ORIGINAL CONTRIBUTIONS

Vaginal Fistula Following Restorative Proctocolectomy Patrick Y. Lee, M.D., Victor W. Fazio, M.D., James M. Church, M.D., Tracy L. Hull, M.D., Kong-Weng Eu, M.B.B.S., Ian C. Lavery, M.D. From the Department of Colorectal Surgery, Cleveland Clinic Foundation, Cleveland, Ohio Vaginal fistula (VF) is a devastating complication following restorative proctocolectomy. PURPOSE: This study was designed to examine the perioperative factors influencing the outcome and management of vaginal fistula. METHOD: Between October 1983 and September 1994, 526 women underwent restorative proctocolectomy. Nineteen develop VF (3.6 percent), and six were referred from other institutions with this complication. These 25 women were followed for a minimum of nine months. RESULTS: Preoperative diagnosis of ulcerative colitis was made in 23 of the patients with VF (92 percent), and indeterminate colitis and familial adenomatous polyposis was determined in the rest of the patients. Postoperatively, 12 of the 23 women (52 percent) with a preoperative diagnosis of ulcerative colitis had clinical/pathologic findings of Crohn's disease, and 1 woman was reclassified as having indeterminate colitis. Postoperative pelvic sepsis was significantly higher in w o m e n with VF than in those without VF (26.3 vs. 6.3 percent; P = 0.003). Median time until occurrence of VF following loop ileostomy closure was later for women with delayed fmdings of Crohn's disease at 16.5 (range, <1-72) months, compared with women without Crohn's disease at 0.5 (range, <1-67) months (P < 0.05). Of the 163 women with handsewn anastomosis performed at our institution, 12 developed VF (7.4 percent). In contrast, 7 of the 363 patients with stapled anastomosis had VF (1.9 percent; P = 0.003). Site of VF was found at the anastomosis in 12 patients, below in 12 patients, and above in 1 patient. Presence of Crohn's disease and anastomotic technique did not influence the site of VF. Initial management of VF consisted of transanal repair in 20 patients (advancement flap, 12; direct repair, 6; and neoileoanal anastomosis, 2), seton in 1 patient, transabdominal approach in 1 patient, transvaginal in 1 patient, observation in 1 patient, and pouch excision in 1 patient. Of the 13 women without Crohn's disease, 12 had transanal repair (10 healed, 1 had recurrence, and 1 had pouch excision), and 1 had successfully repair with transabdominal technique, for an overall success rate of 84.6 percent. Of the 12 women with VF and delayed l~mdings of Crohn's disease, transanal repair was performed on 9, 1 had pouch excision without repair, 1 had seton placement and pouch excision, and 1 underwent

observation. Transanal tedmique of repair in women with Crohn's disease successfully healed three women (33.3 percent). Overall, of the 12 women with delayed findings of Crohn's disease, 6 had pouch excision~ 3 had recurrences, and 3 healed. CONCLUSION: VF is an tmcommon complication following restorative proctocolectomy and is associated with a high incidence of pelvic sepsis and handsewn anastomosis. Late presentation of VF is more common with Crohn's disease and is associated with a poor prognosis and pouch salvage rate. Transanal techniques are an effective means of VF repair. [Key words: Vaginal fistula; Restorative proctocolectomy; Ileoanal anastomosis; Crohn's disease] Lee PY, Fazio VW, Church JM, Hull TL, Eu K-W, Lavery IC. Vaginal fistula following restorative proctocolectomy. Dis Colon Rectum 1997;40:752-759.

R

estorative p r o c t o c o l e c t o m y (RP) with ileal p o u c h c o n s t r u c t i o n has b e c o m e a preferred surgical op-

tion for patients with ulcerative colitis a n d familial a d e n o m a t o u s polyposis. RP is a c o m p l e x p r o c e d u r e with a n associated p o s t o p e r a t i v e m o r b i d i t y of 45 to 63 percent. ~4 O n e of the m o s t distressing c o m p l i c a t i o n s for b o t h patients a n d s u r g e o n s is the d e v e l o p m e n t of vaginal fistula (VF). This is a difficult c o m p l i c a t i o n to treat a n d has a high i n c i d e n c e of r e c u r r e n c e a n d s u b s e q u e n t p o u c h excision. Fortunately, VF is a n u n c o m m o n complication, with a r e p o r t e d i n c i d e n c e of b e t w e e n 5.3 a n d 12 percent. 1'4-7 B e c a u s e it is a n u n c o m m o n p r o b l e m , there are f e w reports o n outc o m e of t r e a t m e n t or m a n a g e m e n t guidelines. T h e p u r p o s e of this s t u d y is to r e v i e w a single institutional e x p e r i e n c e with VF a n d to e x a m i n e the p e r i o p e r a t i v e factors a n d m a n a g e m e n t o u t c o m e associated with this complication.

