Objective To directly compare the efficacy of natalizumab and fingolimod in patients with active relapsing-remitting multiple sclerosis. Methods This phase 4, randomised, rater-and sponsor-blinded, prospective, parallel-group, clinic-based head-to-head study was conducted at 43 sites in nine countries. Patients were randomised (1:1) to intravenous natalizumab 300 mg every 4 weeks or oral fingolimod 0.5 mg once daily for ≤52 weeks. Enrolment-related early study termination precluded assessment of the primary endpoint (evolution of new on-treatment gadolinium-enhancing (Gd+) lesions to persistent black holes). Unplanned exploratory analyses of secondary endpoints evaluated the effects of treatment on the development of new T1 Gd+ lesions and new/newly enlarging T2 lesions, lesion volumes and relapse outcomes. Results The intent-to-treat population comprised 108 patients (natalizumab, n=54; fingolimod, n=54); 63 completed ≥24 weeks of treatment. Due to the limited numbers of events and patients at risk, MRI and relapse outcomes were reported over up to 24 and 36 weeks, respectively. The mean number of new T1 Gd+ lesions was numerically lower with natalizumab than with fingolimod by 4 weeks; accumulation rates were 0.02 and 0.09 per week, respectively, over 24 weeks (p=0.004). The cumulative probability of developing ≥1 lesion at 24 weeks was 40.7% with natalizumab versus 58.0% with fingolimod (HR=0.60; 95% CI 0.31-1.16; p=0.126); the corresponding probabilities for ≥2 lesions were 11.5% vs 48.5% (HR=0.25; 95% CI 0.09-0.68; p=0.007). No significant between-group differences were observed for the other MRI outcomes at 24 weeks. The cumulative probability of relapse over follow-up was 1.9% with natalizumab versus 22.3% with fingolimod (HR=0.08; 95% CI 0.01-0.64; p=0.017). Adverse events were consistent with known safety profiles. Conclusions These results suggest that natalizumab is more efficacious than fingolimod in reducing multiple sclerosis relapses and T1 Gd+ lesion accumulation in patients with active disease. Trial registration numbers NCT02342704; EUCTR2013-004622-29-IT; Post-results.