Injection site reactions (ISRs) are a common side effect of subcutaneous interferon beta therapy, particularly during initiation of therapy. Retrospective analysis of two clinical trials showed that patients using an autoinjector experienced fewer ISRs than patients administering interferon beta manually. This randomized, open-label trial compared the occurrence of ISRs in relapsing-remitting multiple sclerosis patients subcutaneously injecting interferon beta-1a manually or with autoinjector. In total, 1825 patients (autoinjector, 932, manual injection, 893) were included in the intention-to-treat analysis. Significantly fewer patients using the autoinjector experienced ISRs, based on physician assessment, compared with manual injection (78.7 versus 85.4%; P <0.001). There was no statistical difference on primary study endpoint: number of patients experiencing moderate to severe ISRs after 12 weeks' therapy (25.3% versus 23.2%, P =0.449). The patient-reported proportion of any ISR during the treatment period was significantly greater for the manual injection group (71.8% versus 66.1; P<0.001). The decreased incidence of ISRs with the autoinjector compared to manual injection seen in this short-term study, coupled with ease of use of the autoinjector, suggest that it could improve compliance, and therefore therapeutic outcomes in some patients. © 2005 Edward Arnold (Publishers) Ltd.