
often results in redetachment. Silicone oil complications are
minimal and incidentally equal for 1,000 and 5,000 cs when
CHAPTER 19 ■ Diabetic Retinopathy 183
the oil is confined behind an IOL. The emulsification and
silicone oil glaucoma rate is less than 5%.
An cillary Ep ire t in al Me m b ran e Man ag e m e n t
Te ch n iq ue s
Multifunction tools such as the disposable, 20-gauge bipolar
endoilluminator and the Chang end-aspirating laser probe
were developed because of the need of more than two func-
tions and the limitation of having only two hands. Illuminated
instruments have been developed largely to enable so-called
bimanual surgery, which is better termed “forceps stabiliza-
tion of ERMs.” The general concept is that scissors produce
a push-out force and pics produce stress, which can tear
the retina during the dissection of ERM. The purpose of the
forceps is to offset these unwanted forces. Bimanual technique
is often incorrectly used to lift the ERM, which often results in
retinal tears. The author seldom uses bimanual surgery.
Viscodissection can create retinal breaks by pressurizing
the potential space between the ERM and the retina and then
hydraulically rupturing the 100× weaker retina. Viscodissection
has not been shown to improve outcomes, and viscoelastics
can cause retention of cytokines, cells, fibrin, and blood at the
retinal surface, which may lead to glial reproliferation. The
authors do not recommend using viscodissection. As men-
tioned previously, all of authors’ complicated diabetic TRDs are
operated using 25-gauge techniques, and multifunction tools
are currently not available for 25-gauge instruments. Bimanual
techniques require chandelier illumination, which is usually
unnecessary for successful repair of the most severe TRDs.
Re t in e ct o m y
If extremely dense ERM is firmly adherent to atrophic, ele-
vated retina, it is often advisable to perform retinectomy,
removing membrane and retina together. Retinectomy is
often best done under air with the cutter also being used for
fluid-air exchange and internal drainage of SRF. If the retine-
ctomy is small, it can be surrounded by endolaser. If large or
multiple retinectomy areas are present, air-silicone exchange
is a better choice than gas. If silicone oil is used, retinopexy
avoidance is practiced, as this reduces tissue damage and
glial reproliferation. Laser can be performed in a fractional
dose staged manner in the postoperative course if removal
of the oil is anticipated. In many, if not most, cases, the best
practice is to allow the oil to remain in place indefinitely.
SURFACE TENSION MANAGEMENT
Fluid-gas exchange should only be utilized if there are atro-
phic or iatrogenic retinal breaks. The author prefers SF6 to
C3F8 because fibrin may form around long-term bubbles in
diabetic TRD patients. If retinal breaks are noted at surgery, all
ERM should be delaminated followed by simultaneous fluid-
air exchange and internal drainage of SRF. The Alcon 25-gauge
J-wand is an ideal tool for simultaneous internal drainage of
184 SECTION IV ■ Disease-Sp ecific Managem ent
SRF, fluid-air exchange, and removal of preretinal blood. The
25-gauge silicone soft-tip cannula also allows drainage of SRF
with minimum trauma to the intraocular tissues. Portions of
the retina that were attached intraoperatively should not be
treated with PRP because this increases the likelihood of fibrin
syndrome due to overtreatment resulting from residual SRF
and retinal edema. Intraocular air (gas) is only utilized if a
retinal break is suspected or seen. The only role of air (gas)
is to restore the transretinal pressure gradient via surface ten-
sion. Air (gas) surface tension management has no role in the
treatment of TRD without retinal breaks (no tear, no air).
SILICONE OIL SURFACE TENSION
MANAGEMENT
Silicone should be utilized when large or multiple retinal
breaks or retinectomies are present to eliminate the need for
retinopexy and reduce reproliferation, since it allows long-
term (or indefinite) surface tension management of retinal
breaks. Internal fluid-air exchange with the console air
source and internal drainage of SRF with linear extrusion
should precede air-silicone exchange.
Silicone oil acts as a barrier to the anterior diffusion of
VEGF and markedly reduces anterior segment neovascular-
ization if no inferior iridectomy is present. Because silicone
recompartmentalizes the eye, fibrovascular proliferation can
occur at the retina-silicone interface (Fig. 19.13). For this
reason, extensive PRP is our best hope for these cases. Sili-
cone may prevent oxygen diffusion from well-perfused to
ischemic retinal areas, thus causing visual loss. The princi-
pal author used perfluorocarbon liquids in an animal study
in 1976 in the hope that it could be used as a vitreous
substitute in the diabetic patient. It was evaluated because of
high oxygen capacity. Unfortunately, it was discovered that
it caused severe damage to the inferior cornea when left in
place for months.
VISUALIZATION
Wide-angle visualization is ideal for rhegmatogenous reti-
nal detachments, proliferative vitreoretinopathy, and giant
retinal breaks but disadvantageous for most diabetic TRDs.
Flat irrigating contact lenses (Machemer) provide much bet-
ter axial (depth) resolution and lateral resolution than wide-
angle optics and are preferred for most diabetic vitrectomies.
Wide-angle illuminations such as the chandelier and Torpedo
do not permit focal-, specular-, and retro-illumination tech-
niques, which are essential to visualize transparent vitreous.
INFUSION FLUID
Before the availability of intraoperative serum glucose moni-
toring, best practice was to use an intravenous infusion of
dextrose 5% in water to prevent unrecognized hypoglyce-
mia. Serum glucose levels were often in the 300 to 500 mg%
range, necessitating adding dextrose to the infusion fluid.
The author has not added dextrose or anything else to the
infusion fluid since BSS Plus (glutathione bicarbonate Ring-
er’s solution) became available and never observes intraop-
erative or early postoperative posterior subcapsular cataract.
Smaller fluid volumes associate with three-port systems,
Fig u re 1 9 . 1 3 ■ Silico n e o il
and g as cause recom p artm ental-
ization, wh ich seq uesters VEGF,
fibrin, cells, and cytokines at the
retinal-bubble interface.

better fluidics, and better techniques, and now 25-gauge
fluidics have improved the situation as well.
RETINOPEXY
All breaks require retinopexy unless they are macular, papil-
lomacular bundle, or peripapillary in location, in spite of
evidence that an occasional untreated break will not result in
detachment. The laser endophotocoagulator is used to treat
all breaks unless they are very extensive, indicating the need
for long-term silicone oil for rhegmatogenous confinement.
SCLERAL BUCKLING
Although a scleral buckle could potentially support any reti-
nal breaks that cannot reach the RPE after dissection, internal
fluid-air exchange, and internal drainage of SRF, retinectomy
is preferred. Prophylactic encircling bands have not been used
by one of the authors (S.C.) since the early 1980s because of
improved vitreous cutters, fluidics, and techniques.
RESULTS
Greater than 80% of diabetic TRD patients managed in the
previously described manner will sustain visual improvement
with vision greater than 5/200 (31–33). Ninety-seven percent
of the retinas of the patients are attached at the 2-week post-
operative visit, but even after reoperation, 5% of the patients
are blind (34) from AVCFVP and glial recurrence with sec-
ondary retinal detachment (35–37). The incidence of glial
recurrence is less with delamination than with segmentation.
Glial recurrence never occurs in truncation-only cases.
Some aphakic patients with attached retinas ultimately
become blind from NVG in spite of careful management.
NVG correlates with the presence of active retinal neovascu-
larization. AVCFVP causes permanent blindness in some of
the phakic cases (38).
Some of the patients with attached retinas do not have
improved vision because of photoreceptor damage and reti-
nal ischemia. Some of the successfully operated cases ulti-
mately become blind from ischemic optic neuropathy. Some
become blind from open angle glaucoma. Case selection has
a large impact on success rate, but the goal is to help every-
one possible, not to improve the success rate by elimination
of difficult cases. Patients with good results at 6 months typi-
cally have excellent long-term success (39–43).
COMPLICATIONS
He m o rrh ag e
Immediate postoperative intraocular hemorrhage can occur
from ERM vascular attachment points, nontreated new
CHAPTER 19 ■ Diabetic Retinopathy 185
vessels, or sclerotomies. Approximately 50% of phakic cases
develop immediate postoperative vitreous hemorrhages
(44). In aphakic cases, this hemorrhage will typically clear in
1 to 2 weeks, but the phakic cases can take several months.
If ultrasound indicates that the retina is attached, and there
is excellent vision in the other eye, no reoperation is neces-
sary. If ultrasound shows the retina to be detached, immedi-
ate reoperation is indicated. If the patient is bilaterally blind,
emotional and social needs dictate the need for reoperation.
It is advisable to operate on the eye with the highest visual
potential whether it is a first operation or reoperation.
A full vitrectomy setup with standard three-port
25-gauge vitrectomy is preferable to washouts in most
instances. In this way, ERM, persistent vitreous traction to neo-
vascular tufts, and bleeding vessels can be managed and endo-
PRP combined. Two-needle, in-office fluid-fluid exchange can
be used if medical conditions do not permit surgery under
monitored local anesthesia. If any neovascularization is pres-
ent, endo-PRP should be combined with the procedure.
Po st o p e rat ive Cat aract
If a visually significant cataract occurs in the best or
only-vision eye, it should be removed using phacoemulsi-
fication and PCL implantation. If the cataract occurs in the
poorer vision eye, the patient can be observed with ultra-
sound if medical status does not warrant surgery.
Eryt h ro clast ic (He m o lyt ic) Glauco m a
Erythroclastic (hemolytic) glaucoma is best prevented by blunt
cannula extrusion with the 25-gauge soft-tip cannula, trim-
ming of the vitreous skirt, and coagulation of all bleeding ves-
sels. If the pressure exceeds 25 mm Hg, topical alpha-agonists
and beta-blockers will usually control the pressure. On rare
occasions, reoperation may be needed to control the pressure.
Ne o vascular Co m p licat io n s
Just as retinal neovascularization is the most significant
complication of the unoperated PDR eye, NVG and AVCFVP
(AHFVP, RLNV) are the most severe problems in the postvit-
rectomy eye. An understanding of the pathogenesis is neces-
sary to reduce and manage neovascular complications.
Anterior segment and AVC neovascularization are due to
VEGF released from hypoxic but noninfarcted retina (45–49).
PRP is successful in reducing VEGF by causing the destruc-
tion of hypoxic retinal areas, transient release of an inhibitor
substance (50), and increased choroidal oxygenation of the
retina (51). Trabecular meshwork neovascularization with-
out peripheral anterior synechia or apparent iris neovascu-
larization can cause severe glaucoma. It is no longer thought
that anterior segment neovascularization is secondary to a
circulatory disturbance or that iris neovascularization some-
how migrates to the trabecular meshwork. Although vitrec-
tomy can induce changes in the oxygen distribution in the
186 SECTION IV ■ Disease-Sp ecific Managem ent
globe, this observation does not explain the transmissibility
of ocular neovascularization from human vitreous specimens
to bioassay systems, which can only be explained by VEGF.
Intravitreal bevacizumab has been shown by the authors
to cause regression of anterior segment neovascularization
and should be used as necessary to treat active iris neovas-
cularization in diabetic patients. This treatment should be
combined with adequate PRP for long-term control of VEGF
production and recurrence of neovascularization.
THE BARRIER CONCEPT
VEGF encounters sequential barriers in its anterior diffusion
en route to ocular egress through the trabecular meshwork
(Fig. 19.14). In nonoperated eyes, Neovascularization (NVE)
and Neovascularization of the disc (NVD) occur along the
back surface of the PVC. If vitrectomy has removed the PVC,
neovascularization occurs along the back surface of the AVC.
AVCFVP as first reported by one of the authors (S.C.) was
previously incorrectly thought to be due to “fibrovascular
ingrowth” from the sclerotomies.
In aphakic eyes or when present in high concentrations,
VEGF encounters the trabecular meshwork barrier, causing
NVG. Iris neovascularization serves to indicate the presence
of VEGF in the anterior segment. Trabecular meshwork neo-
vascularization, however, has a direct role in NVG. If a suc-
cessful filtering procedure is performed in a diabetic, aphakic,
vitrectomized eye, anterior segment neovascularization will
frequently disappear. This is analogous to the disappearance
of NVE and NVD after removal of the PVC by vitrectomy. In
these filtered cases, neovascularization occurs on the inside
of the bleb, which can be thought of as the final barrier.
If any neovascularization is seen in the postoperative
course, immediate intravitreal bevacizumab and PRP should
be performed. It is not advisable to wait for pressure eleva-
tion, which may obscure the view and lead to irreversible
NVG. While PRP may not affect the intraocular pressure, it
decreases fibrin release and hemorrhage from the iris vessels
that contribute to the phthisis process. Although on occa-
sion iris neovascularization will disappear spontaneously or
stabilize, it is better to treat all cases of iris neovasculariza-
tion with bevacizumab and PRP.
If the eye pressure exceeds 25 mm Hg, topical timolol
may be effective and can be used in combination with bri-
monidine, latanoprost, and topical carbonic anhydrase
inhibitors. If topical treatment cannot keep the pressure in
the mid-20s, glaucoma surgery may be required. Presumably
because of poor perfusion, diabetics have poor tolerance for
elevated pressure. Filtering procedures are effective in some
of these patients but have a tendency to cause hypotony with
resultant repeated intraocular bleeding.
Cyclodestructive procedures can be performed if the
patient cannot tolerate an operating room procedure. Cryo-
therapy on bare sclera, for 6 clock hours, 4 to 5 mm poste-
rior to the limbus to avoid the trabecular meshwork seems
to have best results. The treatments are held at 80°C for
1 minute. Although this can be quite effective in controlling
the pressure using a single treatment, many of these patients
go on to further fibrin release, cyclitic membrane formation,
fibrovascular proliferation, and phthisis bulbi. Transscleral
Nd-YAG laser cyclodestruction has largely replaced cyclocry-
otherapy, since ocular surface damage and inflammation are
greatly diminished, and does not require intraocular sur-
gery. When possible, endocyclophotocoagulation combined
with PRP is better than cryotherapy, transscleral laser, and
Fig u re 1 9 .1 4 ■ Seq uential b arriers
to anterior d iffusion of VEGF. Barriers
also act as sub strates for neovascular
and fibrovascular p roliferation.

ultrasonic cyclodestructive procedures because of reduced
inflammation, less pain, and better visual prognosis.
An t e rio r Vit re o us Co rt e x Fib ro vascular
Pro life rat io n
If AVCFVP develops, VEGF and other cytokines will cause
cellular migration and proliferation on the AVC. The mem-
brane causes a characteristic ring-like equatorial TRD fol-
lowed by total retinal detachment. This configuration can be
noted on ultrasound and must be operated on immediately.
The retrolental, retro-IOL, or cyclitic membrane should be
detected as early as possible by looking obliquely at the slit
lamp to ensure early treatment. Treatment requires ECL,
removal of the capsule and cyclitic membrane with applica-
tion of the bipolar diathermy to the resected edges, internal
fluid-air exchange, and long-term silicone surface tension
management. Extensive PRP is necessary to reduce further
neovascularization.
Rh e g m at o g e n o us Re t in al De t ach m e n t
Peripheral rhegmatogenous retinal detachment is rela-
tively infrequent in diabetic cases. If retinal detachment
occurs, it is usually related to retinal breaks missed at the
time of the original surgery or vitreous incarceration in the
wounds. Postvitrectomy retinal detachments usually can-
not be managed with scleral buckling alone. A greater suc-
cess rate is usually obtained by using a vitrectomy revision
approach with a search for residual traction or glial recur-
rence, internal drainage of SRF, internal fluid-air exchange,
focal endophotocoagulation, and air-gas or air-silicone
exchange.
Glial Re curre n ce
Epiretinal surgery, especially with peeling, can result in a recur-
rence of glial proliferation. Contrary to previous teaching, no
vitreous substrate (often called “scaffold”) is required for a glial
recurrence. Glial tissue can proliferate directly on the retinal
surface. Fibrin from ERM epicenters can constitute a bridge-
like substrate along which glial tissue can reproliferate.
Glial recurrences are managed with scissors delami-
nation and internal drainage of SRF, internal fluid-air
exchange, endophotocoagulation, and long-term silicone oil
for rhegmatogenous confinement if there is a rhegmatoge-
nous component. Frequently, retinectomy is required. These
membranes are tightly adherent to the retina and cannot be
treated with membrane peeling.
References
1. Klein R, Klein BE, Moss SE, et al. The Wisconsin epidemiologic
study of diabetic retinopathy. II. Prevalence and risk of diabetic
retinopathy when age at diagnosis is less than 30 yrs. Arch Oph-
thalmol 1984;102:520–526.
CHAPTER 19 ■ Diabetic Retinopathy 187
2. Klein R, Klein BE, Moss SE, et al. The Wisconsin Epidemio-
logic Study of Diabetic Retinopathy. III. Prevalence and risk of
diabetic retinopathy when age at diagnosis is 30 or more years.
Arch Ophthalmol 1984;102:527–532.
3. The Diabetes control and complications trial research group.
Progression of retinopathy with intensive versus conventional
treatment in the Diabetes Control and complications trial.
Ophthalmology 1995;102:647–661.
4. Chew EY, Klein ML, Ferris FL, et al. Association of elevated
serum lipid levels with retinal hard exudates in diabetic retin-
opathy. Early treatment of Diabetic retinopathy Study report
22. Arch Ophthalmol 1996;114:1079–1084.
5. Early treatment of diabetic retinopathy study research group.
Early photocoagulation for diabetic retinopathy: ETDRS report
9. Ophthalmology 1991;98:766–785.
6. Antcliff RJ, Spalton DJ, Stanford MR, et al. Intravitreal triamci-
nolone for uveitic cystoid macular edema: an optical coherence
tomography study. Ophthalmology 2001;109:765–772.
7. Pollack JA, ISIS Trial Group. Steroid for Diabetic Macular
Edema: The ISIS trial. Presentation in the Retina Annual Sub-
specialty Day, American Academy of Ophthalmology Annual
Meeting, Chicago, IL, October 2005.
8. Vinores SA, Derevjanik NL, Ozaki H, et al. Cellular mecha-
nisms of blood-retinal barrier dysfunction in macular edema.
Doc Ophthalmol 1999;97(3–4):217–228.
9. Ozaki H, Hayashi H, Vinores SA, et al. Intravitreal sustained
release of VEGF causes retinal neovascularization in rabbits
and breakdown of the blood-retinal barrier in rabbits and
primates. Exp Eye Res 1997;64(4):505–517.
10. Funatsu H, Yamashita H, Sakata K, et al. Vitreous levels of vas-
cular endothelial growth factor and intercellular adhesion mol-
ecule 1 are related to diabetic macular edema. Ophthalmology
2005;112(5):806–816.
11. Rosenfeld PJ, Fung AE, Puliafito CA. Optical coherence tomog-
raphy findings after an intravitreal injection of bevacizumab
(avastin) for macular edema from central retinal vein occlu-
sion. Ophthalmic Surg Lasers Imaging 2005;36(4):336–339.
12. The Diabetic Retinopathy Vitrectomy Study Research Group.
Early Vitrectomy for severe vitreous hemorrhage in diabetic
retinopathy: two year results of a randomized trial–DRVS
Study report 2. Arch Ophthalmol 1985;103:1644–1652.
13. Diabetic Retinopathy Vitrectomy Study Research Group. Early
vitrectomy for severe proliferative diabetic retinopathy in eyes
with useful vision: results of a randomized trial: DRVS report
3. Ophthalmology 1988;95:1307–1320.
14. Diabetic Retinopathy Study Research Group. Early vitrectomy
for severe vitreous hemorrhage. Four-year results of a random-
ized trial: DRVS report 5. Arch Ophthalmol 1990;108:958–964.
15. Doft BH, Blankenship GW. Single versus multiple treatment
sessions of argon laser panretinal photocoagulation for prolif-
erative diabetic retinopathy. Ophthalmology 1982;89:772–779.
16. The Diabetic Retinopathy Study Group. Preliminary report
on the effects of photocoagulation therapy. Am J Ophthalmol.
1976;81:383–396.
17. Machemer R, Buettner H, Norton EW, et al. Vitrectomy: a pars
plana approach. Trans Am Acad Ophthalmol 1971;75:813–820.
18. Arevalo JF, Maia M, Flynn HW Jr, et al. Tractional retinal detach-
ment following intravitreal bevacizumab (Avastin) in patients
with severe proliferative diabetic retinopathy. Br J Ophthalmol
2008;92(2):213–216. [Epub 2007 Oct 26.]
19. Cohen HB, McMeel W, Franks EP. Diabetic traction detach-
ment. Arch Opthalmol 1979;97:1268.
20. Flinn C, Charles S. The natural history of diabetic extramacu-
lar traction detachment. Arch Ophthalmol 1981;99:66.
21. Hykin PG, Gregson RM, Stevens JD, et al. Extracapsular cata-
ract extraction in proliferative diabetic retinopathy. Ophthal-
mology 1993;100:394–399.
188 SECTION IV ■ Disease-Sp ecific Managem ent
22. Holekamp NM, Shui YB, Beebe D. Lower intraocular oxygen
tension in diabetic patients: possible contribution to decreased
incidence of nuclear sclerotic cataract. Am J Ophthalmol
2006;141(6):1027–1032.
23. Hartley KL, Smiddy WE, Flynn HW Jr, et al. Pars plana vit-
rectomy with internal limiting membrane peeling for diabetic
macular edema. Retina 2008;28(3):410–419.
24. Blankenship GW. Preoperative iris rubeosis and diabetic vitrec-
tomy results. Ophthalmology 1980;87:176.
25. Scuderi 1, Blumenkranz M, Blankenship G. Regression of dia-
betic rubeosis iridis following successful surgical reattachment
of the retina by vitrectomy. Retina 1982;2:193.
26. Little HL. Rubeosis indis after vitrectomy for complications
of diabetic retinopathy. In: Little HL, Jack RL, Patz A, et al.,
eds. Diabetic retinopathy. New York: Thieme-Stratton, Inc.;
1983:315–340.
27. Lupinacci AP, Calzada JI, Rafieetery M, et al. Clinical out-
comes of patients with anterior segment neovascularization
treated with or without intraocular bevacizumab. Adv Ther
2009;26(2):208–216. [Epub 2009 Feb 20.]
28. Blankenship G, Cortez R, Machemer R. The lens and pars
plana vitrectomy for diabetic retinopathy complications. Arch
Ophthalmol 1979;97:1263.
29. Rice TA, Michels RG, Maguire MG, et al. The effects of
lensectomy on the incidence of ins neovascularization and
neovascular glaucoma after vitrectomy for diabetic retinopathy.
Am J Ophthalmol 1983;95:1.
30. Glaser BM. Relationship of cataract extraction and rubeosis in
diabetes mellitus. Ophthalmology 1983;90:819.
31. Blankenship GW, Machemer R. Pars plana vitrectomy for the
management of severe diabetic retinopathy, an analysis of
results five years after surgery. Ophthalmology 1978;85(6):553.
32. Michels RG. Vitrectomy for the complication of diabetic retin-
opathy. Arch Ophthalmol 1978;96(2):237.
33. Aaberg T. Clinical results in vitrectomy for diabetic traction
retinal detachment. Am J Ophthalmol 1979;88(2):246.
34. Diabetic retinopathy Study Research Group. Four risk factors
for severe visual loss in diabetic retinopathy: the third report
from the DRS. Arch Ophthalmol 1979:97:654–655.
35. Rice TA, Michels RG. Complications of vitrectomy. In: Little
HL, Jack RL, Patz A, et al., eds. Diabetic retinopathy. New York:
Thieme-Stratton, Inc.; 1983:315–340.
36. Michels RG. Vitreous surgery in proliferative diabetic retinopa-
thy. In: Shimizu K, Oosterhuis LA, eds. Acta XXIII Concilium
Ophthalmologicurm, Part 1, Amsterdam/Oxford: Excerpta Med-
ica; 1979:420.
37. Michels RG. Vitrectomy for complications of diabetic
retinopathy. Arch Ophthalmol 1978;96:237.
38. Charles S. Vitreous surgery for diabetic traction detachment.
Presented at Frontiers in Ophthalmology. Phoenix, Arizona,
February 18, 1982.
39. Rice TA, Michels RG. Long-term anatomic and functional
results of initially-successful vitrectomy for diabetic retinopa-
thy. Am J Ophthalmol 1980;90:297.
40. Blankenship CW. Stability of pars plana vitrectomy results
for diabetic retinopathy complications, a comparison of five-
year and six-month postvitrectomy findings. Arch Ophthalmol
1981;99:1009.
41. Rice TA, Michels RG, Palmer L. Late results of initially-successful
vitrectomy in diabetes. Dev Ophthalmol 1981;2:286.
42. Blankenship GW, Machemer R. Long-term diabetic vitrec-
tomy results, report of 10 year follow-up. Ophthalmology
1985;92:503.
43. Blankenship GW. Pars plana vitrectomy for diabetic retinopa-
thy, a report of 8 years’ experience. In: S. Karger AG, ed. Mod-
ern problems in ophthalmology. Basel, Switzerland: Medical and
Scientific Publishers; 1979:376–386.
44. Schachat AP, Oyakawa RT, Michels RG, et al. Complications
of vitreous surgery for diabetic retinopathy. II. Postoperative
complications. Ophthalmology 1983;90:522.
45. Glaser BM, D’Amore PA, Michels RG, et al. Demonstration of
vasoproliferative activity from mammalian retina. J Cell Biol
1980;84:298.
46. Glaser BM, D’Amore PA, Michels RG, et al. The demonstration
of angiogenic activity from ocular tissues: preliminary report.
Ophthalmology 1980;87:440.
47. Glaser BM, D’Amore PA, Lutty GA, et al. Chemical mediators
of intraocular neovascularization. Trans Ophthalmol Soc UK
1980;100:369.
48. Glaser BM, D’Amore PA, Michels RG. The effects of human
intraocular fluid on vascular endothelial cell migration: cor-
relation with intraocular neovascularization. Ophthalmology
1981;88:986.
49. Fett JW, Strydom DJ, Lobb RR, et al. Isolation and characteriza-
tion of angiogenin, an angiogenic protein from human carci-
noma cells. Biochemistry 1985;24:5480–5486.
50. Glaser BM, Campochiaro PA, Davis JL, et al. Retinal epithelial
cells release an inhibitor of neovasculanzation. Arch Ophthal-
mol 1985;103:1870.
51. Wolbarsht ML, Landers MB. The rationale of photocoagulation
therapy for proliferative diabetic retinopathy—a review and a
model. Ophthalmic Surg 1980;11:235.
C H AP TER
20
VENOUS OCCLUSIVE DISEASES
Very little progress has been made in the past
two decades in elucidating the precise patho-
genesis of central retinal vein occlusion (CRVO);
controversy still exists about the mechanism of
arterial compression of the vein as well as the role of throm-
bosis. There is less controversy about the pathogenesis of
branch retinal vein occlusion (BRVO); compression of the
vein by a branch artery is widely accepted, but again the role
of thrombosis is unclear.
Although many physicians order a battery of lab tests,
only homocysteine and anticardiolipin antibodies have been
shown in meta-analysis to be associated with retinal vein
occlusion (RVO) (1). Systemic hypertension, atherosclerotic
cardiovascular disease, and associated risk factors are risk
factors for RVO as well. The role of systemic disease does not
lead to a systemic treatment for RVO, although it emphasizes
the role of wellness-based approaches to prevention.
Systemic anticoagulation does not prevent RVO; many
patients presenting with vein occlusions are adequately anti-
coagulated (2). Similarly, there is no randomized clinical
trial evidence that systemic anticoagulation is effective in the
treatment of RVOs.
PHARMACOTHERAPY
Historical treatments of interest include topical administra-
tion of potassium iodide and pilocarpine, antihyperlipidemic
agents, anticoagulants, fibrinolytic agents, hyperosmotic
agents, carbogen inhalation, and isovolemic hemodilution.
These treatments have been demonstrated to be largely inef-
fective and therefore abandoned.
The modern era of pharmacotherapy for RVO began
in the 1990s with intravitreal triamcinolone acetonide
(Kenalog, Bristol Myers Squibb) injections to treat macu-
lar edema. Many observers noted marked decrease in reti-
nal thickness by optical coherence tomography but modest
increase in best corrected visual acuity presumably because
of ischemia-induced neuronal loss and permanent macular
structural changes. Prior to the widespread use of intravitreal
triamcinolone, it was thought that steroid glaucoma only
occurred in about 6% of the population referred to as steroid
responders. This notion was based on studies using topi-
cal 1.0% prednisolone acetate. Steroid glaucoma has been
shown to occur in approximately 30% of patients receiving
intravitreal triamcinolone (3); unfortunately, efficacy is pro-
portional to the steroid glaucoma and cataract side effects.
Prior to the use of intravitreal triamcinolone, it was thought
that steroid glaucoma was reversible after the drug was with-
drawn, but many patients receiving intravitreal triamcinolone
develop permanent glaucoma. Steroid cataracts occur in at
least 50% of the patients receiving intravitreal triamcinolone
(4). The authors do not use intravitreal triamcinolone for
RVO patients, but controversy exists among physicians still
using this therapy about whether the drug should be used in
patients with a personal or family history of glaucoma. Kena-
log use is off-label in the United States; the manufacturer
sent a warning letter to the retinal community in 2007. Some
surgeons filter and resuspend Kenalog, while others decant
the diluent to reduce the adverse effects related to preser-
vatives. Triesence (Alcon Laboratories, Ft. Worth, Texas), a
preservative-free triamcinolone, received FDA approval in
2007 and addresses the preservative toxicity issue.
Dexamethasone (Decadron) can be used to treat macu-
lar edema secondary to RVO but has a very short duration
of action, with an approximate half-life of 4 hours (5). Short
duration of action reduces the chances of cataract and steroid
glaucoma, but the only potential clinical benefit is to reduce
retinal thickness a few days prior to laser treatment. Sustained-
release dexamethasone (Ozurdex, Allergan), was recently
introduced into the US market.The duration of action is lon-
ger than intravitreal triamcinolone, with a single treatment
producing statistically significant best corrected visual acuity
improvements 90 days after treatment. Kuppermann et al.
(6) have shown that the dexamethasone molecule is less toxic
than triamcinolone without preservative when tested against
human retinal neuronal cells, glial cells, and retinal pigment
epithelial cells in tissue culture (7,8).
189
190 SECTION IV ■ Disease-Sp ecific Managem ent
An intravitreal fluocinolone 0.59 mg sustained-release
device (Retisert, Bausch & Lomb, St. Louis, Missouri) is FDA
approved for posterior uveitis and has been used for the treat-
ment of macular edema by some surgeons. Unfortunately,
the Retisert produces a 90% glaucoma incidence, and at least
30% of the patients ultimately require a glaucoma filtering
procedure (9). This level of adverse effects is simply not jus-
tifiable when treating macular edema secondary to RVO.
Nonsteroidals are effective when used topically for
macular edema after cataract surgery. Postoperative macular
edema (Irvine-Gass syndrome) is inflammatory in nature,
while macular edema in RVO patients is primarily driven
by vascular endothelial growth factor (VEGF) and increased
hydrostatic pressure. There is little direct evidence that
inflammation is a significant factor in macular edema, but the
observation that intravitreal steroids reduce macular edema
in RVO suggests a role for inflammatory mediators. Ketoro-
lac tromethamine 0.5% (Acular, Allergan, Irvine, California)
and diclofenac 0.1% (Voltaren, Novartis Ophthalmics) sup-
press the anterior segment inflammation that causes postop-
erative macular edema but have minimal direct effect on the
macula. Bromfenac 0.09% (Xibrom, Ista Pharmaceuticals,
Irvine, California) and nepafenac (Nevanac, Alcon Labora-
tories, Ft. Worth, Texas) have been shown to reduce mac-
ular edema secondary to retinal vascular disease in several
observational series as well as to be effective for postcataract
surgery of macular edema (10,11). Nepafenac is a prodrug
that has been shown to produce significant drug levels at the
macula in phakic and pseudophakic eyes (12).
Proliferative diabetic retinopathy, RVOs, and retinopa-
thy of prematurity all produce high levels of intravitreal
VEGF (Lloyd Paul Aiello), which is not the case for retinal
detachments or proliferative vitreoretinopathy (PVR) (13).
VEGF increases vascular permeability and is a powerful
angiogenic signaling ligand. Anti-VEGF therapy is effective
for macular edema as well as neovascularization secondary
to RVO and diabetic retinopathy. Bevacizumab (Avastin)
is a safe and effective anti-VEGF compound, has low cost,
and is widely used for retinal vascular disorders worldwide,
albeit off-label in the United States. Although some physi-
cians have expressed concerns about long-term VEGF sup-
pression, this has not proven to be a problem in the 6 to
7 years of anti-VEGF compounds use. Initially, there was
concern that an anti-VEGF antibody would cause uveitis,
but this has not proven to be a problem either. The typical
intravitreal dose, 1.25 mg, is approximately 1/500 the sys-
temic dose used for cancer. Cancer patients using this drug
every 2 week intravenously with a much higher dose than
is used in the eye have a doubling of the thromboembolic
event rate. Although this has been a matter of concern, there
is no evidence of increased thromboembolic event rates after
intravitreal injection.
Most physicians inject Avastin approximately every
2 months based on OCT measurements of macular thick-
ness. Intravitreal Avastin for BRVO demonstrates visual
acuity improvements from 20/200 at baseline to 20/100 at
3 months, with a 25% improvement in central macular thick-
ness (14). Case reports of intravitreal Avastin for CRVO have
shown potential benefit of this treatment in short-term follow-
up (15). Ranibizumab (Lucentis) is the only FDA-approved
anti-VEGF compound other than the minimally effective
pegaptanib sodium (Macugen), but neither is approved for
the RVO indication. The RVO study arm of the Ranibizumab
for Edema of the Macula in Diabetes (READ) demonstrated
a doubling of the visual angle after three monthly injections
(16). The authors have observed excellent results from using
intravitreal Avastin on all BRVOs and CRVOs at the initial
visit and every month until all intraretinal hemorrhage, axo-
plasmic stasis, and edema have regressed as recommended
by Richard Spaide (17). Often these cases have complete
resolution of the BRVO or CRVO appearance even without
evidence of collaterals. It is of interest that VEGF creates a
CRVO picture in animal models (17).
Intracameral and intravitreal Avastin injection results in
rapid and dramatic regression of iris neovascularization sec-
ondary to CRVO; effects are seen within 48 hours and last at
least 4 weeks (18, 19). Typically, if no cicatricial closure of the
iridocorneal angle was seen on gonioscopy, there is a marked
decrease in intraocular pressure as well, presumably because
vessels on the surface of the trabecular meshwork regress
and the viscosity of the aqueous humor decreases markedly
(much less flare). Avastin and presumably any anti-VEGF
compound must be used with panretinal photocoagulation
for sustained effect on anterior segment and retinal neovas-
cularization excluding choroidal neovascular membranes.
Avastin is very effective prior to focal, grid, or sector
photocoagulation for RVOs to decrease retinal thickness and
decrease laser fluence requirements. The anti-VEGF agent
should be injected approximately 1 week prior to laser.
LASER THERAPY
Focal, sector, or light grid laser photocoagulation can be
used for macular edema secondary to BRVO, hemi–vein
occlusion, or CRVO. Treatment with Avastin approximately
1 week prior to laser has dramatically improved results in
the author’s opinion, although there is no randomized con-
trolled trial evidence at this time.
Panretinal photocoagulation is very effective in treating
iris and retinal neovascularization secondary to hemi-RVO
or CRVO but, as pointed out above, is best used combined
with anti-VEGF therapy. Steve Schwartz has coined the term
“targeted retinal photocoagulation (TRP)” to describe the
notion of ablation of angiographically proven midperiph-
eral ischemia and transition zones to treat diffuse diabetic
macular edema (20). Wessing and Shimizu have advocated
this concept for years. The authors believe that this concept
is applicable to RVOs as well.
The PASCAL laser from OptiMedica in Santa Clara,
California, typically uses 30-ms pulses to reduce thermal
diffusion, anteriorly to the nerve fiber layer, laterally causing

