Last Month, a Provocative Article Argued That a Low- Fat Diet May Be Responsible for America's Obesity Epidemic -- and That High-Fat May Be Healthier. The Problem: A Lot of Good Science Suggests Otherwise
The Washington Post - Washington, D.C. Author:
Sally Squires Aug 27, 2002 F.01 HEALTH TAB 4509
Text Word Count:
On Sunday, July 7, science writer Gary Taubes created a big fat splash.
In a cover story in the New York Times Magazine, the award- winning freelance author questioned whether the long- standing public health support for low-fat diets was wrong. Taubes suggested that the Atkins diet -- a low-carbohydrate, high-fat approach permitting nearly unlimited portions of fatty meats but only small amounts of vegetables and virtually no fruit, grains or milk products -- might be healthier and more effective for weight loss. A photo of a steak topped with a pat of butter ran on the cover. The story had the provocative headline: "What if it's all been a big fat lie?"
Not surprisingly, the article spread faster than peanuts at a bar among members of the dieting public as well as the professional nutrition, weight loss and cardiology communities. Sales of the Atkins diet book jumped from number five to number one on the Times's advice books bestseller list and from number 178 to number five on Amazon.com. Taubes got a $700,000 book deal with publisher Alfred Knopf.
But the author now finds himself in the center of a scientific stew of his own making.
Experts criticized Taubes for suggesting the Atkins diet has merit when it has not been rigorously studied, much less been proven effective or safe. He's been accused of downplaying well- substantiated health hazards of saturated fat and of failing to account for some significant and well-known peer-reviewed research supporting the efficacy of low-fat diets. He's been chastised for failing to acknowledge the documented benefits of diets high in fiber and rich in vegetables and whole grains.
Many members of the professional nutrition community, whose work was directly or indirectly questioned by the Times piece, were incensed. In a letter to the New York Times that has not been published, National Heart, Lung, and Blood Institute Director Claude Lenfant and American Heart Association president Robert Bonow asked the paper "to correct several scientific misperceptions that were created" by the story.
"According to the article, a high carbohydrate diet is the villain," they noted in the letter. "However, the scientific evidence shows that weight gain is caused by an excess of total calories, whether fat or carbohydrate. The key to weight loss, or prevention of weight gain, is a combination of exercise and reduced caloric intake."
With many readers newly confused about the benefits and hazards of low-fat and high-fat eating, we decided to take a hard look at Taubes's arguments and examine the broader record of dietary research that he is accused of ignoring or downplaying. We interviewed more than three dozen experts in the field -- many of them the same people Taubes spoke with -- and reviewed the scientific literature. We also spoke to Taubes himself for several hours and reviewed with him some of the research that he used.
(Along the way, we've tried to be mindful of our own potential biases: The Health section has long reported on many of the same studies and public health campaigns Taubes ignores or devalues. Our Lean Plate Club feature advocates a diverse diet that is neither low- fat nor high-fat, but which embraces principles of healthful eating advocated by many of the groups and other institutions Taubes questions or dismisses. And our newspaper's long-standing, largely amicable rivalry with the New York Times may be viewed by some, we think incorrectly, as coloring our report.)
Despite the uproar, even some of Taubes's sharpest critics found merit in the story -- for example, for describing the role that corn sweeteners in soft drinks may play in the obesity epidemic and for forcing scientists to address unanswered questions on fat. "The good part is that Taubes has stimulated discussion," said James O. Hill, director of the Clinical Nutrition Research Unit at the University of Colorado Health Sciences Center in Denver. "The bad part is that once again the poor public is confused."
And some think that the controversy generated by the story will ultimately benefit the scientific study of nutrition, pumping more research dollars into the effort to answer key questions about various nutrition regimens. "If anything, I think this article has increased the ability to get scientific studies on this topic funded," said Eric C. Westman, associate professor of medicine at Duke University Medical Center and author of one of the two published studies on the Atkins diet.
But our reporting also revealed that some of the key assertions in Taubes's story -- that low-fat diets have been proven
failures at weight loss; that eating saturated fat can be beneficial to the heart; and that a low-fat diet isn't a healthy diet -- are contradicted by a significant amount of high-quality research. We've examined these questions below, highlighting the statements from Taubes's article and following with information and expert testimony that appears to contradict them.
The magazine article embraces the hypothesis that "low-fat weight- loss diets have proved in clinical trials and real life to be dismal failures." Near the end, Taubes says, "I have learned that low-fat diets fail in clinical trials and in real life. . . . "
These statements fail to account for dozens of published scientific papers that show low-fat diets to be successful.
