For years, we’ve been reading about studies suggesting that drinking red wine in moderation is good for us. Now comes a less agreeable verdict, written by Timothy Naimi, a School of Medicine and School of Public Health associate professor, and others, and published in the American Journal of Public Health. The authors attribute 6,000 American deaths annually to cancer from moderate drinking, which they define as a drink and a half a day or less. Add in alcohol consumption at all levels and the total surges to 20,000 cancer deaths a year, or 3.5 percent of all cancer deaths in the country.

The researchers reached their conclusion by analyzing risk estimates from hundreds of other studies. Studies linking alcohol to cancer are based on calculations using three types of data: the numbers of people who drink at different levels, the prevalence of various cancers at those various drinking levels, and the number of cancer deaths among people at each level. The American Cancer Society lays out the cancers that have the strongest evidence of an alcohol link, while adding that the precise mechanism for how drinking leads to the disease is not certain. For men, lethal alcohol-caused cancer typically afflicts the mouth, throat, and esophagus, the researchers say. In women, breast cancer is the most common cancer killer linked to alcohol consumption.

Evidence of excessive drinking’s role in cancer is much greater than that for the role of modest drinking, says Naimi, an alcohol epidemiologist specializing in binge and youth drinking and alcohol policy. The idea that limited drinking causes cancer “should be interpreted with caution,” he says. “I have nothing against alcohol. My background is as a physician, and my interest is in seeing harm from alcohol minimized.”

But some doctors say the findings about moderate drinking need to be taken seriously, and Naimi argues that deaths from alcohol “dwarf any small number of people who may derive benefit from low-dose alcohol.” Among all people who start drinking, he says, 5 to 10 times as many die from it as are benefited by it.

Nor is he convinced by studies showing heart benefits from moderate drinking. For one thing, he says, those studies have never included the accepted standard in scientific research: a randomized, controlled study comparing moderate drinkers with teetotalers. Also, moderate drinkers tend to come from higher on the socioeconomic ladder, a rung where people tend to be healthier. In other words, moderate drinking may be “a reflection of people’s social position and good health.”