Rectovaginal Fistula in Crohn's Disease: When and How to Operate?

Abstract

Rectovaginal fistula (RVF) occurring during the course of Crohn's disease (CD) constitutes a therapeutic challenge and is characterized by a high rate of recurrence. To optimize the outcome of CD-related RVF repair, the best conditions for correct healing should be obtained. Remission of CD should be achieved with no active proctitis, the perianal CD activity should be minimized, and local septic complications should be controlled. The objective of surgical repair is to close the fistula tract with minimal recurrence and functional disturbance. Several therapeutic strategies exist and the approach should be tailored to the anatomy of the RVF and the quality of the local supporting tissues. Herein, we review the medical and surgical management of CD-related RVF.

Crohn's disease (CD) is a chronic inflammatory disease of unknown etiology but with important genetic predisposition that mainly leads to patchy transmural inflammation of the gastrointestinal tract. ¹ The chronic inflammatory process can lead to ulcers that communicate with adjacent hollow organs or the skin, leading to different forms of fistula. Perianal fistula in CD was first reported in 1938 and led to the description of the disease. ²

In patients suffering from CD, the cumulative incidence of perianal fistula is 18% after 10 years and 23% after 20 years from diagnosis. ³ Fifty-four percent of all fistulas were reported to be perianal with only 9% being rectovaginal fistulas (RVFs). ⁴ The history of fistulizing CD is characterized by frequent recurrences and prolonged periods of relative quiescence. The incidence of recurrence ranges from 34 ⁴ to 59% ⁵ at 2 years. Treatment of perianal CD is challenging. The goal is to control sepsis and induce mucosal and fistula track healing. However, morbidity of perianal CD remains high and leads to proctectomy in around one in five of these patients. ³

Of note, CD is the second most common cause of RVF, after obstetric trauma (where long labor causes ischemic necrosis of the rectovaginal septum), with other causes including radiotherapy and iatrogenic injury during gynecologic or rectal surgery. ^{6 7 8} Vaginal involvement in cases of CD-related fistula leads to worse outcomes ⁹ and a higher rate of diverting ileostomy or proctectomy. ^{9 10 11 12 13 14} For instance, Löffler et al documented that surgical repair in CD-related RVF only has a 53% success rate at 60 months, and requires proctectomy in 22% of patients. ¹³ Others report that CD-related RVF has a healing rate of only 58 to 63%, as reported in a heterogeneous cohort of patients with a follow-up of at least 3 years. ^{15 16} This might be due to altered quality of rectal and perianal tissue and the fact that this is a poorly vascularized and thin area—as well as the poor healing that characterizes CD and relates to some of its underlying pathogenic mechanisms.

RVF may have devastating consequences on quality of life. Patients complain about discharge of gas, pus, and/or feces through the vagina, dyspareunia, and/or recurrent vaginal infections, therefore leading to significant impact on social and sexual lifes. ¹⁷

Despite its important morbidity, the literature is sparse regarding RVF, especially in CD. ^{6 18} The available evidence remains unclear as to when and how to operate for CD-related RVF. In the present narrative review of the literature, we aim to answer these two critical questions.

When to Operate?

When to operate on CD-related RVF depends on the general status of the patient, but also very specifically on the activity of CD and local anorectal inflammation.

Evaluation of patients with perianal CD and/or RVF should include detailed medical history including quantification of anal continence, using, for example, the Wexner score, ¹⁹ a quality of life evaluation scale, and a detailed sexual history. Clinical examination should include rectal and vaginal examinations, evaluation of the sphincter tone, evaluation for signs of infection, number and localization of fistulas, and an assessment of mucosal inflammatory activity particularly in the rectum. The latter should be performed by proctoscopy, with an additional assessment of the quality of surrounding tissues that will help in deciding the type of repair to be performed. This can sometimes require a formal examination under anesthesia. Endorectal ultrasound may help to assess potential associated sphincter defects/injury (which is documented in up to 30% of women with RVF, notably after obstetric trauma ^{20 21} ), to define the RVF tract and the presence of any associated fistulas. The sensitivity of the examination can be enhanced using hydrogen peroxide or air in a syringe. ²² Magnetic resonance imaging may help complete the examination, especially in complex cases where there are multiple fistulas and/or septic complications. Evaluation of the entire digestive tract to detect areas of CD activity should not be forgotten.

