Вы находитесь на странице: 1из 5

British Journal of Anaesthesia, 116 (2): 27781 (2016)

doi: 10.1093/bja/aev432
Regional Anaesthesia


Epidural distribution of dye administered via

an epidural catheter in a porcine model

Downloaded from http://bja.oxfordjournals.org/ at Gadjah Mada University on November 22, 2016

I. Mowat1, *, R. Tang1, H. Vaghadia1, C. Krebs3, W. R. Henderson2 and A. Sawka1
Department of Anesthesiology & Perioperative Care, 2Department of Critical Care, Vancouver Coastal Health,
899 West 12th Avenue, Vancouver, Canada V5Z 1M9, and 3Department of Anatomy, University of British
Columbia, 2350 Health Sciences Mall, Vancouver, Canada V6 T 1Z3
*Corresponding author. E-mail: imowat@doctors.org.uk

Background: Local anaesthetics are commonly delivered to the epidural space by either intermittent bolus or continuous
infusion. While these methods have been investigated in terms of analgesia and total dose administered, they have not been
compared in terms of their effect on the spread of injectate within the epidural space. This animal study compared the spread of
dye delivered to the epidural space in a porcine model by either bolus or infusion.
Methods: After ethical approval, epidural catheters were placed at three vertebral levels in seven anaesthetized pigs. Aqueous
dye (1 ml) was injected into the catheter as a bolus, or as an infusion over 30 min. Animals were euthanized at the end of the
study and necropsy performed immediately to quantify the extent of dye spread.
Results: In seven animals, 20 catheters were successfully placed in the epidural space. The mean () extent of dye spread was
8.9 (2.6) cm in the infusion group compared with 15.2 (2.7) cm in the bolus group (P<0.001). Segmental spread was signicantly
greater in the bolus group compared with the infusion group (P<0.01).
Conclusions: In the porcine epidural model, spread of one ml of epidural dye solution is more extensive after a single bolus
compared with short term infusion.

Key words: analgesia, epidural; anesthesia, epidural; epidural space; injections, epidural

Editors key points Epidural anaesthesia and analgesia, rst described in 1921, are
commonly used techniques in the management of pain.1 Despite
The effects of delivery of local anaesthetics through epi-
a long history of safety and efcacy, spread of injected local
dural catheters by bolus or infusion techniques on epidural
anaesthetics is difcult to predict,2 and is inuenced by several
spread, were compared in an anaesthetized pig model.
factors3 including dose4 and volume57 of the injectate. The
Bolus injection of dye led to greater rostrocaudal extent of
extent of epidural local anaesthetic spread has been shown to
circumferential spread compared with infusion of the
correlate to the level of sensory blockade obtained.8 However,
same volume.
inadequate analgesia remains an important problem.9 10 Local
These ndings are consistent with greater efcacy and
anaesthetic agents can be delivered to the epidural space by ei-
lower dose requirements for bolus epidural injections of
ther an intermittent bolus, or as an infusion.11 These methods
local anaesthetics.
have been investigated in terms of effect on analgesia and total

Accepted: November 1, 2015

The Author 2016. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.
For Permissions, please email: journals.permissions@oup.com

278 | Mowat et al.

dose administered,11 but have not been compared in terms of protection for the epidural catheters during dissection and
neuraxial spread. This information could guide clinicians as to were fastened to the skin with adhesive tape. All epidural place-
the most effective method of delivery in terms of spread and ments were performed by a single member (IM) of the research
allow for dose reduction of local anaesthetics. team. If it was not possible to enter the epidural space at a chosen
Porcine spinal anatomy has been reported as comparable level, an attempt was made at the immediate cranial or caudal
with that of humans,12 13 and porcine models have been used intervertebral space. Epidural pressure readings were obtained
in studies of the epidural space,14 catheter placement,15 and epi- from each epidural catheter using an electronic transducer
dural anaesthesia.16 17 The primary objective of this study was to (Transpac IV, ICU Medical, Inc. San Clemente, CA, USA).
determine if the delivery of dye to the epidural space, by either Each catheter was either attached to a syringe driver or
bolus or infusion, signicantly inuences spread. We hypothe- syringe to administer blue or green dye via infusion or bolus, re-
sized that delivery of a xed volume of dye via a single bolus spectively. The blue dye was a water-based solution of propyl-
would result in greater spread compared with an infusion. ene glycol and triphenylmethane dye, and the green dye used
was a similar solution with tartrazine. Both infusions and bo-
luses were administered using a 10 ml syringe delivering one
Methods ml of dye. Green or blue dye was alternated between each proto-
col. We used one ml of dye in keeping with previous studies7 12

