ORIGINAL STUDY
Long-term Outcomes and Prognostic Factors for
Trabeculectomy With Mitomycin C in Eyes With Uveitic
Glaucoma: A Retrospective Cohort Study
Keiichiro Iwao, MD, PhD,*w Masaru Inatani, MD, PhD,*z Takahiko Seto, MD, PhD,y
Yuji Takihara, MD, PhD,* Minako Ogata-Iwao, MD,*
Satoshi Okinami, MD, PhD,w and Hidenobu Tanihara, MD, PhD*
Purpose: To elucidate the long-term outcomes and prognostic
factors for trabeculectomy with mitomycin C (MMC) in eyes with
uveitic glaucoma (UG).
Methods: A retrospective, consecutive, comparative cohort study
was conducted with 204 patients who underwent trabeculectomy
with MMC between 1999 and 2008 at 2 Japanese clinical centers.
The study group included 101 eyes with UG and 103 eyes with
primary open-angle glaucoma (POAG). Surgical failure was
dened as intraocular pressure levels of Z21 mm Hg or an addi-
tional glaucoma surgery. Kaplan-Meier survival curves for surgical
failure were compared between UG and POAG eyes, and prog-
nostic factors for surgical failure of trabeculectomy in UG eyes
were analyzed by the Cox proportional hazards model. Secondary
outcome measures included comparisons of the frequency of
additional cataract surgery and other surgical complications after
trabeculectomy between UG and POAG eyes.
Results: The mean follow-up periods ( SD) were 34.7 37.9 and
37.7 34.7 months (median, 24.0 and 27.4 mo) for UG and
POAG, respectively. The subtypes of uveitis were granulomatous
uveitis (n = 20) including sarcoidosis (n = 12), Vogt-Koyanagi-
Harada disease (n = 5) and varicella zoster virus uveitis (n = 3),
Behcet disease (n = 10), Posner-Schlossman syndrome (n = 5), and
other types of UG (n = 12). Fifty-four eyes were diagnosed with
idiopathic UG. The 3-year probabilities of success after trabecu-
lectomy were 71.3% and 89.7% for UG and POAG, respectively
(P = 0.0171). A multivariable model showed that UG eyes with
previous cataract surgery [relative risk (RR) = 2.957, P = 0.0344)]
and granulomatous uveitis (RR = 3.805, P = 0.0106) were asso-
ciated with surgical failure. UG eyes experienced more frequent
cataract surgeries after trabeculectomy than POAG eyes: the 3-year
probabilities of additional cataract surgery of 62.6% and 10.7% for
UG and POAG, respectively (P < 0.0001). There was no signicant
dierence in the frequency of surgical complications such as bleb
Received for publication November 27, 2011; accepted July 3, 2012.
From the *Department of Ophthalmology and Visual Science,
Kumamoto University Graduate School of Medical Sciences, Kuma-
moto; wDepartment of Ophthalmology, Faculty of Medicine, Saga
University, Saga; zDepartment of Ophthalmology, Faculty of Medical
Science, University of Fukui, Fukui; and yDepartment of Oph-
thalmology, Juntendo University School of Medicine, Tokyo, Japan.
Disclosure: Supported by KAKENHI Young Scientists grant (S)
19679008 from the Ministry of Education, Culture, Sports, Science,
and Technology of Japan (M.I.), and KAKENHI Young Scientists
Grant (B) 22791668 from the Ministry of Education, Culture,
Sports, Science, and Technology of Japan (K.I.). The authors
declare no conict of interest.
Reprints: Masaru Inatani, MD, PhD, Department of Ophthalmology,
Faculty of Medical Science, University of Fukui, 23-3 Shimoaizuki,
Matsuoka, Eiheiji, Yoshida, Fukui 910-1193, Japan (e-mail: inatani@
u-fukui.ac.jp).
Copyright r 2012 by Lippincott Williams & Wilkins
DOI: 10.1097/IJG.0b013e3182685167
88 | www.glaucomajournal.com
leakage, hypotensive maculopathy, severe anterior-chamber hem-
orrhage, and infectious endophthalmitis.
Conclusions: Trabeculectomy with MMC was less eective in
maintaining intraocular pressure reduction in UG eyes than in
POAG eyes. The prognostic factors for surgical failure of trabe-
culectomy in UG eyes were previous cataract surgery and gran-
ulomatous uveitis. In addition, UG eyes after trabeculectomy more
frequently required additional cataract surgery.
Key Words: ltering surgery, varicella zoster virus uveitis, rheu-
matoid arthritis, HLA-B27-positive acute anterior uveitis, Fuchs
heterochromic iridocyclitis, human T-lymphotropic virus type
1-associated uveitis
(J Glaucoma 2014;23:8894)
U veitic glaucoma (UG) is one of the main causes of
blindness in patients with uveitis because of the develop-
ment of vision-threatening high intraocular pressure (IOP),
which is often resistant to IOP-lowering treatments such as
antiglaucoma medication and ltering surgeries. Although the
surgical success rate of ltering surgery is lower for UG eyes
than for non-UG eyes, several case series suggest that trabe-
culectomy with antibrotic agents such as mitomycin C
(MMC) or 5-uorouracil (5-FU) results in eective IOP
reduction in eyes with UG.15 However, these studies eval-
uated the surgical prognosis of only 53 eyes at most, despite
the wide range of uveitis subtypes and patient backgrounds.
There is currently insucient evidence about which UG sub-
types or patient backgrounds are more resistant to trabecu-
lectomy. Another major problem after trabeculectomy for UG
is secondary cataract. Although previous clinical studies3,5
reported the high frequency of postoperative cataract pro-
gression in UG eyes (28.3% to 51.6%), non-UG eyes also
often encounter cataract progression after trabeculectomy.610
Whether cataract progresses faster in UG eyes after trabecu-
lectomy rather than in primary open-angle glaucoma (POAG)
eyes after trabeculectomy still remains unknown. The present
study retrospectively compared surgical success and compli-
cations, including cataract progression, after trabeculectomy
with MMC between 101 eyes of 101 UG patients and 103 eyes
of 103 patients with POAG. We also elucidated the risk factors
for surgical failure in UG eyes.
MATERIALS AND METHODS
Patient Selection and Surgical Procedures
We retrospectively reviewed the medical records of
patients with UG who underwent trabeculectomy with
J Glaucoma  Volume 23, Number 2, February 2014
J Glaucoma  Volume 23, Number 2, February 2014
MMC (UG group), and those with POAG who under-
went trabeculectomy (POAG group) between January 1,
1999 and December 31, 2008 at the Kumamoto University
Hospital (Kumamoto, Japan) and Saga University Hos-
pital (Saga, Japan). The study was approved by the Insti-
