Formation of vesicovaginal fistulas in laparoscopic hysterectomy

with electrosurgically induced cystotomy in female mongrel

dogs
Stephanie L. Cogan, MD, Marie Fidela R. Paraiso, MD, and Mohamed A. Bedaiwy, MD
Cleveland, Ohio

OBJECTIVE: The aim of this pilot study was to investigate the relationship between various types of laparo-
scopic bladder injuries and vesicovaginal fistula formation in an animal model.
STUDY DESIGN: Sixteen female mongrel dogs were divided into four groups. All animals underwent a la-
paroscopic hysterectomy. Those assigned to group 1 sustained a 1-cm bipolar cautery injury to the bladder
base without perforation of the bladder mucosa. Animals in group 2 had two sutures of 2-0 polyglactin placed
to incorporate the full thickness of the bladder wall and the vaginal cuff. The bladder injury to group 3 was a
1-cm bladder base laceration induced with monopolar cautery, repaired with two interrupted 2-0 polyglactin
sutures. Group 4 underwent a bladder base cystotomy similar to those in group 3, with the closure incorpo-
rating the anterior vaginal wall. Animals were killed and necropsy was performed at least 28 days after
surgery. The bladder and vagina of each animal were harvested en bloc. Evidence of a vesicovaginal fistula
was determined by two methods: transurethral injection of indigo carmine solution under direct visualization
and air injection during underwater submersion.
RESULTS: The four groups were comparable with regard to postoperative weight changes. No mongrels
showed signs of infection or sepsis. Inspection of the harvested bladder and vagina revealed no fistulas in
groups 1 and 2. One mongrel from group 3 and one from group 4 had evidence of a vesicovaginal fistula.
With 95% CIs, the fistula rate would be at least 2% and as high as 38% if a larger study had been under-
taken.
CONCLUSION: The female mongrel is the first identified animal model of vesicovaginal fistula formation. In
this setting, an electrosurgically induced cystotomy and repair of the bladder during the performance of a la-
paroscopic hysterectomy is associated with the formation of postoperative vesicovaginal fistulas. (Am J Ob-
stet Gynecol 2002;187:1510-4.)

Key words: Vesicovaginal fistula, electrocautery, laparoscopic hysterectomy, female mongrels,

cystotomy

In cases unrelated to obstetric trauma, vesicovaginal fis-
tula formation is caused by inadvertent injury to the blad-
der during surgery.1,2 The reported incidence of this
sequela subsequent to pelvic surgery is 0.5% to 2.0%.1
Since the 1980s, an increasing number of complicated la-
paroscopic surgeries have been developed.3 A review of
the literature reports a 0.02% to 8.3% incidence of blad-
der injury associated with laparoscopy; of those patients
who had vesicovaginal fistulas, the majority occurred sub-
sequent to laparoscopic hysterectomies.4
From the Department of Obstetrics and Gynecology, Cleveland Clinic
Foundation.
Supported by the Research Program Council of the Cleveland Clinic
Foundation.
Presented at the Twenty-eighth Annual Meeting of the Society of Gyneco-
logic Surgeons, Dallas, Tex, March 4-6, 2002.
Reprint requtests: Stephanie L. Cogan, MD, 5444 Chesterton Place, In-
dianapolis, IN 46237. E-mail: coganjhh@aol.com
© 2002, Mosby, Inc. All rights reserved.
0002-9378/2002 $35.00 + 0 6/6/129924
doi:10.1067/mob.2002.129924
The exact mechanism of vesicovaginal fistula forma-
tion remains undiscovered. Two purported etiologies
have been suture placement through the bladder muscu-
laris and electrocautery injury.5 The objective of this
study was to identify an animal model of vesicovaginal fis-
tula formation by inducing two types of bladder injuries:
electrocautery injury or suture placement into the blad-
der during laparoscopic hysterectomy.
Material and methods
Virgin female mongrels were selected for this study be-
cause the anatomic relationship of the uterus, vagina,
and bladder is similar to that of human female reproduc-
tive tract (Fig 1). Animal size was sufficient to make la-
paroscopic hysterectomy feasible. In contrast to the
human anatomy, the bladder is not adherent to the lower
uterine segment so that dissection of the bladder from
the uterus, cervix, and vagina would not be necessary.
This eliminated the possibility of additional uninten-
tional bladder injury.

