International Journal of Colorectal Disease

, 24:1255

Interposition of vital bulbocavernosus graft in the treatment of both simple and recurrent rectovaginal fistulas

Authors

    • Department of Colorectal SurgeryShanghai XinHua Hospital Shanghai JiaoTong University School of Medicine
  • Dawei Chen
    • Department of Colorectal SurgeryShanghai XinHua Hospital Shanghai JiaoTong University School of Medicine
  • Wei Chen
    • Department of Colorectal SurgeryShanghai XinHua Hospital Shanghai JiaoTong University School of Medicine
  • Honghua Jiang
    • Department of Colorectal SurgeryShanghai XinHua Hospital Shanghai JiaoTong University School of Medicine
Review

DOI: 10.1007/s00384-009-0720-4

Cite this article as:
Cui, L., Chen, D., Chen, W. et al. Int J Colorectal Dis (2009) 24: 1255. doi:10.1007/s00384-009-0720-4

Abstract

Objective

The objective of the study is to assess the efficacy of vital bulbocavernosus graft transposition in the treatment of rectovaginal fistula.

Materials and methods

From March 2003 to October 2007, nine consecutive patients diagnosed with rectovaginal fistula were refereed to our institute. All patients were treated using an interposing vital bulbocavernosus graft between rectum and vagina.

Results

Median patient age was 33 years (range, 19–61) and seven of the nine patients had undergone between one and six fistula repair sessions prior to this grafting procedure. The etiology included congenital in three, surgery injury in four, obstetric in one, and radiation in one. No wound infections or abscesses occurred postoperatively, and the in-hospital mortality rate was zero. No recurrence was reported during the follow-up period and all patients had normal fecal continence. Only one patient had mild dyspareunia and no further surgical treatment needed.

Conclusion

Both simple and complex rectovaginal fistula can be reliably repaired using a bulbocavernosus graft.

Keywords

Rectovaginal fistulaBulbocavernosus flapTissue interposition

Introduction

Rectovaginal fistula (RVF), defined as abnormal anatomic communication between vagina and anorectum, is uncommon which accounts for 5% of anal rectal fistulas. However, it brings tremendous physical and psychosocial suffering to the patients. The etiology of this disease can be congenital but mostly acquired, such as trauma (especially obstetric injury) [1, 2], infection [3, 4], inflammatory bowel disease (IBD) [5, 6], tumor, and radiation [7], surgery (especially vaginal hysterectomy and low anterior resection) [8]. Regardless of the cause, RVF is troublesome for both patients and surgeons due to the high recurrent rates after surgical repair [9, 10], which may result from the insufficient blood supply or poor tissue condition at the diseased area. Although several surgical procedures have been proposed including direct repair, the Musset procedure, and procedures using vital tissue grafts, the treatment for RVF, especially for complex fistulas, still remains a clinical challenge.

Recently, more and more surgeons prefer to employ flap advancement or graft transposition between rectum and vagina to treat this disease because of the less failure rate of these procedures [1113]. Here we report our preliminary results from a series of nine RVF patients treated with interposition of vital bulbocavernosus graft, the Martius flap technique originally designed for repairing vesicovaginal communications.

Materials and methods

Patients

From May 2003 to July 2007, nine consecutive patients diagnosed as RVF were referred to the Department of Colorectal Surgery, Shanghai Xinhua Hospital in China and prospectively recruited into this study. This study was approved by the ethic committee in Shanghai Xinhua hospital. Signed consent form was obtained from each patient preoperatively.

The diagnosis of RVF was confirmed with digital rectal examination and vaginal scope in all patients. Anal rectal rest pressure, squeeze pressure, and rectal anal inhibitory reflex were measured by anal rectal manometry (Solar measurement system: Urodynamics and GI Motility, MMS Enschede, Netherlands), either preoperatively or postoperatively, to assess the preoperative anal sphincter function. The reference values of anal rectal rest pressure and squeeze pressure in our laboratory were 35–70 and 95-175 mmHg, respectively. Biopsies of rectal mucosa and the fistula margin were also obtained for pathological evaluation to confirm or exclude the underlying active IBD or malignancy. Pelvic computed tomography was also used to detect if there is any evidence of pelvic tumor. Demographic and clinical pathological data were collected from patients’ medical records.

Procedure

Mechanical bowel preparation with polyethylene glycol solution and oral nonabsorbable antibiotics (kanamycin 1 g tid po.) was given to the patients without previous bowel diverting. For patients with diverting stoma, irrigation of distal rectum was carried out with normal saline through the anal tube. Vaginal washing with normal saline was performed in all the patients just before the operation.

