PromiseSoy: vs. Reality NOVEMBER 2002 VOLUME 29 / NUMBER 9 CENTER FOR SCIENCE IN THE PUBLIC INTEREST $2.50 —page 8—

HEALTH LETTERTM

The Truth About the Atkins Diet

By Bonnie Liebman

“What If It’s All Been a Big Lie?” asked the cover story of the July 7th New York Times Magazine. The arti- cle, by freelance writer Gary Taubes, argues that loading our plates with fatty , , cream, and but- ter is the key not just to , but to a long, healthy life. “Influential researchers are begin- ning to embrace the medical heresy that maybe Dr. Atkins was right,” writes Taubes. Taubes claims that it’s not fatty foods that make us fat and raise our risk of disease. It’s . And to most readers his arguments sound perfectly plausible. Here are the facts—and the fic- tions—in Taubes’s article, which has led to a book contract with a reported $700,000 advance. And here’s what the scientists he quoted —or neglected to quote—have to say about his reporting.

(Continued on page 3) COVER ST

erhaps the most telling state- ment in Gary Taubes’s New York Times Magazine article Pcomes as he explains how diffi- The cult it is to study diet and health. “This then leads to a research lit- Truth erature so vast that it’s possible to find at least some published research to support virtually any theory.” About He got that right. It helps explain why Taubes’s article sounds so credible. the “He knows how to spin a yarn,” says Barbara Rolls, an obesity expert at Pennsylvania State University. “What Atkins frightens me is that he picks and chooses his facts.” Diet She ought to know. Taubes inter- viewed her for some six hours, and she sent him “a huge bundle of papers,” but he didn’t quote a word of it. “If the facts CLAIM #1: The experts recom- ■ According to Taubes, Harvard don’t fit in with his yarn, he ignores mend an Atkins diet. University’s Walter Willett is one of the “small but growing minority of estab- them,” she says. TRUTH: They don’t. Instead, Taubes put together what lishment researchers [who] have come An Atkins diet is loaded with , to take seriously what the low-carb-diet sounds like convincing evidence that , and other foods high in satu- doctors have been saying all along.” carbohydrates cause obesity. rated fat. Taubes implies that many of True, Willett is concerned about the “He took this weird little idea and blew the experts he quotes recommend it. harm that may be caused by high- it up, and people believed him,” says John Here’s what they say: diets (see “What to Eat,” Farquhar, a professor emeritus of ■ “The article was incredibly mislead- page 7). But the Atkins diet? “I cer- medicine at ’s Center ing,” says Gerald Reaven, the pioneer- tainly don’t recommend it,” he says. for Research in Disease Prevention. ing Stanford University researcher, His reasons: heart disease and cancer. Taubes quoted Farquhar, but misrepre- now emeritus, who coined the term “There’s a clear benefit for reducing sented his views. “What a disaster,” says “Syndrome X.” “My quote was cor- cardiovascular risk from replacing Farquhar. rect, but the context suggested that I unhealthy —saturated and trans— Others agree. “It’s silly to say that car- support eating . I was with healthy fats,” explains Willett, bohydrates cause obesity,” says George horrified.” who chairs Harvard’s nutrition depart- Blackburn of Harvard Medical School and ment. “And I told Taubes several times that red meat is associated with a higher the Beth Israel Deaconess Medical Center risk of colon and possibly prostate can- in Boston. “We’re overweight because Gary Taubes cer, but he left that out.” we overeat calories.” tricked us all ■ It’s not clear how Taubes thought he “ “I was greatly offended at how Gary could ignore—or distort—what into coming Taubes tricked us all into coming across as supporters of the Atkins diet,” says researchers told him. “The article was across as Stanford’s John Farquhar. written in bad faith,” says F. Xavier Pi- supporters Taubes’s article ends with a quote Sunyer, director of the Obesity Research of the from Farquhar, asking: “Can we get the Center at St. Luke’s-Roosevelt Hospital Atkins diet. low-fat proponents to apologize?” But Center in New York. “It was irresponsi- that quote was taken out of context. ble.” — John Farquhar “What I was referring to wasn’t that Here’s a point-by-point response to Stanford University” low-fat diets would make a person gain Taubes’s major claims. ››››

