on two papers, a decade apart, discussing the experience of seventeen subjects over two weeks of dieting.
Yudkin was the most prominent advocate of carbohydrate-restricted diets among nutritionists through the 1970s. He also had unconditional faith, however, in the popular misinterpretation of the law of conservation of energy. “The irrefutable, unarguable fact is that overweight comes from taking in more calories than you need,” Yudkin explained in a 1958 diet book entitled This Slimming Business. He reconciled this belief with his advocacy of carbohydrate-restricted diets because he also believed that “much of the extra fat today” in the diet “comes together with carbohydrate in cakes, biscuits, ice cream, and sweetmeats of various sorts.” If we remove the carbohydrates, Yudkin proposed, the fat calories will come down, too.
In 1960, Yudkin provided experimental evidence to support this statement in a Lancet article entitled “The Treatment of Obesity by the ‘High-Fat’ Diet.” He had asked four women and two men to consume a carbohydrate-restricted diet for two weeks. They all lost weight, he reported, by consuming significantly less carbohydrates and no more fat than they typically ate on a balanced diet. The two men ate roughly twenty-nine hundred and thirty-five hundred calories normally, but reported consuming only fifteen to sixteen hundred when they abstained from carbohydrates. Their fat consumption dropped by two hundred calories a day as well. This led Yudkin to the “unequivocal” conclusion that “the high-fat diet leads to weight-loss because, in spite of its unrestricted allowance of fat and protein, it is in fact a low-calorie diet….” Weight is lost by restricting calories, even if calorie-restriction is not required by the diet.
Here again, however, Yudkin was confusing an association with cause and effect. Even if Yudkin’s subjects had reduced their calorie consumption on the carbohydrate-restricted diet, which is a common finding in these studies, it does not mean that the reduction in calories caused the weight loss, only that the diet was associated with a reduction in calories as well as a reduction in weight. The diet could have worked by some other mechanism entirely, but both weight loss and decreased appetite were consequences. The fact that a reduction of appetite associates with weight loss does not mean that it is the fundamental cause.
And, of course, what may have been true, on average, for Yudkin’s seventeen subjects—six in his 1960 study and eleven a decade later—is not necessarily the case for everyone who loses weight on such diets.*105 Even before Yudkin published This Slimming Business, Weldon Walker and the Columbia University physician Sidney Werner had both reported that their subjects lost significant weight while consuming at least twenty-seven hundred and twenty-eight hundred calories a day respectively. In 1954, when the Swiss clinician B. Rilliet discussed his experiences using Pennington’s diet to treat obese patients at the County Hospital of Geneva, he reported that his successes were “numerous and encouraging” with both a twenty-two-hundred-calorie version of the diet and a three-thousand-calorie version. It’s hard to avoid the observation that at least some individuals lose weight on carbohydrate-restricted diets while eating considerably more calories than would normally be consumed in a semi-starvation diet. This is why Werner speculated that his obese subjects must have typically been eating four to five thousand calories a day before he set about experimentally reducing them. But if that is true, why don’t obese patients regularly lose weight on twenty-seven or twenty-eight-hundred calorie balanced diets, and why have clinicians always believed it necessary to semi-starve them with twelve to fifteen hundred calories, or even feed them very low-calorie diets of eight hundred calories or less, to achieve any significant weight loss? Something else is going on here, and it has nothing to do with calories.
The argument that carbohydrate-restricted diets work by the same mechanism as calorie-restricted diets only changes the nature of the dilemma we have to unravel. It does not make it disappear. Even if we accept Yudkin’s notion that all people who lose weight while abstaining from carbohydrates do so because they spontaneously feel the compulsion to eat less, we must then explain why anyone would willingly suffer the symptoms of semi- starvation—hunger, irritability, depression, and lethargy—rather than simply eat another piece of cheese, or steak, or lamb. The standard explanations are that it’s simply too much trouble to do so, or that “all-you-can-eat-diet[s],” as Jane Brody wrote in the New York Times in 1981, “so restrict the dieter’s choices that boredom and distaste automatically produce a calorie cutback.” But these are unacceptably facile. If the obese end up eating less on this diet, the most likely explanation is that they’re less hungry, in the same way that, if we don’t drink water when water is there for the drinking, we’re probably not thirsty. If we don’t feel or act semi-starved, it’s a reasonable bet that we’re not. “The best definition of food deficiency,” as Ancel Keys and his colleagues wrote in The Biology of Human Starvation, “is to be found in the consequence of it.”
