The association between DRD2/ANKK1 and genetically informed measures of alcohol use and problems

In 1990, Blum and colleagues first reported an association between DRD2 and alcoholism. While there have been subsequent replications of this genetic association, there have also been numerous studies that failed to detect an association between DRD2 and alcohol dependence. We propose that one aspect contributing to this inconsistency is the variation in alcohol phenotype used across studies. Within the population-based Finnish twin sample, FinnTwin16, we previously performed multivariate twin analyses to extract latent genetic factors, which account for the variation across seven measures of alcohol consumption (frequency of drinking, frequencyxquantity, frequency of heavy drinking, frequency of intoxication and maximum drinks in a 24-hour period) and problems (the Rutgers Alcohol Problem IndexRAPI and the Malmo-modified Michigan Alcohol Screen TestMmMAST) in 3065 twins. In the present study, we examined the association between 31 DRD2/ANKK1 single-nucleotide polymorphisms (SNPs) and the genetic factor scores generated by twin analyses in a subset of FinnTwin16 (n=602). We focus on two of the genetic factors: a general alcohol consumption and problems factor score, which represents shared genetic variance across alcohol measures, and a alcohol problems genetic factor score, which loads onto the two indices of problematic drinking (MAST and RAPI). After correction for multiple testing across SNPs and phenotypes, of the 31 SNPs genotyped across DRD2/ANKK1, one SNP (rs10891549) showed significant association with the general alcohol consumption and problems factor score (P=0.004), and four SNPs (rs10891549, rs1554929, rs6275, rs6279), representing two independent signals after accounting for linkage disequilibrium, showed significant association with the alcohol problems genetic factor score (P=0.005, P=0.005, P=0.003, P=0.003). In this study, we provide additional positive evidence for the association between DRD2/ANKK1 and alcohol outcomes, including frequency of drinking and drinking problems. Additionally, post hoc analyses indicate stronger association signals using genetic factor scores than individual measures, which suggest that accounting for the genetic architecture of the alcohol measures reduces genetic heterogeneity in alcohol dependence outcomes in this sample and enhances the ability to detect association.