To optimize the outcome of CD-related RVF repair, the best conditions for correct healing should be obtained, aiming at treating the underlying disease and providing best tissue support with minimal inflammation. Therefore, the patient should benefit from optimal nutritional status, smoking should cease, ²³ remission of CD should ideally be achieved for a minimum of 3 months, and the perineal CD activity should be minimized with any local septic complications controlled. The decision to attempt repair should be offered to the patient following joint consensus between the gastroenterology and surgical teams within the setting of an inflammatory bowel disease multidisciplinary meeting.

How to Minimize the Anorectal CD Activity?

Antibiotics or Not?

Most of the literature in the field does not directly address CD-related RVF but perianal fistula disease in CD. Briefly, once septic complications of CD have been ruled out, antibiotics have been assessed for their potential effect of decreasing perineal CD activity. Bernstein et al showed promising results of the use of metronidazole with improvement of fistula-related symptoms in 21 patients with CD-related perineal disease. Ten out of 18 patients with maintenance therapy had complete healing of fistulas. ²⁴ Brandt et al described that 10 out of 26 patients had complete healing of CD-related perianal fistulas after treatment with metronidazole, but documented neurologic side effects (paresthesia) and recurrence of perianal disease when the drug was discontinued. ²⁵ This was similarly observed by other teams. ²⁶ Dejaco et al described an 8-week regimen of ciprofloxacin and metronidazole ± azathioprine in 52 patients suffering from perianal CD. Twenty-six patients (50%) showed improvement and 25% complete healing after 8 weeks according to the perianal disease activity index. At week 20, 35% of patients had a favorable response and 18% maintained complete healing. ²⁷ Interpretation of these studies is limited by their small sample sizes and the absence of a control group without antibiotics.

Maeda et al randomized 74 patients to receive topical metronidazole ointment or placebo for 4 weeks and showed a significant reduction in the perianal disease activity index in the topical metronidazole group. ²⁸ In another randomized controlled trial (RCT), Thia et al showed that patients receiving ciprofloxacin had a 40% response rate versus 12.5% for patients receiving placebo. ²⁹ Further, West et al undertook an RCT comparing patients receiving three courses of infliximab with or without ciprofloxacin for 12 weeks. They showed that patients receiving ciprofloxacin had a 73% response rate compared to 39% response in those without antibiotics, but this difference did not reach statistical significance. ³⁰ Dewint et al performed a similar RCT but compared patients receiving adalimumab for 24 weeks with or without ciprofloxacin for 12 weeks. They found a significant difference between the intervention and the control group, with 71 versus 47% response rates, respectively. ³¹

A systematic review of medical therapies in perianal CD concluded that the evidence supporting addition of antibiotics to anti-tumor necrosis factor (TNF) therapy was of low quality. ³²

Which Immunomodulator to Choose?

Tacrolimus

Tacrolimus is an immunosuppressive drug initially obtained from the bacterium Streptomyces tsukubaensis , which inhibits calcineurin, thereby decreasing the production of interleukin-2 and T-cell activation. Tacrolimus was investigated for its effect in inducing remission of CD, ³³ but the evidence is poor regarding perianal CD. In a multicenter RCT, Sandborn et al randomized 42 patients with CD-related perianal fistulas to receive either tacrolimus or placebo for 10 weeks. They showed that tacrolimus led to initial fistula closure in 50% of patients and maintenance for at least 4 weeks in 9 out of 20 (45%) patients in the tacrolimus group versus 2 of 22 (9%) patients receiving placebo. ³⁴ However, the data were sparse and considered as being of low quality in a recent systematic review. ³² Another RCT by Hart et al did not find any benefit of topical tacrolimus in perianal fistulizing CD, but the study was underpowered. ³⁵ There are no data to specifically guide treatment of RVF with tacrolimus.

Azathioprine and Mercaptopurine

Azathioprine is converted to 6-mercaptopurine and inhibits purine synthesis, therefore leading to inhibition of synthesis of white blood cells. Azathioprine and mercaptopurine (collectively known as thiopurines) are currently used in the management of moderate to severe chronically active CD for the maintenance phase in corticosteroid-dependent patients. ³⁶ The effect of thiopurines on perianal CD has been investigated by several teams but all the studies had small sample sizes. In a RCT, Present et al demonstrated that 6-mercaptopurine allowed fistula closure in 31% of patients when compared to 6% of those taking placebo. ³⁷ Dejaco et al later showed that, among patients treated with ciprofloxacin and metronidazole, those receiving azathioprine had lower perianal activity index than those treated with antibiotics only. ²⁷ In a systematic review and meta-analysis focusing on fistulizing CD, Lee et al did not identify any improvement in fistula response in patients treated with azathioprine or 6-mercaptopurine when compared to placebo. ³² The main role of thiopurines in the management of perianal CD is now considered to be in reducing the risk of immunogenicity to anti-TNF therapy.

