Diverticular disease (DD) remains a major healthcare problem in the United States, with the annual cost of treatment estimated at >2.6 billion dollars.1 This is compounded by the growing prevalence of the disease, with increased hospitalizations from 74.1 per 100,000 persons in 2000 to 91.9 in 2010.2 Although the majority of patients who experience an episode of diverticulitis will not require surgery, fistulizing disease remains an indication for surgical intervention.
In DD, colovesicular fistulas are the most common followed by colocutaneous, colouterine, coloenteric, and colovaginal (CVagF).3,4 Although CVagFs are the rarest type found in DD, this disease process remains the leading cause of CVagFs (DCVagFs), followed by malignancy and IBD.5,6 Similar to rectovaginal fistulas, patients usually present with the passage of air or stool through their vagina. These symptoms can cause significant incontinence, morbidity, distress, and embarrassment to afflicted patients.
Treatment is surgical but can be challenging given the inflammatory process that distorts planes, making dissection notoriously difficult. After definitive resection, recurrence rates are low; however, when they do recur, the morbidity is significant, because reoperative surgery in the setting of pelvic sepsis can lead to poor functional outcomes and sometimes permanent ostomy. Therefore, it is imperative to optimize patient factors and operative strategies during the index operation to achieve the best long-term outcome. Currently, there is no consensus on the optimal surgical approach to treat DCVagFs. Furthermore, to date, most studies are small and include CVagFs secondary to multiple etiologies. The aim of this study was 2-fold, first to examine overall short-term postoperative outcomes and, second, to identify variables associated with successful surgical outcome.
PATIENTS AND METHODS
After institutional review board approval, a single tertiary-referral center retrospective study of a prospectively maintained clinical database was performed to identify all of the women with DCVagF from 2011 to 2016 who underwent elective surgical repair with the intent to close the fistula. Patients were identified using International Classification of Diseases, Ninth Revision, codes for CVagF (code 619.1). Patients were excluded if they underwent emergent surgery, received diversion only as their definitive repair, or had a fistula because of other etiology (ie, IBD, malignancy, or mesh erosion). DCVagF was diagnosed by a combination of preoperative symptoms (passage of stool or air through the vagina), radiologic imaging with either CT or MRI, and physical examination to rule out anovaginal or rectovaginal fistula. Additional testing including colonoscopy, vaginoscopy, and gastrograffin enema was used on a case-by-case basis. Patient characteristics, surgical strategies, perioperative outcomes, conversion rates, postoperative complications, recurrences, readmissions, and reoperations were reported. Patient characteristics included age, sex, BMI, ASA classification, diabetes mellitus, smoking, use of immunosuppressant medications, previous surgery, previous hysterectomy, and history of pelvic radiation. Surgical strategies included laparoscopic, robotic, hand assist, or open techniques; use of omental pedicle flaps, diverting ostomy, pelvic drain, and ureteral stents; splenic flexure mobilization; and primary repair of the vaginal defect. The surgical approach and decision to create a diverting loop ileostomy or place ureteral stents preoperatively was left to the discretion of the operating surgeon. Typically the use of ureteral stents was preferred in cases of significant inflammation directed toward the retroperitoneum on CT and/or in obese patients. Perioperative outcomes included operating room time, estimated blood loss, and length of stay. Postoperative complications were graded using the Clavien-Dindo classification system.7,8
Our primary outcome was DCVagF closure defined as complete resolution of symptoms without the presence of a stoma. Patients were stratified into 2 groups, success and failure. Secondary outcomes included postoperative complications, mortality rates, and intensive care unit (ICU) admission. Outcomes were determined by individual electronic health record review of both inpatient and outpatient records. Those patients with outside hospital encounters captured by our electronic health record were also reviewed.
Univariate analysis was conducted to compare the 2 retrospective cohorts (success versus failure) in terms of demographics, baseline patient characteristics, operative characteristics, and postoperative outcomes. Data are reported as mean (± SD) or frequency (proportion). The Pearson χ2 or Fisher exact test was used for categorical factors and Student t test or Wilcoxon rank-sum test for continuous variables. Two-tailed p values <0.05 were considered statistically significant. All of the analyses were conducted using JMP (version 13, SAS Institute, Inc, Cary, NC).