tutional Review Board of Kumamoto University Hospital.
All study procedures adhered to the 1964 Declaration of
Helsinki.
All 101 UG patients were diagnosed by uveitis spe-
cialists. Eyes that presented with IOP of Z22 mm Hg while
on ocular hypotensive medication before surgery were
included in this study. Exclusion criteria for eyes were as
follows: history of previous glaucoma surgery; intraocular
surgery within the 3 months before trabeculectomy;
combined glaucoma and cataract surgeries; and neo-
vascularization related to uveitis. If both the eyes under-
went glaucoma surgeries, the eye that was treated rst was
investigated.
Trabeculectomy in this study was performed according
to a modication of the technique developed by Cairns,11
and MMC was applied intraoperatively on the scleral ap
and under the conjunctiva.12 Conjunctiva incisions included
limbal-based and fornix-based procedures. After the crea-
tion of a scleral ap, sponges soaked with 0.4 mg/mL MMC
were applied to the posterior surface of the conjunctiva,
Tenon capsule, the adjacent episcleral tissue, and the scleral
ap for 3 to 4 minutes, followed by irrigation with balanced
salt saline. A trabecular block was excised to create a stula
in the anterior chamber, and peripheral iridectomy was
then performed. The scleral ap was closed with 10-0 nylon
sutures, and the conjunctival ap was sutured with 10-0
nylon or 7-0 silk. All patients were required to sign
informed consent forms before surgery.
Main Outcome Measure
The main outcome measure was the probability of sur-
gical success using Kaplan-Meier survival analysis. Before the
data analysis, surgical failure was dened as IOPZ21 mm Hg
with or without antiglaucoma medications, which were veri-
ed at the next visit. IOP data that were obtained using a
Goldmann applanation tonometer were collected from
patients medical records. IOPs recorded up to 2 months after
surgery were not considered because of normal postoperative
IOP uctuations after trabeculectomy. If additional glau-
coma surgery was required, or visual acuity (VA) deteriorated
such that light perception was absent, the eye was regarded as
a surgical failure.
We compared surgical success between the UG group
and the POAG group using Kaplan-Meier survival anal-
ysis. To determine potential risk factors for surgical failure
of UG after trabeculectomy, the following variables were
assessed: sex; age; preoperative IOP; design of conjunctival
incision for trabeculectomy (fornix-based or limbal-based);
previous cataract surgery; inammation area (anterior to
intermediate or panuveitis); peripheral anterior synechiae
index; granulomatous uveitis; preoperative usage of corti-
costeroid administration (ocular instillation and/or systemic
administration); and postoperatively prolonged inamma-
tion. Presence of inammation was graded by slit-lamp bio-
microscopic examinations and scored using a 6-grade scoring
system.13 Postoperative prolonged inammation was dened
as cells that were graded as Z2 + at any examination from
2 to 12 months after trabeculectomy.
r 2012 Lippincott Williams & Wilkins
Trabeculectomy for Uveitic Glaucoma
Secondary Outcome Measures
Kaplan-Meier survival analysis was applied to calculate
the rate of additional cataract surgery after trabeculectomy in
phakic eyes with UG and POAG. Data on other post-
operative complications were also collected from medical
records.
VA
VA for each patient was taken as the best-corrected
visual acuity, and the logarithm of the reciprocal of the
decimal VA was used to approximate the logarithm of the
minimal angle of resolution (logMAR). Eyes without form
vision were classied into one of 4 low-vision categories,
which were assigned decimal equivalents as follows:
counting ngers = 0.00500; hand motions = 0.00250; light
perception = 0.00125; and no light perception = 0.00010.
Statistical Analysis
Data analysis was performed using the JMP version 8
statistical package program (SAS Institute, Cary, NC).
Comparisons of outcomes between the UG group and the
POAG group were statistically analyzed by the log-rank
test. To determine the eects of prognostic factors and
identify the relative risk (RR) of surgical failure, multi-
variable prognostic factor analysis was performed using the
Cox proportional hazards model. Multivariable factors
were selected from variables with a probability value of
P < 0.15 according to the log-rank test for Kaplan-Meier
survival analysis. P < 0.05 was considered statistically sig-
nicant. Data are shown as mean SD.
RESULTS
Patient Characteristics
In total, 101 patients (101 eyes) with UG and 103
patients (103 eyes) with POAG satised the eligibility cri-
teria. All of these patients were Japanese. Seventy-three UG
eyes and 80 POAG eyes were phakic. Table 1 lists the
characteristics of the enrolled patients.
The UG group was signicantly younger (P < 0.001),
had a higher preoperative IOP (P < 0.001), used fewer
antiglaucoma eye drops (P = 0.001), and included more
patients treated with oral carbonic anhydrase inhibitor
(P < 0.001) than the POAG group; these dierences were
also observed in phakic eyes, although the statistical sig-
nicance was P = 0.004 for the latter 2. There was no sig-
nicant dierence in preoperative and nal best-corrected
visual acuities between these 2 groups.
UG diagnosis included sarcoidosis (n = 12), Behcet
disease (n = 10), Vogt-Koyanagi-Harada disease (n = 5),
Posner-Schlossman syndrome (n = 5), varicella zoster virus
uveitis (n = 3), rheumatoid arthritis (n = 3), HLA-B27-
positive acute anterior uveitis (n = 2), Fuchs heterochromic
iridocyclitis (n = 2), human T-lymphotropic virus type 1-
associated uveitis (n = 2), Wegener granulomatosis (n = 1),
psoriatic uveitis (n = 1), and syphilitic uveitis (n = 1). In
addition, 54 patients (53.5%) were diagnosed with idio-
pathic UG that lacked the clinical features characteristic of
any recognized uveitic entities or was not attributed to a
specic systemic disease. Twenty eyes (19.8%) including 12
with sarcoidosis, 5 with Vogt-Koyanagi-Harada disease,
and 3 with varicella zoster virus uveitis were categorized as
granulomatous uveitis.1416
www.glaucomajournal.com | 89
Iwao et al J Glaucoma  Volume 23, Number 2, February 2014

