Research on alcohol consumption is in a pickle. There's no question that pounding one drink after another is bad for your health. Things get murkier when it comes to "moderate" drinking. What does that mean? What's the limit? Can a health-conscious person serenely order a second round?
The alcohol industry has long embraced the notion that alcohol in moderation not only won't harm you but is actually good for you. The hypothesis gained traction in the early 1990s when 60 Minutes reported on what is called the French Paradox. The French have low rates of heart disease despite all the butter, cream, foie gras, etc. in the Gallic diet. Some researchers suggested that compounds in red wine, also favoured by the French, explain the paradox.
Many studies have shown that people who drink any kind of alcohol in moderation — wine, beer, spirits — have lower rates of heart disease than people who abstain or who drink heavily. But the evidence is stubbornly ambiguous.
As reported in the Lancet earlier this year, a survey of the health of nearly 600,000 drinkers in 19 countries found that very moderate drinking — about one drink a day — lowers the rate of certain kinds of heart attacks but raises the risk of other cardiovascular problems. There's no net benefit in life expectancy, the study found.
Alcohol research is notoriously bedeviled by what are called "confounding effects". The most obvious is that the non-drinking population includes people who can't drink because of health problems. Meanwhile, healthy people feel free to drink. This can create a misleading impression of cause and effect.
"People who drink moderately are healthier than people who don't drink. But that doesn't mean the drinking caused them to be healthier," says University of Minnesota social epidemiologist Toben Nelson.
This issue was supposed to be clarified by the 10-year, US$100 million Moderate Alcohol and Cardiovascular Health (MACH) trial, which started to enroll participants earlier this year. The USA's National Institute on Alcohol Abuse and Alcoholism (NIAAA), part of the National Institutes of Health, sponsored the trial.
It would have looked at 7800 people on multiple continents, all older than 50 and at risk of heart disease. Teetotalers and heavy drinkers would have been ineligible to participate. The volunteers would have been randomly assigned into two groups, one that consumed a single alcoholic drink every day for six years and one that abstained from alcohol every day for six years. The goal was to see whether drinking a little bit lowered the rate of heart disease.
But NIH terminated the project on June 15, saying the trial's results would not be viewed as credible. The majority of the funding for the study was to have come from the alcohol industry, funnelled through a non-profit foundation. This group, the Foundation for the National Institutes of Health, is supposed to be a firewall between industry and researchers, but a March story in the New York Times and a subsequent internal NIH investigation found that researchers had engaged in extensive communication with industry representatives before the government's approval of the trial.
The NIH investigation also found flaws with the study design. It didn't have enough participants, and it didn't last long enough, said Michael Lauer, a cardiologist and deputy director for extramural research at NIH. Cancer, for example, can take many years to become detectable.
The study also did not include women deemed at high risk of getting breast cancer. That's because research has shown that any alcohol consumption, even a drink a day, increases breast cancer risk. But the decision not to include such women also could have skewed the results of the study.
The NIH report said the trial, as designed, "could show benefits while missing the harms".
Emails published as part of the NIH report suggest that backers of the trial expected to show that moderate drinking has a health benefit. Researchers are supposed to have what is known as "equipoise" going into a trial. That means "you are approaching a question with a completely neutral attitude", Lauer said.
"It's disappointing that these events occurred. It's disappointing because this is not the high level of stewardship, open competition, that we at NIH are very proud of," Lauer said.
Marion Nestle, a professor emerita of food science at New York University, said, "This is not open-ended research. This is marketing research."
The trial's principal investigator, Kenneth Mukamal, an epidemiologist with Harvard's Beth Israel Deaconess Medical Center, did not respond to an interview request. But he released a statement after NIH's decision, saying he and his team "stand fully and forcefully behind the scientific integrity" of their trial.
Alcohol plays such a prominent role in American culture that it's easy to forget that 98 years ago the prohibition of the sale of alcohol was amended into the Constitution. And teetotaling even today is common: One in three American adults does not drink alcohol. One in two has not had a drink in the last 30 days.
At the other extreme are the millions of Americans who are routinely inebriated. About one in six adults binge-drink four times a month, with an average of seven drinks per binge, according to the Centers for Disease Control and Prevention.
Alcohol can be a killer. The CDC calculates that 88,000 Americans die annually from alcohol-related causes, which include automobile accidents, breast cancer and cirrhosis of the liver.
