BACKGROUND Fampridine leads to significant walking improvements in many persons with multiple sclerosis (PwMS). However, a relevant proportion of PwMS does not respond to fampridine and predictors of initial drug responsiveness are unknown. METHODS Drug response to prolonged-release (PR)-fampridine was assessed in 55 PwMS using the timed 25-foot walk (T25FW), the 6-minute walk test (6MWT) and the 12-item multiple sclerosis walking scale (MSWS-12) as outcome parameters. Patients were treated with PR-fampridine and placebo each for 6 weeks in a randomised, double-blind, placebo-controlled trial with cross-over design (NCT01576354). Possible predictors of drug responsiveness were investigated by multiple correlation analysis and binary logistic regression models. An additional longitudinal analysis followed drug responses of 32 patients treated with PR-fampridine over 3 years to identify potential predictors of long-term drug responsiveness. RESULTS Severity of walking disability was positively correlated with enhanced responses to PR-fampridine. The strongest single predictor of drug responsiveness was poor 6MWT performance at baseline, which was positively correlated with enhanced drug response in the 6MWT (R=-0.541; P<0.001). A multivariable logistic regression model including 6MWT and T25FW baseline performances predicted PR-fampridine responder status with an accuracy of 85.5% (specificity: 90.0%; sensitivity: 73.3%), with a threshold of 211m in the 6MWT best separating responders from non-responders. Enhanced drug responsiveness after 3 years correlated with decline in walking endurance during this period (R=-0.634; P=0.001). CONCLUSIONS Initial walking impairment is a good predictor of therapeutic responsiveness to PR-fampridine. Valid predictors of patients' responsiveness to PR-fampridine are essential for patient stratification and optimisation of MS treatment. This article is protected by copyright. All rights reserved.