TABLE 1. Patients With UG and POAG

Total Phakic Eye
UG POAG UG POAG
n (%) n (%) n (%) n (%)
(n = 101) (n = 103) P (n = 73) (n = 80) P
Female 38 (37.6) 33 (32.0) 0.402* 24 (32.9) 27 (33.8) 0.909*
Right eye 49 (48.5) 54 (52.4) 0.576* 38 (52.1) 38 (47.5) 0.574*
Age (y), mean SD 56.9 15.0 64.4 13.2 < 0.001w 54.3 15.4 62.8 13.1 < 0.001w
Preoperative IOP (mm Hg), mean SD 33.7 8.6 28.1 7.2 < 0.001w 35.3 9.0 28.1 7.8 < 0.001w
Previous cataract surgery 28 (27.7) 23 (22.3) 0.374*
No. antiglaucomatous drops 2.86 0.84 3.21 0.79 0.001w 2.85 0.84 3.23 0.77 0.004w
Use of oral CAI 67 (66.3) 40 (38.8) < 0.001* 50 (68.5) 32 (40) 0.004*
Preoperative BCVA 0.31 0.59 0.34 0.62 0.710 0.25 0.57 0.27 0.55 0.828
BCVA at nal follow-up 0.46 0.86 0.45 0.83 0.976 0.33 0.75 0.38 0.77 0.636
Postoperative follow-up period 34.7 37.9 37.7 34.7 0.560z 37.8 39.1 41.0 34.6 0.589z
*P-values based on w2 for independence test.
wP-values based on Mann-Whitney U test.
zP-values based on Student unpaired t test.
BCVA indicates best-corrected visual acuity; CAI, carbonic anhydrase inhibitor; IOP, intraocular pressure; POAG, primary open-angle glaucoma; UG,
uveitic glaucoma.