PATIENTS

AND

METHODS

B e t w e e n O c t o b e r 1983 a n d S e p t e m b e r 1994, 526 w o m e n u n d e r w e n t RP at the C l e v e l a n d Clinic F o u n d a t i o n (CCF). N i n e t e e n d e v e l o p e d VF (3.6 percent), a n d six others w e r e referred from other institutions for

Read at the meeting of The American Society of Colon and Rectal Surgeons, Montreal, Quebec, Canada, May 7 to 12, 1995. Address reprint requests to Dr. Fazio: Department of Colorectal Surgery A-111, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195. 752

Vol. 40, No. 7

VF FOLLOWING RESTORATIVEPROCTOCOLECTOMY

treatment of this complication. These 25 w o m e n were reviewed in regard to their age, diagnosis, prior abdominal and pelvic surgery, preoperative ileostomy, indication for surgery, type of p o u c h construction and ileal-anal anastomosis, postoperative sepsis, timing and site of VF, methods of fistula repair, associated perineal fistula, and p o u c h outcome. Median age of the 25 w o m e n with VF was 31 (range, 18-53) years. Preoperatively, ulcerative colitis was diagnosed in 23 patients (92 percent), indeterminate colitis in 1 patient (4 percent), and familial adenomatous polyposis in 1 patient (4 percent). Indications for RP were bleeding (36 percent), toxic colitis (32 percent), steroid d e p e n d e n c y (20 percent), dysplasia (8 percent), and urgency (4 percent). Postoperative diagnosis of Crohn's disease by either histologic and/or clinical criteria was found in 12 w o m e n (48 percent), with 8 having both histologic and clinical findings consistent with Crohn's disease and 4 with clinical criteria of Crohn's disease as described by Kotanagi e t al. 8 Thirteen w o m e n (25 percent) had RP performed in three stages, and 11 w o m e n (44 percent) had two-stage ileal pouch construction. History of prior pelvic and perineal surgery was found in seven w o m e n (rectovaginal fistula repair, 3; episiotomy, 2; endoanal pull-through, 1; hysterectomy, 1; 28 percent). Median weight was 129 (range, 90-189) pounds. Pouch construction of these 25 w o m e n with VF are detailed in Table 1. Mucosectomy was performed in 16 (15 from CCF, 1 referred). Rectal cuff length ranged from 0.5 to 7 cm above the dentate line. A diverting ileostomy was constructed at the time of pouch construction in 21 patients (84 percent). Operative records and pathology slides of the 25 w o m e n were reviewed by one surgical pathologist. Postoperative findings of Crohn's disease on biopsy or on excised specimen were also reviewed by the

same surgical pathologist. Healing was defined by absence of symptoms and visual confirmation of the repair by endoscopic and Gastrografin | studies (ER Squibb & Sons, Inc., Princeton, NJ). All VF repairs using the transanal advancement flap or neoileoanal anastomosis were done at CCF as described by Fazio and Tjandra. 9 Direct repair in four of the six w o m e n performed at CCF involved transanal elliptical excision of the internal os, with excision/ curettage of the tract and multilayer closure of the anorectal defect, leaving the vaginal defect open. Abdominal approach of VF repair referred to initial transanal mobilization of the ileoanal anastomosis, followed by a similar excision and repair described for direct repair and abdominal mobilization of the pouch and neoileoanal anastomosis. Transvaginal repair in one patient was performed at an outside institution before presentation to CCF. Statistical analysis was conducted by univariate analysis using the chisquared and Fisher's exact test for small numbers. RESULTS

