A propensity score matched ACS-NSQIP database analysis of primary anastomosis with or without diversion for acute perforated diverticulitis.

BACKGROUND

Resection with primary anastomosis (RPA) has emerged as an accepted alternative to Hartmann procedure for acute perforated diverticulitis with peritonitis, yet the need for routine diversion remains unclear. We compared the short-term outcomes of RPA with and without diversion in this setting.

METHODS

We conducted a retrospective analysis of the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database (2014-2021), including patients who underwent RPA for perforated diverticulitis. Patients who had or did not have a diverting stoma were propensity score matched for baseline characteristics. The primary outcomes were anastomotic leak and 30-day mortality; secondary outcomes included postoperative complications, readmission, length of stay, and operative time. A subgroup analysis was conducted in patients with preoperative sepsis.

RESULTS

Among 1228 eligible patients, 387 patients were matched in each group. Omission of diversion was not associated with an increased risk of anastomotic leak (5.5% vs 4.1%; odds ratio [OR], 1.36; 95% CI, 0.60-3.10; P =.46) or 30-day mortality (3.4% vs 1.8%; OR, 1.89; 95% CI, 0.75-4.78; P =.18). In a subgroup analysis of patients with preoperative sepsis, diversion was associated with a lower rate of clinically significant anastomotic leaks (1.2% vs 5.9%; P =.039). Postoperative morbidity was similar between groups, with no significant differences in infectious, pulmonary, renal, cardiovascular, or thromboembolic complications. Diversion was associated with significantly higher 30-day readmission rates (23.3% vs 16.4%; OR, 1.55; 95% CI, 1.01-2.37; P =.044), longer operative times (159 vs 131 minutes; P <.001), and prolonged hospital stays (9 vs 8 days; P <.001).

CONCLUSION

Based on evidence from observational data, RPA without a diverting stoma might not be associated with increased mortality or anastomotic leak in select patients with perforated diverticulitis. However, because several intraoperative factors that influence the decision to divert are not captured in NSQIP, residual confounding by indication remains possible; these findings should be interpreted with caution.