Surgical strategy for complex perineal small bowel fistula after pelvic malignancy: An etiology-driven approach

Abstract

BACKGROUND

Perineal small bowel fistula (PSF) is a highly debilitating complication of radical pelvic malignancy resections. It is typically driven by empty pelvis syndrome (EPS) and chronic adhesive disease. Current therapeutic guidelines lack standardization.

AIM

To evaluate an etiology-driven surgical algorithm, where procedures are chosen based on the underlying mechanism, aiming to improve closure rates.

METHODS

Retrospective cohort analysis was performed on 28 consecutive patients who underwent definitive repair for complex PSF. Patients were stratified into two surgical groups based on etiology: Group 1 (n = 15) underwent modified pelvic floor reconstruction for structural defects or EPS; group 2 (n = 13) underwent internal intestinal plication for extensive adhesions or tumor recurrence. Overall survival was calculated from the date of definitive surgery to the last follow-up or death.

RESULTS

The study cohort (mean age 56.6 years) included patients with rectal (35.7%), gynecological (35.7%), and various other pelvic malignancies. There were no significant baseline demographic differences between the groups. A significant correlation was observed between fistula etiology and surgical strategy (P = 0.02). While reconstruction was the exclusive surgical approach for patients with EPS (100%), plication was predominantly utilized for cases of tumor recurrence (66.7%). The overall fistula closure rate was 89.3% (25/28), showing no significant difference between group 1 (86.7%) and group 2 (92.3%; P = 1.0). However, group 2 experienced a significantly longer mean time to postoperative flatus (6.8 days vs 4.2 days; P = 0.002). The median follow-up time for the entire cohort was 34.5 months (range: 9.0-107.0 months). Overall survival was comparable across the two surgical groups (32.0 months vs 37.0 months; P = 0.66).

CONCLUSION

Managing complex PSFs requires a tailored approach. Employing an etiology-driven strategy, specifically pelvic floor reconstruction for structural voids (EPS) and internal intestinal plication for “frozen” adhesive conditions, achieves > 89% closure rates with acceptable morbidity.

Key Words: Perineal small bowel fistula; Empty pelvis syndrome; Radiation enteritis; Internal intestinal plication; Pelvic exenteration

Core Tip: This study evaluated an etiology-driven algorithm for repairing perineal small bowel fistulas. By customizing surgical interventions - such as pelvic floor reconstruction and intestinal plication - to the underlying mechanisms of the fistula, researchers achieve a successful closure rate of 89.3% in high-risk patients with a history of pelvic cancer.

INTRODUCTION

Perineal small bowel fistula (PSF) is among the most challenging, severe, and resource-intensive complications in the field of abdominal surgical oncology. Typically following pelvic exenteration or extensive abdominoperineal resections for locally advanced rectal[1], it creates severe, life-altering postoperative morbidity that stands as the harsh physical toll of these radical, life-saving procedures. The reported incidence of fistula formation after such extensive resections varies widely, ranging from 4.5% to over 20% in irradiated pelvises[2]. Although major postoperative complications (Clavien-Dindo grade ≥ 3) have historically occurred in 34.5% of cases, these complications do not negatively impact overall survival (OS)[3].

The pathophysiology of PSF is distinct from other enterocutaneous fistulas due to the unique anatomical environment of the “frozen” or “empty” pelvis. Two primary mechanisms drive PSF formation. First, empty pelvis syndrome (EPS), defined by the large dead space created after total PE, allows small bowel loops to prolapse into the pelvic basin. Without the pelvic floor support, these loops sink into the raw, often infected, perineal wound, causing tissue erosion and eventual fistula formation[4]. Second, chronic radiation enteritis causes progressive endarteritis, submucosal fibrosis, and ischemia, rendering the small bowel fragile and prone to spontaneous perforation or anastomotic leakage[5].

