“Dose matters a lot,” said Dr. Krishna G. Aragam, a preventive cardiologist at Massachusetts General Hospital and an author of the study. “Just realize that, as you go up beyond modest ranges, the risk goes up quite a bit.”
The study, which may help resolve medical disputes over the effects of alcohol on the heart, involved sophisticated analyses of the genes and medical data of nearly 400,000 people who participate in the U.K. Biobank, a British repository that investigators use to study genes and their relation to health. The average age of subjects selected for the alcohol study was 57, and they reported consuming an average of 9.2 drinks a week.
Some researchers have reported that drinking modestly protects the heart because moderate drinkers as a group have less heart disease than those who drink heavily or those who abstain. Dr. Aragam and his colleagues also saw that effect. But the reason, they report, is not that alcohol protects the heart. It is that light to moderate drinkers — those who consume up to 14 drinks a week — tend to have other characteristics that decrease their risk, like smoking less, exercising more and weighing less than those who drink more heavily and those who do not drink.
It’s not known why moderate drinkers tend to be more healthy than nondrinkers, Dr. Aragam said. But the Biobank study did not ask why people drank or abstained. Instead, it attempted to tease apart the effects of alcohol on the heart from the effects of other habits, behaviors and characteristics. To do that, the researchers used a method called Mendelian randomization.
Researchers have found genetic variants that predispose a person to heavier or lighter drinking. Because the variants are distributed randomly in a population, they can serve in a study as the equivalent of randomly assigning people to abstain or to drink at varying levels. Researchers can ask if those with variants that are linked to greater alcohol consumption have more heart disease and high blood pressure than those with variants linked to lower consumption.