BACKGROUND This clinical laboratory study evaluated how the timing of drinking and subjective responses to alcohol are effected by the opioid antagonist naltrexone. METHODS Forty non-treatment seeking alcoholics were randomly assigned to treatment with 50 mg naltrexone or matching placebo for 7 days. On day 7, they were administered an "initial drink" of alcohol (blood alcohol levels of between 20 and 30 mg%) in a bar-like setting. In a random fashion, half of the subjects in each group (naltrexone and placebo) had either immediate or delayed (40 min) access to up to 8 additional mini-drinks over a 2-h period. In the delayed group subjective reactions to alcohol were measured prior to access to more drinks. RESULTS In the immediate access condition, subjects had similar drinking patterns, irrespective of whether they were taking naltrexone or placebo. However, in the delayed condition, naltrexone-treated subjects consumed fewer drinks and had a slower progression of drinking. There was a positive relationship between alcohol-induced stimulation and the number of drinks consumed in the placebo subjects but a negative correlation in the naltrexone subjects. CONCLUSIONS These data suggest that the effectiveness of naltrexone on alcohol consumption may be somewhat dependent on pattern of consumption. Since naltrexone seems to disrupt the connection between alcohol-induced stimulation and further alcohol consumption, there may be a time-critical period between drinks necessary for alcoholics to benefit from its effects. These findings are consistent with clinical trial data that suggest a potential synergistic effect between relapse prevention therapies and opiate antagonist pharmacotherapy.