Downloaded from http://bja.oxfordjournals.org/ at Gadjah Mada University on November 22, 2016

This study was approved by the Animal Research Ethics Board
of the University of British Columbia on April 4, 2014, and regis- in order to allow discrete spread of dye to be studied without
tered with the Ofce of Research Services of the University of spread at one level encroaching onto spread at another level;
British Columbia (A12-0230). The study protocol adhered to the in addition, we surmised that as the cross-sectional area of
Canadian Council on Animal Care guidelines for humane animal the porcine spinal cord and vertebral canal is approximately
use. These guidelines are comparable to the regulations set out in one third that of the human spinal cord, this volume would
the Animal (Scientic Procedures) Act 1986 and the guidance pro- be comparable with a top up dose in humans of X ml 3 ml. 12
vided by the National Centre for the Replacement, Renement The infusion was delivered over 30 min with a Next Advance
and Reduction of Animals in Research. Design, analysis and (Averill Park NY, USA) SP-300 infusion pump, and the bolus
reporting of this study conforms with ARRIVE Guidelines. was injected manually as a fast push over one s. Injection
pressures were measured using a three-way stopcock and elec-
tronic transducer attached between the syringe and catheter.
Anaesthesia Each pig received either two boluses and one infusion proto-
A convenience sample of seven healthy Yorkshire cross breed col, or two infusions and one bolus protocol in alternate fashion.
female pigs (Abbotsford, British Columbia, Canada) was used. For example, infusion protocol via both lumbar and mid-thoracic
Animals were fasted overnight but received water ad libitum. catheters, and a bolus protocol via the low-thoracic catheter.
After sedation with 20 mg kg1 i.m. ketamine, inhalation anaes- Physiological data recorded included arterial pressure, heart
thesia was induced with 45% isourane in oxygen. After tracheal rate, pulse oximetry, end-tidal carbon dioxide, minute ventila-
intubation and establishment of i.v. access, total i.v. anaesthesia tion, tidal volume and peak airway pressure. Epidural pressure
was established with an infusion of 0.40.7 mg kg1 midazolam measurements were repeated immediately after dye delivery.
and 50150 mcg kg1 min1 of propofol. Subsequently, inhalation All animals were euthanized at the end of the experiment
anaesthesia was discontinued, physiologic monitors (ECG, oxygen using an i.v. bolus of pentobarbital sodium (Bimeda-MTC, Cam-
saturation, bp) were established and the animals were placed bridge, Ontario, Canada) 120 mg kg1. Death was conrmed by
prone. Ventilation and anaesthetic gases were delivered using a the absence of cardiac electrical activity on continuous surface
Drager (Luebeck, Germany) AV 2824 anaesthetic machine. Physio- electrocardiography. An anatomist (CK) assisted by one investi-
logical monitoring was displayed by an HP (Palo Alto, CA USA) gator (IM) performed necropsy, with attention to the spine and
78354A X monitor. Capnography was provided using a Datex- spinal cord to assess dye spread and conrm level of epidural
Ohmeda (Madison, Wisconsin, USA) Capnomac Ultima capnograph. needles. Necropsy was completed within one h of the end of
the dye injection to minimize tissue changes associated with
death.7 The longitudinal extent of circumferential dye spread
was measured both in centimeters and in number of spinal
The spinous processes were identied and marked at each level. nerve levels dyed bilaterally. Pattern or patchiness of spread
A total of three epidural catheters were placed in each pig at lum- was documented in descriptive terms. Spine length was mea-
bar (L34), low thoracic (T1011) and mid-thoracic levels (T56). sured in centimeters for each animal.
This allowed the necessary data to be collected, whilst minimiz-
ing the number of pigs required.
Epidural catheters were placed using a paramedian approach
to the epidural space. For lumbar and low-thoracic levels, a 17G, Data were entered into a spreadsheet program (Microsoft
90 mm Touhy needle (Braun Medical Inc, Bethlehem, PA, USA) Excel, Microsoft Corporation, Seattle, WA, USA) and analysed
was used. For deeper, mid-thoracic levels, an 18G, 150 mm by an independent statistician using Statistical Package for
Touhy needle was used. The epidural space was identied the Social Sciences (SPSS, Version 20.0, IBM Corporation,
using a loss-of-resistance to air technique with a glass syringe, Armonk, NY, USA). Tests for normal distribution were per-
with two-three ml of air injected at each level. At each level, a formed using the Shapiro-Wilk test of normality for small
Perix (Braun Medical Inc) epidural catheter was advanced sample sizes. Normally distributed data are presented as
through the epidural needle so that two cm of catheter protruded mean (); non-normally distributed data are presented as me-
into the epidural space. This distance was chosen to ensure that dian (interquartile range). Categorical data were compared
each port of the catheter would be in the epidural space whilst using a 2test analysis. Comparisons of ordinal data were per-
minimizing the risk of the catheter coiling or encroaching upon formed using the Mann-Whitney U-test. The level of signi-
the other catheters. Epidural needles were left in situ to provide cance was set at P<0.05.
Epidural dye spread: bolus vs infusion | 279

