Background: The ideal method of repairing serious small bowel injuries remains unknown. Prior reports suggest a higher rate of enteric anastomotic-related complications (EACs) with stapled posttraumatic bowel anastomosis but did not specifically focus on the small bowel or clarify fully the actual anastomotic construction.
Methods: This was a retrospective review of patients requiring surgical repair of small bowel perforations at a Level I urban American center (Detroit Receiving Hospital [DRH]) and a Canadian provincial trauma center (Vancouver Hospital and Health Sciences Center [VHHSC]). All patients requiring a primary repair and/or resection were included. Anastomoses were hand-sewn, stapled, or combined stapling and sewing with mucosal inversion. Leaks, anastomotic fistulae, and intra-abdominal abscesses were considered specific EACs. A sample size of 53 per group was obtained to detect a 17% difference at alpha = 0.05 (one-sided) and beta = 0.2.
Results: Full-thickness small bowel injuries were repaired in 232 patients (DRH, 165; VHHSC, 67). Injuries were penetrating at DRH (91.5%) and blunt at VHHSC (65.7%). Anastomotic repairs in 127 patients (158 anastomotic repairs [DRH, 113; VHHSC, 55]) were 64 (40.5%) stapled, 38 (24.1%) hand-sewn, and 56 (35.4%) combined. Also, 105 patients had 349 primary closures of an injury. Overall, there were 24 EACs. After anastomosis, there were 11 intra-abdominal abscesses: 6 after stapling, 3 after being sewn, and 2 after a combined construction. There were four small bowel anastomotic fistulae: three after stapled-only anastomosis and one after hand-sewing. After enteroenterostomy, the EAC rate was 10.2% per patient, or 8.4% per anastomosis. After primary repairs, one patient had an anastomotic fistula, which closed spontaneously, and 11 had intra-abdominal abscesses, yielding an EAC rate of 10.6% per patient or 3.4% per repair. A primary repair was significantly less likely to be associated with an EAC than any anastomosis (p = 0.035). No method of anastomosis was statistically safer in relation to EACs, whether analyzed by patient, by anastomosis, or by considering primarily either the use of a linear stapler or the principle of inverting the mucosal approximation. Only damage control procedures and associated pancreaticoduodenal injuries were identified as statistically significant predictors using multiple logistic regression analysis.
Conclusion: Anastomotic complications after enteroenterostomy or primary repair for trauma are uncommon regardless of the technique, but surgeons must be especially cautious during or after damage control. Primary repairs are desirable, but when anastomosis is unavoidable, the method of repair should reflect that with which the surgeon is the most comfortable.