Despite the severity of this condition, high-quality evidence guiding surgical management is scarce. Traditional approaches, such as simple closure or tissue removal, have high failure rates in the challenging pelvic environment. More aggressive strategies, such as using vascularized flaps (ventral rectus abdominis myocutaneous, omental flaps) to fill the dead space, have been proposed to mitigate EPS[6,7]. Alternatively, internal intestinal plication has been utilized to manage extensive adhesions and prevent kinking in radiation-damaged bowels[8].

The core hypothesis of this analysis asserts that surgical success in PSF depends on matching the technique to the specific underlying etiology. This study systematically reviews clinical institutional experience with 28 consecutive cases of complex PSF, evaluating an etiology-driven algorithm that guides the choice between structural reconstruction and plication for adhesive disease. By precisely delineating the underlying specific pathophysiological drivers and successful surgical strategies, this report seeks to establish a standardized clinical protocol with the goal of improving salvage rates, promoting gastrointestinal recovery, and elevating the overall quality of life for this highly complex patient population.

MATERIALS AND METHODS

Study design and patients

This single-center, retrospective cohort study systematically evaluated the medical records of all patients diagnosed with PSF admitted to the Department of Difficult Abdominal Surgery at Shanghai Tenth People’s Hospital (Shanghai, China) between August 2016 and December 2025. The study received ethical approval from the Ethics Committee of Shanghai Tenth People’s Hospital (Approval No. SHSY-LYZX-652) and is registered under China Clinical Trial Registry number ChiCTR2600118378.

Inclusion and exclusion criteria

Patients were included if they met the following criteria: (1) A history of surgery for pelvic malignancy; (2) Clinical and radiological evidence of a small bowel fistula communicating with the perineum or vagina; and (3) Subsequent definitive abdominal surgery for fistula repair. Exclusion criteria were: (1) Fistulas originating from the colon or rectum (unless concomitant with a small bowel fistula); (2) Fistulas caused by benign inflammatory diseases (e.g., Crohn’s disease); and (3) Patients managed solely with conservative therapy without undergoing definitive surgical repair.

Surgical strategies and grouping

Patients were categorized into two surgical groups based on the primary technique employed.

Group 1: Modified pelvic floor reconstruction. This strategy aims to eliminate the pelvic dead space. Techniques included the use of pedicled greater omental flaps, biological meshes (ZR MedTech, Jiangsu Province, China), or myocutaneous flaps. It is typically indicated for EPS or large tissue defects.

Group 2: Internal intestinal plication. This surgical technique is used to manage extensive adhesions by systematically arranging the small bowel (internal plication)[8] to prevent loops from settling into the pelvis and to manage extensive inter-loop adhesions. This is achieved using a double balloon ileus tube (CREATE MEDIC, Kanagawa, Japan). Prior to surgery, pelvic mapping via contrast-enhanced computed tomography or magnetic resonance imaging is used for accurate anatomical assessment.

Race and ethnicity reporting

Patient-reported race and ethnicity data were extracted from hospital electronic health medical records. The study population consisted exclusively of Han Chinese patients (100%), reflecting the demographic distribution of the primary catchment area in Shanghai and the East China region. Ethnicity was documented to comply with reporting standards; however, this surgical investigation prioritized the biomechanical and environmental drivers of fistulation - specifically EPS and radiation - over ethnic variations.

Statistical analyses

Data were analyzed using SPSS version 26.0 (IBM Corp, Armonk, NY, United States). A priori sensitivity analysis framework was applied to determine the appropriate sample size[9]. To detect a 20% difference in clinical success rates, a standard two-proportion comparison requires a sample size of approximately 62 patients per arm. Continuous variables are presented as the mean ± SD and were compared using the Student’s t-test. Categorical variables are presented as n (%) and were compared using the Fisher’s exact test. Odds ratios were calculated to assess the likelihood of selecting specific surgical strategies based on etiology. P < 0.05 was considered statistically significant.