Our study was approved by the Cleveland Clinic Institutional Review Board.
A total of 52 patients underwent definitive surgical treatment of DCVagF by 23 surgeons over 5 years. All of the patients experienced stool per vagina and had either CT or MRI to confirm the diagnosis of diverticulitis. All of the patients had a colonoscopy before surgery within the appropriate screening or surveillance window for each patient. Of the 37 patients who had a colonoscopy during the preoperative period, 6 detected a fistula opening. Of 8 patients who had a vaginoscopy, 6 detected a fistula. Eight patients had a gastrograffin enema, of which 4 detected a fistula. Size of the fistula was not determined preoperatively, because these were all CVagFs, not accessible by physical examination. The operation was successful in 47 patients (90%), whereas 5 (10%) failed. Patients in the success and failure groups did not differ with respect to age, BMI, proportion with diabetes mellitus, ASA grade, chronic steroid use, previous abdominal surgery, or previous hysterectomy. However, patients who failed were more likely to be smokers (n = 3 (60.0%) vs n = 6 (12.8%); p = 0.03). Patient demographics and operative variables are summarized in Tables 1 and 2.
Overall, 23 operations (44%) were performed using a minimally invasive approach, with a conversion rate of 22% (n = 5/23; see Table 3). Reasons for conversion were significant adhesions in 1 patient with a concomitant coloureteral fistula, a large phlegmon in 2 patients, and dense pelvic adhesions in 2 patients.
For clinical outcomes, there were no significant differences between the 2 groups with respect to length of stay, postoperative complications, readmission or ICU admission (Table 4). Postoperative complications included ileus, wound infection, atrial fibrillation, bacteremia, intra-abdominal abscess, acute anemia requiring blood transfusion, acute renal failure, and 1 ureteral injury. The patient with the ureteral injury had preoperative ureteral stents placed. The injury was identified during the index operation and was repaired by the urology team without any postoperative urologic complications. There were no 30-day mortalities in either group. There was only 1 patient who required return to the operating room within 30 days for a midline wound infection, requiring washout and vacuum-assisted closure placement.
Twenty-eight patients (54%) had an ostomy constructed, of which 3 (11%) ultimately failed to meet our success criteria. Of those patients with an ostomy, 25 (100%) and 1 (33%; p = 0.11) were reversed in the success and failure groups. The other 2 patients in the failure group had their ileostomy closed after their index operation, with recurrence of their colovaginal fistula and subsequent creation of a permanent ostomy.
Postoperative outcomes of the patients who failed are shown in Table 5. Median follow-up time was 3 years, with an interquartile range of 2.0 to 4.5 years (range, 1–6 y). Among the failures, there were no anatomic abnormalities or extraordinary difficulties encountered during the index operation. Of these recurrences, 2 were secondarily successful and 3 were secondary failures. Of the secondary successes, patient 1 had either a delayed anastomotic leak or recurrent diverticulitis based on radiologic imaging and operating room findings. Patient 2 had a recurrence 1.5 weeks after the index operation from an anastomotic leak.
Of the secondary failures, patients 3 and 4 had return of symptoms from an anastomotic leak. Given her age and comorbidities, patient 4 had takedown of the fistula with end colostomy during her second operation and was never subsequently reversed. Patient 5 had recurrence 3 years later; however, it was never followed up for definitive repair. Thus, accounting for secondary operations, the ultimate success and failure rates were 94.0% (n = 49) versus 5.7% (n = 3).
We report our institution’s outcomes and surgical approach for the treatment of DCVagF. Overall, we found that the vast majority of DCVagFs can be successfully managed with colectomy and restoration of intestinal continuity. In terms of predicting failure, our study was only able to identify smoking as being associated with adverse outcomes.