Anti-Tumor Necrosis Factor Alpha (Anti-TNF-alpha) Therapy

TNF-alpha is an inflammatory cytokine produced by white blood cells that leads to activation of T lymphocytes and is released as a consequence of active CD. Several antibodies have been produced to bind and neutralize TNF-alpha in patients suffering from CD: infliximab is a chimeric human and murine monoclonal antibody, adalimumab is a recombinant human monoclonal antibody, and certolizumab pegol is also a human monoclonal antibody pegylated for a longer half-life. ¹ Of all the medical therapies used to treat fistulizing perianal CD, anti-TNF therapy is the class with the best evidence of efficacy derived from high-quality RCTs. It is notable that the introduction of anti-TNF therapy has been associated with a decrease in the incidence of fistulizing perianal CD. ^{3 4}

Present et al showed that patients with abdominal and fistulizing perianal CD receiving 5 mg/kg intravenous infliximab at week 0, 2, and 6 had 55% closure rate of fistulas, when compared to 13% in patients receiving placebo. ³⁸ In another trial, Sands et al included 282 patients with perianal fistulizing CD who also received 5 mg/kg intravenous infliximab at week 0, 2, and 6, before being randomized to receive either placebo or 5 mg/kg infliximab every 8 weeks until week 54. The authors demonstrated that, at week 54, 36% of patients in the intervention group versus 19% of patients in the control group had complete absence of draining fistulas. Further, the time to recurrence was significantly longer for patients who received infliximab (40 weeks) than for those receiving placebo (14 weeks). ³⁹ Subgroup analysis suggests that other anti-TNF therapies are also effective but these have not been subjected to trials focused on perianal CD.

Despite their demonstrable benefit in perianal CD, the effect of anti-TNF therapy appears more modest for CD-related RVF, perhaps due to poor vascularization of the rectovaginal septum and decreased penetration of the treatment into the diseased tissues. For instance, Parsi et al found that the rate of complete response to infliximab after at least 3 months of follow-up was significantly lower in 14 patients with RVF (14%) when compared to those with perianal fistulae (78%). ⁴⁰ Other authors also demonstrated poorer response to infliximab induction therapy for patients with CD-related RVF, ⁴¹ but in a subgroup analysis of their study, including 25 patients with RVF, Sands et al showed that 61 and 45% of CD-related RVF were closed at weeks 10 and 14, respectively, after three sessions of infliximab. Maintenance therapy with infliximab allowed induction of a longer remission of RVF (46 weeks) than placebo (33 weeks). ⁴² Thereafter, the same team demonstrated that maintenance therapy with infliximab did not lead to increased incidence of fistula-related abscess. ⁴³ Other authors have reported that immunomodulation does not seem to impact healing after surgical repair. ¹⁶

In a systematic review and meta-analysis, Kaimakliotis et al pooled 23 studies representing 137 patients suffering from CD-related RVF treated with anti-TNF therapy. There was high heterogeneity in the included studies in terms of intervention, outcome measurement, and follow-up, but the authors found that anti-TNF therapy for CD-related RVF led to complete response in 41% of patients, partial response in 22%, and no response in 37%. ⁴⁴

Anti-TNF therapy should be considered in patients with CD-related RVF particularly when local tissue inflammation precludes any surgical repair. If a RVF persists after medical treatment, attempt at surgical repair should only be performed when any inflammation has resolved. We note that fistula tracts might not heal completely after infliximab treatment, ⁴¹ therefore calling for a step-up approach toward surgery in the absence of active proctitis.

Other Biologic Therapies

Newer biologic therapies licensed in CD block integrin-mediated migration of leukocytes to the gut (vedolizumab) or the interleukin 12/23 pathway (ustekinumab). Neither has been prospectively assessed with regard to fistula healing in perianal CD. Post hoc subgroup analysis reported only in abstract form suggested a trend toward benefit of ustekinumab but this did not reach statistical significance. Vedolizumab seems not effective in this regard.

How to Operate?

The objective of surgical treatment for CD-related RVF is to provide relief of symptoms and improvement in quality of life while minimizing morbidity, notably avoiding infectious complications in the perianal area, soiling, and the need for proctectomy. Ultimately, surgery should respect the basic principles for fistula repair: complete excision of the fistula tract, mobilization of the adjacent tissue planes, tension-free closure, and appropriate timing for repair (minimal inflammation). ⁴⁵ As previously mentioned, surgery should be delayed until CD inflammatory activity (particularly in the anorectum) is controlled and then planned during a period of remission. Repair should not be attempted if any evidence of active inflammation is found during clinical examination or investigation. ⁴⁵

What Surgical Treatment in the Acute Phase?

