COLUMN ONE : Refusing to Swallow Diet Myths : A mild-mannered Harvard researcher is causing a stir by debunking common beliefs about nutrition. You may want to put off switching from butter to margarine.

BOSTON —  In the hallway of an aging brick laboratory next to the Harvard School of Public Health, there sits an unusual archive: a metal filing cabinet, unmarked and guarded by lock and key, stuffed with tiny Manila envelopes containing the toenail clippings of nearly 70,000 American nurses.

The clippings will give Harvard researchers clues about whether certain minerals, contained in food and deposited in nails, affect breast cancer. Collecting them was the clever idea of a man with a mission, a mild-mannered Harvard professor who is conducting the world’s largest study of how the foods we consume contribute to the illnesses that consume us.

His name is Walter Willett, and his aim is to put scientific facts behind the old adage “you are what you eat.”

That may sound simple enough. But in an era when one can hardly pick up a newspaper without being assaulted by so much contradictory dietary advice that none of it seems believable, when the President of the United States must break up fights between his Cabinet members over what kind of labels to put on food, experts agree that the scientific evidence linking diet to disease is surprisingly, shockingly, slim.

Within the hallowed halls of Harvard, Willett and his team of several dozen researchers are on a furious quest to fill in the gaps. Again and again, their massive examination of the dietary habits of 121,700 female nurses, launched 12 years ago, has debunked dietary myths. And few take more delight in bashing myths than Willett.

Thought you were doing yourself a favor by drinking decaffeinated coffee? The Nurses’ Health Study has found that caffeine won’t hurt your heart, but decaf might. Cutting out wine and beer? Willett says moderate alcohol consumption actually lowers the risk of heart attack. But women beware; alcohol also raises the risk of breast cancer.

Gave up red meat? Give yourself a gold star. Willett’s research says red meat raises the risk of colon cancer--a finding that, one colleague of Willett recalls, brought the placid professor face to face with “a couple of big tall Texans who came strolling in” to discuss the impact those nasty little studies were having on their cattle ranching business.

Switched from butter to margarine? Demerits are in order. Willett and his group are about to report that margarine is associated with increased risk of heart disease--a finding that, Willett says with more than the slightest hint of glee, is certain to create a stir.

And remember those reports about Chinese women who eat a low-fat diet and rarely get breast cancer? Bunk, says Willett, much to the chagrin of scientists at the National Cancer Institute and elsewhere who continue to place faith in this hypothesis--and to spend money on studies to prove it. The nurses reveal no connection.

Such findings--and there are plenty more on the way--have angered food industry executives, altered government policy and often exasperated dietitians who are struggling through the maze of conflicting reports to formulate nutritional advice. They have also earned Willett a place as perhaps the most respected and provocative researcher in the field of nutritional epidemiology, the relationship between diet and disease.

It is a relatively new scientific arena, and one that Willett has played a major role in developing. In fact, he wrote the book on the topic--a 396-page tome, published in 1990 and titled, appropriately enough, “Nutritional Epidemiology.”

“He is viewed as the leader in the field of nutritional epidemiology in this country or in the world,” said Louise Brinton, who supervises research involving the effect of environment on cancer for the National Cancer Institute. “I never hear nutritional epidemiology being mentioned without reference being made to Walter Willett.

“Really, until these studies had been launched, we weren’t sure which components of diet might either be helpful or hazardous. A lot of what we hear . . . is based not on science but on speculation.”

That speculation, rather than science, is helping to shape beliefs about which foods are healthful and which are not annoys Willett endlessly. Even more distressing, he says, is that this speculation has worked its way into official nutrition advice, which Willett says is often based on guesswork extrapolated from data that is incomplete.

Take, for instance, the American Heart Assn.’s well-known recommendation that no more than 30% of calories come from fat. A nice round number, says Willett, “but it’s not based on much.” More specific information is needed, he says, about how different types of fat--saturated, polyunsaturated, mono-unsaturated--affect disease.

He continues: “When one looks at some of the recommendations about what we should be eating as a whole country, the evidence is often rather appalling. Just for example, there are a lot of recommendations that we should eat fewer eggs. You would think there would be a handful of studies showing that people who ate more eggs have more heart attacks, but there aren’t any.

“And we have all this information about fat. You would think there would be lots of studies that people who had lower fat intake had lower risk of heart attack and cancer. But the evidence is incomplete. Name almost any important topic, and the data are very meager. . . . A lot has been said, rather dogmatically, without a strong scientific basis.”

