Long-term antibiotic treatment for Crohn's disease: systematic review and meta-analysis of placebo-controlled trials.

BACKGROUND We investigated the effectiveness of long-term antibiotic treatment in patients with Crohn's disease. METHODS We performed a systematic review and meta-analysis of randomized clinical trials. Data sources were Medline (from 1966 through June 2009), EMBASE (from 1980 through June 2009), Cochrane Central Register of Controlled Trials (issue 3, 2009), and references from relevant publications. Trials that compared antibiotic therapy during at least 3 months with placebo were included. Outcomes were remission in patients with active disease and relapse in patients with inactive disease. Results from intention-to-treat analyses were combined in a random-effects meta-analysis, stratified by class of drug. Odds ratios (ORs) >1 indicate superiority of antibacterial treatment over placebo. Numbers needed to treat for 1 year to keep 1 additional patient in remission were calculated. RESULTS Sixteen trials that examined 13 treatment regimens in 865 patients were included in the meta-analysis. The median duration of treatment was 6 months (range, 3-24 months). Three trials of nitroimidazoles showed benefit, with a combined OR of 3.54 (95% confidence interval [CI], 1.94-6.47). Similarly, the combined OR from 4 trials of clofazimine was 2.86 (95% CI, 1.67-4.88). For patients with active disease, the number needed to treat was 3.4 (95% CI, 2.3-7.0) for nitroimidazoles and 4.2 (95% CI, 2.7-9.3) for clofazimine. The corresponding numbers needed to treat for inactive disease were 6.1 (95% CI, 5.0-9.7) and 6.9 (95% CI, 5.4-12.0). No benefit was evident for classic drugs against tuberculosis (3 trials; OR, 0.58; 95% CI, 0.29-1.18). Results for clarithromycin were heterogeneous (I(2)=77%; P=.005) and not combined in the meta-analysis. CONCLUSIONS Long-term treatment with nitroimidazoles or clofazimine appears to be effective in patients with Crohn's disease.