In the acute phase of perianal CD, consensus has been reached to treat the local sepsis and to drain any abscess before performing direct repair of the fistulizing disease, and also prior to commencing any immunosuppressive drugs. Standard treatment for ensuring ongoing drainage of the fistulizing tract is to use a seton. If perianal CD is very extensive, it is possible to perform a diverting stoma in order to improve quality of life and decrease local fecal contamination.

Seton Placement

A draining seton consists of a thin silastic band or a surgical suture that is inserted into a fistula tract under direct visual and/or anorectal ultrasound guidance. The objective is to provide selective drainage of the fistula tract and avoid further abscess formation. This will allow fibrosis of the tract, helping for better identification and subsequent removal during future definitive treatment.

Use of a long-term seton was shown to be safe and not lead to increased fecal incontinence in patients suffering from perianal CD. ⁴⁶ Unfortunately, 25 to 44% of patients with a seton in situ were reported to develop recurrence of perianal sepsis. ^{46 47}

de Groof et al performed a systematic review and meta-analysis comparing seton drainage and biologic therapy for perianal CD. Studies including seton drainage alone reported complete closure of the fistulas ranging from 13 to 100% across studies, whereas studies with biologic therapy reported complete closure ranging from 17 to 93% in 1,149 patients. When performing subgroup analysis including only RCTs (4 studies, 1,028 patients), the authors showed no difference between groups in terms of partial or complete closure, except in subgroup analysis including only studies with follow-up > 4 weeks which showed advantage for biologic therapy. ⁴⁸

Best response to seton placement was reached in combination with biologic therapy. For instance, Regueiro and Mardini reported that patients who underwent seton placement in addition to infliximab had better complete response, lower recurrence rate, and longer time to recurrence than patients receiving infliximab alone. ⁴⁷ Another group demonstrated that patients evaluated for perianal CD required on average four to five separate setons and nine infliximab infusions to obtain satisfying results with improvement in quality of life. Setons could be removed in 78% of patients during a mean follow-up of 12 months. ⁴⁹

Some authors have reported that RVF was a poor predictor of success for combined treatment of seton placement and infliximab therapy in patients with CD, ⁵⁰ but studies specifically investigating seton placement for CD-related RVF are lacking. The role of seton in the management of RVF due to CD remains unclear.

What Repair to Propose When the Acute Phase is Over?

Once the acute phase of perianal CD has passed, the objective is to close the fistula with minimal recurrence rate and functional disturbance while providing definitive healing of the perianal area. However, in patients suffering from CD, the quality of the surrounding tissues is an important predictor of the success rate of surgical procedures aiming to treat perianal fistulas. Therefore, surgical treatment of perianal fistulas should be tailored to the local extent and activity of CD. ¹⁸ However, to date, no clear recommendation exists for CD-related RVF.

The anatomical considerations related to RVF render healing of RVF more difficult to achieve, as there is less vascularized surrounding tissue to support the healing process. Therefore, RVF are challenging to treat and have higher rates of recurrence than other CD-related fistulas. Briefly, techniques for repair may include use of tissue glue and plugs, advancement flap through anal, vaginal, or perineal approaches, tissue transposition techniques involving the bulbocavernosus muscle/labial fat pad (Martius flap) and omentum, gracilis, and gluteus maximus muscles with or without possible implantation of a bioprosthesis, the injection of stem cells, and sphincteroplasty/episioproctotomy.

Techniques for the Rectum with Relatively Mild Disease

Fibrin Glue

Fibrin glue for the treatment of perianal fistula is injected into the fistula tract with closure of the internal opening, generally after seton removal. Fibrin glue instillation has shown some promise for perianal fistula treatment, especially in patients with a single tract. ^{51 52} In a multicenter RCT, Grimaud et al randomized 77 patients suffering from CD to receive fibrin glue instillation or not after removal of seton. The authors reported that fibrin glue instillation allowed clinical remission at 8 weeks in 38% of patients treated with fibrin glue instillation versus 16% of patients in the control group. ⁵² However, in patients with RVF, clinical remission was observed in none of three patients in the fibrin glue group, compared with one of four patients in the control group. ⁵² Further, Abel et al reported encouraging results of fibrin glue instillation in some patients with RVF, but not in those suffering from CD. ⁵³

