ADH1B-rs1229984 was significantly associated with all alcohol phenotypes: the high risk group (absence of the protective allele) had higher prevalence of AD, any AUD, and risk drinking, and higher mean values of AD and AUD severity and maxdrinks (Table 3). For example, the high risk group had 2.28 times greater odds of AD, and 1.83 times greater odds of risk drinking (Table 3). All results remained significant (p<0.02) after adjusting for population structure (Table 3) and when controlling for the other SNPs (for rs1229982, p<0.007; for rs1159918, p<0.001; for rs698, p<0.002). The other ADH1B SNPs gave less consistent results. ADH1B-rs1229982 was significantly associated with a subset of phenotypes, including AD, AD and AUD severity, with weaker associations than those for rs1229984 (Table S1). Furthermore, ADH1B-rs1229982 was no longer associated with these phenotypes when controlling for ADH1B-rs1229984. Although the ADH1B-rs1159918 high risk group showed higher prevalence for any AUD, this was not significant after adjusting for demographics in regression analysis (Table S2).