Patient Characteristics Among the 23 w o m e n with ulcerative colitis, 12 were diagnosed postoperatively with Crohn's disease (52 percent) and i with indeterminate colitis. Only 10 of the 23 w o m e n retained the diagnosis of ulcerative colitis following RP (40 percent). Seventeen of the 25 w o m e n with VF had urgent colectomy (i.e., toxicity, n = 8, and bleeding, n = 9; 68 percent). Thirteen of these w o m e n had three-stage RP performed (toxic, n = 8; bleed, n = 5). Median time interval between subtotal colectomy and ileal pouch construction in these 17 w o m e n was 9 (range, 3-24) months. Pouch Construction

Table 1. Details of Pouch Construction in 25 Women with Vaginal Fistulas J-Pouch n=17

S-Pouch n=8

(68%)

(32%)

13 4

0 8

9 8 14 (82)

7 1 7 (88)

Pouch construction

Stabled Handsewn

Anastomosis Handsewn with mucosectomy

Stapled (double and single) Diverting stoma (%)

753

and Anastomosis

Pouch construction details are presented in Table 1. Mucosectomies were performed in 16 women. Rectal cuff length ranged from 0.5 to 7 cm above the dentate line. Cuff abscess was found in one patient. Of the 526 w o m e n w h o had RP performed at CCF, 12 of the 163 handsewn anastomoses developed VF (7.4 percen0 compared with 7 of 363 stapled anastomoses (1.9 percenO. The higher incidence of VF in handsewn vs stapled was significant (chi-squared, P = 0.003). Although the sites of VF appeared to be more c o m m o n below the anastomosis in the 19 w o m e n with ileal pouch-anal anastomosis performed

754

LEE E T A L

Dis Colon Rectum, July 1997 Percentage

at our institution, there were as many VF above and below the anastomosis w h e n the other six referred w o m e n with VF were included (Table 2). Only one patient had VF develop above the ileoanal anastomosis, and she had delayed fndings of Crohn's disease. During the past 11 years, the incidence of VF at CCF has declined. This parallels the decreased use of the handsewn technique of ileoanal anastomosis (Fig. 1).

Timing and Location of VF Median time of VF presentation after loop ileostomy closure was significantly different between w o m e n with and those without a postoperative diagnosis of Crohn's disease. Median time of VF presentation in w o m e n without Crohn's disease was 0.5 (range, < 1 - 6 7 ) months. By contrast, w o m e n with delayed findings of Crohn's disease (Fig. 2) had presentation at 16.5 (range, <1-72) months. This difference in timing of VF presentation was significant by the Fisher's exact test ( P < 0.05). Of the two w o m e n without Crohn's disease w h o presented with VF at greater than 12 months, one had a postoperative diagnosis of indeterminate colitis and the other, ulcerative colitis (she had multiple failed repairs and ultimately had pouch excision without histologic findings of Crohn's disease). Eight w o m e n developed VF before loop ileostomy closure (six were symptomatic, two asymptomatic diagnosed by pouchogram); three of these eight w o m e n had delayed findings of Crohn's disease. The site of VF in w o m e n with and without delayed findings of Crohn's disease were not significantly different ( P = 0.238). Of the 12 w o m e n with VF and Crohn's disease, 6 were found to have the fistula arising at the ileoanal anastomosis, 5 of which were below and 1 above. Of the 13 w o m e n with VF and without Crohn's disease, 8 had fistulas arising below the anastomosis and 5 had fistulas at the anastomosis. Associated perineal fistulas with VF were found in 7

Table 2. Incidence and Site of Vaginal Fistula (VF) Associated with Type of Anastomosis of the Cleveland Clinic Foundation Women (n = 19) Type of Anastomosis

n

Handsewn Staple (double and single)

163 339

Site of VF

VF 12 7

Below

At

Above

6 5

5 2

1 2

83 84 85 86 87 88 89 90 91 92 93 94 Year Figure 1. Overall incidence of vaginal fistulas at the Cleveland Clinic Foundation from October 1983 to September 1994.

72

@ O @

~" >12 e. o

E

P

@ @ @@ @ |

O

O 6-<12

o.
@ O@@

OO 000 0 oOxOo

CD NCD Figure 2. Timing of vaginal fistulas between patients with Crohn's disease (CD) and no Crohn's disease (NCD).

of 12 w o m e n with Crohn's disease and 2 of 13 w o m e n without Crohn's disease.