NUTRITION ACTION HEALTHLETTER ■ NOVEMBER 2002 3 COVER STORY

weight and become obese,” explains CLAIM #4: We’re fat because we Farquhar. Like Willett and Reaven, he’s ate a low-fat diet. worried that too much carbohydrate My quote was TRUTH: We never ate a low-fat can raise the risk of heart disease. diet. “I meant that in susceptible individu- correct, but “ “At the very moment that the govern- als, a very-low-fat [high-carb] diet can the context raise triglycerides, lower HDL [‘good’] ment started telling Americans to eat suggested cholesterol, and make harmful, small, less fat, we got fatter,” says Taubes. dense LDL,” says Farquhar. that I support “We ate more fat-free carbohydrates, which, in turn, made us hungrier and Carbohydrates are not what has eating satu- made us a nation of butterballs, how- then heavier.” ever. “We’re overfed, over-advertised, rated fat. I was It’s hard to believe this claim passed and under-exercised,” he says. “It’s the horrified. the laugh test at The Times. If you enormous portion sizes and sitting in believe Taubes, it’s not the 670-calorie — Gerald Reaven front of the TV and computer all day” Cinnabons, the 900-calorie slices of Stanford University” Sbarro’s sausage-and-pepperoni-stuffed that are to blame. “It’s so gol’darn obvious—how can anyone ignore it?” pizza, the 1,000-calorie shakes or “The Times editor called and tried to Double Whoppers with , the get me to say that low-fat diets were 1,600-calorie buckets of movie theater the cause of obesity, but I wouldn’t,” CLAIM #3: Health authorities popcorn, or the 3,000-calorie orders of adds Farquhar. recommended a low-fat diet as cheese fries that have padded our back- the key to weight loss. sides. It’s only the low-fat Snackwells, CLAIM #2: Saturated fat (with fat-free sauce), and bagels TRUTH: They didn’t. doesn’t promote heart disease. (with no cream cheese). “It’s preposterous,” says Samuel TRUTH “We’ve been told with almost religious : It does. certainty by everyone from the Surgeon Klein, director of the Center for Human If there’s any advice that experts agree General on down, and we have come Nutrition at the Washington University on, it’s that people should cut back on to believe with almost religious cer- School of Medicine in St. Louis. saturated fat. They’ve looked not just tainty, that obesity is caused by the “There’s no real evidence that low-fat at its effect on cholesterol levels, but on excessive consumption of fat, and that diets have caused the obesity epi- its tendency to promote blood clots, raise if we eat less fat we will lose weight and demic.” levels, and damage blood vessels. live longer,” writes Taubes. Taubes argues that in the late 1970s, They’ve issued that advice after exam- It’s true that some diet books, notably health authorities started telling ining animal