Results catheters, nor was any evidence of puncture or subarachnoid

catheter placement seen on dissection. The single catheter
We studied seven pigs with a median weight of 53.6 (IQR 5154) kg
found not to be in the epidural space was identied in the para-
and median spine length of 73 (IQR 7174) cm. Biometric data for
spinal muscles. Distribution of catheter placement with regard to
both groups are summarized in Table 1. Of the 21 catheters at-
vertebral level is shown in Table 2.
tempted, 20 were successfully placed in the epidural space. In
Dye was readily visible and contrasted well with adjacent tis-
all animals, it was possible to advance the catheter past the dis-
sue. Necropsy revealed discrete sections of dye spread between
tance to the tip of the Tuohy needle. Anatomical conrmation of
each injection that allowed accurate measurement of spread
epidural needle level and catheter location was possible in all
(Fig. 1). Mean injection pressures in the bolus protocol were 314
cases during necropsy. Dural puncture, suggested by cerebro-
(12.5) mm Hg which were signicantly greater than in the infu-
spinal uid leak, was not seen during the placement of the
sion protocol (24 (4.9) mm Hg, (P<0.001). Mean longitudinal extent
of circumferential dye spread in the bolus protocol was 15.2 (2.7)
cm, or 5.5 (56). This was signicantly greater than the longitu-
Table 1 Biometric data for pigs dinal extent of circumferential dye spread in the infusion proto-
col, mean (8.9 (2.6 cm), or 3.1 (34 levels, (P<0.001)). Caudad spread
Infusion group Bolus group (n=10) in the bolus protocol was two (one to two) levels compared with

Downloaded from http://bja.oxfordjournals.org/ at Gadjah Mada University on November 22, 2016