1510
Volume 187, Number 6
Am J Obstet Gynecol
Fig 1. Anatomic relationships of the uterus, vagina, bladder, and
urethra in dogs.
This study was approved by the Cleveland Clinic Foun-
dation Animal Research Committee. Mongrels were
housed and maintained in the animal facilities at the
Cleveland Clinic with adherence to institutional guide-
lines for the care and use of animals.
Sixteen virgin female mongrels (19-22 kg) were di-
vided into groups of four to undergo laparoscopic hys-
terectomy with suture or cautery injury to the bladder.
The dogs were anesthetized with intravenous thiopental
sodium at 16 to 20 mg/kg and were then intubated by the
laboratory technicians. A dose of preoperative cefazolin
was administered. The animals were placed in the dorsal
recumbent position and were secured to the operating
table with leg straps. The abdomens were shaved and
prepped with povidone-iodine and alcohol. The animals
were then draped in a sterile fashion.
A 10/12 mm trocar was directly inserted at the umbili-
cus. Insufflation was then performed with carbon diox-
ide. Ancillary ports were placed under direct visualization
lateral to the mammary glands, between the umbilicus
and symphysis pubis. Under direct visualization, the blad-
der was drained by suprapubic aspiration with an 18-
gauge needle with syringe. Bipolar cautery at 30 to 40 W
Cogan, Paraiso, and Bedaiwy 1511
Fig 2. Bipolar cautery injury to bladder base.
was applied across the two uterine horns with kleppingers
at their junction with the uterine corpus, after which they
were transected with laparoscopic scissors. The uterine
vasculature was then skeletonized, cauterized with the
bipolar set at 30 W, and transected. With an endohook at-
tached to monopolar cautery at 30 W, the vagina was en-
tered anteriorly and circumscribed at the cervicovaginal
junction. The anatomy of the dog is such that the cervix
is easily identified and that dissection is not required to
isolate the bladder from the lower uterine segment or
vagina. The vaginal cuff was reapproximated with No. 2-0
polyglactin (Vicryl, Ethicon, Somerville, NJ) interrupted
sutures placed at 5-mm intervals.
The bladder injuries were then induced as follows. In
group 1, a bladder injury of 1 cm in diameter was in-
duced with bipolar cautery at 30 W at the bladder base
until tissue blanching was observed (Fig 2). In group 2,
two sutures of 2-0 polyglactin were placed to incorporate
the full thickness of the bladder wall and the vaginal cuff
(Fig 3). In group 3, a 1-cm horizontal bladder base lac-
eration into the bladder lumen was made with the
monopolar cautery and repaired with two interrupted
2-0 polyglactin stitches (Fig 4). In group 4, a similar blad-
der base cystotomy was created; the repair then incorpo-
rated the anterior vaginal wall with two sutures of 2-0
polyglactin (Fig 5).
Postoperative analgesia was provided with aceta-
minophen and Torbugesic twice daily as needed in accor-
dance with the veterinary guidelines of the Cleveland
Clinic Foundation. Food and water were provided at will.
The dogs were monitored for evidence of infection (de-
creased energy, anorexia, inspection of incisions) and
1512 Cogan, Paraiso, and Bedaiwy
Fig 3. Suture insertion into bladder base with incorporation of
vaginal cuff.
Fig 4. A 1-cm horizontal bladder base laceration into the bladder
lumen was made with the monopolar cautery and reapproxi-
mated with two sutures.
urine leakage. The dogs were then killed at least 28
(range 28-75) days after the surgery by carbon dioxide as-
phyxiation. Necropsy was performed with the abdomen
opened by vertical incision. The bladder and vagina were
harvested en bloc. A transurethral catheter was inserted
through the urethra into the bladder. Detection of fistu-
las was determined in two ways: by instillation of dilute in-
digo carmine through the catheter and assessment for a
December 2002
Am J Obstet Gynecol
Fig 5. Bladder base cystotomy induced with monopolar cautery
then repaired with inclusion of the anterior vaginal wall with two
sutures.
tract between the vagina and the bladder and by sub-
merging the specimen under water and instilling air
through a transurethral catheter while observing for the
presence of air bubbles.
The data were expressed as percentage of fistula for-
mation. The Fisher exact test or Student t test was used as
appropriate for statistical analysis. A value of P < .05 was
considered statistically significant.
Results
Sixteen mongrels divided into groups of four com-
prised this study. The mean preoperative weight of the