studies, population stud- ’s Eat More, Weigh Less, Americans to cut back on fat, and that ies, and clinical studies.1-3 Taubes dis- have encouraged people to eat as much we did. Wrong. misses them with one narrow argument. fat-free food as they want. (Of course, According to the U.S. Department of Saturated fats, he writes, “will elevate Ornish is talking about fruits, vegeta- Agriculture, added fats (oils, shortening, your bad cholesterol, but they will also bles, and whole grains, not fat-free lard, and tallow) have gone up elevate your good cholesterol. In other cakes, cookies, and ice cream.) But steadily since the late 1970s (see words, it’s a virtual wash.” “everyone from the Surgeon General “Hardly a Low-Fat Diet”). Total fats Experts disagree. “Fifty years of on down” is baloney. (which include the fat in meats, cheese, research shows that saturated fat and “The Surgeon General’s report doesn’t and other foods) have also gone up, cholesterol raise LDL [‘bad’] choles- say that fat causes obesity,” says though not as steadily. terol, and the higher your LDL, the Marion Nestle, who was managing edi- So how can Taubes write that “the higher your risk of coronary heart dis- tor of the report and is now chair of the major trends in American diets, accord- ease,” says Farquhar. Yet Taubes has no nutrition and food studies depart-ment ing to USDA agricultural economist qualms about encouraging people to at New York University. “Fat has twice Judith Putnam, have been a decrease in eat foods that raise their LDL. the calories of either or carbo- the percentage of fat calories and a He’s willing to bet that higher HDL hydrate. That’s why fat is fattening ‘greatly increased consumption of car- (“good”) cholesterol will protect them. unless people limit calories from every- bohydrates’”? No experts—at the American Heart thing else.” The key is the word “percentage.” Association; National Heart, Lung, and And health authorities like the The percentage of fat calories in our Blood Institute; or elsewhere—would American Heart Association; National diets declined because, while we ate take that risk. Heart, Lung, and Blood Institute; and more fat calories, we ate even more “The evidence that raising HDL is U.S. Department of Agriculture never carbohydrate calories. protective is less solid than the evi- urged people to cut way back on fat. “We’re eating roughly 500 calories dence that raising LDL is bad,” says Their advice: “Get no more than 30 a day more than we did in 1980,” David Gordon, a researcher at the percent of calories from fat.” At the Putnam told us. “More than a third of National Heart, Lung, and Blood time that advice was issued, the average the increase comes from refined grains, Institute. person was eating 35 percent fat. a fifth comes from added , and a third comes from added fats.”