Postoperative Complications Postoperative pouch-related complications and site of VF in the 25 w o m e n are shown in Table 3. When only the w o m e n with pouch construction done at CCF were considered, there was a higher incidence of postoperative pelvic sepsis associated with those w h o

Vol. 40, No. 7

VF FOLLOWING RESTORATIVEPROCTOCOLECTOMY

Table 3. Relationship Among Postoperative Sepsis, Perineal Fistula, Pouchitis, and the Site of Vaginal Fistula (VF) Site of VF* Type of Complication

Sepsis Pelvic abscess Leak Cuff abscess Unknown

Perineal fistula Pouchitis

n (%)

Below

At

Above

4 2 0 1 1 5 3

6 4 2 0 0 3 2

0 0 0 0 0 1 0

10 (40) 6 2 1 1 9 (36) 5 (20)

* In relation to the anastomosis.

Table 4. Outcome of Initial Vaginal Fistula Repair Type of Surgical Repair

n

Success

Transanal With ileostomy

Without ileostomy Transabdominal With ileostomy Transvaginal Without ileostomy

Pouch excision Seton placement No treatment

15 5 1 1 1 1 1

developed VF than in those without VF (26.3 vs. 6.3 percent; P -- 0.003). Successfully treated postoperative pelvic sepsis did not impact on pouch outcomes and function after VF repair. The anastomotic site was the most c o m m o n location of vaginal fistulas in those with postoperative sepsis. There was no association between Crohn's disease and pelvic sepsis. Incidence of pouchitis during a median follow-up of 42 (range, 8-119) months after pouch construction was similar for those with and without VF (20.8 vs. 23.5 percent). The sites of VF were nearly equally divided between those at and those below the anastomosis.

Pouch Treatments and Outcomes Initial surgical repair of VF was made in 22 of the 25 w o m e n (20 transanal, 1 transabdominal, 1 transvaginal). The other three were found to have Crohn's disease (2 chose against a repair, and 1 had pouch excision because of severe perineal sepsis). Success of initial VF repair are shown in Table 4. Diverting or "protecting" ileostomies for VF repair did not appear to show any statistical benefit over leaving the fecal stream intact.

755

Transanal techniques in the initial repair of the 20 patients with VF (advancement flap, 12; direct repair, 6; neoileoanal anastomosis, 2) resulted in the healing of 10 w o m e n (50 percent). However, 8 of the 20 w o m e n had delayed findings of Crohn's disease. Excluding these 8 w o m e n with Crohn's disease, transanal repair of the 12 w o m e n without Crohn's disease (9 advancement flap, 3 direct repair) resulted in 10 healed fistulas (83.3 percent; 8 after the first repair and 2 after the second repair). Of the two w o m e n w h o did not heal, the one with RP for familial adenomatous polyposis has since undergone a third transanal VF repair and is awaiting radiologic confirmation of the healed fistula. The other patient had her pouch excised after failing multiple repairs, and she retained a pathologic diagnosis of ulcerative colitis. The overall pouch vaginal fistula outcome treatment is shown in Figure 3.

Crohn's Disease and Treatment Outcomes o f the 526 w o m e n w h o underwent RP at CCF between 1983 and 1994, 21 w o m e n had postoperative findings of Crohn's disease (4 percenO, 9 of w h o m developed VF. Comparing these w o m e n with those w h o have Crohn's disease without VF, there is a higher association of p o u c h excision with the develo p m e n t of VF (66.6 v s 25 percent), but was not statistically significant (Fisher's exact test, P -- 0.063). Preoperative features suggestive of but not diagnostic of Crohn's disease (i.e., extraintestinal manifestations, history of perianal fistulas, relative rectal sparing) were found in 6 of the 12 w o m e n with VF and Crohn's disease. However, of the three w o m e n w h o had a history of rectovaginal fistula repair, all three had findings of Crohn's disease following RP. Attempted repair of VF was made in 9 of the 12 w o m e n with Crohn's disease; the other three w o m e n were discussed earlier. Transanal techniques of VF repair were used in all nine women: four had pouch excision after multiple repairs; three healed (two healed after the first repair and the other after a failed transvaginal repair); and two had recurrences (one with permanent fecal diversion and the other has elected to keep the pouch despite vaginal drainage). None of the six w o m e n with Crohn's disease w h o failed to heal after transanal repair remained symptom-free for more than one year before requiring pouch excision or permanent fecal diversion. In contrast, the three w o m e n w h o healed are 9 months, 13 months, and 9 years post-VF repair. In summary, of