retinal pigment epithelial “creep,” and posteriorly to the
choroid, causing pain. Although this concept has most often
been applied to diabetic macular edema, the authors believe
that it is applicable to RVOs as well.
High-power laser pulses were shown by McAllister
and Constable (21) to create retinal-choroidal anastomo-
sis, which seemed to reduce macular edema. Initially oth-
ers had enthusiasm for this approach, but many physicians
noted the development of large disciform scars, vitreous
hemorrhages, and retinal detachments and abandoned this
procedure. McAllister and Constable advocated 100-ms
pulses directed at retinal veins; subsequently, Leonard
(22) proposed 1-second burns and avoiding the veins and
reported a decreased incidence of these complications.
VITRECTOMY-BASED THERAPY
Vitrectomy for vitreomacular traction was introduced by
Hilel Lewis (23) for diabetic macular edema but is effec-
tive for carefully selected cases of RVO as well. Fluorescein
angiography to evaluate macular capillary nonperfusion
is essential, especially on CRVO cases. The presence of
an afferent pupillary defect is a relative contraindication
to pars plana vitrectomy. Optical coherence tomography,
especially spectral domain optical coherence tomography
such as Heidelberg Spectralis, has resulted in diagnosing
a much higher incidence of vitreomacular traction than
is noted even with contact lens examination. Timing is
important in these cases; vitreomacular traction with
macular cysts has better visual outcomes if surgery is per-
formed after a one-month trial of Avastin and nepafenac.
The authors recommend internal limiting membrane
removal in these cases to ensure complete elimination of
vitreomacular traction.
Vitrectomy in patients without vitreomacular traction is
effective in diabetic macular edema. An increasing number
of surgeons believe that pars plana vitrectomy is effective
for macular edema in RVO patients as well. The mechanism
in both disease processes is almost certainly similar because
pars plana vitrectomy permanently increases oxygen tension
by 7 to 12 mm Hg as shown in the laboratory by Stefansson
(24), and later Chang, and clinically by Holekamp (25–26).
Increased oxygen tension is probably the mechanism by
which “sheathotomy” and radial optic neurotomy seem to
be effective.
Branch vein decompression was developed by the
first author and subsequently rediscovered and renamed
“sheathotomy” by Opremcak (27). Seitz and Green have
both emphasized that there is no sheath; the artery and vein
have a common adventitia (28–29). The author abandoned
this procedure after 15 cases but, in retrospect, should
have done a randomized clinical trial. Most surgeons have
now abandoned this procedure as well. Perceived visual
improvement is probably due to the natural history as well
as increased oxygen tension from pars plana vitrectomy.
CHAPTER 20 ■ Venous Occlusive Diseases 191
Furthermore, the results of this surgery are difficult to
evaluate because the studies were small and because of the
lack of analysis between time after onset of the occlusion and
perfusion status prior to surgery. Figueroa has demonstrated
that capillary reperfusion does not improve after branch vein
decompression (30).
Radial optic neurotomy for CRVO was developed by
Opremcak (31), and a similar procedure, laminar puncture,
was developed by D’Amico (32). D’Amico performed a care-
ful internal review board–controlled trial and determined
that the procedure was ineffective, and again there is no
randomized clinical trial evidence of efficacy. In addition to
these issues, there is a significant complication rate and the
procedure has been largely abandoned.
References
1. Janssen MC, den Heijer M, Cruysberg JR, et al. Retinal vein
occlusion: a form of venous thrombosis or a complication
of atherosclerosis? A meta-analysis of thrombophilic factors.
Thromb Haemost 2005;93(6):1021–1026.
2. Mruthyunjaya P, Connor T, Wirostko WJ, et al. Central retinal
vein occlusion in patients on long-term coumadin anticoagula-
tion. Retina 2006;26(3):285–291.
3. Jonas JB, Degenring R, Kreissig I, et al. Intraocular pressure
elevation after intravitreal triamcinolone acetonide injection.
Ophthalmology 2005;112:593–598.
4. Thompson JT. Cataract formation and other complications of
intravitreal triamcinolone for macular edema. Am J Ophthalmol
2006;141(4):629–637.
5. Kwak HW, D’Amico DJ. Evaluation of the retinal toxicity and
pharmacokinetics of dexamethasone after intravitreal injec-
tion. Arch Ophthalmol 1992;110:259–266.
6. Kuppermann BD, Blumenkranz MS, Haller JA, et al. Random-
ized controlled study of an intravitreous dexamethasone drug
delivery system in patients with persistent macular edema.
Arch Ophthalmol 2007;125:309–317.
7. Nabih M, Peyman GA, Tawakol ME, Naguib K. Toxicity of
high dose intravitreal dexamethasone. Int Ophthalmol 1991:15:
234–235.
8. Narayanan R, Mungcal JK, Kenney MC, et al. Toxicity of tri-
amcinolone acetonide on retinal neurosensory and pigment
epithelial cells. Invest Ophthalmol Vis Sci 2006;47:722–728.
9. Jaffe GJ, Martin D, Callanan D, et al. Fluocinolone Acetonide
Uveitis Study group. Fluocinolone acetonide implant (Retisert)
for noninfectious posterior uveitis; thirty-four-week results
of a multicenter randomized clinical study. Ophthalmology
2006;113:1020–1027.
10. Gross JG. Xibrom (bromfenac) reduces cystoid macular edema
associated with vein occlusions, macular epiretinal membrane,
diabetic retinopathy and age related macular degeneration.
Poster F-3 presented at: Annual Meeting of the Retina Society.
Boston, MA, USA, September 27–30, 2007.
11. Hariprasad SM, Callanan D, Gainey S, et al. Cystoid and dia-
betic macular edema treated with nepafenac 0.1%. J Ocul Phar-
macol Ther 2007;23(6):585–590.
12. Walters T, Raizman M, Ernest P, et al. In vivo pharmacokinet-
ics and in vitro pharmacodynamics of nepafenac, amfenac,
ketorolac, and bromfenac. J Cataract Refract Surg 2007;33:
1539–1545.
13. Aiello LP, Avery RL, Arrigg PG, et al. Vascular endothelial
growth factor in ocular fluid of patients with diabetic retin-
opathy and other retinal disorders. N Engl J Med 1994;331:
1480–1487.
192 SECTION IV ■ Disease-Sp ecific Managem ent
14. Rabena M, Pieramici DJ, Castellarin AA, et al. Intravitreal
bevacizumab (Avastin) in the treatment of macular edema
secondary to branch retinal vein occlusion. Retina 2007;27:
419–425.
15. Iturralde D, Spaide FR, Meyerle CB, et al. Intravitreal beva-
cizumab (Avastin) treatment of macular edema in central
retina vein occlusion: a short term study. Retina 2006;26:
279–284.
16. Campochiaro PA, Hafiz G, Shah SM. Ranibizumab for macular
edema due to retinal vein occlusions; implication of VEGF as a
critical stimulator. Mol Ther 2008.
17. Ferrara DC, Koizumi H, Spaide RF. Early bevacizumab treat-
ment of central retinal vein occlusion. Am J Ophthalmol
2007;144:864–871.
18. Grisanti S, Biester S, Peters S, et al. Intracameral bevacizumab
for iris rubeosis. Am J Ophthalmol 2006;142:158–160.
19. Iliev ME, Dommig D, Wolf-Schnurrbursch U, et al. Intravitreal
bevacizumab (Avastin) in the treatment of neovascular glau-
coma. Am J Ophthalmol 2006;142:1054–1056.
20. Schwartz S. Ultra-wide angle angiographically targeted retinal
photocoagulation (TRP) combined with anti-VEGF pharmaco-
therapy for the treatment of retinal vascular macular edema.
The Macula Society 30th Annual Scientific Program. June 2,
2007.
21. McAllister IL, Constable IJ. Laser-induced chorioretinal venous
anastomosis for non-ischemic central retinal vein occlusion:
evaluation of the complications and their risk factors. Am J
Ophthalmol 1998;126:219–229.
22. Leonard BC, Coupland SG, Kertes PJ, Bate R. Long-term
follow-up of a modified technique for laser-induced chorioret-
inal venous anastomosis in nonischemic central retinal vein
occlusion. Ophthalmology 2003;110(5):948–954.
23. Lewis H, Abrams GW, Blumenkranz MS, Campo RV. Vitrectomy
for diabetic macular traction and edema associated with poste-
rior hyaloidal traction. Ophthalmology 1992;99(5):753–9.
24. Stefansson E. The therapeutic effects of retinal laser treatment
and vitrectomy. A theory based on oxygen and vascular physi-
ology. Acta Ophthalmol Scand 2001;79(5):435–440.
25. Barbazetto IA, Liang J, Chang S, et al. Oxygen tension in the
rabbit lens and vitreous before and after vitrectomy. Exp Eye
Res 2004;78(5):917–924.
26. Holekamp NM, Shui YB, Beebe DC. Vitrectomy surgery increases
oxygen exposure to the lens: a possible mechanism for nuclear
cataract formation. Am J Ophthalmol 2005;139:302–310.
27. Opremcak EM, Bruce RA. Surgical decompression of branch
retinal vein occlusion via arteriovenous crossing sheathotomy.
Retina 1999;19:1–5.
28. Seitz R. Die Netzhautgefäße. Stuttgart, Germany: Georg Thieme
Verlag, 1962.
29. Green WR, Chan CC, Hutchins GM, Terry JM. Central retinal
vein occlusion: a prospective histopathologic study of 29 eyes
in 28 cases. Trans Am Ophthalmol Soc 1981;79:371–422.
30. Figueroa MS, Torres R, Alvarez MT. Comparative study
of vitrectomy with and without vein decompression for
branch retinal vein occlusion: a pilot study. Eur J Ophthalmol
2004;14(1):40–47.
31. Opremcak EM, Bruce RA, Lomeo MD, et al. Radial optic neu-
rotomy for central retinal vein occlusion: a retrospective pilot
study of 11 consecutive cases. Retina 2001;21:408–415.
32. D’Amico DJ, Lit ES, Viola F. Lamina puncture for central reti-
nal vein occlusion: a pilot study. Arch Ophthalmol 2006;124:
972–977.
C H AP TER
21
TREATMENT OF CHOROIDAL
NEOVASCULAR MEMBRANES
The management of choroidal neovascular
(CNV) membranes in age-related macular
degeneration (AMD), ocular histoplasmosis
syndrome, myopia, and uveitis has advanced
greatly in the past few years. In general terms, there has been
a progression from nihilistic observation to laser ablation
to submacular surgery, photodynamic therapy (PDT) with
and without intravitreal steroids to modern anti–vascular
endothelial growth factor (VEGF) therapy.
Many potential treatments have been promoted by
positive pilot studies ultimately followed by randomized
clinical trials showing no significant efficacy. Examples of
this sequence include alpha one interferon, thalidomide,
brachytherapy, anecortave acetate, transpupillary thermo-
therapy for occult CNV, and external beam radiation.
Pegaptanib sodium (Macugen) and PDT (Visudyne)
therapy resulted in modest slowing in the rate of visual loss
but rarely resulted in visual improvement. Ranibizumab
(Lucentis), a humanized Fab fragment of an anti-VEGF
antibody, was the first medication to achieve reliable visual
gains in a significant percentage of patients. Visual results
in several pivotal multicenter, randomized clinical trials
were simply outstanding compared to the results previously
achieved with earlier therapies: 95% of patients avoided
moderate visual loss (2 ETDRS lines or 25 letters), 30% to
40% of the patients experienced improved vision, and 75%
achieved true stability. Phillip Rosenfeld introduced the con-
cept off-label use of bevacizumab (1,2) (Avastin), a related,
full-length anti-VEGF antibody, as a significantly less expen-
sive alternative to Lucentis. Retrospective, prospective, and,
more recently, small randomized trials and a large amount
of anecdotal experience using Avastin have demonstrated
similar outcomes to Lucentis. The National Eye Institute–
sponsored Comparisons of Age-Related Macular Degenera-
tion Treatments Trials (CATT) is a head-to-head comparison
of Lucentis to Avastin and has both protocol and PRN treat-
ment arms for both drugs.
Most patients are currently treated with monotherapy
with either bevacizumab or ranibizumab. There are patients
were combination with other therapies (like laser) can be
entertained.
FOCAL THERMAL LASER ABLATION
The goal of laser therapy is to achieve complete ablation of
the CNV complex. Laser therapy produces concurrent over-
lying retinal ablation with an immediate scotoma.
Photocoagulation of juxtafoveal lesions has been shown
to produce greater than 50% recurrence rates, frequently
on the foveal side (3). Many patients present with subfoveal
lesions; therefore, laser treatment of these lesions was inves-
tigated in the Macular Photocoagulation Study (MPS) as well.
It was shown that treated patients suffer an immediate loss of
greater than three lines of vision but have slightly better vision
after 18 months compared to patients who did not receive laser
therapy. Few physicians ever became comfortable with causing
immediate loss of central vision, and this treatment is no lon-
ger performed since Lucentis and Avastin became available.
Laser treatment can still be considered for extrafoveal
CNVs if they are at a substantial distance from the fovea and
the total size of the lesion (and the secondary scotoma) is small.
In other words, the smaller and more peripheral the lesion is,
the more amenable it is to laser therapy. It is important for the
lesion to have clearly defined borders on angiography to guar-
antee that all the CNV is ablated whenever laser is preformed.
The advantage of laser above anti-VEGF therapy is
single-session treatment with long-term control. The authors
often recommend injection of Avastin or Lucentis concur-
rently with focal laser therapy for extrafoveal lesions with the
hopes of decreasing recurrence rates.
PREVENTION STRATEGIES
Die t ary Sup p le m e n t at io n
Oral antioxidants such as zinc, beta-carotene, zeaxanthin,
lutein, vitamin A, vitamin E, selenium, and others have been
193
194 SECTION IV ■ Disease-Sp ecific Managem ent
advocated for the prevention of AMD and treatment of early
stages of AMD or atrophic (dry) AMD. The Age Related Eye
Study showed that zinc plus antioxidants reduced AMD
progression risk by 26% in patients with intermediate AMD.
The Beaver Dam Eye Study looked at the dietary and smok-
ing history of individuals with AMD compared to matched
normals and concluded that spinach and, to a lesser extent,
other dark leafy green vegetables slightly reduce the progres-
sion to AMD (4). Smoking was found to have a doubling
effect on the incidence of AMD (5). The authors predict that
some specific antioxidants may be proven to have minor
value if given to individuals at genetic risk for AMD (drusen).
It is likely that this effect will be greatest in patients who do
not eat dark leafy green vegetables, other vegetables, whole
grains, and fruits. Beta carotene has been shown to increase
the risk of lung cancer in smokers in a Swedish study (6). At
this time, the authors counsel the children of patients with
AMD, drusen patients, and patients with early AMD to stop
smoking and eat spinach and other dark leafy green veg-
etables at least five times per week.
The Age Related Eye Disease Study (AREDS) 2 Study is
underway, which is structured to assess the effects of oral sup-
plementation of macular xanthophylls (lutein and zeaxanthin)
and/or long-chain omega-3 fatty acids (docosahexaenoic acid
and eicosapentaenoic acid) on the progression to advanced
AMD. An additional goal of the study is to assess whether
forms of the AREDS nutritional supplement with reduced zinc
and/or no beta-carotene work as well as the original supple-
ment in reducing the risk of progression to advanced AMD.
Plasm ap h e re sis
A randomized trial of plasmapheresis did not reach the pri-
mary endpoint, and there is no scientific rationale for this
treatment. Unfortunately, many patients have been treated
using this unproven, expensive treatment.
Lase r Tre at m e n t o f Druse n
Laser treatment of drusen has been shown to cause faster dis-
appearance of drusen than the spontaneous disappearance
rate (7–9). The reader is reminded that drusen are “associ-
ated with” AMD; they do not “cause” AMD. Threshold treat-
ment has been shown to result in higher CNV rates in two US
studies now investigating subthreshold treatment (10,11).
Two large randomized clinical trails failed to show a benefit
of subthreshold laser treatment (Friberg, Fine CAPT).
Ph o t o d yn am ic Th e rap y
PDT was arguably the first therapy other than focal laser
therapy available for the treatment of choroidal neovascu-
larizations (12–20). Since the introduction of anti-VEGF
therapy, the use of PDT has decreased dramatically. There are
still physicians who recommend the use of PDT for certain
patients, often in combination with anti-VEGF therapy. The
use of PDT as monotherapy has largely been abandoned for
choroidal neovascularizations.
The authors currently do not recommend PDT therapy
in the management of CNVs, since they aim to avoid any
further injury to the choriocapillaris, which is often dam-
aged with PDT therapy.
ANTI-VEGF THERAPY STRATEGIES
The injection frequency selected for the pivotal Lucentis trials
was based on preclinical pharmacokinetic data and assump-
tions about the length of time untreated CNV membranes
remain active. This strategy resulted in patients having an
average of 21 injections over 24 months. Rosenfeld introduced
the concept of treat and observe based on time domain opti-
cal coherence tomography (OCT) evidence of subretinal fluid
(SRF), Pigment epithelial detachment (PED), or retinal edema
(Prospective OCT Imaging of Patients with Neovascular AMD
Treated with Intraocular Ranibizumab (PrONTO) Study).
More recently, Spaide et al. introduced the concept of treat and
extend; generally, the notion is to give a “booster shot” after
the CNV is not leaking based primarily on OCT evidence and
a 6-week rather than a 1-month appointment. If the lesion is
not leaking at the 6-week appointment, another injection is
given and the patient is scheduled to return in 8 weeks. If the
lesion remains inactive at this point, no injection is given and
follow-up is done at 1- to 2-month intervals. More recently,
many physicians, including the authors, use clinical decision
making and include vision, visual complaints, exudates, and
hemorrhage in addition to spectral domain OCT (Spectralis)
to determine the need for retreatment. Fluorescein angiogra-
phy (confocal using Spectralis), formerly the primary tool in
determining the need for treatment, is still used anytime there
is a visual complaint or decreased vision and SD OCT fails to
reveal the etiology. Indocyanine green angiography is used
less often but remains a valuable tool.
COMBINATION THERAPY
Combination therapy is an appealing concept but is appro-
priate in some instances and not in others. Combination che-
motherapy in oncology is utilized because the agents have
narrow windows between effective and toxic drug levels as
well as to provide multiple barriers to the evolution of cancer
cells. Combination therapy for infectious disease is utilized
in severe infections when the infectious agent has not been
identified and delayed treatment would produce bad out-
comes. Combination therapy for infectious disease creates
multiple barriers for the evolution of the infectious agent but
unfortunately leads to higher incidence of resistance.
Focal and/or panretinal photocoagulation laser plus
anti-VEGF therapy is very effective in diabetic retinopathy
and can be broadly defined as combination therapy. Topical
nonsteroidal (Nevanac) therapy in combination with laser and
anti-VEGF compounds is effective for diabetic macular edema
 CHAPTER 21
because of multiple mechanisms: VEGF and inflammation.
Combination therapy has been advocated for CNV, but in the
authors’ view, it has very little place with current therapeu-
tic options in view of the extraordinary success of Lucentis
and Avastin monotherapy. PDT requires intravitreal steroids,
which results in an approximately 30% incidence of steroid
glaucoma and approximately 90% incidence of cataract while
only modestly reducing the need for anti-VEGF therapy.
Induction with Lucentis or Avastin followed by main-
tenance with pegaptanib sodium (Macugen) is based on the
notion that pan-isoform suppression (Lucentis, Avastin) may
result in neuronal loss but suppressing only the so-called
pathological 165 isoform would not. Long-term pan-isoform
suppression has not been demonstrated to have risk in
patients.
Vitrectomy to enable epiretinal radiation with a Stron-
tium 90 probe (NeoVista) is currently being studied but
requires anti-VEGF therapy to be effective, produces cata-
ract in 90% of the AMD patients because of the vitrectomy,
reduces the effectiveness of the anti-VEGF agent by eliminat-
ing the inherent sustained-release action of vitreous, creates
retinal detachment risk, and potentially will produce radia-
tion retinopathy.
Vit re ct o m y Surg e ry fo r Sub m acular
He m o rrh ag e Disp lace m e n t
Submacular hemorrhages have been surgically removed
since the early 1990s. These cases have widely varying
outcomes based on the underlying pathology and surgical
trauma involved (21–23). Some investigators have injected
tissue plasminogen activator (tPA) under the retina at the
time of surgery and wait approximately 45 minutes for
apparent liquefaction of the clot before aspiration (24–27).
Others inject tPA into the vitreous approximately 24 hours
before surgical removal in an attempt to liquefy the blood
clot (28). It has been reported that tPA can be injected into
the vitreous cavity in the office followed by a gas injection
and face-down positioning (29). Hilel Lewis has shown that
tPA does not cross the primate retina (30). This less invasive
method was thought to displace the blood away from the
macula; however, there is no randomized clinical trial evi-
dence of efficacy. The authors no longer recommend intra-
vitreal tPA.
The authors have positive experience with a technique
developed by Hilel Lewis to displace submacular hemor-
rhage from underneath the fovea, with the aim of decreas-
ing foveal photoreceptor damage caused directly from the
submacular hemorrhage. Thin submacular hemorrhages
(those that do not have mass effect and do not cause foveal
elevation) do not require surgery and can be managed with
anti-VEGF monotherapy, as the hemorrhage will invariably
clear spontaneously and the final visual outcomes are not
worsened by the presence of thin submacular hemorrhage.
Massive submacular hemorrhages (those that cause
macular elevation from submacular mass effect), on the
■ Treatm ent of Choroidal Neovascular Mem branes 195
other hand, do cause foveal damage and are amenable to
hemorrhage displacement. The authors recommend per-
forming spectral domain OCT on all of these patients,
since it is often difficult to determine the presence of
an underlying retinal pigment epithelium (RPE) detach-
ment or sub-RPE hemorrhage that could give the clini-
cal impression of massive macular elevation. If the OCT
demonstrates a subfoveal RPE detachment or sub-RPE
hemorrhage with a thin overlying subretinal hemorrhage,
the authors do not recommend surgical displacement and
treat with anti-VEGF monotherapy as described previ-
ously. Only those patients in whom the OCT demonstrates
thick subfoveal hemorrhage should be treated with surgi-
cal displacement.
Another important clinical consideration for surgical
displacement is the duration between first symptoms and
surgery. The natural history of massive submacular hemor-
rhages is first to appear with dark red submacular blood that
changes in color over the course of weeks to a slate gray
appearance as the blood is dehemoglobinized. By the time
the hemorrhage appears gray, the clot is frequently too dense
and hard to be amenable for displacement, and often there is
already photoreceptor damage. Given these considerations,
it is important to proceed with surgical displacement within
2 weeks of initial symptoms, while the hemorrhage is still
dark red in clinical appearance.
The surgical sequence is as follows: A core vitrectomy
is performed with 25-gauge sutureless technique, a poste-
rior vitreous separation is carefully created if not present,
and tPA is injected into the clot using a 38-gauge polyamide
cannula (developed for macular translocation surgery). The
tPA can be injected using the viscous fluid injection system
of the Accurus and Constellation vitrectomy machines to
allow the surgeon to use foot control of injection pressure.
The VFC system should be set to a maximum of 12 mm Hg
to prevent overpressuring the subretinal space. A slow, con-
trolled injection is then achieved and a pocket of subretinal
tPA resembling an exudative retinal detachment is created
underneath the macula and further expanded below the
inferior vascular arcade to create a path for the hemorrhage
to be easily displaced.
Total fluid-air exchange is performed, and the patient is
kept in the supine position for 45 minutes after surgery. The
patient is then placed in the seated position so that the air
bubble will displace the liquefied blood clot inferiorly. Verti-
cal position should be maintained for 48 to 72 hours to keep
the displaced hemorrhage away from the fovea while the
injected SRF is reabsorbed. Timing is important; blood clots
over 2 weeks old should not be operated. Surgical aspiration
of subretinal blood results in damage to the photoreceptor
outer segments and Retinal Pigment Epithelium (RPE) apical
processes and should be avoided. Intravitreal Avastin should
be used to treat underlying CNV membranes after successful
displacement of the blood.
Surgical success should be defined as only thin subfo-
veal hemorrhage on postoperative day one.
196 SECTION IV ■ Disease-Sp ecific Managem ent
SUBMACULAR SURGERY
AND MACULAR TRANSLOCATION
In 1991, Thomas and Kaplan reported submacular surgery
for subretinal neovascular membranes secondary to pre-
sumed ocular histoplasmosis syndrome (31). Subsequently,
these authors and many other surgeons applied their tech-
nique to idiopathic submacular membranes as well as
membranes secondary to AMD and many other disease pro-
cesses. In 1993, the principal author developed a simplified
approach to submacular surgery, which will be described
herein along with a discussion of indications in the context
of the negative results of the Submacular Surgery Trial and
the advent of effective anti-VEGF therapy.
Although the initial article referred to removal of “neo-
vascular” membranes, it is actually a fibrovascular scar that
is removed. It is common for the scar (membrane) to be
significantly larger than the “net” seen on the angiogram
presumably because of vascular regression in some areas
(Fig. 21.1).
INDICATIONS FOR SUBMACULAR
SURGERY FOR SUBFOVEAL MEMBRANES
Presumed photoreceptor and RPE viability are an absolute
requirement when considering submacular surgery. The
membrane acts as a diffusion barrier between the photore-
ceptors and the RPE and leads first to degeneration of the
outer segments and ultimately the inner segments. Free iron
from blood in the subretinal space for an extended period of
time is toxic to the photoreceptor inner segments. Cases with
Fig ure 2 1 .1 ■ It is com m on for the fibrovascular scar (neovas-
cular m em brane) to b e significantly larger than the “net” seen
on angiography.
membranes present for over 6 months have almost no chance
of visual improvement after membrane removal because of
irreversible photoreceptor damage. Type I membranes seen
in AMD cases are under the RPE and cannot be removed
without removing the overlying RPE. Type II membranes,
as seen in idiopathic and histoplasmosis cases, are on the
anterior surface of the RPE and can have a reasonable prog-
nosis. Anti-VEGF therapy is the first-line treatment for all
CNV membranes, but submacular surgery can be effective
for larger subfoveal, nonleaking membranes.
AMD cases have a very poor visual prognosis after mem-
brane removal. Approximately 20% of the AMD patients
have improved vision, 60% have unchanged vision, and
20% have worse vision after submacular surgery. Poor vision
after submacular surgery in these patients can be explained
by several factors: (a) diffuse, ongoing RPE disease, (b) high
incidence of recurrences, (c) photoreceptor and RPE damage
from scar and hemorrhage, (d) surgical removal of the RPE,
and (e) absence of the choriocapillaris. The membrane must
be located between the retina and the RPE, not under the
RPE, in order to permit surgery without inadvertent removal
of a segment of RPE. Most, if not all, AMD cases have type I
membranes, which are under the RPE. The authors have not
performed submacular surgery on patients with AMD since
1994. Determination that the patient has AMD is not made
by age alone but by the presence of drusen in the affected
and/or other eye. A 50-year-old patient with advanced
drusen has “AMD,” while a 70-year-old patient with a uni-
lateral membrane and no drusen in either eye probably has
an idiopathic CNV membrane. The Submacular Surgery
Trial results were released in November 2004 and did not
demonstrate a benefit for any AMD cases with vision better
than 20/100.
Idiopathic cases have the best prognosis, while his-
toplasmosis is the most frequent indication in the cen-
tral United States. The rare myopic patient with a recent,
medium to large well-defined lesion can benefit from sub-
macular surgery. Selected angioid streaks and trauma cases
are candidates for surgery as well.
SURGICAL SEQUENCE
A core vitrectomy is performed (Fig. 21.2) to facilitate pas-
sage of the instruments and to enable fluid-air exchange.
An aggressive attempt to create a posterior vitreous detach-
ment (PVD) in submacular surgery cases is unnecessary and
a potentially dangerous step. Many surgeons believe that
forceful creation of a PVD is a required step, while in fact it
may increase the chance of retinal detachment and/or dam-
age the optic nerve. The principal author has had no rheg-
matogenous detachments or retinotomy-related problems in
a series of over 250 cases with no aggressive attempt to make
a PVD.
Thomas and Kaplan recommended injection of balanced
salt solution (BSS) under the retina to create a small retinal
 CHAPTER 21 ■ Treatm ent of Choroidal Neovascular Mem branes 197

Fig ure 2 1 .2 ■ Core vitrectom y

is p erfo rm ed with n o sp ecific
attem p t to m ake a PVD.

detachment before removing the membrane (1). The senior
author (S.C.) observed that this step occasionally results in an
acute hydraulic foveal hole. The author has learned that other
surgeons have experienced this problem as well. For this rea-
son, the principal author has omitted the BSS injection step
since 1992. In addition to preventing hydraulic rupture of
the fovea, omitting BSS injection has the added advantage of
preventing damage from shearing photoreceptors from the
RPE in an area larger than the membrane.
A very small retinotomy is made with the 25-gauge
microvitreoretinal (MVR) blade at the outer margin of
the membrane (Fig. 21.3), usually in the superotemporal
quadrant. The retinotomy is made by teasing the nerve
fibers apart along the orientation of the nerve fibers rather

Fig u re 2 1 .3 ■ A nonm od ified

MVR blade is used to create a reti-
notom y by teasing the nerve layer
ap art. The retinotom y should be
m ade at the outer edge of the CNV
m em brane. If the CNV m em brane
extends from the fovea und er the
p ap illom acular (PM) b un d le, a
radial incision should be m ade in
the PM bundle. There is no need
to create a retinal detachm ent by
injecting BSS, thereby avoiding a
hydraulic m acular hole.
198 SECTION IV ■ Disease-Sp ecific Managem ent
than cutting across the fibers. This concept is similar to
the making of incisions along Langer’s lines in facial plas-
tic surgery. Diathermy is not applied to the retinotomy site
because bleeding is not an issue, as retinal vessels are never
transected. Diathermy causes an unnecessary nerve fiber
layer defect and may stimulate scarring.
The 25-gauge forceps are used to grasp the anterior
surface of the outer margin of the membrane (Fig. 21.4).
Forceps used with one blade under the membrane are more
likely to strip the RPE or damage the choriocapillaris in the
authors’ opinion. Pics are never used because of the risk of
RPE avulsion and bleeding from choriocapillaris trauma.
Grasping of the anterior surface of submacular membranes
with end-opening forceps was developed to reduce damage
to the RPE in an analogous manner to that developed for
epiretinal membranes. The intraocular pressure (IOP) is ele-
vated to 60 mm Hg using the Constellation alternative infu-
sion feature, using another console air source, or increasing
the bottle height prior to removing the membrane to reduce
bleeding. The membrane is rotated slowly to determine if
there is excessive adherence to the retina or RPE. Exces-
sive adherence is a common problem in patients with prior
photocoagulation, but much less frequent since anti-VEGF
therapy became available. Changing the direction of pull-
ing can usually resolve this problem, although 25-gauge fine
curved scissors are occasionally necessary to delaminate the
membrane from the retina or RPE. The membrane should be
removed very slowly to reduce retinal tearing and to enable
close observation of the retina and RPE to prevent damage
(Fig. 21.5). If the retina becomes more elevated during
removal, the endoilluminator can be used to gently push
Fig u re 2 1 . 4 ■ En d -g rasp in g
forcep s are used to g rasp th e
subfoveal m em brane at the outer
edge on the top surface. Pics and
forcep s b lad es sh ou ld n o t b e
p laced und er the retina to avoid
avulsion of th e RPE. Th e m em -
b ran e should b e rotated slig htly
to test for adherence to the retina
and to free up the m em brane.
the retina away from the membrane and toward the RPE.
Suction should not be applied to the retinotomy because
of the possibility of shearing damage to the photoreceptors
and RPE.
After the membrane is removed from the subreti-
nal space, it is retained in the grasp of the forceps and
used to gently push the retinotomy margins back together
(Fig. 21.6) and express fluid from the subretinal space.
The membrane is then moved to the anterior vitreous
space and approximated to the port of the vitreous cutter.
The membrane is then removed with the vitreous cutter
(Fig. 21.7) using coaxial illumination from the operating
microscope. Dragging the membrane through the periph-
eral vitreous and pars plana for removal with the forceps
may create vitreoretinal traction and subsequent retinal
detachment. There is no need to submit the membrane to
pathology, as the histologic appearance of these lesions is
well known.
Fluid-air exchange is used for surface tension manage-
ment for the retinotomy site to eliminate the need for laser
retinopexy. Laser retinopexy creates a scotoma from nerve
fiber layer damage and increases the chance of a new neovas-
cular membrane at the laser site. The exchange is performed
using the vitreous cutter for fluid egress and the Landers
or equivalent high minus lens for visualization in phakic
or pseudophakic eyes. The Constellation or equivalent air
pump is used as an air source.
The IOP should be maintained at higher than normal
levels during wound closure to prevent bleeding. Conjunc-
tival closure and subconjunctival antibiotics and steroids are
used in the manner described elsewhere in the text.
 CHAPTER 21 ■ Treatm ent of Choroidal Neovascular Mem branes 199

Fig ure 2 1 .5 ■ Rem ove the m em brane slowly to reduce retinal tearing and RPE avulsion. Scissors
can be used to sever the adherent areas if necessary.

