New concepts of clinical trials in rheumatoid arthritis: a boom of noninferiority trials

The purpose is to describe the most recent randomized controlled trials (RCT) in patients with rheumatoid arthritis that had a noninferiority design, and to focus on methodological aspects of noninferiority. In 2014 and 2015 10 different RCTs with a noninferiority-design could be identified, in comparison to only a few in the decade before. Most RCTs had a rather small sample size, and had ill-defined noninferiority-margins, or noninferiority-margins without comprehensible clinical meaning. Six of the 10 trials indeed arrived at a conclusion of 'noninferiority'; four did not. Interestingly, many of the RCTs were pragmatic studies comparing strategies, and the investigators were neither blind to the treatment nor to the outcome. In addition, the treatments were often adaptive (e.g. treat-to-target approach). These characteristics are considered built-in incentives for noninferiority. In the competitive pharmaceutical landscape of rheumatoid arthritis, with many effective drugs and strategies, it is no surprise that the number of noninferiority-trial (sharply) rises. But noninferiority trials are difficult to design, conduct, and interpret, and many principles of noninferiority-trial designs are currently ignored, which may jeopardise their conclusions to some extent