Tall and lanky, with a wiry mustache and deep-set blue eyes, Willett looks the part of the tousled scientist. His office is devoid of diplomas and awards, although his chairs--bearing the Harvard motto, Veritas (truth)--give a clue to his philosophy. He favors Oxford cloth shirts and tweed blazers. He rides his mountain bike to work--a 12-minute jaunt across the Charles River from his home in Cambridge to Boston--and carries his lunch to the office in a backpack.

Peanut butter and jelly? No way. Willett practices what he preaches. Lunch is a spinach and tomato sandwich, on whole grain bread brushed with olive oil. Willett is big on olive oil. He even talks of baking cookies with it--a remark that elicits groans from his students, who know as well as any plump grandmother that olive oil cookies just won’t taste the same.

As his studies expand, Willett’s culinary repertoire narrows.

Steak is no longer part of the Willett lexicon; he gave up red meat after the colon cancer study. He has also quit eating margarine, and has started taking Vitamin E supplements, based on a recent finding that the vitamin helps lower the risk of heart disease. The optimal diet, he says, is the Mediterranean menu--plenty of fruits and vegetables, very little meat or chicken and olive oil as the main form of dietary fat.

Says one colleague of Willett: “He puts his money where his mouth is.”

He is also a big believer in carrying his work out of the ivory tower and into the realm of public policy.

A case in point: Shortly before Thanksgiving, armed with a finding that folic acid helps prevent birth defects such as spina bifida, Willett went to Washington. There, he and a panel of experts recommended to the Food and Drug Administration that the nation’s food supply be fortified with the vitamin, in much the same way that flour is fortified with iron.

Should this recommendation be adopted--and it is likely to be--the move would be extraordinary on two counts. First, such a sweeping decision is extremely rare. And, according to Willett, it would represent a complete turnabout for federal policy-makers, who--based on what he described as a “serious misinterpretation” of the data--had been considering urging Americans to consume less folic acid, not more.

This may seem an obscure matter, and certainly it has not attracted much public attention. But if birth defects decline as a result, then Willett’s work will have paid off.

“Most of us just don’t bother with that,” said Meir Stampfer, a colleague of Willett at Harvard. “I’ve presented data to the FDA but I’ve never really taken the time and energy to push something through. I figure well, it’s not my job. . . . It’s a thankless task. There’s no recognition, no glory in it at all.”

Willett hardly seems on a quest for glory. He is soft-spoken and self-effacing, almost to the point of seeming shy. He doesn’t foist his opinions on others, but when asked, he is unafraid to speak his mind--a rare quality in the cautious world of academic research, where caveats litter nearly every phrase.

“I was astounded to hear him be quoted in major newspapers saying that the ideal amount of meat consumption may be zero,” said Bonnie Liebman, a nutritionist at the Center for Science in the Public Interest. “Researchers are rarely that blunt. They say, ‘Well, we need further research, or perhaps we should cut back. . . . It’s going out on a limb to say that the healthiest amount of meat is no meat at all. And yet, he’s willing to be bold like that.”

Such boldness sometimes irks Willett’s peers.

“What he has done is told the truth, and the truth has happened to be at odds with some of the political correctness in the field,” says Saxon Graham, a forerunner of Willett in the field of nutritional epidemiology who is now retired from the State University of New York at Buffalo.

“Some of the leaders in the field,” Graham continued, “are sure that fat causes breast cancer, and Walter had the nerve to come out and say that his studies didn’t show that. He just comes out and says what he finds, and let the chips fall where they may. And that’s what gets them all shook up.”

One of those who is “all shook up” is Neal D. Barnard, president of the Physicians Committee for Responsible Medicine. Barnard, along with other critics of the breast cancer study, argues that Willett’s work is flawed because the nurses whom he considered on a low-fat diet still consumed as much as 29% of their calories from fat.

“In effect,” Barnard said recently, “Willett studied a high-fat diet versus a very high-fat diet.’

Willett dismisses such criticism, saying that though low-fat diets are a good idea in general, there is no evidence that an even lower fat diet would have made any difference in breast cancer. He accuses the medical establishment of allowing politics to interfere with research, and takes particular aim at the National Cancer Institute, which is embarking on a massive women’s health study that will include an examination of the breast cancer-fat hypothesis.

“There is a lot of frustration about breast cancer, and there has been a lot of public pressure to do something about it,” says Willett. “The unfortunate political response was, ‘Here, let’s spend a lot of money on doing this huge trial and we can say we’re doing something about it.’ ”

Left to sort out the truth from this scientific flap are the nation’s nutritionists, among them Dale M. Ogar, managing editor of the Wellness Letter, a highly respected nutrition newsletter published by UC Berkeley.

