Introduction

Rectovaginal fistulae (RVF) are a potential complication of Crohn’s disease (CD) associated with significant morbidity and increased risk of proctectomy.1,2 Crohn’s related RVF have a significant impact on quality of life and is a source of considerable social embarrassment for affected women. In a recent Crohn’s population study, the cumulative risk for developing any fistula was 33% after 10 years and 50% after 20 years, with at least one recurrent fistula occurring in 34% of the patients.3 Up to 10% of women with CD will develop RVF.4 Common symptoms include chronic vaginal discharge, dyspareunia, and the passage of flatus or stool through the vagina.

Treatment choices depend on several factors, which include the fistula characteristics, sphincter status, anal canal CD such as strictures and ulcerations, presence of active CD in the rectum, and the degree of impairment on quality of life. Many techniques have been developed in the attempt to treat RVF with a wide range of success rates quoted in the literature.4 Crohn’s related RVF in particular have a higher propensity for recurrence than other fistulae with reported recurrence rates ranging from 25% to 50%.5–8

Data examining long-term surgical success rates and associated information regarding its effect on quality of life (QOL), fecal incontinence (FI), and sexual function is limited.

The aim of this study was to obtain long-term follow-up of surgically treated women with Crohn’s related RVF to examine variables influencing surgical success and to determine the effect of surgery on QOL and sexual function.

Methods

Patients

Clinico-pathological data were collected from an IRB-approved pelvic floor database, which was supplemented by a review of medical records, patient administered questionnaires, and direct patient contact in the form of a telephone call by a trained research nurse. Patient’s functional outcome and quality-of-life parameters were obtained from the prospectively maintained database for all available follow-up visits. Patients were asked to complete a self- administered, structured questionnaire on each return visit to the office. If the patient did not attend for more than 1 year, the information was requested by means of the same questionnaire delivered by mail.

All women with Crohn’s related RVF who underwent surgical repair with intent to close the fistula from 1997 to 2007 were contacted for long-term follow-up. Variables assessed to determine their effect on surgical success rates included age, body mass index (BMI), smoking, CD activity (defined clinically by the presence of active inflammation in the rectum or anal canal, anal/or perineal ulceration, or anal strictures), type of surgical procedure performed, use of a preoperative seton or postoperative diverting stoma, number of previous surgical procedures performed, time interval between last repair and current repair, and use of immunomodulators and steroids within the 3 months prior to surgery.

Long-term closure of RVF was determined by clinical examination in the office or by more invasive evaluation such as examination under anesthesia in suspicious cases. Riddance of preoperative symptoms was also verified by a telephone questionnaire. RVF were considered closed if all preoperative symptoms attributable to the fistula had resolved at the time of follow-up and no fistula was detected by physical exam at the last office visit.

Patients were excluded from the study if a surgical procedure was performed where the intent was not fistula closure such as seton placement, diverting stoma alone, or definitive proctectomy without reconstruction.

Quality of Life and Sexual Function

QOL was assessed using the SF-12 Health Survey,9 The Irritable Bowel Disease Quality of Life Instrument (IBD-QOL),10 and Fecal Incontinence Quality of Life Scale (FIQL).11

In determining sexual function, we initially asked patients at the time of direct telephone contact if they were currently sexually active. For patients who were sexually active, we further asked if they experienced pain or discomfort with sexual intercourse. Patients who were sexually active were then forwarded the Female Sexual Function Index (FSFI) validated questionnaire. The FSFI assesses domains of sexual functioning such as sexual arousal, orgasm, satisfaction, and pain. This provides a domain score range of (0–36) with a score of zero indicating no sexual activity and a score of 36 indicating best sexual function.12

Statistics

Fisher’s exact test, Chi-square test, and multivariable logistic regression model were used to identify the variables associated with success or failure. A p value ≤0.05 was considered as significant.

Results

Patient Demographics

Over a 10-year period between 1997 and 2007, 65 women with Crohn’s related RVF who had surgical procedures with intent to close the fistula at our institution were identified. Median follow-up was 44.6 months (interquartiles, 13.1–79.1) with a mean age of 42.3 ± 2 years and BMI of 27.5 ± 12 kg/m2. Information regarding whether a patient’s RVF had healed at follow-up was available for all 65 patients. At the time of follow-up, 30 patients (46.2%) were successfully healed (Table 1). Twenty-nine (45%) patients (15 healed and 14 unhealed) agreed to complete the QOL questionnaires, and of the sexually active women, 57.1% completed the FSFI questionnaire.

Table 1 Demographic and Patients Characteristics
Full size table

Preoperative Symptoms

The most common complaints were fluid drainage per vagina (75.4%), gas per vagina (64.6%), stool per vagina (56.9%), perineal pain (13.8%), and fecal incontinence (9.2%). We did not assess preoperative dyspareunia.