Alcohol is a chemical compound that has multiple and complex effects on many parts of the body, so it's impossible to reduce its effects, positive or negative, to a simple metric. A generic description of alcohol is that it is a central nervous system depressant. It impairs perception and slows reaction time. Those effects can increase the risk of accidents or make the intoxicated person more vulnerable to violence and abuse.
"At extreme levels of consumption, alcohol can shut down even the most basic bodily functions, like respiration," says Nelson, the Minnesota researcher.
The body breaks down alcohol into a chemical, acetaldehyde, that can damage cellular DNA and potentially cause cancer. Alcohol is associated with cancers of the mouth and throat, larynx, esophagus, colon and rectum, liver and breast, according to the CDC.
Alcohol consumption is included in the US dietary guidelines, which say that if you don't already drink, you shouldn't start. But they also say that alcohol can be consumed safely in moderation. That's defined as no more than two drinks a day for a man and one a day for a woman (zero if pregnant, because alcohol disrupts fetal development).
Stop the presses: Two-thirds of drinkers fail to drink moderately, says Robert Brewer, head of CDC's alcohol programme.
"We think of the dietary guidelines on alcohol as aspirational," Brewer said. "Current compliance with the dietary guidelines is poor."
Different countries have dramatically different guidelines for moderate drinking. One prominent variant is in Britain, which does not issue distinct standards for men and women. The study published in the Lancet backed this unisex approach: It showed that life expectancy begins to drop for both men and women if they drink more than about 100 grams of alcohol a week. That's equal to about seven 12-ounce beers or five-ounce glasses of wine.
The US guidelines take into account the fact that men are larger than women, on average, and have more water as a fraction of their body mass. The average weight of an American adult man is 196 pounds (88.9kg), compared with 169 pounds (76.7kg) for the average American woman. Aaron White, chief scientist at NIAAA, said men are, on average, 60 percent water, and women are 50 percent water.
"You're pouring your drink into more water in the male," White said. "You've got more fluid to dilute the alcohol in."
By White's calculation, men should be able to drink about 1.5 times as much as women, on average, to reach the same blood alcohol content. But not twice as much.
"People clearly didn't come up with these moderate drinking guidelines based on math," White said.
The dietary guidelines do not address alcoholism. That's a separate, parallel universe of definitions and standards. Alcoholism has evolved as a medical concept, and the DSM-5 — the Diagnostic and Statistical Manual of Mental Disorders — has adopted the term alcohol use disorder, or AUD. That disorder is divided into mild, moderate and severe levels, depending on how many of 11 criteria are met by the drinker.
Among the scientists who argue that moderate drinking provides a health benefit is epidemiologist R. Curtis Ellison of the Boston University School of Medicine. He says drinkers should drink a little bit every day, without any days off, because alcohol makes blood platelets less sticky and keeps other clotting factors low.
In the opposite camp is Michael Holmes, a University of Oxford epidemiologist who led a 2014 study that found no benefit. Alcohol can cause a rise in blood pressure, coronary-artery calcification and higher body fat, all factors in heart disease, he said. "The burden of evidence is toward alcohol having a detrimental effect on heart disease, even in small quantities," he said.
The collapse of the MACH trial left researchers where they've been for years, Holmes said: "We're at an impasse."
Where does the field go now?
Lauer said that when patients ask him if moderate drinking is good for them, he can only answer, "I don't know."
He said it would be hard to set up a perfect clinical trial because alcohol has so many different effects. But he said it's clear that heavy drinking leads to all kinds of negative outcomes, including alcoholism, stomach ulcers, bleeding in the intestinal tract and liver disease.
Lauer continued: "Lots of different kinds of cancer, including cancer of the oral pharynx and the liver, motor vehicle accidents, poor judgments, ruined marriages."
The CDC's Brewer says it's time to think of new approaches to curbing excessive drinking. Among the possibilities: increasing alcohol taxes, limiting the density of alcohol retailers, limiting the hours alcohol can be sold and creating a public awareness campaign on the dangers of binge drinking.
Given the known dangers and catastrophic consequences of binge drinking, the questions surrounding moderate drinking are arguably esoteric, or at least not an urgent public health issue. Nelson, the Minnesota researcher, said he doesn't think the drink-a-day riddle is "a worthwhile scientific question to pursue".
The end-point of such research might show a benefit, he said, but society doesn't have a compelling need to get people who don't drink to start drinking. It's more important, he said, to persuade heavy drinkers to moderate.
"People like to drink, like to get buzzed, and people like to get drunk, too. But there are major risks associated with heavy drinking," he said. His informed message to the public, delivered without any hint of moralising, is a simple one: "Drinking is not good for you."