Fistula Plug

The mode of action of fistula plug is similar to that of fibrin glue, as it consists of occluding the fistula tract and providing a scaffold for healing. Several authors reported anal fistula closure rates up to 85% in patients suffering from CD. ^{54 55} A systematic review including 12 studies representing 84 patients reported that fistula plug allowed closure of the fistula tract in 58% of patients suffering from CD-associated anal fistula. The success rate was 40% in patients with recurrent fistulas. ⁵⁶ To our knowledge, only a small cohort study investigated the fistula plug in patients suffering from CD-related RVF. The authors reported a long-term success rate of 44% (for a median follow-up of 118 weeks). ⁵⁷ However, encouraging results reported by a small series using the fistula plug should be weighed against the most recent literature in the field. For instance, in a multicenter RCT comparing the plug technique with advancement flap surgery for trans-sphincteric cryptoglandular fistulas in 94 patients, Bondi et al showed that patients who underwent the plug procedure had higher recurrence rate (66%) within a median follow-up of 12 months when compared to patients who underwent the flap procedure (38%). ⁵⁸ We believe that this technique has clear limitations due to the very short tract seen in RVFs and the poor quality of surrounding tissues making the normal use of plugs even more challenging.

Fistulectomy

Fistulectomy has been advocated by some groups for treating superficial fistulas in CD. However, failure of the treatment occurs in up to 40% of patients ⁵⁹ and an important fecal incontinence rate has been documented. ⁶⁰ Fistulectomy associated with sphincteroplasty showed some success in small studies including incontinent patients with CD-related RVF, ⁶¹ but, due to anatomical considerations, we would not routinely recommend fistulectomy for CD-related RVF.

Episioproctotomy

Episioproctotomy consists of a transperineal repair with layered closure. Briefly, a curvilinear incision is made around the anus or a linear incision made in the perineum, the sphincters are mobilized, the rectovaginal septum is dissected, and the external sphincter muscles are approximated before closure. Hull et al compared the outcomes of 50 patients who underwent episioproctotomy with those of 37 patients who underwent rectal advancement flap for non-CD-related RVF. They reported that 78% of patients healed after episioproctotomy, which was similar to the healing rate of patients who underwent rectal advancement flap. However, fecal and sexual functions were better after episioproctotomy. ⁶² These data are limited to a single-center study without widespread reported use elsewhere. Further, this technique has not been attempted in patients suffering from CD-related RVF, so would not be routinely recommended in this setting.

Advancement Flaps

Local flap repair is generally performed through a rectal or vaginal approach ( Fig. 1 ). The creation of a mucosal, submucosal, or full thickness flap allows for coverage of the fistula internal opening and for provision of healthy tissue for healing if surrounding tissues are of unsatisfactory quality. The flaps can be curvilinear, linear, or a sleeve advancement.

Fig. 1.

Repair of the rectovaginal fistula through a vaginal approach. (image courtesy of Dr S. Steele).

Sliding rectal flap repair was first reported by Noble in 1902. It consisted of a full thickness flap involving the mucosa, the submucosa, and a portion of the internal sphincter muscle. The internal sphincter was approximated on the fistula tract, the flap was advanced down and sutured in place. This was changed to a curvilinear flap for better vascularization of the flap tip. Of note, dehiscence due to tension and/or poor vascularization might lead to RVF recurrence. ¹⁰ Hull and Fazio previously described that the severity of inflammation of the rectum influences the type of flap to choose for treating CD-related RVF, although it is now accepted that medical therapy should be optimized to treat any local CD inflammation prior to surgery for RVF: curvilinear advancement flap was advised for patients with minimal anorectal disease activity, linear advancement rectal flap for long fistula or significant ulceration, and advancement sleeve flap for minimal to mild disease with severe anal ulceration and poor tissues. ⁸ Using various flap techniques, the authors reported a 68% success rate in 35 patients with and without fecal diversion. ⁸ Briefly, for the curvilinear advancement flap, the flap is mobilized in a curvilinear fashion starting at the dentate line, the fistula tract undergoes curettage and is closed in layers, the distal diseased portion of the flap is removed, and the remaining flap is sutured back. The vaginal mucosa can be left open. A full thickness flap consisting of mucosa, submucosa, and some muscle is brought down. ⁸ Hull and Fazio reported that 13/24 patients healed after curvilinear advancement flap, and two more after a redo procedure, representing a 67% healing rate. ⁸ Linear advancement rectal flap is used for a long fistula that would require excessive resection of the diseased part of the flap. The fistula is excised perpendicular to the dentate line, closed in layers leaving the vaginal side open, and mucosa and submucosa are mobilized laterally and closed together. A 50% success rate was achieved in a small group of six patients. ⁸ Advancement sleeve flaps are reserved for patients with severe anal disease but spared rectum. Mucosectomy of any diseased mucosa is performed circumferentially below the dentate line, exposing the sphincters, until reaching the supralevator space with minimal rectal mobilization. The fistula tract is cored out and closed, and the rectum is pulled down after sufficient mobilization and sutured to the anoderm without tension. ^{8 63} Out of five patients, Hull and Fazio reported three successes and one further success after redo. ⁸ Marchesa et al performed rectal sleeve advancement flap in 13 patients suffering from CD-related perianal disease, including 11 patients with RVF. They reported healing without the need for additional procedure in five of the patients (38%) after a follow-up of 1 year. ⁶⁴ Advancement flaps can also be considered in patients with an anal stricture. ⁸ Other groups have reported favorable outcomes using rectal advancement flaps for CD-related RVF. ^{65 66}

