Understanding Pouch-Related Fistulae: Don’t Lose Track of Time

Introduction

Pouch-related fistulae (PRF) occur in up to 15% of patients with ulcerative colitis (UC) after total proctocolectomy (TPC) with ileal pouch-anal anastomosis (IPAA) and may be secondary to inflammation, anastomotic leakage, cryptoglandular disease, or malignancy.^1–5 Delineating the etiology of PRF is important as it dictates management. PRF secondary to anastomotic leakage is primarily managed surgically with local procedures such as seton placement, fistulotomy, endoanal pouch advancement flaps, and ligation of the fistula tract in combination with antibiotics. In contrast, PRF secondary to inflammation as in Crohn’s disease like pouch inflammation (CDLPI) is primarily managed medically with advanced therapies such as biologics or small molecules in combination with surgical therapy. Unfortunately, no standardized diagnostic criteria for PRF exist, and many patients with PRF secondary to anastomotic leakage may be inappropriately diagnosed and treated unnecessarily with advanced therapies.

We have previously proposed a classification system by which the etiology of PRF is delineated by the timing of development.⁶ We hypothesize that early-onset PRF that present within 12 months of IPAA are more likely to be related to anastomotic leakage, while late-onset PRF that present greater than 12 months after IPAA are more likely to be secondary to CDLPI. To investigate this, we conducted a review of patients with UC who underwent TPC with IPAA and subsequently developed PRF at our institution.

Methods

Patients with UC who underwent TPC with IPAA at Mount Sinai Hospital between 2007 and 2022 were identified from a prospectively maintained database that was last updated on June 2024. CDLPI was defined as 1 or more of the following features: inflammation that extended >10 cm into the pre-pouch ileum, strictures (exclusive of sites of anastomosis), and fistulae involving the pouch, proximal small bowel, or perineum. PRF were defined as unintended connections from 1 epithelial surface (ileal pouch, cuff, or dentate line) to another (vagina, bladder, abdominal, perianal, or perineal skin). The time from the final IPAA surgical stage to PRF presentation was calculated as time to onset. PRF remission was defined as the resolution of pain and drainage and closure of all external openings confirmed by cross-sectional imaging.

Categorical variables were summarized as frequencies and percentages, and quantitative variables were summarized as medians with interquartile range (IQR). Tests for association between patients with early- versus late-onset PRF were performed using the chi-squared test and Wilcoxon rank-sum method.

Results

Of the 772 patients who underwent IPAA, 98 (12.7%) developed PRF over a median follow-up of 3.8 [IQR 1.1-7.4] years. Early-onset PRF developed in 34 (34.7%) patients and late-onset PRF in 64 (65.3%). Demographics and clinical characteristics are summarized in Table 1. The median time to PRF development was 2.3 [IQR 0.76-5.39] months and 4.34 [1.91-8.21] years in the early- and late-onset groups, respectively. The most common types of PRF were perianal (64, 59.2%), pouch-vaginal (22, 20.4%), enterocutaneous (12, 11.1%), and entero-enteric (7, 6.5%). Patients with early-onset PRF had significantly higher rates of postoperative complications compared to those with late-onset PRF (82.4% vs 53.1%, P = .004), specifically with higher rates of anastomotic leakage (23.5% vs 7.8%, P = .03) and abscess formation (32.4% vs 14.0%, P = .03). In contrast, patients with early-onset PRF had lower rates of subsequent CDLPI diagnosis compared with those with late-onset PRF (32.4% vs 71.9%, P < .001).

Table 1.

Demographics and clinical characteristics

Variable	Early-onset PRF, N = 34	Late-onset PRF, N = 64	P-value
Sex
Female, n (%)	18 (52.9%)	28 (43.8%)	.39
Male, n (%)	16 (47.1%)	36 (56.2%)
Age at diagnosis, years, mean ±SD	30.9 +/- 13.1	26.7 +/- 10.9	.11
Indication for surgery
Medically refractory disease	24 (70.6%)	58 (90.6%)	.04
IBD-U	2 (5.9%)	1 (1.6%)
Dysplasia or cancer	8 (23.5%)	5 (7.8%)
Steroid use within 3 months	19 (55.9%)	43 (67.2%)	.27
Biologic and small molecules use within 3 months	21 (61.8%)	40 (62.5%)	.94
Surgery
• 1 stage	1 (2.9%)	6 (9.4%)	.34
• 2 stages	8 (23.6%)	19 (29.7%)
• 3 stages	25 (73.5%)	39 (60.9%)
Type of anastomosis
• Handsewn	13 (38.2%)	21 (32.8%)	.59
• Stapled	21 (61.8%)	43 (67.2%)
All postop complications	28 (82.4%)	34 (53.1%)	.004
• Anastomotic leak	8 (23.5%)	5 (7.8%)	.03
• Abscess formation	11 (32.4%)	9 (14%)	.03
CDLPI diagnosis	11 (32.4%)	46 (71.9%)	<.001
• Pre-pouch ileal inflammation	9 (81.8%)	32 (69.6%)
• Stricture	0	3 (6.5%)
• Fistulae	11 (100%)	46 (100%)

Abbreviations: CDLPI, Crohn’s disease like pouch inflammation; IBD-U, inflammatory bowel disease unclassified; PRF, pouch-related fistula.