One of the most prominent: The 1998 Clinical Guidelines on the Identification, Evaluation and Treatment of Overweight and Obesity in Adults, conducted by the National Heart, Lung, and Blood Institute and published in the journal Obesity Research.
The study's 12-member expert panel, led by F. Xavier Pi-Sunyer, director of the Obesity Research Center, St. Luke's- Roosevelt Medical Center in New York, identified four dozen published, well- designed, randomized controlled clinical trials on low-fat diets. A review of this literature led the panel to conclude that cutting dietary fat can "help promote weight loss by producing a reduced calorie intake" and that "reducing fat as part of a low calorie diet is a practical way to reduce calories."
Taubes rejects these findings based on what he acknowledges is a negative personal assessment of the panel's chair. "Anything that Pi- Sunyer is involved with, I don't take seriously," Taubes said in an interview during which he mistakenly identified Pi-Sunyer as a psychiatrist. " . . . He just didn't strike me as a scientist. He struck me as a grand old man. . . . Pi-Sunyer was one of the people who in the course of interviewing I decided was not a good scientist."
Pi-Sunyer, an endocrinologist, received his medical degree at Columbia University's College of Physicians and Surgeons, earned a master's of public health from Harvard, is on the faculty at the Institute of Human Nutrition at Columbia, served as president of both the American Diabetes Association and the American Society of Clinical Nutrition and is a current member of the National Academy of Sciences committee that sets recommended food intakes for the nation.
Likewise, Taubes failed to mention conclusions from the 2001 Diabetes Prevention Program (DPP), a $164 million, federally funded study whose results were widely published earlier this year.
DPP found that a diet limiting fats to 25 percent of total calories, combined with exercise, not only helped trim pounds but also prevented type 2 diabetes, an epidemic that is closely associated with the spread of obesity in the United States (and that Taubes suggests at several points in his article is partly the result of the medical establishment's endorsement of low-fat diets). The low-fat dietary approach worked equally well in dropping pounds for men and women, regardless of age or race.
Run by the National Institute of Diabetes and Digestive and Kidney Diseases, the six-year DPP trial included 3,200 adults who were overweight and had impaired glucose tolerance, a condition that usually leads to full-blown diabetes. They were randomly assigned to follow a low-fat diet and exercise; take the drug metformin (which lowers blood sugar levels); or participate in a control group that received the generic medical advice to lose weight and exercise more. The study found that those in the low-fat diet and exercise group not only lost about 15 pounds, on average about 7 percent of their body weight -- they kept it most of it off for four years. They also cut their risk of developing diabetes by 58 percent. The results provide "very compelling evidence that following a low-fat diet in combination with physical activity is very effective in losing weight and in reducing the risk of diabetes," concluded Rena Wing, professor of psychiatry and behavior at Brown University and a co-author of the study, which was published in the New England Journal of Medicine in February.
Again, Taubes disagrees with the mainstream scientific findings. He said he does not believe the study proves low-fat diets lead to weight loss, faulting the researchers for failing to tease out separate effects of low-fat diet and exercise.
The DPP researchers noted when they released the findings last year that they would continue to analyze the data to separate out the effects of diet and exercise, an analysis that is still underway. But many experts say that this calculation is unnecessary, since numerous studies have already proved exercise alone to be ineffective in producing weight loss.
"It's not a mystery why DPP worked," said Danish researcher Arne Astrup, director of the Centre for Advanced Food Studies at the Royal Veterinary and Agricultural University in Frederiksberg and author of two extensive reviews showing the merits of low-fat diets.
Astrup's two studies, which Taubes also dismisses, analyzed results from 20 studies involving nearly 2,000 people who followed low-fat diets. In one report published in 2001 in the International Journal of Obesity and Related Metabolic Disorders, Astrup found that low-fat diets "prevent weight gain in normal weight subjects and produce weight loss in overweight individuals." In the other report, published this year in the Proceedings of the Nutrition Society, Astrup and his colleagues concluded that low-fat diets "consistently demonstrate a highly significant weight loss" of seven to nine
pounds in normal-weight and overweight participants.
The evidence, they said, "strongly supports the low-fat diet as the optimal choice for the prevention of weight gain and obesity," noting that the data for use of higher fat diets, even those rich in healthy types of fat, "is unsubstantiated."
"Astrup is problematic," Taubes said. "I interviewed him." Taubes said he doesn't like the fact that Astrup "chose studies that put people on low-fat, low-calorie diets and compared them to people who ate normally." Simply by enrolling people in a weight loss study, Taubes said, "you turn them into healthier people and intervene in their lives," and that's why he sayshe didn't mention the findings in his article. But Astrup notes that many of the studies he reviewed do in fact control for this "intervention effect," something he said he explained to Taubes during a half-hour telephone conversation. "I reviewed all the evidence suggesting that low-fat diets are the best documented diets and was extremely surprised to see that he didn't use any of that information in his article," Astrup said.