4 NUTRITION ACTION HEALTHLETTER ■ NOVEMBER 2002 COVER STORY

Government surveys show no The result is Hardly a Low-Fat Diet change—or a slight decrease—in fat hunger and a Added Fats & Oils in the Food Supply consumption since the late 1970s. But craving for more they don’t look at how much fat is pro- carbohydrates.” duced, how much is sold, and how It sounds con- much is wasted. The surveys simply vincing, but ask consumers what they eat. And it’s there’s a problem: possible that once people were told to “It’s not proven eat less fat, they (consciously or uncon- at all,” says Penn sciously) started under-reporting how State’s Barbara much they ate. Rolls. “We have Says Putnam: “People don’t ade- no firm data that quately report added fats, added sugars, and refined grains.” affects body The bottom line: Taubes blames the weight or how obesity epidemic on a low-fat diet that full people feel the nation never ate. after eating.” Harvard’s CLAIM #5: Carbs, not fats, cause David Ludwig has obesity. done a few stud- TRUTH: The evidence blaming ies on glycemic obesity on carbs is flimsy. index and weight. In the largest, he The evidence that carbohydrates make found that 64 According to Taubes, a low-fat diet has made us fat. Yet our con- you fat can be called “Endocrinology overweight ado- sumption of all added fats combined (red line) is higher than ever 101,” says Taubes, implying that it’s lescents who were before. Estimates of total fat (not shown), which includes the fats well-established fact. In a nutshell, in meats, dairy, etc., also show a rise since the late 1970s. The bot- told to eat lower- tom line: Americans never went on a low-fat diet. Endocrinology 101 says that “we’re hun- glycemic-index grier than we were in the ‘70s” because foods lost an Source: USDA/Economic Research Service. we’re eating more carbohydrates. average of four “ and starches like potatoes and pounds, while 43 , or anything made from flour, like overweight adolescents who were told there is no good evidence that insulin a slice of white ,” are “known in to make modest cuts in calories and fat triggers weight gain. “Insulin crosses the the jargon as high-glycemic-index car- gained three pounds.4 blood-brain barrier and turns off food bohydrates, which means they are “It’s hard to tease apart what led to intake,” says Pi-Sunyer. “That makes absorbed quickly into the blood,” the weight loss in that study,” explains sense. You’ve just eaten, so you don’t explains Taubes. Rolls, “because calorie density, fiber, need to eat for a while. If anything, “As a result they cause a spike of and glycemic index all go hand in insulin should lower food intake.” blood sugar and a surge of insulin hand.” within minutes. The resulting rush of In other words, foods with a low CLAIM #6: The Atkins diet is insulin stores the blood sugar away glycemic index—most vegetables, the best way to lose weight. and a few hours later, your blood sugar fruits, and whole grains—are also high TRUTH is lower than it was before you ate.... : We don’t know the best in fiber and low in calorie density. way to lose weight. What’s more, Ludwig’s study didn’t randomly assign children to one diet or “Until we have more research, no one another, so the two groups weren’t has the solution to the safest and most It’s preposterous. comparable. “The low-glycemic-index effective weight loss,” says Washington University’s Samuel Klein. There’s no real group had fewer minorities,” says “ Columbia’s Pi-Sunyer. Whites in both “Preliminary data from several stud- evidence that groups were more likely to lose weight. ies suggest that, at least over the short- term, the Atkins diet is superior to a low-fat diets And he and others question the whole glycemic index theory.5 Among low-fat diet in a free-living environ- have caused his criticisms: “People eat meals, where ment,” he says. “But it’s too early to the obesity low-glycemic foods balance out high- say that the Atkins diet is better.” Even if ongoing studies show that epidemic. glycemic foods.” For example, “people don’t eat pasta the Atkins diet promotes weight loss, —Samuel Klein alone,” he explains. “They eat it with we won’t know if other diets—ones Washington” University olive oil, clams, tomatoes, or other high in unsaturated fat or protein or School of Medicine foods, and that dampens the differ- vegetables and whole grains, for exam- ences in their effects on insulin.” ple—would work as well or better. And, contrary to Taubes’s claims, ››››

