MARGARET WARNER: For years now, researchers have tried to pin down whether there are health benefits from drinking, particularly red wine. Now a study out earlier this month in the New England Journal of Medicine has stirred new controversy. Researchers who studied 38,000 men for 12 years found that regular, moderate drinking of beer, wine, or hard liquor correlated with a significant reduction in the risk of heart attack. Alcohol consumption appeared to reduce the risk of heart attack by 12 percent for those who drank one to two days a week; 32 percent for those who drank three to four days a week; and 37 percent for those who drank five to seven days a week.
Here to discuss the study's findings and implications are Dr. Kenneth Mukamal, the study's lead author. He's an internist at Beth Israel Deaconess Medical Center in Boston, and assistant professor of medicine at Harvard Medical School; and Dr. Ira Goldberg, medical professor at Columbia University, and chief of its division of preventive medicine. Welcome to you both. Dr. Mukamal, before we really get into this, define your terms a little more for us. When you say your subjects were moderate drinkers, how much did they drink on the days they drank?
DR. KENNETH MUKAMAL: There was a wide variety of how much men were drinking. We typically defined it as about a glass of, say, four to five ounce glass of wine, red or white, about 12 ounces of beer and about a one to one-and-a-half shot of spirits.
MARGARET WARNER: Did you define it as one drink a day?
DR. KENNETH MUKAMAL: We looked at really a range. Some men were drinking less than a full drink on those days, some one to two and some even more than two, although there were a very heavy drinkers in the study.
MARGARET WARNER: What's really new in your study is, one, it didn't matter what they drank. Is that right? Wine, beer…
DR. KENNETH MUKAMAL: That's exactly correct. It didn't matter. In fact, these men tend to consume beer and lick inner the greatest quantities as is true for other American men. And those were the beverages that were most strongly associated with lower risk. Red wine and white wine were least associated with lower risk.
MARGARET WARNER: Were red and white wine the same? There's been an ongoing debate about that.
DR. KENNETH MUKAMAL: They tend to be similar. Once you got to a level of consumption that was about a drink a day or more, all of the beverages pretty much looked alike.
MARGARET WARNER: But what was important was how much days a week they drank. In other words, the more days a week they drank, the lower the risk of heart attack?
DR. KENNETH MUKAMAL: That's correct. In our analyses, frequency really trumped everything including quantity. What was really dominant was how many days per week men said they were drinking. It mattered considerably less how much they were drinking on those days -- again what type of beverage they were drinking or whether they were consuming their alcohol with meals or not.
MARGARET WARNER: Did your study also look at how the alcohol works on the body? Why frequent drinking makes a body less susceptible to a heart attack?
DR. KENNETH MUKAMAL: We didn't look at the issues of mechanism directly in our study but there's been a wide body of evidence looking at that and previous work including an experimental trials where people are directly given alcohol. In most of that work, about half of the benefit of alcohol is generally attributed to HDL, the so-called "good cholesterol." We know that alcohol raises good cholesterol pretty consistently. Our findings about frequency of use also point perhaps toward an effective alcohol as a blood thinner whereby making the platelets less sticky in the blood we might prevent them from clogging arteries and therefore prevent heart attacks.
MARGARET WARNER: Dr. Goldberg, your thoughts on the study first of all whether you consider it medically scientifically valid as demonstrating the benefits of moderate drinking.
DR. IRA GOLDBERG: Well, of course, the study is valid in a sense that it's epidemiology looking at a large number of people and finding an association. The question is whether it's valid for therapy or how we should be approaching our patients. Doctors really use three kinds of information to decide on treating patients. Epidemiology, this kind of association, is really often the first kind of information we have. The second kind of information often is experimental or mechanistic kind of information, trying to say, is this reasonable? But most doctors really to treat patients want to have definitive proof where two groups of patients were equivalent and part of them were given one type of therapy and part of them were given another. That's the kind of information we need, for example, to use a new drug on patients to protect them from heart disease or any other disease. This is not that kind of data yet.
MARGARET WARNER: So you're saying in other words just because there's a correlation that doesn't prove a cause?
DR. IRA GOLDBERG: Absolutely.
MARGARET WARNER: So are there other things about people who drink that might explain quite irrespective of this, why they were... why they had lower rate of heart attack? In other words, why wouldn't correlation be a cause?
DR. IRA GOLDBERG: Why would some people drink and some people not drink? There must be some differences from them. The investigators of this study have done whatever they could to try to figure out what the differences were, but we've learned in the last decade on two other major types of studies-- one looking at vitamin E intake and several studies looking at women taking hormone replacement-- that there are differences between people that sometimes can't be ferreted out when you collect data like this on a large population. So there's always a bit of uncertainty.
