Doctors have long known that heavy alcohol consumption increases the risk of stroke, but new research suggests that it's even more of a factor in middle-age than previously thought. A study published in the journal Stroke this week says that consuming more than two alcoholic drinks a day on average increases the stroke risk for individuals in their 50s and 60s more than traditional factors like high blood pressure and diabetes.

The study examined more than 11,000 twins from the Swedish Twin Registry, sorting them into the categories of non-drinkers, light drinkers, moderate drinkers or heavy drinkers based on a questionnaire issued in 1967. Non-drinkers were excluded from the study as they might have been alcoholics who later gave up. Light drinkers were classified as those that had less than half a drink a day on average while heavy drinkers had more than two drinks a day. Moderate drinkers fell in between these two poles.

The researchers reported that after controlling for factors including diabetes, high blood pressure, smoking, exercise, and weight, individuals that were heavy drinkers in their middle age were 34 percent more likely to suffer a stroke than their light-drinking peers. However, for individuals between the ages of 75 and 80, blood pressure was the greatest factor, and for those older than 80, diabetes mattered most.

"The main message is that for people in middle age it's harmful to drink more than two drinks of alcohol a day," says Pavla Kadlecová, a statistician who worked on the study. Kadlecová notes that as the study looked at twins it also controlled for genetic and lifestyle factors, although it did not take into account how twins' lifestyles might diverge after a young age.

"They're making the important point that in relative terms alcohol is a more important risk factor for people in middle age than high blood pressure or diabetes," says Jonathan Mant, a professor at the University of Cambridge's Primary Care Unit and an expert in stroke prevention. "And in relative terms it seems to match smoking very closely."

However, as the measure of alcohol consumption in the study only came from a single questionnaire, the researchers might have even underestimated the risk says Mant. "If the truth is that there is genuinely a link between alcohol consumption and strokes and you have a very good way of measuring alcohol consumption then you will detect that link. If you have a very bad measure of alcohol consumption then you will underestimate the association."