756

LEE E T A L

Dis Colon Rectum,July 1997

25(12c)

1(lc)

Observation Seton Transvaginal Transanal l(lc) Placement Repair Repair

Abdominal Approach

20(8c)

~

Rlei!Crr)ed ~

~

9(5C) ~

Pouchexcision 7(6c)

10(2C)~~

NoRepair 2ridRepair l(lc)

1

~

Healed 14(3c) 3(3c) 3'(lc) ,~__+

1(lc)

Recurred7(5C) V Additional Repairs

2(lc)

V 6(4c)

Recurred4(3c) Figure 3. Pouch vaginal fistula treatment outcomes of the 25 women (12 with Crohn's disease), c = Crohn's disease patients. *One patient is healed without radiologic confirmation.

the 12 women with delayed findings of Crohn's disease, 6 had their pouches excised, 3 experienced recurrences, and 3 healed. Overall, treatment outcome is greatly influenced by Crohn's disease (Table 5). The overall success rate of repair in women without Crohn's disease is 84.6 percent (92 percent were repaired with transanal techniques). Delayed findings of Crohn's disease accounted for 85.7 percent of the pouches excised. Transanal techniques successfully repaired 33.3 percent of women with Crohn's disease and accounted for 25 percent of the healed fistula in this group of women. DISCUSSION The incidence of VF in our study was 3.6 percent, which is lower than the reported incidence of 5.3 to 12 percent. 4-7 The origin of this complication is unknown, but our study showed a higher incidence of VF associated with handsewn anastomosis (P = 0.003), pelvic sepsis (P = 0.003), and delayed findings of Crohn's disease (P < 0.05). The incidence of pelvic sepsis after ileal pouch construction varies between 5 and 25 percent. < 7, 9 In

Table 5. Overall Treatment Outcome Between Crohn's (CD) and

Non-Crohn's Disease (NCD) Patients with Vaginal Fistulas Outcome

n

NCD (n = 13)

CD (n = 12)

Healed Recurred Pouch excision

14 4 7

11 1 1

3 3 6

the 19 women who developed VF following RP performed at our institution, the incidence of pelvic sepsis was significantly higher than those without vaginal fistula (26.3 vs. 6.3 percent; P = 0.003). Similar association between pelvic sepsis and vaginal fistulas has been reported by others. 2' 4 Wexner e t a l . 5 theorized that pelvic sepsis is the primary cause of vaginal fistula. A number of predisposing factors leading to development of pelvic sepsis has been cited in the literature, including tension on the anastomosis, long rectal cuff, inadequate drainage, mucosectomy, preoperative toxic colitis, and pelvic contamination. 1~ Our study did not assess the technical difficulties in pouch construction. However, we did not find length

Vol. 40, No. 7

VF FOLLOWING RESTORATIVE PROCTOCOLECTOMY

of rectal cuff, mucosectomy, or urgent indications such as toxic colitis or bleeding to be associated with an increased incidence of postoperative pelvic sepsis. Pelvic sepsis has b e e n reported to account for 38 to 50 percent of p o u c h failures. 3' 11 In our study p o u c h failure was mostly related to a delayed diagnosis of Crohn's disease. There was no association b e t w e e n pelvic sepsis and Crohn's disease. We did not find any sequelae of p o u c h dysfunction following successful treatment of pelvic sepsis as suggested b y Keighley e t al. 3

