Background: The overall risk for infection with contemporary biological agents in treating Crohn's disease (CD) and ulcerative colitis (UC) has not been systematically assessed.
Methods: We performed a PubMed and Cochrane database literature search to evaluate randomized, placebo-controlled trials of biologics in treating UC and CD. Meta-analysis was performed using a DerSimonian and Laird random effects model. We determined relative risk (RR) of harm against placebo; number needed to harm (NNH) was reported when appropriate. Heterogeneity and publication bias were assessed.
Results: Fourteen trials (6 UC and 8 CD) evaluating 5107 patients were included. For anti-tumor necrosis factor agents used in the treatment of UC, golimumab {NNH of 9.3, RR = 1.4 (95% confidence interval [CI], 1.04-1.8)} and pooled studies of infliximab and adalimumab (NNH = 17.2, RR = 1.2 [95% CI, 1.0-1.3]) had a statistically significant higher risk for any infection versus placebo. Risk was not significantly increased in anti-tumor necrosis factor trials in CD (RR = 1.1 [95% CI, 0.8-1.5]). By contrast, anti-integrin agents in UC (RR = 1.0 [95% CI, 0.9-1.2]) or CD (RR = 1.1 [95% CI, 0.97-1.3]) did not confer a statistically significant excess risk of infection versus placebo.
Conclusions: Anti-tumor necrosis factor therapy but not anti-integrin therapy is associated with a greater infection risk than placebo in treating UC. Neither class of therapy is associated with increased infection risk over placebo in treating CD. Our findings can help guide patient-centered discussions regarding the risk for infection with biological agents.