Keys’s starvation studies suggest where the “no bread, no butter” logic will take us. We know from these studies that if we feed people a carbohydrate-rich diet of fifteen or sixteen hundred calories a day, they will be obsessed with the “persistent clamor of hunger,” so much so that they might be willing to mutilate themselves to escape the ordeal. Meanwhile, if those same people were allowed to consume unlimited calories of only meat, cheese, and eggs, this school of thought dictates, they will voluntarily restrict their consumption to the same fifteen or sixteen hundred calories—or at least they will if they’re obese or need to lose ten or twenty pounds—because in this case, as Harvard endocrinologist George Cahill suggested, the “nonappetizing nature” of this meat-egg-and- cheese diet will overcome the urge to amply satisfy their desire for food. Our subjects will voluntarily starve themselves, as though hunger itself, and all its regrettable side effects, have been rendered impotent in the face of monotony, which is to say, a diet that these experts define as unappetizing because it does not allow consumption of starches, flour, sugar, or beer.
But Keys had also severely restricted the choice of foods he fed his subjects. Remember, he had wanted to simulate the foods available during wartime in Eastern Europe and so had allowed his conscientious objectors only bread, potatoes, cereals, turnips, cabbages, and “token” amounts of meat and dairy products. Yet, in the entire fourteen hundred pages of his Biology of Human Starvation, there is not the slightest hint that his semi-starved subjects, or those starving populations he discusses in his comprehensive history of famine, would have turned down more cabbages, bread, or turnips had they been available, not to mention meat, cheese, fish, or eggs. The notion that hunger can be relieved or eliminated simply by limiting the choice of food is exceedingly difficult to embrace.
Over the years, a common way to avoid thinking about the paradox of a diet that allegedly restricts calories but does not induce hunger is to attribute the suppression of appetite to a factor that these authorities consider irrelevant to the bigger picture of weight and health—to ketosis, the condition produced when the liver increases its production of ketone bodies to replace glucose as a fuel for the brain and nervous system. Once ketone bodies are produced, “their appetite-depressing activity takes effect,” as Richard Spark of Harvard Medical School claimed in 1973. “Substances called ketones will accumulate in your bloodstream [during carbohydrate restriction] and can make you slightly nauseated and light-headed and cause bad breath,” wrote Jane Brody in the New York Times in 1996. “This state is not exactly conducive to a hearty appetite, so chances are you will eat less than you might otherwise have of the high-protein, high-fat foods permitted on the diet.”
But this, too, fails as a viable explanation. The liver increases ketone-body synthesis only when carbohydrates are unavailable and the body is relying predominantly on stored fat for its fuel. Ketone bodies could be responsible for appetite suppression, as Spark and Brody suggested, but so could the absence of carbohydrates or the burning of fat, or something else entirely. All of these are associated with the absence of hunger. In fact, the existing research argues against the claim that ketone bodies suppress appetite. Individuals with uncontrolled diabetes, for example, will suffer from ketoacidosis, during which ketone-body levels can be tenfold or even forty-fold higher than the mild ketosis of carbohydrate restriction, and yet these people are ravenous. “It is not clear why the sensation of hunger subsides [in starvation studies], but the disappearance is apparently not related to ketosis,” wrote Ernst Drenick in 1964 about his fasting studies at UCLA. Hunger sensations often disappeared in his subjects before ketone bodies could be detected in their blood or urine, “and it did not reappear” in those periods when ketone body levels were low. The same dissociation between ketone bodies and hunger was reported