VF d e v e l o p e d equally at or b e l o w the ileoanal anastomosis, regardless of the timing of presentation. This was an u n e x p e c t e d finding, because it is generally believed that early presentations of VF are usually related to anastomosis and late ones are cryptoglandular or related to Crohn's disease. G r o o m e l al. 6 also reported VF b e l o w the anastomosis in 2 of 17 w o m e n following RP and postulate cryptoglandular disease and retained ulcerative mucosa as the cause. There is indirect evidence that anal canal changes occurring after RP m a y cause fibrosis and degeneration of the anal sphincter muscles and vaginal septum. 13 There may, therefore, be a "zone" of instability b e t w e e n the ileoanal anastomosis and the anal verge created by surgical trauma that predisposes to local sepsis and fistulization. This m a y also explain w h y there was a significant incidence of VF associated with h a n d s e w n vs. stapled anastomosis (7.4 vs. 1.9 percent; P = 0.003). Handsewn anastomoses are technically more difficult and are associated with a higher incidence of anal trauma and incontinence. 14' 15 Multivariate analysis of factors influencing the results of RP by Keighley e l al. 3 have also s h o w n a higher incidence of p o u c h failure associated with h a n d s e w n anastomosis. However, prospective studies have not s h o w n any differences in complications or p o u c h outcome b e t w e e n stapled and h a n d s e w n ileoanat anastomosis. 1r A delayed diagnosis of Crohn's disease was the most important cause of p o u c h failure. Twenty-one of 526 w o m e n (4 percent) w h o underwent RP at CCF b e t w e e n 1983 and 1994 were found to have Crohn's disease. Nine of these 21 w o m e n (43 percent) develo p e d VF. Similar findings were reported in Toronto, where three of five w o m e n with delayed findings of Crohn's disease d e v e l o p e d VF. 19 Pouch excision owing to delayed findings of Crohn's disease are reported to be b e t w e e n 25.4 and 44 percent 1' 19, 20 and are d e p e n d e n t on length of follow-up. In our o w n institutional experience, w o m e n with delayed find-

757

ings of Crohn's disease w h o d e v e l o p e d VF are more likely to loose their p o u c h c o m p a r e d with those without VF, but this was not statistically significant (66.7 vs. 25 percent; P = 0.063). The difficulty in identifying patients with Crohn's disease before RP has b e e n detailed by H y m a n e t al. 2~ In the current study, 12 of the 23 w o m e n with preoperative diagnoses of ulcerative colitis had their diagnoses changed postoperatively to Crohn's disease. Extraintestinal manifestation of Crohn's disease is not reliable in distinguishing from ulcerative colitis. Of six w o m e n with preoperative features suggestive of but not diagnostic of Crohn's disease ( i . e . , mouth ulcers, skin erythematous patches, arthritic complaints, relative rectal sparing), only three were found after RP to have Crohn's disease; two of the others had ulcerative colitis, and one had indeterminate colitis. A history of rectovaginal fistulas b e f o r e RP was f o u n d to h a v e the highest correlation with the delayed findings of Crohn's disease. T h r e e patients in our series with a history of rectovaginal fistula repair w e r e f o u n d to h a v e Crohn's disease after RP. This association b e t w e e n rectovaginal fistulas a n d Crohn's disease has also b e e n n o t e d b y B a n d y e t a l . 21 w h o f o u n d an 11 percent incidence of Crohn's disease in their series of 138 patients with rectovaginal fistulas and 1 patient with ulcerative colitis w h o h a d a history of perirectal abscesses with s p o n t a n e o u s drainage into the vagina, h e m o r r h o i d ectomy, and l y m p h o g r a n u l o m a v e n e r e u m . O t h e r features associated with d e l a y e d findings of Crohn's disease w e r e late p r e s e n t a t i o n of VF ( > 1 2 m o n t h s ) and d e v e l o p m e n t of other perianal fistulas. Despite a 50 p e r c e n t loss of p o u c h function in our w o m e n with VF and Crohn's disease, transanal techniques of VF repair in these w o m e n can still achieve a 33.3 p e r c e n t healing rate, with no recurrence at follow-up from nine m o n t h s to nine years. Similar successes h a v e also b e e n reported. 19' 20 H o w e v e r , in the six w o m e n with Crohn's disease w h o had p o u c h excision or p e r m a n e n t fecal diversion, all had recurrence within o n e year of the repair. In contrast, 2 of the 12 w o m e n without Crohn's disease w h o had transanal repair of VF d e v e l o p e d recurrences within on year, one of w h o m s u b s e q u e n t l y h a d her p o u c h excised. Therefore, d e l a y e d findings of Crohn's disease does not contraindicate attempt to repair the VF, but vaginal refistualization within o n e y e a r after transanal repair is an o m i n o u s sign and is likely to b e associated with Crohn's disease and a p o o r prognosis.

