Yoshida et al present an overview of different designs and ‘failure definitions’ in biologic discontinuation studies in rheumatoid arthritis (RA).1 We feel it is a very important review, as the number of discontinuation studies is increasing and, therefore, awareness of the heterogeneity in these study designs as demonstrated in this review is essential. We have, however, a few comments. First, one of the difficulties in interpreting discontinuation studies is that different protocols can be used, discontinuation protocol or dose-tapering protocols, and that interpretation is sometimes not straightforward. This is exemplified by …