The vaginal approach offers better visualization of a RVF than the rectal approach. Further, the vagina constitutes a lower pressure cavity than the rectum and its more lax tissue allows a higher chance of proceeding to a tension-free repair. It might be more adaptable to CD-related RVF repair as the surrounding vaginal tissue, which serves as a support to the sutured flap, is generally not involved in the CD-related inflammatory process. Briefly, a curvilinear incision is performed below the fistula opening, the flap is raised, the fistula is excised, the rectal opening is closed, the sphincter approximated, the flap sutured in place, and a stoma may be considered by some authors. ⁶⁷ Bauer et al reported the technique to be successful in 12 out of 13 patients with CD-related RVF after an average follow-up of 50 months. ⁶⁷ The same team then published the cohort with longer follow-up and showed one patient with a RVF remained unhealed after a mean follow-up of 55 months. ¹¹

Ruffolo et al performed a systematic review and meta-analysis aimed at comparing the vaginal and rectal approaches to repair of RVF in patients with CD. Eleven observational studies were included for a total of 219 flap procedures. The pooled rate of primary fistula closure was 54% (range 33–100%) after the rectal approach and 69% (range 0–92.9%) after the vaginal approach, without significant difference. ⁶⁸

Techniques for Rectum with Advanced Disease

Advanced rectal disease, with tissues surrounding the RVF being fibrotic or of poor quality and also a large RVF orifice render repair with advancement flap unlikely to be successful. Therefore, tissue interposition techniques have been developed. The objective is to bring healthy vascularized tissue between the vagina and the rectum to close the fistula tract and separate the two mucosal aspects with a healthy scaffold to facilitate healing. The first muscle transposition was reported by Byron and Ostergard in 1969 in a patient with radiation-induced RVF. ⁶⁹ It has been suggested that tissue interposition techniques may require fecal diversion for optimal healing, ⁷⁰ but in our practice it is not seen as mandatory. Alternatives have been more recently developed and include interposition of a bioprosthesis, which is not vascularized but might offer some similar benefits. In a cohort of 99 patients with minimal follow-up of 3 years who underwent different surgical procedures for CD-related RVF, best results were achieved with tissue interposition techniques, but patient heterogeneity limits comparison between the techniques. ¹⁶

Muscle Transposition

Several muscles from the local area can be transposed into the rectovaginal septum to facilitate fistula closure as well as a vascularized scaffold for healing, which is of importance if surrounding tissues are of low quality and fibrotic. The bulbocavernosus, gracilis, and gluteus maximus muscles have been described for use in this setting.

The evidence regarding skin-sparing gracilis muscle transposition in patients with CD-related RVF mostly arises from small series of patients in otherwise heterogeneous cohorts, which render firm conclusions difficult to draw. ^{71 72 73 74} Underlying CD is associated with a higher failure rate. ^{73 74} Wexner et al performed gracilis muscle transposition in 15 patients with RVF, of which approximately 50% were due to CD. All patients had fecal diversion. In patients with CD-associated fistulas, three healed and did not recur after stoma closure. ⁷¹ Among the six patients with CD-related RVF included by Kersting et al, gracilis muscle transposition only had a 50% success, when compared to the 85% healing rate in patients without CD. ⁷⁴ To our knowledge, only three studies investigated gracilis muscle transposition in cohorts containing only CD-related RVF. Fürst et al performed gracilis muscle transposition in 23 patients with CD-associated RVF. Thirteen could undergo stoma reversal due to favorable healing. ⁷⁴ Korsun et al performed gracilis muscle transposition in 21 patients with CD-related RVF. They described a RVF closure rate of 71% and a stoma reversal rate of 55%. ⁷⁰ Fürst et al performed gracilis muscle transposition in 12 patients with recurrent CD-related RVF. Diverting ileostomy was performed in all patients. Only one RVF recurrence was reported at a mean follow-up of 3.4 years. ⁷⁵

