BACKGROUND: There is strong evidence for the role of the endogenous opioid system in alcohol reinforcement and consumption; however, recent human laboratory studies and clinical trials have reported mixed effects of naltrexone (a nonselective opioid antagonist) on alcohol-related behaviors. This paper reports a secondary data analysis of a human laboratory study that examines the relationship between serum levels of 6-beta-naltrexol, the major, biologically active metabolite of naltrexone, and subjective effects of alcohol. METHODS: The study used a within-subjects design to examine the effects of naltrexone (0, 50, and 100 mg/day) on subjective responses to alcohol (none, moderate, and high dose) in heavy drinkers (n = 23). Each subject received three doses of naltrexone in random order; each naltrexone dose was administered over an 8 day period on an inpatient unit, with a 1 week outpatient washout between doses. After stabilization at each of the naltrexone doses, subjects participated in three alcohol challenge sessions (none, moderate, and high dose) in random order; thus, each subject participated in a total of nine alcohol administration sessions. RESULTS: Doubling the naltrexone dose (50 vs. 100 mg/day) doubled the mean serum 6-beta-naltrexol levels. At each naltrexone dose, there was a 4-fold range in 6-beta-naltrexol levels across subjects. Before alcohol administration, higher 6-beta-naltrexol levels were associated with higher ratings of sedation. After high-dose alcohol administration, higher 6-beta-naltrexol levels were associated with significantly lower ratings of liking and best effects. CONCLUSIONS: These findings provide further evidence of the involvement of the opioid system in the modulation of alcohol effects and suggest that serum 6-beta-naltrexol concentrations may be important in predicting therapeutic response to naltrexone.