animals in group 1 was 18.8 ± 0.9 kg (mean ± SD), in
group 2 the mean was 18.4 ± 1.4 kg, in group 3 the mean
was 20.0 ± 1.2 kg, and in group 4 the mean was 19.5 ± 1.1
kg (P = .13). At necropsy, the animal weights were as fol-
lows: group 1, 18.9 ± 2.1 kg; group 2, 18.4 ± 1.5 kg; group
3, 19.8 ± 2.5 kg; and group 4, 20.2 ± 1.2 kg (P = .44). When
groups 1 and 2 were grouped for comparison with groups
3 and 4, groups 1 and 2 had significantly lighter dogs be-
fore the study (P = .04) but did not differ in amount of
weight gained or lost. The mean changes in weight from
preoperative to postoperative assessment for all groups
were not statistically significant.
All animals survived the postoperative course without
any complications. Findings at necropsy revealed no evi-
dence of infection. There was no evidence of abscesses or
purulence. The stitches from the surgery were not visible.
Inspection of the harvested bladders and vaginas re-
vealed one fistula in each of groups 3 and 4 for a 25% fis-
tula rate (P = 1.00). No fistulas were detected in groups 1
and 2. When groups 1 and 2 were combined versus
groups 3 and 4, the difference was not statistically signifi-
cant (P = .47).
Volume 187, Number 6
Am J Obstet Gynecol
Comment
Vesicovaginal fistulas can be devastating for the pa-
tients afflicted; there can be physical, psychologic, and
social implications.1 Despite the efforts of Sims6 and his
successors, repair of genital fistulas remains a signifi-
cant challenge.1 In the Western world, the majority of
fistulas occur as a result of inadvertent bladder injury
during gynecologic surgery.1,7 Obstetric vesicovaginal
fistulas remain a major problem in developing coun-
tries; modern obstetric care has made this a rarity in de-
veloped countries.8
The exact mechanism of fistula formation remains
uncertain. In an effort to elucidate whether absorbable
suture placement through the bladder during the vagi-
nal cuff closure step of a hysterectomy could account
for vesicovaginal fistula formation, a study performed
by Meeks et al9 used the rabbit as an animal model.
There were no demonstrable fistulas in any of the 32
rabbits studied.
An electrosurgically induced cystotomy and repair of
the bladder during performance of a laparoscopic hys-
terectomy is associated with the formation of postopera-
tive vesicovaginal fistulas in this pilot study. All the
animals tolerated the surgery well without postoperative
infections. We used two tests to assess for the presence of
fistulas; both dogs with positive findings had evidence of
a fistulous tract in accordance with both the instillation
of dilute indigo carmine through a transurethral
catheter as well as the presence of bubbles when air was
instilled with the specimen submerged under water. Al-
though the number of fistulas obtained in the groups
with monopolar-induced bladder lacerations is not sta-
tistically significant relative to the other induced blad-
der injuries, it is probable that this is due to the small
number of animals used in this study. However, this in-
vestigation provides the first animal model of vesicovagi-
nal fistulas.
In most cases, definitive cure mandates surgical in-
tervention.10 The greatest success occurs with the initial
attempt; each successive repair results in more scar tis-
sue and impairment of the local vascular supply.2 This
underscores the importance of devising a management
protocol for recognizing cystotomy at the time of initial
injury. The discovery of an animal model opens the
doors to new investigations of reparative surgery as well
as preventive measures for fistula formation. Perhaps
the repair of cystotomy should be performed in two lay-
ers. Consideration should also be made to the lateral
spread of monopolar cautery so that excision of the lac-
eration margins should be carried out before the re-
pair. The interposition of an omental flap, the
technique often used in reparative measures, might
also serve as a preventive measure at the time of cysto-
tomy. An investigation of these preventive modalities is
currently underway.
Cogan, Paraiso, and Bedaiwy 1513
REFERENCES
1. Smith GL, Williams G. Vesicovaginal fistula. Br J Urol Int
1999;83:564.
2. Rizvi JH. Genital fistulae: a continuing tragedy. J Obstet Gy-
naecol Res 1999;25:1.
3. Jansen FW, Kapteyn K, Trimbos-Kempter T, Hermans J, Trimbos
JB. Complications of laparoscopy; a prospective multicentre ob-
servational study. Br J Obstet Gynaecol 1997;104:595.
4. Ostrzenski A, Ostrzenska KM. Bladder injury during laparo-
scopic surgery. Obstet Gynecol Surv 1998;53:175.
