For example, studies from the U.S. National Data Bank for Rheumatic Diseases and the U.S. Medicare system have found no increased risk of infection with TNF antagonists. While one observational study from the German Biologics Register (RABBIT) found an increased risk, the infection rate in the controls was surprisingly low; this raises the possibility that controls were drawn from a different source population or were followed in a different manner from the TNF antagonist patients.
The current study has some notable aspects. The authors found that 4% to 5% of patients had serious infections per year; this infection rate is consistent across several studies. (See Table 3, page 27, for details on the rates of serious infection in this study.) As expected, the severity of disease among patients taking TNF antagonists was greater than those on non-biologic DMARDs. Before severity adjusting, the rates of infection were higher among TNF antagonists, but the rate ratios equalized after accounting for severity.
However, the rates of intracellular infection were much higher among the TNF antagonists and the presentation of tuberculosis was atypical. In addition, there was a significant increase in the risk of soft tissue and skin infections among patients on TNF antagonists.
I will continue to discuss the possible risk of infection with patients starting TNF antagonists. My review of systems must continue to include infectious symptoms, focusing on the skin, soft tissue, and atypical manifestations of opportunistic infections. However, the current study adds to my sense that the benefits of TNF antagonists outweigh the risks.