(n=10) one level in the infusion protocol (P=0.03). Cephalad spread was
Median IQR Median IQR four (three to ve) levels in the bolus protocol compared with
two (one to three) levels in the infusion protocol (P=0.029). Figure 2
Weight (kg) 53.6 5154 (3) 53.7 51.454 (2.6)
shows the injection pressures and longitudinal extent of circum-
Spine length (cm) 73 7174 (3) 73 7274 (2)
ferential spread for each correctly placed catheter. Spinal pos-
Depth to epidural 6 5.57.5 (2) 6 58 (3)
ition of each catheter is also shown. No signicant correlation
space (cm)
between spinal position and extent of spread was found. There
was no signicant difference in the recorded physiological data
between groups Table 3.
Dye within the epidural space showed no evidence of crystal
Table 2 Epidural catheter placement levels in infusion and bolus formation or distortion of soft tissue structures. A clear line of de-
groups. 2 value for comparison of catheter placement level=0.286, marcation could be seen at the dye border, to suggest that dye
P value (two-sided)=0.867 distribution was primarily by bulk ow rather than by diffusion.
Seven of the injections demonstrated a combination of longitu-
Vertebral level Infusion group (n) Bolus group (n)
dinal circumferential and non-circumferential spread from the
Mid-thoracic 3 3 catheter site. For each of these, circumferential spread began
Low-thoracic 3 4 nearest the catheter tip, to then extend as a patchy, non-circum-
Lumbar 4 3 ferential spread. This extension of spread was variable so that
Total 10 10 more spinal nerves were dyed on one side compared with the
other. No anatomical barriers were noted that could explain the

Fig 1 Photograph of porcine spinal cord still within the vertebral canal after necropsy. Discrete sections of dyed and non-dyed spinal cord allowed for accurate
measurement of spread at three levels. Circumferential and unilateral spread can be seen.
280 | Mowat et al.

unequal distribution of dye. This study did not yield sufcient used an animal model to negate the risk of potential neurological
data to determine whether this spread pattern was more com- injury posed to humans by a study of this type and to allow full
mon in bolus or infusion protocols, but it was seen in both. necroscopic examination. Additionally, we used anaesthetised
animals to ensure compliance with local and national animal
care guidelines, and to provide optimal conditions for epidural
Discussion placement. Necroscopy examination of the spinal cord was cho-
sen because it provides a robust assessment of the spinal cord,
Delivery of a xed volume of dye to the epidural space by single
spinal nerves and extent of dye spread, and has been previously
bolus produced a greater longitudinal extent of circumferential
validated in this animal model.7 19
spread, compared with delivery of the same volume via an infu-
Access to the epidural space was easily achieved in all ani-
sion in a porcine model. Cephalad and caudad spread of dye was
mals on rst attempt. No signicant difculty was experienced
more extensive after bolus administration compared with an in-
in advancing the epidural catheter and other manoeuvres to as-
fusion protocol.
sist catheter advancement were not required.19 Unlike Gamble
The porcine model was chosen because of its close anatomic-
and colleagues,19 we did not experience difculty advancing
al similarity to the human spine.12 13 Epidural placement in pigs
the catheter when it was just beyond the tip of the epidural nee-
is comparable with that used in human clinical practice.18 We
dle. The dye used in this study is commonly used in our anatomy

Downloaded from http://bja.oxfordjournals.org/ at Gadjah Mada University on November 22, 2016

laboratory for tissue marking, is easily injected through an epi-
dural catheter, does not distort tissue or crystallize, and can be vi-
20 sualized easily. Hogan reported that dye spread in the epidural
space does not occur as a universal front, but rather as rivulets
Longitudinal extent of circumferential dye spread

18 through numerous small channels,20 and that solution preferen-

tially travels along the nerve root sheath.20 Our ndings support
these observations in that the margins delineating dye spread did
not appear uniform and were irregular akin to rivulets. Like
Hogan, we also found that dye spread occurred along natural tis-
12 sue planes in a non-uniform manner. The pattern we observed of
circumferential spread extending as patchy, non-circumferential

