This scenario is the most difficult to sort out clinically, because when these investigators measure insulin resistance in humans they invariably do so on a whole-body level, which is all the existing technology allows. Any disparities between the responsiveness of fat and muscle tissue to insulin cannot be measured. This is critical, because for the past thirty-five years the American Diabetes Association has recommended that diabetics eat a diet relatively rich in carbohydrates based on the notion that this makes them more sensitive to insulin, at least temporarily, so the diet appears to ameliorate the diabetes. This effect was initially reported in 1971, by the University of Washington endocrinologists Edwin Bierman and John Brunzell,*119 who then waged a lengthy and successful campaign to persuade the American Diabetes Association to recommend that diabetics eat more carbohydrates rather than less. If Neel’s third scenario is correct, however, a likely explanation for why carbohydrate-rich diets appear to facilitate blood-sugar control after meals is that they increase the insulin sensitivity of the fat cells specifically, while the muscle tissue remains insulin-resistant.

One of the few attempts, if not the only one, to measure the insulin sensitivity of fat cells and muscle cells separately in human subjects was made by the University of Vermont investigator Ethan Sims, in his experimental obesity studies of the late 1960s. Sims and his colleagues surgically removed fat samples from their subjects before, during, and after the periods of forced overeating and weight gain. They reported that high-carbohydrate diets had the unique ability to increase the insulin sensitivity of fat cells, and particularly so in fat cells that were already large and overstuffed. They had no similar effect, however, on the insulin resistance of the muscle tissue.

If this observation is correct, it means carbohydrates are uniquely capable of prolonging this lipid-trapping condition by keeping the fat cells sensitive to insulin when they might otherwise become insulin-resistant. This might lower blood-sugar levels temporarily and delay or improve the appearance of diabetes—or “mask” the diabetes, as von Noorden put it—but would do so at the cost of increasing fat accumulation and obesity. Sims’s observation suggests that Neel’s third scenario for the genesis of obesity and diabetes was astute, and it suggests that a carbohydrate-rich diet might temporarily improve the symptoms of diabetes only by furthering the fattening process. Sims’s studies have not been repeated in humans, but they have been reproduced and confirmed in animals. Brunzell says he refuses to believe that Sims got this measurement correct, but he also says that he has never tried to do the measurements himself because they’re too difficult. But the question of whether Sims got it right requires a definitive answer. Without one, there’s no way to know if the ADA recommendations have been helping diabetics or hurting them, let alone to understand the pathology of obesity and diabetes. The impact on the public health could be immense.