The surgical procedures were summarized as follows: patient was placed in lithotomy position, and a “U”-shaped skin incision was created between vagina and rectum. The rectum was separated from the posterior wall of vagina by electrocautery. Dissection did not proceed until beyond 3 cm upper opening of fistula within rectovaginal septum in order to facilitate a tension-free rectal closure and to create enough space to accept the interposing flap. The wound was washed, and homeostasis was achieved by using the electrocautery, and then the rectal and vaginal defect was closed, respectively. Another 3 cm length skin incision was subsequently made parallel to the one side of the labia majora to harvest the graft. The bulbocavernosus muscle was mobilized from the labial, preserving its posterior vascular pedicle. A subcutaneous tunnel connecting the labial and perineal incisions was created, and the graft transposed, ensuring proper orientation. The graft was loosely sutured in place over the rectal closure. Two drainage tubes were placed in the labial and vaginal incisions, respectively.

The first dose of prophylactic antibiotic (ceftriaxone, 1 g) was administered intravenously at the time anesthesia was induced, and then the same dose of the agent was administered after 12 h. The patients who had no colostomy refrained from any food for 3 days and then received elemental diet for another 4 days postoperatively before starting semisolid food. Six months after the surgery, the continuity of digestive tract was restored for the patients who had colostomy for fecal diversion. The patient was considered to be cured if there was no passage of intestinal waste through vagina, confirmed by vaginal scope. At least 1 month after the continuity of digestive tract was restored, the patients returned to have normal diet. Meanwhile, patients abstained from vaginal intercourse for 6 months at least.

Follow-up

Postoperatively, patients were followed up by clinic or telephone interview with specific questionnaires to collect the information about status of fecal control, flatus or fecal leakage from vagina, and quality of personal and social life. The interview was scheduled twice a month for the first 3 months, followed by once a month for 1 year and finally once a year. Follow-up study was carried out by the surgeons who took part in the operation.

Statistical analysis was performed with the two-tailed t test for paired samples (preoperation vs. postoperation).

Results

The median age of the patients was 33 years, ranging from 19–61 years. The etiology of RVF included congenital deformation in three, surgical injury in four, obstetric injury in one, and pelvic radiation in one (Table 1). No IBD or tumor was seen. Seven of the nine patients underwent fistula surgical repair for one to six times, and six of them received diverting colostomy after the failure of the first repair attempt (Table 1). All patients had symptom of passage of flatus or liquid stool through vagina, and five had signs of vaginitis. Openings of the fistula were all located above the anal sphincter and varied between 5 and 25 mm in diameter. The preoperative rest anal pressure, squeeze pressure and rectoanal inhibitory reflex were normal in all patients (Table 2).
Table 1

Patient characteristics and follow-up data of 9 patients

No. of patients

Age (years)

Multiparous

Etiology

Stoma

Repair session

Follow-up (months)

1

19

No

Congenital

No

0

14

2

21

No

Congenital

No

0

15

3

42

Yes

Anterior rectal resection for endometriosis

No

1

6

4

23

No

Congenital

Yes

6

48

5

27

No

Obstetric injury

Yes

1

12

6

33

No

Hysterectomy

Yes

2

12

7

57

Yes

Anterior rectal resection for rectal cancer

Yes

2

20

8

61

Yes

Radiation after anterior rectal resection for rectal cancer

Yes

3

18

9

52

Yes

Stapled hemorrhoidopexy

Yes

1

6

Table 2

Comparisons of patient’s anorectal manometry (mmHg) between preoperative and postoperative

No. of patients

Resting pressure

Squeeze pressure

Rectoanal inhibitory reflex

Preoperative

Postoperative

Preoperative

Postoperative

Preoperative

Postoperative

1

56

61

120

107

Yes

Yes

2

50

47

148

136

Yes

Yes

3

65

69

134

117

Yes

Yes

4

58

52

110

98

Yes

Yes

5

53

55

137

152

Yes

Yes

6

47

41

101

148

Yes

Yes

7

35

39

117

99

Yes

Yes

8

37

35

136

106

Yes

Yes

9

62

56

149

113

Yes

Yes

Ave

51

51

128

119

  

SD

10

11

17

21

  

P

0.584

0.343

NS

Results of the two-tailed t test for paired samples (preoperation vs postoperation). The reference values of anal rectal rest pressure and squeeze pressure in our laboratory were 35–70 and 95–175 mmHg, respectively

Symptoms occurring after the surgery included discomfort in the anus in all patients, tenesmus without diarrhea in three patients, and feeling of incomplete defecation in one patient. All these symptoms spontaneously disappeared in 2 weeks after surgery. Wound swollenness was seen in three patients on the third postoperative day, which was caused by hematoma in one patient and infection in two patients. However, there was no wound abscess and incision disruption, and all surgical wounds were healed in 10 days after surgery. Output of drainage tube was less than 10 ml during the first 24 h postoperatively, and the tube was removed within 48 h. No postoperative mortality was seen.

