Meta-analysis of the role of granulomas in the recurrence of Crohn disease

Abstract

Purpose: This study used meta-analytical techniques to compare the recurrence of granulomatous vs nongranulomatous Crohn disease.

Methods: Comparative studies published between 1954 and 2007 of granulomatous vs nongranulomatous Crohn disease were included. Using a random effects model, end points evaluated were the number of recurrences and reoperations, and the time to recurrence and reoperation, of granulomatous vs nongranulomatous Crohn disease. Heterogeneity was assessed and sensitivity analysis was performed to account for bias in patient selection.

Results: Twenty-one studies (14 nonrandomized retrospective, 7 nonrandomized prospective) reported on 2236 patients with Crohn disease, of whom 1050 (47.0%) had granulomas (granulomatous group) and 1186 (53.0%) had no granulomas (nongranulomatous group). The number of recurrences and reoperations was found to be significantly higher in the granulomatous group compared to the nongranulomatous group (odds ratio 1.37, P = .04; odds ratio 2.38, P < .001; respectively), with significant heterogeneity between studies (P = .06; P < .001; respectively). The time to recurrence and reoperation was significantly shorter in the granulomatous group compared with the nongranulomatous group (hazard ratio 1.63, P = .001; hazard ratio 1.62, P = .002; respectively), with no significant heterogeneity between studies. The number of recurrences and reoperations remained significantly higher in the granulomatous group compared to the nongranulomatous group during sensitivity analysis of higher-quality studies, more recent studies, and studies with a larger group of patients.

Conclusions: Granulomatous Crohn disease appears to be associated with a higher number of recurrences and reoperations and a shorter time to recurrence and reoperation compared to nongranulomatous Crohn disease. Because of significant heterogeneity between studies, further studies should be undertaken to confirm these findings.