UG Versus POAG so a Cox proportional hazards model was performed

Nineteen eyes in the UG group (18.8%) and 10 in the
POAG group (9.7%) were classed as surgical failures.
Kaplan-Meier survival analyses of the 2 groups are pre-
sented in Figure 1. At the point of 3-year after trabeculec-
tomy, the numbers at risk were 23 in UG group and 43 in
POAG group. Before this time point, 17 and 8 eyes were
judged as failure, and 61 and 52 eyes were censored in UG
group and POAG group, respectively. The UG group was
found to have a signicantly lower cumulative probability
of success than the POAG group (P = 0.0171). The prob-
abilities of success 1, 2, 3, and 5 years after trabeculectomy
in the UG group and POAG group were 89.5% versus
91.5%, 82.6% versus 89.7%, 71.3% versus 89.7%, and
61.7% versus 82.5%, respectively.
Signicant dierences were observed between the
preoperative data of the UG group and the POAG group,
(Table 2). The multivariable model determined that trabe-
culectomy in UG eyes and previous cataract surgery were
independently associated with a worse prognosis compared
with the same procedure in POAG eyes, even after adjust-
ing for confounding factors (RR = 2.362, P = 0.0483 for
UG; and RR = 3.898, P = 0.0011 for previous cataract
surgery).
Prognostic Factors for Failure of Trabeculectomy
With MMC for UG Eyes
Potential prognostic factors inuencing survival time are
listed in Table 3. Univariable analysis showed that female sex
(P = 0.0097), previous cataract surgery (P = 0.0002), and
granulomatous uveitis (P < 0.0001) were signicant prog-
nostic factors for surgical failure. Older age (P = 0.0788) was
a factor at the P < 0.15. Figure 2 shows Kaplan-Meier sur-
vival curves for UG eyes with or without previous cataract
surgery, and with or without granulomatous uveitis. The
Cox proportional hazards model including these variables
revealed that the prognostic factors for surgical failure were
previous cataract surgery (RR = 2.957, P = 0.0344) and
granulomatous uveitis (RR = 3.805, P = 0.0106) (Table 4).

TABLE 2. Cox Proportional Hazards Model Determining

Likelihood of Surgical Failure for Patients With UG and POAG
Who Underwent Trabeculectomy With MMC
Variables RR 95% CI P
Uveitis-related glaucoma 2.362 1.006-5.915 0.0483
Female sex 1.951 0.920-4.206 0.0812
Right eye 0.997 0.450-2.154 0.9939
Age (per year) 1.029 0.997-1.065 0.0752
Preoperative IOP (per mm Hg) 1.016 0.962-1.067 0.5425
Previous cataract surgery 3.898 1.738-8.756 0.0011
No. antiglaucomatous eye drops 0.733 0.489-1.138 0.1614
FIGURE 1. Kaplan-Meier survival curves of surgical failure of Use of oral CAI 1.340 0.615-3.092 0.4674
trabeculectomy with mitomycin C in patients with uveitic glau- CAI indicates carbonic anhydrase inhibitor; CI, condence interval;
coma (UG) (solid line) versus primary open-angle glaucoma IOP, intraocular pressure; MMC, mitomycin C; POAG, primary open-angle
(POAG) (dotted line). UG eyes had a significantly lower cumu- glaucoma; RR, relative risk; UG, uveitic glaucoma.
lative probability of success than POAG eyes (P = 0.0171).

90 | www.glaucomajournal.com r 2012 Lippincott Williams & Wilkins

J Glaucoma  Volume 23, Number 2, February 2014 Trabeculectomy for Uveitic Glaucoma