Surgical Repair

The overall healing rate of surgical repair at the time of follow-up was 46.2%. The most common surgical procedures performed were mucosal advancement flaps (72.3%), episioproctotomy (12.3%), proctectomy and pull-through procedure with colo-anal anastomosis (Turnbull–Cutait procedure) (10.8%), fibrin glue (3.1%), and fistula plug placement (1.5%). There was no significant difference in type of RVF repairs between healed and unhealed patients (p = 0.6). The median number of attempted repairs between the healed and unhealed groups was similar (p = 0.5). Eighteen patients received more than three repairs, with eight patients in this subgroup (44.4%) having their RVF healed at follow-up. No patients who had five or more attempts at repair were healed at follow-up.

Healing rates were not significantly affected by age (p = 0.5), BMI (p = 0.4), comorbidity (p = 0.6), presence of active anorectal CD (p = 0.5), time interval between last repair and most recent repair (p = 0.1), use of a preoperative seton (p = 0.08), or postoperative diverting stoma to protect the RVF repair (p = 0.2) (Table 2). On multivariate analysis, use of immunomodulators such as the biologics infliximab (Remicade®) and adalimumab (Humira®) as well as 6-mercaptopurine and azathioprine within the 3 months prior to surgery was the only variable associated with successful healing (p = 0.009). Smoking and corticosteroids within the 3 months prior to surgery were both associated with failure (p = 0.04).

Table 2 Preoperative and Operative Details
Full size table

Quality of Life and Sexual Function

Twenty-nine (45%) patients, consisting of 15 healed and 14 unhealed RVF, agreed to complete the QOL questionnaires. The SF-12 questionnaire showed modest scores in both the healed and unhealed groups with no significant difference in the categories of physical health (p = 0.6) and mental health (p = 0.7) (Table 3). The IBD-QOL also showed modest scores in both healed (53.7 ± 33.2) and non-healed (42.6 ± 27.1) groups with no significant difference (p = 0.4) (Table 3). Results from the FIQL questionnaire demonstrated no significant difference in the overall scores between the healed and unhealed groups (p = 0.9). Likewise, when the individual components of the FIQL were examined comparing healed and unhealed women, there was no significant difference in the areas of lifestyle (p = 0.7), coping behavior (p = 0.9), depression and self-perception (p = 0.6), and embarrassment (p = 0.5) (Table 3).

Table 3 Patient’s Quality of Life
Full size table

Of our total cohort of 65 women, 28 (43.1%) were sexually active at follow-up. This included 15 patients with healed RVF and 13 patients with unhealed RVF. Of this subgroup of sexually active women, nine women (30%) complained of dyspareunia on direct questioning at the time of telephone contact. All women who complained of dyspareunia were in the unhealed group. Of the 28 women that were sexually active, 16 (57.1%) agreed to complete the FSFI sexual function questionnaire. There was no significant difference between healed and unhealed patients in either the overall FSFI score (p = 0.9) or the separate domains of the FSFI: desire (p = 0.8), arousal (p = 0.4), lubrication (p = 0.3), orgasm (p = 0.5), satisfaction (p = 0.4), and pain (p = 0.5) (Table 4).

Table 4 Female Sexual Function Index (FSFI)
Full size table

Discussion

Patients with Crohn’s related RVF often have significant symptoms which affect their quality of life. When possible, this group of patients should be offered surgical repair in an attempt to improve symptoms. There is no ideal treatment option suitable for all patients, and many techniques have been reported with a wide range of success. Previous publications on Crohn’s RVF repair from our institution and others have addressed short term outcomes. This is the first study to address long-term follow-up for this unique group of patients.

Published healing rates for Crohn’s associated RVF in large studies have ranged from 40% to 60%.4,13–18 Sonoda et al. reported 50% failure rate for Crohn’s related RVF.8 Hull et al. retrospectively reviewed 35 patients with low anovaginal fistulae in CD.15 Overall, an initial healing rate of 54% was reported following primary surgical repair. Ultimately, 68% of patients healed their fistula with the use of additional repairs. Another study reported 60% of the Crohn’s RVF were successfully repaired using a sleeve advancement flap.19 In another report from our institution, six out of 12 patients (50%) with Crohn’s related RVF healed after a total of 21 operations.6

In the present study, after variable numbers and types of repairs, 46.2% of RVF’s healed. This is lower than the healing rate reported for non-Crohn’s related RVF,8 which reflects the complexities of the underlying disease process. This rate of healing is also lower than an earlier report from our group, which showed healing rates as high as 68%.15 We believe that patients with Crohn’s associated RVF initially may heal but with longer follow-up, higher failure rates may be seen. This likely reflects the recrudescent nature of Crohn’s disease and the variable responses to medical treatment. This fact is not fully appreciated in studies with shorter follow-up.

