"Prospective Population Studies  on Wine and Health"

Morten Grønbæk. M.D., PhD
Recipient of the Society's 2001
Leon D. Adams Wine Research Award
April 25, 2001
Mark Hopkins Hotel, San Francisco

Mr. President, Ladies and Gentlemen, dear Colleagues:

I must admit that until a few month ago I did not know about your Society, but having been taught about the efforts of Leon Adams by Al de Lorimier and Ron Unzelman, I now know what a great honour it is to receive this prize.

When I prepared this talk, I had an ambition of giving a general overview of all population studies on the relation between wine and health, but due to this short time, I have chosen to give a short overview of my own studies. Nevertheless, I would like to start out with mentioning two types of the studies which inspired me to take up this research, namely the initial ones of the relation between alcohol and mortality and the correlation studies that showed that wine drinking countries had a lower coronary heart disease mortality than others. I am sure you have all seen the curves by Boffetta et al showing the U-shaped relation between alcohol and mortality, which has been reproduced in many large prospective studies from all over the world. Boffetta was one of the first who very elegantly showed that the high mortality among non-drinkers cannot only be attributable to disease at baseline nor to underestimation of alcohol intake. First, if disease at baseline would be a cause of the descending leg of the curve, it seems rather odd that this should only be for coronary heart disease. Secondly, if under-reporting or even lying about one's intake should be a cause of the descending leg, it also seems rather strange that people who under-report or lie have no increased risk of causes other than coronary heart disease unless you can lie to your body to prevent cirrhosis. Further, it is also well known that there are several biological mechanisms that can explain this apparent cardio-protective effect of alcohol itself.

Many years ago St. Leger showed that there seems to be a negative correlation between wine intake in different countries and death from ischemic heart disease in the same. Not that much attention had been paid to this, until Serge Renaud took up the issue in 1992, which was actually a response to the first of the 60 Minutes programmes on the French Paradox. This curve, showing the low incidence of coronary heart disease in wine drinking countries and a high incidence in the non-wine drinking countries, inspired me to go a little more into details with this in prospective population studies. One must emphasize the large differences in ability to infer causality in looking at either a correlational study or prospective population study. In the correlational study many other factors may as well be correlated to the low incidence of coronary heart disease. For instance, it could be many other differences between the different countries causing the lower coronary heart disease risk in France as compared to the US. When looking at prospective population studies, we are actually able to assess the risk of death or risk of getting a certain disease at an individual level and thereby differentiate between, for instance, categories of wine intake. In the Copenhagen City Heart Study several years ago, we showed that there was the same U-shaped relation between total alcohol intake and coronary heart disease as in many other studies. We took a step further and looked at the different effect of the different types of alcohol and found that beer intake at the level from non-drinking to light to moderate daily intake did not influence mortality, while wine seemed to have a beneficial effect on all cause mortality and spirits intake in the high intake group was even associated with a higher risk. The relation between beer, wine and spirits and coronary heart disease mortality was also shown in the same study published in the British Medical Journal several years ago. Later, several studies both from Sweden, California and UK have reached the same conclusion. Hence, Wannamethee and Shaper showed that in the British Regional Heart Study wine drinkers also seemed to be at a lower risk than beer or spirits drinkers. Several mechanisms may explain this apparent additional beneficial effect of wine compared to beer or spirits. For instance, a few studies have shown that the intake of flavonoids in the diet may have a protective effect against coronary heart disease; secondly, studies from University of California, Davis, have shown that LDL oxidation is inhibited by phenolic compounds in wine. In the last few years we have gone through the effect of wine versus beer and spirit on many different diseases. For some reason many of those were published in 1998. The first one was by my colleague Dr. Truelsen who showed that neither beer nor spirits seemed to have an effect on ischemic stroke, while wine had a beneficial effect on ischemic stroke. The second one, which we published in the British Medical Journal, also in 1998, showed that the relation between alcohol and upper digestive tract cancer was the same as known in many other studies, but when we looked at the percentage of wine of total alcohol intake it turned out that only those who did not include wine in their total alcohol intake had an increased risk of upper digestive tract cancer while those who had a large percentage wine of the total intake may even have had a protective effect on upper digestive tract cancer. This was not due to different total mean alcohol intake in the different percentage groups. Also in 1998 a colleague of mine, Dr. Prescott, published a paper on the relation between beer, wine and spirits and lung cancer. We found that alcohol itself was positively related to lung cancer, meaning that those who had a high alcohol intake had a high risk of lung cancer. The study was well controlled for tobacco consumption. Now, these studies, and the last one I am going to present, the one from Annals of Internal Medicine from last year, is about to bring me closer to the conclusion that there are differences in the effect on health of wine, beer and spirits. It may not be on coronary heart disease, hence, we are aware of the cardioprotective effect of ethanol in all the different types of alcoholic beverages. But when we differentiate between the different causes of death and look at the effect of including and not including wine in one's alcohol intake it certainly looks as if wine itself has a beneficial or maybe a less harmful effect on several types of cancers.

Again, I should very much like to thank the Society for having bestowed upon me this prize.