For years, we have been seeing very large differences among countries in reported rates of death from CHD. The highest rates are now in Scotland, Northern Ireland, Finland and Eastern Europe. The lowest, with only one-fifth the number of heart disease deaths as Northern Europe, is Japan, due primarily to the very low amounts of fat in the traditional Japanese diet. We have always explained the lower rates of heart disease in Italy, Spain and Greece from the lower fat diet and other features of what is known as the Mediterranean diet. But France has even lower rates of coronary heart disease than Italy or Greece. Why do the French have such low rates of heart disease? We do not know. Their diet is higher in fat, their hypertension rates are the same, their blood cholesterol worse. Further, genes studied so far seem to be about the same as in countries with much higher rates of heart disease.

The explanation of the low rates in France include more intake of fruits and vegetables containing high levels of anti-oxidant vitamins, and a lower percentage of fat intake from red meat, as the meat in France is very low in fat and smaller portions are generally served then in the US. But the theory that has received the most scientific support is that the French consume large amounts of alcohol, on a regular basis, and particularly in the form of wine. This is not a new story, as we know from paper by St Leger et al back in 1979, that of all the lifestyle factors related to CHD in different countries, the strongest relation was with the average wine consumption in the country. Since then there have been a large number of studies on alcohol consumption among individuals and their risk of CHD, and the results have been remarkably consistent: individuals who consume alcohol moderately have fewer heart attacks.

How does alcohol produce this protective effect?

We now know many of the reasons why alcohol reduces the risk of CHD, as we have identified many of the biologic, physiologic effects of alcohol. Alcohol in any form affects the blood lipids. Alcohol will increase HDL-cholesterol, the so-called good cholesterol that lowers the risk of heart disease. Alcohol also tends to decrease slightly the LDL or bad cholesterol 'which increases the risk of CHD. But alcohol, and red wine in particular, also favourably affect the coagulation of blood within the arteries. For example, it decreases platelet aggregation, and it also has favourable effects on fibrinogen and fibrinolysis, factors related to blood clotting and the removal of clots within the arteries. The effects of alcohol on blood clotting may be just as important as the effects on the HDL-cholesterol.

The effects on coagulation are short-term effects, lasting for a day or so. The short-term beneficial effects were well-demonstrated by Jackson et al. These investigators compared the risk of heart attack and cardiac death among regular drinkers and found that if a regular drinker had had one or more drinks in the preceding 24 hours, his or her risk was reduced (to about 75% for men and 61% for women) when compared with drinkers who had not had anything to drink in the preceding 24 hours.

These results suggest that you should consume alcohol on a regular basis, perhaps daily. Unfortunately, many do not have good drinking patterns, and tend to drink nothing all week, then drink heavily at weekends, which is a very unhealthy way to consume alcohol. Many Europeans drink wine with their meals every day, it is not the amount that you drink, but when you drink it that is important in preventing CHD. In a recent study in Italy, people who drank wine with their meals were much less likely to die of any cause than those who consumed wine at other times. The best pattern is regular consumption with meals every day, but only a few drinks each day.

I am often asked, Is wine, particularly red wine, better than spirits or beer in terms of preventing heart disease? We do not have definitive data on this, and generally all types of alcoholic beverages have a protective effect against heart disease. However, we are accumulating new data suggesting that red wine may have additional benefits. A number of scientists have shown that red wine contains, in addition to alcohol, a large number of substances that are powerful anti-oxidants, tend to reduce blood clotting, and have other effects that should reduce heart disease risk. For example, we have data from the US showing that moderate amounts of red wine will prevent blood platelets from clumping together to form a blood clot which may lead to a heart attack. Such protection does not occur from white wine or alcohol, so it must be something in red wine that gives the added protection. On the other hand, in the US, wine drinkers are different in many ways from beer or spirits drinkers. They tend to be better-educated, have higher incomes, smoke less, and exercise more than beer drinkers. So it is difficult for an epidemiologist to be sure that wine drinkers are healthier because they drink wine, or whether it is just that people who have healthier lifestyles drink wine.

Should physicians prescribe alcohol to their non-drinking patients to lower their risk of heart disease?

