Correction to: Neuropsychopharmacology advance online publication, 17 May 2017; doi:10.1038/npp.2017.74

Following publication, the authors reported that the legends for Figures 3 and 4 were inadvertently switched. This error does not change the content, statistics, or conclusions of the report.

The correct legend for Figure 3 is: Percent heavy drinking days during the 16-week treatment period and the 24-week post-treatment follow-up by OPRM1 A118G genotype and medication group. Regression lines were fit separately for each period. The genotype by medication by time interaction was significant, such that G-allele carriers who received naltrexone had an increased rate of heavy drinking once medication was stopped, whereas the other groups did not. Figures are estimated marginal means±SE’s.

The correct legend for Figure 4 is: Percent heavy drinking days by smoking status at study entry and medication group. There was a significant interaction between smoking, medication, and time, such that %HDD increased over time among smokers, but not non-smokers, and naltrexone, relative to placebo, ablated this increase among smokers. Figures are estimated marginal means±SE’s. N’s indicate the number of subjects who provided drinking data at each time point. *p<0.05 for the simple effect of medication among smokers during each month.