Diet-based approaches to weight loss are "a disaster" in terms of permanent reduction and should be abandoned, a Harvard Medical School physician told some of the world's top obesity researchers and clinicians last week.

On the basis of published reports of diet-based programs, said Dr. William Bennett of the Harvard Medical School's department of continuing education and editor of the Harvard Medical School Health Letter, "I can see no ethical basis for continuing research or treatment."

Speaking at the final session of a three-day conference on human obesity sponsored by the New York Academy of Science, Bennett said that "these diet-based studies involve a great deal of effort on the part of patients and investigator. They lead to almost inevitable failure on the part of participants -- and the failure is blamed on the participants, not on the researcher."

Bennett's indictment included one of today's most popular weight control techniques -- behavior modification.

"Strictly speaking," he said, "behavior therapy is a set of techniques that might be used to alter any habitual behavior from nose-picking to child abuse. In practice, as it is applied to fat people, it is a dietary approach."

"Moreover," he said, "most of this research does not include serious study of potential adverse effects, particularly of the dropouts, about whom we know nothing. We can hope that they are angry at the investigators, but in fact, many of them are suffering from severe loss of self-esteem."

The international community of obesity researchers and clinicians gather at regular international conferences to bemoan the increased fatness of western civilization and to repeat the same warnings about the life-threatening dangers and genuine social disability of obesity -- which they define "as body mass-index" over 30. The index is determined by dividing a person's weight in kilograms by the square of his or her height in meters. (There are about 39 inches in a meter, and 2.2 pounds in a kilogram.)

Yet they concede that there are only hypotheses about how people become fat and little to cheer about in terms of prevention -- much less cure.

But one area of progress is that researchers are beginning to learn that obesity is not a single disease, but a group of diseases with different causes and different manifestations, requiring different treatments.

Bennett's position -- that diets don't work -- was first popularized in his 1982 book, "The Dieter's Dilemma." It has been regarded as "nihilistic" by many of his colleagues, especially those who fear obese patients will simply give up, leaving clinicians very little to offer instead.

However, Bennett dismisses that criticism. "Blood-letting," he said, "was abandoned in the treatment of pneumonia about a century before penicillin was discovered. There is no indication in the literature that pneumonia patients suffered as the result of that bit of therapeutic nihilism."

Although many of his colleagues are hardly prepared to give up the behaviorally modified diet programs, and, at the conference bridled perceptibly at the suggestion that promoting diets was in any way unethical, some of them are beginning to concede that the so-called "set point theory," for which Bennett is one of the chief advocates, has some basis in clinical experience. The set point theory holds that each individual has a particular weight that reflects a balance between energy input and output, and that artificial attempts to alter it are quickly defeated.

Obesity expert Dr. Albert Stunkard, a University of Pennsylvania psychiatrist, said there is some evidence that something like a set point is working in some humans, and more clearly in some animals.

However, he said, "it is not appealing from the therapeutic point of view, because it sounds kind of . . . hopeless.

"If you're fat and your set point is elevated," he said, "you're in bad shape."

Stunkard himself has done some of the studies that show disappointing results from diet-based eating programs and most recently has conducted studies showing a clear genetic element in obesity. One published earlier this year in The New England Journal of Medicine took meticulous adoption statistics kept in Denmark and demonstrated that adults who had been adopted as children were thin, plump or fat according to the size of their biological parents, not their adoptive parents.

Still, Stunkard maintains that genetics are only part of the picture, representing a complicated, inherited vulnerability, rather than one specific gene that might be susceptible to eventual genetic engineering. "I haven't been able to interest the geneticists," he said. Nor, he says, has he been able to interest commercial weight loss programs in scientific studies of their results.

"There is not one single commercial weight loss program that makes available any data on its results or even wants to know what they are," he contends. "It's fantastic -- you take three of the largest groups, and they might allow you to study a small select group if you wouldn't tell anybody what you found . . . It isn't happenstance that there's not a single bit of scientific evidence that they are effective. The studies are actively opposed by the weight reducing industry."

Stunkard also has found that as unhappy as obesity makes people, and as victimized as overweight people are by prejudice and discrimination, dieting itself can produce emotional symptoms that are even more difficult to manage.

In a study he did recently at the University of Iowa, he said, "with a group at least 100 percent overweight, we found that the prevalence of emotional symptoms -- depression, anxiety, preoccupation with food, irritability, things like that, were even higher" than among mildly overweight dieters. Up to 80 percent of the morbidly obese people had these symptoms. "Very few of them could diet and lose any significant weight without feeling bad."

This may be an indication that in trying to lower their weight, they are eating less than their natural set point requries, he said. "Many of these are symptoms of starvation."

