News — Moderate alcohol use does not reduce cardiometabolic disease risk among veterans of European, African, or Hispanic ancestry, a new study suggests. The findings add to growing evidence that traditional research methods applied to drinking levels and certain disease outcomes have created illusory and misleading results. Heavy drinking is known to be linked to coronary heart disease (CHD) and type 2 diabetes (T2D). Traditional observational studies have, however, associated moderate drinking with the lowest risk and abstinence with a moderate risk (the U-curve or J-curve effect). In recent years, the U-curve has been increasingly attributed to confounding errors—when study results are distorted by other factors. In this case, the abstinence category is implicated since it establishes a false equivalence between study participants with widely differing risk factors (lifelong non-drinkers, those who stopped drinking for health or other alcohol-related problems, and those who falsely reported not drinking). Scientists have become able to control for these and other confounders using Mendelian randomization (MR), which draws on genetic data to explore disease risk. Most MR studies—largely limited to people of Asian and European origin—have found no link between moderate alcohol use and CHD or T2D, though some suggest a raised risk. For the study in Alcohol: Clinical & Experimental Research, investigators applied old and new analytical methods to a large, diverse data pool of US veterans.
Researchers worked with participants from the Million Veteran Program (MVP), which incorporates genetic information, alcohol exposure data, and electronic health records. The researchers identified more than 33,000 patients with CHD diagnoses, whom they matched to 165,000 controls without CHD, with a median age of 61 and 97% male. Seventy-four percent were European American, 18% African American, and 6% Hispanic American. There were more than 28,000 patients with T2D, matched to 141,000 controls without T2D, with a median age of 59, 93% male, and 65% European American, 25% African American, and 8% Hispanic American. They performed both observational and MR analyses exploring the relationship between alcohol consumption, CHD, and T2D. The models accounted for demographic factors, blood pressure, smoking, and other factors. For ancestry groups of sufficient size, genetic risk scores linked to alcohol use, blood pressure, and smoking were incorporated.
The observational analysis yielded the familiar U-shaped curve linking alcohol use with CHD and T2D among European and African Americans—but not among Hispanic Americans, a much smaller sample. The MR analysis of the full sample found no evidence of drinking level as a causal risk factor for CHD or T2D. This was the case when working with variants of ADH1B, a gene with well-established links to alcohol use, and when using genetic risk scores based on other relevant variants (for European Americans). Adjusting for the effects of blood pressure and smoking similarly showed no link between alcohol use and cardiometabolic disease risk.
This study provides persuasive additional evidence that moderate drinking is not protective against cardiometabolic disease and that the U-shaped curve is a product of confounding errors. It provides the most extensive exploration to date of the relationship between drinking level and CHD and T2D in African American and Hispanic American populations. Additional MR analysis is needed involving larger samples of non-European populations.
A Mendelian randomization study of alcohol use and cardiometabolic disease risk in a multi-ancestry population from the VA Million Veteran Program. R. Kember, C. Rentsch, J. Lynch, M. Vujkovic, B. Voight, A. Justice, T. L. Assimes, H. Kranzler.
ACER-24-6063.R1
RSS