Opiate antagonist medications have been shown to improve alcoholism treatment, but few human laboratory-based studies investigating mechanisms for these effects have been conducted on alcohol dependent persons. The present study was designed to determine the impact of two opiate antagonists on alcohol consumption among nontreatment-seeking alcoholics (n=125) and social drinkers (n=90). Participants were randomly assigned to receive placebo, naltrexone (titrated to 50 mg/day), or nalmefene (titrated to 40 mg/day) for 8 days with an alcohol laboratory session on the final day. Alcohol consumption was monitored in the natural environment during the first 5 medication days, and during a choice consumption paradigm following a standard 'priming' alcohol dose in a bar-laboratory setting. Social drinkers consumed less alcohol than alcoholics during the prelab medication period and the laboratory choice consumption paradigm, and they attained lower blood alcohol levels than alcoholics following the priming drink. Both opiate antagonist medications equally reduced drinking amounts and frequency among alcoholics but not social drinkers, relative to placebo, during natural environment and bar-lab alcohol consumption evaluations. Greater medication side effects, mostly mild in nature, were observed in participants taking nalmefene. These findings demonstrate that both naltrexone and nalmefene can lead to reductions in alcohol consumption among alcoholics who are not attempting to reduce drinking. Similar laboratory paradigms may offer substantial advantages for observing these effects during evaluation of other medications as well.