Fig ure 2 1 .6 ■ Use the m em brane to gently reap proxim ate the retinotom y and exp ress SRF.
200 SECTION IV ■ Disease-Sp ecific Managem ent
RESULTS
Underlying, ongoing pathology, surgical damage to the RPE
and retina, and recurrent membranes determine the outcome.
Careful case selection, as described previously, is essential in
producing better outcomes. The membrane recurrence rate has
been reported from 25% (the principal author’s series) to 45%.
An occasional patient will develop an extrafoveal recurrence
and benefit from laser photocoagulation. Reoperation is indi-
cated for a well-defined subretinal membrane in patients who
had visual improvement after previous successful surgery.
TRANSPLANTATION OF RETINAL
PIGMENT EPITHELIUM
Although tremendous advances have been made in the treat-
ment of retinal diseases, macular degeneration still results in
poor visual outcomes in most patients and there is no treat-
ment for dry AMD. It has been postulated that a contributing
factor to the poor visual outcome following removal of CNV
membranes in macular degeneration patients is atrophy of
the subfoveal choriocapillaris (32). It has been reported that
the area of atrophy can continue to enlarge 1 year after sur-
gery. The stimulus for the choriocapillaris atrophy may be
the failure of the RPE to repopulate the surgical bed (33).
The extent of perfusion in the fovea is related to the visual
prognosis and therefore of great importance (34,35).
Unfortunately, tightly integrated RPE cells are removed
along with subfoveal neovascular membranes in Age Related
Fig u re 2 1 . 7 ■ Re m o ve t h e
m em brane with the vitreous cut-
ter; d o n ot d rag in throug h the
p erip heral vitreous. Coaxial illu-
m ination from the m icroscop e is
sufficient for this step in the ante-
rior vitreous.
Macular Degeneration (ARMD) patients during submacular
surgery. It has been shown in numerous studies that RPE
removal will lead to choriocapillaris atrophy (36–41). While
partial RPE regeneration may occur in some areas, other areas
develop choriocapillaris atrophy and resultant disorganized
photoreceptors. If one could insert new RPE cells during the
submacular surgical procedure, perhaps the inevitable atro-
phy could be prevented, or at least minimized.
It is not difficult to imagine the inherent problems of
RPE transplantation. Complex issues include transplanted
RPE cell viability and function, the need for lifelong immune
suppression to prevent rejection, choriocapillaris viabil-
ity, and RPE cell adherence to Bruch’s membrane. For over
25 years, researchers have been investigating these and the
many other complex issues regarding RPE transplantation.
News stories in the media have caused an excitement among
patients, and therefore it is very important for physicians to
be educated in order to counsel their patients effectively.
In 1975, researchers discovered that autotransplanted
RPE cells underwent metaplasia after their injection into the
vitreous cavity. The RPE cells first transformed into mac-
rophages and then later into spindle-shaped cells with col-
lagen production (42,43).
In 1989, physicians described a pars plana approach for
the transplantation of autologous RPE cells from a periph-
eral chorioretinal biopsy to prepare Bruch’s membrane at the
posterior pole of the same eye (44).
In 1991, Peyman reported his technique for RPE trans-
plantation in two patients with extensive subfoveal scar-
ring secondary to macular degeneration (45). His technique
involved the preparation of a large retinal flap encompassing
 CHAPTER 21
the macula and the arcades, removal of the submacular scar,
and replacement of the RPE cells using either an autologous
pedicle graft or homologous RPE cells and Bruch’s mem-
brane. One patient, who had undergone a pedicle graft,
had an improvement of visual acuity from count fingers to
20/400 at 14 months. The other patient developed encap-
sulation of his homologous graft without any improvement
in vision.
In 1992, scientists in Japan reported on the histology of
transplanted RPE cells in New Zealand White rabbits (46).
They found that by 1 week, the transplanted cells had formed
a monolayer. By 3 weeks, grafted RPE cells had formed api-
cal microvilli and tight junctions with adjacent cells. Their
contact with Bruch’s membrane appeared to be composed of
basal infoldings that were well formed. Their findings dem-
onstrated the functional appearance of the transplanted RPE
cells. The same year, a group of researchers reported that
RPE transplants stabilized retinal vasculature and prevented
neovascularization in the RCS (Royal College of Surgeons)
rat (47). Another study demonstrated that the transplanta-
tion of normal RPE cells reversed pathological changes in
the photoreceptors that had already occurred by the time of
transplantation in the RCS rat (48).
In 1994, a Swedish group led by Algvere published
their results on RPE transplantation in patients with exu-
dative ARMD performed in Sweden with RPE harvested by
investigators from Columbia University (49). Fetal RPE was
placed beneath the neurosensory retina after the removal of
submacular neovascularization in five patients with ARMD.
The preoperative vision in all five patients was very poor.
Surgical complications included cystoid macular edema and
macular pucker. Microperimetry demonstrated that all five
patients were able to fixate over the area of the RPE graft
immediately after surgery, but an absolute scotoma devel-
oped in this region within several months. There is no
evidence that the transplanted RPE cells survived in the
subretinal space. It is noteworthy that these patients did not
receive any immune suppression.
While progress was being made in transplantation tech-
niques, the topic of rejection was also being studied. In 1997,
Algvere’s group published another study comparing the fate
of fetal (13–20 weeks of gestational age) transplanted RPE
cells in the subretinal space of five patients with fibrovascu-
lar membranes with those transplanted in four patients with
atrophic ARMD (50). In patients with disciform lesions, all
grafts were rejected over a 6-month period. In patients with
nonexudative disease, however, three out of four transplants
showed little change in shape or size at 12 months. Visual
acuity remained stable in these patients. The authors con-
cluded that human RPE allografts are not invariably rejected
in the subretinal space and that an intact blood-retinal bar-
rier is likely to protect against rejection. More recent stud-
ies demonstrate a slow but significant effect of the systemic
immune system in the subretinal space, and therefore sci-
entists are cautioning investigators against considering the
subretinal space to be immunologically privileged (51).
■ Treatm ent of Choroidal Neovascular Mem branes 201
The latest development in the field of RPE transplantation
involves the cotransplantation of intact sheets of fetal retina
with RPE (52). Investigators at the University of Louisville
transplanted intact cografts into the subretinal space of RCS
rats. After 6 to 7 weeks, transplanted photoreceptors, with
the support of the cografted RPE cells, developed fully in
organized, parallel layers in the subretinal space. They con-
cluded that such transplants have the potential to benefit
retinal diseases with dysfunctional RPE and photoreceptors.
A tremendous amount of RPE transplantation research
has been conducted and continues to be undertaken. While
this is an exciting area of research with tremendous potential
benefits, we must remember that currently this remains an
area of research, not treatment.
MACULAR TRANSLOCATION
Lindsey and Finklestein first reported macular transloca-
tion for the purpose of studying the relationship between
the macula and the submacular RPE (53). The principal
author developed the concept of macular translocation for
the treatment of submacular CNV. Both retinal rotation and
small flap translocation were developed and investigated in a
Hanover pig model with the late Scott Langdon. It was deter-
mined that there was a high incidence of retinal detachment
and proliferative vitreoretinopathy (PVR). This work was
reported at the Bascom Palmer Eye Institute Annual Alumni
meeting in 1987. The principal author stated at the meeting
that this procedure should not be undertaken in humans
because of retinal detachment and PVR risk. Machemer, who
was present at the presentation, later reported performing
this work in humans. Subsequently, DeJuan, Tano, Toth,
Lewis, Eckardt, and others began clinical studies. All of
this work confirmed the original concerns of the principal
author, which were retinal detachment and PVR. In addition
to retinal detachment and PVR, macular holes, new CNV
membranes at the BSS injection sites, hemorrhage, cyclo-
version, diplopia, phthisis, multiple reoperations, macular
folds, and decreased or unchanged vision in spite of macu-
lar translocation were reported (54–57). The authors do not
believe that this procedure is indicated because of the high
incidence of complications.
DeJuan developed a scleral resection method and later
an imbrication method called “limited macular translocation”
(58). Complications associated with this method include ret-
inal detachment, PVR, hemorrhage, macular hole, new CNV,
phthisis, multiple reoperations, marked astigmatism, aniso-
coria, diplopia, ptosis, enophthalmos, and failure to improve
vision in spite of translocation. The authors do not believe
that this method is indicated because of unacceptable com-
plication rates. Hilel Lewis has developed an outpouching
method using clips that may prove to be more effective than
DeJuan’s technique of scleral imbrication, although most of
the same complications will not be reduced by this tech-
nique change.
202 SECTION IV ■ Disease-Sp ecific Managem ent
After approximately a 2-year period of intense interest
and promotion of this technique, the vast majority of sur-
geons have given up this procedure. In spite of extensive
early experimental work, the principal author has never
elected to perform this procedure on humans.
References
1. Michels S, Rosenfeld PJ, Puliafito CA, et al. Systemic bevaci-
zumab (Avastin) therapy for neovascular age-related macular
degeneration twelve-week results of an uncontrolled open-label
clinical study. Ophthalmology 2005 Jun;112(6):1035–1047.
2. Avery RL, Pieramici DJ, Rabena MD, et al. Intravitreal bevaci-
zumab (Avastin) for neovascular age-related macular degen-
eration. Ophthalmology 2006 Mar;113(3):363–372. e5. [Epub
2006 Feb 3.]
3. Macular Photocoagulation Study Group. Laser photocoagula-
tion of subfoveal neovascular lesions in age-related macular
degeneration: results of a randomized clinical trial. Arch Oph-
thalmol 1991;109:1220–1231.
4. Klein BE, Brady WE, Palta M. Association between antioxidant
and zinc intake and the 5-year incidence of early age-related
maculopathy in the Beaver Dam Eye Study. Am J Epidemiol
1998;148(2):204–214.
5. Klein R, Klein BE, Moss SE. The Beaver Dam Eye Study. Rela-
tionship of smoking to the incidence of age-related maculopa-
thy. Am J Epidemiol 1998;147(2):103–110.
6. De Luca LM, Ross SA. Beta-carotene increases lung cancer inci-
dence in cigarette smokers. Nutr Rev 1996;54(6):178–180.
7. Figueroa MS, Regueras A, Bertrand J. Laser photocoagulation to
treat macular soft drusen in ARMD. Retina 1994;14(5):391–396.
8. Wetzig PC. Photocoagulation of drusen-related macular degen-
eration: a long-term outcome. Trans Am Ophthalmol Soc 1994;92:
299–303.
9. Olk RJ, Friberg TR, Stickney KL, et al. Theraputic benefits of
infared (810 nm) diode laser macular grid photocoagulation in
prophylactic treatment of nonexudative ARMD. Ophthalmology
1999;106(11):2082–2090.
10. The Choroidal Neovascularization Prevention Trial Research
Group. Choroidal neovascularization in the Choroidal Neo-
vascularization Prevention Trial. Ophthalmology 1998;105(8):
1364–1372.
11. The Choroidal Neovascularization Prevention Trial Research
Group. Laser treatment in eyes with large drusen. Short term
effects seen in a pilot randomized clinical trial. Ophthalmology
1998;105(1):11–23.
12. Treatment of Age-related Macular Degeneration With Photo-
dynamic Therapy (TAP) Study Group. Photodynamic ther-
apy of subfoveal choroidal neovascularization in age-related
macular degeneration with verteporfin. Arch Ophthalmol
1999;117:1329–1345.
13. Schmidt-Erfurth U, Hasan T, Gragoudas E, et al. Vascular tar-
geting in photodynamic occlusion of subretinal vessels. Oph-
thalmology. 1994;101:1953–1961.
14. Allison BA, Waterfield E, Richter AM, et al. The effects of
plasma lipoproteins on in vitro tumor cell killing and in vivo
tumor photosensitization with benzoporphyrin derivative.
Photochem Photobiol 1991;54:709–715.
15. Schmidt-Erfurth U, Hasan T, Schomacker K, et al. In vivo
uptake of liposomal bezoporphyrin derivative and photo-
thrombosis in experimental corneal neovascularization. Lasers
Surg Med 1995;17:178–188.
16. Young LH, Howard MA, Hu LK, et al. Photodynamic therapy of
pigmented choroidal melanomas using a liposomal preparation of
bezoporphyrin derivative. Arch Ophthalmol 1996;114:186–192.
17. Miller JW, Walsh AW, Kramer M, et al. Photodynamic therapy
of experimental choroidal neovascularization using lipoprotein-
delivered benzopophyrin. Arch Ophthalmol 1995;113:810–818.
18. Kramer M, Miller JW, Michaud N, et al. Liposomal benzopor-
phyrin derivative verteporfin photodynamic therapy: selective
treatment of choroidal neovascularization in monkeys. Oph-
thalmology 1996;103:427–438.
19. Husain D, Miller JW, Michaud N, et al. Intravenous infusion of
liposomal benzoporphin derivative for photodynamic therapy
of experimental choroidal neovascularization. Arch Ophthalmol
1996;114:978–985.
20. Weishaupt KR, Gomer CJ, Dougherty TJ. Identification of
singlet oxygen as the cytotoxic agent in photo-inactivation of
murine tumor. Cancer Res 1976;36:2326–2329.
21. Steinhorst UH, Theischen M, Winter R. Subretinal lavage: a
technique of continuous irrigation for removal of traumatic sub-
macular hemorrhage. Ophthalmologica 1997;211(6):399–401.
22. Capone A. Submacular surgical procedures. Int Ophthalmol
Clin 1995;35(4):83–93.
23. Ibanez HE, Williams DF, Thomas MA, et al. Surgical man-
agement of submacular hemorrhage. Arch Ophthalmol
1995;113(1):62–69.
24. Humayun M, Lewis H, Flynn HW, et al. Management of sub-
macular hemorrhage associated with retinal arterial macroan-
eurysms. Am J Ophthalmol 1998;126(3):358–361.
25. Claes C, Zivojnovic R. Efficacy of tissue plasminogen activator
in subretinal hemorrhage removal. Bull Soc Belge Ophthalmol
1996;261:115–118.
26. Hawkins WR. Intraocular fibrinolysis of submacular hem-
orrhage with tPA and surgical drainage. Am J Ophthalmol
1994;118(5):559–568.
27. Lim JI, Drews-Botsch C, Sternberg P, et al. Submacular hemor-
rhage removal. Ophthalmology 1995;102(9):1393–1399.
28. Chaudhry NA, Mieler WF, Han DP, et al. Preoperative use of
tissue plasminogen activator for large submacular hemorrhage.
Ophthalmic Surg Lasers 1999;30(3):176–180.
29. Johnson MW. Pneumatic displacement of submacular hemor-
rhage. Curr Opin Ophthalmol 2000;11(3):201–206.
30. Kamei M, Misono K, Lewis H. A study of the ability of tPA to
diffuse into the subretinal space after intravitreal injections in
rabbits. Am J Ophthalmol 1999;128(6):739–746.
31. Thomas MA, Kaplan HJ. Surgical removal of sub-foveal neo-
vascularization in POHS. Am J Ophthalmol1991;111:1–7.
32. Castellarin A, Nasir M, Sugino IK, et al. Decreased chorio-
capillaris perfusion following surgical excision of choroidal
neovascular membranes in age-related macular degeneration.
Br J Ophthalmol 1997;81:481–489.
33. Castellarin A, Nasir M, Sugino IK, et al. Progressive choriocap-
illaris atrophy after surgery for age-related macular degenera-
tion. Retina 1998;18:143–149.
34. Akduman L, Del Priore LV, Desai VN, et al. Perfusion of the
subfoveal choriocapillaris affects visual recovery after submac-
ular surgery in the presumed ocular histoplasmosis syndrome.
Am J Ophthalmol 1997;123:90–96.
35. Pollack JS, Del Priore LV, Smith ME, et al. Postoperative abnor-
malities of the choriocapillaris in exudative age-related macu-
lar degneration. Br J Ophthalmol 1996;80:314–318.
36. Korte GE, Reppucci V, Henkind P. RPE destruction causes
choriocapillaris atrophy. Invest Ophthal Vis Sci 1984;25:1135–
1145.
37. Kuwabara T, Ishikawa Y, Kaiser-Kupfer MI. Experimental
model of gyrate atrophy in animals. Ophthalmology 1981;88:
331–334.
38. Takeuchi M, Itagaki T, Takahashi K, et al. Changes in the inter-
mediate stage of retinal degeneration after intravitreal injection
of ornithine. Nippon Ganka Gakkai Zasshi. Acta Soc Ophthalmol
Jpn (Tokyo) 1993;97:17–28.
 CHAPTER 21
39. Valentino A, Kaplan HJ, Del Priore LV, et al. Retinal pigment
epithelium repopulation in monkeys after submacular surgery.
Arch Ophthalmol 1995;113:932–938.
40. Del Priore LV, Kaplan HJ, Hornbeck R, et al. Retinal pigment
epithelium debridement as a model for the pathogenesis
and treatment of macular degeneration. Am J Ophthalmol
1996;122:629–643.
41. Del Priore LV, Hornbeck K, Kaplan HJ. Debridement of the
pig retinal epithelium in vivo. Arch Ophthalmol 1995;113:
939–944.
42. Mueller-Jensen K, Machemer R, Azarnia R. Autotransplanta-
tion of retinal pigment epithelium in intravitreal diffusion
chamber. Am J Ophthalmol 1975;80:530–537.
43. Mueller-Jensen K, Mandelcorn MS. Membrane formation by
autotranplanted retinal pigment epithelium. Mod Probl Oph-
thalmol 1975;15:228–234.
44. Lane C, Boulton M, Marshall J. Transplantation of retinal pigment
epithelium using a pars plana approach. Eye 1989;3:27–32.
45. Peyman GA, Blinder KJ, Paris CL, et al. A technique for reti-
nal pigment epithelium for age related macular degeneration
secondary to extensive subfoveal scarring. Ophthalmol Surg
1991;22(2):102–108.
46. Yamaguchi K, Yamaguchi K, Young RW, et al. Vitreoretinal sur-
gical technique for transplanting retinal pigment epithelium in
rabbit retina. Jpn J Ophthalmol 1992;36(2):142–150.
47. Seaton AD, Turner JE. RPE transplants stabilize retinal vascula-
ture and prevent neovascularization in the RCS rat. Invest Oph-
thalmol Vis Sci 1992;33(1):83–91.
48. Lavail MM, Li L, Turner JE, et al. Retinal pigment epithelial
cell transplantation in RCS rats: normal metabolism in rescued
photoreceptors. Exp Eye Res 1992;55(4):555–562.
■ Treatm ent of Choroidal Neovascular Mem branes 203
49. Algvere PV, Berglin L, Gouras P, et al. Transplantation of fetal
retinal pigment epithelium in age-related macular degenera-
tion with subfoveal neovascularization. Graefes Arch Clin Exp
Ophthalmol 1994;232(12):707–716.
50. Algvere PV, Berglin L, Gouras P, et al. Transplantation of RPE
in age-related macular degeneration: observations in disciform
lesions and dry RPE atrophy. Graefes Arch Clin Exp Ophthalmol
1997;235(3):149–158.
51. Zhang X, Bok D. Transplantation of retinal pigment epithelial
cells and immune response in the subretinal space. Invest Oph-
thalmol Vis Sci 1998;39(6):1021–1027.
52. Aramant RB, Seiler MJ, Ball SL. Successful cotransplantation
of intact sheets of fetal retina with retinal pigment epithelium.
Invest Ophthalmol Vis Sci 1999;40(7):1557–1564.
53. Bressler NM, Finklestein D, Sunness JS, et al. Retinal pigment
epithelial tears through the fovea with preservation of good
visual acuity. Arch Ophthalmol 1990;108(12):1694–1697.
54. American Academy of Ophthalmology. Macular translocation.
Ophthalmology 2000;107(5):1015–1018.
55. Lewis H, Kaiser PK, Lewis S, et al. Macular translocation for
subfoveal choroidal neovascularization in ARMD: a prospec-
tive study. Am J Ophthalmol 1999;128(2):135–146.
56. Ohji M, Fujikado T, Saito Y, et al. Foveal translocation:
a comparison of two techniques. Semin Ophthalmol 1998;
13(1):52–62.
57. Ninomiya Y, Lewis JM, Hasegawa T, et al. Retinotomy and
foveal translocation for surgical management of subfoveal chor-
oidal neovascular membranes. Am J Ophthalmol 1996;122(5):
613–621.
58. DeJuan E, Vander JF. Effective macular translocation without
scleral imbrication. Am J Ophthalmol 1999;128(3):380–382.
C H AP TER
22
INTRAVITREAL
INJECTION TECHNIQUE
Never inject if blepharitis or conjunctivitis is
present; external infections should be treated
intensively for several days before injecting and
the patient examined at the slit lamp before
injecting to verify eradication of the infection.
Never stop anticoagulants before intravitreal injec-
tions or vitrectomy; the risk of bleeding is exception-
ally low, bleeding is rare and minimal, and the risk of
thromboembolic events is increased. Many patients as
well as their medical doctors and nurses stop anticoagu-
lants before injections and surgery even though they were
not instructed to do so, increasing thromboembolic risk,
potentially creating a scenario in which an anti–vascular
endothelial growth factor compound is wrongly blamed for
a stroke or myocardial infarction (1).
The technician and injecting ophthalmologist should
wear a mask because of the proximity of the nose and mouth
bacteria to the needle when adjusting the quantity of drug in
the tuberculin syringe. Gloves should be worn by the techni-
cian and injecting ophthalmologist largely for patient percep-
tion. Sterile gloves are also advantageous if the sterile speculum
is manipulated to improve access to the injection site.
Subconjunctival lidocaine is not required; the authors
believe that perforation risk and the pain of anesthetic
injection exceed the value. Lidocaine jelly (2%) in prefilled
syringes should be applied before 5% povidone iodine
(Betadine) both to improve patient comfort and because
multiuse packaging could theoretically become contaminated.
The povidone iodine should be applied to lashes, lid mar-
gins, injection site, and cul-de-sac. Povidone iodine swabs
are used on both upper and lower lids.
204
Betadine should be used for all patients because the
risk of endophthalmitis is greatly reduced. The risk of topi-
cal allergic reactions is exaggerated; an allergy to shellfish
is not sufficient to avoid povidone iodine, nor is mild skin
irritation. It is important to use sterile saline (EyeStream) to
wash off lids and lashes after the procedure.
A sterile, bladed speculum must be used to expose the
injection site as well as contain the lashes. The patient must
be supine at elbow height for the surgeon to maximize inject-
ing ophthalmologist ergonomics. Measure 3.5 mm posterior
to the limbus using the Luer Lok on a sterile tuberculin
syringe without the needle. The tip of the index finger on
the opposite hand should be placed against the needle hub
and used to steer the 30 or 32 gauge needle and maintain a
trajectory perpendicular to a tangent plane at the injection
point. The patient should be given a fixation point on the
ceiling, the room should be silent, and no one should move
or enter or leave the room during the injection to eliminate
stimuli for saccades.
Indirect ophthalmoscopy should be performed after the
injection to verify circulation rather than intraocular pres-
sure measurement because of the risk of contamination. The
authors use topical antibiotics after injection (Vigamox q2h
on the day of injection and qid for 3 days) but no pretreat-
ment with topical antibiotics.
Reference
1. Charles S, Rosenfeld PJ, Gayer S. Medical consequences of stop-
ping anticoagulant therapy before intraocular surgery or intravit-
real injections. Retina. 2007;27(7):813-815.
C H AP TER
23
TRAUMA
Vitreous surgery techniques have greatly
improved the management of ocular trauma.
Work-related activity, home maintenance, auto-
mobile and motorcycle accidents, fireworks,
hunting, violence, and recreational activity create a contin-
ued threat of severe ocular injury. The vast range of objects
and velocities implicated in ocular trauma lead to the wide
spectrum of injuries seen (1). In this chapter, the emphasis
will be placed on several stereotypic subgroups.
WOUND-RELATED CELLULAR
MIGRATION/ PROLIFERATION
Any interruption of tissue results in proliferation of the inter-
rupted cell groups. Planar cell groups respond to loss of contact
inhibition by a migration/proliferation of the cells adjacent to
the interruption. Misalignment of the proliferating plane may
result in reduplication of the original tissue layer. Migration/
proliferation ceases when contact inhibition is restored by con-
tinuity of the new cell group with similar cells representing the
margin of defect. This new structure can be called a membrane;
however, it is actually a reparative extension of previously nor-
mal tissues. The widely used term “fibrovascular ingrowth”
implies that wound-related cellular proliferation originates from
extraocular tissues. The severe proliferation that occurs in blunt
trauma after choroidal rupture (retinitis sclopetaria) is clinically
indistinguishable from “fibrovascular ingrowth.” There is little
direct evidence of extraocular origin for the majority of cases
of wound-related cellular proliferation. Because the reparative
mechanism stems from tissue disruption and destruction, the
additional damage of retinopexy should be avoided unless nec-
essary for the treatment of a retinal defect. Silicone oil can be
used for what the authors term “retinopexy avoidance” in the
setting of severe trauma with retinal breaks.
SUBSTRATES FOR PROLIFERATION
Cellular proliferation occurs on preexisting support sub-
strates such as the cornea, retina, vitreous, lens, and iris.
The vitreous has been referred to as a “scaffold” (2,3), but
this is a substrate. Substrate better describes the manner in
which cellular migration and proliferation occur on the vit-
reous collagen matrix. More accurately, it is critical to rec-
ognize that proliferation usually occurs along the anterior
vitreous cortex (AVC) and posterior vitreous cortex (PVC),
although a foreign object can make a path through the vit-
reous along which apparent transvitreal proliferation can
occur. As the retina is an ideal substrate for proliferation,
it is not necessary to implicate the vitreous when epiretinal
cellular migration and proliferation occur.
TIMING OF VITRECTOMY
Immediate vitrectomy in penetrating ocular trauma cases
should be avoided unless certain types of intraocular for-
eign bodies (IOFBs) are present. The experience of the Iraq
war has shown that primary closure of entry wounds with
delayed removal of the IOFB can have good outcomes (4).
Arterial bleeding, choroidal swelling, leaking wounds, striate
keratopathy, corneal edema, and lack of preparation add to
the difficulty of immediate vitrectomy (5,6). More impor-
tantly, penetrating trauma frequently occurs in the young
patient not having a posterior vitreous detachment (PVD).
Without adequate vitreoretinal separation, iatrogenic retinal
breaks, difficult surgery, and postoperative contraction of
residual vitreous can ensue.
Typically, the hemorrhage and inflammation that accom-
pany trauma induce a PVD in 7 to 14 days, allowing safer,
more effective vitreous surgery. Cellular proliferation starts at
10 to 14 days, making this the ideal time to intervene (7,8). If
the retina can be seen, these cases can be watched at weekly
intervals and vitrectomy avoided if cellular proliferation
does not occur. Cases with opaque media require vitrectomy
at this 10- to 14-day point because further delay could lead
to cellular proliferation, which is initially difficult to deter-
mine with B-scan ultrasound. Decreased vitreous mobility
observed on ultrasound examination indicates hypocel-
lular vitreous collagen contraction and/or early cellular
proliferation.
205
206 SECTION IV ■ Disease-Sp ecific Managem ent
INTRAOCULAR FOREIGN BODIES
Vitreous surgery allows excellent visualization; prevention
of postoperative transvitreal proliferation; and removal of
blood, lens materials, and organisms if present. Bronson
or giant magnet foreign body removal has virtually disap-
peared because of the widespread availability of vitreous
surgery training, technology, techniques, and the Machemer-
Parel diamond-coated IOFB forceps (Fig. 23.1). Intraocular
magnets can be used to pick up foreign bodies for transfer
to diamond-coated forceps and removal, although this is
seldom necessary (9–15).
In e rt Ve rsus To xic Fo re ig n Bo d ie s
Most iron- and copper-containing foreign bodies should
be removed immediately (16,17). Stainless steel, alumi-
num, and lead foreign bodies are much less toxic and
can be tolerated in selected cases. Indication for removal
of these less toxic materials must be individualized on a
clinical basis. Occasionally, iron- and copper-containing
foreign bodies will be overlooked and only discovered
after they have become encapsulated. If no evidence of
toxicity is seen in these cases, frequent follow-up should
be undertaken rather than surgery. Clinical examination
of adjacent structures (cornea, iris, and lens) is probably
a better indication of toxicity than electroretinography.
Plastic materials can be left in place in selected cases.
Submacular foreign bodies and those embedded in the
optic nerve should be managed on an individual basis
because of the extreme hazards of removal. Any exog-
enous biologic material such as vegetable matter should
be removed immediately because of the risk of infection
and inflammation. Endogenous cilia and bone fragments
are usually well tolerated and need not be removed unless
vitrectomy is performed for other reasons or they appear
to be the cause of inflammation.
Fig ure 2 3 .1 ■ Diam ond -coated foreign body forceps prevent
dropping and grasp-induced m ovem ent of the IOFB.
Tim in g o f Surg e ry
All large, toxic, biologic, or sharp IOFBs should be removed as
soon as the patient can be safely taken to the operating room.
This approach decreases secondary mechanical trauma, rapid
toxicity, and endophthalmitis. Plastic, glass, and lead shotgun
pellets can be observed until vitrectomy is indicated for other
reasons. Although late-night and weekend vitreous surgery
creates logistical and cost problems, it decreases the risk of
endophthalmitis and toxic damage and should be undertaken
if possible.
Surg ical Se q ue n ce an d Te ch n iq ue s
Wound Repa ir
Any visible corneal or scleral wound should be sutured
before proceeding with the vitrectomy. Running shoelace
monofilament nylon sutures distribute stress evenly, are
elastic and well tolerated, and may be rapidly placed. Silk
sutures are inelastic and lead to wound leaks during the vit-
rectomy, while absorbable sutures are inelastic and not per-
manent. Interrupted sutures can cause striate keratopathy
and take longer to place. Small sutures (10-0) are used for
central cornea, 9-0 for midcornea, and 8-0 for peripheral
cornea or sclera.
Surgical judgment should guide the decision-making
process concerning excision of prolapsed tissue. Viable-
appearing iris or ciliary body in a very recent injury can
be irrigated and repositioned, while any sign of infection
or tissue destruction indicates excision. Retinopexy should
not be used anterior to the muscle ring (ora serrata) and
should be applied only to definite retinal breaks located
posteriorly. Excessive and unnecessary retinopexy leads to
greater wound-related cellular proliferation and inflamma-
tion. Endolaser retinopexy should be applied only to specific
breaks identified during vitrectomy.
Exploration of/for a posterior wound should be done
only if pressure on the globe can be completely avoided.
The vitreous and retina can be prolapsed from a posterior
wound by surgical manipulation. If vitrectomy is completed
first, the location of the wound will be known, and the eye
can be softened and filled with air (gas) before proceeding
with posterior wound repair in the rare instance that it is
thought to be necessary. Most posterior wounds are self-
sealing, and wound closure does not decrease the incidence
of wound-related cellular proliferation. Retinopexy increases
the incidence of wound-related cellular proliferation and
should be avoided.
Conjunctiva l Incisions
The authors currently use 25-gauge sutureless vitrectomy for
all trauma cases and open the conjunctiva selectively only
to repair scleral wounds and to construct a one clock-hour
incision when a superotemporal 25-gauge wound is enlarged
to 20 gauge or larger for IOFB removal.

Sclerotomies
The incision for the vitrectomy instrument, infusion cannula,
and endoilluminator should be made in the usual position,
3 mm posterior to the limbus if the lens is to be removed
or 4 mm posterior to the limbus if not. Extra care must be
taken to avoid choroidal infusion in trauma cases because
hypotony and choroidal edema are frequently present.
The incision for the foreign body removal should usu-
ally be the superotemporal vitrectomy instrument incision,
enlarged after the vitreous is removed using a diamond or
disposable knife.
Lens Remova l
If the lens is clear, it should be allowed to remain unless a
very large foreign body requires translimbal removal. Many
small, localized traumatic cataracts do not progress and
therefore may not require removal. The vitrectomy instru-
ment should be used for anterior vitrectomy if any vitreous
is in the anterior chamber or capsular bag. The aspirating
fragmenter is faster than vitrectomy instruments for lensec-
tomy but should never be applied to the vitreous. If vitreous
enters the capsular bag, it should be removed with a vit-
rectomy instrument and the lensectomy completed with the
fragmenter. In most cases, the capsule should be removed
with the diamond-coated or end-grasping forceps.
Vitrectomy
The vitreous may be clear or have significant hemorrhage in
acute IOFB cases, but in either instance, it must be completely
removed. The first goal is to remove enough vitreous to
remove all vitreous attachments to the foreign body. Although
Fig u re 2 3 .2 ■ The free h and
should use a d iam ond or d isp os-
ab le knife to enlarg e sclerotom y
for foreign bodies too large to go
through 20-gauge m icrovitreoret-
inal (MVR) incision. This should be
d on e after g rasp in g th e foreig n
body and bringing it near the pars
plana to avoid a wound leak while
grasping the IOFB.
CHAPTER 23 ■ Traum a 207
it is tempting to remove the foreign body as soon as it is seen,
it is far better to complete the vitrectomy first. A big advan-
tage of the vitrectomy approach is the avoidance of removal-
induced vitreous traction. The need for a relatively complete
vitrectomy before a foreign body removal is absolute.
Foreign Body Remova l
It is not recommended that an extraocular magnet be used
in conjunction with vitrectomy; the advantage of a slow,
controlled removal without vitreous traction would be lost.
Some surgeons have recommended intraocular magnets,
but the authors have not used these devices because of the
availability of the diamond-coated forceps. When all the vit-
reous around the IOFB has been removed, the superotempo-
ral incision is enlarged to accommodate the largest outside
diameter of the diamond IOFB forceps.
Moderately Sized Foreign Bodies
After the foreign body is grasped in the forceps and
brought to the anterior vitreous area, its size can be better
assessed. If the foreign body will not fit safely through the
wound, but is not larger than 6 mm in its smallest diameter,
the scleral wound should be enlarged with a knife (Fig. 23.2).
It is best to plug the other sclerotomy so that the surgeon’s
hand is free to enlarge the wound. A knife at the opposite
end of the circumferential sclerotomy from the shank of the
forceps should be used to extend the sclerotomy while stabi-
lizing the other end of the wound with the shank of the for-
ceps. Using this approach, it is not necessary to release the
grasp of the foreign body, and a large, leaky incision need not
be present while searching for and picking up the IOFB.
208 SECTION IV ■ Disease-Sp ecific Managem ent
Cylindrical Foreign Bodies
Long, cylindrical, small-diameter foreign bodies such as
wire frequently lie tangential to the retinal surface. The initial
pickup of the foreign body will then result in the IOFB being
nearly perpendicular to the shank of the forceps. If it were
removed with this orientation, the scleral opening would have
to be unnecessarily large. It is therefore best to use a second
forceps to bimanually transfer or regrasp the foreign body to
permit removal along the long axis of the IOFB (Fig. 23.3).
Fig u re 2 3 .3 ■ A secon d p air
of forcep s is used to reg rasp a
long, thin IOFB so that it can be
rem oved along its long axis.
Large Foreign Bodies
If the foreign body is brought to the anterior vitreous
and appears too large to be removed through the pars plana,
translimbal removal should be performed (Fig. 23.4). The
grasp on the foreign body need not be lost as the lens is
removed with the aspirating fragmenter in the surgeon’s other
hand. A cataract-type limbal section is then initiated with a
blade, again in the surgeon’s other hand. The infusion system
prevents hypotony during the initial incision but will cause
Fig ure 2 3 .4 ■ Very larg e IOFBs
sh o u ld b e re m o ved t h ro u g h
the lim b us after vitrectom y and
len sectom y, g rasp in g th e IOFB
an d b rin g in g it in to th e an te-
rior cham b er without d ang erous
regrasping.
 CHAPTER 23 ■ Traum a 209

Fig u re 2 3 . 5 ■ En cap su lat ed

IO FBs sh o u ld b e rem o ved b y
perform ing com plete vitrectom y,
then incising the capsule with an
MVR blade.

iris prolapse as the wound is enlarged. If the pupil is large and
iridectomy is not required, the infusion should be turned off
to avoid iris prolapse. If an iridectomy is required, the infu-
sion-induced iris prolapse can be used to the surgeon’s advan-
tage, permitting ab-externo iridectomy with one hand. When
the incision is large enough, the IOFB should be passed retro-
grade through the wound and removed with a second pair of
forceps through the limbal wound. The wound should then
be closed with a running shoelace, 8-0 nylon suture.
Encapsulated Foreign Bodies
The fibrous capsule must be incised to safely remove
an encapsulated IOFB (Fig. 23.5). The vitrectomy should be
completed first and all vitreous traction to the foreign body
site severed. The capsular incision should be performed with
20-gauge scissors (Fig. 23.6). A cruciate incision will allow
the IOFB to be completely free before the IOFB forceps are
used to remove the fragment (Figs. 23.7 and 23.8). Stan-
dard removal techniques as described previously are used,

Fig u re 2 3 .6 ■ The top of the

cap sule should be rem oved with
the scissors.
210 SECTION IV ■ Disease-Sp ecific Managem ent
depending on the size of the foreign body. Subretinal lOFBs
are removed with diamond-coated forceps (Fig. 23.9).
After the IOFB has been removed, the wound should be
closed with a running shoelace, 8-0 nylon suture, leaving a
20-gauge–size opening around a scleral plug.
The vitrectomy should then be completed, removing
any fibrin, capsular material, or hemosiderin left in the
residual vitreous or at the IOFB site. In acute cases in young
Fig ure 2 3 .7 ■ The foreign body
should b e g ently m oved to free
up adhesions.
people, it may not be possible to remove a portion of the
PVC on the macula and optic nerve; however, all connec-
tions between these areas and to the peripheral retina must
be severed. Theoretically, it would be beneficial to remove
the entire PVC, but this is not possible in many young people
without damaging the optic nerve or retina. Any subsequent
cellular proliferation along these remaining portions of PVC
can be managed, if necessary, at a later date.
Fig ure 2 3 .8 ■ The IOFB is
rem oved with diam ond-coated
forceps after m ob ilization.