62.6% versus 10.7%, respectively. The Cox proportional

TABLE 3. Influence of Prognostic Factors on Survival Time of UG hazards model (Table 5) determined that trabeculectomy in
Patients Who Underwent Trabeculectomy With MMC
UG eyes rather than POAG eyes, and trabeculectomy for
80% older patients were independently associated with additional
No. Survival cataract surgery, after adjusting for confounding factors
Patients Time (d) P* [RR = 9.470, P < 0.0001 for UG; and RR = 1.033, P =
Sex 0.0097 0.0136 for age (per year)]. Univariable analysis within the UG
Female 38 297 group revealed no signicant variables for additional cataract
Male 63 873 surgery after trabeculectomy. Fornix-based conjunctival
Age (y) 0.0788 incision for trabeculectomy was the only variable with
Younger than 50 28 > 3821 P < 0.15 (P = 0.1471; Table 6).
50 or older 73 608
Preoperative IOP (mm Hg) 0.6720
< 30 31 766
Postoperative Complications
Z30 70 713 There was no signicant dierence in the frequency of
Preoperative IOP (mm Hg) 0.5276 postoperative complications between the UG and POAG
< 40 79 720 groups. Excess ltration requiring additional sutures
Z40 22 713 occurred in 1 eye of both the UG and the POAG group
Design of conjunctival 0.2477 (1.0%), aqueous leakage requiring additional sutures
incision occurred in 3 eyes (3.0%) and 2 eyes (1.9%), hypotony
Fornix-based 47 377 maculopathy occurred in 3 eyes (3.0%) and 1 eye (1.0%),
Limbal-based 54 766
Previous cataract surgery 0.0002 and postoperative hyphema occurred in 5 eyes (5.0%) and 3
Yes 28 237 eyes (2.9%), respectively. One UG eye (1.0%) developed a
No 73 873 malignant glaucoma and 1 POAG eye (1.0%) developed
Inammation area 0.2725 infectious endophthalmitis.
Anterior to intermediate 79 593
Panuveitis 22 715
PAS index (%) 0.3538 DISCUSSION
< 25 76 730 This study compared the success rates and frequencies
Z25 25 540 of postoperative cataract surgery and other surgical com-
PAS index (%) 0.6335 plications of trabeculectomy between UG and POAG eyes,
< 50 75 730 and evaluated the prognostic factors for surgical failure of
Z50 26 218
trabeculectomy for UG. Trabeculectomy showed a sig-
Granulomatous uveitis <0.0001
Yes 20 279 nicantly lower cumulative probability of surgical success
No 81 >3821 (P = 0.0171), and a signicantly higher cumulative proba-
Preoperative use of steroid eye 0.5694 bility of postoperative cataract surgery (P < 0.0001), in UG
drops patients than in POAG patients. In addition, the Cox
Yes 77 720 proportional hazard model revealed that previous cataract
No 24 713 surgery (RR = 2.957, P = 0.0344) and granulomatous
Preoperative oral 0.7517 uveitis (RR = 3.805, P = 0.0106) were prognostic factors
administration of steroid for surgical failure of trabeculectomy. There was no sig-
Yes 21 713
nicant dierence in the frequency of other surgical com-
No 80 706
Postoperative prolonged 0.3656 plications between the 2 groups.
inammation Several studies have reported the results of glaucoma
Yes 13 377 surgery for UG patients15,17,18; however, there have been no
No 88 720 large-scale examinations of the long-term outcomes of tra-
beculectomy with MMC for UG. Ceballos et al3 reported an
CI indicates condence interval; IOP, intraocular pressure; MMC,
mitomycin C; PAS, peripheral anterior synechiae; UG, uveitic glaucoma.
overall success rate of 78% at 1 year and 62% at 2 years for
*P-values are based on log-rank test. 44 eyes from UG patients who underwent trabeculectomy
with MMC or 5-FU; however, this study lacked a control
group without uveitis. Noble et al4 found that 21 UG eyes
Postoperative Cataract Surgery that underwent trabeculectomy with MMC showed worse
Thirty UG eyes (29.7%) and 8 POAG eyes (7.8%) outcomes, and that UG increased the risk of requiring
required additional cataract surgery after trabeculectomy postoperative therapeutic interventions to stabilize IOP; few
because of postoperative cataract progression, and the patients in this study achieved an IOP reduction of Z30%
Kaplan-Meier survival analysis of these data is shown from baseline. Park et al17 reported that 23 UG eyes showed
in Figure 3. At the point of 3-year after trabeculectomy, the signicantly lower success rates than 43 non-UG eyes that
numbers at risk were 11 in UG group and 38 in POAG underwent phacotrabeculectomy with MMC. Kaburaki et al5
group. Before this time point, 28 and 6 eyes were judged as observed a 64.7% ve-year success rate of IOPr15 mm Hg
failure, and 34 and 36 eyes were censored in UG group and with antiglaucoma medications in 53 UG eyes after trabe-
POAG group, respectively. The UG group had a sig- culectomy compared with 80 POAG eyes (65.9%).
nicantly higher cumulative probability of additional cat- To our knowledge, the present retrospective study is the
aract surgery than the POAG group (P < 0.0001). The largest investigation of UG patients to date. Our multi-
cumulative probabilities 1, 2, 3, and 5 years after trabecu- variable model demonstrates that trabeculectomy with MMC
lectomy in the UG group and POAG group were 21.3% in UG eyes is independently associated with a worse prog-
versus 3.2%, 47.6% versus 10.7%, 62.6% versus 10.7%, and nosis than the same procedure in POAG eyes, even after

r 2012 Lippincott Williams & Wilkins www.glaucomajournal.com | 91

Iwao et al J Glaucoma  Volume 23, Number 2, February 2014

FIGURE 2. Kaplan-Meier survival curves of surgical failure of trabeculectomy with mitomycin C in uveitic glaucoma (UG) eyes with
versus without previous cataract surgery (A), and with versus without granulomatous uveitis (B). UG eyes with previous cataract surgery
(P = 0.0002) and UG eyes with granulomatous uveitis (P < 0.0001) had a significantly lower cumulative probability of success.