RESULTS

Patient demographics and baseline characteristics

The study cohort comprised 28 patients (13 males, 15 females) with a mean age of 56.6 ± 12.8 years. The most common primary malignancies were rectal/anal cancer (35.7%) and gynecological cancer (35.7%). A history of pelvic radiotherapy was present in 64.3% (18/28) of the population. Patients were stratified into two surgical groups: Group 1 (modified pelvic floor reconstruction, n = 15) and group 2 (internal intestinal plication, n = 13). Comparative analysis revealed no statistically significant differences in age (P = 0.36), sex (P = 0.47), primary malignancy (P = 0.73), or history of radiotherapy (P = 0.69) between the two groups, ensuring a comparable baseline for evaluating outcomes (Table 1).

Table 1 Comparison of patient characteristics by surgical group, n (%)/mean ± SD.

Characteristic	Group 1: Reconstruction (n = 15)	Group 2: Plication (n = 13)	P value
Race/ethnicity
Han Chinese	15 (100)	13 (100)
Age in years	58.7 ± 13.2	54.2 ± 11.9	0.36
Sex			0.47
Male	6 (40.0)	7 (53.8)
Female	9 (60.0)	6 (46.2)
Primary malignancy			0.73
Rectal/colorectal	8 (53.3)	9 (69.2)
Gynecological: Cervix/ovary	6 (40.0)	4 (30.8)
Other: Sarcoma/bladder	1 (6.7)	0 (0.0)
History of radiotherapy			0.69
Yes	9 (60)	9 (69.2)
No	6 (40)	4 (30.8)
Primary fistula etiology			0.02a
Empty pelvis syndrome	3 (20.0)	0 (0.0)
Tumor recurrence	4 (26.7)	8 (61.5)
Radiation enteritis	5 (33.3)	4 (30.8)
Surgical/iatrogenic injury	2 (13.3)	1 (7.7)
Other: Internal radiation	1 (6.7)	0 (0.0)

^aP < 0.05.

Data are n (%) unless indicated otherwise, presenting comparative demographic and baseline clinical variables between patients undergoing pelvic floor reconstruction (Group 1) and internal intestinal plication (group 2). Statistical significance was evaluated using the independent Student's t-test for continuous variables and Fisher’s exact test for categorical variables. While demographic parameters were statistically identical, the primary etiology of the fistula demonstrated a significant divergence (P = 0.02), validating the core premise that the surgical strategy was heavily driven by the underlying pathophysiological mechanism.

Correlation between etiology and surgical strategy

The selection of surgical technique was strongly driven by the underlying fistula etiology, as confirmed by a statistically significant difference in the distribution of etiologies between the two groups (P = 0.02; Table 1). Specifically, all patients presenting with pure EPS (n = 3) were managed exclusively (100%) with modified pelvic floor reconstruction (group 1) to eliminate the pelvic dead space (Table 2). Conversely, patients with fistulas associated with tumor recurrence (n = 12) were preferentially managed with internal intestinal plication (group 2) in 66.7% of cases (8/12) to address the frozen pelvis adhesive pattern. For patients with radiation enteritis (n = 9), surgical intervention was split between reconstruction (55.6%) and plication (44.4%), depending on the available space within the pelvic cavity. Odds ratios analysis showed that patients with tumor recurrence had 4.00 times the odds (95% confidence interval: 0.77-20.8) of undergoing internal intestinal plication rather than reconstruction, whereas EPS was a perfect predictor that a patient would receive reconstruction (Table 3).

Table 2 Correlation between fistula etiology and surgical strategy, n (%).

Etiology	Total, n	Group 1: Reconstruction	Group 2: Plication	Dominant strategy
Empty pelvis syndrome	3	3 (100)	0 (0)	Reconstruction
Tumor recurrence	12	4 (33.3)	8 (66.7)	Plication
Radiation enteritis	9	5 (55.6)	4 (44.4)	Mixed
Surgical/iatrogenic injury	3	2 (66.7)	1 (33.3)	Reconstruction
Internal radiation: Seeds	1	1 (100)	0 (0)	Reconstruction
Total	28	15 (53.6)	13 (46.4)

Stratification of patients based on the dominant pathophysiological driver of the perineal small bowel fistula. The data illustrate strict adherence to an etiology-driven algorithm, where pure structural defects (empty pelvis syndrome) universally trigger reconstructive flap techniques, whereas recurrent, adhesive malignant disease predominantly necessitates the physiological stenting provided by intestinal plication.