In the context of the DCVagF literature, this is one of the largest studies to date looking specifically at DCVagFs. Hjern et al9 studied a similar number of patients (n = 57) with diverticular-related fistulas to the female genital tract but included colouterine fistulas as well. They had satisfactory outcomes with sigmoid resection and primary anastomosis but had a follow-up period of only 1 year. They noted that previous hysterectomy maybe a risk factor for DCVagF; however, it may also occur in those patients with an intact uterus.9 Seven patients in our series did not have a previous hysterectomy, including 5 in the success group and 2 in the failure group. Only 1 of these patients (who was in the failure group) underwent a hysterectomy during the index operation for unrelated gynecologic issues. With 5 patients having full resolution of their DCVagF with sigmoid resection alone, we do not recommend routine hysterectomy to treat DCVagFs.
Overall, we were highly successful, with an ultimate fistula closure rate exceeding 90%. This was achieved with similar outcomes whether the procedure was done with a minimally invasive or an open approach. The question of laparoscopy in the treatment of CVagFs was recently addressed by Wen et al,10 who did a retrospective review of 27 patients with DCVagFs and found no difference in recurrence, postoperative complications, ICU admission, or length of stay when comparing the laparoscopic with the open approach. They concluded that laparoscopy was safe for the surgical treatment of CVagFs and should be used when feasible. The role of laparoscopy has been more extensively studied with larger numbers in the diverticular colovesicular literature. Cirocchi et al3 performed a systematic review of retrospective studies comparing laparoscopic and open surgery for the treatment of a variety of diverticular fistulas but primarily colovesicular. They found no difference in fistula recurrence, early reintervention, or need for Hartmann or proximal diversion between laparoscopy and open, therefore concluding that laparoscopy is safe for the treatment of diverticular-related fistulas.3 Our results echo these findings and add incrementally to the literature on this topic.
In our study, the only characteristic that was associated with poor surgical outcomes was smoking. Patients who recurred were more likely to be smokers (n = 3 (60.0%) vs n = 6 (12.8%); p = 0.03). This result is not surprising, because numerous studies have shown the deleterious effects of smoking on wound healing and anastomotic leak, particularly in colorectal anastomosis.11,12 Therefore, patients who smoke should be strongly encouraged to stop before embarking on an elective repair.
In our institution, the use of an omental pedicle flap has traditionally been considered favorable to improve success rates. The omentum possesses a rich vascular and lymphatic supply that facilitates healing, even in the presence of infection, all while allowing epithelialization to occur over its surface.13 These properties, in particular the latter, make the omentum an extremely useful adjunct in difficult situations where tissues may be thin and friable in an infected hostile environment. Its use is well described in the genitourinary literature, particularly for the repair of vesicovaginal fistulas, where there is often no tissue to reapproximate.14 Because of the retrospective nature of our study, there was no specific criteria set forth to determine when an omental pedicle flap would be used. In review of the operative reports, we could not elucidate why an omental pedicle flap was or was not created. Overall, our data failed to show that the use of omental pedicle grafts for DCVagF decreased recurrences. This may be because of the small number of patients in this study and possible selection in the use of omental flaps for more difficult cases. Despite this, given the natural healing properties of the omentum and the relative ease of mobilizing a flap, it seems reasonable that when there is an inflamed area of the vagina, interposition of the omental pedicle flap between the vagina and new anastomosis may be done. However, if not feasible or if the revised anastomosis is away from the vaginal defect, the omental pedicle flap maybe omitted.
Interestingly, in our study proximal diversion at the index operation did not correlate with surgical success of DCVagFs. There is no consensus on whether a proximal diversion is necessary for the treatment of diverticular-related fistula. Most studies are small retrospective reviews that report conflicting results, with some authors promoting the liberal use of diverting stoma, while others are proponents of a single-stage procedure.15,16 Smeenk et al16 looked at 40 patients with a combination of colovesicle and CVagFs. They had a very high anastomotic leak rate of 44% (8/18) in those patients with primary anastomosis without a diversion, and therefore recommended a 2-stage procedure for this patient population. In our series, there were 24 patients who underwent primary anastomosis without diversion, with an anastomotic leak rate of only 4.1% (1/24). These findings demonstrate that primary anastomosis is feasible and safe in the correctly selected patient. In this series and as a protocol for our institution, we advocate selective use of proximal diversion for DCVagFs and other diverticular related fistulas if there are concerns about the healing of the anastomosis, specifically in immunosuppressed patients, or because of intraoperative findings (emergent procedure, poor tissue quality, or positive leak test), particularly when the surgeon anticipates life-threatening consequences deriving from an anastomotic leak.
