Randomized data evaluating the impact of the extracorporeal cardiopulmonary resuscitation (ECPR) approach on long-term clinical outcomes in patients with refractory out-of-hospital cardiac arrest (OHCA) are lacking. The objective of this follow-up study was to assess the long-term clinical outcomes of the ECPR-based versus CCPR approach. The Prague OHCA trial was a single-center, randomized, open-label trial. Patients with witnessed refractory OHCA of presumed cardiac origin, without return of spontaneous circulation, were randomized during ongoing resuscitation on scene to conventional CPR (CCPR) or an ECPR-based approach (intra-arrest transport, ECPR if ROSC is not achieved prehospital and immediate invasive assessment). From March 2013 to October 2020, 264 patients were randomized during ongoing resuscitation on scene, and 256 patients were enrolled. Long-term follow-up was performed 5.3 (interquartile range 3.8-7.2) years after initial randomization and was completed in 255 of 256 patients (99.6%). In total, 34/123 (27.6%) patients in the ECPR-based group and 26/132 (19.7%) in the CCPR group were alive (log-rank P = 0.01). There were no significant differences between the treatment groups in the neurological outcome, survival after hospital discharge, risk of hospitalization, major cardiovascular events and quality of life. Of long-term survivors, 1/34 (2.9%) in the ECPR-based arm and 1/26 (3.8%) in the CCPR arm had poor neurological outcome (both patients had a cerebral performance category score of 3). Among patients with refractory OHCA, the ECPR-based approach significantly improved long-term survival. There were no differences in the neurological outcome, major cardiovascular events and quality of life between the groups, but the trial was possibly underpowered to detect a clinically relevant difference in these outcomes. Trial registration ClinicalTrials.gov Identifier: NCT01511666, Registered 19 January 2012.