Good Calories, Bad Calories (57 page)

The one that has been most influential is the association between heart disease, obesity, and diabetes. If heart disease is caused by high-fat diets, as is commonly believed, then so are obesity and diabetes, since these diseases appear together in both individuals and populations. But there is no evidence linking obesity to dietary-fat consumption, neither between populations nor in the same populations.*91 And, of course, if dietary fat is not responsible for heart disease, then it’s unlikely that it plays a role in obesity and diabetes.

Second, laboratory rats wil become obese on a high-fat diet. This is the evidence that convinced George Bray that excessive dietary fat would cause obesity in humans, too, and Bray has been among the most influential obesity authorities and the foremost proponent of this dietary-fat/obesity hypothesis. According to Bray, the rats used in his laboratory experiments would grow reliably obese on high-fat diets. “I could feed them any kind of composition of carbohydrates I want,” Bray said, “and in the presence of low fat, they don’t get fat. If I raised the fat content, particularly saturated fat, in susceptible [my italics] strains I would get obesity regularly.”

But some strains of rats, perhaps most of them, wil not grow obese on high-fat diets, and even those that do wil grow fatter on a high-fat, high-carbohydrate diet than a high-fat, low-carbohydrate diet. Moreover, to induce obesity even in susceptible rodents, the percentage of fat in the diet has to be greater than 30 percent, and usual y closer to 40 or even 60 percent (which stil makes only some strains of rats fat). Though 30 percent sounds like a low-fat or moderate-fat diet for humans, it’s far greater than anything rats would normal y consume, either in the wild or in the laboratory. It’s what researchers wil cal a pharmaceutical dosage of fat. Rat chow is typical y 2–6 percent fat calories. Rats wil also fatten when fed large amounts of carbohydrates in the form of sugar. Moreover, other animals fatten on carbohydrates, including pigs—whose digestive apparatus is most like that of humans among experimental animals—cattle, and monkeys.

In the 1970s, Anthony Sclafani of Brooklyn Col ege demonstrated that rats get “super obese” if al owed to freely consume a selection of foods from the local supermarket. This made their eating habits and subsequent obesity seem particularly like ours in character. But, as Sclafani explained, his rats fattened preferential y on sweetened condensed milk, chocolate-chip cookies, and bananas. Among the foods they didn’t eat to excess were cheese, pastrami, and peanut butter—the items that were high in fat and low in carbohydrates.

The third supporting leg of the hypothesis that fat is particularly fattening is an assumption that the density of fat calories fools people into eating too many. Density was original y invoked to explain why some rats would eat fat to excess and become obese. Because the fat used in these experiments is typical y an oil—Crisco cooking oil poured over the rat chow—it was hard to imagine that palatability was the deciding factor. As a result, researchers suggested that the density of the fat calories—nine calories per gram, compared with four for protein and carbohydrates—fooled the rats into consuming too many.

This was in line with the belief that we match our intake to expenditure by simple mechanisms such as those that limit the volume of food consumed in a single meal. It also led to the notion that eating fiber-rich, leafy vegetables wil prevent weight gain by fil ing our stomachs with fewer digestible calories than if we consumed the densely packed calories of fat or refined carbohydrates. The more rigorous experiments with laboratory animals, however, suggest otherwise. The seminal experiments on this question were done by the University of Rochester physiologist Edward Adolph back in the 1940s.

Adolph diluted the diets of his rats with water, fiber, and even clay, and noted that the rats would continue to eat these adulterated diets until they consumed the same amount of calories they had been eating when he had fed them unadulterated rat chow. The more Adolph diluted the chow with water, the more the rats consumed—until the meals were more than 97 percent water. At these very low dilutions, the rats apparently expended so much energy drinking that they couldn’t consume enough calories to balance the expenditure. When Adolph put 90 percent of their daily calories directly into the rats’

stomachs, “other food was practical y refused for the remainder of the twenty-four hour period.” Putting water directly into their stomachs had no such effect. Adolph’s conclusion was that rats adjust their intake in response to caloric content, not volume, mass, or even taste, and this is presumably true of humans as wel .