Table 3 Odds ratio analysis for the selection of internal intestinal plication.

Variable: Predictor	Reconstruction (n = 15)	Plication (n = 13)	OR	95%CI	P value
Tumor recurrence	4	8	4.00	0.77-20.8	0.09
Radiation enteritis	5	4	0.75	0.13-4.23	0.74
Empty pelvis syndrome	3	0	0.00	0.00-NA	0.23
Surgical injury	2	1	0.43	0.03-5.56	0.52

Multivariate evaluation assessing the likelihood of a specific pathology dictating the use of internal intestinal plication. Tumor recurrence represents the strongest clinical predictor for selecting plication, reflecting the surgical necessity to manage diffuse, malignant adhesive disease rather than attempting to fill a non-existent pelvic void. Patients with tumor recurrence had 4-fold higher odds of undergoing internal intestinal plication (group 2) compared to reconstruction, reflecting the preference for plication in frozen, malignant pelvises. OR: Odds ratio; CI: Confidence interval; NA: Not applicable.

Clinical outcomes

The overall successful fistula closure rate was 89.3% (25/28). Group 1 (reconstruction) and group 2 (plication) achieved closure rates of 86.7% (13/15) and 92.3% (12/13), respectively. There was no statistically significant difference in closure efficacy between the two strategies (P = 1.0), indicating that both techniques are highly effective when tailored to the appropriate pathology (Table 4). The overall 90-day mortality rate was 7.1% (2/28).

Table 4 Postoperative clinical outcomes, n (%)/median (minimum-maximum)/mean ± SD.

Outcome measure	Group 1: Reconstruction (n = 15)	Group 2: Plication (n = 13)	Total (n = 28)	P value
Fistula closure rate	13 (86.7)	12 (92.3)	25 (89.3)	1.0
Duration of follow-up/OS in months1	32.0 (9.0-107.0)	37.0 (11.0-57.0)	34.5 (9.0-107.0)	0.66
90-day mortality	1 (6.7)	1 (7.7)	2 (7.1)	1.0
Postoperative flatus in days	4.2 ± 1.5	6.8 ± 2.1	5.4 ± 2.2	0.0021
Complications, by Clavien-Dindo				0.65
Grade I-II: Minor	4 (26.7)	5 (38.5)	9 (32.1)
Grade III-IV: Major	4 (26.7)	3 (23.1)	7 (25.0)
Stoma status at discharge
Stoma free/reversed	3 (20.0)	1 (7.7)	4 (14.3)
Permanent/temporary stoma	12 (80.0)	12 (92.3)	24 (85.7)

¹Continuous variables are presented as the mean ± SD; P-value is from the Student’s t-test.

Summary of critical postoperative outcomes. Both techniques demonstrated exceptionally high success rates in achieving durable fistula closure with no statistical superiority. However, group 2 (plication) experienced significantly longer time to flatus, consistent with the postoperative ileus often being associated with extensive intestinal manipulation. Continuous variables are presented as the median (range) due to non-normal distribution; P-values were calculated using the two-sample Wilcoxon/Mann-Whitney U test. OS: Overall survival.

For long-term outcomes across all patients, the median follow-up (representing OS time) was 34.5 months (range: 9.0-107.0 months). Survival times were comparable between the two strategies. Group 1 had a median survival of 32.0 months (range: 9.0-107.0 months) and group 2 had a median of 37.0 months (range: 11.0-57.0 months). This difference was not statistically significant (P = 0.66; Table 4). These findings indicate that while etiology-driven surgical approaches successfully manage the mechanical drivers of fistulation and yield high closure rates, OS remains comparable between both cohorts. Instead, long-term survival appears to be dictated by the behavior of the primary pelvic malignancy.

