RT Journal Article SR Electronic T1 Analysis of clinical outcomes according to original treatment groups 16 years after the pivotal IFNB-1b trial JF Journal of Neurology, Neurosurgery & Psychiatry JO J Neurol Neurosurg Psychiatry FD BMJ Publishing Group Ltd SP 907 OP 912 DO 10.1136/jnnp.2009.204123 VO 81 IS 8 A1 G C Ebers A1 A Traboulsee A1 D Li A1 D Langdon A1 A T Reder A1 D S Goodin A1 T Bogumil A1 K Beckmann A1 C Wolf A1 A Konieczny A1 for the Investigators of the 16-year Long-Term Follow-Up Study YR 2010 UL http://jnnp.bmj.com/content/81/8/907.abstract AB Background Evidence for efficacy of disease-modifying drugs in multiple sclerosis (MS) comes from trials of short duration. We report results from a 16 y, retrospective follow-up of the pivotal interferon β-1b (IFNB-1b) study.Methods The 372 trial patients were randomly assigned to placebo (n=123), IFNB-1b 50 μg (n=125) or IFNB-1b 250 μg (n=124) subcutaneously every other day for at least 2 y. Some remained randomised for up to 5 y but, subsequently, patients received treatment according to physicians' discretion. Patients were re-contacted and asked to participate. Efficacy related measures included MRI parameters, relapse rate, the Expanded Disability Status Scale, the Multiple Sclerosis Functional Composite Measure and conversion to secondary progressive MS.Results Of the 88.2% (328/372) of patients who were identified, 69.9% (260/372) had available case report forms. No differences in outcome between original randomisation groups could be discerned using standard disability and MRI measures. However, mortality rates among patients originally treated with IFNB-1b were lower than in the original placebo group (18.3% (20/109) for placebo versus 8.3% (9/108) for IFNB-1b 50 μg and 5.4% (6/111) for IFNB-1b 250 μg).Conclusions The original treatment assignment could not be shown to influence standard assessments of long-term efficacy. On-study behaviour of patients was influenced by factors that could not be controlled with the sacrifice of randomisation and blinding. Mortality was higher in patients originally assigned to placebo than those who had received IFNB-1b 50 μg or 250 μg. The dataset provides important resources to explore early predictors of long-term outcome.