NUTRITION ACTION HEALTHLETTER ■ NOVEMBER 2002 5 COVER STORY

“We need lots more randomized the other if it can hurt your bones?” referring to. In 1998, the National controlled trials to evaluate the differ- The problem: All the protein that Heart, Lung, and Blood Institute issued ent permutations,” says Walter Willett. Atkins recommends leads to acidic guidelines to help doctors treat obesity.7 (He and Blackburn are embarking on a urine.6 “And there’s no dispute that an Its conclusion: People who are told study testing a high-unsaturated-fat acid urine leaches calcium out of to cut fat (but not calories) lose some Mediterranean diet, not the high- bones,” says Blackburn. weight because they inadvertently eat saturated-fat Atkins diet, as Taubes “You can buffer the diet by taking a fewer calories. But people who cut fat implies.) couple of Tums a day, but now we’re and watch calories lose more. “What’s important is not theories, into medical supervision of people on “A low-fat diet helps people eat fewer but evidence.” the diet,” he adds. calories,” says Rena Wing, a professor Blackburn and others also want to of psychiatry and human behavior at CLAIM #7: The Atkins diet know whether an Atkins diet makes the the Brown University Medical School works because it cuts carbohy- blood vessels less elastic. “Studies sug- in Providence, Rhode Island. “Maybe drates. gest that a diet high in animal fats may people want to hear that if they eat a TRUTH: If the Atkins diet cause blood vessels to constrict,” he lower-fat diet they don’t have to eat works, it’s not clear why. says. “That’s a root cause of atheroscle- fewer calories, but that’s not true.” rosis.” What about Taubes’s claim that low- If the Atkins diet does work, it may In preliminary studies, the LDL fat diets are a failure “in real life”? have nothing to do with the glycemic (“bad”) cholesterol of people on the Wing’s National Weight Loss index or Atkins’s promises. “It’s Atkins diet didn’t go up. That’s com- Registry keeps track of people—so far, unlikely to be related to the explana- forting. (Of course, LDL didn’t go down about 3,000—who report having lost at tion in Atkins’s book,” says Klein, either, as it usually does with weight least 30 pounds and having kept the “because that doesn’t make physio- loss.) weight off for at least six years.8 The logical sense.” “The harm caused by saturated fat registry can’t “prove” which diet is best Other possibilities: In one study, the could be overcome by weight loss,” because it’s not a controlled experi- people on a low-carb diet were told to Klein explains. But what happens once ment. But it does offer evidence of follow Dr. Atkins’ New Diet Revolution, people stop losing weight and start try- what works in the long run. which could have been more persua- ing to maintain the loss? Will their “People on low-carbohydrate diets sive than what the people on a lower- LDL climb? “We don’t know.” like Atkins’s are very rare in the reg- fat diet got—a manual designed by istry,” says Wing. academics. “The people in our registry consis- Or, says Klein, “it may simply be eas- tently report eating around 24 percent ier to cut carbs.” Everyone knows what of calories from fat,” she adds. They also they are: bread, pasta, rice, potatoes, It’s silly to say expend roughly 2,800 calories a week— sweets, etc. that carbohy- that’s like walking four miles a day. Or, the monotony of a low-carb diet Furthermore, a low-fat diet aided could have curbed the dieters’ appetites. “ drates cause weight loss in a six-year study of 3,200 “You lose a lot of foods when you cut obesity. We’re people called the Diabetes Prevention out carbs,” says Klein. And with less Program.9 overweight variety, says Blackburn, “people eat “Patients were put on a low-fat diet less, so they lose more weight.” because we with about 25 percent of calories from “It’s also possible that a chemical is overeat calories. fat and they participated in 150 minutes released by a high-fat diet that sup- of physical activity a week,” says Wing. presses the appetite,” adds Klein. “We —George Blackburn “They lost about seven percent of just don’t know.” ” their body weight and kept most of it off for four years. And they reduced their CLAIM #8: The Atkins diet is safe. risk of diabetes by 58 percent.” TRUTH: It isn’t. Of course, it was both diet and exer- Taubes not only neglects to mention cise that led to their success. But if a low-fat diet promotes weight gain, as that the meat in an Atkins diet may CLAIM 9 promote cancer. He ignores some # : Low-fat diets don’t Taubes argues, the exercise—only researchers’ concerns about other help people lose weight. about 20 minutes a day—would have adverse effects. TRUTH: Low-fat diets work if had to not only counter the fattening “The Atkins diet may produce more dieters cut calories. effects of the low-fat diet, but actually lead to weight loss. Unlikely. weight loss in the first three weeks, but “Low-fat weight-loss diets have proved it’s not spectacular,” says Harvard’s in clinical trials and real life to be dis- George Blackburn. “Who cares if one mal failures,” writes Taubes. group loses a few more pounds than It’s not clear which clinical trials he’s