In our study, the use of immunomodulators was significantly associated with successful healing. The use of immunomodulators has also been reported to aid healing of Crohn’s RVF in other studies.20–23 The post hoc analysis of the ACCENT II trial by Sands et al.20 examined 25 women with RVF treated with infliximab. They reported that 72.2% had healed RVF at 14 weeks. However, follow-up of this same group at 54 weeks found that healing had decreased to 44.4%. In our study, 61.5% of patients that were administered immunomodulators within 3 months of their definitive surgery were healed at a median follow-up significantly longer than the Sands study. Of note, healing of RVF in our study involved the use of immunomodulators as well as definitive surgical management. The Sands study involved immunomodulator use only. Another study by Topstad et al.21 looked at using setons with infliximab in eight patients with RVF and showed only 13% healed after the setons were removed. We postulate that preoperative use of immunomodulators followed by curative surgical repair (as in this study) may provide better RVF closure rates than immunomodulators alone or immunomodulators combined with some element of seton drainage. However, direct comparison between our study and others is difficult.

In our study, smoking and use of steroids were associated with a higher rate of failure. Smoking has previously been shown to negatively affect the successful outcome of mucosal advancement flaps in patients with perianal fistulae presumably due to reduced rectal mucosal blood flow.24,25 Likewise, other studies have found steroid use to be associated with a higher failure rate.8

It is unclear how the number of attempted repairs influence outcome. In this study, patients having four or five repairs had healing rates similar to the overall healing rate. This suggests that multiple surgical attempts may still offer a successful outcome. This is tempered by the finding that no patients having more than five repairs were healed at follow-up. Scarring from previous repairs may impede healing with five repairs being the limit in this study. Other reports evaluating all types of perineal fistulae, including RVF, have shown conflicting results on whether the number of repairs affects the eventual healing rate.24,25 It is possible that our study, despite enrolling 65 patients, may have lacked the power to make a definitive statement.

Placement of a preoperative seton has previously been reported as a factor that improves the healing rate of perineal fistulas, including RVF.8,26 The benefit has been speculated to result from drainage of sepsis and promotion of fibrosis in the tract. This study did not show a significant benefit, but shows a trend toward increased healing with seton use (p = 0.08).

The response rate for mailed questionnaires was 45% for QOL and 57.1% for FSFI. This is consistent with typical response rates for mailed questionnaires in the literature, which ranges from 40% to 60%.28

Due to the systemic impact of Crohn’s disease, QOL is generally lower in these patients versus unaffected individuals.27 This may account for the modest QOL scores in the three QOL questionnaires used in our study. When comparing the group of women that had healed RVF versus non-healed, there was no significant difference in QOL overall or separately in any of the domains within the questionnaires. This may reflect that the systemic symptoms of Crohn’s disease have a more overriding influence on QOL rather than any one individual complication of the disease.

Due to the sensitive nature of the topic, evaluation of female sexual function is difficult and has resulted in a paucity of data. Moody et al.29 performed a structured interview of fifty CD patients and compared them to controls. They showed in the CD group that 24% of women had infrequent or no sexual intercourse compared with just 4% of controls. Reasons for sexual inactivity in Moody’s study included abdominal pain (24%), diarrhea (20%), and fear of fecal incontinence (14%). In our study, nearly 67% of women abstained from sexual intercourse. Our higher abstinence rate may reflect that all women had a RVF, which may deter them from having sexual intercourse. In the Moody study, dyspareunia was significantly more common in women with CD compared to controls, and this was independent of the enteric site of disease. Interestingly, they found when women had only perianal disease or fistulae that there was no significant difference in dyspareunia versus the control group. Similar to Moody’s results, this study found that the FSFI scores were similar between the healed and non-healed RVF groups. In particular, subset analysis of dysparuenia with the FSFI found no difference between the two groups. These findings must be interpreted with caution due to the small number of people answering this questionnaire. When we specifically asked all 65 women about dysparuenia at the time of telephone interview, 30% of sexually active women in the unhealed group admitted to experiencing dysparuenia versus zero in the healed group. This may reflect the difficulty in assessing dysparuenia since we found apparent differences in questioning patients on paper versus specifically asking over the phone. We do believe however that an unhealed RVF does contribute to dysparuenia.

Conclusions

Crohn’s related RVF continue to be difficult to treat. Healing increased when immunomodulators were used within 3 months prior to surgery. Further prospective trials are needed to help surgeons decide if and when to consider immunomodulators before surgical repair of Crohn’s RVF. Smoking and steroids were predictors of repair failure. Regardless of successful healing, QOL and sexual function were similar. Dyspareunia appears to be higher for women with unhealed fistulas.