adjusting for confounding factors. However, previous case Previous cataract surgery in UG eyes is another prog-
studies suggest that trabeculectomy with antibrotic agents nostic factor for surgical failure. Several previous studies
enhances the postoperative IOP reduction in UG eyes. reported that earlier surgical scarring in the conjunctiva is a
The current study reveals that UG eyes with gran- potential risk factor for bleb failure.2226 In most cases in
ulomatous uveitis (RR = 3.805, P = 0.0106) have a greater glaucoma with or without uveitis, phacoemulsication by
risk of surgical failure after trabeculectomy with MMC. clear corneal incision should be selected to avoid making
Although no studies have yet investigated which subtypes of unnecessary conjunctival scar. However, prior conjunctival
uveitis are more resistant to trabeculectomy, the reported sur- damage develops into a major problem in the case of glau-
gical success rates for UG with Behcet disease, which is a major coma and/or uveitis that occur after surgery. In the present
nongranulomatous uveitis, are 82.6% to 83.3% at 1 year19,20 study, all pseudophakic eyes had been treated with phacoe-
and 76.2% at 2 years after trabeculectomy with MMC.20 These mulsication, which results in minimal invasion to the
rates are comparable with the surgical success of UG eyes with
nongranulomatous uveitis in the present study (94.4% and
85.6% at 1 and 2 y after surgery, respectively). By contrast, UG
eyes with granulomatous uveitis showed a signicantly worse
prognosis for IOP control in our study (72.4% and 54.3% at 1
and 2 y after surgery, respectively). To understand this nding,
it is necessary to perform histologic examinations of UG eyes
with granulomatous uveitis. Hamanaka et al21 reported that
such eyes accumulate brotic tissue and granuloma containing
Langhans giant cells in the trabecular meshwork and Schlemm
canal. This continuous granuloma formation in the anterior
chamber might obstruct the ltering pathway that is created by
trabeculectomy.

TABLE 4. Cox Proportional Hazards Model Determining

Likelihood of Surgical Failure of Trabeculectomy With MMC in All
Patients With UG
Variables RR 95% CI P
Female 1.753 0.646-5.037 0.2717
Age (per year) 1.016 0.976-1.061 0.4486
FIGURE 3. Cumulative probabilities of additional cataract sur-
Previous cataract surgery 2.957 1.085-8.101 0.0344
gery after trabeculectomy with mitomycin C in patients with
Granulomatous uveitis 3.805 1.368-11.025 0.0106
uveitic glaucoma (UG) (solid line) versus primary open-angle
CI indicates condence interval; MMC, mitomycin C; RR, relative risk; glaucoma (POAG) (dotted line). UG eyes had a significantly
UG, uveitic glaucoma. higher cumulative probability of additional cataract surgery than
POAG eyes (P < 0.0001).

92 | www.glaucomajournal.com r 2012 Lippincott Williams & Wilkins

J Glaucoma  Volume 23, Number 2, February 2014 Trabeculectomy for Uveitic Glaucoma

