(CNN) The science is clear: Drinking too much alcohol is bad for your health, but exactly how low-risk is light drinking?
A study published in the journal PLOS Medicine on Tuesday provides some new insight. It found that those who drink the most have the highest risks of death and cancer.
Yet the study also found that a person's combined risk of dying younger or developing cancer is lowest among light drinkers: those consuming only one to three alcoholic drinks per week. That risk increases with each additional drink consumed per week.
Light drinkers appeared to have a lower combined risk of overall mortality or cancer compared with those who never drink, the study found, but more research is needed to determine why.
The study only showed associations between alcohol and such risks, and not any causal relationships.
"We had expected light drinkers to be at a similar combined risk to never drinkers, so the reduced risk in light drinkers was surprising," said Andrew Kunzmann, a research fellow at Queen's University Belfast in Northern Ireland and lead author of the study.
"The reasons for the reduced risk in light drinkers compared to never drinkers are still open to debate amongst the scientific community. Some have suggested that alcohol may have cardio-protective effects that may reduce risk of cardiovascular disease," Kunzmann said, referring to certain studies on red wine and heart health.
Yet "others have indicated that light drinkers may be at a lower risk of cardiovascular disease for other reasons, as light drinkers tend to be wealthier and more health-conscious in other ways," he said.
Those who never drink also might do so for other health reasons that could put them at an overall higher risk of death compared with light drinkers.
In general, a better understanding of the health risks that come with drinking alcohol can help inform clearer guidelines on how much you should limit your alcohol consumption, the researchers wrote in their study.
The new study included data on 99,654 adults in the United States, who were 55 to 74 years old. The data, collected between 1993 and 2001, came from the US Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial.
The adults completed a diet history questionnaire, which assessed their alcohol consumption. They also were followed up with during an average 8.9 years, and their cancer diagnoses were ascertained through medical records.
After analyzing the data, the researchers found that the average lifetime alcohol intake reported among the adults was 1.78 drinks per week, with men reporting that they drink more -- 4.02 drinks per week -- than women -- 0.80 drinks per week.
In both the men and women, risk of death was lowest among those who consumed less than 0.5 drinks per day, the data showed. How much alcohol was consumed over a lifetime had a J-shaped relationship with overall mortality: Those who never drank had a slightly higher risk, and those who had more than 0.5 drinks per day had a much higher risk.
But the average amount of alcohol someone drank in a lifetime was linearly associated with total risk of cancer, in that risk went up the more someone drank, the data showed.
The study had some limitations, including that the data were limited to adults 55 and older, and the data on alcohol consumption were based on self-reported questionnaire responses, which tends to be the case for epidemiology studies of alcohol.
The study could have been more valuable if it provided specific risk factors for individual cancers, said Dr. Anne McTiernan, a cancer prevention researcher at the Fred Hutchinson Cancer Research Center in Seattle, who was not involved in the research.
For instance, "alcohol increases risks for many cancers including, breast, colorectal, esophagus, liver, head and neck, and stomach," said McTiernan, author of the book "Starved: A Nutrition Doctor's Journey from Empty to Full."
"Equally problematic was that they classified men and women the same in relation to categories of alcohol use. Women metabolize alcohol differently than men, so a 'light' drinker for men may be more of a 'moderate' drinker for women," she said.
"The study strengths included the large sample size, including both women and men, and a lifetime history of alcohol use," she said.
In the United States, the American Cancer Society and the American Institute for Cancer Research both recommend that adults who drink alcohol limit their intake to no more than two drinks per day for men and one drink per day for women.
A "drink" can be counted as 12 ounces of beer, 5 ounces of wine or 1.5 ounces of hard liquor.
In the United Kingdom, the Chief Medical Officers' alcohol guidelines recommend drinking no more than 14 units per week -- 140 milliliters, or less than one drink per day -- on a regular basis.
The study added information on specific risk estimates according to lifetime consumption of alcohol, said Dr. Otis Brawley, chief medical and scientific officer and executive vice president of the American Cancer Society, who was not involved in the research.
"It also reinforces the linear relationship between drinking alcohol and cancer risk: The more you drink, the higher the risk," Brawley said.
"Alcohol is estimated to be the third-largest modifiable risk factor for cancer, responsible for 5.6% of cancers," he said.
"Alcohol is also estimated to be the third-largest contributor to overall cancer deaths in both men and women," he added. "The evidence that alcohol increases the risk of cancer has existed for many years."
A statement from the American Society of Clinical Oncology, published in the Journal of Clinical Oncology in January, endorsed certain public health strategies in an effort to minimize excessive alcohol exposure in a way that could boost cancer prevention efforts.
"The purposes of that statement was twofold. One was to just raise awareness about the link between alcohol and cancer," said Dr. Noelle LoConte, an oncologist and associate professor of medicine at the University of Wisconsin-Madison who was lead author of that statement.
The second was that the statement supported certain policy strategies for the first time, such as "raising taxes, limiting hours of sale, reducing youth exposure to alcohol marketing as a way of reducing high-risk drinking," said LoConte, who was not involved in the new PLOS Medicine study.
As for that research, she said, "I think it reinforces what we already knew, which is moderate and heavy drinking is bad universally for cancer."