BACKGROUND: Refractory status epilepticus (RSE) has a mortality of 16-39%; coma induction is advocated for its management, but no comparative study has been performed. We aimed to assess the effectiveness (RSE control, adverse events) of the first course of propofol versus barbiturates in the treatment of RSE. METHODS: In this randomized, single blind, multi-center trial studying adults with RSE not due to cerebral anoxia, medications were titrated toward EEG burst-suppression for 36-48 h and then progressively weaned. The primary endpoint was the proportion of patients with RSE controlled after a first course of study medication; secondary endpoints included tolerability measures. RESULTS: The trial was terminated after 3 years, with only 24 patients recruited of the 150 needed; 14 subjects received propofol, 9 barbiturates. The primary endpoint was reached in 43% in the propofol versus 22% in the barbiturates arm (P = 0.40). Mortality (43 vs. 34%; P = 1.00) and return to baseline clinical conditions at 3 months (36 vs. 44%; P = 1.00) were similar. While infections and arterial hypotension did not differ between groups, barbiturate use was associated with a significantly longer mechanical ventilation (P = 0.03). A non-fatal propofol infusion syndrome was detected in one patient, while one subject died of bowel ischemia after barbiturates. DISCUSSION: Although undersampled, this trial shows significantly longer mechanical ventilation with barbiturates and the occurrence of severe treatment-related complications in both arms. We describe practical issues necessary for the success of future studies needed to improve the current unsatisfactory state of evidence.