Multiple Sclerosis (MS) annualized relapse rates (ARRs) in trials may be declining due to changes in diagnostic criteria, MS etiology, study criteria, and selection biases. This review examines if there is a trend in the ARR for relapsing-remitting MS patients (RRMS) over time and if so, why. A comprehensive literature search was performed using PubMed, Web of Science, and the Cochrane Library using electronic searches, screen scraping for abstracts, and hand searching of references for randomized trials conducted between 1960 and 2008. Out of 72 randomized trials, 56 (77.8%) defined relapse. This study uses 32 placebo relapsing-remitting studies out of the 37 (66.1%) with RRMS. The mean ARR for the treatment arms was 0.68 and the one for the placebo groups was 1.002. The year of publication was negatively associated with the ARR (p = 0.0001). The annual reduction amounts to 0.36 relapses over a 10-year period. Age and duration of symptoms were negatively associated with the ARR. Year of publication was significantly negatively associated with ARR after controlling for covariates. ARRs have fallen with relapse definition, entrance criteria remain important, but time exceeds all these variables and reflects two likely sources, selection of patients for trials by clinicians and rescue of patients truncating the number of multiple relapses. The impact of truncating the number of relapses on the falling rates is important, not only on the ARRs, but also on the impact of informative censoring in drop-outs. © The Author(s) 2010.