In a placebo-controlled, double-blind German multicenter study (seven sites) the efficacy of naltrexone as an adjunctive treatment in alcoholism to maintain abstinence was assessed for 12 weeks. A total of 171 detoxified patients (97.7% met the DSM-III-R criteria for alcohol dependence) were included. Patients had been abstinent for a mean of 19.5 +/- 9.4 days at study entry. Eighty-four and 87 patients were randomized to receive naltrexone (50 mg/day) and placebo, respectively. Each site was instructed to provide its usual psychosocial alcohol treatment program. The primary effectiveness measure was the time to first heavy drinking as derived from self-reports of drinking (timeline-follow-back method). Secondary effectiveness measures included time to first drink, amount of alcohol consumption, intensity of craving, severity of alcoholism problems, and liver enzymes. Thirty-three (38%) placebo patients and 28 (33%) naltrexone patients discontinued the study. At endpoint, 62% of the patients in each group did not have an episode of heavy drinking. Also, there were no significant differences between the study groups concerning secondary effectiveness measures as well as compliance and adverse clinical events--with the exception of the gamma-GT, which was significantly greater reduced in the naltrexone group throughout the study. Based upon an intention-to-treat population, this study confirms the safety but not the efficacy of naltrexone in prevention of alcohol relapse. Nevertheless, the question arises whether self-reports of drinking are more reliable than gamma-GT as a measure of recent alcohol consumption.