10 spread could explain the asymmetric difference in dermatomal

sensory block, sometimes seen when assessing epidural anaes-
8 thesia on each side of the body.
The applicability of our results to humans is limited by the ob-
6 vious anatomical differences between pigs and humans. Of rele-
vance in this study is the smaller volume of the porcine spinal
4 canal.12 This, and the specic uid properties of the dye, could
explain why a spread of 15 cm was achievable with only one
2 ml, or approximately 0.02 ml kg1, of dye injected. Other anatom-
ical differences that have been described include the greater thor-
0 acic and lumbar curvatures seen in humans.12 Thus extent of dye
0 50 100 150 200 250 300 350
Injection pressures (mmHg) spread in our model might not reect spread in humans, but
basic principles might still be relevant. The increase in spread
seen in the bolus protocol could result from the greater injection
Fig 2 Scatter plot of injection pressures against longitudinal extent of
pressures with this delivery method.
circumferential dye spread. Infusion group in blue. Bolus group in green.
Limitations in our study included the following. The place-
Different spinal levels of epidural catheter placement are mid-thoracic
(triangles), low- thoracic (crosses) and lumbar (squares). ment of multiple catheters in each pig meant we could not
control for potential effects of one injection on the spread of

Table 3 Respiratory and epidural physiological data for infusion and bolus groups. PAwP, Pulmonary Artery wedge Pressure

Infusion group (n=10) Bolus group (n=10) Mann-Whitney P (two-tailed)

Median (IQR) Median (IQR) U-test value

PAwP start cm H20 32 (2240) 28 (2740) 48.0 0.878

PAwP nish cm H20 33 (2740) 28 (2740) 48.0 0.878
S pO2 start (%) 99 (9799) 98 (9899) 49.0 0.934
S pO2 nish (%) 98 (9799) 99 (9899) 36.0 0.255
Et CO2 start (%) 5.0 (3.76.3) 5.2 (3.76.3) 42.5 0.567
Et CO2 nish (%) 5.0 (3.65.8) 5.2 (4.85.8) 46.5 0.786
Thoracic Epidural Pressure start (mm Hg) 17 (1620) 18 (1220) 47.5 0.849
Thoracic Epidural Pressure nish (mm Hg) 17 (1720) 17.5 (1020) 48.0 0.877
Lumbar Epidural Pressures start (mm Hg) 15 (1416) 14.5 (1016) 41.5 0.877
Lumbar Epidural pressures nish (mm Hg) 15 (1416) 14 (916) 43.0 0.512
Minute Ventilation (litre min1) 12.9 (8.417.5) 12 (8.417.5) 45.5 0.731
Epidural dye spread: bolus vs infusion | 281

another. However, the benet of this method included limiting volume regimen of infusion. Acta Anaesthesiol Scand. 2009;
the number of pigs required and ensuring similar physiological 53: 4838
parameters, such as epidural pressure, between groups. The 5. Liu SS, Ware PD, Rajendran S. Effects of concentration and
expenses of studying each pig were signicant enough to warrant volume of 2-chloroprocaine on epidural anesthesia in volun-
placing three epidural catheters per animal. It is likely that some teers. Anesthesiology. 1997; 86: 128893; discussion 7A
dye spread continued at the end of each injection, possibly up 6. Kaneko T, Iwama H. The association between injected
until the completion of necropsy. There was no evidence of con- volume of local anesthetic and spread of epidural anesthesia:
tinued spread after the epidural space was exposed. In real terms a hypothesis. Reg Anesth Pain Med. 1999; 24: 1537
this reects clinical practice as the establishment of maximum 7. Johnson RA, Lopez MJ, Hendrickson DA, et al. Cephalad dis-
block height after epidural injection is seen some time later. tribution of three differing volumes of new methylene blue
While we compared the injection pressure of manual hand bo- injected into the epidural space in adult goats. Vet Surg : VS
lused dye with infused dye, we did not include boluses adminis- 1996; 25: 44851
tered by infusion pump. As bolus injection rates via mechanical 8. Yokoyama M, Hanazaki M, Fujii H, et al. Correlation between
pumps are generally slower and injection pressures generally the distribution of contrast medium and the extent of block-
lower than manual boluses, we believe the spread of dye would ade during epidural anesthesia. Anesthesiology. 2004; 100:
likely be somewhere in between our two groups. 150410

Downloaded from http://bja.oxfordjournals.org/ at Gadjah Mada University on November 22, 2016

