A controversial new British report challenges the popular belief, which is supported by more than a dozen medical studies, that drinking alcohol in moderation reduces the chance of dying from heart disease.

The report and an editorial in a recent issue of The Lancet assert that the supposed benefit of alcohol on the heart is a "myth" created by a flawed interpretation of death rates among drinkers and non-drinkers.Several American researchers discounted the British findings and maintained that the evidence for alcohol's protective effect remains strong and persuasive.

At issue is the meaning of the "U-shaped curve," a pattern that has emerged in many studies that attempt to chart the relationship between alcohol intake and the death rate from heart disease. At the left of the graph are "non-drinkers," at the right are "heavy drinkers" and in the middle are those who drink in moderation. Because the heart disease death rates were highest at the left and right ends and lowest at the middle, the resulting curve is U-shaped.

Many experts have interpreted the U-shaped curve to mean that moderate drinking reduces the risk of heart disease, perhaps because alcohol increases the level of high-density lipoprotein (HDL), the so-called "good" cholesterol.

But the report by Dr. A. Gerald Shaper and two colleagues at London's Royal Free Hospital School of Medicine suggests that an illusion of a higher death rate from heart disease among non-drinkers is created because many people counted as "non-drinkers" in studies are people who gave up alcohol after learning they had heart disease.

Doctors, the public and the liquor industry have been too quick to embrace the idea that moderate drinking is beneficial because they want to believe it, said Mary Walker, one of the study's authors.

"The message has tended to be adopted very rapidly," she said. "We are not saying specifically that light and moderate drinking is bad for you. We are simply saying that it should not be interpreted as protective."

The British researchers' findings were based on a survey of 7,735 middle-aged men who were questioned in detail about their diet, drinking and smoking habits and their medical history. The researchers kept track of the men for more than seven years, recording the causes of all deaths.

The death rate from all causes and the portion from heart disease reflected the pattern seen in previous studies, with lower rates for light and moderate drinkers than for non-drinkers. In this study, the alcohol-consumption levels considered "light" and "moderate" were higher than in some American studies. Light drinking was defined as consuming between one and 15 drinks a week, moderate drinking between 16 and 42. A "drink" was equal to one shot of liquor, one beer or one glass of wine, all of which are roughly equivalent in alcohol content.

The researchers then divided their subjects into two groups: 1,873 men who had a history of heart disease or related conditions when they entered the study, and 5,856 men who did not. Then they studied the relationship of alcohol intake to death rates in each group.

In the men with pre-existing heart disease, light or moderate drinkers still showed lower death rates than non-drinkers. But in the "healthy" group, there was no relationship between alcohol use and death rates from heart disease.

Previous findings from the study had shown that men, upon learning they have heart disease or another illness, tend to reduce their drinking. So the researchers suggested that the higher death rates seen among non-drinkers occurred because that group contained a disproportionate number of people who had given up drinking because they knew they were ill.

Walker said other studies may have been flawed because they paid too little attention to such shifts in drinking habits. "It's a dynamic process that has been looked at in rather a snapshot way," she said.

Two American experts took issue with that view, saying that the evidence that moderate drinking protects the heart is persuasive.

"It's very, very impressive," said Dr. Thomas B. Turner, dean emeritus of the Johns Hopkins School of Medicine.

Turner said that, contrary to allegations by the British group, a number of studies have allowed for the effect of shifting drinking patterns. He noted that the Framingham Study enrolled almost 5,000 young adults in Framingham, Mass., and recorded their habits and medical histories in detail for 24 years. It found that moderate drinking was associated with lower rates of heart disease in men and in women smokers, although not in non-smoking women.

Dr. Meir J. Stampfer of Harvard Medical School reported this year that in a study of 87,000 female nurses, heart disease rates were about 50 percent lower in light and moderate drinkers than in non-drink-ers. The levels of drinking in the nurse study were considerably lower than those in the British study, with the heaviest drinkers having about two drinks a day.