conjunctiva. Notably, we recently showed that open-angle

TABLE 5. Cox Proportional Hazards Model Determining glaucoma eyes that had undergone phacoemulsication had
Likelihood of Postoperative Cataract Surgery for Patients With UG
and POAG Who Underwent Trabeculectomy With MMC lower surgical success rates of trabeculectomy with MMC
than phakic eyes with open-angle glaucoma,26 which is con-
Variables RR 95% CI P sistent with our present data. Conjunctival incision during
UG 9.470 3.916-25.123 <0.0001 phacoemulsication might recruit broblasts and inamma-
Female sex 0.831 0.388-1.699 0.6167 tory cells to the conjunctiva distant from surgically scarred
Right eye 1.080 0.555-2.110 0.8209 regions, as reported by Broadway et al.25 An alternative
Age (per year) 1.033 1.007-1.063 0.0136 explanation is that breakdown of the blood-aqueous barrier
Preoperative IOP (per mm Hg) 1.019 0.971-1.065 0.4290 in pseudophakic eyes increases the chemoattractant activity of
No. antiglaucomatous eye drops 1.345 0.854-2.145 0.2031 broblasts in the aqueous humor, which results in the devel-
Use of oral CAI 0.810 0.392-1.757 0.5835
opment of subconjunctival brosis after trabeculectomy.27
CAI indicates carbonic anhydrase inhibitor; CI, condence interval; Third, prior development of cataract requiring surgery may
IOP, intraocular pressure; POAG, primary open-angle glaucoma; RR, rel- be a maker of an eye with worse uveitis, which would have a
ative risk; UG, uveitic glaucoma. poorer outcome whether or not it underwent prior cataract
surgery.
Many more UG eyes than POAG eyes required post-
operative cataract surgery in the present study, despite the
signicantly younger average age of the UG group. These
data are consistent with previous studies of UG eyes treated
TABLE 6. Influence of Prognostic Factors on Survival Time Before with trabeculectomy. Ceballos et al3 reported that 16
Additional Cataract Surgery in UG Patients Who Underwent (51.6%) of 31 phakic UG eyes were treated with phacoe-
Trabeculectomy With MMC mulsication after trabeculectomy with MMC or 5-FU.
No. 80% Survival
Patitsas et al28 found that cataract progression after lter-
Patients Time (d) P*
ing surgery occurred in 9 of 10 phakic UG eyes, 7 of which
were treated with cataract surgery.
Sex 0.7312 In addition to the evidence that trabeculectomy per se
Female 24 372 is a risk factor for cataract progression,610 chronic ocular
Male 49 245 inammation and continuous corticosteroid administration
Age (y) 0.2084
Younger than 50 25 468
in UG patients cause cataract development to progress
50 or older 48 351 more rapidly.29,30 Phakic UG eyes therefore require addi-
Preoperative IOP (mm Hg) 0.3532 tional phacoemulsication more frequently than phakic
< 30 19 213 POAG eyes. For the management of coexisting cataracts in
Z30 54 447 UG eyes, Park et al17 retrospectively reported on the sur-
Preoperative IOP (mm Hg) 0.9237 gical outcomes of phacotrabeculectomy with MMC in 23
< 40 57 279 phakic UG eyes including 1 with sarcoidosis and 1 with
Z40 16 447 Vogt-Koyanagi-Harada disease, which are recognized as
Design of conjunctival incision 0.1471 granulomatous uveitis. The surgical procedure oered VA
Fornix-based 29 231
Limbal-based 44 317
improvement and favorable IOP reduction in phakic UG
Inammation area 0.2050 eyes. Combined surgery might therefore be an option for
Anterior to intermediate 59 447 the management of coexisting cataracts in UG eyes.
Panuveitis 14 201 Univariable analyses in previous studies reported
PAS index (%) 0.3433 black ethnicity,2 male sex,3 and postoperative inamma-
< 25 54 245 tion5 as risk factors for IOP reduction after trabeculectomy
Z25 19 279 in UG patients. However, our multivariable analysis of
PAS index (%) 0.4483 Japanese UG patients did not detect these risk factors.
< 50 61 372 Instead, we demonstrated lower success rates in female UG
Z50 12 162
Granulomatous uveitis 0.8586
patients, although sex was not a signicant prognostic
Yes 13 441 factor in the multivariable analysis. This could be because
No 60 231 women are more susceptible to sarcoidosis, and 8 of the 12
Preoperative use of steroid eye 0.8191 sarcoidosis-induced UG patients were female in the present
drops study. UG eyes with postoperative prolonged inammation
Yes 52 351 showed a lower success rate in the univariable analysis,
No 21 213 although it was not statistically signicant according to the
Preoperative oral 0.3147 log-rank test. This could be because we retrospectively
administration of steroid quantied the extent of inammation in the anterior
Yes 14 215
No 59 351
chamber based on medical charts.
Postoperative prolonged 0.7062 This study had some limitations caused by the retro-
inammation spective design. First, the selection bias of surgical oppor-
Yes 9 89 tunity for UG might have aected the surgical result. In this
No 63 351 study, the number of patients with UG and POAG
underwent trabeculectomy were similar, although UG
IOP indicates intraocular pressure; MMC, mitomycin C; PAS, peri-
pheral anterior synechiae; UG, uveitic glaucoma.
occurs less often than POAG. As 2 hospital involved in
*P-values are based on log-rank test. present study has specialized outpatient clinic for severe
uveitis patient, this could lead to large number of patients