The fourth piece of evidence is thermodynamic. The idea dates to the late nineteenth century and its revival by the University of Massachusetts nutritionist J. P. Flatt was coincident with the rise of the dietary-fat/heart-disease hypothesis in the 1970s. According to Flatt’s calculations, the “metabolic cost” of storing the calories we consume in adipose tissue—the proportion of energy dissipated in the conversion-and-storage process—is only 7

percent for fat, compared with 28 percent for carbohydrates. For this reason, when carbohydrates are consumed in excess, as the University of Vermont obesity researchers Ethan Sims and El iot Danforth explained in 1987, the considerable calories expended in converting them to fat wil “blunt the effect on weight gain of high-carbohydrate, high-caloric diets.” High-fat diets, on the other hand, would lead “to a metabolical y efficient and uncompensated growth of the fat stores.” Flatt’s analysis omitted al hormonal regulation of fuel utilization and fat metabolism (as wel as a half-century’s worth of physiological and biochemical research that we wil discuss shortly) but it has nonetheless been invoked often during the last twenty years to make the point, as Sims and Danforth did, that obesity is yet another “penalty for living off the fat of the land rather than the carbohydrate.”

Like much of the established wisdom on diet and health, this conclusion was based on very little experimental evidence. In this case, its only supporting evidence came from Sims’s overfeeding studies. These began in the mid-1960s with four smal trials that led to the observation that some people wil gain weight easily and others won’t, even when consuming the same quantity of excess calories. Another half dozen trials fol owed, each with only a handful of subjects, intending to shed light on what Sims and his col aborators cal ed the “obvious question” of whether a carbohydrate-rich diet, independent of the calories consumed, could raise insulin levels, cause obesity, and induce hyperinsulinemia and insulin resistance. Sims and his col aborators varied the composition of the diets that their volunteers would then eat to excess. Some diets were “fixed carbohydrate” regimens, in which the amount of fat was increased as much as possible but the carbohydrates were limited to what the subjects would have normal y consumed in their pre-experiment lives; others were “variable carbohydrate” regimens, in which both fat and carbohydrates were added to excess.

In the mid-1970s shortly after finishing their research, Sims and Danforth believed that obesity was most likely caused by chronical y elevated levels of insulin, and that the elevated levels of insulin were likely the product of carbohydrate-rich diets. In the 1980s, their opinions changed and fel into step with the prevailing consensus on the evils of dietary fat. Sims and Dansforth now found in their decade-old results an observation that supported Flatt’s argument that it was thermodynamical y more efficient to fatten on fat than on carbohydrates. When excess calories were provided in the form of fat alone, they now explained, the subjects converted a greater proportion of the excess into body fat than when the excess calories included both fat and carbohydrates. “Simply stated, when taken in excess, fat is more fattening than carbohydrate,” Danforth wrote in 1985. “Therefore, if one is destined to overeat and desires to suffer the least obesity, overindulgence in carbohydrate rather than fat should be recommended.” “In view of these considerations and the tendency toward overnutrition in most affluent societies,” he added, “main attention should be toward reducing both caloric and fat intake.”

What the Vermont investigators failed to take into account, however, was their own previous observation that the nutrient composition of the diet seemed to affect profoundly the desire to consume calories to excess. One potential y relevant observation that Sims and his col eagues neglected to publish, for example, was that it seemed impossible to fatten up their subjects on high-fat, high-protein diets, in which the food to be eaten in excess was meat. According to Sims’s col aborator Edward Horton, now a professor of medicine at Harvard and director of clinical research at the Joslin Diabetes Center, the volunteers would sit staring at “plates of pork chops a mile high,” and they would refuse to eat enough of this meat to constitute the excess thousand calories a day that the Vermont investigators were asking of them. Danforth later described this regimen as the experimental equivalent of the diet prescribed by Robert Atkins in his 1973 diet book, Dr. Atkins’ Diet Revolution. “The bottom line,” Danforth said, “is that you cannot gain weight on the Atkins diet. It’s just too hard. I chal enge anyone to do an overfeeding study with just meat. You can’t do it. I think it’s a physical impossibility.”

Getting their volunteers to add a thousand calories of fat to their daily diet also proved surprisingly difficult. Throughout their numerous publications, Sims and his col eagues comment on the “difficult assignment of gaining weight by increasing only the fat.” Those fattening upon both carbohydrates and fat, on the other hand, easily added two thousand calories a day to their typical diet. Indeed, subjects in some of his studies, Sims and his col eagues reported, experienced “hunger late in the day…while taking much greater caloric excesses of a mixed diet”—as much as ten thousand calories a day.