Fig ure 2 3 .9 ■ Subretinal IOFBs
are rem o ve d wit h d iam o n d -
coated forcep s throug h a retinal
break or retinotom y.
Subretinal surgery is occasionally required in late trauma
cases, but the majority of these membranes are highly vascu-
lar and should be left in place.
Retinopexy
Laser retinopexy should only be performed if a definite reti-
nal break is seen, not for prophylaxis along trauma contact
and wound sites. If the break is located contiguous with
the optic nerve, papillomacular bundle, or macula, it is not
treated because breaks in these areas almost never result in
detachment, and retinopexy in these areas can decrease cen-
tral vision.
Surfa ce Tension Ma na gement
Internal fluid-air exchange and subsequent air-gas exchange
should be used if a definite retinal break is present. Inter-
nal drainage of subretinal fluid (SRF) should be performed
simultaneously with internal fluid-air exchange, if detach-
ment exists. Postreattachment endolaser retinopexy can then
be used to treat the break(s).
Although perfluorocarbon (PFC) liquids have been rec-
ommended for IOFB removal and management of coexisting
retinal detachment, the authors have not found these agents
to be necessary in most cases.
Air-gas exchange with SF6 after postreattachment endo-
laser retinopexy should be used for small to moderate retinal
defects.
Large retinal defects or especially severe trauma should
be managed by air-silicone exchange without retinopexy.
CHAPTER 23 ■ Traum a 211
The purpose of postoperative surface tension management
is to prevent aqueous humor from passing through nonvisi-
ble retinal defects, new traction-induced breaks, and defects
that are intentionally untreated. Retinopexy avoidance is
an excellent approach for the prevention of proliferative
vitreoretinopathy (PVR) and fibrovascular proliferation.
Sclera l Buckling
The authors find no benefit in placing buckle elements for
trauma cases, either for primary repair or for secondary
retinal detachments. These complex cases are usually best
repaired with vitrectomy.
Segmental buckling can be used for peripheral
detachments/breaks but is unnecessary for posterior
detachments/breaks. Circumferential segmental explants
are preferable to radial buckles in all instances. The
authors no longer buckle trauma eyes. Vitreous traction
from initial IOFB impact, egress of the foreign body, or
vitreous removal creates a relatively high incidence of
dialyses and aphakic-like retinal breaks. These can be dif-
ficult to recognize at the time of surgery, thus emphasizing
the need for a careful search and air-silicone exchange in
difficult cases. Late traction from vitreous incarceration in
the sclerotomies plays a role in postoperative detachment
as well.
Antibiotic-Steroid Thera py
Subconjunctival antibiotic therapy against both Gram-
positive and Gram-negative organisms should be utilized.
212 SECTION IV ■ Disease-Sp ecific Managem ent
Subconjunctival repository steroids such as triamcinolone
should be used to decrease postoperative fibrin formation
and scarring.
DOUBLE PENETRATING INJURY
High-velocity objects from shotgun and explosive injuries
typically enter the eye anteriorly and exit posteriorly, creat-
ing two ocular wounds (18). Low-velocity injuries such as
hammering metal-on-metal typically cause IOFBs (19–25).
In it ial Re p air
Any corneal or anterior scleral wound should be closed as
soon as the patient is seen, and any visible wound should be
closed before attempting further exploration. Great care must
be taken to avoid tissue prolapse from surgical manipula-
tion. Running shoelace monofilament nylon sutures should
be utilized as described above. Posterior exploration is rarely
indicated and should be done only if it can be done without
exerting any pressure on the globe.
Tim in g o f Vit re ct o m y
Vitrectomy should be performed between 7 and 14 days
unless angle closure from lens swelling or endophthalmi-
tis is an issue. Ultrasonic evidence of retinal detachment
does not necessarily indicate early vitrectomy. Delay for 7 to
14 days permits PVD to occur, decreases choroidal swelling,
decreases bleeding, and provides better corneal clarity.
Surg ical Se q ue n ce an d Te ch n iq ue s
Lens Ma na gement
Clear lenses or small, localized cataracts should be allowed to
remain. Any central, posterior subcapsular opacity warrants
lensectomy because this will almost invariably progress after
vitrectomy. Wounds at the ciliary body–pars plana level per-
mit wound-related cellular proliferation to proceed along the
AVC-lens interface, creating a cyclitic membrane. Although
prophylactic lens removal is not indicated, these cases must
be watched weekly postoperatively for any evidence of cyclitic
membrane growth. The endocapsular fragmenter lensectomy
technique should be utilized, with care taken to avoid vitre-
ous in the fragmenter. The vitrectomy instrument should be
used to remove any vitreous in the anterior chamber or cap-
sular bag. Diamond-coated or end-grasping forceps should be
used for capsular removal. Iris surgery should be performed
only if absolutely necessary to see posteriorly; overzealous
removal creates postoperative inflammation and glare.
Vitrectomy
The AVC should be removed to decrease the chance of
cyclitic membrane formation unless the trauma is very
posterior and the anterior retina and vitreous are normal.
An opening should then be made through the PVC nasally
or in any area known to have attached retina. Vacuum
cleaning–extrusion may be necessary through this initial
opening to remove free blood products in the sub-PVD
space. When an adequate view of the retina is obtained,
vitrectomy can proceed by enlarging this opening in a
circumferential fashion. If retinal detachment is present,
extremely low suction force should be used. If a retinal
break is seen, vitrectomy should be continued with inter-
mittent drainage of SRF. Layer-by-layer removal of the vit-
reous increases the chance of retinal breaks and is more
time-consuming than the full-thickness, circumferential
method described previously.
All vitreous attached to the exit wound should be
removed if it can be accessed without further damage to
the retina. The vitreous applied to the surface of the wound
should be allowed to remain because its removal can create
bleeding, retinal breaks, wound leaks, and enhancement of
cellular proliferation (Fig. 23.10). At the 14-day point, it is
extremely unusual for an exit wound to leak. As in vitrec-
tomy for diabetes, the goal of cone truncation is to eliminate
the anteroposterior traction, which is the critical element in
successful management of these cases.
Surfa ce Tension Ma na gement
Fluid-air exchange is utilized as in other clinical situations
when a retinal break is present. Internal drainage of SRF fol-
lowed by fluid-air exchange and completion of SRF drain-
age should be performed if detachment is present. Air can
also be utilized to confine bleeding and to help seal leaky
wounds by a surface tension effect. The use of PFC liquids,
gas, and silicone is described earlier.
Retinopexy
Laser endophotocoagulation is used only if retinal breaks
are apparent. Most posterior exit wounds do not give rise to
retinal detachment, and retinopexy serves only to increase
wound-related cellular proliferation.
Sclera l Buckling
The authors no longer use prophylactic encircling bands.
This change was made because buckles increase operating
time, postoperative pain, refractive error, strabismus, ptosis,
and cost and may develop late intrusion into the sclera.
CORNEAL-SCLERAL LACERATION
The great variability of corneal-scleral lacerations makes
generalization difficult, but certain principles merit discus-
sion. Microsurgery has advanced the success rate in these
cases, as has vitrectomy technology.
 CHAPTER 23 ■ Traum a 213

Fig u re 2 3 .1 0 ■ Con e trun ca-

tion (rem oval of PVC) elim inates
the substrate for transvitreal p ro-
liferation and reduces the chances
of retinal detachm ent.

Wo un d Re p air cases must be observed very frequently for evidence of vitre-

ous “orientation” and loss of vitreous mobility. Taut vitreous
As discussed previously, running monofilament sutures are
with stress lines directed toward a wound is associated with a
best for closing corneal-scleral lacerations. Running sutures
high incidence of retinal detachment. If this condition coexists
distribute stress evenly, stimulate less vascularization, leak
with opaque media, frequent ultrasound examinations, includ-
less, and are placed rapidly. All knots should be buried if
ing the periphery, are necessary to rule out retinal detachment.
interrupted sutures are used.
In virtually all instances of taut vitreous with hypocellular gel
The scleral portion should be closed by sequential sutur-
contraction, vitrectomy will be necessary. If the vitreous is clear,
ing, exploration, and further suturing. Unless extremely large
the retina must be observed frequently for signs of detachment
pieces of sclera are absent, primary closure is preferable to
and/or cellular proliferation. True vitreous “organization” is a
scleral grafting. As these cases frequently develop wound-
late sign and should not be allowed to occur.
related cellular proliferation, scleral resection-like shortening
that occurs from primary closure of a defect can be beneficial.
COMPLICATIONS
Tim in g o f Le n s Re m o val
In fe ct io n
Many factors relate to the question of whether lens removal
should occur at the time of primary surgical repair or later. Immediate surgical repair with subconjunctival antibiotic
Hypotony-induced choroidal edema, striate keratopathy, prophylaxis and removal of all necrotic exteriorized tissue
and miosis make safe lens removal difficult at the time of leads to a surprisingly low incidence of endophthalmitis.
initial repair. Arterial bleeding is also common, as is surgical- If suspected endophthalmitis does occur, it should be man-
induced corneal stromal swelling. Delayed lensectomy can aged as described elsewhere in this book.
be easier because the lens imbibes water, sliding endothelial
cells have closed the cornea, the pupil may dilate better, and
Co rn e al Op acifi cat io n
arterial bleeding has ceased. If ideal circumstances permit
lensectomy at the time of primary repair, this, of course, Modern wide-angle contact lens systems have nearly elimi-
would avoid two trips to the operating room. nated the need for penetrating keratoplasty and temporary
intraoperative keratoprosthesis in the management of severe
intraocular trauma with corneal lacerations or opacifications.
Ro le o f Po st e rio r Vit re ct o m y
A single 1-mm area of clear cornea is sufficient for adequate
Lacerations anterior to the ora serrata (muscle ring) do not cre- viewing of the retina for vitrectomy if a trained assistant is
ate a need for posterior vitrectomy in the early stages. These holding the contact lens. The coupling agent also neutralizes
214 SECTION IV ■ Disease-Sp ecific Managem ent
the irregular astigmatism created by the corneal wound and
improves visualization in comparison to noncontact indirect
visualization systems. If total corneal opacification is encoun-
tered, a temporary keratoprosthesis (26) could be utilized to
permit vitrectomy. This device allows visualization without a
fundus contact lens and allows replacement with the original
blood-stained cornea or a donor button if permanent opac-
ity is suspected. The principal author prefers, however, to
trephine the cornea and perform the entire procedure “open
sky” (Fig. 23.11), that is, lens (remnant) removal, vitrectomy,
IOFB removal, epiretinal and subretinal dissection, instilla-
tion of silicone oil through the corneal opening, and suturing
of the donor button. This approach is faster than use of the
temporary epikeratoprosthesis. It allows removal of very large
IOFBs, gentle dissection of epiciliary tissue, bimanual surgery,
subretinal dissection, aspiration of all intraocular fluid, and
easy instillation of silicone. Although endolaser can easily be
used open sky, many of these cases are candidates for retin-
opexy avoidance using medium- to long-term silicone. Per-
fluoron (PFO) can also be used open sky and provides inertial
and gravitational stabilization compared to air. PFO removal
will allow air to fill the eye at the end of the case, enabling sili-
cone to be poured into the eye through the corneal incision.
Glauco m a
Erythroclastic (hemolytic) glaucoma can be managed effectively
by vitrectomy (27) if medical management fails. “Phacogenic
glaucoma” is a general term including phacolytic and lens-
induced pupillary block. If unresponsive to medical treatment,
these forms of glaucoma can be managed effectively by
Fig u re 2 3 .1 1 ■ Op en -sky vit-
rectom y is p referred to th e use
of tem porary ep ikeratop rosthesis
if corneal op acification p revents
sufficient visualization.
vitrectomy and removal of lens material as well. Other forms of
traumatic glaucoma are well beyond the scope of this book.
References
1. Goldblum D, Frueh BE, Koerner F. Eye injuries caused by cow
horns. Retina 1999;19(4):314–317.
2. Cleary PE, Minckler DS, Ryan SJ. Ultrastructure of traction
retinal detachment in rhesus monkey eyes after a posterior
penetration ocular injury. Am J Ophthalmol 1980;90:829.
3. Cleary PE, Ryan SJ. Experimental posterior penetrating eye
injury in the rabbit 11. Histology of wound, vitreous, and ret-
ina. Br J Ophthalmol 1979;63:312.
4. Colyer MH, Weber ED, Weichel ED, et al. Delayed intraocular
foreign body removal without endophthalmitis during Opera-
tions Iraqi Freedom and Enduring Freedom. Ophthalmology
2007;114(8):1439–1447. [Epub 2007 Feb 28.]
5. Ryan SJ, Allen AW. Pars plana vitrectomy in ocular trauma. Am
J Ophthalmol 1979;88:483.
6. Ryan SJ. Results of pars plana vitrectomy in penetrating ocular
trauma. Int Ophthalmol 1978;1:5.
7. Ryan SJ. Guidelines in the management of penetrating ocular
trauma with emphasis on the role and timing of pars plana
vitrectomy. Int Ophthalmol 1979;1:105.
8. DeJuan E, Sternberg P, Michels RG, et al. Timing of vit-
rectomy after penetrating ocular injuries. Ophthalmology
1984;91:1072.
9. Chiquet C. Intraocular foreign bodies. Factors influencing final
visual outcome. Acta Ophthalmol Scand 1999;77(3):321–325.
10. Coday MP. Nailing down the diagnosis: imaging intraocular
foreign bodies Arch Ophthalmol 1999;117(4):548.
11. De Souza S, Howcroft MJ. Management of posterior segment
intraocular foreign bodies: 14 years’ experience. Can J Ophthalmol
1999;34(1):23–29.
12. Chiquet C, Zech J, Gain P, et al. Visual outcome and prognostic
factors after magnetic extraction of posterior segment foreign
bodies in 40 cases. Br J Ophthalmol 1998;82(7):801–806.

13. Kozielec GF, To K. Penetrating eye injury from a metal wedge.
Ophthalmic Surg Lasers 1999;30(1):59–60.
14. Azad R, Sharma VR, Mitra S, et al. Triple procedure in pos-
terior segment intraocular foreign body. Indian J Ophthalmol
1998;46(2):91–92.
15. Pavlovic S, Schmidt KG, Tomic Z, et al. Management of intra-
ocular foreign bodies impacting or embedded in the retina.
Aust N Z J Ophthalmol 1998;26(3):241–246.
16. Michels RG. Surgical management of non-magnetic intraocular
foreign bodies. Arch Ophthalmol 1975;93(10):1003.
17. Michels RG. Closed vitrectomy in trauma: selected intraocu-
lar foreign bodies. In: Freeman HM, ed. Vitreous surgery and
advances in fundus diagnosis and treatment. New York: Appleton-
Century-Crofts; 1977:335–344.
18. Cleary PE, Ryan SJ. Vitrectomy in penetrating eye injury.
Results of a controlled trial of vitrectomy in an experimental
posterior penetrating eye injury in the rhesus monkey. Arch
Ophthalmol 1981;99:287.
19. Hutton WL, Snyder WR, Vaiser A. Vitrectomy in the treatment of
ocular perforating injuries. Am J Ophthalmol 1976;B1(6):733.
CHAPTER 23 ■ Traum a 215
20. Mandelcorn MS. Results after vitrectomy for trauma. Can J
Ophthalmol 1977;12(1):34.
21. Benson WE, Machemer R. Severe perforating injuries treated
with pars plana vitrectomy. Am J Ophthalmol 1976;B1(6):728.
22. Michels RG. Early surgical management of penetrating ocu-
lar injuries involving the posterior segment. South Med J
1976;69(9):1175.
23. Conway BP, Michels RG. Vitrectomy techniques in the man-
agement of selected penetrating ocular injuries. Ophthalmology
(Rochester) 1978;85(6):560.
24. Michels RG, Conway BP. Vitreous surgery techniques in penetrat-
ing ocular trauma. Trans Ophthalmol Soc UK 1978;98(4):472.
25. Abrams GW, Topping TM, Machemer R. The effect of vitrec-
tomy on intraocular proliferation following perforating injuries
in rabbit eyes. Arch Ophthalmol 1978;96:521.
26. Landers MB, Foulks G, Landers DM, et al. Temporary keratopros-
thesis for pars plana vitrectomy. Am J Ophthalmol 1981;91:615.
27. Brucker AJ, Michels RG, Green WR. Pars plana vitrectomy in
the management of blood-induced glaucoma with vitreous
hemorrhage. Ann Ophthalmol 1978;10:1427.
C H AP TER
24
ENDOPHTHALMITIS
A broad definition of endophthalmitis includes
any severe intraocular inflammation. Toxic sub-
stances, necrotic tumors, noninfectious uveitis,
and infarction can create the clinical picture of
vitreitis, hypopyon, and ocular pain.
Infectious endophthalmitis can be of bacterial, fungal,
or parasitic etiology. Vitreous surgery reduces the number of
organisms; reduces the intravitreal load of neutrophils, mac-
rophages, lymphocytes, and soluble mediators; enhances the
penetration and diffusion of antibiotics; and aids in identifi-
cation of the pathogen. Late complications related to cellular
proliferation on the vitreous matrix are reduced as well.
Early diagnosis and treatment are of paramount impor-
tance when managing a patient with endophthalmitis. It is
strongly recommended that all ocular surgery postoperative
patients be examined on the first postoperative morning
as well as immediately if the patient complains of pain or
decreased vision. If the physician cannot personally exam-
ine the patient, immediate and definite arrangements can
be made with another qualified physician. The welfare of
the patient is always the surgeon’s responsibility. Once
endophthalmitis is suspected, one can examine the patient
often and take definitive action when indicated. While the
more common signs of endophthalmitis are well known
by physicians, less frequent signs are also important. These
include chemosis, lid edema, fibrin membrane formation on
the intraocular lens (IOL), and retinal hemorrhages.
ETIOLOGIC SUBGROUPS
En d o g e n o us En d o p h t h alm it is
Endogenous endophthalmitis accounts for a minority of
cases (1). Predisposing conditions include immune defi-
ciency, immune suppression, diabetes mellitus, chronic renal
failure, IV drug abuse, and patients receiving hyperalimenta-
tion. These patients can develop endophthalmitis without
prior ocular disease. Such cases may be bilateral, increasing
both the impact on the patient and the management diffi-
culties. Systemic workup and therapy play an extensive role
216
in the management of these patients and require infectious
disease consultation. Metastatic infection accounts for appro-
ximately 8% of endogenous bacterial endophthalmitis. At
times, vitrectomy is indicated in this patient group but is
quite difficult because of the associated anesthesia risk and
the potential bilaterality of the disease. Eyes with endog-
enous endophthalmitis have increased penetration of sys-
temic antibiotics through breakdown of the blood-retinal
barrier, and intravenous antibiotics may be sufficient to treat
the disease if the intraocular inflammation is not severe. On
the other hand, whenever bilateral endogenous endophthal-
mitis is diagnosed and rapid progression of the diseases is
noticed or one of the eyes has progressed to severe visual
loss, the authors often recommend vitrectomy for the less
involved eye to prevent bilateral blindness. Adequate coordi-
nation with the infectious disease consultants is paramount.
If the patient has a known systemic infection, the appropri-
ate intravitreal antibiotic therapy can be chosen. If the infec-
tious agent has not been identified, vitrectomy can permit
isolation of the bacteria and guide systemic treatment.
Exo g e n o us En d o p h t h alm it is
Ocular trauma (2,3) accounts for approximately 20% of bac-
terial endophthalmitis (4,5). Ocular surgery (6–8) accounts
for the vast majority (~70%) as it facilitates the introduction
of organisms into the eye. Management of trauma cases usu-
ally requires vitreous surgery and may involve removal of
intraocular foreign bodies. While conventional surgical wis-
dom suggests removal of any implanted materials if infection
occurs, this logic does not apply to the IOL. Removal of an
IOL, especially in an endophthalmitis case, has a risk of iris
avulsion, endothelial trauma, intraocular bleeding, choroidal
expulsive hemorrhage, and retinal detachment. Vitrectomy
with intraocular antibiotics without implant removal can be
successful in a high percentage of cases. This is probably
because the smooth surface of the lens implant coupled with
the high fluid throughput of vitrectomy removes all organ-
isms on the lens surface, but organisms may remain in the
peripheral capsular bag.

Fortunately, the incidence of acute postoperative
endophthalmitis remains relatively low. Extracapsular cata-
ract surgery with or without an IOL insertion carries an inci-
dence of 0.072% (9). Vitrectomy (0.051%) and penetrating
keratoplasty (0.11%) have a far less risk than the insertion
of a secondary IOL (0.30%). Sutureless clear corneal inci-
sions and temporal placement of phacoemulsification inci-
sions may have an increased risk of postoperative infectious
endophthalmitis (10). The most frequent organisms include
coagulase-negative Staphylococcus, Staphylococcus aureus, Strep-
tococcus species, and Gram-negative bacteria.
Delayed-onset, post–cataract surgery endophthalmitis
is most often caused by Propionibacterium acnes, coagulase-
negative Staphylococcus, or Candida. In the setting of P. acnes
endophthalmitis, inflammatory plaque and associated capsule
must be removed at a minimum; in most instances, the IOLmust
be removed as well (11).
The incidence of trauma-related endophthalmitis varies
with or without the presence of a retained foreign body. Follow-
ing a penetrating injury, the incidence of endophthalmitis ranges
from 3.2% to 7.4%. With a retained foreign body, the incidence
jumps to 6.9% to 13% (metallic—7.2%, nonmetallic—7.3%,
organic—6.3%). The most frequent organisms include Bacillus
sp., Staphylococcus sp., and Streptococcus sp.
Streptococcus, coagulase-negative Staphylococcus, and
Haemophilus are the most common organisms in delayed-onset
filtering bleb–related cases. Immediate vitrectomy interven-
tion is mandatory in these often rapidly deteriorating cases.
Therapeutic intravitreal injections with anti–vascular
endothelial growth factor (VEGF) compounds are wide-
spread but fortunately produce a low incidence of postinjec-
tion endophthalmitis when meticulous sterile technique is
utilized. Fortunately, the incidence of endophthalmitis after
intravitreal triamcinolone injections, although higher than
anti-VEGF compounds, appears to be low (~0.1%–0.9%)
(12,13).
TASS
Toxic anterior chamber syndrome must be distinguished
from infectious endophthalmitis because the management
and causation implications are vastly different. The inci-
dence after cataract, glaucoma, and corneal surgery is 0.1%
to 2.0%, increasing (14) worldwide, and occurring in clus-
ters because of institutional outbreaks. Toxic Anterior Seg-
ment Syndrome (TASS) usually is diagnosed within a few
hours after surgery; all patients have blurred vision, severe
limbus-to-limbus corneal edema (surgical trauma produces
localized edema), 4+ flare, possible hypopyon, fibrin, mini-
mal conjunctiva or episcleral hyperemia, and minimal or
no pain (75%–80% of endophthalmitis patients have pain).
Although TASS typically presents on the first postopera-
tive day, virulent bacterial endophthalmitis can also present
on the first postoperative day. Outcomes include glaucoma
due to Peripheral Anterior Synechiae (PAS) and trabecular
CHAPTER 24 ■ Endophthalm itis 217
meshwork damage, chronic inflammation, CME, corneal
endothelial damage, and iris damage (fixed pupil and tran-
sillumination defects).
Causes in general include drugs and devices, process
issues, and surgical issues. Device and drug factors include
poor-quality infusion solutions (Cytosol and others), intraoc-
ular lidocaine, intraocular epinephrine (pH, dose, preser-
vative, stabilizing agents: bisulphites or metabisulphites),
intraocular antibiotics (pH, concentration, dose), intraocu-
lar Kenalog (active drug and preservative), and poor-quality
viscoelastics. Process issues include contamination within
steam sterilizer (filters, water chambers, and the inside of
autoclave), denatured viscoelastics from resterilized cannu-
las, bacterial endotoxins from Gram-negative bacteria killed
when cannulas are autoclaved, and contamination of ultra-
sonic cleaner and water baths with enzymes or detergents.
Devices designated for single use should not be reused.
Additional process causes include endotoxins from tap water,
detergent or chemical residue on instruments, and incom-
plete cleaning of reusable cannulas. Surgical factors include
retained lens cortex, iris trauma, iris retractors, viscoelastics,
intracameral agents, IOLs, and endocyclophotocoagulation.
There are many advantages of disposable forceps, scis-
sors, pics, and cannulas: small-diameter 25- or 23-gauge
tools are easily damaged during cleaning and sterilization,
optimal gripping, and cutting performance beginning to end
of every case, and there is no risk of contamination (Transmis-
sible Spongiform Encephalopathy (TSE)/Bovine Spongiform
Encephalopathy (BSE)/prions, bioburden, hepatitis, HIV/
AIDS, denatured proteins, bacterial endotoxins). Scissors and
forceps all have a lumen, and because the IOP is greater than
atmospheric pressure, proteins are forced into the lumen.
Disposable tools result in a reduced cost per case because of
the elimination of cleanup, sterilization, packaging, storage,
backup, and inventory costs including labor and materials.
The initial workup should include an anterior cham-
ber tap if there is any suspicion of bacterial endophthal-
mitis; it will be negative if TASS is the cause. Vitreous tap
should be performed if there is any vitreous clouding or
clumped inflammatory deposits. If an infectious etiology is
suspected, treat with tap and inject intravitreal antibiotics;
vitrectomy is unnecessary. If the surgeon has a significant
level of confidence in the diagnosis of TASS, frequent topical
antibiotics and close observation are usually enough to con-
trol inflammation and recuperate vision. On the other hand,
a surgeon will never be faulted for performing a vitreous tap
and intravitreal antibiotic injection if the diagnosis of infec-
tious endophthalmitis is considered.
ANTERIOR CHAMBER VERSUS
VITREOUS ASPIRATION
Vitreous taps have a much higher incidence of positive
cultures than anterior chamber taps (15). Both types of
taps risk pain, wound disruption, intraocular hemorrhage,
218 SECTION IV ■ Disease-Sp ecific Managem ent
and retinal detachment. Because of these problems, many
clinicians recommend performing the tap in the operating
room or a minor procedure room. Vitreous samples obtained
at the time of therapeutic vitrectomy have higher yields and
are safer than vitreous taps. The authors use tap and inject
intravitreal antibiotics to manage most cases. Busy office and
operating room schedules and the concern about contamina-
tion of other cases can lead to a delay between the suspected
diagnosis and the performance of the tap. The rapid progres-
sion of this disease does not warrant any delay in the onset
of therapy. Given the typical delay between the diagnosis
of endophthalmitis and the start of vitrectomy, the authors
nearly always choose to perform an injection of intravitreal
antibiotics immediately after the diagnosis of endophthal-
mitis even if the situation warrants vitrectomy. Taps should
only be done if they are accomplished immediately and with
the intent of not performing vitreous surgery at that time.
Suspicion of the operating room environment, IOL, or surgi-
cal materials as an etiologic agent demands an epidemiologic
approach to the workup. In general, the goal should be to
initiate treatment immediately.
TECHNIQUE FOR IN-OFFICE
VITREOUS TAP AND INJECTION
OF INTRAVITREAL ANTIBIOTICS
Anesthesia with a peribulbar injection of lidocaine and topi-
cal viscous lidocaine is necessary for this procedure. Retrob-
ulbar anesthesia in the office without intravenous sedation
in a very anxious patient with acute ocular pain is usually
not recommended.
Sterile techniques should be utilized the same way that
intravitreal injections are performed, with 5% iodine on the
conjunctival surface and eyelids and the use of a bladed
speculum to cover the lashes. The main difference in tech-
nique revolves around these factors: (a) multiple medications
are usually injected (typically vancomycin, ceftazidime, and
dexamethasone), (b) these medications should not be mixed
in the same syringe since they can precipitate, (c) a success-
ful vitreous tap requires a larger caliber needle, usually 25
gauge or larger, and (d) the acutely inflamed eye is very dif-
ficult to anesthetize using topical anesthetics.
A vitreous tap and injection of all the required medi-
cations can be performed with a single needle penetration
of the sclera. As the first step, a 3-mL syringe with a short
25-gauge needle is firmly grasped at the hub of the needle
with a locking mosquito hemostat, perpendicular to the axis
of the needle. The needle is introduced into the pars plana
and gentle suction is applied to the embolus, withdrawing
the vitreous sample. In a nonvitrectomized eye, it is usu-
ally difficult to withdraw more than 0.1 mL of vitreous. In a
previously vitrectomized eye, if the tip of the needle is intro-
duced beyond the vitreous base, it is easy to withdraw up to
0.3 mL of liquid vitreous. The surgeon can then firmly hold
the hemostat that is immobilizing the hub of the needle and
unscrew the syringe with the vitreous sample and proceed
to sequentially screw the syringes with the medications and
inject all the antibiotics and steroids necessary without per-
forming a second penetration of the sclera. Great care should
be taken to lock the surgeon’s hand holding the hemostat
to the patient’s face to avoid injury to the eye if the patient
unexpectedly moves his or her head.
TIMING OF VITRECTOMY
Vitreous surgery is not required for every case of infectious
endophthalmitis. Patients with acute-onset postoperative
endophthalmitis and vision of hand motion or better may
be treated with tap and injection of intravitreal antibiotics
alone, as shown by the Endophthalmitis Vitrectomy Study
(EVS) study. If the vision is worse than hand motion, vitre-
ous surgery should be undertaken immediately, regardless of
the hour of the day or night.
Advanced cases with corneal decompensation cannot
have vitreous surgery under safe conditions because of the
visualization requirements. Open-sky vitrectomy would be
fraught with complications and is not recommended. Imme-
diate intracameral antibiotics offer the best hope in this poor
prognosis group and in patients with medical problems pre-
venting surgical intervention.
EVS findings do not apply to endophthalmitis cases
that are endogenous, delayed onset postoperative, traumatic,
or filtering bleb related.
SURGICAL SEQUENCE AND
TECHNIQUES
The operating room should be alerted to prepare for a
dirty case with all appropriate isolation and postsurgical
cleanup precautions. This should not be used as an excuse
for delay of therapy, however. At times, it is best to operate
in a nonophthalmic operating room with only the minimal
equipment required for the case.
Monitored local anesthesia (MAC) is appropriate in
these cases. Intravenous antibiotics should be started at this
time if a presurgical medical treatment plan has not been
utilized, but the role of systemic antibiotics is controversial.
If a cataract wound, surgical wound, or ruptured filter-
ing bleb is present, it must be secured before the initiation
of vitreous surgery. Absorbable sutures should be removed
and replaced with 10-0 or 9-0 monofilament nylon sutures.
Careful attention to a tight wound prior to vitrectomy can
decrease subsequent problems.
Tran s–Pars Plan a Ve rsus Tran slim b al Ap p ro ach
As many endophthalmitis cases can and should be handled
by predominantly anterior segment surgeons, the lim-
bal approach can be considered. Unfortunately, the limbal
approach prevents adequate visualization for posterior

