Twelve placebo-controlled RCTs were included in the review: three rituximab trials (n=1,143), five abatacept trials (n=2,945) and four anakinra trials (n=2,771). All the trials were double-blind and performed an intention-to-treat analysis. Ninety-four percent of the included participants were followed-up.
Meta-analyses indicated that overall risk of serious infection was not significantly increased by rituximab (odds ratio 1.45, 95% CI: 0.56, 3.73), abatacept (odds ratio 1.35, 95% CI: 0.79, 2.32), or anakinra (odds ratio 2.75, 95% CI: 0.90, 8.35). The risk was significantly increased for high-dose anakinra (i.e. 100mg or more) relative to placebo (odds ratio 3.40, 95% CI: 1.11, 10.46) and low-dose anakinra (odds ratio 9.63, 95% CI: 1.31, 70.91), but these results were not significant after excluding participants with comorbidity factors.
None of the investigated factors (age, concomitant steroid intake, rheumatoid factor status) appeared to be confounding factors of infections.