Resection with primary anastomosis vs nonrestorative resection for perforated diverticulitis with peritonitis: a...

It is still controversial whether the optimal operation for perforated diverticulitis with peritonitis is primary anastomosis (PRA) or nonrestorative resection (NRR). The aim of this systematic review and meta-analysis was to evaluate mortality and morbidity rates following emergency resection for perforated diverticulitis with peritonitis and ostomy reversal, as well as ostomy nonreversal rates.

The Pubmed, EMBASE, Cochrane Library, MEDLINE via Ovid, CINAHL and Web of Science databases were systematically searched. Mortality was the primary end-point. A subgroup meta-analysis of randomized controlled trials was performed in addition to a meta-analysis of all eligible studies. Odds ratios (ORs) and mean difference (MD) were calculated for dichotomous and continuous outcomes, respectively.

Seventeen studies, including three randomized controlled trials (RCTs), involving 1016 patients (392 PRA vs 624 NRR) were included. Overall, mortality was significantly lower in patients with PRA compared with patients with NRR [OR (95% CI) = 0.38 (0.24, 0.60), P < 0.0001]. Organ/space surgical site infection (SSI) [OR (95% CI) = 0.25 (0.10, 0.63), P = 0.003], reoperation [OR (95% CI) = 0.48 (0.25, 0.91), P = 0.02] and ostomy nonreversal rates [OR (95% CI) = 0.27 (0.09, 0.84), P = 0.02] were significantly decreased in PRA. In the RCTs, the mortality rate did not differ [OR (95% CI) = 0.46 (0.15, 1.38), P = 0.17]. The mean operating time for PRA was significantly longer than for NRR [MD (95% CI) = 19.96 (7.40, 32.52), P = 0.002]. Organ/space SSI [OR (95% CI) = 0.28 (0.09, 0.82), P = 0.02] was lower after PRA. Ostomy nonreversal rates were lower after PRA. The difference was not statistically significant [OR (95% CI) = 0.26 (0.06, 1.11), P = 0.07]. However, it was clinically significant [number needed to treat/harm (95% CI) = 5 (3.1, 8.9)].

This meta-analysis found that organ/space SSI rates as well as ostomy nonreversal rates were decreased in PRA at the cost of prolonging the operating time.