vitrectomy and causes more corneal and iris trauma.
Translimbal vitrectomy should be reserved for the novice
surgeon only in a true emergency. Translimbal vitrectomy
is virtually impossible in patients with posterior chamber
IOLs, which represents the biggest subgroup of endophthal-
mitis cases.
Vit re ct o m y
The vitrectomy instrument should be used with the lowest
possible suction force, with the highest cutting rate, preferably
5,000 cuts per minute, and proportional suction control to
reduce the chance of tearing necrotic retina and iris. A 25-gauge
vitrectomy is ideal for this application, but 23-gauge vitrec-
tomy can be used as well. Utilize the highest possible cutting
rate, and never pull the cutter away from the retina while suc-
tion is being applied. The anterior vitreous cortex should be
removed first, with special care taken to avoid iris contact and
peripheral vitreoretinal traction. The iris tissue can be quite
necrotic and may be easily shredded or avulsed. On occasion,
hypotony will lead to oozing from iris vessels, requiring bipo-
lar diathermy. If a fibrin membrane covers the anterior surface
of the IOL, it can be removed through a peripheral iridectomy
constructed with the vitreous cutter. Viscoelastic can also be
injected through a small limbal paracentesis to displace fibrin
from the surface of the intraocular lens. Limbal incisions fre-
quently leak, causing hypotony and miosis.
Endophthalmitis is one of the rare situations in which
only a “core” vitrectomy should be done to avoid traction
on the potentially necrotic retina. At least 50% of the
formed vitreous is removed with this approach. The dense
initial vitreous aspirate should be removed for smear,
culture, and sensitivity testing (16,17). Blood culture bottles
have been shown by Joondeph and Flynn et al. (18) to be
equally effective at achieving a positive culture compared to
inoculating culture plates and tubes in the operating room. The
endoilluminator is essential to adequate visualization. Vacuum
cleaning (extrusion) and membrane peeling should never be
utilized in these cases because of the necrotic retina. Creation
of an intraoperative posterior vitreous detachment can also
lead easily to iatrogenic retinal tears. If there is severe retinal
necrosis, the retina will appear white and rough surfaced with
the vessels appearing dull. Such retinas are extremely prone to
retinal breaks and can be seen to move with the probe several
millimeters away from the retina, even with the mild pulsatile
suction force of the vitrectomy instrument.
USE OF ANTIBIOTICS
In t racam e ral An t ib io t ics
Intraocular antibiotics are used in virtually all cases (19).
There is currently less disagreement in the literature about
safe doses of intraocular antibiotics than in previous years.
Because of the risk of retinal toxicity, the lowest published
CHAPTER 24 ■ Endophthalm itis 219
effective dose should be used. Intraocular antibiotics should
always be used if an IOL is present. Antibiotics diluted in the
infusion fluid are not recommended because of the toxicity
question and the difficulty in assessing total retinal dose with
this method. Vancomycin 1.0 mg/0.1 mL and ceftazidime
2.25 mg/0.1 mL are the most commonly used agents today.
It has been shown by D’Amico et al. (20) that pharmacists
mix antibiotics more accurately than nurses and the nurses
are far more accurate than physicians. Many toxicity cases
are probably due to incorrect concentrations and/or volumes
being injected. Intraocular dexamethasone 0.4 mg/0.1 mL
has been recommended by many investigators and has been
shown to produce better outcomes (21).
The antibiotic should then be injected slowly in the mid-
vitreous cavity through a cannula with a 25-gauge needle. If
two to three separate injections are utilized, as is usually the
case, these can be applied sequentially in the same wound
with this method. Mixing the antibiotics in the same syringe
is poor practice and results in dilution and precipitation.
Sub co n jun ct ival An t ib io t ics an d St e ro id s
After removal of 25-gauge cannulas, subconjunctival antibi-
otics can be considered. Vancomycin 25 mg and ceftazidime
100 mg are the most widely used agents at the time of this
writing. Detailed discussion of antibiotic options is left to
other publications because of their rapidly changing status.
It is recommended that an antibiotic that is primarily for
Gram-positive organisms and effective on penicillinase pro-
ducers be combined with an antibiotic that is primarily for
Gram-negative organisms. Substances released from inflam-
matory cells, as well as bacteria, cause severe tissue destruc-
tion and inflammation. The inflammation suppression role
of steroids is thought to far outweigh the potential enhance-
ment of infection. If the case is strongly thought to be fungal
in origin, then steroids should probably be omitted.
Although the authors recommend using prophylactic
subconjunctival antibiotics routinely after vitrectomy sur-
gery (whether for the treatment of endophthalmitis or other-
wise), subconjunctival antibiotics do not alter the course of
endophthalmitis adequately treated with intravitreal antibi-
otic injections or surgery (22).
To p ical An t ib io t ics
Topical antibiotics and steroids are typically used in all cases,
primarily to inhibit potential, associated lid, cul-de-sac, bleb,
and wound infections (23). Care should be taken to prevent
cross contamination by doctors, nurses, and technicians tak-
ing care of other patients.
RESULTS
The causative organisms and pretreatment delay constitute
the most important factors in therapeutic success. If cases
with mild pathogens that could have been managed with
220 SECTION IV ■ Disease-Sp ecific Managem ent
subconjunctival and systemic antibiotics are operated upon,
the success rate appears to improve. Even with the best man-
agement, a significant percentage of these eyes will be lost. In
view of the overall poor prognosis in endophthalmitis cases,
major emphasis should be placed on prevention. Meticulous
preparation of the operative field is vital, including a drape
that covers the lashes and lid margins, Betadine prep, micro-
scope draping, use of the highest quality infusion fluids (Alcon
BSS, BSS Plus), and extreme care if tubing or any other instru-
ments with a lumen or cavity are resterilized. Subconjunctival
antibiotics are used at the end of vitrectomy cases, which may
decrease the incidence of this dreaded complication.
COMPLICATIONS
Corneal edema, glaucoma, and retinal detachment can be
seen after otherwise successful endophthalmitis manage-
ment with or without vitrectomy. Prompt therapy is prob-
ably the single most important factor in the prevention of
these nonspecific complications.
Retinal breaks and detachment related to retinal necro-
sis and surgically induced retinal traction occur in at least
10% of the cases. Postoperative follow-up includes periph-
eral retinal examination by indirect ophthalmoscopy because
of the relatively high frequency of retinal breaks.
References
1. Romero CF, Rai MK, Lowder CY, Adal KA. Endogenous
endophthalmitis: case report and brief review. Am Fam Physi-
cian 1999;60(2):510–514.
2. Brinton GS, Topping TM, Hyndiuk RA, et al. Post-traumatic
endophthalmitis. Arch Ophthalmol 1984;102:547.
3. Forster RK. Endophthalmitis. In: Duane TD, ed. Clinical oph-
thalmology, Vol. 4. New York: Harper & Row, 1981:1–20.
4. Abu el-Asrar AM. Post-traumatic endophthalmitis: causative
organisisms and visual outcome. Eur J Ophthalmol 1999;9(1):
21–31.
5. Meredith TA. Posttraumatic endophthalmitis. Arch Ophthalmol
1999;117(4):520–521.
6. Forster RK. Etiology and diagnosis of bacterial post-operative
endophthalmitis. Ophthalmology 1978;85:320.
7. Schmitz S, Dick HB, Krummenauer F, Pfeiffer N. Endophthal-
mitis in cataract surgery: results of a German study. Ophthal-
mology 1999;106(10):1869–1877.
8. Solomon A, Ticho U, Frucht-Pery J. Late onset bleb associ-
ated endophthalmitis following glaucoma filtering surgery
with or without antifibrotic agents. J Ocul Pharmacol Ther
1999;15(4):283–293.
9. Cottingham AJ, Forster RK. Vitrectomy in endophthalmitis;
results of study using vitrectomy, intraocular antibiotics, or a
combination of both. Arch Opthalmol 1976;94:2078.
10. Nagaki Y, Hayasaka S, Kadoi C, et al. Bacterial endophthal-
mitis after small-incision cataract surgery. Effect of incision
placement and intraocular lens type. J Cataract Refract Surg
2003;29(1):20–26.
11. Clark WL, Kaiser PK, Flynn HW Jr, et al. Treatment strate-
gies and visual acuity outcomes in chronic postoperative
Propionibacterium acnes endophthalmitis. Ophthalmology
1999;106(9):1665–1670.
12. Westfall AC, Osborn A, Kuhl D, et al. Acute endophthal-
mitis incidence: intravitreal triamcinolone. Arch Ophthalmol
2005;123(8):1075–1077.
13. Moshfeghi DM, Kaiser PK, Scott IU, et al. Acute endophthalmi-
tis following intravitreal triamcinolone acetonide injection. Am
J Ophthalmol 2003;136(5):791–796.
14. Kim SY, Park YH, Kim HS, Lee YC. Bilateral toxic anterior
segment syndrome after cataract surgery. Can J Ophthalmol
2007;42(3):490–491.
15. Results of the Endophthalmitis Vitrectomy Study. Arch Oph-
thalmol 1995;113(12):1479–1496.
16. Frederick AR. A modification of the Ocutome setup to permit
the sterile collection of intraocular specimens. Ocutome Frag-
matome Newsletter No. 3,1978:2.
17. Smith RE. A technique for collecting uncontaminated speci-
mens for culture during vitrectomy for endophthalmitis. Ocu-
tome Fragmatome Newsletter No. 3, 1978:2.
18. Joondeph BC, Flynn HW, Miller D, Joondephet HC. A new cul-
ture method for infectious endophthalmitis. Arch Ophthalmol
1989;107(9):1334–1337.
19. Peyman GA, Vashine DW, Crouch ER, Herbst RW. Clinical use
of intravitreal antibiotics to treat bacterial endophthalmitis.
Trans Am Acad Ophthalmol Otolaryngol 1974;8:862.
20. D’Amico DJ, Caspers-Velu L. Comparitive toxicity of intravit-
real aminoglycoside antibiotics. Am J Ophthalmol 1985;100(2):
264–275.
21. Park SS, Vallar RV, Hong CH, et al. Intravitreal dexamethasone
effect on Vancomycin elimination in endophthalmitis. Arch
Ophthalmol 1999;117(8):1058–1062.
22. Smiddy WE, Smiddy RJ, Ba’Arath B, et al. Subconjunctival
antibiotics in the treatment of endophthalmitis managed with-
out vitrectomy. Retina 2005;25(6):751–758.
23. Bannerman TL, Rhoden DL, McAllister SK, et al. The source of
coagulase-negative staphylococci in the EVS. Arch Ophthalmol
115(3):357–361.
C H AP TER
25
SURGICAL MANAGEMENT
OF THE UVEITIS PATIENT
VITREORETINAL SURGERY
IN THE MANAGEMENT OF
UVEITIC DISORDERS
Vitrectomy is recommended in certain situations for the
management of uveitic diseases. Vitrectomy may be needed
for diagnostic purposes as well as therapeutic purposes.
Diag n o st ic Vit re ct o m y
Indica tions
The three main indications for diagnostic vitrectomy in
uveitis are (a) severe vitreitis that precludes visualiza-
tion of the fundus and diagnosis is not clear, (b) suspi-
cion of intraocular lymphoma, and (c) retinochoroiditis of
unclear etiology that threatens vision in the short term and
(d) cystoid macular edema (CME) unresponsive to medical
management.
Vitreous Sa mpling Techniques
Whenever diagnostic vitrectomy is performed, great care
should be taken with the vitreous sample, which, after all, is
the main goal of the surgery. There are variations in the sur-
gical technique that can provide adequate vitreous samples.
The authors will present the techniques that they find pref-
erable in their hands. Surgery begins by placing the infer-
otemporal 25-gauge cannula and attaching to it the infusion
cannula in the off position (without infusion). Following the
placement of the superior 25-gauge cannulas, removal of
undiluted vitreous sample can begin. For this step, the aspi-
ration line coming from the vitrectomy probe is disconnected
from the Constellation vitrectomy system and attached to a
syringe. With adequate visualization with the endoillumina-
tor and plano irrigating contact lens, the surgeon engages
the cutting action of the vitrectomy probe and immediately
after orders the assistant to begin gentle aspiration with the
syringe on the line. It is important to have the assistant wait
for cutting to begin before aspirating on the syringe, thereby
preventing peripheral retinal tears. The surgeon can then
direct the sampling of the vitreous to the area that has the
most apparent cellular infiltration. This increases the yield
of the vitreous biopsy and increases safety since the retina is
visualized at all times during the dry aspiration. As soon as
the eye appears to become soft, the surgeon can direct the
assistant to stop manual aspiration and then disengages the
cutter action. Immediately the fluid infusion can be initiated
and the intraocular pressure normalized. Appropriate order
of steps is important to decrease the potential for iatrogenic
retinal tears.
Some surgeons prefer to use air infusion during the
undiluted vitreous biopsy, trying to decrease intraopera-
tive hypotony. The authors’ main objection to this tech-
nique is that it compromises visualization during the
vitrectomy and prevents adequate positioning of the cutter
on the most pathological area of the eye. Once the undi-
luted vitreous sample has been obtained, standard vitrec-
tomy can continue with normal pars plana infusion. The
vitrectomy cassette can then be sent directly to the labora-
tory or a syringe can be used to aspirate from the cassette
to be sent.
As in all vitrectomy cases, great care should be taken
at the end of the surgery to visualize the retinal periph-
ery and verify the absence of peripheral retinal tears. The
creation of a posterior vitreous detachment (PVD) dur-
ing surgery has potential pitfalls. If there are any areas of
retinitis, the likelihood of retinal tears during PVD creation
is high. On the other hand, patients with dense vitreitis
who undergo core vitrectomy will have postoperative PVDs
and can often bitterly complain of floaters. Surgical safety
should override minor issues such as postoperative float-
ers. The authors do not recommend performing PVDs rou-
tinely during diagnostic vitrectomies, but do explain to
patients the possibility of floaters. As long as the patient
understands that the reason for the surgery is to prevent
blindness, issues such as floaters can be placed in their
appropriate perspective.
221
222 SECTION IV ■ Disease-Sp ecific Managem ent
La bora tory Studies
The authors cannot overemphasize the importance of
personal communication with the pathologist prior and
following diagnostic vitrectomies. Given the small volumes
of the authors’ samples, protocol-based testing can poten-
tially waste highly valuable samples on unimportant tests.
Since clinical laboratory techniques are constantly changing
and improving, the vitrectomy surgeon may be unaware of
the availability of certain tests that can provide the needed
diagnosis. The authors strongly advocate personally explain-
ing to the pathologist the differential diagnosis and following
the pathologist’s recommendations regarding handling of the
material and prioritizing of studies.
Cytology and Flow Cytom etry for Diagnosis
of Intraocular Lym phom a
The diagnosis of intraocular B-cell lymphoma requires
vitreous biopsy for cytology and flow cytometry analysis (1).
Whereas many surgeons in the past have delayed diagnostic
vitrectomy in lieu of empirical steroid therapy and/or neu-
roimaging with lumbar puncture, the authors routinely per-
form 25-gauge diagnostic vitrectomies on elderly patients
with new-onset vitreitis if a syndromic clinical diagnosis is
unclear. Steroid therapy risks partial clinical improvement
with delayed diagnosis. Neuroimaging and lumbar puncture
should be recommended on patients with headaches or neu-
rological signs or symptoms (2), but they are of low yield
on patients who have isolated vitreitis without neurological
complaints. Cytological examination should be done by an
experienced cytopathologist, and no delay in the transfer
of the sample should occur, since the lymphoma cells can
die and render a biopsy useless if they are left out waiting
for someone to pick up the sample to take it to the labo-
ratory. Since the authors communicate with the pathologist
prior to every diagnostic vitrectomy for possible lymphoma,
the pathologist comes into the operating room and receives
the undiluted sample for immediate processing before the
authors finish the core vitrectomy, increasing the yield of the
study.
Gene Rearrangem ent Studies for Lym phom a Diagnosis
Modern cytogenetic studies can be performed on vit-
reous lymphocytes whenever the diagnosis of intraocular
lymphoma is unclear. Experienced laboratories can provide
a final lymphoma diagnosis with only a few cells analyzed
based on gene rearrangement studies (3).
Polym erase Chain Reaction for Infectious Diseases
Polymerase chain reaction (PCR) testing can confirm the
presence of viral DNA in the setting of unusual viral retinitis
such as atypical acute retinal necrosis syndrome, progressive
outer retinal necrosis syndrome, and cytomegalovirus retini-
tis (4). Despite the authors’ opinions against protocol-based
testing, personal experiences with misdiagnosed viral retini-
tis in which PCR testing was not performed since they were
low on the authors’ initial differential diagnosis have led the
authors to recommend PCR testing for herpes simplex virus,
varicella zoster virus, and cytomegalovirus whenever there
is sufficient sample to run these tests after vitreous biopsy.
Again, direct communication with the pathologist is the
authors’ best tool to increase efficient utilization of valuable
and scant vitreous material.
Re t in al Bio p sy
There are clinical scenarios that require retinal and even
choroidal biopsy for final diagnosis. The potential risks of
postoperative retinal detachment and intraoperative hemor-
rhage should limit the frequency of these complex cases.
Indica tions
The indication for performing a retinal biopsy is progressive
undiagnosed relentless retinitis that cannot be diagnosed
with less invasive means. The availability of PCR testing for
viral infections and modern flow cytometry and gene rear-
rangement studies as discussed previously have decreased
the need for retinal biopsies. The authors usually only per-
form a retinal biopsy in the setting of bilateral retinitis in
which one eye has progressed to profound vision loss and
the contralateral eye has good vision but is in danger. In this
circumstance, retinal biopsy of the blind eye has little risk of
worsening the clinical condition and can potentially provide
a diagnosis that can save the good eye.
Techniques for Retina l Biopsy
Follow the steps mentioned previously for vitreous biopsy
(the authors recommend to always send vitreous for labora-
tory analysis in cases where retinal biopsies are performed).
The next step is the selection of the biopsy site. The surgeon
should remember that biopsies of necrotic retinal tissue are
usually not diagnostic. The best location for biopsy is at the
leading edge of the retinitis, where viable retinal cells with
pathological changes can increase diagnostic yield. If there
is no leading edge of disease, perhaps from diffuse retinal
involvement, the most comfortable location for biopsy can
be selected. Immunohistochemical staining of necrotic tissue
can still be positive for viral antigens.
Retinal biopsies are technically easier when performed
around the posterior pole. The authors usually choose a site
just peripheral to the vascular arcades or nasal to the optic
nerve. The authors first perform three rows of retinal laser
surrounding the designated biopsy site and later cut the reti-
nal biopsy free hand with 25-gauge scissors. Retrieval of the
biopsy tissue can sometimes be challenging. The authors
prefer to remove a 25-gauge cannula, perform a small con-
junctival incision for exposure of the sclerotomy site, and
enlarge the sclerotomy with a 20-gauge microvitreoretinal
blade. 25-G ILM forceps can then be used to grasp the reti-
nal tissue and retrieve it atraumatically from the eye through
 CHAPTER 25
the 20-gauge incision, which is later sutured. End-grasping
forceps should not be used because they have a larger
grasping platform and can damage the retinal tissue.
A similar technique can be performed to remove chor-
oidal tissue, but it may require endodiathermy of the chor-
oid surrounding the biopsy site to decrease intraoperative
and postoperative hemorrhage. Following retinal biopsy
on previously attached retina, the authors perform fluid-air
exchange to decrease postoperative retinal detachment.
The fixation media should be chosen after discussion
with the pathologist to verify that appropriate special stains
are not ruined.
Th e rap e ut ic Vit re ct o m y
Vitrectomy should be seen as a potential treatment strategy
in the therapy of severe posterior uveitis. The most common
scenarios are described in the following sections.
Adjuva nt Thera py in the Ma na gement of
Intermedia te Uveitis
Vitrectomy surgery can positively affect the course of chronic
idiopathic intermediate uveitis (pars planitis) (5). Patients with
unresponsive CME can have decreased edema and improve-
ment in vision following vitrectomy (6). The explanation for
the improved clinical course in pars planitis following vit-
rectomy is not well understood. The authors postulate that
a vitreous antigen may be the target of the immune response
and that vitrectomy may debulk this antigen and improve
the clinical inflammation. Considering that 25-gauge vitrec-
tomy is a safe procedure when done by an experienced sur-
geon, the authors strongly support the idea of progressing to
complete pars plana vitrectomy before immunosuppressive
therapy is considered. The risk profile of a well-performed
vitrectomy is much better than the systemic complications
associated with immunosuppressive therapy.
Ma na gement of Complica tions of Severe Uveitis
Vitreoretinal surgery may be required for the management
of complications of posterior segment uveitis. The following
scenarios are the most important clinically.
Rhegm atogenous Retinal Detachm ents
in Necrotizing Retinitis
Viral retinitis that develops retinal necrosis can frequently
complicate with large retinal holes in the areas of the retinal
necrosis that progress to rhegmatogenous retinal detachment.
Prophylactic Retinopexy Around Attached
Necrotic Retina
Many authors have suggested performing laser demar-
cation posterior to the areas of retinal necrosis to decrease
the likelihood of ulterior retinal detachments (7). Despite
acknowledging this technique, the authors do not routinely
■ Surgical Managem ent of the Uveitis Patient 223
recommend laser demarcation in necrotizing retinitis for
the following reasons: (a) if done during active disease,
the surgeon may underestimate the final extent of the reti-
nal necrosis, (b) if the area of retinal necrosis is small, the
risk of detachment is low, and (c) on the other hand, those
patients with extensive retinal necrosis who have a high risk
of progressing to retinal detachment have such compromise
of visual fields that the surgeons do not advocate further
encroachment of their field of vision by laser photocoagula-
tion. Rhegmatogenous retinal detachments that present after
retinal necrosis are usually relatively easy surgical repairs,
which, if done, promptly may not damage visual function
significantly. The authors do, therefore, thoroughly explain
to patients the risk of retinal detachment and the need for
immediate communication with them in case of visual
changes suggestive of retinal detachment.
Technical Considerations for Repair of Retinal
Necrosis–Associated Retinal Detachm ents
These retinal detachments usually have broad areas
of retinal holes and atrophy that extend posteriorly. Scleral
buckling of these large holes is usually impossible. The
authors perform 25-gauge vitrectomy and fluid-air exchange
with simultaneous drainage of subretinal fluid through the
retinal holes to achieve retinal reattachment. The authors
usually do not require perfluorooctane or posterior drainage
retinotomies given the presence of posterior retinal defects.
Use of long-term silicone oil rhegmatogenous confinement is
usually required for permanent retinal reattachment. Periph-
eral vitrectomy should be undertaken with great care, at the
highest cut rate possible and with the lowest suction that
is sufficient for vitreous removal, since iatrogenic retinal
tears are easily created on these atrophic retinas. Intraop-
erative laser retinopexy versus postoperative laser treatment
is decided based on the clinical likelihood of proliferative
vitreoretinopathy. The authors tend to defer laser retinopexy
whenever there are any remaining inflammatory signs in the
eye until total resolution of inflammation is determined. This
decreases the stimulus for PVR and limits the extent of the
laser treatment only to the areas that absolutely require it.
Traction Retinal Detachm ents in Interm ediate
and Posterior Uveitis
Severe pars planitis can develop neovascularization of
the disc (NVD) and retinal periphery (8). Similar to NVDs
in diabetes, contraction of the NVD with attached posterior
hyaloid can develop tractional retinal detachments. Traction
retinal detachment without retinal neovascularization can
also be seen in any uveitic condition that develops severe
vitreous inflammation in an eye without preexisting PVD.
This may occur in pars planitis, toxoplasmosis, toxocariasis,
and sarcoidosis.
Technical Considerations
The surgical goal for repair of traction retinal
detachments in uveitis, as in any traction retinal detachment,
224 SECTION IV ■ Disease-Sp ecific Managem ent
is relief of all traction vectors that are acting on the retina
without creation of retinal tears in the process. Contraction of
NVDs in pediatric patients can present as an unusual traction
detachment where the papillomacular bundle and the nasal
retina immediately adjacent to the optic nerve adhere to each
other, covering the underlying optic nerve head. An inexpe-
rienced examiner may assume that the origin of this traction
configuration is the peripheral retina and attempt to repair
this configuration through risky peripheral vitrectomy. These
detachments are best repaired by careful scissors dissection at
the posterior pole where the NVD contraction occurred.
Peripherally originated vitreous traction, as in toxo-
cariasis or toxoplasmosis, can create peripheral retinal folds
that reach the optic nerve. These folds may not be readily
apparent through the vitreous opacification, but the surgeon
should infer their location based on the contours of the sur-
rounding retina to avoid iatrogenic retinal tears. Sometimes
the vitreous traction takes the configuration of broad pos-
terior attachment centered on the optic nerve with a cone
of posterior hyaloid creating anteroposterior traction. Cone
truncation, similar to diabetic vitrectomies, can release this
traction, but the authors always proceed to peel the posterior
vitreous from the retinal surface to decrease postoperative
epimacular membranes and macular striae from contraction
of the posterior hyaloids left over the macula.
Vitreom acular Interface Disorders
Epimacular membranes, vitreomacular traction syn-
drome, and macular holes are frequent in the setting of
severe uveitis. There is no significant difference in surgical
approach to these macular disorders compared to nonuveitis
eyes. The surgeon should only attempt whenever possible to
defer vitrectomy until clinical quiescence of the inflamma-
tion is observed.
Many uveitic patients can develop CME, and the vitre-
ous surgeon can be presented with the question of whether
to recommend vitrectomy surgery in the patient with CME.
Certain issues should be analyzed: (a) Is medical therapy
optimized to control the inflammation? The main cause for
CME in uveitis is diffuse breakdown of the blood-retinal bar-
rier due to inflammatory mediators. Other than the specific
setting of intermediate uveitis where, as mentioned previ-
ously, vitrectomy can improve the inflammatory reaction
and decrease CME, the best course of treatment for uveitic
CME is medical control of the inflammation. (b) Is there any
tractional component to the CME? Modern spectral domain
optical coherence tomography (OCT) can help answer this
question. The presence of macular traction on OCT, based
not only on the physical presence of a membrane but also on
the observation of sharp contour changes on the macula due
to traction, suggests that vitrectomy with membrane peeling
can help control CME.
Im p lan t at io n o f Drug De live ry Syst e m s
At the time of this writing, there are two classes of medications
that are available on implantable drug delivery formulations:
intravitreal steroids and intravitreal ganciclovir. The specific
indications for the use of these implants are beyond the
scope of this book.
LENS MANAGEMENT
IN SEVERE UVEITIS
There has been a push for increased use of immunosup-
pressants in severe uveitis to allow phacoemulsification with
intraocular lens (IOL) placement in the setting of cataracts.
While a patient with mild uveitis can undergo safe IOL
placement with topical steroids only, the authors believe that
patients with severe uveitis and cataracts are best served by
performing pars plana vitrectomy, lensectomy, and capsulec-
tomy. This avoids the possibility of lens-induced uveitis,
chronic CME, posterior synechiae and secondary glaucoma,
and cyclitic membrane formation. Immunosuppressants
should not be seen as a therapy for aphakia, which creates a
single-chamber eye that allows easier access of topical medi-
cations and prevents synechiae and chronic CME. Aphakic
spectacles or contact lenses have a much safer risk profile
than antimetabolites and high-dose steroids.
References
1. Davis JL. Diagnostic testing of vitrectomy specimens. Am
J Ophthalmol 2005;140(5):822–829.
2. Herrlinger U, Schabet M, Bitzer M, et al. Primary central
nervous system lymphoma: from clinical presentation to
diagnosis. J Neurooncol 1999;43(3):219–226.
3. White VA, Gascoyne RD, Paton KE. Use of the polymerase
chain reaction to detect b- and t-cell gene rearrangements in
vitreous specimens from patients with intraocular lymphoma.
Arch Ophthalmol 1999;117:761–765.
4. Knox CM, Chandler D, Short GA, Margolis TP. Polymerase
chain reaction-based assays of vitreous samples for the diagno-
sis of viral retinitis. Use in diagnostic dilemmas. Ophthalmology
1998;105(1):37–44.
5. Becker M, Davis J. Vitrectomy in the treatment of Uveitis. Am
J Ophthalmol 2005;140(6):1096–1105.
6. Wiechens B, Nölle B, Reichelt JA. Pars-plana vitrectomy in
cystoid macular edema associated with intermediate uveitis.
Graefe’s Arch Clin Experimen Ophthalmol 2001;474–481.
7. Han DP, Lewis H, Williams GA, et al. Laser photocoagula-
tion in the acute retinal necrosis syndrome. Arch Ophthalmol
1987;105:1051–1054.
8. Kalina PH, Pach JM, Buettner HM, Robertson DM. Neo-
vascularization of the disc in pars planitis. Retina 1990;10:
269–273.
C H AP TER
26
RETINOPATHY OF PREMATURITY
AND PEDIATRIC 25-GAUGE
VITRECTOMY

Retinopathy of prematurity (ROP) is a disease of screening services and (b) timely and appropriate laser
of premature, low–birth-weight infants that is treatment for active ROP disease. Vitrectomy surgery should
characterized by the presence of peripheral avas- be viewed as salvage therapy for selected cases.
cular retina, neovascularization at the junction
of the vascular to the avascular retina, and cicatricial traction
on the retina with subsequent traction retinal detachments. NORMAL RETINAL DEVELOPMENT
ROP has a distinct epidemiology worldwide that depends
on the availability of sophisticated neonatal intensive care Re t in al Vascular De ve lo p m e n t
units (NICUs), trained ophthalmologists for screening and
To understand the pathophysiology of ROP, we must start
treatment, adequate coordination of services, and obstet-
with a basic framework of normal retinal vascular develop-
rical care. In general, ROP is not a significant problem in
ment. Mesenchymal spindle cells grow from the optic nerve
countries where the lack of intensive neonatal services leads
toward the retinal periphery and differentiate centrifugally
to poor survival of very low–birth-weight infants, as in cer-
into retinal blood vessels. Vasculature development starts at
tain countries in sub-Saharan Africa. Moving up in the eco-
the optic nerve at gestational age 16 weeks. Retinal blood
nomic development, middeveloped countries, such as many
vessels reach the nasal ora serrata at week 36. Infants born
countries in Latin America and Eastern Europe, have a high
before full retinal vascular development have peripheral reti-
incidence of severe ROP, given that NICUs are available, but
nas devoid of fully developed blood vessels.
tight postnatal oxygen control, sufficiently trained ophthal-
mic consultants, and adequate coordination of services are
difficult to provide. Whereas ROP had decreased in inci- PATHOPHYSIOLOGY
dence in the United States as neonatal services had improved
OF RETINOPATHY OF PREMATURITY
and coordination with ophthalmologists trained in screen-
ing and treatment of ROP had been achieved, two issues are
Vascular En d o t h e lial Gro wt h Fact o r
increasing the incidence of the disease: (a) increased sur-
in No rm al Re t in al Vascular De ve lo p m e n t
vival of extremely low–birth-weight infants, who commonly
develop ROP, and (b) increased incidence of multiple preg- Vascular endothelial growth factor (VEGF) is one of the major
nancies (twins, triplet, quadruplet pregnancies), in large part molecules involved in retinal vascular development and is
due to aggressive fertility treatments and in vitro fertilization critical in the pathophysiology of ROP. Normal VEGF pro-
(1). A more recent problem is that ophthalmologists willing duction is required for vascular development. Both phases
and able to treat ROP are becoming less common due to of ROP have dysregulation of VEGF.
fears of liability and due to the practical concerns of coordi-
nating an ROP service.
Dire ct io n al Vasculo g e n e sis
The basic premise that this chapter will try to convey is
that, beyond the specific technical details of vitreous micro- An important aspect of retinal vascular development is its
surgery, the most cost effective and reliable way to decrease “directionality.” In other words, vessel development has to
ROP blindness is determined by (a) adequate coordination progress from the nerve toward the periphery. Although the
225
226 SECTION IV ■ Disease-Sp ecific Managem ent
biological details of vessel development are beyond the scope
of this book, some simple concepts should be understood.
The VEGF gene promoter has “hypoxia response elements”
that bind hypoxia-induced factors (2). This creates a feed-
back loop that increases VEGF production in hypoxic tissues
and decreases VEGF levels as tissue oxygen levels increase.
This relation between hypoxia and VEGF also plays a role in
diabetic retinopathy, sickle cell retinopathy, and other isch-
emic retinopathies. In the developing retina, areas absent of
blood vessels are relatively hypoxic in comparison to vascu-
larized retina. This creates a gradient of VEGF concentrations
that provide the “directional signal,” akin to chemotaxis, for
blood vessel development.
Ph ase 1 o f ROP: Arre st o f No rm al
Vascular De ve lo p m e n t
The association of ROP and O2 became clear after 10,000
babies had vision loss in the 1940s due to aggressive O2
therapy. In 1956, clinical trials demonstrated 3× risk of
ROP in babies with O2 supplementation (3). The incidence
of ROP decreased after O2 therapy was strongly discour-
aged in the late 1950s, but NICU mortality increased
concurrently.
The intrauterine environment has a PaO 2 of 22 to
24 mm Hg, whereas the human adult has an arterial PaO2 of
70 to 90 mm Hg. In essence, the intrauterine environment is
relatively hypoxic. In addition, many very low–birth-weight
premature babies have pulmonary disease that requires
hyperoxic therapy for survival. This hyperoxic environment
provides retinal oxygenation of the thin premature avascular
retina through the subjacent choriocapillaris and abolishes
the hypoxic-driven VEGF gradient required for normal retinal
vascular development.
Ph ase 2 o f ROP: Re t in al Ne o vascularizat io n
Weeks after birth, when growth of the eye and retina increase
oxygen demand and the choriocapillaris cannot supply
enough oxygen to keep up with this increased demand, the
peripheral avascular and ischemic retina becomes hypoxic.
While the initial phase of ROP depended on decreased
VEGF production, the second phase of ROP is caused by
pathologically high levels of VEGF secondary to peripheral
retinal hypoxia. As in other forms of proliferative retin-
opathies, such as diabetes, the increased concentration of
vitreous VEGF leads to retinal neovascularization. In ROP,
the localization of the neovascularization is at the border of
the vascular and avascular retina.
De ve lo p m e n t o f Tract io n
Re t in al De t ach m e n t
Cicatricial regression of the retinal neovascularization leads
to retinal traction and secondary traction retinal detachment.
An understanding of the vectors of traction is important to
surgical relief of traction and repair of the retinal detach-
ments. The origin of all traction vectors is the circumferen-
tial ring of retinal neovascularization at the ridge between
the vascular and the avascular retina. From this point, the
traction, as in the anterior loop traction (radial vitreous
collagen fibers) in proliferative vitreoretinopathy (PVR), is
directed toward the anterior vitreous base, the lens, and the
diametrically opposite neovascular ridge. Since these infants
do not have a posterior vitreous detachment, another vector
that can play a role in traction is directed from the ridge to
the posterior pole and optic nerve.
MANAGEMENT OF THE ROP INFANT
The CryoROP study demonstrated that ROP was amenable to
successful therapy through ablation of the peripheral avascu-
lar retina (4). It defined threshold disease as retinal neovas-
cularization (stage 3 ROP), in 5 contiguous or 8 total clock
hours and plus disease, and it determined that cryotherapy
for this level of disease decreased blindness approximately
25%. Since the CryoROP study, despite the absence of large
multicenter, randomized trials, laser therapy has largely
replaced cryotherapy. A significant caveat of the CryoROP
trial was that it was not directed toward determining differ-
ent clinical levels of disease progression and lumped zone 1
disease infants with zone 2 disease infants. It is now clear that
zone 1 disease, since by definition has a larger proportion of
avascular to vascularized retina and occurs in the most imma-
ture of babies, has a worse prognosis than zone 2 disease and
therefore requires earlier intervention. The Early Treatment
of Retinopathy of Prematurity Study (ETROP) trial demon-
strated that simpler and earlier guidelines for therapy, based
mostly on the presence of plus disease (retinal blood vessel
dilation and tortuosity at the posterior pole) and/or neovas-
cularization at zone 1, led to improved outcomes (5). Plus
disease in particular has become the main focus of clinical
determination of vascularly active ROP disease. Despite a
degree in subjectivity is accepted in the diagnosis of plus dis-
ease, the biological implication is clear. Since the junctional
neovascular ridge acts as an arteriovenous shunt, increased
shunting through the ridge increases retinal venous blood
flow and congestion. Venous congestion at the posterior pole
can then be used as a clinical surrogate for determining the
magnitude of shunting through the ridge and likelihood of
progression of the disease.
 CHAPTER 26 ■ Retinopathy of Prem aturity and Pediatric 25-Gauge Vitrectom y
The single important scenario where the clinician should
not wait for overt plus disease to recommend peripheral laser
retinal ablation is the infant with zone 1 disease. Zone 1
disease can progress through stages that are not the typical
clinical ROP stages: (a) line, (b) ridge, (c) neovascularization.
Particularly, the neovascularization, rather than growing in a
perpendicular plane to the retina, can grow along the retinal
surface and can be very difficult to diagnose. The clinician
should always follow the retinal vessels from the nerve to
the avascular retina and look for the normal dichotomous
branching. Nondichotomous branching, with vessels grow-
ing in a different direction other than centrifugally in the
posterior retina, often signifies retinal neovascularization.
The speed of progression of disease from zone 1 with flat
neovascularization to plus disease and later to traction retinal
detachment can be very fast, often taking 2 to 3 weeks. If the
clinician fails to diagnose the presence of nonelevated neo-
vascularization and waits for the development of plus dis-
ease in these infants with posterior disease, peripheral laser
ablation may not stop the progression of retinal detachment.
This is the previously called “Rush disease,” which implies a
retrospective diagnosis of fast progression to retinal detach-
ment. The concept that the clinician should now have is
aggressive posterior ROP (APROP), which is characterized
by severe vessel immaturity (zone 1 or posterior zone 2) with
flat neovascularization. In contrast to the retrospective “Rush
disease,” APROP has prognostic implication and mandates
immediate laser ablation to the avascular retina.
Lase r Tre at m e n t fo r Re t in o p at h y
o f Pre m at urit y
Indica tions
As mentioned previously, the indication for laser treatment
in ROP is zone 2 disease with plus disease (regardless of the
extent of neovascularization) and zone 1 neovascularization,
regardless of the presence of plus disease.
Anesthesia Considera tions
The authors prefer to do laser treatments with light, moni-
tored intravenous sedation, and analgesia with opiates
and without endotracheal intubation in the NICU when-
ever possible. Efficient laser techniques are important to
decrease the duration of treatment and likelihood of sys-
temic problems during laser therapy. Unfortunately, often
the babies require intubation due to the severity of their
cardiopulmonary disease and intraoperative oxygen desat-
uration and bradycardia. Still, in many babies, intubation
can be avoided if the laser therapy is started and tolerance
of the procedure is determined. The authors require all
babies to be either in the operating room with an experi-
enced pediatric anesthesiologist or at the NICU, with an
available ventilator at bedside and constant supervision by
the neonatology team.
227
Techniques
The authors perform their laser therapies with the near-
infrared 810-nm laser. The advantage of the near-infrared laser
is decreased absorption through blood vessels in the tunica
vasculosa lentis, which can potentially lead to cataract for-
mation. The goal of therapy is near confluent laser to all
the avascular retina. It is imperative to perform 360-degree
indirect ophthalmoscopy at the end of the laser treatment
to confirm that all the avascular retina is treated and no
gaps in the laser are present. The sclera depressor is used
both to manipulate the globe and rotate it to the required
direction and to depress and expose the far peripheral retina
for laser treatment. Scleral depression should be gentle to
decrease oculocardiac reflex and progressive media opaci-
fication during the laser. The authors rarely need to repeat
laser treatments after a single session of near confluent laser.
The exception is APROP, where the extent of the avascular
retina can be covered by the nonelevated neovasculariza-
tion. These infants require reexamination a week after the
laser therapy, since regression of the neovascularization can
expose untreated avascular retina that may require ablation.
Postoperatively, the authors treat all eyes with topical antibi-
otics and steroids (Tobradex) and mydriatics (homatropine
2%) for 2 weeks.
Man ag e m e n t o f St ag e 4a Re t in o p at h y
o f Pre m at urit y
The concept of lens-sparing vitrectomy for retinal detach-
ments in ROP that spare the macula (stage 4a ROP)
has been popularized in recent years, with the goal of
improved visual outcomes (6). The authors maintain a
sense of caution toward stage 4a vitrectomy. The authors
feel that many eyes that may have good visual poten-
tial without vitrectomy are being operated on and that
the risks of vitrectomy, namely, iatrogenic retinal tears
with rhegmatogenous retinal detachments and systemic
complications from general anesthesia in these delicate
infants—including death, should not be minimized. In
addition, many inexperienced surgeons are feeling pushed
to recommend surgery for 4a ROP.
Issues with Exuda tive Versus
Tra ctiona l Retina l Deta chment Dia gnosis
Whereas tractional retinal detachments are the typical and
most ominous type of detachment in ROP, not infrequently
will exudative retinal detachments be observed. These
detachments are convex in appearance and, in ROP, tend to
form in the peripheral retina, below areas of active plus dis-
ease. It is often very difficult to determine if a small amount
of subretinal fluid represents early tractional 4a detachment
or exudative detachment. Exudative detachments do not
require surgery, since they are self-limited by their nature as
the vascular activity of the disease subsides.
228 SECTION IV ■ Disease-Sp ecific Managem ent
Progression of 4a Retina l Deta chment
The natural history of stage 4a retinal detachments is not
clearly established in the literature. There are 4a detach-
ments that can be nonprogressive and do not require sur-
gery. It is likely that earlier gestational age at the time of
diagnosis of tractional 4a retinal detachment as well as active
vascular dilatation with plus disease at the time of diagnosis
may be associated with the higher likelihood of progression
to macular detachment.
The authors do not recommend early stage 4a vitrec-
tomy for all infants. They advocate close observation upon
the diagnosis of peripheral retinal detachment, and they only
perform vitrectomy in cases where the tractional detachment
is progressive and threatens the macula. As in all their pedi-
atric vitrectomies, they perform 25-gauge techniques.
Man ag e m e n t o f 4b Re t in o p at h y
o f Pre m at urit y
The most important indication for vitrectomy in ROP is reti-
nal detachment with macular involvement (stage 4b ROP).
As mentioned previously, the authors perform all their sur-
geries using 25-gauge techniques. The technical details will
be discussed later in this chapter.
Man ag e m e n t o f St ag e 5 Re t in o p at h y
o f Pre m at urit y
The most important problem with total retinal detachment
(stage 5 ROP) is the poor visual prognosis of these patients.
The senior author has published a large series (7) of stage 5
vitrectomy that has determined very poor visual outcomes
(only 4% of patients have final visual acuity 5/200 or bet-
ter). Although these data largely preceded the introduction
of peripheral retinal ablation and had a high proportion of
vascularly active eyes, they underscore the difficulties of suc-
cessful functional retinal reattachment. A number of problems
plague vitrectomy for total retinal detachments in ROP: (a)
high frequency of postoperative epiretinal reproliferation after
epiretinal membrane peeling, (b) diffuse retinal foreshortening
after infantile retinal detachment (maturation in the detached
configuration), (c) high likelihood of surgical failure after iat-
rogenic retinal breaks, and (d) amblyopia ex anopsia, since
retinal reattachment may not occur until after an interlude of
detachment during a critical cortical developmental period.
The issue of lens-sparing versus lens removal vitrec-
tomy depends more on the anatomic configuration of the
retinal detachment and the surgical space available between
the lens and the retina for successful removal of all traction.
In other words, if the space allows for safe entry of the sur-
gical instruments and accomplishment of all surgical goals
without removal of the lens, the authors certainly avoid per-
forming a lensectomy. Unfortunately, certain retinal detach-
ments simply are not amenable to surgical repair without
removal of the lens.
The authors currently do not recommend vitrectomy
for total closed-funnel retinal detachments with epiretinal
membranes. Recent open-funnel detachments in which
epiretinal membrane dissection is not required and careful
vitrectomy can relieve all transvitreal traction, on the other
hand, the authors do think should be operated on.
Re t in o p at h y o f Pre m at urit y Lase r Failure s
The authors’ anecdotal experience makes them believe that
true ROP laser failures, that is, after timely and adequate
laser therapy by an experienced surgeon, are uncommon.
Although severe immaturity with zone 1 disease (APROP)
can progress to retinal detachment despite well-performed
laser, many cases of laser failure are due to late treatment,
insufficient treatment, or both. The likelihood of “laser fail-
ure” is very likely low when laser is administered before
significant growth of neovascularization, since ROP traction
detachment progression is dependent on the contraction of
preexisting extraretinal proliferation, as in diabetic traction
retinal detachments. In the authors’ view, the main reasons
for laser failure are (a) lack of adequate follow-ups during the
screening stage, for example, inpatients lost during hospital
transfers and outpatients not brought by parents to clinic
visits as scheduled, and (b) misdiagnosis of APROP. This
leads the authors to the discussion of “The ROP Service.”
Th e Re t in o p at h y o f Pre m at urit y Se rvice
Since most failures of therapy are caused by nonpathological
causes, the best protection against ROP blindness is a well-
run ROP service.
Retinopa thy of Prema turity Service
An ROP service should have these characteristics:
1. Multiple physicians trained in screening and laser
therapy, to provide adequate cross-coverage and be
able to provide timely therapy regardless of an indi-
vidual physician’s schedule
2. At least one or two retina surgeons knowledgeable in
medical therapy and surgical therapy of ROP
3. Open communication with the neonatology services
of all NICUs served by the ROP service
4. Ancillary staff that keeps track of all babies in a ser-
vice. Again, overlapping and redundant tracking
lists decrease the likelihood of loss of follow-ups
5. Adequate coordination between outpatient clinics
and neonatology wards
6. Adequate attending physician supervision of all
these activities
Unfortunately, this level of workforce coordination is very
difficult to provide in many midsized or smaller cities in the
United States and almost impossible in many developing
countries. The authors run an ROP service that consists of
 CHAPTER 26 ■ Retinopathy of Prem aturity and Pediatric 25-Gauge Vitrectom y
seven NICUs spread in multiple hospitals in the metropolitan
Memphis region, one designated attending surgeon coordi-
nating all activities, two surgeons experienced in pediatric ret-
ina, screening services facilitated by two vitreoretinal surgery
fellows, one board-certified pediatric ophthalmologist with
ROP expertise, and two secretarial staff assisting in updating
the ROP rounding list and constantly communicating with
the NICUs and outpatient parents. The authors are proud
to say that despite the large size of their service, they only
need to perform one or two ROP vitrectomies a year from the
local NICU population that they follow. On the other hand,
the authors perform more vitrectomies on patients from the
hospitals outside their metropolitan region.
Te le m e d icin e in Re t in o p at h y o f Pre m at urit y
The development of the Retcam system for pediatric retinal
photography has kindled interest in the idea of telemedi-
cine in ROP precisely to increase the coverage area of ROP
services from metropolitan regions to hospitals outside of
these regions where skilled and willing ophthalmologists are
difficult to find. In the authors’ view, the most appropriate
application for ROP telemedicine is screening of neonates
in satellite locations (suburban or rural) too far from a cen-
tralized ROP service to provide a skilled examiner but close
enough for a timely transfer of a patient in need of laser ther-
apy. After all, the diagnosis of prethreshold ROP is useless if
an ophthalmologist skilled in ROP laser therapy is immedi-
ately unavailable. A secondary application of Retcam retinal
imaging is medicolegal protection by photographic docu-
mentation of ROP pathology. As the technology decreases in
size and technical difficulty and costs of the equipment drop,
the authors may see a stronger drive toward telemedicine in
ROP using skilled technicians and nurses in the NICU and
centralized imaging interpretation centers.
An t i–Vascular En d o t h e lial Gro wt h Fact o r
Th e rap y in Re t in o p at h y o f Pre m at urit y
Since the introduction of bevacizumab into the retina surgery
field, the question of anti-VEGF therapy in ROP has been
entertained. At the time of this writing, only anecdotal reports
are available that suggest that bevacizumab intravitreal injec-
tions in ROP are not associated with grossly obvious compli-
cations and may have adequate anatomic outcomes (8). While
the jury is still out, the main issue to ponder is: Is there a
need for anti-VEGF therapy given the good outcomes with
timely and adequate laser therapy? The authors believe that
there is a need. As mentioned previously, laser therapy, despite
being successful, requires a highly skilled surgeon and has
a very short window of opportunity to stop the progression
of ROP. If anti-VEGF therapy decreases the complexities of
therapy and expands the window of opportunity for interven-
tion, many more infants could be properly treated who are
not being treated properly today. In addition, the high costs
of lasers are barriers of entry in many parts of the world to
229
adequate ROP management, and these costs are significantly
lowered by using intravitreal injection therapy. Still, there are
many caveats of neonatal anti-VEGF therapy: (a) potential
systemic side effects and (b) potential for rapid progression to
traction detachment as in diabetic retinopathy.
Man ag e m e n t o f t h e Ch ild wit h
In ad e q uat e Pre o p e rat ive Lase r
As suggested previously, many infants who come to an ROP
referral practice have inadequate laser at the time of exami-
nation, often with coexisting retinal detachment. This is one
of the most difficult situations in a referral practice, since
there are no large studies that demonstrate the best course
of action. The authors’ approach, whenever possible, is to
perform laser therapy to any part of the avascular retina that
is still attached to attempt to decrease the vascular activity of
the eye and perform the vitrectomy 4 to 7 days following the
laser. This is another setting in which combination therapy
with anti-VEGF medications can provide a way to stop the
progression of the disease.
25-GAUGE PEDIATRIC VITRECTOMY
After the discussion of ROP, the authors will now focus on
the technical aspects of pediatric vitrectomy. Since the incor-
poration of 25-gauge technique as the authors’ nearly exclu-
sive system for adults, they have come to adopt 25-gauge
techniques exclusively in all pediatric vitrectomies, from
ROP to congenital cataracts and trauma. There are certain
technical considerations that need to be kept in mind when
using 25-gauge technique in children that are not relevant to
adults. On the other hand, the same benefits that 25-gauge
surgery offers adult patients are applicable to pediatric vit-
rectomies. Particularly, the lack of postoperative pain and
ocular surface discomfort is highly desirable in children.
In t ro d uct io n o f Tro car Can n ulas
The elasticity of the pediatric sclera makes the introduction
of the 25-gauge trocar cannulas significantly more difficult
than in adults. If any anterior segment surgery is planned in
conjunction with the vitrectomy surgery, the authors prefer
to preplace the inferotemporal infusion cannula prior to the
first anterior segment incision. Since vitreous infusion may
create “positive pressure” to the anterior segment surgeon, the
authors either place a plug on the cannula or introduce the
infusion system in the off position into the infusion cannula.
This avoids the introduction of the first 25-gauge cannula into
a soft eye, which may be very difficult.
Whereas in adults the authors feel it is very important
to introduce the cannulas in an oblique orientation, this
type of entry may be too difficult on the pediatric sclera. The
authors prefer to introduce all 25-gauge trocar cannulas near
perpendicularly to the sclera in pediatric cases. The inherent
230 SECTION IV ■ Disease-Sp ecific Managem ent
elasticity of the sclera actually closes the sclerotomies faster
in children than in adults, and the authors have never seen
choroidal effusions in pediatric vitrectomies during the post-
operative period.
A problem inherent to children is the lack of space in the
interpalpebral fissure for the external profile of the 25-gauge can-
nulas. This is usually managed by placing all three cannulas near
the horizontal meridian. In very small infants and those with
eyelid malformations, one may need to do a small canthotomy
(without cantholysis) in the temporal raphe. If done properly
with a single snip of the Wescott scissors, adequate space can
be achieved with minimal difficulties. The canthotomy can be
approximated at the end of the case with a single absorbable plain
gut suture and is rarely noticeable after 2 weeks from surgery.
The thin pediatric sclera does not provide enough grasp
of the infusion cannula to prevent torquing of the cannula
from the weight of the infusion line. To prevent the rotation
of the cannula from injuring the lens or the peripheral retina,
the surgeon should hold the cannula with bishop forceps
until the assistant secures the infusion line with Tegaderm
in its appropriate position. For the same reason, the surgeon
should take great care when rotating the eye, since the eyelid
can push on the infusion cannula and alter its orientation
with respect to the sclera, pushing it into the suprachoroidal
space, subretinal space, or the lens.
25-Gaug e Surg e ry Wit h o ut Can n ulas
If the surgeon is approaching the anterior vitreous in a lens-
sparing case (so that the contralateral entry cannula cannot
be used to vitrectomize the opposite anterior vitreous with-
out lens injury), the internal profile of the 25-gauge can-
nula system may preclude the ipsilateral approach with the
vitreous cutter to the anterior vitreous. Some surgeons have
recommended using 20-gauge techniques in these patients
to avoid the internal profile of the 25-gauge cannula. This
opinion does not take into consideration that the 25-gauge
cannulas are not required for proper functioning of the vit-
rectomy probe. The real need for the cannula is to perform
transconjunctival vitrectomy. On the other hand, after per-
forming a conjunctival incision, the 25-gauge cutter tip can
enter a sclerotomy without the cannula and be used for vit-
rectomy in a similar way as larger 20-gauge instruments. In
other words, if the cannula is in the way of the tip of the
cutter, then the surgeon can remove the cannula, perform
a small conjunctival incision to expose the sclerotomy, and
introduce the cutter and endoilluminator through the scle-
rotomy sites. Closure will only then require reapproximation
of the conjunctiva without sclera suturing.
Th re e -Po rt Ve rsus Two -Po rt Vit re ct o m y
Whenever the space allows for three-port vitrectomy, the
authors prefer to perform surgery in this way. If needed, the
surgery can be performed by two-port techniques: a cannula
for infusion and a superotemporal cannula for vitrectomy
instrument. The surgeon can hold a plano lens on his free
hand, and coaxial microscope illumination can be used to
see the vitreous anatomy, given the small size of the eye. The
lens has to be angled to direct the reflected light from the
microscope away from the surgeon’s view.
MANAGEMENT OF EPIRETINAL
MEMBRANES
In ROP as well as in rhegmatogenous detachments with
PVR, the postoperative reproliferation rate is very high, and
this limits anatomic postoperative outcomes. The surgeon
should use a minimalist approach to surgery and limit the
epiretinal dissection to the minimum required to relieve
retinal traction. The more a surgeon dissects, the higher the
stimulus for postoperative reproliferation.
SCLERAL BUCKLING
The authors do not advocate scleral buckling in ROP. Given
the pathophysiology of internal vitreous traction, the best
approach is to relieve the traction through vitrectomy tech-
niques when warranted.
On the other hand, scleral buckles are a good way to
treat uncomplicated pediatric retinal detachments with inf-
erotemporal dialyses or other anterior retinal pathology. The
low likelihood of cataract formation after buckling makes
this technique attractive for these cases. On the other hand,
if PVR is present, the authors proceed with vitrectomy with-
out scleral buckling.
References
1. Bergh T, Ericson A, Hillensjö T, et al. Deliveries and children
born after in-vitro fertilisation in Sweden 1982–95: a retrospec-
tive cohort study. Lancet 1999;354(9190):1579–1585.
2. Kimura H, Weisz A, Kurashima Y. Hypoxia response element of
the human vascular endothelial growth factor gene mediates tran-
scriptional regulation by nitric oxide: control of hypoxia-inducible
factor-1 activity by nitric oxide. Blood 2000;95(1):189–197.
3. Kinsey VE. Retrolental fibroplasia: cooperative study of ret-
rolental fibroplasia and the use of oxygen. Arch Ophthalmol
1956;56:481–543.
4. Group CF. Multicenter trial of cryotherapy for retinopathy of pre-
maturity: preliminary results. Arch Ophthalmol 1988;106:471–
479.
5. Group ET. Final results of the early treatment for retinopathy of
prematurity (etrop) randomized trial. Trans Am Ophthalmol Soc
2004;102:233–250.
6. Capone A. Lens-sparing vitreous surgery for tractional stage 4A
retinopathy of prematurity retinal detachments. Ophthalmology
2001;108(11):2068–2070.
7. Cusick MM, Charles MK, Agrón EM, et al. Anatomical and visual
results of vitreoretinal surgery for stage 5 retinopathy of prema-
turity. Retina 2006;26(7):729–735.
8. Kong L, Mintz-Hittner HA, Penland RL, et al. Intravitreous beva-
cizumab as anti–vascular endothelial growth factor therapy for
retinopathy of prematurity: a morphologic study. Arch Ophthal-
mol 2008;126(8):1161–1163.
C H AP TER
27
PEDIATRIC TRACTION
RETINAL DETACHMENTS

