Rectovaginal fistula (RVF) accounts for less than 5% of all perianal fistulas, but leads to not only physical problems, including inflammation and irritation, but also emotional distress and social and sexual dysfunction. Previous reports found birth injury to be the most common cause of RVF accounting for 60%–80%. It may result from a third- or fourth-degree perineal laceration during vaginal delivery, and anal incontinence can develop, as with sphincter injury. Perianal infection or prior perianal surgery can be a cause of RVF. Inflammatory bowel diseases are also causes of RVF and should be suspected when a fistula is complex or recurrent. Pelvic irradiation is not an uncommon cause in cancer patients. Because cancer invasion can also develop a fistula, evaluating tumor recurrence at the fistula site is important in these patients [1234].

Numerous surgical procedures have been developed to resolve RVFs; the approach chosen in each case depends on fistula size, the presence of a sphincter defect, the existence of combined fistulas, and the causes of the disease. However, the location of the fistula opening is the most important feature to consider when selecting an appropriate procedure. In a simple low fistula, transvaginal or transanal local repair can be performed and are safe and feasible methods, though the transanal rectal advancement flap is more familiar to colorectal surgeons. More complex procedures, including gracilis muscle transposition and rectal sleeve advancement flap repair, may be necessary to repair complex fistulas. Often, high fistula should be treated with abdominal operations, such as proctectomy or diversion colostomy [56].

The purpose of this study was to investigate the outcome of surgical treatments for RVF according to the various causes and procedures and to analyze the prognostic factors for successful RVF closure.

Between 1998 and 2016, 92 patients underwent 128 operations for RVF at our institute. We prospectively registered these patients and retrospectively reviewed the medical records. This study was approved by the Institutional Review Board of Seoul National University Hospital (approval number: H-1205-107-410).

Diagnosis was based on patient history, digital rectal examination, and vaginal examination, as well as radiological findings, such as CT or fistulogram. Low fistula was defined as a fistula located from near the dentate line to the posterior vaginal fourchette and around the anal sphincter complex and was easy to detect. High fistula was defined as a fistula located from near the peritoneal reflection to the cervix and was detected, with difficulty, by digital examination through the anus or vagina. A sphincter defect was evaluated preoperatively by digital rectal examination or transrectal ultrasonography. Operative methods were selected according to disease cause, location of the fistula opening, and patient clinical condition. Successful closure was confirmed based on patient symptoms, digital rectal examination, or fistulogram during follow-up. The risk factors for failure of closure, which are fistula continuance or recurrence, were also analyzed.

Statistical analysis was conducted using IBM SPSS Statistics ver. 21.0 (IBM Co., Armonk, NY, USA). Student t-test was used to compare the averages of variables. Univariate analysis was performed using Pearson chi-square test and Fisher exact test. Multivariate analysis was performed by logistic regression. In statistical analyses, P < 0.05 was regarded as indicative of significance.

RVFs have various causes, the most common of which were found to be RT and pelvic operation in our study. Fistula location differed according to the causes, and low fistula was more common. A variety of operations was performed in patients with persistent or recurrent RVF, and complex procedures were more frequent with repeated operations. The overall successful closure rate was 42.4%, and better outcomes were achieved for RVF from birth injury or perineal operation. However, RVF from RT, pelvic operation, or Crohn disease presented poor prognosis. Although diverting ostomy alone could close some of the RVFs from trauma or pelvic operation, it was not found to be effective for RVF from RT or cancer invasion.

In this study, RT for cancer was the most common cause of RVF. We assumed that because our hospital is a tertiary hospital, there may have been more patients with causes of RT or pelvic operation from cancer than at other centers. Radiation damage is thought to be related to radiation dose and the fistulas induced by RT are difficult to heal via local repair because irradiated tissue exhibits friability or ischemic fibrosis. Combined fistulas, such as enterovaginal, vesicovaginal, and rectovesical fistulas, are also common among recipients of RT. For the 2 patients in our study who had multiple fistulas, the cause was RT. Resection of the affected bowel or diverting colostomy is frequently required in these patients. Unfortunately, colostomy may be permanent because spontaneous fistula closure is rare [78]. Most of the patients who received RT had high fistulas and underwent diverting ostomy alone in this study. For the patient who was treated with diverting ostomy alone, it was not possible to take down the ostomy during the follow-up period. Successful closure was achieved in only 2 patients (6.5%) with low fistulas after transanal rectal advancement flap and gracilis muscle transposition.

Pelvic operation with rectal resection can lead to iatrogenic injury of the rectovaginal septum due to the double-stapling technique, or develop postoperative leakage at the anastomosis site. RVF from pelvic operation with rectal resection is also difficult to close, and complex procedures may be needed [910]. Among our patients, successful closure was achieved in only 23.4% of cases, despite repeated operations from the diverting ostomy alone to gracilis muscle transposition.