Morbidity and recovery

The overall rate of severe complications (Clavien-Dindo grade III-IV) was comparable between group 1 (26.7%) and group 2 (23.1%), yielding an overall rate of 25.0%. Despite this similarity, the two groups significantly differed in their gastrointestinal recovery. Patients in group 2 (plication) had a significantly longer mean time to postoperative flatus compared to group 1 (6.8 ± 2.1 days vs 4.2 ± 1.5 days; P = 0.002). This delay is consistent with the extensive intestinal manipulation and obligatory ileus associated with the internal plication technique (Table 4).

DISCUSSION

This comprehensive, single-center retrospective analysis of 28 consecutive cases addresses one of the major challenges in surgical oncology: Complex PSFs following pelvic malignancy. The findings unequivocally demonstrate that a standardized, etiology-driven surgical algorithm yields an exceptional overall fistula closure rate of 89.3%, coupled with an acceptable 90-day mortality rate of 7.1% and a major morbidity profile of 25.0%. Historically, the management of enteroperineal fistulas following radical pelvic exenteration was associated with devastating outcomes. Early foundational studies, such as those by Berman et al[10], demonstrated significant mortality rates associated with fistula, with death rates climbing to 13% and major morbidity exceeding 34% in similar demographic cohorts. Our mortality rate dropped significantly to 7.1%. This improvement is largely attributed to advances in perioperative care, total parenteral nutrition, and more targeted surgical approaches. Based on a series of 15 consecutive cases, Turrini et al[11] outlined guidelines for treating perineal sinuses following pelvic exenteration, noting that while surgical management is feasible, it poses significant challenges. Our study expands on this groundwork by matching specific surgical techniques to distinct patient characteristics, ultimately offering a clearer clinical pathway to improve and standardize outcomes. The 25% major complication rate underscores the severity of the population. Our findings challenge the traditional “one-size-fits-all” approach for treating the hostile pelvis, suggesting instead that the choice between pelvic floor reconstruction and internal intestinal plication should be dictated by whether the primary pathology is a structural defect (EPS) or diffuse adhesive disease (radiation enteritis).

The empty pelvis and role of reconstruction

The dead space created by pelvic exenteration is a distinct pathological entity that increases the risk of fluid accumulation and bowel obstruction. In the absence of supporting viscera, the rigid pelvic cavity allows small bowel loops to prolapse and adhere to the perineal wound. Without adequate support, the combined effects of gravity and peristalsis generate shearing forces, which can result in pressure necrosis and fistula formation[12]. In our cohort, patients with EPS or large defects (group 1) were managed with modified pelvic floor reconstruction. Unlike the intestinal bypass with distal limb exclusion recently reported by Aw et al[13] as a surgical management strategy for PSF, our approach focuses not only on obliterating the dead space but also on resecting the affected intestinal segment. By completely removing the fistula, this technique resolves the associated symptoms. Utilizing vascularized tissue (pedicled omentum) or biological mesh to recreate the pelvic floor stops the small bowel from entering the perineal space, which prevents recurrent erosion and fistulas. Additionally, the pedicled omentum is easier to use for intrapelvic filling than a small-bowel seromuscular flap[14].

Radiation enteritis and the revival of plication

Unlike EPS, patients with chronic radiation enteritis exhibit a frozen pelvis characterized by dense fibrosis rather than empty space[15]. In these cases (group 2), recurring fistulas are primarily caused by the reformation of pathological adhesions triggering bowel obstruction and anastomotic failure, rather than tissue prolapse. Unlike the Nobel procedure[16], which uses external sutures (seromuscular stitches) to align loops, internal plication uses a tube to create a stable, temporary scaffold. Our data suggest that it has specific utility in the salvage setting. Inserting a guiding tube into the intestinal lumen transforms uncontrolled, obstructive adhesions into controlled, non-obstructive ones. This “internal stenting” protects the repaired fistula from high intraluminal pressure and recurrent obstruction, a benefit that simple adhesiolysis cannot provide[17].

Clinical implications and contrasting alternative therapies

In this cohort, 85.7% of patients were discharged with either a permanent or temporary proximal diversion stoma (Table 3). This highlights that while the etiology-driven algorithm successfully manages the acute, life-threatening sepsis of a perineal fistula, the underlying pelvic musculature and sphincter complexes are often permanently destroyed by the primary malignancy and initial exenteration. Therefore, clinicians must use a cautious definition of “success” in these cases; while surgery preserves life and stops chronic perineal drainage, it rarely restores normal gastrointestinal function or continence.