Diverse disease processes may create traction Surg ical Se q ue n ce an d Te ch n iq ue s

retinal detachments in the pediatric age group.
Microincisional vitrectomy with 25-gauge or possible
Special approaches are required to manage
23-gauge technology is ideal for PFV. The vitrectomy instru-
these difficult problems. These young patients
ment is introduced into the lens substance (Fig. 27.1),
have many years ahead of them and require the most aggres-
except in the rare instance that there is a clear lens. If the
sive attempts at restoration of vision. Patients in the pedi-
lens is clear, occasionally it is possible to introduce just the
atric age group having retinal detachments of several years’
scissors through the pars plana without infusion and to
duration can have remarkable visual recovery, and this
transect the membrane behind the lens, which can allow
emphasizes the need to proceed with such cases. By con-
the traction detachment to resolve. More commonly, the
trast, the very young patient, especially with retinopathy of
lens is cataractous or a large retrolental membrane is pres-
prematurity, is a high medical risk patient requiring careful
ent, requiring lens removal. The vitrectomy instrument is
assessment of the risk-benefit ratio of surgery. From the ocu-
used to remove the retrolental membrane centrally, being
lar standpoint, unilateral disease in the pre-6-year-old patient
careful not to amputate a ciliary process and thus cause
has an extremely high incidence of amblyopia, diminishing
bleeding. After a large opening is made, almost to the tips
the visual impact of successful surgery.
of the ciliary processes, the curved, 25-gauge scissors are
introduced and are used to cut between ciliary processes in
PERSISTENT HYPERPLASTIC order to segment the ring. If this is not done, the circum-
PRIMARY VITREOUS, PERSISTENT ferential traction will keep the ciliary processes detached in
the future.
FETAL VASCULATURE
As the anterior portion of the stalk extending poste-
riorly is resected, bleeding will frequently occur from the
Persistent hyperplastic primary vitreous is also known as per-
remnants of the hyaloid artery system. This should be
sistent fetal vasculature (PFV). It is usually a unilateral phe-
treated with bipolar diathermy. Although the stalk can be
nomenon accompanied by a smaller eye (1,2). The almost
resected midway in the vitreous cavity, it is usually necessary
uniform incidence of amblyopia means that these cases should
to delaminate the posterior termination of the stalk away
be operated on early (3). An additional argument for early sur-
from the retinal surface. Care should be taken to avoid the
gery is the prevention of long-term traction detachment and
retina, which can be pulled up into the central portion of the
pupillary block chamber (4,5). This condition is usually recog-
stalk around the optic nerve area. If there is a small “table-
nized early in life, and the patient should be operated upon if a
top” traction detachment surrounding the optic nerve, the
traction detachment is recognized, the cataract is sufficient to
25-gauge fine, curved scissors should be used to delaminate
cause visual loss and amblyopia, or there is shallowing of the
the membrane.
anterior chamber secondary to pupillary block. The embryo-
logical explanation for the syndrome is the lack of regression
of the primary vitreous and hyaloid vasculature, although the
Re sult s
primary cause is still unclear. Bilateral cases in males are usu-
ally associated with Norrie’s syndrome. Norrie’s cases should Anatomic results are excellent in this group, well over 95%.
not be operated on because the retina is dysplastic and the Bleeding and postoperative retinal detachment are exceed-
vitreoretinal interface cannot be delineated at surgery. ingly rare.

231
232 SECTION IV ■ Disease-Sp ecific Managem ent
Amblyopia is such a frequent accompaniment of this
disorder that aggressive treatment by early surgery, early
contact lens fitting, and patching of the other eye is neces-
sary. On occasion, a patient will achieve an excellent visual
result, but amblyopia is more common. If aggressive contact
lens fitting and patching are not done, amblyopia is the uni-
form result.
TOXOCARA CANIS
The parasite, Toxocara canis, can migrate from the choroidal
vessels through the choroid and retina into the vitreous cav-
ity (6). As it migrates through the eye, it creates an intense
inflammatory reaction, which may simulate endophthalmi-
tis or severe uveitis (7). After this initial phase, a fibrous
proliferation, more extensive than that seen with vascular
retinopathy, enters the eye. It may come from both the
posterior entrance site of the parasite and the pars plana
exit site. If this membrane, created by the inflammatory
reaction, does not cause traction detachment, it is not nec-
essary to operate. If there is a small traction detachment
away from the macula, this also can be followed up without
surgery.
These cases have a high incidence of late rhegmatog-
enous detachment secondary to long-standing traction.
Although scleral buckling may be able to reattach some of
these detachments, vitreous surgery with resection of the
stalk might be beneficial in certain cases. Occasionally, the
macula will become detached in the relatively early stages
Fig u re 2 7 .1 ■ The vitrectom y
instrum ent is used to rem ove the
lens and retrolental m em brane.
of this disease in a nonrhegmatogenous fashion. When this
occurs, vitrectomy is necessary.
Surg ical Se q ue n ce an d Te ch n iq ue s
Standard 25-gauge cannula placement is performed. The vit-
rectomy instrument is introduced superotemporally and the
endoilluminator superonasally. The vitreous is usually coni-
cally shaped and the posterior vitreous cortex (PVC) is con-
tinuous with the tract (Fig. 27.2). Although a casual examiner
would think that this tract progresses through the vitreous
cavity, in fact, posterior vitreous detachment usually occurs
with the proliferation on the vitreous surface. If resection of
the dense tract alone is performed, the clear vitreous may cause
the traction detachment to remain. For this reason, the vitrec-
tomy is completed first and then attention is directed to the
stalk. Care should be taken to avoid a steep radial fold of retina
underlying the stalk, which is present in many cases. Scissors
segmentation and delamination techniques should be utilized
to allow the release of retinal traction (Figs. 27.3–27.5). The
principles are similar to those used for traction detachments
from diabetic retinopathy, although the proliferation is much
more dense and localized. Bleeding may occur from this stalk
and can be treated with bipolar diathermy, paying close atten-
tion to avoid the retina and optic nerve.
Scleral buckling is not utilized in these young patients
because of the possibility of later intrusion of the encircling
element. If a rhegmatogenous component is present, it must
be managed by internal drainage of subretinal fluid, fluid-
gas exchange, and laser endophotocoagulation.
 CHAPTER 27 ■ Pediatric Traction Retinal Detachm ents 233

Fig ure 2 7 .2 ■ The vitreous is usually conically shaped and the PVC is continuous with the tract.

Fig ure 2 7 .3 ■ Multip le radial cuts are m ade in an inside-out direction.

234 SECTION IV ■ Disease-Sp ecific Managem ent

Fig ure 2 7 .4 ■ Scissors are used to delam inate the resulting segm ents.

Fig ure 2 7 .5 ■ Curved scissors are used to rem ove the d elam inated seg m ents.
 CHAPTER 27
Re sult s
The visual outlook primarily depends upon the involvement
of the macula in the full-thickness chorioretinal destructive
process. If the macula is secondarily detached but has not
been invaded by the parasite, excellent visual recovery is
possible. Fortunately, recurrence of postoperative inflamma-
tion is not a problem in these cases.
References
1. Yanoff M, Fine BS. Ocular pathology-a text and atlas. Hagerstown,
MD: Harper & Row, 1975:698.
■ Pediatric Traction Retinal Detachm ents 235
2. Goldberg MF. Persistent fetal vasculature (PFV): an integrated
interpretation of signs and symptoms associated with persistent
hyperplastic primary vitreous (PHPV). LIV Edward Jackson
Memorial Lecture. Am J Ophthalmol 1997;124(5):587–626.
3. Hunt A, Rowe N, Lam A, et al. Outcomes in persistent hyper-
plastic primary vitreous. Br J Ophthalmol 2005; 89(7):859–863.
4. Gass JDM. Surgical excision of persistent hyperplastic primary
vitreous. Arch Ophthalmol 1970;83:163.
5. Smith RE, Maumenee AE. Persistent hyperplastic primary vitre-
ous. Trans Am Acad Ophthalmol Otolaryngol 1974;78:911.
6. Hogan MJ, Kimura SJ, Spencer WH. Visceral larval migrans and
peripheral retinitis. JAMA 1965;194:1345.
7. Stewart JM, Cubillan LD, Cunningham ET Jr. Prevalence, clini-
cal features, and causes of vision loss among patients with ocu-
lar toxocariasis. Retina 2005;25(8):1005–1013.
C H AP TER
28
MANAGEMENT OF VITREORETINAL
COMPLICATIONS ASSOCIATED
WITH KERATOPROSTHESIS
The management of severe anterior segment
disease and intractable corneal opacities has
evolved rapidly in the recent years. Whereas in
the late 1990s there was a push for limbal stem
cell transplantation in conjunction with aggressive systemic
immunosuppression, the complications and practical diffi-
culties associated with immune therapy have decreased the
interest in this technique. Keratoprosthetic devices have been
in development for decades. Claes Dohlman, at the Massa-
chusetts Eye and Ear Infirmary, has developed and perfected
the Boston keratoprosthetic device. His relentless pursuit of
technical improvements of the device has allowed the recent
popularization of the Boston keratoprosthesis and its more
common utilization in a variety of disorders of the cornea
and anterior segment (1).
One may classify vitreoretinal disorders associated with
keratoprosthesis implantation into
1. Preexisting vitreoretinal conditions
Keratoprosthesis can be associated with various vitreoretinal
disorders that may require vitreous microsurgery (2). The eyes
that require keratoprosthetic implantations have commonly
been operated on multiple times unsuccessfully and have
sustained long-standing inflammation. These eyes may har-
bor preexisting retinal conditions that may be unmasked by
the clear optics of the keratoprosthetic device, such as dense
epiretinal membranes, retinal detachment, or macular holes,
that may require surgery with the keratoprosthesis in situ.
2. Disorders secondary to the presence of the keratopros-
thetic device
Keratoprosthesis can commonly present with retroprosthetic
membranes that severely limit visual outcomes. These mem-
branes are usually more frequent and severe in children but
can also present in adults. Retroprosthetic membranes are
usually too thick and adherent to be amenable to YAG laser
236
and frequently require vitrectomy techniques for removal of
the retroprosthetic membrane.
In addition, since the Boston keratoprosthesis never
integrates with the host cornea or sclera, there will always be
a permanent open wound around the keratoprosthesis that
may be the route for bacteria to develop endophthalmitis.
The Alphacor keratoprosthesis and the osteo-odontogenic
keratoprosthesis designs are attempts to increase integra-
tion of the device. Dohlman has noticed that placement of
a contact lens over the keratoprosthesis and chronic topical
antibiotics decreases endophthalmitis rates.
Glaucoma is very common after keratoprosthesis, and
many surgeons recommend primary placement of Ahmed
valve glaucoma implants (3). In eyes that have undergone
iridectomy, lensectomy, and keratoprosthesis that also have
an Ahmed valve, vitreous occlusion of the tip of the Ahmed
valve can require emergent vitrectomy for relief of the acute
glaucoma attack.
3. Disorders associated with the surgical technique of ker-
atoprosthesis implantation
Keratoprosthesis implantation is a difficult procedure that
may require, beyond replacement of the cornea, total iridec-
tomy, lensectomy, or intraocular lens explantation and open-
sky vitrectomy. The iridectomy may cause intraoperative
and postoperative vitreous hemorrhage, and the vitrectomy
and lens removal may cause inadvertent retinal tears with
subsequent retinal detachment. Early postoperative vitreous
hemorrhages are often difficult to manage, since although
the most likely cause for the hemorrhage is iridectomy in
a hypotonic eye, one cannot usually rule out intraoperative
retinal tear until surgery is undertaken. Another surgical
complication of keratoprosthesis implantation is intraopera-
tive suprachoroidal hemorrhage. Since these eyes undergo
open-sky surgery, choroidal expulsive hemorrhages are a real
risk. Unfortunately, it is almost impossible to repair these
 CHAPTER 28 ■ Managem ent of Vitreoretinal Com plications Associated with Keratoprosthesis
eyes even if the surgeon is able to suture the keratoprosthesis
quickly enough to prevent expulsion of the intraocular con-
tents, since the retina and anterior vitreous can become
adherent to the corneal wound and the back plate of the
keratoprosthetic device, making successful retinal detach-
ment repair almost impossible.
TECHNICAL CONSIDERATIONS
REGARDING OFFICE EXAMINATIONS
IN KERATOPROSTHESIS PATIENTS
Office examinations of patients with keratoprosthesis can be
challenging. While examination of the posterior pole with
an indirect ophthalmoscope can usually be accomplished,
examination of the retinal periphery may be more difficult.
The authors prefer the use of high plus wide-angle lenses
on the slit lamp for examination of the retinal periphery on
patients with keratoprosthesis, since the oblique slit beam
reduces reflections and glare from the surface of the ker-
atoprosthesis. If B-scan ultrasonography is needed due to
vitreous hemorrhage, the examiner may need to place the
transducer directly adjacent to the keratoprosthesis and
perform transscleral imaging, avoiding artifacts from the
keratoprosthesis.
VITRECTOMY IN
KERATOPROSTHESIS PATIENTS
The most significant technical considerations while doing
vitrectomies in eyes with keratoprosthesis are (a) introduc-
tion of 25-gauge trocar-cannula systems, (b) peripheral reti-
nal visualization and management, and (c) management of
retroprosthetic membranes.
In t ro d uct io n o f t h e 25-Gaug e Tro car Can n ulas
Eyes that require keratoprosthesis may not have clearly iden-
tifiable limbal structures from which to measure posteriorly
the location of the pars plana. Entry into the eye should obvi-
ously avoid the anterior retina but is constrained anteriorly
by the presence of the back plate of the keratoprosthesis.
The 25-gauge cannula needs to be at least 2 mm posterior to
the edge of the back plate to allow clearance below the back
plate, required to approach the posterior side of the optic in
case of retroprosthetic membrane removal. The only identifi-
able structure on the anterior segment of keratoprostheses is
the edge of the front plate. The back plate measures 8.5 mm
in diameter, but 7-mm diameter back plates are also avail-
able for small pediatric eyes. It is important to remember
that the edge of the back plate is not usually visible, so the
surgeon should estimate its location from the edge of the
anterior optic. Since the radius of the front plate is 2.5 mm,
the edge of the back plate is 1.75 mm posterior to the edge
of the optic. If 2 mm clearance posterior to the back plate
237
is the target, then cannula entry should be performed 3.75
to 4 mm posterior to the edge of the optic (3 mm in case of
smaller pediatric back plates). This entry location likely cor-
responds to pars plicata rather than pars plana. The authors
do not recommend more posterior entry since the surgeon
has no assurance of perfect centration of the keratoprosthetic
device. If the device is decentered, posterior entry through
the anterior retina can be unfortunately performed. If the
surgeon can visualize the edge of the back plate, then direct
measurement from this location overrides the previous dis-
cussion. The authors do not have experience with translid
keratoprostheses but would recommend a similar analysis
for sclerotomy placement if confronted with the situation.
The authors find the 25-gauge system excellent of surgery
in keratoprosthesis cases, since it avoids conjunctival dissec-
tion on these eyes that have preexisting ocular surface scar-
ring. Transconjunctival sclerotomy closure with a single 8-0
Biosorb suture may be needed at the end of the surgery since
the conjunctiva rarely covers the sclerotomies.
Pe rip h e ral Re t in al Visualizat io n
an d Man ag e m e n t
The configuration of the optic is a cylinder of 3 mm diam-
eter, with a thickness of slightly greater than 3 mm. Direct
visualization of the posterior pole structures is usually excel-
lent with the eye on primary position. Rotation of the globe,
though, creates tilting of the optical cylinder and decreases
the optical aperture for visualization. In essence, permanent
keratoprosthetic vitrectomies require surgery in the primary
position. Given this constraint, contact-based wide-angle lens
(Volk) visualization is imperative for peripheral vitrectomy.
Even in these conditions, visualization of the far periphery
may be very difficult, and this is one of the reasons that rheg-
matogenous detachments are difficult to repair. Endoscopic
techniques may be required for complete retinal peripheral
evaluation and laser.
If peripheral visualization impairs adequate manage-
ment of a retinal detachment, a useful technique to consider
is medium-term perfluorooctane with 360-degree peripheral
retinal endophotocoagulation. Perfluoro-octane (PFO) can
be injected over the optic nerve and used to fill the vitreous
cavity and reattach the retina as described in Chapter 14. Pos-
terior retinotomy should be avoided in these eyes. Once the
retina is attached, the surgeon can proceed with laser of the
retinal periphery circumferentially with at least three to four
rows of confluent laser as far as visualization allows. The PFO
can then be left in the vitreous cavity for 2 weeks and can be
later removed. If permanent posterior pole reattachment is
accomplished, despite peripheral subretinal fluid, this quali-
fies as a successful repair in an eye that would otherwise have
proceeded to blindness. The authors discourage the use of sil-
icone oil in keratoprosthesis retinal detachments. These eyes
are already prone to glaucoma, intraocular pressure measure-
ments are impossible to perform reliably, and the silicone oil
can prevent fluid egress through the Ahmed valve.
238 SECTION IV ■ Disease-Sp ecific Managem ent
Man ag e m e n t o f Re t ro p ro st h e t ic Me m b ran e s
Retroprosthetic membranes are often very thick and adherent
to the posterior surface of the optic. Forceps, scissors, and
the vitreous cutter are usually unable to start an opening
on the membrane. The authors use a 25-gauge needle or
25-gauge microvitreoretinal (MVR) blade with the tip bent
to form a sharp pick that is introduced through the 25-gauge
cannula. The sharp bent needle tip is then used to engage
the membrane away from the center of the optic (in case the
needle scratches the posterior surface of the device) and an
initial opening on the membrane is created. Once an edge
is found, further removal with forceps, scissors (segmenta-
tion), or the vitreous cutter can be performed (Fig. 28.1).
The authors attempt to remove the membrane beyond the
edge of the optic along the surface of the back plate, with
the goal of decreasing recurrence of the membrane over the
optical surface of the device.
The authors are currently working with their cornea
specialist colleagues to refine the initial implantation of the
.
Fig ure 2 8 .1 ■ After an opening
is m ade with an MVR blad e, the
retroprosthetic m em brane can be
rem oved with the vitreous cutter.
keratoprosthetic device. The authors’ goal is to eliminate
postoperative retinal detachments and vitreous hemorrhages
and intraoperative suprachoroidal hemorrhages. Their most
recent approach is placement of 25-gauge cannulas and
anterior vitrectomy (limited by visualization) prior to the
removal of the corneal button. Once the device is in place,
pars plana infusion can be reinitiated to pressurize the globe.
The authors believe that this technique decreases iatrogenic
retinal tears and decreases the time of globe depressurization
and hypotony that can lead to intraoperative hemorrhages.
References
1. Dohlman CH, Harissi-Dagher M, Khan BF, et al. Introduction
to the use of the Boston keratoprosthesis. Exp Rev Ophthalmol
2006;1(1):41–48.
2. Subhransu Ray, Khan BF, Dohlman CH, et al. Management of
vitreoretinal complications in eyes with permanent keratopros-
thesis. Arch Ophthalmol 2002;120:559–566.
3. Netland PA, Terada H, Dohlman CH. Glaucoma associated with
keratoprosthesis. Ophthalmology 1998;105(4):751–757.
C H AP TER
29
INADVERTENT PENETRATION
OF THE EYE
Inadvertent penetration of the globe can occur in
association with many ocular procedures. Most,
if not all, of these incidents can be prevented
with strict attention to technique. The keys to
prevention are awareness of the factors likely to cause pen-
etration and constant vigilance. Many factors are associated
with inadvertent penetration of the eye. They include patient
movement, posterior staphylomas, myopic eyes, and poor
technique (1–3).
OFFICE INJECTIONS
Subconjunctival injection in the lower fornix is safer than
sub-Tenons (infra-Tenons) or so-called periocular injection.
Subconjunctival injection under the bulbar conjunctiva is
potentially dangerous. Myopic eyes are more likely to be
penetrated than emmetropic or hyperopic eyes because of
greater ocular length, thin sclera, and staphylomas. Injec-
tions should be performed with patient supine. Viscous
lidocaine provides better anesthesia than topical anesthe-
sia. Both physicians’ hands should be braced against the
patient’s facial bones with the needle entering laterally and
nearly parallel to the lid margins. If the needle is directed
posteriorly, the globe can be penetrated if the patient moves
forward.
PRESURGICAL ANESTHESIA
Topical and intraocular anesthesia is rapidly increasing in
popularity for cataract surgery. Although some straightfor-
ward core vitrectomy procedures can be performed with
topical anesthesia, the extraocular muscles must be blocked
for macular and complex surgery. Peribulbar anesthesia
using a blunt, curved, flexible cannula and small conjunc-
tival incision are theoretically safer than multiple injec-
tions into the anterior orbital tissue. The patient should be
directed to position the eye in the primary position during
retrobulbar injection. This position has been shown by
Grizzard to minimize displacement and increased tension
on the optic nerve, which increase the risk of penetration of
the nerve (4).
Oversedation is a significant cause of patient motion
and inadvertent penetration of the eye during the adminis-
tration of anesthesia. Explanation is frequently better than
sedation, especially for elderly patients. There is no evidence
that blunt, so-called retrobulbar needles are safer than the
standard, sharp, 27-gauge, 1.25-inch needles used by the
authors. There is cadaver CT evidence that 1.5-inch needles
are too long and increase the likelihood of optic nerve dam-
age at the apex of the orbit. Sharp needles require less force
than blunt needles and are therefore less likely to advance
abruptly through the lid septum or cause pain and patient
movement. The entry point should be at the outer “corner”
of the orbit, not the more traditional junction of the inner
2/3 and outer 1/3 of the lower lid.
TOXICITY
Intraocular gentamicin and other aminoglycoside antibiotics
are very toxic to the retina (5–7). Lidocaine is apparently
relatively safe but Wydase has been shown to be very toxic
to the retina (8). Many randomized trials have not demon-
strated any beneficial effect of Wydase (9,10). The authors
never use Wydase and suspect that many surgeons are
unaware of the lack of efficacy and potential hazards and use
this agent on a “routine” basis.
RECOGNITION
Corneal edema occurs instantaneously when the intraocular
pressure is elevated to very high levels. Some surgeons have
reported that the patients experience marked pain, anxiety,
and/or nausea if intraocular injection or penetration of the
eye occurs. Hypotony is a variable and somewhat delayed
239
240 SECTION IV ■ Disease-Sp ecific Managem ent
finding in these cases. Hemorrhage into the orbital tissue,
eye wall, or eye can occur if a scleral laceration or double
penetrating injury occurs.
IMMEDIATE ACTION
The fundus should be immediately inspected with the indi-
rect ophthalmoscope by the operating surgeon if there is
any suspicion of inadvertent penetration. If a penetration
is recognized, elective anterior segment surgery should be
cancelled in most instances and a vitreoretinal surgeon
contacted immediately. An immediate view of the fundus
can determine if there has been a single or double pen-
etration and if the macula, posterior pole, optic nerve, or
retinal vessels have been damaged. Although some sur-
geons recommend immediate vitrectomy if an aminogly-
coside antibiotic is injected into the eye, the diffusion rate
is very rapid and it is highly likely that retinal damage will
occur before a vitrectomy can be accomplished. The laser
indirect ophthalmoscope (LIO) is ideal to treat the pen-
etration site(s) before diffusion of vitreous hemorrhage
can occur. Cryopexy is less desirable than LIO because it
requires pressure on the eye and a conjunctival incision
and probably causes more reparative scarring. Exploration
of the sclera and suturing of the scleral penetration sites
are not indicated in these cases in the authors’ opinion.
DELAYED VITRECTOMY
Although some surgeons recommend immediate vitrectomy,
the authors believe that it is better to delay vitrectomy and
find vitrectomy unnecessary in most cases. Indications for
vitrectomy include the development of a tract, retinal detach-
ment, or formation of an epimacular membrane. Vitreous
hemorrhage should be observed frequently for the develop-
ment of hypocellular collagen contraction as evidenced by
decreased vitreous mobility. Ultrasonic imaging can deter-
mine the presence of hypocellular collagen contraction via
decreased vitreous mobility during saccades. B-scan can also
determine if retinal and/or choroidal detachment is present.
Scleral buckling is virtually never indicated.
SURGICAL INDICATIONS
If vitreous hemorrhage persists over 10 days or if hypocellular
vitreous collagen contraction as evidenced by decreased vitre-
ous mobility, or a fibrous tract, or retinal detachment occurs,
vitrectomy is indicated. If the eye is No Light Perception
Vision (NLP) from optic nerve damage, surgery is not indi-
cated. Surgery is not indicated for removal of subretinal blood
or intravitreal anesthetic agents in the authors’ opinion.
Surg e ry Ap p ro ach
Standard vitrectomy methods for trauma are used as previ-
ously described. Retinopexy around the penetration sites is
only indicated if a retinal break is observed. Exploration of
the site, cryopexy, or scleral buckling is not required.
References
1. Edge R, Navon S. Scleral perforation during retrobulbar and
peribulbar anesthesia: risk factors and outcome in 50,000
consecutive injections. J Cataract Refract Surg 1999;25(9):
1237–1244.
2. Ginsburg RN, Duker JS. Globe perforation associated with ret-
robulbar and peribulbar anesthesia. Semin Ophthalmol 1993;
8(2):87–95.
3. Modarres M, Parvaresh MM, Hashemi M, Peyman GA. Inad-
vertent globe perforation during retrobulbar injection in high
myopes. Int Ophthalmol 1997–1998;21(4):179–185.
4. Grizzard WS, Kirk NM, Pavan PR, et al. Perforating ocular
injuries caused by anesthesia personnel. Ophthalmol 1991;
98(12):1757.
5. Campochiaro PA, Lim JI. Aminoglycoside toxicity in the treat-
ment of endophthalmitis. The aminoglycoside toxicity study
group. Arch Ophthalmol 1994;112(1):48–53.
6. Peyman GA. Aminoglycoside toxicity. Arch Ophthalmol 1992;
110(4):446.
7. Campochiaro PA, Conway BP. Aminoglycoside toxicity- a sur-
vey of retinal specialists. Implications for intraocular use. Arch
Ophthalmol 1991;109(7):946–950.
8. Gottlieb JL, Antoszyk AN, Hatchell DL, Saloupis P. The safety
of intravitreal hyaluronidase. A clinical and histologic study.
Invest Ophthalmol Vis Sci 1990;31(11) 2345–2352.
9. Bowman RJ, Newman DK, Richardson EC. Is hyaluronidase
helpful for peribulbar anaesthesia? Eye 1997;11:385–388.
10. Crawford M, Kerr WJ. The effect of hyaluronidase on peribulbar
block. Anaesthesia 1994;49:907–908.
C H AP TER
30
MANAGEMENT OF
SUPRACHOROIDAL HEMORRHAGE
Suprachoroidal hemorrhage is a devastating
complication of intraocular surgery. This com-
plication is very difficult to anticipate, prevent,
and manage (1–5) (Fig. 30.1). Intraoperative
management is complex, as is the decision to intervene in
the postoperative period. These cases are also called chor-
oidal hemorrhages because blood dissects into the spongy
choroidal tissue. They are termed expulsive hemorrhages if
the choroid and retina are forced out of the eye by high pres-
sure in the suprachoroidal space.
INCIDENCE OF SUPRACHOROIDAL
HEMORRHAGE
The incidence of suprachoroidal hemorrhage in extracapsu-
lar cataract surgery and phacoemulsification is approximately
0.15%, filtering procedures 0.15%, penetrating keratoplasty
0.56%, vitrectomy 0.41%, and the principal author’s vit-
rectomy series is 0.01% (5/26,000). Small-incision cata-
ract surgery does not necessarily reduce the incidence of
this complication as the intraocular pressure (IOP) must
be reduced to atmospheric pressure during intraocular lens
(IOL) insertion, although it certainly makes wound closure
faster and safer. Clearly, the duration of low IOP is less with
phacoemulsification than with intracapsular surgery. Small-
incision surgery with self-sealing wounds construction facil-
itates rapid wound closure and normalization of the IOP.
Filtering procedures such as trabeculectomy, setons, and
valves remain a common cause of this complication. Late
hemorrhages are common if antimetabolites (mitomycin) are
used producing a sustained, very low IOP.
PATHOGENESIS AND PREVENTION
A key factor in the pathogenesis of suprachoroidal hemor-
rhages is a high trans–arterial wall pressure gradient due to
acute lowering of the IOP to atmospheric pressure in the
presence of hypertension. The bleeding is probably due to
shearing of vessels traversing the suprachoroidal space
caused by scleral stretch. Penetration of the eye is probably
frequent and often unrecognized because of suprachoroidal
hemorrhage. Hypertension and arterial disease are critical
factors in the pathogenesis. Patients should be normotensive
during cataract, penetrating keratoplasty, secondary IOL,
and filtering procedures. If general anesthesia is being used,
it is probably advisable to ask the anesthesiologist to use
neuromuscular blockade for open eye procedures to prevent
high blood pressure secondary to “bucking on the tube.”
COMPLICATIONS OF
SUPRACHOROIDAL HEMORRHAGE
Bad outcomes in nonexpulsive cases are usually not directly
due to the hemorrhage but are secondary to retinal detach-
ment from hypocellular collagen contraction and adherence
of vitreous to anterior structures (iris, wound, capsule). Vit-
reoretinal traction increases in the weeks and months after
surgery after the choroidal blood absorbs and the buckle-
like effect disappears (Fig. 30.2). Many patients suffer optic
nerve damage secondary to the acute increase in IOP or
possibly high intrasheath pressures.
ACUTE MANAGEMENT
It is not advisable to open the anterior chamber to implant,
reposition, or replace the IOL if a suprachoroidal hemorrhage
occurs. It is better not to cut down on the sclera if a hemorrhage
occurs, but instead, the focus should be on closing the wound
with 8-0 nylon sutures as rapidly as possible. A viscoelastic
can occasionally be used to reposition the iris. The surgeon
should close the wound and delay surgery until inflammation
is gone if a suprachoroidal hemorrhage is recognized.
241
242 SECTION IV ■ Disease-Sp ecific Managem ent
INDICATIONS FOR VITRECTOMY
Rhegmatogenous or traction retinal detachment secondary
to hypocellular collagen contraction and adherence of vitre-
ous to anterior structures is the most common indication
for vitrectomy. A flat chamber is also an indication for surgi-
cal intervention. An arbitrary duration of the suprachoroidal
Fig u re 3 0 .1 ■ Sup rach oroid al
h em o rrh ag e p u sh es th e retin a
centrally and anteriorly and com -
presses rem aining vitreous.
hemorrhage is often cited as an indication for vitrectomy.
The authors disagree with this indication and use the indi-
cations described previously. Appositional (kissing) supra-
choroidal hemorrhage is not an indication for surgery in the
authors’ opinion. There is no scientific evidence for retinal
detachment due to retinal-retinal adherence. There is usually
a layer of vitreous interposed between the retinal surfaces.
Fig u re 3 0 .2 ■ Retin al d etach -
m ent m ay be caused by hypocel-
lular con traction of the vitreous
com bined with adherence to ante-
rior structures.
 CHAPTER 30
Fig ure 3 0 .3 ■ Scleral cut-downs
are not required because the blood
d rain s out th roug h p ars p lan a
sclerectom ies during vitrectom y.
SURGICAL MANAGEMENT
OF SUPRACHOROIDAL HEMORRHAGE
The infusion should be initiated with a 25-gauge needle
or the long, handheld, angulated, blunt cannula (May) or
the end-irrigating endoilluminator if a 20-gauge case. The
suprachoroidal hemorrhage usually drains through each
pars plana sclerotomy during the early part of the vitrectomy
(Fig. 30.3). Scleral cut-downs are not required to drain the
blood and may cause severe acute bleeding. Manipulation of
the sclera with a scleral depressor breaks up clots, allowing
faster egress of blood through the standard pars plana sclero-
tomies. Removal of all the blood is not necessary as the blood
acts as a scleral buckle and will slowly absorb in the postop-
erative course.
MANAGEMENT OF ASSOCIATED
RETINAL DETACHMENT
Silicone is usually preferred to gas in conjunction with vit-
rectomy. The purpose of silicone surface tension manage-
ment is rhegmatogenous confinement for unrecognized
■ Managem ent of Suprachoroidal Hem orrhage 243
retinal breaks and subsequent retinal breaks secondary to
collagen contraction and epiretinal membrane formation.
SUMMARY
As in many medical problems, prevention, followed by early
recognition and conservative management, is the key to
reducing bad outcomes from suprachoroidal hemorrhage.
References
1. Chu TG, Green RL. Suprachoroidal hemorrhage. Surv Ophthal-
mol 1999;43(6):471–486.
2. Glazer LC, Williams GA. Management of expulsive choroidal
hemorrhage. Semin Ophthalmol 1993;8(2):109–113.
3. Beatty S, Lotery A, Kent D, et al. Acute intraoperative suprachor-
oidal haemorrhage in ocular surgery. Eye 1998;12:815–820.
4. Tabandeh H, Sullivan PM, Smahliuk P, et al. Suprachoroidal
hemorrhage during pars plana vitrectomy. Risk factors and out-
comes. Ophthalmology 1999;106(2):236–242.
5. Wirostko WJ, Han DP, Mieler WF, Pulido JS. Suprachoroidal
hemorrhage: outcome of surgical management according to
hemorrhage severity. Ophthalmology 1998;105(12):2271–2275.
C H AP TER
31
COMPLICATIONS OF
VITREORETINAL SURGERY
Most complications of vitreous surgery, after an
initial surgical learning phase, result from bio-
logic problems associated with specific disease
states. Implicit in the discussion of management
of complications is their recognition. The importance of
frequent follow-up of the complex vitreous surgery patient
cannot be overemphasized. In some cases, the primary oph-
thalmologist, less familiar with vitreous surgery, will not ini-
tiate appropriate action if complications occur. It is therefore
quite important for the surgeon to follow up the patient with
a general ophthalmologist. Every attempt should be made to
educate other ophthalmologists in the recognition and man-
agement of these complications, but the primary responsi-
bility rests with the surgeon.
CORNEAL COMPLICATIONS
Poor epithelial adherence can persist for weeks if the epithe-
lium is removed at the time of vitrectomy (1–4). Care dur-
ing the prep, frequent irrigation of the cornea with balanced
salt solution (BSS), and shorter operating times dramatically
reduce the need to remove the epithelium. The authors
remove the epithelium in less than 2% of cases. When neces-
sary, the epithelium should be mobilized rather than scraped.
The rounded blade used for mobilization should never touch
Bowman’s membrane. The defect should be made the same
size as the pupil, avoiding the peripheral cornea. Some sur-
geons use Gelfoam, tear substitutes, or viscoelastics during
surgery, but the authors have not found these to be necessary
or efficacious. Bandage contact lenses or pressure patches
are not necessary in the typical postoperative management.
In the rare cases requiring epithelial removal, the epithelium
is healed within several days postvitrectomy and is always
healed on the 2-week office visit. If the patient develops a
recurrent epithelial defect, cessation of topical medications
and taping the lid closed for 12 to 24 hours will usually
be sufficient. A rare patient will require a therapeutic soft
244
contact lens. Infection of the corneal epithelial defect may
occur in conjunction with soft lens usage.
Endothelial cell damage is rare after vitrectomy with
decreased turbulence three-port and 25-gauge systems, less
fluid throughput, better lens removal techniques, and better
irrigating fluids (BSS Plus). If those patients with previous
surgical trauma, glaucoma, or inherited endothelial dys-
trophies are excluded, postoperative corneal edema should
not happen. Prolonged contact of a gas bubble or a silicone
bubble totally filling the anterior chamber can damage the
corneal endothelium and must be avoided by proper post-
operative positioning. Epithelial edema can be managed by
topical hyperosmotic medications, although this is largely
symptomatic treatment. Fortunately, sliding of endothelial
cells and the regaining of function of the remaining cells
cause clearing of corneal edema in most cases. If corneal
edema persists and the eye is required for the patient’s visual
function, Descemet’s Stripping Endokeratoplasty (DSEK) or
penetrating keratoplasty should be performed.
Precipitates on the endothelium are frequently inter-
preted as evidence of inflammation, while in many cases,
they represent pigment released from iris and retinal pigment
epithelium or erythroclasts.
INFLAMMATION
Most cells in the anterior chamber are erythroclasts released
either from the vitreous lamella or as a result of intraop-
erative or postoperative bleeding. If no retinopexy is per-
formed, vitrectomy results in little inflammation. Iris trauma
in conjunction with vitreous surgery results in inflamma-
tion and should be avoided. Retained nucleus appears to
play a greater role in postoperative inflammation than does
persistent cortex. All cases of iris neovascularization and
most cases of retinal detachment have protein release in the
anterior chamber, which is visible as “flare.” Treatment of
the basic disease process by reattachment of the retina or

