Comparison of the risk of infections in different anti-TNF agents: a meta-analysis

Abstract

Background: Anti-tumor necrosis factor α (anti-TNF-α) treatments are widely used in patients with rheumatoid arthritis (RA); however, the increased risk of infections is one of the most important side effects of anti-TNF-α agents. This study evaluated the differences between monoclonal antibodies and the soluble receptor for infections in patients with RA by direct comparison of observation studies. Methods: A systemic literature search was conducted in March 2014 and an up-to-date search was conducted in August 2014. All studies reporting infections in RA patients treated with the soluble receptor (ETA [etanercept]) and at least one of monoclonal antibodies (INF [infliximab], ADA [adalimumab]) were included. Results: Twelve articles were finally included. The meta-analysis revealed that compared with monoclonal antibodies, the soluble receptor had a lower incidence rate of serious infections (relative risk [RR] = 0.63 [0.40–0.97] P = 0.04), but we have to notice that the heterogeneity was high (I2 = 85%) and publication bias might exist. As to tuberculosis, the pooled analysis revealed that the soluble receptor had a lower risk (RR = 0.19 [0.06–0.56] P = 0.003) and its heterogeneity was low (I2 = 0%) while no publication bias was observed. For general infections, ETA had a lower risk compared with mono-antibodies and its heterogeneity was high (RR = 0.66 [0.49–0.89] P < 0.00001 I2 = 79%). Conclusion: Compared with mono-antibodies, the soluble receptor has a lower risk for tuberculosis and general infections. But as to serious infections, the answer is uncertain due to its high heterogeneity and possibility of publication bias. More well-designed long-term prospective studies would be important to strengthen these findings.