Men who consume more alcohol on drinking days are at increased risk of death even if their average amount of alcohol over time isn't excessive, according to researchers who said that drinking quantity and frequency influence mortality more than average consumption over time.

Researchers Rosalind A. Breslow, Ph.D., of the National Institute on Alcohol Abuse and Alcoholism's (NIAAA) Division of Epidemiology and Prevention Research, and Barry I. Graubard, Ph.D., of the Division of Cancer Epidemiology and Genetics at the National Cancer Institute reported that their findings "reinforce the importance of drinking in moderation. In drinkers who are not alcohol dependent, the majority of U.S. drinkers, alcohol quantity and frequency might be thought of as modifiable risk factors for mortality."

The researchers studied 44,000 people who had participated in a national health survey in 1988, about half of whom were current drinkers who consumed at least 12 alcoholic drinks within the previous year. By 2002, 2,500 of these individuals had died.

Breslow and Graubard found that men who consumed five or more drinks on the days they drank were 30 percent more likely to have died than those who consumed only one drink per drinking day, and those who drank more heavily increased their risk of cancer.

Drinking frequency was associated with a lower risk of cardiovascular disease among men, but raised the risk of cancer among women, the researchers found. As with men, women who drank more alcohol at a sitting were more likely to have died.

Breslow said the study was an improvement over past research on alcohol consumption and mortality risk that focused on average alcohol intake. "Average intake makes no distinction between the individual who has seven drinks one day each week, for example, and someone who has just one drink, every day," he said. "Our study is the first to look at how both quantity and frequency components of alcohol consumption independently influence cause-specific mortality within a single cohort representing the US population."

The study appears in the March 2008 issue of the journal Alcoholism: Clinical and Experimental Research.