panretinal photocoagulation (PRP) is more effective than
topical steroids in reducing the flare due to neovasculariza-
tion. Anti–vascular endothelial growth factor (VEGF) com-
pounds have proven to be effective in these cases. Topical
and intraoperative subconjunctival, long-acting steroids are
used in all patients who are not steroid glaucoma responders
(5). This is primarily because any severe inflammation can
result in the development of a cyclitic membrane, periretinal
membranes, and, ultimately, phthisis bulbi. Steroids do not
appear to significantly retard the healing of any of the ocu-
lar structures and should be used to reduce inflammation.
The authors rarely prescribe systemic steroids for primary
ocular conditions. Inflammation not responsive to frequent
topical steroids is treated with repeated subconjunctival
triamcinolone.
IRIS NEOVASCULARIZATION
The cause and treatment of iris neovascularization have been
extensively discussed in Chapter 6. Occasionally, iris neo-
vascularization will appear when retinal detachments with
severe periretinal proliferation are operated on unsuccess-
fully. Intravitreal bevacizumab can be used to cause regres-
sion of the neovascularization postoperatively, followed by
PRP (6). If anti-VEGF therapy and PRP cause regression
of iris neovascularization, the large vessels will not disap-
pear because of rheologic considerations. The examiner
must concentrate on the presence of capillary activity and
endothelial budding on the iris surface rather than the more
impressive large vessels. Peripheral anterior synechiae and
ectropion uveae are late changes and never disappear, even
when capillary involution occurs. Although some clinicians
have emphasized the difficulty in differentiating iris neovas-
cularization from dilation of preexisting stromal vessels, as a
rule, this is not a difficult problem.
GLAUCOMA
Increased intraocular pressure (IOP) from varied mecha-
nisms is an all too frequent complication of vitreous sur-
gery. A high incidence of suspicion and frequent follow-up
is a necessity for recognizing and managing this dreaded
complication (7).
Eryt h ro clast ic (He m o lyt ic) Glauco m a
Erythroclastic (hemolytic) glaucoma was quite frequent after
vitrectomy before vacuum cleaning–extrusion techniques
and intraoperative coagulation methods were utilized (8).
This type of glaucoma is transient and self-limited. The vast
majority of cases can be treated with ocular hypotensive
medications such as topical carbonic anhydrase inhibitors,
beta-blockers, and other agents. Extreme caution should be
used in the administration of systemic hyperosmotic agents to
CHAPTER 31 ■ Com plications of Vitreoretinal Surgery 245
diabetic patients. The authors never use these agents because
of the risk of stroke, myocardial infarction, and ketoacidosis.
Extreme caution should also be taken to prevent the IOP
from going above 30 mm Hg in the patient with vascular
disease, systemic hypotension, and poor retinal perfusion.
Air (Gas) Pup illary Blo ck
When air (gas) is utilized in the vitrectomized, aphakic or
pseudophakic patient, the surface tension effect of the bub-
ble can seal the pupil, just as it does a retinal break, resulting
in a transiris pressure gradient. The continued production of
aqueous then forces the iris forward against the cornea, clos-
ing the angle and elevating the IOP. This can be prevented by
proper postoperative positioning and typically occurs when
the instructions to patients are disregarded. In most cases,
reinforcement of the instructions given to the patient with
assistance from the office and nursing staff can reverse this
pupillary block immediately. If it is unrecognized for several
days, the iris can become adherent to the cornea, requiring
reoperation with a chamber deepening procedure through
the pars plana. Iridectomies do not prevent this complica-
tion and are not required in most vitrectomy procedures.
Ne o vascular Glauco m a
The treatment of neovascular glaucoma (NVG) can be sepa-
rated into components. The basic process of VEGF release
from the retina, which causes the iris neovascularization,
must be treated by anti-VEGF agents, followed by PRP and/
or retinal reattachment. If the pressure is highly elevated,
this must be managed aggressively. Timolol, brimonidine,
and latanoprost are effective in many of these patients in
combination with a topical carbonic anhydrase inhibitor.
In others, these medications appear to have little effect. If
medical management is unable to achieve pressure control
in the sub-35 mm Hg region, endocyclophotocoagulation
can be combined with endo-PRP. Trabeculectomy proce-
dures usually fail unless there is near complete involution of
neovascularization and inflammation. Ahmed valves are the
preferred method of filtering most of these patients. Patients
without light perception are usually taken off their pressure
drops and treated only with topical steroids and pain medi-
cations. High IOP destroys the pain fibers to the eye just
as it destroys the optic nerve; this approach eliminates the
need for enucleation. Enucleation is only necessary if there
is a risk of systemic infection from endophthalmitis, certain
tumors, or a cosmetic appearance that cannot be managed
with a scleral shell.
CYCLOCRYOTHERAPY
In-office cyclocryopexy is usually reserved for No Light
Perception Vision (NLP) cases. Cyclocryotherapy may
cause phthisis bulbi, but it can be used for diseases that
246 SECTION IV ■ Disease-Sp ecific Managem ent
will result in phthisis bulbi without cyclocryotherapy. In the
vitrectomized aphakic eye, there is no matrix in the pupil-
lary plane along which a cyclitic membrane can develop
unless fibrin is present. It is important to use large doses
of periocular steroids in conjunction with cyclocryotherapy.
Cyclocryotherapy can be performed on bare sclera so as to
cause more predictable destruction of the ciliary processes as
well as less postoperative conjunctival and Tenon’s capsule
damage and discomfort. If the cryopexy is used at −80°C and
confluent treatments are held for 1 minute, the effect appears
to be reasonably predictable. It is better to use 180-degree
treatment and additional treatment only in reoperation situ-
ations. If the cryopexy is directed over the ciliary body with
only minimal anterior extension, there is less damage to the
remaining functioning trabecular meshwork. If the iceball
reaches the limbus, this means it is reaching the trabecular
meshwork internally and has extended too far. At this point,
the probe should be reapplied and treatment directed more
posteriorly. Therapeutic ultrasound or transscleral 1,064-nm
laser cyclodestructive procedures may be effective in lieu of
cyclocryotherapy in these cases. Endocyclophotocoagula-
tion is more effective and has fewer complications than cryo,
laser, or ultrasound but requires intraocular surgery.
FILTERING PROCEDURES FOR
NEOVASCULAR GLAUCOMA
Trabeculectomies (9), valve, filtering shunt procedures
(10–14), and pars plana filtering procedures (15) have been
utilized with moderate success in selected patients with
NVG. The Ahmed valve is the preferred procedure at this
time. It appears that in the filtering procedures, there is a
dynamic dilutional effect whereby VEGF is decreased due
to increased fluid throughput. Because of this, the iris neo-
vascularization may involute rapidly and neovascularization
may appear in the bleb.
Extensive PRP to cause involution of iris vessels should
precede filtering procedures whenever possible. Frequent
topical or subconjunctival steroids should be used to suppress
inflammation before and after filtering procedures.
St e ro id Glauco m a
The efficacy of subconjunctival short-acting steroid (Decad-
ron) is so great in the postvitrectomy patient that it should
be universally used, unless the patient is a known steroid
responder. Subconjunctival Kenalog was used in most cases
in earlier years, but the more recent observation that steroid
glaucoma is often not reversible has changed this practice.
Intravitreal triamcinolone results in approximately 30%
incidence of steroid glaucoma, which is often irreversible. In
addition, the incidence of steroid cataracts is approximately
90%. The incidence of steroid glaucoma after fluocinolone
implant (Retsisert, Bausch & Lomb) is approximately 90%,
and over 30% of the patients require a filtering procedure.
Op e n -An g le Glauco m a
Open-angle glaucoma (OAG) can be exacerbated by trans–pars
plana vitrectomy. It appears that this type of OAG is second-
ary to the trauma suffered by the trabecular meshwork from
infusion fluid, cells, cytokines, protein, and debris. In this
way, it is comparable to the metabolic trauma suffered by
the endothelium from infusion fluid and other intraocular
agents. OAG can usually be managed by topical medications
but may require surgery. Chang has postulated that increased
oxygen tension after vitrectomy plays a role.
ENDOPHTHALMITIS
Endophthalmitis was relatively frequent in the early days
of vitrectomy at some institutions; it now fortunately has
become an infrequent complication. The principal author
has had three cases of endophthalmitis in over 26,000 vit-
rectomies. Better equipment, shortened operating times,
povidone iodine prep, microscope drapes, one-piece drapes,
and better irrigating solution (BSS Plus) have probably con-
tributed to the rarity of endophthalmitis now seen after vit-
rectomy. The authors recommend the use of subconjunctival
antibiotics with coverage for both Gram-positive and Gram-
negative bacteria as the final step in surgery. Even though
these agents will only infrequently play a role, the devastat-
ing effect of endophthalmitis should be prevented at all costs.
The authors are currently using Tobramycin and Cefazolin.
If endophthalmitis does occur in the postvitrectomy patient,
there is no need to perform further vitrectomy. A needle may
be placed through the pars plana into the vitreous cavity and
fluid aspirated for culture and sensitivity testing. This same
needle can then be used for intraocular antibiotic injection.
Some have recommended the use of antibiotics routinely in
the infusion fluid. The question about toxicity of these sub-
stances, especially with cumulative dosing, coupled with the
relative infrequency of endophthalmitis makes this approach
unnecessary and possibly dangerous.
RETINAL COMPLICATIONS
Removal of the vitreous causes minimal, if any, inflam-
mation or cystoid macular edema (CME). Postvitrectomy
inflammation is caused primarily by iris trauma, retinopexy,
retinal trauma, laser treatment, agents injected into the eye,
and residual lens material. Vitrectomy can reduce or elimi-
nate CME by debulking inflammatory factors in the vitre-
ous. Vitrectomy, especially when combined with aphakia,
decompartmentalizes the eye, facilitating egress of cells,
proteins, and cytokines through the trabecular meshwork.
Oral and topical nonsteroidal anti-inflammatory agents
(Nevanac) as well as subconjunctival and topical steroids
can have a beneficial effect on postvitrectomy inflammation
and CME (16).

Vascular Occlusio n
The elevation of IOP associated with vitrectomy can create
vascular occlusion and permanent visual loss. Strict atten-
tion to intraoperative and postoperative IOP is required to
prevent this severe complication. Children and adults with
low blood pressure are at high risk for retinal and optic nerve
ischemia from elevation of the IOP.
Re t in al Bre aks an d De t ach m e n t
Retinal breaks may precede vitrectomy, occur at the time of
vitrectomy, or appear in the postoperative course. Any breaks
present at the time of vitrectomy should have been identified
and treated, but, on occasion, surgical visualization prob-
lems prevent this. Retinal breaks occurring at the time of
vitrectomy can be divided into two groups: those caused by
direct retinal trauma and those related to vitrectomy trac-
tion. Any time that instruments are brought near the retinal
surface, the potential exists for the creation of a retinal break.
A retinal tear can be made with any intraocular instrument,
although it is most frequently found as a result of epiretinal
membrane (ERM) surgery in diabetic traction retinal detach-
ment cases. Vitreous removal results in an obligatory trac-
tion on the retina from the suction force on the vitreous fiber
framework. This mechanism is similar to that of an aphakic
retinal detachment occurring from anterior movement of the
vitreous after capsule rupture during cataract surgery. With
this form of detachment, there are small breaks at the poste-
rior edge of the vitreous base. These are frequently difficult
to recognize at the time of vitrectomy or are only of partial
thickness at that time, becoming full thickness later. Thus,
there is a built-in incidence of aphakic-like retinal detach-
ment as a result of vitreous removal by any method. More
direct trauma to the peripheral retina occurs from the entry
of instruments through the pars plana. Large instruments
with abrupt increases in diameter can cause dialyses, while
small instruments can push the vitreous base in front of
them and create dialyses or small breaks as well.
Late retinal breaks occur from several mechanisms.
Trauma to the retina at the time of vitrectomy may result in
partial-thickness retinal loss with subsequent retinal break
formation. Postoperative traction can occur from residual
vitreous particularly that incarcerated in the sclerotomies.
Any form of periretinal proliferation can cause retinal short-
ening and retinal breaks from tangential traction. In some
cases, large, atrophic retinal breaks will occur without any
evidence of traction or surgical trauma. These are seen in the
context of ischemic retinopathies and appear to be due to
retinal necrosis from vascular occlusive disease.
No n rh e g m at o g e n o us Re t in al De t ach m e n t
Periretinal migration and proliferation from glial, retinal pig-
ment epithelial, or wound-related cellular proliferation can
lead to elevation of the retina without a rhegmatogenous
CHAPTER 31 ■ Com plications of Vitreoretinal Surgery 247
component. Required ERM peeling, segmentation, and
delamination elicit a reparative effort with further prolifera-
tion, causing retinal elevation. Residual ERM can contract
also, creating nonrhegmatogenous retinal elevation. If these
elevated areas are small, they can be observed without surgi-
cal intervention. At times, vitreous will unintentionally be
left, causing residual traction on the retina. If extramacular
retina is elevated, it can be observed postoperatively for an
extensive period with much the same criteria as those used
at the time of the original surgery. Nonrhegmatogenous mac-
ular elevation or any elevation accompanied by iris neovas-
cularization should indicate the need for reoperation with
scissors segmentation and delamination methods. At times,
segmental scleral buckling is required.
Man ag e m e n t o f Re t in al De t ach m e n t
Rhegmatogenous detachment repair after vitrectomy almost
always requires gas or silicone for surface tension manage-
ment. Because vitreous removal markedly decreases the vis-
cosity of the vitreous fluid, a very small break will rapidly
result in total retinal detachment. With the use of the sur-
face tension effects of air (gas), these cases can be reattached
more expeditiously and safely. In virtually all instances, the
procedure is set up for full vitrectomy with the customary
incisions, the endoilluminator, cutter, and tools. Any residual
vitreous traction is removed and the visualization improved
by removal of any blood or debris. If ERM is causing reti-
nal foreshortening, peeling, segmentation, and delamina-
tion are utilized. After removal of the causative traction,
internal drainage of subretinal fluid (SRF), internal fluid-air
exchange, and completion of internal drainage of SRF are
utilized. This is followed by endolaser retinopexy unless the
break is in the macular or peripapillary region.
INTRAOCULAR HEMORRHAGE
Postoperative intraocular hemorrhage is rare except in the
diabetic patient. Diabetic patients develop postoperative
hemorrhage approximately 50% of the time. Resected vas-
cular tissue, sclerotomy wounds, iris neovascularization, and
incomplete regression of retinal neovascularization are the
most frequent causes of postoperative intraocular hemor-
rhage. As discussed in this chapter, hemorrhage cases should
be followed up with ultrasound and reoperated only if reti-
nal detachment occurs. If the patient is blind in an only eye
or both eyes, reoperation to remove blood should be consid-
ered for the better eye.
If indicated, blood removal should be performed in all
instances using the full setup and three-port, usually 25-gauge
technique. This permits extrusion, bipolar diathermy, endo-
photocoagulation, and dissection of any significant ERM to
be done. If iris neovascularization is present, blood removal
should be performed to accomplish retinal reattachment
and/or completion of PRP.
248 SECTION IV ■ Disease-Sp ecific Managem ent
CATARACT
Many surgical factors add to the baseline incidence of cataract
associated with the ocular diseases requiring vitreous sur-
gery (17). Lens opacities are a frequent but easily manage-
able postvitrectomy complication. Lens removal should be
considered to address the patient’s visual needs or if the view
of the retina is lost, preventing good management.
Phaco-vit is preferred to pars plana lensectomy approach
for cases requiring combined vitreoretinal surgery.
SUTURE EXTRUSION
On occasion, 8-0 monofilament nylon sclerotomy sutures
will erode through the conjunctiva. This problem has been
virtually eliminated with the advent of sutureless 25-gauge
vitrectomy. Because this is a running suture, it is quite dif-
ficult to remove at the slit lamp. It is better to use topical
anesthesia and the laser to melt the protruding end. Alterna-
tively, a disposable electrothermal cautery can be used if the
lids are retracted securely. If loops of suture erode, they can
be severed with a blade or Vannas scissors at the slit lamp
in the office. The ends of 5-0 nylon scleral buckle sutures
can be trimmed in the office, but this may result in buckle
extrusion. With the use of hard silicone explants, copious
antibiotic flushing, and a 1.0-mm conjunctival flap, infected
buckles are extremely rare. If infection occurs, it should be
managed by removal of the explant materials and irrigation
with saline and antibiotics.
FOLLOW-UP INTERVALS
Typically, vitrectomy patients are treated as outpatients and
therefore go home on the day of surgery. The first post-
operative visit should be on the day after surgery, and the
subsequent exam should be in 3 weeks in most instances.
It would be unusual to experience any reason to reoperate
in the first 3 weeks, and a 3-week follow-up exam discov-
ers the highest incidence of treatable postoperative com-
plications. If this period were extended to 1 month, in the
diabetic vitrectomy patient, there would be a much higher
incidence of NVG. Most PVR recurrences occur in the first
3 weeks. Most rhegmatogenous postvitrectomy retinal
detachments will also present by the 3-week mark and can
be effectively managed at that time. The first postoperative
day visit is usually sufficient to cover the acute endophthal-
mitis risk.
If the patient is doing well at the 3-week visit, he or she
can be given a 6-week visit unless the patient is a diabetic.
The diabetic patient should return in 3 weeks for further
examination for iris neovascularization and glial recurrence.
In all patients, it is important to emphasize that they should
check their vision every day and should contact the phy-
sician immediately if visual loss or pain occurs. It is criti-
cal to instruct the office staff about the access these patients
must be afforded to postoperative care. The complexity of
postvitrectomy patients simply prohibits the use of a wait-
ing list. As a rule, diabetics should be followed up at 6-week
to 3-month intervals until they are stable for 1 year. At this
point, the interval can be lengthened.
References
1. Perry HD, Foulks GN, Thoft RA, et al. Corneal complications
after closed vitrectomy through the pars plana. Arch Ophthal-
mol 1978;96(8):401.
2. Brightvill FS, Myers FL, Bresnick GH. Postvitrectomy keratop-
athy. Am J Ophthalmol 1978;85:651.
3. Aaberg TM, Van Horn DL. Late complications of pars plana
vitreous surgery. Ophthalmology (Rochester) 1978;85(2):116.
4. Kenyon KR, Stark WJ, Stone DL. Corneal endothelial degen-
eration and fibrous proliferation after pars plana vitrectomy.
Am J Ophthalmol 1976;8(4):486.
5. Hall AJ. Effective local treatment for severe ocular inflam-
matory disease: towards the holy grail. Clin Exp Ophthalmol
2001;29(1):1.
6. Lupinacci AP, Calzada JI, Rafieetery M, et al. Clinical outcomes
of patients with anterior segment neovascularization treated
with or without intraocular bevacizumab. Adv Ther 2009;
26(2):208–216. [Epub 2009 Feb 20.]
7. Campbell DG, Simmons RL, Tolentino FL, et al. Glaucoma
occurring after closed vitrectomy. Am J Ophthalmol 1977;
83(1):63.
8. Brucker AJ, Michels RG, Green WR. Pars plana vitrectomy in
the management of blood-induced glaucoma with vitreous
hemorrhage. Am J Ophthalmol 1978;10(10):1427.
9. Herschler J, Agness D. A modified filtering operation for vascu-
lar glaucoma. Arch Ophthalmol 1979;97:2339.
10. Krupin T, Kaufman P, Mandell A, et al. Filtering valve implant
surgery for eyes with neovascular glaucoma. Am J Ophthalmol
1980;89:338.
11. Krupin T, Kaufman P, Mandell A, et al. Long-term results of
valve implants in filtering surgery for eyes with neovascular
glaucoma. Am J Ophthalmol 1983;95:775.
12. Molteno ACB, Van Rooyen MMB, Bartholomew RS. Implants for
draining neovascular glaucoma. Br J Ophthalmol 1977;61:120.
13. Schocket SS, Lakhanpal V, Richards RD. Anterior chamber
tube shunt to an encircling band in the treatment of neovascu-
lar glaucoma. Ophthalmology 1982;89:1188.
14. Schocket SS, Nirankari VS, Lakhanpal V, et al. Anterior cham-
ber tube shunt to an encircling band in the treatment of neo-
vascular glaucoma and other refractory glaucomas; a longterm
study. Ophthalmology 1985;92:553.
15. Sinclair SH, Aaberg TM, Meredith TA. A pars plana filtering
procedure combined with lensectomy and vitrectomy for neo-
vascular glaucoma. Am J Ophthalmol 1982;93:185.
16. Sjaarda RN, Glaser BM, Thompson JT, et al. Distribution of
iatrogenic retinal breaks in macular hole surgery. Ophthalmol-
ogy 1995;102(9):1387–1392.
17. Blankenship G, Cortez R, Machemer R. The lens and pars
plana vitrectomy for diabetic retinopathy complications. Arch
Ophthalmol 1979;97(7):1263.
C H AP TER
32
SURGICAL SELF-EDUCATION
While there are many excellent courses, articles,
and textbooks available concerning vitreo-
retinal surgery, improvement in judgment and
surgical skills must principally come about by
self-education. The complexity of high-technology vitre-
ous surgery on high-risk patients demands a continued
assessment of surgical and biologic results. Vitreous surgery
requires an excellent training in microsurgery and retinal
diseases. Eye bank eyes, porcine eyes, and molded rubber
practice eyes can be used for surgical practice (1–6). It is
simply poor judgment to begin vitreous surgery or a new
technique on the human patient without sufficient practice.
After reading the available literature, visiting other surgeons,
and attending appropriate courses, progress will be made in
the laboratory. When sufficient competency is obtained in
the laboratory, the beginning surgeon should assemble the
disposables and equipment required for simulated surgery.
Simulator technology has improved, gaining substantial
fidelity in recent years, and has the advantages that disease
states and intraoperative complications can be simulated. The
EyeSi simulator has an excellent vitreoretinal module and is
highly recommended. Regardless of the presence of other
vitreous surgeons at the same institution, it is the responsi-
bility of the beginning surgeon to go through this practice
surgery approach. It is absolutely the responsibility of each
surgeon to make certain that all equipment is available and
functioning. Unfortunately, many surgeons fall into the trap
of placing this responsibility upon technicians, nurses, and
ancillary personnel. Practice surgery in the actual operating
room should be repeated on the days preceding vitreous sur-
gery if the case in question has not been approached before
or the procedures are done infrequently.
The great complexity of vitreoretinal surgery requires
an honest assessment of the surgeon’s own capabilities. It is
simply inadequate to perform vitrectomy without stereopsis.
There are many areas of medicine that are less demanding
in the requirements for stereopsis, and the surgeon should
not perform vitreous surgery without stereopsis. Red-green
color blindness is a major handicap because it makes the
diagnosis of iris neovascularization, small retina breaks, and
the regression of neovascularization much more difficult to
see. It is even important to attempt an assessment of one’s
temperament. Vitreous surgery requires a calm but rapid
and efficient approach. A surgeon who becomes very tense
and inefficient in times of surgical stress has no place in vit-
reous surgery. A person so compulsive and rigid that nec-
essary changes in the game plan produce overwhelming
stress probably should not be performing vitreous surgery.
Although ego and economic factors unfortunately influence
some surgeons’ decisions, the pleasure is short lived if the
results are poor, resulting in a damaged and unhappy patient
and possibly a lawsuit. It therefore becomes important to
look realistically at the demands for vitreous surgery in the
individual’s practice, with an intent to determine if certain
procedures can be done frequently enough to attain suffi-
cient surgical skill.
OUTCOMES ANALYSIS
The collection of preoperative, operative, and postopera-
tive information is essential to the self-assessment of surgi-
cal techniques and skills. While some feel that this is the
obligation of so-called academic institutions, it is, in fact,
the obligation of each and every surgeon. Some surgeons
are able to achieve series large enough for publication, but
each individual must produce a series for comparison. The
importance is not so much in contributing to the literature
as in being apprised of one’s own outcomes. Outcomes
research begins with careful preoperative evaluation and
prospective recording of this information. It is best to have
a format on which these data can be recorded, and exten-
sive use of abbreviations and recognized grading systems
should be utilized. With the use of a scribe accompanying
the examiner, the information can be dictated in abbrevia-
tion format and recorded extremely rapidly without the
unavoidable errors that occur with trying to recall this later
at the time of dictation. Although it is permissible to fill
out sheets at this time, if they are secondary to the primary
charting method, details tend to be overlooked in a busy
practice. It is critical to determine the parameters that one
wishes to follow preoperatively and postoperatively and
249
250 SECTION IV ■ Disease-Sp ecific Managem ent
to record all of this information on each patient to permit
biostatistical evaluation at a later date.
It is best to compile this information by disease cat-
egory so that discrete biologic groups may be identified.
This compilation can be as simple as single sheets that list
patients with a given disease state and entrance criteria, with
columns for preoperative, operative, and postoperative find-
ings and all complications. These sheets should be filled out
immediately following each day’s examinations so that miss-
ing findings can be determined at that point. The surgical
part of these forms should be filled out immediately follow-
ing surgery so that it will not have the inaccuracies inher-
ent in subsequent abstraction from postoperative dictation.
Similarly, the postoperative findings should be recorded after
each visit, preferably while the patient is still present, so that
any missing information can be obtained.
The computers utilized in office practice could make
this process more efficient, but useful computerized medical
record software is just becoming available and if one waits
until computer skills or capabilities are sufficient, much data
and useful information will be lost.
It is essential to determine average success rates from the
data forms and to compare complication rates to published
outcomes. This should be done at least every 3 months, with
a stimulus being provided by a quarterly report, an upcom-
ing paper, or an upcoming lecture. The data then become
very helpful in predicting the outcome of surgery for patients
and in self-assessment for surgical improvement. Care must
be taken to compare similar biologic groups that are selected
with given entrance criteria. For example, a vitreous hemor-
rhage patient should not be compared to a traction detach-
ment patient with respect to visual improvement because
simple clearing of the media improves the vision in a differ-
ent manner than macular reattachment.
SURGICAL DATA
It is important to dictate an extremely complete and honest
operative note at the end of each procedure. This should be
done describing every aspect of the technique and all surgi-
cal findings. While some physicians use surgical forms, they
are usually too stereotyped and inflexible for this complex
type of surgery. Similarly, while photographs and drawings
can complement the description, carefully described surgical
findings are very helpful in following up the patients in the
office. This approach helps not only in outcomes research but
also in understanding complications on an individual basis.
It is suggested that in addition to the copies of the operative
notes in the hospital and office records, a third copy be kept
in the computer or separate files. These sequential operative
notes should then be abstracted with cross-referencing for
certain findings or techniques that will benefit from subse-
quent analysis. For example, all macular hole patients who
underwent peeling of the cortex from the optic nerve can be
assessed for visual results to determine if optic nerve damage
is a complication of this method. Again, a computer-based
approach can be quite effective but requires additional work
to set up, and its unavailability should not be used as an
excuse for not having continuous data monitoring.
It is helpful to have one member of the surgical assistant
team monitor these data with the surgeon so as to increase
his or her involvement and understanding. If a surgical team
member helps by abstracting operative records, it adds imp-
artial credibility and enhances surgical understanding as
well. If this same individual is involved in preoperative and
postoperative clinical photography and other special exami-
nation techniques, it increases concern for the patient and
understanding of prognostic and management factors.
CORRESPONDENCE
All postoperative follow-up information from referring doctors
should be coded immediately into the file sheets as well as
on the chart. In this way, a retrospective chart review with all
its inherent inadequacies is avoided entirely. The quarterly or
pretalk data compilation will uncover many patients who had
inadequate follow-up. At this time, the referring doctor should
be contacted by phone or postoperative follow-up forms
should be mailed in order to complete all follow-up informa-
tion. These frequent checkups serve a purpose in stimulating
follow-up by the surgeon and referring physicians. Preprinted
follow-up forms can be made available to referring physicians,
which can act as a stimulus for better data retrieval. Unfortu-
nately, many practitioners do not refract other physician’s post-
operative patients, making the visual acuity data inadequate.
One then must specifically inquire and encourage the use of
best-refracted visions.
GROUP EDUCATION
The use of effective outcomes research as described previ-
ously contributes to the quality of presentations to other sur-
geons and at meetings. It is important to frequently attend
meetings with surgeons performing similar work in order to
upgrade medical and surgical knowledge. Unfortunately, the
literature is months to years behind in reporting newer meth-
ods; more rapid education is possible by attending meetings.
If all speakers emphasize this approach to outcomes research
with at least a modicum of knowledge concerning biostatis-
tics, better communication is possible. Certainly, everything
cannot be studied in a randomized masked study, but accu-
rate compilation of results is nevertheless mandatory.
INTERACTION WITH COMPANIES
The careful analysis of results with different surgical
techniques permits better communication with medical
equipment manufacturing companies. This should not be

done so as to make the equipment a scapegoat for surgical
failure but rather to provide constructive advice concern-
ing equipment improvement. It is usually better to go to
major equipment manufacturers for prototype equipment
than to local machine shops. If these devices are made in
local shops, they are not accessible to colleagues and are
not prototyped with a view to future manufacturing meth-
ods. It is very important for the surgeon to have frequent
and open communication with companies so as to cre-
ate a climate of intellectual cooperation that encourages
equipment improvement. An extension of this attitude
toward the surgical team and colleagues benefits all those
involved.
CHAPTER 32 ■ Surgical Self-Education 251
References
1. Michels RG. Intraocular fluorescein in experimental vitrectomy.
Ophthalmic Surg 1977;8:139.
2. Bensen WE. Vitrectomy in rabbit eyes (appendix). In: Vitrectomy,
a pars plana approach. New York: Grune & Stratton, 1975.
3. O’Malley C. Learning surgery without risk or anxiety. Ocutome
Newsletter1977; 2(3).
4. Borirak-Chanyavat S, Lindquist TD, Kaplan HJ. A cadaveric eye
model for practicing anterior and posterior segment surgeries.
Ophthalmology 1995;102(12):1932–1935.
5. Eckardt U, Eckardt C. Keratoprosthesis as an aid to learning sur-
gical techniques on cadaver eyes. Ophthalmic Surg 1995;26(4):
358–359.
6. Moorehead LC. Practice vitrectomy. Arch Ophthalmol 1980;
98(7):1297–1298.
 Index

Note: Page numbers followed by “f” indicate figures; page numbers followed by “t” indicate tables.