Myocutaneous flaps: Myocutaneous flaps are primarily designed to reconstruct skin defects, with the associated muscle bulk also serving to partially obliterate the pelvic cavity. However, the use of muscle tissue in these flaps can introduce additional morbidity, such as flaps dehiscence[18]. Although a retrospective study[19] of 130 consecutive cases found that the use of the vertical rectus abdominis myocutaneous (VRAM) flap was significantly associated with higher rates of flap failure, early return to the operating room, wound dehiscence, and infection compared to the inferior gluteal artery myocutaneous flap, a separate systematic review identified VRAM as the most commonly used myocutaneous flap. The review also reported that both inferior gluteal artery myocutaneous and VRAM flaps are associated with relatively high overall morbidity rates[20]. Unlike VRAM flaps, which present a high risk of donor site morbidity, including hernia rates up to 16.7%, and may be contraindicated in patients with compromised abdominal walls due to stomas or previous incisions[21], the omentum serves as a pliable, vascularized biological filler. Our group 1 strategy (modified reconstruction) often utilized the greater omentum or biological mesh. Our results suggest that for PSF repair - unlike initial exenteration - an omental flap often provides sufficient bulk and vascularity to fill the dead space without the added morbidity of a myocutaneous harvest. This assumes the omentum is uncompromised by prior surgeries. If the omentum was completely removed during previous gynecologic tumor surgery, biological mesh was used instead for pelvis reconstruction. A large observational cohort study[22] found that omentoplasty was associated with a reduction in pelvic bowel obstruction, while the use of biological mesh resulted in lower reconstruction-related morbidity compared to perineal flaps. These findings support the effectiveness of our group 1 strategy (modified reconstruction).

Simple resection and anastomosis: A common pitfall in fistula surgery is performing a simple segmental resection and primary anastomosis without optimizing the pelvic environment. In patients who have undergone irradiation, this approach is particularly susceptible to anastomotic memory, in which the new anastomosis adheres to existing areas of pelvic inflammation, increasing the risk of fistula recurrence[23]. The strategy employed in our study is specifically designed to actively mitigate the risk of recurrent fistula formation by addressing the anatomical and physiological vulnerabilities inherent to the post-exenteration pelvis. Pelvic floor reconstruction creates a neoperineum that isolates abdominal organs from the empty pelvic cavity. This barrier prevents bowel loops from herniating or prolapsing into the pelvis, protecting the intestines from contacting the raw perineal wound surface. By restoring mechanical support and compartmentalization, this approach not only decreases the likelihood of bowel obstruction and fistula formation but also reduces the potential for postoperative adhesions and related complications. Overall, the reconstructive strategy provides a more stable and protective environment for healing, significantly improving surgical outcomes in patients undergoing pelvic exenteration.

Limitations of the study

The limitations of this study include its retrospective nature and the relatively small sample size (n = 28), although this represents a significant volume for such a rare pathology. Although the surgical approach was driven by surgeon preference, its strong correlation with etiology suggests the implicit application of a consistent decision-making process. Additionally, the lack of prospectively collected quality of life data limits our ability to compare these findings with non-operative management in the future. The retrospective nature of the study precluded robust multivariate regression to control for key systemic confounders, such as preoperative albumin and diabetes, that undoubtedly impact tissue healing.

CONCLUSION

Surgical management of PSF is feasible with high success rates when an etiology-driven approach is adopted. We recommend pelvic floor reconstruction for patients with EPS to obliterate dead space, and internal intestinal plication for patients with radiation enteritis to manage adhesive disease. Standardizing this algorithm across major hospitals can drastically reduce variability in surgical care, limit postoperative mortality, and optimize outcomes for this highly vulnerable patient population.

ACKNOWLEDGEMENTS

We thank Dr. Sinthu Kosasih for his advice on language editing.