5. Margolis T, Mercer LJ. Vesicovaginal fistula. Obstet Gynecol Surv
1994;49:840.
6. Sims JM. On the treatment of vesico-vaginal fistula. Am J Med
1852;23:59.
7. Gerber GS, Schoenberg HW. Female urinary tract fistulas. J Urol
1993;149:229.
8. Elkins TE. Surgery for the obstetric vesicovaginal fistula: a review
of 100 operations in 82 patients. Am J Obstet Gynecol 1994;
170:1108-20.
9. Meeks GR, Sams JO, Field KW, Fulp KS, Margolis MT. Formation
of vesicovaginal fistula: the role of suture placement into the
bladder during closure of the vaginal cuff after transabdominal
hysterectomy. Am J Obstet Gynecol 1997;177:1298.
10. Lee RA, Symmonds RE, Williams TJ. Current status of genitouri-
nary fistula. Obstet Gynecol 1988;72:313.
Discussion
DR ROBERT L. HARRIS, Jackson, Miss. Vesicovaginal fis-
tula remains a significant surgical challenge even for the
most skilled surgeon; thus, prevention should continue
to be our goal. Recent reviews and observational studies
have suggested rates of bladder injury with laparoscopy at
up to 8%, with almost 5% of these resulting in fistula for-
mation. Most commonly, injury seems to occur at laparo-
scopically assisted vaginal hysterectomy with monopolar
cautery as the culprit, usually during mobilization of the
bladder flap. Unfortunately, only about one half of these
injuries are recognized intraoperatively.
The authors present a study to investigate the relation-
ship between common areas of potential injury to the
bladder during laparoscopic pelvic surgery and the sub-
sequent risk for development of vesicovaginal fistula
using a reasonable animal model. Their findings further
support the implication of monopolar cautery in causing
irreversible tissue damage and subsequent fistula forma-
tion. Certainly, a basic understanding of electrosurgical
systems is imperative to any pelvic surgeon using laparos-
copy. The more concentrated energy of monopolar cur-
rent does lead to greater thermal tissue damage with
significant lateral spread. Many times this results in ab-
lated tissue that will subsequently undergo necrosis. The
study findings support that irreversible lateral tissue dam-
age likely did occur in some of the dogs in which
monopolar cautery was used. This resulted in fistula for-
mation even after suture closure. I would ask the authors
to further discuss the differences in biophysics of bipolar
and monopolar cautery. Also, please make suggestions on
how we as pelvic surgeons should make use of these find-
ings to improve our ability to recognize bladder and tis-
sue damage and subsequently prevent fistula formation.
1514 Cogan, Paraiso, and Bedaiwy
DR COGAN (Closing). In response to the first question
of bipolar versus monopolar cautery, that is a very impor-
tant issue that needs to be taken into consideration when
laparoscopic surgery is performed. With monopolar en-
ergy, there is an active concentrating electrode that deliv-
ers electrical energy through the patient and then
returns to a grounding pad, via the patient, to the
electrosurgical unit. One must be aware that the active
electrode is in close proximity to other conductive instru-
ments and tissues, which may lead to injury via stray elec-
trical currents. Electrons like to travel the path of least
resistance, and this can cause direct injury. Insulation fail-
ure can be the cause of stray electrical current, as well as
capacitive coupling.
In bipolar energy, there is an isolated circuit. One
prong of the forceps is active and the other serves as the
return plate. In this system, the intervening tissue acts as
part of the circuit. There is less lateral spread with bipolar
December 2002
Am J Obstet Gynecol
energy versus monopolar energy, as Dr Harris stated.
With bipolar cautery, a pure cutting modality is not gen-
erated and is thus the reason for the use of monopolar
energy.
The key to preventing tissue damage is to be cognizant
that there is lateral spread of thermal energy. When total
laparoscopic hysterectomy is being performed, when cir-
cumscribing the cervicovaginal junction for amputation
of the uterus, it is crucial to be very wary of the location of
the bladder and bowel. If you accidentally touch the blad-
der, you may not actually visualize that you have damaged
it but there can be injury with subsequent necrosis, as Dr
Harris stated. So, I basically encourage gynecologic sur-
geons to exercise caution when using monopolar cautery
adjacent to the viscera. Of interest to note is that the blad-
der histologic features of our specimens demonstrated
similar levels of injury, regardless of the thermal energy
modality.