BACKGROUND: Naltrexone is a nonselective opioid receptor antagonist used as a treatment for alcohol use disorder. However, only modest clinical effects have been observed, possibly because of limited knowledge about the biological variables affecting the efficacy of naltrexone. We investigated the potential role of the kappa opioid receptor (KOR) in the therapeutic effect of naltrexone. METHODS: A total of 48 non-treatment-seeking heavy drinkers (16 women) who met DSM-IV criteria for alcohol dependence participated in two alcohol drinking paradigms (ADPs) separated by a week of open-label naltrexone (100 mg daily). Craving, assessed with the Alcohol Urge Questionnaire and the Yale Craving Scale, and drinking behavior were recorded in each ADP. Prior to naltrexone initiation, KOR availability was determined in the amygdala, hippocampus, pallidum, striatum, cingulate cortex, and prefrontal cortex using positron emission tomography with [ 11 C]LY2795050. RESULTS: Participants reported lower levels of craving (Yale Craving Scale: 211 6 1, p , .0001; Alcohol Urge Questionnaire: 26 6 0.6, p , .0001) and consumed fewer drinks (23.7 6 4, p , .0001) during the second ADP following naltrexone therapy. The observed reduction in drinking was negatively associated with baseline KOR availability in the striatum (p = .005), pallidum (p = .023), and cingulate cortex (p = .018). Voxelwise analysis identified clusters in the bilateral insula, prefrontal, and cingulate cortex associated with the reduction in drinking (p , .0001). In addition, KOR availability in all evaluated brain regions was associated with craving measured in both ADPs. CONCLUSIONS: The KOR is implicated in drinking and craving following naltrexone therapy in alcohol use disorder.