Most recently, a huge statistical snapshot of eating habits in the United States found that people who eat a low-fat diet are more likely to weigh less than those who eat a high-fat diet. Known as the Continuing Survey of Food Intakes by Individuals, the continuing study includes 10,000 adults randomly selected to be representative of the nation and interviewed from 1994 to 1996 as part of a U.S. Department of Agriculture (USDA) effort. Participants were interviewed in their home about what they ate during the past 24 hours.
Using those data, USDA scientists Shanthy A. Bowman and Joseph T. Spence divided participants into four broad categories, from those who ate a very-low-carbohydrate, high-fat diet similar to the Atkins program to people who followed a high-carbohydrate, low-fat approach recommended by the U.S. Dietary Guidelines.
The study found that all four groups ate the same amount of food. But because carbohydrates contain fewer calories than fat, the researchers found, people in the low-fat, high-carbohydrate group consumed on average 200 to 300 fewer calories per day than the others. This helps explain why these people weighed less than the other groups.
More than half the women in the low-fat, high-carbohydrate group had a body mass index (BMI) of less than 25, compared with 45 percent of women in the high-fat, very-low-carbohydrate group, whose approach is similar to the Atkins diet. That same low-fat eating pattern also paid off for men, although not in such a big way: 45 percent of those on a low-fat, high-carb diet had a BMI of less than 25, compared with 35 percent of the men in the high-fat, very-low- carbohydrate group. Results were published in the June issue of the Journal of the American College of Nutrition.
Taubes argues that the data from the study are flawed because they rely so heavily on both epidemiology -- the science of studying large populations rather than individual outcomes that Taubes rejects as unreliable -- and food intake records, which can be inaccurate. "As soon as you start asking people what they ate, you get extraordinary biases," he said. "Epidemiologists don't really have a clue" about this. (Nutrition experts agree that asking someone what they ate over the past couple of days has limitations, but note that the USDA questionnaire has been tested and shown to be more accurate than others. And it is landmark epidemiology studies such as Harvard's Nurses Health Study that have shaped U.S. public health for decades.)
And finally there's the 1998 review of more than two dozen clinical studies of low-fat diets that was also dismissed by Taubes.
"Ample research from animal and clinical studies, from controlled trials and from epidemiological and ecologic analyses provides strong evidence that dietary fat plays a role in the development and treatment of obesity," noted the authors, Louisiana State University (LSU) researchers George Bray and Barry Popkin in the American Journal of Clinical Nutrition. Their review of 28 clinical trials showed that a 10 percent reduction in dietary fat was associated with a weight loss of 16 grams per day -- or roughly a quarter- pound per week, or 13 pounds per year. Taubes faults that paper for the studies that the LSU researchers included and complains that they lowered the bar on success. "You're looking at a weight loss of about five, or six or seven pounds," he said, too little to matter to someone who weighs 300 pounds. Studies suggest, however, that trimming even a few pounds can help improve health, if not appearance.
As for Taubes's suggestion that low-fat diets have also been proven "dismal failures" in "real life," again there is significant evidence contradicting this. The National Weight Control Registry -- an ongoing research project of more than 3,000 people who have lost an average of 60 pounds and kept it off for an average of five years -- also points to the success of the low-fat approach.
To get into the registry, participants must prove their weight loss with pictures, medical records or other documentation. Even so, because it's based on self-reports, the survey is considered less scientifically rigorous than randomly controlled prospective trials. But the data have been considered of high enough quality to merit publication in the peer-reviewed Journal of the American Dietetic Association, American Journal of Clinical Nutrition, International Journal of Obesity Related Metabolic Disorders, Annual Review of Nutrition and the journal Obesity Research.
Led by Wing and the University of Colorado's Hill, the registry shows that those who succeed at losing weight and keeping it off generally follow the low-fat approach (about 25 percent of fat as total calories), limit daily calories and get plenty of exercise. That was true whether participants shed pounds with commercial weight loss programs, liquid diets or
just toughing it out solo.
What the registry hasn't uncovered, however, are many examples of those who have lost weight by the low-carb, high- fat methods Taubes's story implicitly supports. "We can't find more than a handful of people who follow the Atkins program in the registry and, believe me, we've tried," Hill said.
Taubes said he is familiar with the registry findings, but rejects any findings based on them. "I don't believe that [the registry] has any scientific merit at all," he said.