Sims and his col aborators evidently did not wonder why anyone would lose appetite—develop “marked anorexia,” as they put it—on a diet that includes eight hundred to a thousand excess fat calories a day, and yet feel “hunger late in the day” on a diet that includes six to seven thousand excess calories of fat and carbohydrates together. It would seem there is something about carbohydrates that al ows the consumption of such enormous quantities of food and yet stil induces hunger as the night approaches.

By perceiving obesity as an eating disorder, a defect of behavior rather than physiology, and by perceiving excessive hunger as the cause of obesity, rather than a symptom that accompanies the drive to gain weight, those investigators concerned with human obesity had managed to dissociate the perception of hunger and satiety from any underlying metabolic conditions. They rarely considered the possibility that hunger, satiety, and level of physical activity might be symptomatic of underlying physiological conditions. Imagine if diabetologists had perceived the ravenous hunger that accompanies uncontrol ed diabetes as a behavioral disorder, to be treated by years of psychotherapy or behavioral modification rather than injections of insulin. These researchers simply never confronted the possibility that the nutrient composition of the diet might have a fundamental effect on eating behavior and energy expenditure, and thus on the long-term regulation of weight.

There is one way to test this latter notion, and, in fact, such tests were done from the 1930s onward. Alter the proportion of fats and carbohydrates in experimental diets and see what happens. Test low-fat diets versus low-carbohydrate diets, keeping in mind that a diet low in fat must be high in carbohydrates and vice versa. This would test the notion that these nutrients have unique metabolic and hormonal effects that influence weight, hunger or satiety, and energy expenditure. Such trials provide the means of answering these fundamental questions: What happens when we eat a diet restricted in carbohydrates, but not calories? Do we lose or gain weight? Are we as hungry as we are when calories are restricted? Do we eat more or less? Do we expend more or less energy? And what about when fat is restricted, but carbohydrates or calories are not? What are the effects on hunger, energy expenditure, and weight?

REDUCING DIETS

Concentrated carbohydrates, such as sugars and breadstuffs, and fats must be restricted. Diets, therefore, should exclude or minimize the use of rice, bread, potato, macaroni, pies, cakes, sweet desserts, free sugar, candy, cream, etc. They should consist of moderate amounts of meat, fish, fowl, eggs, cheese, coarse grains and skimmed milk.

ROBERT MELCHIONNA of Cornel University, describing the reducing diet prescribed at New York Hospital in the early 1950s THE AMERICAN HEART ASSOCIATION TODAY insists that severe carbohydrate restriction in a weight-loss diet constitutes a “fad diet,” to be taken no more seriously than the grapefruit diet or the ice-cream diet. But this isn’t the case. After the publication of Banting’s Letter on Corpulence in 1863, physicians would routinely advise their fat patients to avoid carbohydrates, particularly sweets, starches, and refined carbohydrates, and this practice continued as the standard treatment of obesity and overweight through the better part of the twentieth century. Only after the AHA itself started recommending fat-restricted, carbohydrate-rich diets for heart disease in the 1960s and this low-fat prescription was then applied to obesity as wel , was carbohydrate restriction forced to the margins. “In the instruction of an obese patient,” as Louis Newburgh of the University of Michigan explained in 1942, “it is a simple matter to teach him to omit sugar because sweet flavors are not easily disguised. It is also relatively simple to teach him to limit the use of foods high in starch.”

Those early weight-loss diets were meant to eliminate fat tissue while preserving muscle or lean-tissue mass. The protein content of the diet would be maximized and calories reduced. Only a minimal amount of carbohydrates and added fats—butter and oils—would be al owed in the diet, because these were considered the nonessential, i.e., nonprotein, elements. When physicians from the Stanford University School of Medicine described the diet they prescribed for obesity in 1943, it was effectively identical to the diet prescribed at Harvard Medical School and described in 1948, at Children’s Memorial Hospital in Chicago in 1950, and at Cornel Medical School and New York Hospital in 1952. According to the Chicago clinicians, the “general rules” of a successful reducing diet were as fol ows: