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 well, 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 allowed in the diet, because these were considered the nonessential, i.e., non protein, 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 Cornell Medical School and New York Hospital in 1952. According to the Chicago clinicians, the “general rules” of a successful reducing diet were as follows:
1.Do not use sugar, honey, syrup, jam, jelly or candy.
2.Do not use fruits canned with sugar.
3. Do not use cake, cookies, pie, puddings, ice cream or ices.
4. Do not use foods which have cornstarch or flour added such as gravy or cream sauce.
5. Do not use potatoes (sweet or Irish), macaroni, spaghetti, noodles, dried beans or peas.
6. Do not use fried foods prepared with butter, lard, oil or butter substitutes.
7. Do not use drinks such as Coca-Cola, ginger ale, pop or root beer.
8. Do not use any foods not allowed on the diet and [for other foods use] only as much as the diet allows.
With the carbohydrates and added fats minimized in these diets, meat was inevitably the primary constituent. This would provide the protein necessary to ensure that weight loss came mostly from the patient’s fat and not the muscle. The idea was to keep the body in what is called nitrogen equilibrium, with the nitrogen consumed from the protein in the diet balancing out the nitrogen being excreted in the urine from the breakdown of muscle protein.
When these clinicians discussed what plant foods they would allow in their diets, they typically did so on the basis of the carbohydrate content: potatoes are nearly 20 percent carbohydrate by weight (the rest is mostly water), so they were known as 20-percent vegetables. Green peas and artichokes are 15-percent vegetables. Onions, carrots, beets, and okra are 10-percent vegetables. Most of the green vegetables—including lettuce, cucumbers, spinach, asparagus, broccoli, and kale—are 5-percent, which means carbohydrates constitute at most 5 percent of their weight. These weight-loss diets allowed only 5-percent vegetables, which ruled out all starchy vegetables, like potatoes. Because a one-cup serving of a 5-percent vegetable will yield only twenty to thirty calories, as the University of Toronto physician Walter Campbell wrote in 1936, “the inclusion of an extra portion or omission of an undesired portion is of little moment in the [dietary] scheme as a whole.” Some of these diets did allow an ounce or two of bread—usually whole-grain, because white bread had too few vitamins to make it worth including. But most did not. “All forms of bread contain a large proportion of carbohydrate, varying from 45–65 percent,” noted H. Gardiner-Hill of London’s St Thomas’s Hospital Medical School in 1925, “and the percentage in toast may be as high as 60. It should thus be condemned.”
When these physicians talked about lean meat as the basis of a weight-reducing diet, they did not mean a chicken breast without the skin, as has been the iconic example for the past twenty years. They meant any meat, fish, or poultry (bacon, salt pork, sausage, and duck occasionally excepted) in which the visible fat had been trimmed away.
Once weight was satisfactorily lost, weight-maintenance diets were also restricted in carbohydrates, although not so drastically. For maintaining a reduced weight, as described by the Pittsburgh physician Frank Evans in the 1947 edition of the textbook Diseases of Metabolism, the daily diet should include at least one egg, a glass of skimmed milk, a portion of raw fruit, “a generous portion of any cut of lean fresh lamb, beef, poultry or fish,” and a portion of each of three 5-percent vegetables. Individuals trying to maintain their weight loss could then eat anything else they wanted, Evans wrote, but they could do so only as long as they maintained a stable weight and were sufficiently “sparing with” alcohol, added fats and oils, “concentrated carbohydrate foods,” “starches,” “mealy vegetables, which are potatoes, beans, peas,” and “cereals, used as vegetables, which are: macaroni, spaghetti, rice, corn.”
Evans provided one of the few variations on this regime that caught on as an obesity therapy in the years before World War II. This was a very low-calorie diet, of 360 to 600 calories a day, rather than the common prescription of 1,200 to 1,500 calories, then considered the minimal amount that a patient would tolerate and that would produce a safe and consistent weight loss. Evan’s diet could induce a loss of up to five pounds a week, rather than the two pounds predicted for the more typical semi-starvation diets. The daily menu, explained Evans in 1929, was “composed of fresh meat and egg white. Approximately 100 [grams] of lean steak was the backbone of each of the two largest meals. When necessary, fresh fish was given at intervals.” No starches or sugars were allowed, but the patient could eat a few ounces of 5-percent vegetables and one ounce of bread each day. These minimal carbohydrates—perhaps twenty grams—were included to “spare” the protein in the diet, so that it would be utilized for balancing out nitrogen losses rather than having some of it converted to glucose to fuel the brain and central nervous system. “The secret of the success of this procedure depends, almost certainly, on giving enough protein,” Russell Wilder of the Mayo Clinic wrote after first prescribing the diet for his patients in 1931. Evans’s very low-calorie diet may also have been popular because it appealed to the puritanical sense of those clinicians like Louis Newburgh, who believed that gluttony had to be vigorously curbed in obese patients. One of the fundamental rules of Evans’s diet was: “No concession to gustatory sensualism is permitted.”
In the century before the medical community began prescribing fat-restricted, carbohydrate-rich diets for weight loss, one point of controversy was whether carbohydrates should be avoided because they are uniquely fattening or perhaps even cause obesity—as Jean Anthelme Brillat-Savarin and William Banting would have suggested—in which case they would be the only nutrient restricted, or because they constitute superfluous calories, in which case dietary fat was restricted as well, by avoiding oils, lard, and butter. “The next question to decide,” wrote the Chicago physician Alfred Croftan in the Journal of the American Medical Association in 1906, “is whether the carbohydrates or the fats are to be chiefly restricted.”
One observation made repeatedly through the 1960s was that the obese favor carbohydrates, and that these constitute the great proportion of all calories they consume. Though the obese did not appear to eat more calories, on average, than the lean, they did consume more carbohydrates. Such a dietary assessment was inevitably difficult to make with any accuracy, explained Sir Derrick Dunlop of the Royal Infirmary in Edinburgh, when he reported in 1931 on the lessons he had learned from treating 523 obese patients. Nonetheless, Dunlop believed that “obesity does occur in persons without showing any direct relationship to food intake, and that a certain group of patients do become overweight on an apparently normal well-balanced diet,” and, second, “that an outstanding dietetic abnormality was an excessive intake of carbohydrate.” “In some extreme cases,” he noted, “the diet had consisted almost exclusively of sweet tea, white bread and scones.”
This observation was echoed in The Lancet in 1935 by the British physician John Anderson, and in the 1940s by Hilde Bruch, Hugo Rony, and the Harvard physician Robert Williams and his colleagues, all of whom had questioned their fat patients extensively about their diets. Their common finding was an excessive consumption of starches and sweets. Rony reported that the craving for sweets and starches among his patients was so common that it suggested an underlying physiological mechanism at work, possibly related to a greater need for or reduced availability of glucose. “It is easier to induce the gluttonous obese to control his general appetite than to control his craving for sweets,” Rony noted. One common rationale for restricting carbohydrates in weight-reducing diets was that it eliminated a disproportionately large share of the calories that the obese would normally eat.