In conclusion, delivery of dye by bolus injection in a porcine 9. Hermanides J, Hollmann MW, Stevens MF, et al. Failed epi-
model provided greater spread within the epidural space than de- dural: causes and management. British J of Anaesth. 2012;
livery by infusion. In clinical terms, this supports the greater ef- 109: 14454
cacy and lower drug requirement offered by a bolus protocol.11 It 10. Motamed C, Farhat F, Remerand F, et al. An analysis of post-
is reasonable to assume that patients receiving epidural anal- operative epidural analgesia failure by computed tomog-
gesia but who are experiencing pain because of narrow bands raphy epidurography. Anesth Analg 2006; 103: 102632
of analgesia, might benet from an additional bolus dose of 11. Capogna G, Camorcia M, Stirparo S, et al. Programmed inter-
local anaesthetic rather than an increase in infusion rate alone. mittent epidural bolus versus continuous epidural infusion
for labor analgesia: the effects on maternal motor function
and labor outcome. A randomized double-blind study in nul-
Authors contributions
liparous women. Anesth Analg 2011; 113: 82631
Study design/planning: I.M., R.T., C.K. 12. Busscher I, Ploegmakers JJ, Verkerke GJ, Veldhuizen AG. Com-
Study conduct: I.M., R.T., C.K., W.R.H. parative anatomical dimensions of the complete human and
Data analysis: I.M. porcine spine. Eur Spine J 2010; 19: 110414
Writing paper: I.M., R.T., H.V., A.S., W.R.H. 13. Sheng SR, Wang XY, Xu HZ, Zhu GQ, Zhou YF. Anatomy of
Revising paper: all authors large animal spines and its comparison to the human
spine: a systematic review. Eur Spine J 2010; 19: 4656
Acknowledgements 14. Chiang HK, Zhou Q, Mandell MS, et al. Eyes in the needle:
novel epidural needle with embedded high-frequency
We thank Boris Kuzeljevic, Statistician, for help with statistical
ultrasound transducerepidural access in porcine model.
Anesthesiology 2011; 114: 13204
15. Tsui BC, Tsui JH, Corry GN. Estimation of equivalent thresh-
Declaration of interest old currents using different pulse widths for the epidural
stimulation test in a porcine model. Can J Anesth 2014; 61:
None declared.
16. Missant C, Rex S, Claus P, et al. Thoracic epidural anaesthesia
Funding disrupts the protective mechanism of homeometric autore-
This work was supported by Department of Anaesthesiology gulation during right ventricular pressure overload by cardiac
funding only. sympathetic blockade: a randomised controlled animal
study. Eur J Anaesth 2011; 28: 53543
17. Vagts DA, Iber T, Puccini M, et al. The effects of thoracic epi-
References dural anesthesia on hepatic perfusion and oxygenation in
1. Pages F. Anestesia Metamerica. Rev Sanid Mil Madr 1921; 11: healthy pigs during general anesthesia and surgical stress.
35196 (in Spanish) Anesth Analg 2003; 97: 182432
2. Burn JMB, Guyer PB, Langdon L. The Spread of Solutions In- 18. Carrie LES. The approach to extradural space. Anaesthesia
jected Into The Epidural Space: A Study using Epidurograms 1971; 26: 2523
in Patients with the Lumbosciatic Syndrome. British J 19. Gamble JJ, Ambros B, Sguin P, Benmansour P, Simko E.
Anaesth 1973; 45: 33845 Stimulating thoracic epidural placement via a lumbar
3. Higuchi H, Adachi Y, Kazama T. Factors affecting the spread approach causes signicant spinal cord damage in a porcine
and duration of epidural anesthesia with ropivacaine. model. Can J Anesth 2014; 61: 30611
Anesthesiology. 2004; 101: 45160 20. Hogan Q. Distribution of solution in the epidural space:
4. Danelli G, Venuti FS, Zasa M, et al. Continuous lumbar epi- examination by cryomicrotome section. Reg Anesth Pain Med
dural infusion of levobupivacaine: effects of small-or large- 2002; 27: 1506

Handling editor: H. C. Hemmings