r 2012 Lippincott Williams & Wilkins www.glaucomajournal.com | 93

Iwao et al
requiring surgical treatment subsequently. However, it may
be undeniable that there is a selection bias. Second, because
of a certain number of censored cases, comparisons in the
long-term result between UG and POAG group require
appropriate caution. In fact, there was only 23 UG patients
compared with 43 POAG patients after 3 years.
In conclusion, the present study demonstrated that
trabeculectomy with MMC was less eective in maintaining
IOP reduction in UG eyes than in POAG eyes. Gran-
ulomatous uveitis and pseudophakia were more susceptible
to surgical failures of trabeculectomy with MMC, and UG
eyes more frequently required additional cataract surgery
after trabeculectomy with MMC.
REFERENCES
1. Jampel HD, Jabs DA, Quigley HA. Trabeculectomy with 5-
fluorouracil for adult inflammatory glaucoma. Am J Oph-
thalmol. 1990;109:168173.
2. Towler HM, McCluskey P, Shaer B, et al. Long-term follow-
up of trabeculectomy with intraoperative 5-fluorouracil for
uveitis-related glaucoma. Ophthalmology. 2000;107:18221828.
3. Ceballos EM, Beck AD, Lynn MJ. Trabeculectomy with
antiproliferative agents in uveitic glaucoma. J Glaucoma.
2002;11:189196.
4. Noble J, Derzko-Dzulynsky L, Rabinovitch T, et al. Outcome
of trabeculectomy with intraoperative mitomycin C for uveitic
glaucoma. Can J Ophthalmol. 2007;42:8994.
5. Kaburaki T, Koshino T, Kawashima H, et al. Initial trabecu-
lectomy with mitomycin C in eyes with uveitic glaucoma with
inactive uveitis. Eye (Lond). 2009;23:15091517.
6. Popovic V, Sjostrand J. Course of exfoliation and simplex
glaucoma after primary trabeculectomy. Br J Ophthalmol.
1999;83:305310.
7. Daugeliene L, Yamamoto T, Kitazawa Y. Cataract develop-
ment after trabeculectomy with mitomycin C: a 1-year study.
Jpn J Ophthalmol. 2000;44:5257.
8. AGIS (Advanced Glaucoma Intervention Study) Investigators.
The advanced glaucoma intervention study: 8. Risk of cataract
formation after trabeculectomy. Arch Ophthalmol. 2001;119:
17711779.
9. Lazaro C, Benitez-del-Castillo JM, Castillo A, et al. Lens
fluorophotometry after trabeculectomy in primary open-angle
glaucoma. Ophthalmology. 2002;109:7679.
10. Husain R, Aung T, Gazzard G, et al. Effect of trabeculectomy
on lens opacities in an east Asian population. Arch Ophthalmol.
2006;124:787792.
11. Cairns JE. Trabeculectomy. Preliminary report of a new
method. Am J Ophthalmol. 1968;66:673679.
12. Kitazawa Y, Kawase K, Matsushita H, et al. Trabeculectomy
with mitomycin. A comparative study with fluorouracil. Arch
Ophthalmol. 1991;109:16931698.
13. Nussenblatt RB, Whitcup S, Palestine AG. Examination of the
patients with uveitis. In: Nussenblatt RB, Whitcup SM,
94 | www.glaucomajournal.com
J Glaucoma  Volume 23, Number 2, February 2014
Palestine AG, eds. Uveitis Fundamental and Clinical Practice.
St Louis: Mosby; 1996:5868.
14. Moorthy RS, Inomata H, Rao NA. Vogt-koyanagi-harada
syndrome. Surv Ophthalmol. 1995;39:265292.
15. Fischer RE. Granulomatous uveitis occurring in identical twins
with sarcoidosis; report of two cases. Portland Clin Bull. 1957;
11:4648.
16. Hedges TR III, Albert DM. The progression of the ocular
abnormalities of herpes zoster. Histopathologic observations
of nine cases. Ophthalmology. 1982;89:165177.
17. Park UC, Ahn JK, Park KH, et al. Phacotrabeculectomy with
mitomycin C in patients with uveitis. Am J Ophthalmol.
2006;142:10051012.
18. Stavrou P, Murray PI. Long-term follow-up of trabeculectomy
without antimetabolites in patients with uveitis. Am J
Ophthalmol. 1999;128:434439.
19. Elgin U, Berker N, Batman A, et al. Trabeculectomy with
mitomycin C in secondary glaucoma associated with behcet
disease. J Glaucoma. 2007;16:6872.
20. Yalvac IS, Sungur G, Turhan E, et al. Trabeculectomy with
mitomycin-c in uveitic glaucoma associated with behcet
disease. J Glaucoma. 2004;13:450453.
21. Hamanaka T, Takei A, Takemura T, et al. Pathological study
of cases with secondary open-angle glaucoma due to sarcoi-
dosis. Am J Ophthalmol. 2002;134:1726.
22. Mandal AK, Bagga H, Nutheti R, et al. Trabeculectomy with
or without mitomycin-c for paediatric glaucoma in aphakia
and pseudophakia following congenital cataract surgery. Eye
(Lond). 2003;17:5362.
23. Fontana H, Nouri-Mahdavi K, Caprioli J. Trabeculectomy
with mitomycin C in pseudophakic patients with open-angle
glaucoma: outcomes and risk factors for failure. Am J
Ophthalmol. 2006;141:652659.
24. Prata Junior JA, Minckler DS, Baerveldt G, et al. Trabecu-
lectomy in pseudophakic patients: postoperative 5-fluorouracil
versus intraoperative mitomycin C antiproliferative therapy.
Ophthalmic Surg. 1995;26:7377.
25. Broadway DC, Grierson I, Hitchings RA. Local effects of pre-
vious conjunctival incisional surgery and the subsequent outcome
of filtration surgery. Am J Ophthalmol. 1998;125:805818.
26. Takihara Y, Inatani M, Seto T, et al. Trabeculectomy with
mitomycin for open-angle glaucoma in phakic vs pseudophakic
eyes after phacoemulsification. Arch Ophthalmol. 2011;129:
152157.
27. Joseph JP, Grierson I, Hitchings RA. Chemotactic activity of
aqueous humor. A cause of failure of trabeculectomies? Arch
Ophthalmol. 1989;107:6974.
28. Patitsas CJ, Rockwood EJ, Meisler DM, et al. Glaucoma
filtering surgery with postoperative 5-fluorouracil in patients
with intraocular inflammatory disease. Ophthalmology. 1992;
99:594599.
29. Belfort R Jr, Nussenblatt RB. Surgical approaches to uveitis.
Int Ophthalmol Clin. 1990;30:314317.
30. Foster CS, Fong LP, Singh G. Cataract surgery and intra-
ocular lens implantation in patients with uveitis. Ophthalmo-
logy. 1989;96:281288.
r 2012 Lippincott Williams & Wilkins