Good Calories, Bad Calories

The laboratory evidence that carbohydrate-rich diets can cause the body to retain water and so raise blood pressure, just as salt consumption is supposed to do, dates back well over a century. It has been attributed first to the German chemist Carl von Voit in 1860. In 1919, Francis Benedict, director of the Nutrition Laboratory of the Carnegie Institute of Washington, described it this way: “With diets predominantly carbohydrate there is a strong tendency for the body to retain water, while with diets predominantly fat there is a distinct tendency for the body to lose water.” The context of Benedict’s discussion was the weight loss that occurs in the first few weeks of any calorie-or carbohydrate-restricted diet, and particularly the latter. As Benedict pointed out, this weight loss is to a large extent water, not fat, which has to be factored into any discussion of the apparent benefits of a reducing scheme. In the late 1950s, a new generation of investigators rediscovered the phenomenon, and it was then used to rationalize the popularity of carbohydrate-restricted diets as due not to the ease of losing fat, but entirely to the water lost in the first few weeks of the diet.

The “remarkable sodium and water retaining effect of concentrated carbohydrate food,” as the University of Wisconsin endocrinologist Edward Gordon called it, was then explained physiologically in the mid-1960s by Walter Bloom, who was studying fasting as an obesity treatment at Atlanta’s Piedmont Hospital, where he was director of research. As Bloom reported in the Archives of Internal Medicine and The American Journal of Clinical Nutrition, the water lost on carbohydrate-restricted diets is caused by a reversal of the sodium retention that takes place routinely when we eat carbohydrates. Eating carbohydrates prompts the kidneys to hold on to salt, rather than excrete it. The body then retains extra water to keep the sodium concentration of the blood constant. So, rather than having water retention caused by taking in more sodium, which is what theoretically happens when we eat more salt, carbohydrates cause us to retain water by inhibiting the excretion of the sodium that is already there. Removing carbohydrates from the diet works, in effect, just like the antihypertensive drugs known as diuretics, which cause the kidneys to excrete sodium, and water along with it.

This water loss leads to a considerable drop in blood pressure, so much so that it led critics of these diets, such as Philip White, author of a nutrition column in the The Journal of the American Medical Association, to worry publicly about the “low blood pressure resulting from…losses of…fluid, sodium, and other minerals.” Discussions of the treatment of obesity with very low-carbohydrate diets would address the need to retain some carbohydrates in the diet to maintain “fluid balance” and “avoid large shifts in weight due to changes in water balance.” By the early 1970s, researchers had demonstrated that the water-retaining effect of carbohydrates was due to the insulin secreted, which in turn induced the kidneys to reabsorb sodium rather than excrete it, and that insulin levels were indeed higher, on average, in hypertensives than in normal individuals. Finally, by the mid-1990s, diabetes textbooks, such as Joslin’s Diabetes Mellitus, contemplated the likelihood that chronically elevated levels of insulin were “the major pathogenetic defect initiating the hypertensive process” in patients with Type 2 diabetes. But such speculations rarely extended to the potential implications for the nondiabetic public.

There are several possible explanations for why this phenomenon rarely entered into the discussions of hypertension and heart disease. Those investigators concerned with the dangers of hypertension might simply have considered the obesity literature or even the diabetes literature of little significance to their research, other than the obvious observation that obese and diabetic patients tend to be hypertensive and vice versa. Another possibility is that by the 1960s hypertension and high cholesterol were two of the three major risk factors associated with premature coronary heart disease (the third was smoking), so it was difficult to imagine that eating carbohydrates might be beneficial for one risk factor, cholesterol, while being detrimental for another, blood pressure.