Aims: The orexigenic peptide ghrelin may enhance the incentive value of food-, drug-and alcoholrelated rewards. Consistent with preclinical findings, human studies indicate a role of ghrelin in alcohol use disorders (AUD). In the present study an a priori hypothesis-driven analysis was conducted to investigate whether a Leu72Met missense polymorphism (rs696217) in the prepro-ghrelin gene (GHRL), is associated with AUD, alcohol consumption and subjective responses to alcohol. Method: Association analysis was performed using the National Institute on Alcohol Abuse and Alcoholism (NIAAA) clinical sample, comprising AUD individuals and controls (N = 1127). Then, a post-hoc analysis using data from a human laboratory study of intravenous alcohol selfadministration (IV-ASA, N = 144) was performed to investigate the association of this SNP with subjective responses following a fixed dose of alcohol (priming phase) and alcohol selfadministration (ad libitum phase). Results: The case-control study revealed a trend association (N = 1127, OR = 0.665, CI = 0.44-1.01, P = 0.056) between AUD diagnosis and Leu72Met. In AUD subjects, the SNP was associated with significantly lower average drinks per day (n = 567, β = −2.49, 95% CI = −4.34 to −0.64, P = 0.008) and significantly fewer heavy drinking days (n = 567, β = −12.00, 95% CI = −19.10 to −4.89, P < 0.001). The IV-ASA study further revealed that 72Met carriers had greater subjective responses to alcohol (P < 0.05) when compared to Leu72Leu both at priming and during ad lib selfadministration.