758

LEE E T A L

The low incidence of VF following RP makes it difficult to determine the o p t i m u m m a n a g e m e n t of this complication. A n u m b e r of techniques have b e e n advocated for the treatment of this complication, including endoanal flap, direct repair, seton placement, transvaginal repair, fistulotomy, sleeve resection with neoileoanal anastomosis, m y o c u t a n e o u s flap with gracilis muscle, and ileal diversions alone or with repairs.i, 2, 6, 7, 9-11 All of these techniques were develo p e d for the treatment of rectovaginal fistulas. H o w ever, there are important anatomic differences and considerations b e t w e e n the rectum and the ileal pouch. The wall of the ileum is thinner than the rectum, and the blood supply of the ileal p o u c h is from the superior mesenteric vessel, in contrast to the m a n y collaterals in the native rectum The decision over which type of transanal technique to use is d e p e n d e n t on w h e r e the fistula arises in relation to the anastomosis, the integrity of the sphincter and perineal body, and the presence of associated fistulas. We have favored a full-thickness a d v a n c e m e n t flap or total mobilization of the ileoanal anastomosis with neoileoanal anastomosis w h e n the fistula is at or near the anastomosis. In those b e l o w the anastomosis, a direct repair with or without plication of the sphincter and perineal b o d y is favored. The use of a diverting ileostomy during the repair of the VF did not a p p e a r to influence the healing rate, although it is likely that fecal diversion was used in the "worst" fistulas. Successful healing was found to be nearly equally divided, regardless of whether a diverting ileostomy was used. Conclusions about the significance of these findings must be cautioned in view of the limitation of a retrospective study, but it raises a question as to the importance of a diverting ileostomy during VF repair. G r o o m e t al. 6 s h o w e d that VF repair with or without ileostomy had an equally p o o r outcome in p o u c h patients. Mazier e t al. 22 recently reported their techniques of repair of complex anovaginal and rectovaginal fistulas in 95 patients without fecal diversion and achieved a 97 percent excellent or g o o d outcome. Wexner and colleagues, 5 however, s h o w e d that diversion of ileal content alone m a y result in a 13 percent healing rate without any repair. A diverting ileostomy is obviously advocated if the patient has severe perineal fistula and sepsis, but the exact role of a covering stoma during an elective VF repair remains to be determined.

Dis Colon Rectum, July 1997 CONCLUSIONS

Vaginal fistula following restorative proctocolectomy is an u n c o m m o n complication. It has a significant association with a history of pelvic sepsis and h a n d s e w n anastomosis. It can occur at or b e l o w the ileoanal anastomosis. Delayed presentation of VF is associated with Crohn's disease and has a p o o r prognosis for p o u c h salvage. However, a measurable degree of success can still be achieved in these women. Transanal repair is an effective technique of VF repair in w o m e n without Crohn's disease. REFERENCES 1. Fazio VW, Ziv Y, Church JM, et al. The ileal pouch anal anastomoses: complications and function in 1005 patients. Ann Surg 1995;222:120-7. 2. Marcello PW, Roberts PL, Schoetz DJ, Coller JA, Murray JJ, Veidenheimer MC. Long-terra results of the ileoanal pouch procedure. Arch Surg 1993;128:500-4. 3. Keighley MR, Winslet MC, Flinn R, Kmiot W. Multivariate analysis of factors influencing the results of restorative proctocolectomy. Br J Surg 1989;76:740-4. 4. Mathey P, Ambrosetti P, Morel P, et al. Experience Suisse de l'anastomoe ileo-anale avec reservoir (AIA). Ann Chir 1993;47:1020-5. 5. Wexner SD, Rothenberger DA, Jensen L, et al. Ileal pouch vaginal fistulas: incidence, etiology and management. Dis Colon Rectum 1989;32:460-5. 6. Groom JS, Nicholls RJ, Hawley PR, Phillips RK. Pouchvaginal fistula. Br J Surg 1993;80:936-40. 7. O'Kelly TJ, Merrett M, Mortensen NJ, Dehn TC, Kettlewell M. Pouch vaginal fistula following restorative proctocolectomy: aetiology and management. Br J Surg 1994;81:1374-5. 8. Kotanagi H, Kramer K, Fazio VW, Petras RE. Do microscopic abnormalities at resection margins correlate with increased anastomotic recurrence in Crohn's disease? Prospective analysis of 100 cases. Dis Colon Rectum 1991;34:909-16. 9. Fazio VW, Tjandra JJ. Pouch advancement and neoileoanal anastomosis for anastomotic stricture and anovaginal fistula complicating restorative proctocolectomy. Br J Surg 1993;79:694-6. 10. Dozois RR. Pelvic and perianastomotic complications after ileoanal anastomosis. Perspect Colon Rectal Surg 1988;1:113-21. 11. Gemlo BT, Wong D, Rothenberger DA, Goldberg SM. Ileal pouch-anal anastomosis patterns of failure. Arch Surg 1992;127:784-7. 12. Galandiuk S, Scott NA, Dozois RR, et al. Ileal pouchanal anastomosis, reoperation for pouch-related complications. Ann Surg 1990;212:446-54. 13. Emblen R, Erichsen AA, Morkrid L, Ganes T, Stien R,