Limitations of this study include its retrospective design that inherently encompasses selection bias; specifically, the management of these patients, such as when to divert or create an omental pedicle flap, was surgeon driven. In addition, despite being one of the largest studies reported to date, our overall number of patients is not large enough to adequately power the study to make steadfast conclusions. Although the results are from 23 surgeons, our data are from a single tertiary institution and thus may not be generalizable to the general surgical or colorectal community. Despite these limitations, we report one of the largest series of DCVagFs and emphasize certain techniques that may contribute to the successful outcome of surgically treated DCVagFs.
Surgery is highly effective in achieving successful closure of DCVagFs through either minimally invasive or open techniques. Patients who smoke should be strongly encouraged to stop before embarking on an elective repair. Although the use of fecal diversion and omental pedicle flaps did not correlate with success, they may be considered when deemed clinically appropriate.
1. Sandler RS, Everhart JE, Donowitz M, et al. The burden of selected digestive diseases in the United States. Gastroenterology. 2002;122:1500–1511.
2. Feuerstein JD, Falchuk KR. Diverticulosis and diverticulitis
. Mayo Clin Proc. 2016;91:1094–1104.
3. Cirocchi R, Arezzo A, Renzi C, et al. Is laparoscopic surgery the best treatment in fistulas complicating diverticular disease of the sigmoid colon? A systematic review. Int J Surg. 2015;24(pt A):95–100.
4. Goldberg JB, Moses RA, Holubar SD. Colosplenic fistula
: a highly unusual colonic fistula
. J Gastrointest Surg. 2012;16:2338–2340.
5. Grissom R, Snyder TE. Colovaginal fistula
secondary to diverticular disease. Dis Colon Rectum. 1991;34:1043–1049.
6. Berger MB, Khandwala N, Fenner DE, Burney RE. Colovaginal
fistulas: presentation, evaluation, and management. Female Pelvic Med Reconstr Surg. 2016;22:355–358.
7. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg. 2004;240:205–213.
8. Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg. 2009;250:187–196.
9. Hjern F, Goldberg SM, Johansson C, Parker SC, Mellgren A. Management of diverticular fistulae to the female genital tract. Colorectal Dis. 2007;9:438–442.
10. Wen Y, Althans AR, Brady JT, et al. Evaluating surgical management and outcomes of colovaginal
fistulas. Am J Surg. 2017;213:553–557.
11. Richards CH, Campbell V, Ho C, Hayes J, Elliott T, Thompson-Fawcett M. Smoking
is a major risk factor for anastomotic leak in patients undergoing low anterior resection. Colorectal Dis. 2012;14:628–633.
12. Midura EF, Hanseman D, Davis BR, et al. Risk factors
and consequences of anastomotic leak after colectomy
: a national analysis. Dis Colon Rectum. 2015;58:333–338.
13. Wein AJ, Malloy TR, Greenberg SH, Carpiniello VL, Murphy JJ. Omental transposition as an aid in genitourinary reconstructive procedures. J Trauma. 1980;20:473–477.
14. Reisenauer C. Vesicovaginal fistulas: a gynecological experience in 41 cases at a German pelvic floor center. Arch Gynecol Obstet. 2015;292:245–253.
15. Carvajal Balaguera J, Camuñas Segovia J, Peña Gamarra L, et al. Colovesical fistula
complicating diverticular disease: one-stage resection. Int Surg. 2006;91:17–23.
16. Smeenk RM, Plaisier PW, van der Hoeven JA, Hesp WL. Outcome of surgery for colovesical and colovaginal
fistulas of diverticular origin in 40 patients. J Gastrointest Surg. 2012;16:1559–1565.
Colectomy; Colostomy; Colovaginal; Diverticulitis; Fistula; Ileostomy; Omental pedicle flap; Rectovaginal; Risk factors; Smoking
Supplemental Digital Content
© 2019 The American Society of Colon and Rectal Surgeons