(The New York Times News Service) -- Even moderate alcohol use may substantially raise the risk of dying from cancer, according to a study released Thursday offering the first comprehensive update of alcohol-related cancers deaths in decades.

"People don't talk about the issue of alcohol and cancer risk," said Dr. David Nelson, director of the Cancer Prevention Fellowship Program at the National Cancer Institute and lead author of the study.

"Alcohol has been known to be related to causing cancer for a long period of time. We talk about cancer prevention, screenings and tests. This is one of those things that seems to be missing in plain sight."

Alcohol use accounts for about 3.5 percent of all U.S. cancer deaths annually, according to the study. The majority of deaths seemed to occur among people who consumed more than three alcoholic drinks a day, but those who consumed 1.5 beverages daily may account for up to a third of those deaths, the researchers found.

In 2009, 18,000 to 21,000 people in the United States died of alcohol-related cancers, from cancer of the liver to breast cancer and other types, the researchers said. That's more that the number of people who die every year of melanoma or ovarian cancer, which caused over 9,000 and 14,000 U.S. deaths respectively in 2009.

How alcohol causes cancer is not fully understood, the study notes. Previous research has shown alcohol appears to work in different ways to increase cancer risk, such as affecting estrogen levels in women and acting as a solvent to help tobacco chemicals get into the digestive tract.

The study, published in the American Journal of Public Health, is the first major analysis of alcohol-attributable cancer deaths in more than 30 years. Researchers said the lack of recent research on the subject may contribute to a lack of public awareness of the cancer risks.

"People are well aware of other risks, like the impact of tobacco on cancer, and are not as aware alcohol plays quite a bit of a role," said Thomas Greenfield, one of the study's authors and scientific director of Public Health Institute's Alcohol Research Group, in Emeryville.

Researchers examined seven types of cancers known to be linked to alcohol use: cancers of the mouth and pharynx, larynx, esophagus, liver, colon, rectum and female breast. To link the cancer to alcohol use, they relied on surveys of more than 220,000 adults, 2009 U.S. mortality data, and sales data on alcohol consumption.

Breast cancer accounted for the most common alcohol-related cancer deaths among women -- alcohol contributed to 15 percent of all breast cancer deaths. Among men, cancers of the mouth, pharynx, larynx and esophagus accounted for the most alcohol-linked cancer deaths.

The study drew some criticism.

Dr. Curtis Ellison, professor of medicine and public health at Boston University School of Medicine, said the study failed to take into account several important factors, such as the pattern of drinking rather than just the amount of alcohol consumed. He said consuming small, consistent amounts of alcohol is much healthier than occasional binge drinking.

"They're mixing alcohol abuse, which leads to all of these cancers as they've clearly shown, with the casual drinker, where the risk is very small," said Ellison, also co-director of the International Scientific Forum on Alcohol Research.

Dr. Arthur Klatsky, adjunct investigator at the Kaiser Permanente Northern California Division of Research, said many studies have shown moderate drinkers in older age groups may be healthier than those who abstain. Klatsky has long researched the effects of alcohol on health, but declined to comment directly on the study because he had not yet read it.

"Advice needs to be individualized," Klatsky said. "The advice one would give to 60-year-old man who has no problem with alcohol but is at high risk of heart disease due to family history is quite different than the advice we give to a 25-year-old woman whose mother died of breast cancer."

The study's authors acknowledged alcohol can have health benefits, but claimed alcohol causes 10 times as many deaths as it prevents. There's no known safe level of drinking, they said.

"The safest level for cancer prevention is that people don't expose themselves to any potential risk," said Nelson of the National Cancer Institute. "The bottom line means for people who choose to drink, their cancer risk will be lower if they drink lower amounts."

Copyright 2013 The New York Times News Service. All rights reserved.