Dr. William Dietz, a specialist in childhood obesity, is not at all convinced that genetics play much of a role at all. Although it is true fat parents tend to have fat children, he said, studies show that they also tend to have fat pets. Nor, he says, is set point applicable to children whose weight and body fat distribution is constantly changing as they grow.

Dietz is best known for his finding that watching television is a predictor of childhood obesity. (However, contrary to expectations, fat TV watchers were not found to consume perceptibly more calories per day than thin TV watchers.)

Confronted with these apparently paradoxical findings and with mounting evidence about the serious health risks of obesity, there is intense activity to find new techniques, new drugs, new therapies and new understanding. Mostly, scientists have found new complications. Not the least of these is the obsessive and "fat-phobic" behavior among people of normal weight -- especially women -- and, frighteningly, among pre-pubertal girls.

"There is too much emphasis made on weight control for people who are not really at risk," said Dr. Trevor Silverstone, a London psychopharmacologist. "I think there is an unholy alliance between the fashion industry, the media and us, the doctors, to coerce what are plump, but not physically at risk, young women into the corset of conformity of the slim-line dress. I see much more of a problem, more hearteache and the whole train of eating disorders as a result.

"After all," he said, "anorexia nervosa and bulimia do not occur in socieities where thinness is not next to godliness."

Some 40 researchers and clinicians from the United States, Canada, England, Scotland, France, Sweden, Switzerland, Israel, Italy and Chile attended the New York conference, which was organized by neuroendocrinologist Richard J. Wurtman and biochemist Judith Wurtman, both from the Massachusetts Institute of Technology's department of applied biological sciences.

Some of the presenters also participated in last year's consensus development conference at the National Institutes of Health in Bethesda. However, where the NIH conference concentrated on the health risks of obesity, the New York sessions were, as Dr. Richard Wurtman noted in his opening remarks, designed to promote some optimism about "actually doing something about obesity -- something that has not, unfortunately but perhaps justly, characterized earlier conferences on obesity."

Wurtman said he had polled some 30 of his colleagues involved in obesity research, asking them for suggestions for the conference's focus. "It is interesting and perhaps symptomatic that each of them proposed a different focus," he said. "Virtually each one saw obesity as resulting from a different mechanism, a different problem from the other 29.

"Some said obesity equals fat cells; others said it results from people eating too fast, not chewing properly. Others saw it as a result of impaired energy utilization or as a hedonic disease involving the pleasure taken in taste, texture, smell and looks of food perhaps resulting from a disorder of brain opiates. Others saw it as a systems engineering problem involving set points. Still others viewed it as a characteristic of passive aggressive behavior as a way of manipulating one's associates. Some saw it as an enzyme defect; anthropoligsts saw it as a vestige of earlier periods in man's evolutionary history where every calorie was precious and it wasn't a bad idea to lay down some excessive fat."

The Wurtmans and members of their research team have been involved with a subset of overweight patients they have identified as craving carbohydrates. (They believe this may be the largest obesity subset of all.) The craving has been biochemically linked to the brain chemical serotonin, which is responsible, among other things, for communicating a feeling of satiety after eating.

The Wurtmans have hypothesized that some overweight people have too little serotonin and are directed by the brain to ingest more carbohydrate. To make serotonin, the brain cells need an amino acid called tryptophan. To get into the brain, tryptophan must compete with other, more plentiful, amino acids. When carbohydrates are consumed, insulin is secreted, interfering with other amino acids -- but not tryptophan. This allows more triptophan to get into the brain and stimulate serotonin production.

Serotonin affects mood and sleep, eases pain, and performs other necessary functions. If the system is malfunctioning, there can be mood as well as appetite disorders, the Wurtmans have hypothesized.

Indeed, some of the newest research to come out of the conference involves serotonin, including work with a group of seasonally depressed patients who suffered depression and strong carbohydrate cravings during the winter. Dr. Norman Rosenthal of the National Institute of Mental Health, who has identified and treated seasonal affective disorder (SAD) patients for about five years, noted that they report weight gains of up to 20 pounds in the winter, often, but not always, lost in the summer.

In France, pharmacologists have developed a drug, D-fenfluramine, which specifically increases serotonin activity and has built up a good record in weight control. Clinical trials in this country are just getting under way, although the drug is already on the market in France. Another similar drug, fluoxetine, is also being tested.

And Dr. Benjamin Caballero, working in the Wurtmans' MIT laboratory, has found that certain obese individuals have a resistance to insulin so that triptophan loses out in the competition with other amino acids, possibly causing a serotonin deficiency.

There were also presentations on energy storage and utilization -- cases in which even vigorous exercise did not reduce fat storage, with an enzyme abnormality possibly implicated, and some new information on "mechanical" controls from gastric bubbles, now being tested in New York and elsewhere, to tight gold chains that London models solder around their waists. Noted one researcher: "They'd rather be cut in half than gain weight."