The Martius flap was first described in 1928. It consists of the transposition of the bulbocavernosus muscle, or later modified to the labial fat pad (modified Martius flap) graft into the rectovaginal septum. The Martius flap has been used successfully in patients with RVF and CD ^{76 77} and a success rate of up to 50% has been reported. ⁷⁸ Persistent or recurrent fecal incontinence and dyspareunia may result in view of the underlying perineal graft harvest and repair. ⁷⁶ As an alternative to other muscle transposition, Onodera et al reported bilateral gluteus maximus muscle patching by approximation of the muscle to bolster repair of RVF in four patients with RVF caused by obstetric injury or non-CD-related fistula. ⁷⁹

Omental Interposition

Omental interposition was proposed as an alternative to muscle transposition to bring a well vascularized healthy tissue into the scarred area of RVF, in order to help repair and healing. Briefly, a laparoscopic anterior dissection and a transperineal dissection of the rectovaginal septum are simultaneously carried out. An omental flap retaining blood supply from the left gastroepiploic artery is then brought down in into the rectovaginal space and secured. ⁸⁰ de Bruijn et al performed omental interposition under the protection of a diverting stoma in seven patients suffering from non-CD-related RVF. The authors reported that none of the patients had a fistula recurrence after a median follow-up of 15 months. ⁸⁰ Similarly, van der Hagen et al performed omental interposition in a heterogeneous cohort of 38 patients (including CD) without systematic fecal diversion. Only two patients (5.3%) developed a recurrence within a median follow-up of 28 months, but we note that one patient required redo surgery for necrosis of the omental flap. ⁸¹ Other teams have reported the technique in association with resectional surgery. ⁸² Omental interposition seems to be a promising technique for treating RVF, but needs more evidence, notably including patients suffering from CD-related RVF and documenting longer follow-up. Moreover, interposition techniques including an abdominal approach are more suited to higher RVF.

Interposition of a Bioprosthesis

Interposition of a biological mesh between the vagina and rectum as an alternative to tissue transposition has been proposed to treat RVF by closing the fistula tract even in fibrotic areas and to provide a scaffold for healing. ^{83 84 85} After a preliminary report, ⁸⁶ Schwandner et al implanted a biological mesh made from porcine intestinal submucosa (Surgisis, Cook Medical, USA) placed transvaginally after transrectal and transvaginal excision of RVF in 21 patients, most of them having recurrent RVF with failure of previous treatment. Nine of 21 patients from the cohort had CD. The authors reported a success rate of 71% at 12 months. ⁸⁷ Further, Ellis reported 27 patients with RVF who underwent intersphincteric interposition of a Surgisis (Cook Medical, USA) biological mesh, sutured to the levator ani muscles laterally and to the external sphincter distally to prevent migration. Two patients had CD. Five patients (19%) had recurrence after a mean follow-up of 12 months. ⁸⁸ Despite poor evidence in patients with CD-associated RVF, this technique might constitute an option for these patients. One major limitation of this technique is that, in cases of recurrence, local tissues become very fibrotic and thus increase the difficulty of a redo repair. ⁸⁹

To Divert or Not to Divert?

In cases of severe CD ⁹⁰ or of complex repair of perianal fistula, fecal diversion helps to reduce the bacterial load in the perianal area and to improve conditions for healing. Fecal diversion might constitute the sole treatment for perianal CD and RVF, ⁹¹ or can be performed as an adjunct to attempted repair. ⁶⁷ In a systematic review and meta-analysis including 16 studies (556 patients), Singh et al showed that temporary fecal diversion performed for refractory perianal CD contributed to reaching clinical response in 64% of patients. ⁹² Although there is no consensus regarding the indication to perform fecal diversion during RVF repair surgery, which is left to the discretion of the surgical team, ^{8 93} Bauer et al reported a 92% success rate in a cohort of patients who underwent a transvaginal repair associated with systematic fecal diversion. ⁶⁷ Fecal diversion can be temporary and reversed when local conditions have improved, or definitive in cases of severe refractory perianal CD and/or in patients suffering from fecal incontinence. The proportion of patients requiring permanent fecal diversion was 31% in patients suffering from perianal CD and 54% in patients suffering from CD-related RVF. ⁹¹ In their meta-analysis, Singh et al documented that only 34.5% of patients attempted restoration of bowel continuity after fecal diversion for control of severe perianal disease, which was successful in only 17% of them. ⁹²

When to Perform Proctectomy?

