Did we need another “drinking isn’t bad for you” study? Apparently so. But the National Health Service (NHS) in the United Kingdom said don’t pour that second glass of wine just yet.

The study involves health records of more than 70,000 Danish people, and found that those who drink small amounts of alcohol throughout the week—rather than a lot at one sitting—have the lowest risk of diabetes. In fact, the moderate drinkers had less risk of developing diabetes than people who didn’t drink at all, according to results appearing in the journal Diabetologia.

The study looked data from the Danish Health Examination Survey from 2007 to 2008. The group included 28,704 men and 41,847 women who were followed for 4.9 years. Researchers wanted to go beyond the overall amount people drank in a week and examine when people were drinking and how much at each sitting. Then, they matched this data with who developed diabetes from the Danish National Diabetes Register.

At follow-up, 859 men and 887 women were diagnosed with diabetes; men who drank 14 drinks per week had the lowest risk, a 43% reduction, while women who had 9 drinks per week had a 58% reduction compared with those who did not drink at all. Moderate drinkers who consumed alcohol 3 to 4 days a week also fared better than those who drank less than a day a week.

The study did find that the moderate drinkers who drank a glass or 2 on separate days throughout the week fared better those who drank heavily on a few days, but abstained on other days. Drinking wine was especially beneficial, and beer seemed to be fine for men. Not enough women in the study drank beer to evaluate its benefits.

Many studies have evaluated the effects of drinking alcohol on diabetes and cardiovascular disease. It is well-established that drinking too much is harmful, but studies have differed on whether some amount of alcohol is beneficial or at least not harmful. It is important to note that this Danish study is based on self-reports, and people may underreport their alcohol use. And it’s not known if the results transfer to populations outside Denmark that may be at greater risk for diabetes.

But besides the problem of using self-reports, the NHS pointed out that the Danish study relies on data from a single point in time. While there is an association between moderate drinking and less diabetes, it doesn’t prove having a drink has a protective quality.

“It is not proof that starting to drink more, especially for those who do not currently drink, is useful in preventing diabetes,” the health service warned.

Reference

Holst C, Becker U, Jamenson ME, Gronbaek, M, Tolstup JS. Alcohol drinking patterns and risk of diabetes: a cohort study of 70,551 men and women from the general Danish population [published online July 27, 2017]. Diabetologia. https://doi.org/10.1007/s00125-017-4359-3.