MARGARET WARNER: Dr. Mukamal, address that point: That you established correlation but not a cause.
DR. KENNETH MUKAMAL: I think Dr. Goldberg is exactly correct. What we're looking at are associations indeed although obviously there are other sources of information regarding alcohol intake and the mechanistic factors that he alluded to. We certainly do not have the level of randomized evidence that we ideally would like to have; whether we're going to have that any time soon is unclear, I think. Along the same lines it's certainly true that there have been well publicized examples of situations where observational evidence and randomized evidence did not agree.
MARGARET WARNER: Like the hormone therapy, for instance?
DR. KENNETH MUKAMAL: An ideal example is hormone replacement. Indeed vitamin E is another good one. There are certainly many others where they have come together. There are a variety of other dietary factors, for example fish intake and omega 3 fatty acids and the risk of fatal heart disease where randomized evidence and observational evidence seem to agree. So although it's certainly true we don't have that kind of level of evidence yet for alcohol, I'm hoping some day we will. In the meantime I think it's difficult to simply ignore the evidence that we have in front of us.
MARGARET WARNER: And Dr. Mukamal quickly, another criticism of your study or at least maybe the way it was reported was that there are no women in your study. Would you like to comment on that and does this... do you believe this study has any validity for women?
DR. KENNETH MUKAMAL: I think it's very important not to generalize our results to women indiscriminately. It is certainly true that there have been studies of women looking at moderate alcohol use and risk of drinking and those have suggested similar findings to men although the level of drinking that seems to confer lower risk is probably lower in women, that is to say, women need to drink less than for men. At the same time, there are a lot of risks of alcohol and some additional ones in women -- for example, risk of breast cancer that can't be ignored when trying to extrapolate results like ours to women. Some day we would like to extend these sorts of results to women. I hope we will be able to do that in the future.
MARGARET WARNER: Dr. Goldberg, for better or worse, this study is out there, got a lot of press attention. The headline was moderate drinking everyday is good for you. What should an individual do with this information? Should people who don't drink start drinking? Should people who drink on special occasions increase the frequency?
DR. IRA GOLDBERG: I think people who are concerned about heart disease or come from families that have high risk of heart disease should take this information, put it in their back pocket, and then go and do what has been proven to help them and what's been proven is to lower your cholesterol if it's elevated, to lower your blood pressure if it's elevated, to exercise, to raise your HDL by doing exercise -- to eat a healthy diet. Those are proven therapies and they are therapies that don't have a down side risk that alcohol maybe does have to it. I think the concern in a public health arena is that the message seems like a quick fix. Don't do everything that you should do that is right but rather have a couple of drinks. That's the concerning thing that when you have a line like this, people will do the easy thing instead of doing the correct thing.
MARGARET WARNER: Dr. Mukamal, what about that?
DR. KENNETH MUKAMAL: I agree that there are a variety of ways in which people can lower heart disease risk. I wouldn't discourage anybody from doing those. I don't think that they're mutually exclusive. That is to say that doing exercise doesn't preclude one from say eating a healthy diet and potentially from considering moderate alcohol consumption. I suppose my take on it would be that this is something people can bring to their doctors. I don't think anybody should take up drinking simply on the basis of results of a single study like ours but perhaps in consultation with their physicians, they can make informed decisions together.
MARGARET WARNER: But if a study like this doesn't suggest action, doesn't recommend an action, then of what value is it?
DR. KENNETH MUKAMAL: You know, it's difficult, I think to make a single one-size-fits-all public health recommendation from this. I agree very much with Dr. Goldberg in that regard. I don't think though that this is useless information. I think doctors can and ultimately need to use information like this until we have the sort of randomized trial evidence that ideally we need. Since we don't have that now and in fact we may not have that for decades, this is I think the best evidence that's available to date.
MARGARET WARNER: Dr. Goldberg, very briefly, do you think there's any risk that headlines or stories like this or reports on this will encourage people to drink who shouldn't drink at all?
DR. IRA GOLDBERG: Oh, yeah, I think absolutely. I think people come to my office and every other physician's office and they'll say, gee, I read this study. I'm not going to take my blood pressure medicine and I'm not going to take my cholesterol medicine. I'm not going to watch my diet because I read this study and it suggests that you could have one or two drinks a day. I think neither of us think that that is what patients should do. I think patients should realize this is research study published in a research journal. And this is the way we learn. This is the process of doctors and the public also learning. It very well may be that in the decades ahead, this might be an additive therapy. But currently we have other things and other things that are better proven and other things that are safer.
MARGARET WARNER: Doctors Goldberg and Mukamal, thank you both.