Among the 98 patients with PRF, 136 therapeutic procedures were performed, 37 (27.2%) in the early-onset group and 99 (72.8%) in the late-onset group. There was no significant difference in the number of procedures performed in the early- versus late-onset group. The most common procedures performed were EUA with or without drainage and seton placement (90, 66.2%), fistulotomy (18, 13.2%), advancement flap (14, 10.3%), seton re-insertion (5, 3.7%), fistulectomy (4, 3%), transvaginal repair (3, 2.2%), plug/glue (1, 0.7%), and ligation of intersphincteric fistula tract (1, 0.7%). Significantly less patients with early- compared to late-onset PRF were prescribed biologic therapy (15.6% vs 29.4%, P = .04). The most common biologics prescribed were antitumor necrosis factor agents (32, 71.1%), ustekinumab (9, 20%), and vedolizumab (4, 8.9%). PRF remission rates were significantly greater in patients with early- compared to late-onset PRF (79.2% vs 47.1%, P = .01); however, pouch failure rates were not significantly different (29.2% vs 33.3%, P = .53).

Discussion

In this single-center cohort study of patients with UC who underwent TPC with IPAA and subsequently developed PRF, approximately one-third of patients presented as early-onset within 12 months of the final surgical stage. Consistent with prior reports, patients with early-onset PRF were more likely to have postoperative anastomotic leakage while patients with late-onset PRF were more likely to be diagnosed with CDLPI.⁷ These data have significant diagnostic and therapeutic implications.

First, they suggest the timing of PRF development may be helpful to delineate etiology and guide evaluation. In patients with early-onset PRF, a thorough review of the surgical intra- and postoperative course should be conducted with attention paid to subtle signs and symptoms such as abdominal pain, fever, and tachycardia that may indicate underlying anastomotic leakage. Evaluation with EUA, pouchoscopy, and cross-sectional imaging should be promptly performed. The location of PRF noted on EUA and pouchoscopy may provide further insight into etiology—those originating from the anastomosis are likely related to anastomotic leakage, while those originating from the cuff or anoderm are likely related to CDLPI.

Second, these data suggest the timing of PRF development may inform therapy. Advanced therapies such as biologics or small molecules should not be reflexively prescribed for early-onset PRF unless there are other concerns for CDLPI such as significant pre-pouch ileal inflammation. If anastomotic leakage is confirmed, therapy should consist of a combination of antibiotics and source control with consideration for pouch advancement or reconstruction. As noted in this study, early-onset PRF have higher rates of remission than late-onset PRF when treated appropriately.

Although this study may be viewed as an over-simplification of PRF, it provides clear insights for clinical care that are relevant and practical. The strengths of the study include the use of a prospectively managed database with prespecified definitions and long-term follow-up. The limitations include the possibility of early- versus late-onset PRF misclassification dependent on when symptoms were recognized and the perceptual bias of late-onset PRF deemed to be secondary to CDLPI and managed with biologics. To that end, an effort was taken to clearly delineate the timing of the first PRF symptom and biologic therapy relative to the final surgical stage. Certainly, the proposed timing is not absolute and patients with CDLPI may present with early-onset PRF just as patients with anastomotic leakage may present with late-onset PRF. Ultimately, additional multicenter, prospective studies are necessary to validate these findings and supplement them with biomarker, imaging, and histopathologic criteria to further develop a PRF classification schema.

The frequency of pouch failure in patients with PRF remains significant at 30%, and prompt recognition and management are needed to mitigate this risk and optimize pouch longevity.⁸ We propose the timing of PRF presentation serve as initial guidance for evaluation, particularly for early-onset PRF where a high index of suspicion for anastomotic leakage is needed. Given the complexity of PRF, a multidisciplinary approach with input from gastroenterologists and colorectal surgeons is necessary.

Funding

This work was supported by NIH grant K23DK127241 (M.K.).

Conflicts of Interest

S.B., H.W., M.P., S.H., S.K., P.S., and A.G. have no conflicts to disclose. S.Y.W. has a research contract with Takeda/Trinetx; advisory board: BMS. J.F.C. reports receiving research grants from AbbVie, Janssen Pharmaceuticals, Prometheus, Takeda, and Bristol Myers Squibb; receiving payment for lectures from AbbVie, and Takeda; receiving consulting fees from AbbVie, Amgen, AnaptysBio, Allergan, Arena Pharmaceuticals, Boehringer Ingelheim, Bristol Myers Squibb, Celgene Corporation, Celltrion, Eli Lilly, Ferring Pharmaceuticals, Galmed Research, Glaxo Smith Kline, Genentech (Roche), Janssen Pharmaceuticals, Kaleido Biosciences, Immunic, Invea, Iterative Scopes, Merck, Landos, Microba Life Science, Novartis, Otsuka Pharmaceutical, Pfizer, Protagonist Therapeutics, Prometheus, Sanofi, Seres, Takeda, Teva, TiGenix, Vifor; and hold stock options in Intestinal Biotech Development. M.K. has served as a consultant or received advisory board fees from AbbVie, Pfizer, BMS, Johnson & Johnson, and Fresenius Kabi.