6 NUTRITION ACTION HEALTHLETTER ■ NOVEMBER 2002 What to Eat I told Taubes Judging by Magazine article, you’d think that experts were several times “ in a tug-of-war over whether to endorse low-fat or low-carbohydrate diets. that red meat Not so. Here’s what they agree on...and where they differ: is associated with a higher Cut saturated (and trans) fat. Forget Atkins. Experts agree that people should cut back on saturated (and trans) fat. That includes burgers, french fries, risk of colon and pizza, ice cream, and sweets made with butter, shortening, or stick margarine. possibly prostate “There’s a clear benefit from replacing unhealthy fats with healthy fats,” says cancer, but he 1Harvard’s Walter Willett. “The fat in poultry, fish, and nuts is much better than left that out. the fat in red meat and dairy.” Healthy fats also include salad dressings, mayon- naise, cooking oils, and fish oils. —Walter Willett Harvard University” But the sky’s not the limit, as Atkins would argue. “We’re not working in the fields and burning calories all day, so we need to pay attention to all forms of calo- ries,” says Willett. “You can’t eat unlimited quantities of fats or you’ll gain weight.” CLAIM #10: Taubes examined Don’t overdo carbohydrates. A high-carb diet can cause trouble for the the evidence objectively. estimated 25 percent of Americans who have the Metabolic Syndrome, also called TRUTH: He let his biases rule. Syndrome X or insulin resistance (see “Read My Lipids,” October 2001). The New York Times Magazine isn’t the “Too much carbohydrate will raise triglycerides, lower HDL cholesterol, and National Enquirer. Readers expect The 2make LDL small and dense, all of which raises the risk of heart disease,” says Times to run articles that are honestly Stanford University’s John Farquhar. reported and written. Yet in August, That doesn’t happen to everyone. Syndrome X doesn’t show up in people The Washington Post revealed that who are not genetically susceptible, or in people who get too few calories or too Taubes simply ignored research that didn’t agree with his conclusions. much exercise to be overweight. (In China, Japan, and other Asian nations, diets For example, The Post asked Taubes are high in carbohydrates, yet heart disease rates are rock–bottom low.) But why he made no mention of a review many Americans are genetically susceptible, pudgy, couch potatoes. of nearly 50 studies on weight loss in “A high-carb diet is worse for overweight, underexercised people and for peo- the National Heart, Lung, and Blood ple from racial groups—Latino, Asian, Indian—in whom a higher proportion have Institute’s 1998 Clinical Guidelines on a genetic disposition to Type 2 diabetes,” Farquhar explains. treating obesity. The panel of experts But will a high-carbohydrate diet make you fat? Most researchers say no. was chaired by ’s Pi-Sunyer, who has served as president Even people who get higher insulin levels on a high-carb diet don’t gain weight. of both the American Society of “If anything, more studies show that insulin resistance protects against weight Clinical Nutrition and the American gain,” says Stanford’s Gerald Reaven. Diabetes Association. Willett isn’t sure. “It may be easier to control weight if you cut back on refined “Anything that Pi-Sunyer is involved starches, sugars, and potatoes,” he says. His new study is testing that theory. with, I don’t take seriously,” said In any case, it would be foolish to assume that the calories in fat-free carbohy- Taubes. “He just didn’t strike me as a scientist.” drates will bounce off your body like Teflon. And it’s clear that some carbs—like If Taubes had written a news article vegetables, fruits, and whole grains—are healthier than refined carbs like white for the front page of The Times, com- bread, soft drinks, and sweets. ments like those would have ended his “The type of carbohydrate matters,” says Willett, “just as the type of fat matters.” career. But when it comes to reporting Look for a weight-loss strategy that works for you. Until more about diet, the bar is set lower. Surely, the public deserves better. studies are done, it’s too early to say which diet makes it easiest to lose weight. Some people may find it easier to cut back on bread, pasta, rice, potatoes, and 1 Circulation 102: 2284, 2000. sweets, while others find it easier to cut back on fried foods, oils, salad dressings, 2 Circulation 103: 1034, 2001. mayonnaise, and margarine. Just make sure that you cut calories, and that the fats 3 3 www.nhlbi.nih.gov/guidelines/ and carbs you do eat are healthy. “Most everyone agrees that we need to eat cholesterol/index.htm. 4 Arch. Pediatr. Adolesc. Med. 154: 947, 2000. more fruits and vegetables, that our grains should be whole rather than refined, that 5 Am. J. Clin. Nutr. 76(suppl): 2908, 2002. our protein foods should be lean, and that our oils should come from plants or 6 Am. J. Kidney Dis. 40: 265, 2002. fish,” says Penn State’s Barbara Rolls. 7 www.nhlbi.nih.gov/guidelines/ “To say that experts don’t know what people should eat is deliberately mis- obesity/ob_gdlns.htm. 8 Amer. J. Clin. Nutr. 66: 239, 1997. leading.” 9 New Eng. J. Med. 346: 393, 2002.

NUTRITION ACTION HEALTHLETTER ■ NOVEMBER 2002 7