Assessment of morbidity and predictors of wound complications following perineal wound closure after radical anorectal oncologic resection: retrospective cohort study

Authors: Bercz A, Alvarez J, Rosen R, Drescher M, Sonoda H, Karagkounis G, Wei I, Widmar M, Nash GM, Weiser MR, Paty PB, Allen RJ, Nelson JA, Coriddi M, Dayan JH, McCarthy C, Shahzad F, Matros E, Disa JJ, Cordeiro PG, Mehrara BJ, Garcia-Aguilar J, Smith JJ, Pappou EP 

Affiliation: Memorial Sloan Kettering Cancer Center, Colorectal Surgery Service and Plastic Surgery Service, New York, NY 

Journal: BJS Open, May 2025 

PMID: 40503607 

Key takeaways 

• Tissue flap reconstruction (TFR) was used in more complex resections and showed higher unadjusted 90-day dehiscence and infection than primary closure (PC). 

• In subgroup analysis, dehiscence was significantly higher in patients who underwent APR. There was no significant difference between TFR and PC in cases of total pelvic exenteration (TPE). 

• Flap type (VRAM, gluteal and gracilis) was not associated with the extent of resection, and there were no differences in wound infection, dehiscence or reintervention within 90 days between flap subtypes 

• Omental flaps in PC did not change wound outcomes 

• Independent predictors of dehiscence were anal cancer and extralevator abdominoperineal resection.  

• Pelvic exenteration independently predicted wound infection 

Background 

Perineal wound complications after abdominoperineal resection and pelvic exenteration remain common, particularly in irradiated fields and large dead-space defects, delaying recovery and adjuvant therapy. 

Objective 

Assess postoperative morbidity and identify predictors of wound complications across perineal closure strategies after radical anorectal oncologic resection. 

Methods 

• Design: Retrospective cohort at a comprehensive cancer center, January 2012–December 2020; STROBE-adherent; Therapeutic Level III. 

• Population: 414 patients with rectal (364) or anal (50) malignancy undergoing abdominoperineal resection (including extralevator abdominoperineal resection) or pelvic exenteration. 

• Interventions and comparators: 

• Tissue flap reconstruction: 150 patients, including VRAM (101), gluteal V-Y advancement (37), gracilis (12). 

• Primary closure: 264 patients, with pedicled omental flap in 81. 

• Endpoints: 

• 90-day outcomes: wound dehiscence, wound infection, transfusion, readmission, invasive reintervention, Clavien–Dindo complications, flap loss. Definitions were explicitly provided. 

• Long-term outcomes: chronic hernia and non-healing wound beyond 6 months. 

• Statistics: Fisher exact and Wilcoxon rank-sum for group comparisons; multivariable logistic regression for independent predictors; significance threshold P < 0.05. 

Results 

• Cohort characteristics: Median age 61 years; median follow-up 34.3 months. 

• Case-mix differences: Tissue flap reconstruction preferentially used in higher-complexity scenarios including exenteration, sacrectomy, vaginectomy, and intraoperative radiation. 

• Unadjusted 90-day morbidity (tissue flap reconstruction vs primary closure): 

• Wound dehiscence: 27% vs 11%. 

• Wound infection: 25% vs 14%. 

• Grade 3 or higher Clavien–Dindo: 32% vs 17%. 

• Flap loss: 1% overall in tissue flap reconstruction (2 patients). 

• Flap subtype comparisons: No differences in wound infection, wound dehiscence, readmission, or reintervention among VRAM, gluteal, and gracilis; VRAM associated with higher transfusion requirement. 

• Primary closure with omental flap vs without: No differences in wound infection, dehiscence, readmission, reinterventions, or long-term hernia/non-healing wound. 

• Independent predictors (multivariable): 

• Wound dehiscence: anal cancer (OR 5.24) and extralevator resection (OR 3.09). 

• Wound infection: pelvic exenteration (OR 17.8). 

• Closure method and intraoperative radiation were not independent predictors of dehiscence or infection in the full cohort. 

Conclusion 

In this cohort, higher raw wound morbidity tracked with reconstructive flap use because flaps were used in more complex resections; after adjustment, closure method did not independently predict dehiscence or infection, while anal cancer, extralevator resection, and pelvic exenteration predicted complications. 

Strengths 

• Large, detailed single-institution dataset with standardized definitions of wound outcomes and 90-day morbidity. 

• Includes multiple flap options and an omental-flap primary-closure subgroup, allowing clinically relevant comparisons. 

Limitations 

• Strong selection bias and confounding by indication: flap reconstruction was preferentially used for exenteration, sacrectomy, vaginectomy, and intraoperative radiation, which inflates unadjusted morbidity comparisons. 

• Key reconstructive drivers like defect size and dead-space volume were not captured, limiting mechanistic inference about technique choice and outcomes. 

• Functional recovery and patient-reported outcomes were not consistently available, leaving the most patient-important differences unanswered. 

Clinical relevance 

For reconstructive planning after abdominoperineal resection or exenteration, this paper reinforces that the patient’s oncologic context and extent of extirpation drive wound risk more than the type of flap or method of closure. It does not directly answer the operative question of how a given defect should be closed; it retrospectively reports outcomes based on how surgeons elected to close defects in practice. In straightforward abdominoperineal resection, primary closure appears reasonable when tension and dead space are controlled; in anal cancer salvage and extralevator resections, anticipate higher dehiscence risk and plan resources accordingly. 

Editorial Notes 

• The unadjusted comparison is expected to favor primary closure because flaps were reserved for larger, higher-risk defects. The multivariable analysis is the clinically actionable message. 

• The regression model cannot correct for unmeasured confounders that dominate reconstructive decision-making: defect size, pelvic dead-space volume, levator geometry, tissue quality, and prior perineal wound history. 

• The flap subtype analysis is likely underpowered for meaningful comparisons, especially for gracilis. Grouping abdominal based muscle flaps and fasciocutaneous V-Y flaps confounds comparisons as these are generally chosen for differing reasons.  

• The anal cancer predictor has a wide confidence interval, suggesting limited events and potential model instability. 

• What was the institutional algorithm for selecting VRAM versus gluteal versus gracilis, and did it