AIM: To compare the efficacy of acamprosate and naltrexone in the treatment of alcohol dependence. DESIGN: A double-blind, placebo-controlled trial. SETTING: Three treatment centres in Australia. PARTICIPANTS: A total of 169 alcohol dependent subjects were given naltrexone (50 mg/day), acamprosate (1998 mg/day) or placebo for 12 weeks. INTERVENTION: All subjects were offered manualized compliance therapy, a brief intervention that targets problems that may affect treatment compliance such as ambivalence and misperceptions about medication. MEASUREMENTS: Time to the first drink, time to first relapse, drinks per drinking day and cumulative abstinence. FINDINGS: In intention-to-treat analyses, there were no differences between groups on outcome measures of drinking, craving or biochemical markers. Similarly, analyses of the 94 subjects that completed the study in full and demonstrated 80% compliance, revealed no significant treatment effects. Differential treatment effects were identified after stratification according to scores on the Alcohol Dependence Scale (ADS) and Depression Anxiety and Stress Scale (DASS). A significant beneficial treatment effect on time to first relapse was revealed for subjects with 'no depression' allocated to naltrexone (n = 56; P < 0.01). In addition, a significant beneficial treatment effect was revealed in subjects with 'low dependence' allocated to naltrexone (n = 34; P < 0.05). CONCLUSIONS: The results of this study support the efficacy of naltrexone in the relapse prevention of alcoholism amongst those with low levels of clinical depression and alcohol dependence severity. No effect of acamprosate was found in our sample.