The median duration of follow-up was 14 months with a range of 6–48 months. Three patients without colostomy returned to normal diet 1 month after the surgery and had normal defecation of stool in 3 months. The other six patients underwent stoma closure 6 months after surgery and recovered eventually. They all had normal bowel movement subsequently. Only one patient complained of mild dyspareunia, which requires no further surgical management. The results of anal pressure measurement postoperatively showed no difference compared with preoperative ones (Table 2). By the time of the last follow-up, there was no recurrence of RVF, and all patients reported normal fecal continence and returned to a normal life.

Discussion

The most common etiologies of rectovaginal fistula have been reported to be IBD and obstetric or surgical injuries [9, 14, 15]. However, probably due to the extremely low incidence of IBD in China, no RVF caused by IBD was seen in our series. Moreover, there was no RVF related to tumor. Therefore, the etiology of our patients was relatively simple, which may contribute to the excellent outcome in these patients. Notably, one RVF patient was caused by procedure for prolapse and hemorrhoidectomy (PPH). Rectovaginal fistula has been reported to be an occasional complication after PPH [16] and occurred in one case (0.2%) in a series of 449 patients [17]. Attention should be also paid to low anterior rectal resection with assumption of safe procedure by using staplers due to the same anatomic reason [18, 19].

Patients with RVF are usually symptomatic. Various symptoms were seen in all of our patients. The clinical manifestation largely depends on the size, location, and etiology of the fistula. Discharge of intestinal waste from the vagina is the predominant symptom and discharge of malodorous vaginal may accompany with recurrence episodes of vaginitis. In our patients, the most common symptom was the passage of flatus or liquid through vagina, which is similar to those described in the literature [20]. Additionally, intestinal symptoms of diarrhea, rectal bleeding, mucous discharge, and abdominal cramping secondary to patients’ underlying disorder may dominate the clinical picture. With regard to the diagnosis of RVF, it is quite easy depending on digital examination and vaginal scope. More importantly, the underlying disorders such as IBD or tumor should also be diagnosed by the same time because these conditions can lead to repeated failure of a correct surgical procedure. Therefore, biopsy of the mucosa from the rectum or the margin of the fistula, pelvic CT, and colonoscopy should be considered mandatory prior to surgical repair.

Several surgical approaches have been introduced to treat rectovaginal fistula, such as simple repairs, rectal mucousal flap advancement, interposing a viable tissue between rectum and vagina [2123]. Local direct fistula repairs and flap advancements are suitable for simple fistula, while a procedure with interposing viable tissues can be applied to repair complex fistula, and the successful rate ranged between 60% and 85% [15]. For complex RVF, due to the fibrosis and chronic inflammatory which compromise the blood supply to the tissue surrounding the fistula and, in turn, reduces the ability of tissue healing, procedures with grafts of omentum and gracilis, gluteus maximus, or rectus muscles have all been successfully used [12, 13, 19]. In recent years, surgeons prefer to use bulbocavernosus muscle along with its fat pad to repair complex rectovaginal fistula as an option of the treatment [11, 14]. We have successfully cured all of nine patients without any significant morbidity, such as incontinence, abscess, and recurrence of fistula.

No fecal incontinence and recurrence were found during the time of follow-up. Only one patient experienced mild dyspareunia, for whom further treatment was not needed. Based on the results from our patients, it is plausible that both simple and complex rectovaginal fistula could be reliably repaired with good outcome by interposing the vital bulbocavernosus graft. Some reasons contributed to this success. One was that interposing a health and well-vascularized tissue avoids direct apposition of two suture lines and introduces well-vascularized tissue to the area. Another was that none of our patients were found to have the anal sphincter injury in the preoperational examination, although 78% (seven out of nine) patients underwent fistula surgical repair for one to six times. Our results was in coincidence with Tsang CB’s suggestions that local tissues are inadequate for endorectal advancement flap repairs in patients with sphincter defects and a history of previous repairs and that patients with clinical or anatomic sphincter defects should be treated by sphincteroplasty with levatoroplasty [24]. Other reasons of the simple procedure are the surgeons’ understanding about the anatomy of rectum, anus, and pelvic floor, and the operators’ skill in meticulous dissections for protecting the blood vessel of the pedicle.

Conclusion

Though the rectovaginal fistula is troublesome to the surgeons, it could be cured by using the interposition of vital bulbocavernosus graft between rectum and vagina. Our results, despite the series that is small, seem to demonstrate that this technique is both simple and effective, especially for recurrent rectovaginal fistula without anal sphincter injury.

Copyright information

© Springer-Verlag 2009