" . . . even saturated fat -- a.k.a. the bad fats -- are not nearly as deleterious as you would think . . . a Porterhouse steak, for instance, will definitely improve your cholesterol profile (at least in comparison with the baked potato next to it). . . . The same is true for lard. If you work out the numbers you come to the surreal conclusion that you can eat lard straight from the can and conceivably reduce your risk of heart disease."
The 2000 Report of the Dietary Guidelines Advisory Committee cites "many years of epidemiological research," including 14 different studies, showing that diets high in saturated fats raise serum cholesterol and thus carry a high risk of coronary heart disease. An analysis of six clinical trials involving 6,356 participants found that decreasing saturated fat cut blood cholesterol levels and reduced coronary heart disease by 24 percent.
Scott Grundy, director of the Center for Human Nutrition at the University of Texas Southwestern Medical Center in Dallas, noted these very points in a letter to the editor of Science magazine in 2001 after Taubes made similar assertions about a porterhouse steak in an article for that publication.
"The significance of saturated fatty acids has been demonstrated by an enormous number of high-quality studies carried out with dietary fat in the fields of animal research, epidemiology, metabolism and clinical trials," Grundy noted. "Although all questions have not been answered, a clear picture of the metabolic and health effects has emerged."
That evidence clearly fingers LDL (low-density lipoprotein) as the leading culprit in heart disease. Lowering LDL -- by diet or drugs such as statins -- also clearly reduces heart disease risk, according to the U.S. Dietary Guidelines report.
Taubes says he bases his conclusion about the ability of lard to decrease heart disease risk on two things: an equation developed by Dutch researcher M. B. Katan of Limburg University in Maastricht and a 1999 interview with Harvard University epidemiologist Walter Willett.
In an interview with The Post, Taubes said that Willett told him, "if you look objectively at exchanging lard for pasta, it would not affect your total cholesterol/HDL (high density lipoprotein)ratio. And your triglycerides will actually lower with lard."
Willett recalls the conversation and the calculation, but doesn't come to the same conclusion about lard as Taubes. As Willett explains, the equation assigns a number to each type of fat in a food. It counts bad fats -- palmitic acid, for example -- as lower than good fats, such as stearic acid, which is found in chocolate and, yes, lard. Then it's possible to assign a theoretical number to a food that suggests what it might do to blood cholesterol levels.
But Willett's interpretation of the results for lard are significantly different from Taubes's. He thinks the numbers might theoretically come out closer to a wash for lard, which is to say any benefits are canceled out by increased risks. But to Willett, that's a moot point. The way to reduce heart disease risk, Willett said in an interview with The Post, "is not [to] eat a lot of lard, but to replace unhealthy fats with healthy fats," such as those found in fish, nuts, olives and avocados.
"That way," Willett said, "you reduce the bad cholesterol, but don't reduce [the protective] HDL cholesterol at the same time and so you actually, really substantially improve the risk of heart disease. And we have seen that in our large epidemiological studies. . . .
"And I have gone over this a number of times with Gary, but he barely mentioned it in the article."
So instead of cooking with lard or other unhealthy fats such as butter, Willett says, "use canola or soybean oil or something like that. I don't think that lard is part of a healthy diet."
Taubes contends in the article, "a low-fat diet is not by definition a healthy diet."
Taubes said he was led to this conclusion by research into a condition called Syndrome X. Also known as the metabolic syndrome, it is characterized by extra abdominal fat, elevated cholesterol and other blood lipids, high blood pressure and insulin resistance, which significantly raises the risk of heart disease and diabetes. Estimates are that 25 to 30 percent of the adult population may have Syndrome X.
There is some credible evidence that people with this syndrome may help reduce their risk of heart disease by adding some healthier fats to their diet. The benefit appears to come because healthy fats -- such as the omega-3 fatty acids
found in fish -- are able to get into heart muscle cells where they have been shown to help reduce the risk of a fatal irregular heartbeat that often accompanies a heart attack.
For this reason, the National Heart, Lung, and Blood Institute suggests in its latest cholesterol guidelines that doctors treating people diagnosed with Syndrome X may consider raising their total fat to as much as 35 percent of total calories, provided that it's done carefully and by adding healthy fats. They also warn about limiting saturated fat to no more than 8 to 10 percent of total calories. Dean Ornish, an advocate of very-low-fat diets, says he has added tiny amounts of fish oil supplements to all participants in his program for the same reason.
But for the rest of the population, the dietary recommendations are to aim for about 30 percent of daily calories to come from fat and to eat as little saturated fat as possible. That's because saturated fat (which raises LDL, the most damaging form of cholesterol) "is not required for human health," as the 2000 Dietary Guidelines committee noted in its report.
