Baclofen has been shown to reduce alcohol consumption in alcohol-dependent individuals, but there is marked heterogeneity in response. An association between GABBR1 rs29220 and alcohol dependence has been demonstrated previously. The present study evaluated whether the response to baclofen is moderated by a single nucleotide polymorphism (rs29220) in the GABAB receptor subunit 1 gene (GABBR1). Double-blind, placebo-controlled study. Australia. Seventy-two alcohol-dependent men and women receiving 12 weeks of 30 mg/day of baclofen, 75 mg baclofen or placebo. Primary outcomes included time to lapse (any drinking) and relapse (> 5 drinks per day in men and > 4 in women). We also examined alcohol consumption at follow-up (drinks per drinking day, number of heavy drinking days and percentage days abstinent). We observed significant medication × genotype interaction effect for time to relapse (P = 0.049) and a near-significant interaction effect for time to lapse (P = 0.055). For the CC genotype group, the relapse hazard ratio for baclofen versus placebo was 0.32 [95% confidence interval (CI) = 0.14-0.75] and for the G- group it was 1.07 (95% CI = 0.43-2.63). There was also a significant medication × genotype interaction for follow-up alcohol consumption (drinks per drinking day, heavy drinking days and days abstinent) (P = 0.02). Covarying for baseline levels of craving, aspartate aminotransferase and abstinence before enrolment reduced the medication × genotype effect for time to lapse and relapse but not for alcohol consumption at follow-up. The GABBR1 rs29220 polymorphism may influence treatment response and possibly predict adverse effects to baclofen in the treatment of alcohol dependence.