January 11, 2006

To Drink, or Not to Drink

1501 page8 capHE SHARING OF A GLASS OF WINE has long been considered sociable and sophisticated. In recent years, however, research has suggested that moderate drinking may also be good for you! Can this be true? Or could the widely publicized benefits of a daily glass of wine or a can of beer be largely the product of hype? Is it possible that the advice to drink in moderation--?for your health?--is all a mistake?

It all started some 18 years ago with the description of the ?French Paradox.?1, 2 Why, some scientists wondered, did the French have so few heart attacks when they were not exactly committed to enthusiastic exercise and they ate a diet high in both cholesterol and fat?

The scientists concluded that this had to do with the type of fat (monounsaturated olive oil) that the French ate, and the fact that they drank alcohol daily, often in the form of red wine.

Many articles followed, all agreeing that moderate wine drinking was healthful as well as enjoyable. One particularly convincing study--known as the J-curve study--described the likelihood of death from heart attacks and strokes.3, 4 The J came from the appearance of a graph showing the chances of dying each year compared to daily alcohol intake. What the scientists discovered was that both heavy drinkers and those who drank no alcohol at all had a higher death rate than those who drank alcohol in moderation. Moderate drinkers had the lowest death rate of all the groups. The result of this study was a graph shaped like a big letter J--and so it became known as the J curve of mortality.

The evidence did seem convincing. And all subsequent studies have shown the same thing. Those who drink moderately live longer than those who don?t drink at all. This was unexpected and, until recently (18 years later), had not been satisfactorily explained.

What Is Alcohol?
Alcohol is neither a vitamin nor a mineral. It contains no essential building blocks for body tissues. True, the body can use it in small quantities for energy--but that is so for every other food we eat as well. The body can do just fine without alcohol. So why should alcohol protect against untimely death?

1501 page8 link

I Just Said NoThe Benefits of Being Alcohol-freeI Prefer a Different Form of FunWhy I Don't DrinkStraight A's and HangoversAlcohol--The Real Picture

The apparent life-prolonging effects of alcohol are well established and virtually unquestioned. This has caused a real problem for both public health officials and practicing physicians. Public health officials are responsible for advising cities, states, or countries on health matters. Physicians are responsible for advising individual patients. What should they say? Alcohol has a dark side--a very dark side indeed!

? Alcohol abuse is the third leading cause of death in the U.S., killing more than 75,000 every year.5

? It indirectly increases death rates from other major diseases, including breast cancer.6

? More than 20 percent of moderate drinkers report episodes of binge drinking (five or more drinks at one time)--thus increasing risk of injuries and violence.7

? Alcohol is involved in 40 percent of highway fatalities.8

? Approximately 10 percent of those who start drinking after age 20 (and 40 percent who start drinking before age 14) will become alcohol-dependent.9

Understandably, public health officials and physicians were perplexed. Since alcohol is the cause of so much suffering and so many unnecessary deaths, what advice should they give? Should they advise nondrinkers to start drinking--moderately, of course--so that they would live longer? Or should they advise against drinking? Most health counselors ended up advising those who were already drinking to drink only in moderation, and those who had not started drinking to remain teetotalers. For most health advisers, the ratio of risks to apparent benefits just didn?t seem to add up.

Apparent, but Not Certain
Apparent health benefits! Why are these health benefits only apparent, and not 100 percent certain? After all, every scientist who had studied the matter agreed that those who drink in moderation live longer. Why put ?apparent? in front of health benefits? The answer is tricky because it has to do with statistics and equations--you know, those pesky things that you skipped over in school and figured that they fell into the category of things in life that were fine in their place but unnecessary for you! Here is where statistics become very important--maybe even life-and-death important.

You see, although everyone agreed that people who drink in moderation lived longer, no one had shown that it was because they drank in moderation that they lived longer. Why had no one bothered to check this? The simple answer is that it is very difficult to do with people. With mice it is simple: with humans it is almost impossible.

1501 page8Imagine taking a large group of people and pairing them up so that each member of a pair matched the other in every characteristic that would be important for such a study (e.g., age, sex, health status, etc.). Then imagine putting one member of each pair in a group that would drink no alcohol with the other pair member in a group that drank moderately. Then, for a period of several years, imagine keeping both groups on identical diets!

That sort of an experiment would be very difficult and very expensive, and would take years to complete. However, it is precisely the sort of experiment that would show whether moderate amounts of alcohol lengthen lifespan.

That experiment has never been done--nor is it likely that it will be done any time soon. Instead, we assume that moderate drinkers live longer because of the alcohol they drink. Behind this assumption is a very large--and very well-hidden--additional assumption. (It must be well hidden, because no scientist has checked it in the past 18 years.) This second assumption is that moderate drinkers and teetotalers are identical in every way except for alcohol intake. But if they, in fact, differ in ways other than alcohol intake, there is a possibility that it is these other differences that are causing moderate drinkers to live longer.