In the past, this kind of regimen -- 30 percent total fat, 10 percent or less saturated fat -- has been considered a low-fat diet. And there's been enough evidence of its health benefits in controlling heart disease, obesity and diabetes for expert committees from the U.S. Dietary Guidelines to the researchers and physicians who wrote dietary recommendations for the American Heart Association and the National Cholesterol Guidelines of the National Heart, Lung, and Blood Institute to recommend it.
These days, that same diet is called moderate in fat. "This change is not accompanied by a change in the numerical recommendation (30 percent)," the Dietary Guidelines Committee noted, but rather by a concern "that the previous priority given to a low-fat intake may lead people to believe that as long as fat intake is low, the diet will be entirely healthful."
But whether it is called low fat or moderate fat, the recommendations to eat little or no saturated fat put this dietary approach directly at odds with Atkins' philosophy of eating liberal amounts of such high-saturated fat foods as cream and meat.
"Many years of epidemiological research have shown that populations consuming diets high in saturated fats have relatively high levels of serum cholesterol and carry a high prevalence of coronary heart disease," notes the 2000 Report of Dietary Guidelines Advisory Committee.
Taubes, often critical of epidemiological research, does not agree with that conclusion.
Taubes also rejects findings of reported health benefits from the very-low-fat approaches advocated by the Pritikin Program and by physician Dean Ornish, who runs the nonprofit Preventive Medicine Research Institute in Sausalito, Calif. Results of their findings have been peer-reviewed and published in the first rank of scientific publications, including the New England Journal of Medicine, JAMA and The Lancet. Ornish's work suggests that arterial blockages can be reversed by a very-low-fat diet.
Taubes said he believes that the very-low-fat approach of both Pritikin and Ornish is not only unhealthy, but dangerous. And he calls Ornish "a faith healer, not a scientist." (Ornish earned his medical degree from Baylor College of Medicine and did a residency at Harvard. He was a member of the White House Commission on Complementary and Alternative Medicine Policy, which was charged with advising federal agencies about integrative care.)
The fact that these programs include more than just diet (Ornish's program also includes group support and meditation, for example) makes it difficult, Taubes said to know what each component contributes. "I wish my colleagues in the press would go to better people than Ornish to defend low-fat diets," he said. "Somehow Ornish managed to get published in reasonable [scientific] journals."
Taubes also doesn't think much about a study presented this year at the American Heart Association's epidemiology meeting and published this summer in the peer-reviewed Preventive Cardiology, which is edited by cardiologist Ezra Amsterdam, director of the coronary care unit at the University of California, Davis.
Conducted by physician Richard Fleming, who runs the Fleming Heart & Health Institute in Omaha, the year-long study assigned 100 men and women to four different diets for a year: 28 people followed a moderate fat 30 percent fat approach without restricting calories; 16 people followed a reduced calorie diet with about 15 percent of calories as fat; 38 people ate a reduced calorie diet with about 30 percent fat; and 18 people followed the high fat, low-carb diet similar to Atkins.
The study found that calories count: The only group not to lose weight was the one that didn't limit calories. Among the others, the best weight loss -- about 18 percent of body weight -- occurred in the low-fat group eating a diet of 15 percent fat. Next best: the Atkins-like approach which achieved a 14 percent weight loss. But the 30 percent fat group lost nearly as many pounds as the Atkins group. "So maybe we don't have to go to just 15 percent fat," Fleming said.
As for risk factors, "only patients following the high-fat diets showed a worsening of cardiovascular risk factors," including LDL cholesterol, total cholesterol, triglycerides and lowered levels of the protective form of cholesterol, HDL, Fleming
For Taubes, the data is simply "too perfect." He asks whether Ornish may have had a hand in the paper, even suggesting that Ornish may have funded the study or the Fleming Institute--something that Ornish and Fleming both adamantly deny.
"Frankly, I think Fleming made it [the data] up," Taubes said. (Fleming said that is unequivocally wrong. "I wish the results were perfect," he said when told Taubes' reaction. "Fortunately, they are not...This was the research that we did here." His editor, cardiologist Ezra Amsterdam, said that the paper was peer-reviewed and is a very solid study.)
But that is how Taubes said that data like this make him feel. He said that he "doesn't trust Fleming." "That's my gut feeling," he said. "My 20 years in the business."
And then he catches himself.
"I know," he said. "I sound like if somebody finds something I believe in, then I don't question it." Reproduced with permission of the copyright owner. Further reproduction or distribution is prohibited without permission.