Flupenthixol, with its broad receptor profile, interacts with a variety of dopamine and serotonin binding sites which are important in the neurobiology of alcohol dependence. Its pharmacology, together with encouraging results from both animal studies and clinical trials with cocaine users, led us to postulate that flupenthixol would significantly prevent relapse in detoxified alcohol-dependent individuals. We conducted a prospective, randomized, double-blind, placebo-controlled, multi-centre trial with two parallel groups and appropriate statistical evaluation. Subjects met criteria for moderate to severe alcohol dependence (DSM-III-R), without any concomitant psychiatric disorder. After complete detoxification, 281 women and men received either 10 mg of flupenthixol decanoate or placebo as i.m. injection every second week for 6 months on an out-patient basis, followed by 6 months of follow-up. Efficacy was based on absolute abstinence, with relapse being defined as consumption of any alcohol after inclusion in the study. In contrast to the hypothesis, flupenthixol did not reduce, but was associated with more, relapses. Though well tolerated, relapse rates after 6 months of treatment were 85.2% (flupenthixol) versus 65.5% (placebo), a highly significant difference from the medication. Flupenthixol was also inferior to placebo with regard to other secondary criteria of efficacy (cumulative abstinence duration, relapse rate after 12 months). These results indicate that a 10 mg dose of flupenthixol decanoate does not have a beneficial effect on abstinence maintenance in alcohol-dependent individuals.