Birth injury is known to be a common cause of RVF, but the number of RVFs resulting from this cause has been declining due to a decrease in the number of severe obstetric perineal lacerations resulting from the minimal use of episiotomy and operative vaginal delivery [11]. Most fistulas resulting from birth injury can be successfully treated by transanal local repair with rectal advancement flap, and sphincteroplasty or perineoplasty may be necessary in some cases involving sphincter injury [12]. In this study, birth injury was the third most common cause, and transanal rectal advancement flap or sphincteroplasty with perineoplasty was performed in most patients. The success rate was 87.5%, representing the best results achieved for all fistula causes. Transvaginal approaches may be better in terms of exposure and are easier to perform by obstetric and gynecologic surgeons, and success rates are reportedly as high as those achieved using the transanal approach [1314]. Although insufficient studies comparing these 2 approaches have been conducted, it is assumed that because the fistula opening in the anus represents the site experiencing the greatest pressure, closure at the anal site is important for achieving successful outcomes [1516].

Inflammatory bowel diseases are common causes of RVF. Although transanal local repair may be acceptable in patients with Crohn disease without rectal inflammation, the recurrence rate is reportedly as high as 25%–50%, and further complex procedures are likely necessary [1718]. Although combined medical treatments using anti-TNF-α and infliximab are promising, these treatments have not yet been found to be satisfactory in terms of long-term outcome [1920]. Recently, adipose-derived stem cells, representing a novel treatment, have been used for recurrent Crohn anal fistula and are reportedly safe and feasible [212223]. We performed stem cell injections in 3 patients, and complete closure was achieved in 2 of them (66.7%). Although the long-term effects of stem cells have not yet been determined, this treatment may also represent an effective option for RVF treatment. Fistula after IPAA for UC is also difficult to manage due to the high recurrence rate after local repair [24]. Only 2 patients (28.6%) with pouch vaginal fistulas achieved successful closure despite repeated operation, including pouch excision and reformation of the IPAA in 1 patient.

Although numerous surgical procedures have been developed to treat RVF, no gold standard procedure currently exists. The choice of procedure should depend on the clinical features of the patients, including fistula cause. The use of an individualized approach involving a precise surgical technique can yield successful closure of RVF [115]. The success rates obtained in previous studies have been diverse, ranging from 30% to approximately 90%, because the patients presented with various causes of RVF and a variety of surgical procedures were used [252627]. The operative treatment for recurrent RVF can be more complex, and the results of local repair with rectal flap have been disappointing [28]. The use of a muscle transplant, such as gracilis muscle transposition, has been attempted for fistula repair, and an additional variety of procedures, such as rectal sleeve advancement flap, Martius flap, proctectomy, and diversion colostomy, may be necessary [252930]. Recently, the aggressive use of diverting ostomy and major procedures has been suggested [26]; however, in this study, we tended to select gracilis muscle transposition or Martius flap for repeated operations, with success rates of 42.9% and 33.3%, respectively. Although these success rates were not as high as that obtained with transanal rectal advancement flap, we considered that the success rate may not depend on the procedures used but on the cause of RVF, which were mainly RT, pelvic operation, and Crohn disease in the present study. Achieving closure of recurrent RVF is important, and surgeons should aim to improve the outcomes of operative treatments for RVF even in patients who undergo repeated complex procedures.

This study has the limitations associated with a retrospective review, and the clinical features of RVF resulting from individual causes were not fully represented due to the small number of patients involved. As the number of patients with each cause of RVF was also small, we could not analyze the separate results from the homogenous populations with the same causes of fistula. Further research with a large number of patients is necessary. Also, the follow-up periods were considered insufficiently long to evaluate the outcomes; however, the patients who were treated successfully for simple low fistula in our study did not need to return to the hospital for a long time, and the patients who had fistula associated with RT or cancer invasion may have died early due to their underlying disease. Despite the lack of high-quality prospective studies evaluating the outcomes of surgical treatment for RVF, we verified that the RVF cause may be the most important prognostic factor and that diverting ostomy alone was sufficient to close an RVF resulting from RT or pelvic operation.

In conclusion, RVF results from various causes and can be treated successfully by a variety of surgical procedures. The treatment outcome differed according to clinical presentation, fistula cause, and operative procedure, and the cause of RVF was the most important prognostic factor for successful closure. RT, pelvic operation or Crohn disease is increasing at tertiary hospitals and the surgical procedure involving diverting ostomy alone is not enough in these high-risk fistulas. Taking an individualized approach is necessary for achieving successful outcomes, and surgeons should attempt to close fistulas with repeated procedures, including complex or novel procedures, even for persistent or recurrent fistulas with a high risk of causes of RVF.