In early reports, proctectomy was required in as many as 12 out of 28 (43%) patients with RVF arising during the course of CD. ⁹⁴ Nowadays, improvements in medical treatment has reserved proctectomy to cases when other techniques have failed, and additional treatment is needed in cases of refractory severe disease, especially if associated with fecal incontinence. ⁸ Overall attempts at RVF surgical repair only have a 53% success rate at 60 months and may take several attempts to be successful, with proctectomy still required in 22% of patients. ¹³ The risk of an unhealed perineal wound being symptomatic after proctectomy should always be discussed with the patient.

What is the Place for Innovative Therapies?

Stem Cell Transplantation

Adipose-derived stem cells have the ability to differentiate into muscular cells ^{95 96} and have therefore been proposed as an adjunct to treat perianal fistulas. In a preliminary safety study, fertility and pregnancy outcomes were not affected by local treatment of CD-related perianal fistulas using adipose-derived stem cells, including three patients with RVF. ⁹⁷ In a multicenter prospective cohort study, de la Portilla et al treated 24 patients suffering from CD-related perianal fistulas with one or two injections of expanded adipose-derived allogeneic mesenchymal stem cells in addition to standard care. The authors showed that 69% of treated patients had a reduction in the number of draining fistulas and 56% achieved complete fistula closure at week 24. ⁹⁸

Garcia-Olmo et al performed a RCT investigating treatment with fibrin glue associated with adipose-mesenchymal stem cells, and confirmed the benefit of precursor cells, showing complete healing of perianal fistulas (including some with CD) in 71% of patients versus 16% with fibrin glue alone. ⁹⁹ In a multicenter double-blinded RCT including 212 patients with CD perianal fistulas, Panés et al showed that a significantly greater proportion of patients treated with adipose-derived mesenchymal stem cells versus placebo-achieved combined remission (50% vs. 34%) at week 24 and 52. ¹⁰⁰

Therefore, despite evidence lacking in CD-related RVF, we could imagine that adipose-derived mesenchymal stem cells may constitute an adjunct to conventional surgical repair ¹⁰¹ and this approach deserves future study despite concerns over potential cost.

Hyperbaric Oxygen

Hyperbaric oxygen has been shown to be a useful adjunct to stimulate healing of difficult, hypoxemic, and scarred tissues. Hull reported that hyperbaric oxygen improved healing of CD-related RVF after 40 sessions. ⁸⁹ However, this technique, which could be used in association with conventional attempts at surgical repair, is logistically demanding and costly, thus potentially limiting its indication.

CO ₂ Laser

A sole case report has shown improvement of symptoms and size of a CD-related RVF in a patient refractory to infliximab who underwent intravaginal CO ₂ laser treatment over three sessions. ¹⁰² The data are too limited to be able to make any firm recommendations.

Conclusion

RVF constitutes one of the most difficult manifestations of perianal CD to treat, has high rates of recurrence and symptoms that can be very debilitating for patients. Of note, the decreased vascularization and the thin rectovaginal septum do not provide optimal conditions for healing. Treatment should be medical initially, aimed at inducing disease remission and providing optimal conditions for healing, and subsequently surgical to close the fistula tract and reduce the likelihood of recurrence. Several authors or professional societies have attempted to propose an algorithm to choose the most suitable treatment. ^{18 103} However, there is no high quality evidence on the best surgical repair procedure, as treatment is usually multimodal, authors report small cohorts that are often heterogeneous in terms of included patients, treatments, outcome measures, and follow-up, and mostly reflect the personal experience of the authors. Nevertheless, we propose a therapeutic management algorithm in Fig. 2 . Briefly, in cases of sepsis, seton drainage should be performed and antibiotics given if there is associated cellulitis. Surgical repair should only be considered once local inflammation and/or sepsis has resolved, and adapted to the anatomy of the fistula, to its size, and to the quality of surrounding tissues. In cases of complex redo repairs, fecal diversion should be considered. If the local disease if too advanced and symptomatic despite defunctioning, proctectomy should be performed.

Fig. 2.

Decision-making algorithm for Crohn's disease-related rectovaginal fistula treatment. CD, Crohn's disease; RVF, rectovaginal fistula.

Funding Statement

Funding None.