A for ciliary block glaucoma, 91–92 Blood pressure

Adjuvant therapy congenital and pediatric cataracts, 90 during general anesthesia, 40–41
in management of intermediate uveitis, 223 for cystoid macular edema, 92–93 Bright flash electroretinography, 15
Age infusion sources for, 91 B-scan ultrasonography, 12
in preoperative assessment, 2 instrumentation for, 76 Buckling. See Scleral buckling
Air intraocular lenses, 93–96, 94f, 95f
vitrectomy under, 68, 69f limbal versus pars plana approach in, 78 C
Air-gas exchange pars plana lensectomy, 78–90 Cannulas
and general anesthesia, 43 for phacolytic glaucoma, 93 extrusion
for giant retinal breaks, 139 pupillary membranes, 90–91 for subretinal fluid drainage, 60f
pupillary block and, 245 translimbal, 91 infusion, 22, 113–115, 114f, 115f
subretinal fluid drainage and, 62, 63f traumatic cataracts, 90 soft-tip
Air/silicone exchange. See Silicone infusion ultrasonic fragmentation in, 76–78 for subretinal fluid drainage, 60f
Anesthesia in uveitis, 93 trocar system of, 103–104, 104f, 105f
anticoagulation and, 43 for vitreocorneal touch, 91 withdrawal and wound closure, 111
carbon dioxide issues during, 42–43 for vitreous loss in cataract surgery, Capsulectomy
diabetic retinopathy, 175 92, 92f pars plana
for endophthalmitis, 43 Antibiotics after lensectomy, 83
facial nerve blocks, 42 in infusion fluid, 28 Capsulotomy
general intracameral pars plana
air/gas exchange and, 43 for endophthalmitis, 219 after lensectomy, 83
blood pressure during, 40–41 in posterior vitrectomy, 73–74 Carbon dioxide
versus local, 40 subconjunctival intraoperative build-up of, 42–43
intraconal for endophthalmitis, 219 Cardiopulmonary disease
technique for, 42 topical preoperative assessment in, 3
local for endophthalmitis, 219 Cataract(s)
choice of, 41–42 toxicity of, 239 after vitrectomy for epimacular
psychological preparation for, 41 Anticoagulation, 43 membranes, 162
reblocking during, 42 Aphakic keratoplasty congenital and pediatric, 90
sedation during, 41 anterior vitrectomy and, 93 postoperative
monitoring during, 40 A-scan ultrasonography, 12 in diabetic retinopathy, 185
with open globe, 43 Aspiration traumatic, 90
and operating room efficiency, 40–44 technology for in vitreoretinal surgery, 248
postoperative pain, 43–44 cutter movement, 25–26 Cataract surgery
preoperative evaluation for, 40 port configuration, 26 capsular defects in
presurgical self-sharpening, 26 recognition and management of, 84–85
for penetration of eye, 239–240 vitreous cutter design issues, 26–27 posterior lens dislocation in, 83–90, 86f–89f
for scleral buckling, 43 with ultrasonic fragmentation, 77 vitreous loss in
for vitreoretinal surgery, 42 Avastin anterior vitrectomy for, 92, 92f
Angioscopy in choroidal neovascular membranes, 193 cellulose sponge vitrectomy, 85, 86f
fluorescein Cellular migration/proliferation
in posterior vitrectomy, 73–74 B traumatic, 205
Anterior chamber aspiration Barcode reader wand, 35, 36f Cellulose sponge vitrectomy, 85, 86f
for endophthalmitis, 217–218 Barrier concept Children
Anterior segment in diabetic retinopathy, 186–187, 186f cataracts in, 90
problems of Bimanual bipolar diathermy traction retinal detachment in, 231
vitrectomy techniques and, 76–96 in posterior vitrectomy, 72 Choroidal hemorrhage, 241
Anterior vitrectomy Biologic modifiers Choroidal neovascular membranes
aphakic filtering procedures and, 92 for macular holes, 169 anti-VEGF therapy strategies, 194
aphakic keratoplasty and, 93 Bipolar diathermy systems, 30 combination therapy, 194–195
for aphakic pupillary block, 91 Bleeding. See Hemorrhage dietary supplementation for, 193–194

253
254 Index

Choroidal neovascular membranes (Continued) cysts, 175 Endophthalmitis

drusen, laser treatment of, 194 imaging for, 171, 172f anesthesia and, 43
focal thermal laser ablation, 193 infusion fluid, 184–185 anterior chamber versus vitreous aspiration,
macular translocation, 201–202 macular edema, 175 217–218
photodynamic therapy, 194 medical issues in, 171 antibiotics for, 219
plasmapheresis, 194 nonproliferative retinopathy complications of, 220
results of, 200 corticosteroids, 172–173 endogenous, 216
retinal pigment epithelium transplantation, macular edema, 172–173 epimacular membranes and, 162
200–201 vascular endothelial growth factor exogenous, 216–217
submacular surgery antagonists, 173 in-office vitreous tap and intravitreal
and macular translocation, 196, 196f proliferative retinopathy, 173–174 antibiotics injection, 218
for subfoveal membranes, 196 results of, 185 postoperative, 111
surgical sequence for, 196–198, 197f–200f retinal detachment results of, 219–220
Color discrimination ancillary epiretinal membrane management surgical sequence and techniques for
preoperative assessment of, 4 techniques, 183 trans–pars plana versus translimbal
Companies anti-vascular endothelial growth factor approach, 218–219
interaction with, 250–251 therapy, 183 vitrectomy, 219
Congenital cataracts, 90 epiretinal membranes, 181, 182f TASS, 217
Conjunctiva hemostasis, 182–183, 182f vitrectomy for, 218
displacement of, 46–47 inside-out versus outside-in technique, 181 in vitreoretinal surgery, 246
incisions of, 112 panretinal photocoagulation, 183 Endoscope systems, 38–39
in trauma, 206 retinectomy, 183 Entoptic phenomenon
and tenon’s capsule, 116 scissors segmentation and delamination, preoperative assessment of, 4
Constellation architecture 180–181, 180f, 181f Epimacular membranes
auto gas fill, 34, 34f silicone oil, 183 vitreomacular traction syndrome
barcode reader wand, 35, 36f viscodissection, 181–182 case selection, 157–158
embedded PurePoint 532-nm laser, 34, 34f retinal thickness measurement, 171 cataract, 162
power forceps, 35, 35f retinopexy, 185 coexistent cataract, 160–161
power scissors, 35, 35f scleral buckling, 185 cystoid macular edema, 161–162
UltraVit 5,000 cpm vitreous cutter, 32, 33f screening for, 171 endophthalmitis, 162
xenon illuminations systems, 33, 33f submacular exudates, 175 epimacular proliferation, 162
Contrast sensitivity surface tension management, 183–184 epiretinal membrane removal, 159–160,
in preoperative assessment, 4 surgical sequence and techniques 159f–161f
Cornea anesthesia, 175 history of, 157
preoperative slit lamp examination of, 5 incisions, 175 membrane peeling, 158–159
Corneal complications, 244 lens management, 175–176 nonrhegmatogenous proliferative
Corneal opacification, 213–214, 214f posterior vitreous detachment, 176–178, vitreoretinopathy, 158
Corneal-scleral laceration, 212–213 177f–179f pathogenesis, 157
Correspondence vitrectomy, 176, 176f proliferative vitreoretinopathy
management of, 250 traction retinal detachment, 174 recurrence, 161
Corticosteroids visualization, 184 retinal breaks, 160, 161
diabetic macular edema, 172–173 vitrectomy, 174 retinal whitening, 162
Cryopexy, 119 vitreous hemorrhage, 174 rhegmatogenous retinal detachment, 161
Cutters. See Tissue cutting Diathermy, 30 surgical sequence and techniques, 158–161
Cyclocryotherapy, 245–246 bimanual bipolar visual loss and, 157
Cystoid macular edema in posterior vitrectomy, 72 visual results, 161
anterior vitrectomy for, 92–93 transscleral, 73 Epiretinal membrane dissection
epimacular membranes and, 161–162 Direct (transscleral) needle drainage bimanual surgery, 53, 54f
Cysts of subretinal fluid, 70–71, 70f, 71f “chopsticks” membrane peeling, 56
in diabetic retinopathy, 175 Discission, 90 cutter delamination, 56, 56f, 57f
Dislocation en bloc in, 53
D lens flowchart, 55
Diabetes mellitus in cataract surgery, 83–90, 86f–89f hemostasis, 54
preoperative assessment in, 2–3 Double penetrating injury, 212 membrane peeling in, 49–50, 49f
Diabetic retinopathy Drainage retinotomy, 61 power scissors and forceps in, 54
barrier concept in proliferative vitreoretinopathy, 146–148,
anterior vitreous cortex fibrovascular prolif- E 148f
eration, 187 Education scissors segmentation and delamination in,
glial recurrence, 187 group, 250 50–53, 51f–53f
rhegmatogenous retinal detachment, 187 surgical self-education, 249–251 simplified, 49
VEGF encounters, 186, 186f Electroretinography suction, 56
complications bright flash, 15 surgery stabilization in, 54, 55f
erythroclastic (hemolytic) glaucoma, 185 Endocapsular lensectomy, 80–83, 80f–82f vacuum cleaning–extrusion, 56–58, 57f
hemorrhage, 185 Endodiathermy, 72–73. See also Diathermy viscodissection, 53–54
neovascular complications, 185–186 Endoillumination, 28, 28f Epiretinal membranes
postoperative cataract, 185 Endophotocoagulation, 29. See also Laser in diabetic retinopathy, 181, 183
contraindications, 175 endophotocoagulation preoperative assessment of, 8
 Index 255

removal of, 159–160, 159f–161f anterior vitrectomy for, 93 Intraocular hemorrhage, 247
retinopathy of prematurity, 230 steroid, 246 Intraocular lens
Epithelial ingrowth, 96 in trauma, 214 dislocated
Erythroclastic (hemolytic) glaucoma, 245 Glial recurrence lens removal, 96
Extrusion method in diabetic retinopathy, 187 repositioning of, 93–96, 94f, 95f
in epiretinal membrane dissection, 56–58, 57f Gray scale ultrasonography, 13 epithelial ingrowth of, 96
Group education, 250 fogging during fluid-air exchange, 62
F implantation of, 90
Facial nerve blocks, 42 H removal of
Fluid Hemorrhage before trans–pars plana vitrectomy, 96
infusion, 28 in diabetic retinopathy, 185 retrolental membranes and, 93
diabetic retinopathy, 184–185 intraocular, 247 Intraocular lymphoma
posterior vitrectomy, 45 submacular, 195 cytology and flow cytometry for, 222
Fluid-air-gas exchange. See also Air-gas exchange suprachoroidal, 241–243 Intraocular pressure
giant retinal breaks, 139 vitreous slit lamp preoperative assessment of, 5–6
in-office, 122–123, 123f in diabetic retinopathy, 174 Intraoperative video recording, 19
Fluidics, 104–106, 106f Intravitreal antibiotics, 218
aspiration, 22–23 I Intravitreal injection technique, 204
infusion, 21–22, 21f Illumination, 37–38 Iris
surgical, 21 Indirect ophthalmoscopy anterior chamber
Fluorescein angiography preoperative, 6 slit lamp preoperative examination
in diabetic retinopathy, 171 Inflammation of, 5
Fluorescein angioscopy in proliferative vitreoretinopathy, 155 Iris neovascularization
in posterior vitrectomy, 73–74 in vitreoretinal surgery, 244–245 in vitreoretinal surgery, 245
Follow-up Infusion fluid, 28 Irrigation
intervals for, 248 diabetic retinopathy, 184–185 saline, 88, 88f
Forceps posterior vitrectomy, 45
constellation architecture, 35, 35f Infusion instruments K
power scissors and, 54 alternative, 27–28 Keratoplasty
Foreign body. See Intraocular foreign body cannulas, 27, 27f aphakic
Fragmentation enter sites for, 27 anterior vitrectomy and, 93
in pars plana lensectomy, 78–79 Infusion needles, 28 Keratoprosthesis
ultrasonic Infusion systems classification of, 236
in anterior vitrectomy, 76–78 for anterior vitrectomy, 76, 77f office examinations in, 237
Fragmenter, 79, 79f Injuries. See Trauma vitrectomy in
avoidance of vitreous in, 83, 84f In-office fluid-air-gas exchange, 122–123, 123f 25-gauge trocar cannulas, 237
Full function probes, 27 Instruments peripheral retinal visualization and
for anterior vitrectomy, 76 management, 237
G cutters, 19–22 retroprosthetic membranes, 238, 238f
General anesthesia. See Anesthesia, general ergonomic issues and, 28–29
Giant retinal breaks infusion, 27 L
case selection, 138 peak electron avalanche knife, 20 Laser endophotocoagulation, 72–73
definition of, 138 port configuration, 26 Laser interferometry, 4
management of, 143 presurgical testing of, 19 Laser photocoagulation
proliferative vitreoretinopathy and, 138 self-sharpening, 26 cryopexy and, 119
subretinal fluid drainage, 64 storage of, 18–19 PASCAL laser, 118
vitrectomy for, 138–139, 139f Intracameral antibiotics, 219 retinal breaks, prophylaxis for, 118–119
fluid-air-gas exchange in, 139 Intraconal anesthesia, 42 technique for, 119
incisions in, 138 Intraocular foreign body Laser retinopexy, 141
lens management in, 138 antibiotic-steroid therapy, 211–212 Laser therapy
patient education for, 138 conjunctival incisions, 206 for retinopathy of prematurity, 227
perfluorocarbon liquids in, 139, 140f, cylindrical, 208, 208f for venous occlusive diseases, 190–191
141, 141f encapsulated, 209–211, 209f–211f Lasers, tissue cutting, 20
postoperative surface tension inert versus toxic, 206, 206f Lens
management, 142 large, 208–209, 208f management
results of, 142–143 lens removal, 207 in diabetic retinopathy, 175–176
retinopexy, 141–142, 142f localization of, 8 in double penetrating injury, 212
Glaucoma moderately sized, 207, 207f in giant retinal breaks, 138
air (gas) pupillary block, 245 retinopexy, 211 in proliferative vitreoretinopathy, 145
ciliary block scleral buckling, 211 in severe uveitis, 224
anterior vitrectomy for, 91–92 sclerotomies, 207 removal of
erythroclastic (hemolytic), 185 surface tension management, 211 conjunctival displacement, 46–47
erythroclastic (hemolytic) glaucoma, 245 surgery for eye, stabilization of, 46
neovascular, 245 sequence and techniques for, 206–212 timing of, 213
filtering procedures for, 246 timing of, 206 in trauma, 207
open-angle, 246 vitrectomy for, 207 wound construction, 46
phacolytic wound repair, 206 wound leaks, 47
256 Index

Lens capsule defects Operating microscope, 29–30, 29f presurgical anesthesia for, 239
in recognition and management of, 84–85 diathermy, 30 recognition, 239–240
Lens loops endophotocoagulation, 29 surgical indications for, 240
vitreous traction from, 88, 88f Operating room environment, 18–29 toxicity, 239
Lens material aspiration fluidics and vitreous cutting, Perfluorocarbon liquids, 151–152, 152f
posterior dislocation of 22–23, 23f in floating of lens material, 89, 89f
in cataract surgery, 86, 87f–89f, 88–90 instrument storage, 18–19 in vitrectomy
Lens opacities personnel, 19 for giant retinal breaks, 139, 140f,
slit lamp preoperative examination and, 5 presurgical equipment testing, 19 141, 141f
Lensectomy. See Pars plana lensectomy scheduling, 18 Perfluoron, 135–136, 135f, 136f
Limbal approach tissue cutting, 19–22, 20f, 21f Peribulbar (extraconal) anesthesia, 42
in anterior vitrectomy, 78 video recording, 19 Persistent fetal vasculature, 231–232, 232f
Liquid perfluorocarbon techniques, 62–64, 63f vitrectomy technique Persistent hyperplastic primary vitreous. See
Local anesthesia. See Anesthesia, local aspiration technology, 25–27, 25f Persistent fetal vasculature
cutter technology, 24–25, 24f Phacoemulsification
M endoillumination, 28, 28f with pars plana lensectomy, 79
Macular edema infusion devices, 27–28 vitrectomy procedures, 99t
cystoid, 92–93, 161–162 infusion fluid, 28 concepts of, 99–100
diabetic, 172–173 physics of, 24 principles of, 100–101
Macular holes role of, 23–24, 24f silicone oil issues, 99
arcuate retinotomy, 169, 169f Operating room setup, 30–31, 30f for silicone oil-filled eyes, 101, 102
biologic modifiers, 169 prep technique, 31 techniques, 101
hole closure mechanisms, 166 sterilization systems, 31 Phacolytic glaucoma, 93
indications for, 164 surgical effciency, 31 Photocoagulation
internal limiting membrane peeling, 166, Ophthalmoscopy laser. See Laser photocoagulation
167f, 168 indirect Photodynamic therapy
ocular coherence tomography, 164–165 preoperative, 6 for choroidal neovascular membranes, 194
pathogenesis of, 164 Optical coherence tomography, 13–14, 14f, 15f Phototoxicity, 38
posterior vitreous detachment creation Pneumatic retinopexy, 121–122, 122f, 123f
methods, 166, 167f P Polymerase chain reaction, 222
postoperative positioning, 169 Pain Ports
results of, 169 during vitreoretinal surgery, 42 configuration of, 26
surface tension management, Panretinal photocoagulation, 183 Posterior segment surgery. See Posterior
168–169, 168f Pars plana approach vitrectomy
surgery for, 165–166 in anterior vitrectomy, 78 Posterior vitrectomy
Macular translocation, 196, 201–202 Pars plana capsulectomy under air, 68, 69f
Maddox rod after lensectomy, 83 antibiotic injection in, 73–74
in preoperative assessment, 4 Pars plana capsulotomy in corneal-scleral laceration, 213
Medical equipment companies after lensectomy, 83 endodiathermy, 72–73
interaction with, 250–251 Pars plana lensectomy epiretinal membrane dissection in, 48–58. See
Membrane peeling development of, 78 also Epiretinal membrane dissection
methods of. See Epimacular membranes; endocapsular, 80–83, 80f–82f fluorescein angioscopy, 73–74
Epiretinal membrane dissection fragmenter performance, 79, 79f infusion fluid in, 45
Membranectomy, 90–91 fragmenter techniques in, 78–79 laser endophotocoagulation, 72–73
dense, 91 indications for, 78 lens removal in, 46
Membranotomy, 90 infusion options in, 79 objectives of, 47–48, 48f
Metamorphopsia lens capsule removal in, 79 pupillary dilation in, 45–46
in preoperative assessment, 4 phacoemulsification with, 79 retinopexy, 73
Microscope sulcus in, 80 scleral buckling in, 74
operating, 29f, 29–30 Patient speculum, 45
placement of, 30 draping of, 31 stabilized surgery in, 54, 55f
prep technique for, 31 strategies of, 45
N Patient movement viscodissection in, 53–54
Needles prevention of, 40–41 visualization in, 45–47
infusion, 28 Patient preparation Post-operative follow-up, 248
Neovascular glaucoma, 245 psychological, 41 Povidone-iodine (Betadine), 31
filtering procedures for, 246 Peak electron avalanche knife, 20 Power scissors and forceps, 54
Neovascularization Pediatric cataracts, 90 Practice surgery, 249
in diabetic retinopathy, 185–186 Pediatric 25-gauge vitrectomy, 229–230 Preoperative assessment
Pediatric traction retinal detachments bright flash electroretinography in, 15
O persistent fetal vasculature, 231–232, 232f in cardiopulmonary disease, 3
Obesity, 3 results of, 235 color discrimination, 4
Obstructive sleep apnea syndrome, 3 Toxocara canis, 232, 233f, 234f, 235 contrast sensitivity, 4
Ocular coherence tomography, 164–165 Penetrating injury in diabetes mellitus, 2–3
Open globe delayed vitrectomy in, 240 entoptic phenomenon, 4
anesthesia in, 43 immediate action in, 240 indirect ophthalmoscopy in, 6
Open-angle glaucoma, 246 office injections, 239 laser interferometry, 4
 Index 257

Maddox rod, 4 Retinal breaks/detachment, 8 Retinotomy

metamorphopsia, 4
obesity and obstructive sleep apnea
syndrome, 3
projection and field testing, 4
psychophysical testing in, 4
pupil testing, 5
retinal configuration, 7
slit lamp biomicroscopy, 6
slit lamp examination in, 5–6
two-point discrimination, 4
visual acuity, 3–4
visual function testing, 3
vitreous clarity, 7
vitreous configuration, 6–7
Pressure
intraocular
slit lamp preoperative assessment of, 5–6
Presurgical evaluation and imaging, 2–15
Projection and field testing, 4
Proliferative retinopathy
diabetic retinopathy, 173–174
Proliferative vitreoretinopathy
compartmentalization in, 145
and intentional incarceration, 153–154
lens management in, 145
operability, 145
retinal sutures, 153–154
retinectomy for, 150–152, 151f, 152f
retinopexy in, 154, 154f
retinoplasty in, 154
scleral buckling for, 145
silicone oil, 153
surgical sequence in, 145
tacks, 153–154
vitrectomy for
anterior dissection in, 146, 147f
complications of, 154–155
epiretinal membranes in, 146–148, 148f
results of, 154
subretinal proliferation, 148–150,
149f, 150f
viscoelastics as vitreous substitutes, 153
Psychological preparation
for local anesthesia, 41
Psychophysical testing, 4
Pupil
preoperative assessment of, 5
Pupillary block
aphakic, anterior vitrectomy for, 91
Pupillary dilation
in posterior vitrectomy, 45–46
Pupillary membranes
dense membranectomy, 91
membranectomy, 90–91
membranotomy (discission), 90
R vascular development, 225
Radio frequency identification, 33
Real-time ultrasonography, 12
Reattachment experiment, 61–62, 150–151,
151f
Retina
configuration of
preoperative assessment of, 7, 7t
slit lamp biomicroscopy of, 6
Retinal biopsy, 222–223
diabetic retinopathy and, 174, 179–183
epimacular membrane and, 160, 161
laser prophylaxis for, 118–119
management of, 247
nonrhegmatogenous, 247
postoperative endophthalmitis and, 111
prevention of, 126
prophylaxis of, 126
rhegmatogenous, 161, 187
in necrotizing retinitis, 223
suprachoroidal hemorrhage, 243
traction, 226
in intermediate and posterior uveitis,
223–224
pediatric, 231–235
treatment of
evidence-based, 126
vitrectomy for
indications for, 131
medium-term perfluoron for, 135–136,
135f, 136f
silicone oil in, 136
surface tension management for, 135
surgical sequence for, 131, 132f–134f,
134–135
Retinal neovascularization
presurgical evaluation in, 8
for retinopathy of prematurity, 226
Retinal pigment epithelium transplantation,
200–201
Retinectomy, 150–152, 151f, 152f, 183
Retinopathy of prematurity
epiretinal membrane management, 230
25-gauge pediatric vitrectomy
three-port versus two-port
vitrectomy, 230
trocar cannulas, 229–230
without cannulas, 230
infants
anti-vascular endothelial growth factor
therapy, 229
4b management, 228
with inadequate preoperative laser, 229
laser failures, 228
laser treatment for, 227
service, 228–229
stage 5 management, 228
stage 4a management, 227–228
telemedicine in, 229
pathophysiology of
directional vasculogenesis, 225–226
retinal neovascularization, 226
traction retinal detachment, 226
vascular development, 226
vascular endothelial growth factor, 225
scleral buckling, 230
Retinopexy
for diabetic retinopathy, 185
for double penetrating injury, 212
for giant retinal breaks, 141–142, 142f
posterior vitrectomy, 73
in proliferative vitreoretinopathy, 154, 154f
prophylactic, 126
in trauma, 211
Retinoplasty, 154
drainage, 61
Retrolental membranes
intraocular lens, 93
Retroprosthetic membranes, keratoprosthesis,
238, 238f
Rhegmatogenous retinal detachment, 161, 187
in necrotizing retinitis, 223
S
Scissors
power, 35, 35f, 54
segmentation and delamination, 180–181,
180f, 181f
Scleral buckling
anesthesia for, 43
for diabetic retinopathy, 185
for double penetrating injury, 212
in posterior segment techniques, 74
in proliferative vitreoretinopathy, 145, 152
prophylaxis of, 126–130, 127f–129f
with broad encircling buckle, 129f, 130
case selection, 126–127
circumferential explants, 128, 129f
with encircling band, 127
implants versus explants, 127
sponges versus hard silicone, 127, 127f
with vitreous surgery, 127, 128f
retinopathy of prematurity, 230
subretinal fluid drainage, 70–71, 70f, 71f
in trauma, 211
Scleral defects, 130
Scleral incisions, 112–113, 113f
Scleral laceration, 212–213
Sclerotomies
in trauma, 207
wound construction and closure, 115–116
Silicone infusion, 64–65, 65f
Silicone oil, 101, 183
intraocular lens calculation issues, 102
phacoemulsification vitrectomy
procedures, 99
in proliferative vitreoretinopathy, 153
in vitrectomy for retinal detachment, 136
Slit lamp biomicroscopy
preoperative, 6
Slit lamp examination
preoperative, 5–6
Sonification
with ultrasonic fragmentation, 77
Sterilization systems, 31
Steroid glaucoma, 246
Steroids
for endophthalmitis, 219
in posterior vitrectomy, 73–74
Subfoveal membranes
submacular surgery for, 196
Subluxation
lens
during vitrectomy, 83, 84f
Submacular exudates
in diabetic retinopathy, 175
Submacular hemorrhage
displacement of, 195
Submacular surgery
and macular translocation, 196, 196f
for subfoveal membranes, 196
258 Index
Subretinal fluid drainage
air-gas exchange, 62, 63f
direct transscleral needle, 70–71, 70f, 71f
external versus internal, 70
fluid-air exchange and, 59, 60f
giant retinal breaks, 64
indications for, 69–70
interface vitrectomy, 65–68, 66f–68f
interfacial tension agents, 59
internal, 61, 61f
intraocular lens fogging during, 62, 62f
liquid perfluorocarbon techniques, 62–64, 63f
medium-term perfluoron, 64, 64f
postoperative positioning, 69
in reattachment experiment, 61–62
retinotomy for, 61
silicone infusion, 64–65, 65f
silicone removal, 65, 66f
silicone reoperations, 65
soft-tip extrusion cannula, 60f
surface tension management, 58–59, 58f
vitrectomy under air, 68, 69f
Suprachoroidal hemorrhage
acute management, 241
complications of, 241, 242f
incidence of, 241
pathogenesis and prevention, 241
retinal detachment, 243
surgical management of, 243, 243f
vitrectomy, indications for, 242
Surface tension management, 58–59, 58f
for diabetic retinopathy, 183–184
for double penetrating injury, 212
macular holes, 168–169, 168f
postoperative, 142
for retinal breaks/detachment, 135
in trauma, 211
Surgical data, 250
Surgical self-education
companies, interaction with, 250–251
correspondence, 250
group education, 250
outcomes analysis, 249–250
surgical data, 250
Sutures
extrusion of
in vitreoretinal surgery, 248
T lens management in, 224
Tenon’s capsule, 116
Tissue cutting
elongation and, 20
enzyme-assisted approaches, 20–21
high IOP effects, 22
inertial cutting in, 20
infusion cannula placement, 22
infusion fluidics, 21–22, 21f
infusion system technologies, 22
low IOP effects, 22
sharpness and, 20
shear and, 20
surgical fluidics, 21
Tomography
ocular coherence, 164–165
optical coherence, 13–14, 14f, 15f
Toxocara canis
traction retinal detachments, 232, 233f,
234f, 235
Traction retinal detachment
in diabetic retinopathy, 174
pediatric, 231
Transscleral diathermy, 73
Trauma
cellular migration/proliferation in, 205
complications of
corneal opacification, 213–214, 214f
glaucoma, 214
infection, 213
corneal-scleral lacerations
lens removal, timing of, 213
posterior vitrectomy, 213
wound repair, 213
double penetrating injury
initial repair, 212
surgical sequence and techniques, 212
vitrectomy, 212
intraocular foreign bodies and, 206–212
substrates for proliferation, 205
vitrectomy
timing of, 205
Traumatic cataracts, 90
Triamcinolone
intravitreal
venous occlusive diseases, 189
medium-term perfluoron and, 64, 64f
Trocar-cannula system, 103–104, 104f, 105f, 237
retinopathy of prematurity, 229–230
Two-point discrimination
preoperative assessment of, 4
U wound leak issues, 109, 109f, 110f, 111
Ultrasound
diagnostic
A-scan versus B-scan, 12
B-scan, 12
clinical acoustic physics, 9, 9f–11f, 12
examination methods in, 13
gray scale, 13
image archiving and viewing, 13
limitations, 13
real-time, 12
three-dimensional, 12
vector scanning, 13
fragmentation
in anterior vitrectomy, 76–78
Uveitis, 93
vitreoretinal surgery in
diagnostic vitrectomy, 221–222
drug delivery systems, 224
retinal biopsy, 222–223
therapeutic vitrectomy, 223–224
V Vitreoretinal complications
Vacuum cleaning–extrusion, 56–58, 57f
Vacuum systems
for anterior vitrectomy, 77
Vascular endothelial growth factor (VEGF)
antagonists
diabetic macular edema, 173
in retinal vascular development, 225
Vascular occlusion
in vitreoretinal surgery, 247
Venous occlusive diseases
laser therapy for, 190–191
pharmacotherapy for, 189–190
vitrectomy for, 191
Video recording
intraoperative, 19
Viscodissection
in diabetic retinopathy, 181–182
in posterior vitrectomy, 53–54
Viscoelastics
in proliferative vitreoretinopathy, 153
Visual acuity testing, 3–4
Visual evoked potential, 15
Visual function testing, 3
Visualization
25-gauge vitrectomy, 109
and illumination, 37–39
posterior vitrectomy, 45–47
Vitrectomy
under air, 68, 69f
cellulose sponge, 85, 86f
for diabetic retinopathy, 174, 176
for double penetrating injury, 212
for endophthalmitis, 218, 219
25-gauge
cannula withdrawal and wound closure, 111
case selection, 103
cutter design options, 106
fluidics, 104–106, 106f
20-gauge combined, 107–108, 107f, 108f
postoperative endophthalmitis and retinal
detachments, 111
tool issues, 106–107, 106f, 107f
trocar-cannula system, 103–104, 104f, 105f
visualization, 109
interface, 65–68, 66f–68f
keratoprosthesis, 237–238
outcomes analysis, 249–250
phacoemulsification, 98–102
phacoemulsification and, 98–102
for retinal detachment, 131–136
subluxated lenses during, 83, 84f
submacular hemorrhage displacement, 195
for suprachoroidal hemorrhage, 242
techniques of
and anterior segment problems, 76–96
in trauma
intraocular foreign bodies, 207
timing of, 205
uveitis
diagnostic, 221–222
therapeutic, 223–224
for venous occlusive diseases, 191
Vitreocorneal touch
anterior vitrectomy for, 91
Vitreomacular interface disorders, 224
Vitreomacular traction syndrome, 157–162
keratoprosthesis, 236–238
Vitreoretinal examination
indirect ophthalmoscopy in, 6
Vitreoretinal surgery
anesthesia and, 42
complications of
cataract, 248
corneal, 244
cyclocryopexy, 245–246
endophthalmitis, 246
follow-up intervals, 248
glaucoma, 245
 Index 259

inflammation, 244–245 hypocellular contraction, 6 20-gauge sutures

intraocular hemorrhage, 247
iris neovascularization, 245
neovascular glaucoma, 246
retinal complications, 246–247
suture extrusion, 248
in uveitis management, 221–224
Vitreoretinal traction
minimization of
in vitrectomy, 24, 24f
Vitreoretinopathy
proliferative. See Proliferative
vitreoretinopathy
Vitreous
clarity of
preoperative assessment of, 7
configuration of
preoperative assessment of, 6–7
in fragmenter
avoidance of, 83, 84f
persistent hyperplastic primary, 231–232,
232f
Vitreous bands, 6
Vitreous cortex
fibrovascular proliferation
in diabetic retinopathy, 187
Vitreous cutters. See Tissue cutting
Vitreous hemorrhage
in diabetic retinopathy, 174
Vitreous infusion suction cutter, 27
Vitreous loss
in cataract surgery, 85–86, 86f, 87f
Vitreous microsurgery. See specific techniques and
indications
W Zonulorhexis
Wound construction and closure
conjunctiva and tenon’s capsule, 116
conjunctival incisions, 112
infusion cannula placement, 113–115,
114f, 115f
scleral incisions, 112–113, 113f
instrument sclerotomies, 115–116
and postoperative medication, 116, 116f
traction sutures, 112
Wound leaks, 47
Wound repair
in corneal-scleral laceration, 213
in trauma, 206
X
Xenon illuminations systems, 33, 33f
Z
in anterior vitrectomy, 83, 83f