It takes only about one drink per day, or even every other day, to get most of the benefit in terms of reducing coronary heart disease (CHD). Since this is the leading cause of death throughout the developed world, why not prescribe a glass a day to everyone? Some scientists argue that we do not have to use alcohol to prevent CHD, there are other ways to prevent it: losing weight, lowering blood pressure and cholesterol, eating a healthy diet. But they do not appreciate how hard it is to lose 10kg of body weight, or to modify diet. And it is not enough to simply tell people to eat better. Several good studies show that people who get lots of vitamins, especially vitamin E and folate, may lower their risk of CHD. But recent research by Rimm et al from the Nurses Health Study found that the reduction in risk of CHD associated with folate was many times greater for those who consumed at least 15g of alcohol per day versus non-drinkers. So a little wine along with a healthy diet will be preferable to just a healthy diet.

We know that alcohol abuse is very harmful, and the question I want to consider is what we refer to as the bottom line, i.e. total mortality. In other words, if you are a drinker, are you likely to have a longer or shorter life than you would if you did not consume alcohol? We all realise that even moderate drinking can have adverse effects if someone quickly consumes several drinks then takes their car on the road. So I always refer to moderate and responsible drinking. But in terms of diseases, most are related to heavy drinking. The only condition that may relate to even light to moderate drinking is breast cancer in women.

A number of research reports suggest that a small increase in risk begins to appear among women who normally consume only one or two drinks per day. This is not found consistently in all studies. At our Institute at Boston University we have just completed a study of wine, beer and spirits as they relate to breast cancer, using data from the Framingham Study that has been following more than 5,000 women for 25 to 45 years. We found that women who have consumed wine or other alcoholic beverages moderately over a lifetime do not have increased rates of breast cancer in fact, they tend to have slightly lower rates.

Let's turn to the total mortality rates. In other words, are you more likely, or less likely, to die of any cause during a specified period if you drink or if you do not drink? In essentially every prospective study carried out, the net effect on total mortality of consuming one or two drinks per day is a lower death rate. As long as the alcohol intake is light to moderate, up to a couple of drinks per day, the risk of most diseases is lower, and the risk of dying of any cause is reduced.

Data from the Copenhagen Heart Study, that followed over 13,000 men and women for over 10 years, illustrate how reducing moderate drinking might increase mortality. The results from the Copenhagen Heart Study show the typical J-shaped relationship between the relative risk of death from any cause and alcohol consumption. Drinkers who stated that they averaged one to six drinks per week had about 40% lower death rates than non-drinkers; heavy drinkers had higher death rates.

Whatever we do to reduce alcohol consumption among heavy drinkers, irresponsible drinkers and alcoholics, from the point of view of public health, we should be sure that we do not discourage moderate use.

How should all this information relate to alcohol policy?

What should the public be advised about alcoholic beverages? What should physicians advise their patients, especially those who do not drink? Colleagues and I are preparing a report using official death rates in the US to determine the absolute risk of dying for adults 45-75 years of age, according to whether they are abstainers or consume alcohol moderately, which by our governments official definition is an average of no more than two drinks (US 'drink' = 1 1/2 UK units) per day for men, and one drink per day for women. Based on an extensive review of the scientific literature, we assumed that if a non-drinking person began to consume alcohol moderately, he or she would reduce his or her risk of CHD by 33.3%, and would reduce their risk of stroke by almost as much. Other diseases related to arteriosclerosis would be reduced, but women might show a 10% increase in their risk of breast cancer.

We know that there is always the possibility that a patient will not follow his or her doctors suggestions and begin to drink very heavily. We factored into our equations that approximately 5% of patients might do this, and if so, would increase their risk of cirrhosis of the liver, death from drunk driving, and an increase in the risk for breast cancer for women.

Taking all these risks into our calculations, we found that for all women from age 55 and for all men aged 45-75, individuals who began to drink had lower mortality rates. The results indicate that a non-drinking patient would actually have a reduced risk of dying over the next 10 years if he begins to drink moderately, even if 5% of such subjects became alcohol abusers. As an example, for a 65 year old non-drinking man in the US, government statistics show that 278 will die over the next 10 years; if these men all became moderate drinkers, only 251 would die, a reduction of about 10%. Our data are based on the consumption of any type of beverage.

Physicians should no longer be so afraid of alcohol abuse that when they find middle-aged men or post-menopausal women who do not consume alcohol, or only do so rarely, that they stick to the old adage of leaving well alone. I believe that for many patients, doctors can encourage moderate alcohol consumption, especially the daily consumption of a glass of wine, without any real risk of the patient abusing wine. And in doing so, their risk of heart disease could be markedly reduced.

In summary, the scientific data are quite clear: moderate alcohol consumption can be considered a part of healthy lifestyle.