BACKGROUND

Alcohol consumption behaviors and alcohol use disorder risk and presentation differ by sex, and are associated with blood concentrations of the steroid sex hormones, testosterone and estradiol, and their regulatory binding proteins, sex hormone binding globulin (SHBG) and albumin. Genetic variation is also associated with alcohol consumption, alcohol use disorder, and levels of these hormones and binding proteins.

METHODS

To assess the contribution of genetic factors to previously described phenotypic associations between alcohol-use traits and sex-hormone levels, we estimated genetic correlations (r) using summary statistics from prior published, large sample size genome-wide association studies (GWAS) of alcohol consumption, alcohol dependence, testosterone, estradiol, SHBG, and albumin. We defined statistical significance at p < 0.005 and trends at p < 0.05.

RESULTS

For alcohol consumption, we observed positive genetic correlation (i.e. genetic effects in the same direction) with SHBG in females (r = 0.089, p = 0.004) and a trend toward negative genetic correlation (i.e. genetic effects in opposite directions) with bioavailable testosterone (r = -0.064, p = 0.032); however there were only trends toward positive genetic correlation with total testosterone in males (r = 0.084, p = 0.007) and with albumin in a sex-combined cohort (r = 0.082, p = 0.015). For alcohol dependence, we observed trends toward negative genetic correlation with total testosterone in females (r = -0.106, p = 0.024) and positive genetic correlation with BMI-adjusted SHBG in males (r = 0.119, p = 0.017). Some of these genetic correlations were different than the corresponding phenotypic associations, and some may suggest differences between females and males.

CONCLUSIONS

Shared genetic effects might contribute to positive associations of alcohol consumption with albumin and between alcohol dependence and SHBG in males; however, most of the phenotypic associations between alcohol-use traits and levels of sex hormones and their binding proteins did not correspond to broadly shared genetic effects in the same direction. Some even corresponded to genetic effects in the opposite direction. Future studies of these traits should include GWAS on larger cohorts by sex and investigation of localized correlations of genetic effects and the relative contributions of heritable and environmental factors.