Genomic Markers of Heavy Alcohol Consumption Outperform Self-Report and Other Common Predictors of Mortality in Two...

Heavy alcohol consumption (HAC) may contribute to elevated mortality risk in middle age, but its contribution to mortality in middle age among Black Americans is poorly characterized, suggesting the potential for genomic variables to improve prediction. We examined whether a DNA methylation-based measure of HAC (HAC/ATS) increased prediction of mortality beyond Chronic Illnesses and other controls and compared its predictive abilities to that of self-reported HAC.

Two longitudinal samples were examined to test hypotheses. First, we examined the Family and Community Health Study (FACHS) sample, comprising 360 individuals who were 95.8% female, with a mean age of 47.12 at the time of their blood draw, and followed for 17 years. Second, we examined the Promoting Strong African American Families (ProSAAF) sample comprising 336 individuals who were 54.8% female, with a mean age 44.64 at the time of their initial blood draw, and followed for 5 or more years. To overcome potential biases arising from small sample sizes we used Firth's penalized likelihood logistic regression analyses to predict mortality in both samples.

In both samples, the DNA-methylation-based indicator of HAC (HAC/ATS) significantly contributed to the prediction of mortality beyond the significant contribution of age, Number of Chronic Illnesses, and other control variables. This was in contrast to the non-significant contribution of self-reported HAC. In both samples, high blood pressure was also a significant predictor of mortality, albeit not as powerful a predictor as DNA-methylation-based indicator of HAC.

A DNA-methylation-based indicator of HAC (HAC/ATS) made a substantial contribution to mortality risk in two longitudinal samples of middle-aged African Americans. Significant contributions of Chronic Illness suggested the value of future exploration of genomic indicators of HAC in combination with chronic conditions.