Confounded Wisdom
The technical term for other differences between the groups is confounders. In 2004 a group of scientists went looking for confounders in the studies on alcohol and longevity--and they found them, lots of them. The study findings were published early this year under the intimidating title of ?Cardiovascular Risk Factors and Confounders Among Nondrinking and Moderate-Drinking U.S. Adults.? 10 In this report, Timothy S. Naimi and his collaborators (11 coauthors) analyzed whether nondrinking and moderate-drinking populations are comparable. The answer is--they are not! The moderate drinkers are much less likely to suffer from common diseases (see Table 1).

The moderate drinkers are much more likely to be well-educated, earn more money, etc. (see Table 2). In fact, the moderate drinkers did significantly ?better? on 27 out of 30 factors looked at.

All this information was obtained from the 2003 Behavioral Risk Factor Surveillance System, which surveyed 250,496 individuals. The Centers for Disease Control (CDC) in Atlanta, Georgia, funded and supervised the data collection for purposes other than the effects of alcohol on health. They were looking for developing diseases or health threats in the population. After excluding the sick and the heavy drinkers (more than two drinks a day for men, one a day for women) from the analysis, 116,841 nondrinkers and 118,889 moderate drinkers remained.

1501 page8There were confounders aplenty in this very large study of the differences between nondrinkers and moderate drinkers. These two populations are not very similar at all! But surely alcohol doesn?t account for all the differences. Moderate drinking isn?t likely to cause a person to finish college, earn more money, and have more hobbies! Rather, better-educated people with more money and time for hobbies are more likely to drink moderately in social situations than to be either teetotalers or heavy drinkers. They are also more likely to exercise regularly, weigh less, watch their cholesterol levels . . . and live longer! Indeed, that is just what the study showed.

Those with more college degrees and higher income (the same group who drank moderately) were less obese, had lower cholesterol levels, and, more often than the teetotalers, had a very low risk of heart attacks and stroke.

What are we to conclude? Science being what it is, other similar data sets will certainly be examined closely to see if these findings hold up. Since alcohol-related disease is a significant burden on the health of the U.S., it is even possible that large, properly controlled, long-term studies will now be funded to settle this question once and for all.

Meanwhile, it would seem wise to behave like the moderate-drinking group. Get:

  • More education, which will probably bring . . .
  • Higher income
  • More knowledge of health and dietary matters
  • More hobbies, etc.

And skip the alcohol: that way you will have the best of both worlds. In short, act like an Adventist!

1 J. L. Richard, ?Coronary Risk-Factors--the French Paradox,? Archives des Maladies du Coeur et des Vaisseaux 80 (1987): 17-21.
2 S. Renaud and M. de Lorgeril, ?Wine, Alcohol, Platelets, and the French Paradox for Coronary Heart Disease,? Lancet 339 (1992): 1523-1526.
3 A. L. Klatsky, G. D. Friedman, and A. B. Siegelaub, ?Alcohol and Mortality--A 10-Year Kaiser-Permanente Experience,? Annals of Internal Medicine 95, No. 2 (1981): 139-145.
4 A. L. Klatsky, M. A. Armstrong, and G. D. Friedman, ?Alcohol and Mortality,? Annals of Internal Medicine, Oct. 15, 1992, pp. 646-654.
5 M. A. Stahre, R. D. Brewer, T. S. Naimi, J. W. Miller, L. T. Midanik, F. J. Chaloupka, R. Saitz, T. L. Toomey, J. L. Fellows, M. Defour,
M. Landen, and P. J. Brounstein, ?Alcohol-Attributable Deaths and Years of Potential Life Lost--United States, 2001,? U.S. Morbidity and Mortality Weekly Report, Sept. 24, 2004, pp. 866-870.
6 Ibid.
7 Substance Abuse and Mental Health Services Administration, Results From the 2003 National Survey on Drug Use and Health: National Findings (2004) (Office of Applied Studies, NSDUH Series H-25, DHHS Publication No. SMA 04-3964), Rockville, Md. 
 8 Persons Killed, by State and Highest Blood Alcohol Concentration in Crashes (2003). National Center for Statistics and Analysis (accessed 5-26-2005 at http://www.fars.nhtsa.
 9 B. F. Grant and D. A. Dawson, ?Age at Onset of Alcohol Use and Its Association With DSM-IV Alcohol Abuse and Dependence: Results From the National Longitudinal Alcohol Epidemiologic Survey,? Journal of Substance Abuse 9 (1997): 103-110.
10 T. S. Naimi, D. W. Brown, R. D. Brewer, W. H. Giles, G. Mensah, M. K. Serdula, A. H. Mokdad, D. W. Hungerford, J. Lando, S. Naimi, and D. F. Stroup, ?Cardiovascular Risk Factors and Confounders Among Nondrinking and Moderate-drinking U.S. Adults,? American Journal of Preventive Medicine 28, No. 4 (2005): 369-373.

Brian Bull, M.D., is professor and chair of the Department of Pathology and Human Anatomy at Loma Linda University School of Medicine in California.