Perforated diverticulitis: To anastomose or not to anastomose? A systematic review and meta-analysis.

BACKGROUND: No consensus has been reached in the management of perforated diverticulitis. Many surgeons opt for a Hartmann's procedure to avoid the risk of an anastomotic leak. We hypothesise that resection with primary anastomosis is a safe alternative in selected patients. We aim to conduct a systematic review and meta-analysis on the available literature.

METHODS: Studies that compared emergency Hartmann's with primary anastomosis in perforated left sided colonic diverticulitis were systematically reviewed. The search strategy included all study types that compared primary anastomosis to Hartmann's in perforated diverticulitis and reported on morbidity and mortality. 5 databases (PubMed, MEDLINE via PubMed, OVID, EMBASE via OVID and The Cochrane Collaboration). The Cochrane's Bias Methods Group tool was used to assess the risk of bias and a meta-analysis of the relevant studies was conducted.

RESULTS: The review retrieved 1933 abstracts of which 14 studies (2 RCTs, 4 prospective non-randomised and 8 retrospective non-randomised) with 765 patients in total, 482 in the Hartmann's group and 283 in the primary anastomosis group, met the inclusion criteria. This showed a significantly lower mortality with primary anastomosis (10.6%) compared to Hartmann's (20.7%) (p = 0.0003). Morbidity was also significantly lower (41.8% vs. 51.2%) (p = 0.0483). The RR for mortality was 0.92 in favour of primary anastomosis (p = 0.0019). The average anastomotic leak rate was 5.9%.

CONCLUSION: Resection and primary anastomosis should be considered as a feasible and safe operative strategy in selected patients with perforated diverticulitis. There is however a paucity of high level evidence and further research is needed.