“Oh, the breast cancer stuff,” Ogar said wearily, when asked about Willett. “That’s this week’s scare. Every day there is a new study, a new headline, a new scare. Eat this and don’t eat this and then eat this again”

Ogar’s message on the controversial finding: Ignore it. A high-fat diet is not good for you, whether it causes breast cancer or not. And that, says Willett, is just fine with him--so long as science is driving the decision.

At 47, Willett has devoted the better part of his career to this line of work. He joined the Nurses’ Health Study in 1977, one year after it was founded by Frank Speizer, a Harvard medical professor who wanted to examine the relationship between oral contraceptives and cancer. Willett thought the subjects of the study--all female nurses, ages 30 to 55 at the outset of the study--would make ideal candidates for a long-running examination of the relationship between diet and illness.

The nutritional component of the Nurses’ Health Study was launched in 1980. More recently, with the original corps of nurses now growing older, Willett has launched a second study of younger nurses, as well as a study of male health professionals.

All together, these three studies enable Willett and his Harvard crew to examine the health and eating patterns of 300,000 people--a population bigger than the city of Santa Ana.

Conducting work on this vast scale is not without drawbacks. For starters, the research--which relies primarily on the subjects’ responses to questionnaires, along with samples of their blood, their drinking water and, of course, the toenails--takes up a lot of room. Already, Willett’s group has filled 17 cylindrical freezers with 33,000 blood samples, and it is running out of space.

Then there are the logistics. On a recent Monday morning, Willett and his team spent nearly two hours grappling with the details of how to obtain yet another 20,000 samples of blood and drinking water, this time to study the effects of diet on prostate cancer in men.

Among the familiar questions on the table: Would Federal Express refuse to send blood samples through the mail? Would United Parcel Service be cheaper? How long would it take to put together the supplies for the blood sample kit--20,000 sets of tubes, needles, Band-Aids and other sundries?

Once collected, the samples will be stored. Over time, some of the men will inevitably develop prostate cancer. Then, the scientists will be able to compare the blood and drinking water, as well as the dietary information gleaned from questionnaires, of those who remained healthy to that of those who get sick.

The toenails are used in much the same way. The clippings, which were gathered 10 years ago, retain mineral deposits for decades. Researchers rely on them to determine how much of the trace element selenium the nurses consumed in their diets. As some nurses develop breast cancer, the scientists can return to the toenails to compare selenium intake of the sick to the healthy. So far, despite studies showing a link between the mineral and breast cancer in animals, the nails have revealed no connection.

In scientific circles, this type of work is known as a “prospective cohort study,” meaning a large group of people--a “cohort”--is followed over a long period of time. A major advantage of these studies is that researchers can collect a broad range of data on their subjects before they become ill, rather than backtracking after illness or death occurs, as in “retrospective” studies; a drawback is that they take years to produce findings.

This is common practice for epidemiologists, people who study causes and patterns of disease. But until Willett came along, the sophisticated methods of epidemiology were not often applied to the study of nutrition. Many scholars theorized that the work would be flawed, based on the subjects’ faulty memory of what they had eaten.

Willett blew holes in that theory when he devised a dietary questionnaire that has come to be considered the gold standard of his business. Then, he proved the questionnaire results were valid by comparing them to an extremely detailed examination of the eating habits of 200 nurses who, under supervision, weighed and measured everything they ate.

Now, after 12 years of probing the link between diet and health, Willett’s work is bearing fruit. But as he readily admits, a host of questions remain, and answering them could keep him busy for the rest of his academic life.

“How much and what type of fat should we be eating?” he asks. “Should we be taking more vitamin A in our diet? The evidence, in a very general way, that fruits and vegetables are good for us is strong. But the evidence is rather vague as to what types of fruits and vegetables are important, how much is optimal?”

Stay tuned, and be patient. The myth-basher of Harvard Square is at work.

The Word From Willett

Red meat: May cause colon cancer. The safest amount of meat may be no meat at all.

Dietary fat: No apparent links with breast cancer, but a culprit in other diseases.

Margarine: May increase the risk of heart disease. Stick to olive oil.

Alcohol: Moderate consumption--slightly less than two drinks a day--may cut the risk of heart attack, but increase the risk of breast cancer. A conundrum.

Decaffeinated coffee: Large amounts may increase the risk of heart disease. Drink in moderation.

Caffeine: Little if any risk.