Vol. 40, No. 7

14.

15.

16.

17.

VF FOLLOWING RESTORATIVE PROCTOCOLECTOMY

Bergan A. Failed ileoanal anastomosis: correlations between clinical function and anal canal neurophysiology and histologic examinations. Scand J Gastroenterol 1989;24:623-31. Tuckson WB,McNamara MJ, Fazio VW, Lavery IC, Oakley JR. Impact of anal manipulation and pouch design on ileal pouch function. J Natl Med Assoc 1991;83: 1089-92. Sun WM, Read NW, Katsinelos P, Donelly TC, Shorthouse AJ. Anorectal function after restorative proctocolectomy and low anterior resection with coloanal anastomosis. Br J Surg 1994;81:280-4. Seow-Choen A, Tsunoda A, Nicholls RJ. Prospective randomized trial comparing anal function after handsewn ileoanal anastomosis with mucosectomy versus stapled ileoanal anastomosis without mucosectomy in restorative proctocolectomy. Br j Surg 1991;78:430-4. Luukkonen P, Jarvinen H. Stapled versus hand-sutured ileoanal anastomosis in restorative proctocolectomy a

759

prospective randomized study. Arch Surg 1993;128: 437-40. 18. McIntyre PB, Pemberton JH, Beart RW Jr, Devine RM, Nivatvongs S. Double-stapled vs handsewn ileal pouchanal anastomosis in patients with chronic ulcerative colitis. Dis Colon Rectum 1994;37:430--3. 19. Deutsch AA, McLeod RS, Cullen J, Cohen Z. Results of the pelvic-pouch procedure in patients with Crohn's disease. Dis Colon Rectum 1991;34:475-7. 20. Hyman NH, Fazio VW, Tuckson WB, Lavery IC. Consequences of ileal pouch-anal anastomosis for Crohn's colitis. Dis Colon Rectum 1991;34:653-7. 21. Bandy LC, Addison A, Parker RT. Surgical management of rectovaginal fistula in Crohn's disease. Am J Obstet Gynecol 1983;147:359-63. 22. Mazier WP, Senagore AJ, Schiesel EC. Operative repair of anovaginal and rectovaginal fistulas. Dis Colon Recturn 1995;38:4-6.

A M E S S A G E TO OUR SUBSCRIBERS

Williams& Wilkins and most other publishers seal issues of professional journals in polywrap bags to mail to subscribers. Although these bags are veW effective in protecting issues from damage during transport, they are not biodegradable and pose serious environmental problems. A number of you have written to us to suggest that we change to biodegradable plastic or paper wrappers or no wrappers at all. We have considered the alternatives and have chosen the one imposing the least environmental threat--no wrappers for issues mailing to addresses within the United States. Second class postage regulations require that wrappers be used to mail issues outside the United States. We hope your issues of tile DISEASES OF THE COLON & RECTUM arrive in good condition. If they do not, please call us at 1-800-638-6423. ALMAJ. WILLS President Periodical Publishing