Good Calories, Bad Calories

Sugar and candies do not cause diabetes, but contribute to the burden on the pancreas and so should be used sparingly…. Carbohydrates are best taken in starchy forms: fruits, vegetables and cereals. The absorption is slower and the functional strain minimal.

GARFIELD DUNCAN, Diabetes Mellitus and Obesity, 1935

OF ALL THE DISEASES OF CIVILIZATION that may have been linked to the consumption of sugar and the refining of carbohydrates, diabetes was certainly a prime suspect. Here is a disease in which a conspicuous manifestation is the body’s inability to use for fuel the carbohydrates in the circulation—known as blood sugar or, more technically, glucose or serum glucose. This glucose accumulates in the bloodstream, effectively overflows the kidneys, and spills over into the urine, causing a condition referred to as glycosuria. One symptom is a constant hunger, specifically for sugar and other easily digestible carbohydrates. Another is frequent urination, and the urine not only smells like sugar but tastes like it. For this reason, diabetes was often known as the sugar sickness. Hindu physicians two thousand years ago suggested it was a disease of the rich, caused by indulgence in sugar, which had only recently arrived from New Guinea, as had flour and rice.

“This ancient belief has a point in its favor,” noted the American diabetologist Frederick Allen in his 1913 textbook Studies Concerning Glycosuria and Diabetes. “It originated before the time of organic chemistry, and there was no way for its authors to know that flour and rice are largely carbohydrate, and that carbohydrate in digestion is converted into the sugar which appears in the urine. This definite incrimination of the principal carbohydrate foods is, therefore, free from preconceived chemical ideas, and is based, if not on pure accident, on pure clinical observation.”

By the end of the nineteenth century, researchers had established that the pancreas was responsible for the disease. By the 1920s, insulin was discovered and found to be essential for the utilization of carbohydrates for energy. Without insulin, diabetic patients could still mitigate the symptoms of the disease by restricting the starches and sugar in their diet. And yet diabetologists would come to reject categorically the notion that sugar and refined carbohydrates could somehow be responsible for the disease—another example of powerful authority figures winning out over science.

In the era that predated the discovery of insulin—a hormone that plays the crucial role in the carbohydrate hypothesis we will be discussing—the leading authorities on the treatment of diabetes could be divided into three groups: those firmly convinced that sugar and other carbohydrates played no causative role (among them Carl von Noorden, the pre-eminent German authority); those who thought the evidence ambiguous (including the German internist Bernhard Naunyn) and wouldn’t put the blame on sugar itself but would concede, as Allen remarked, that “large quantities of sweet foods and the maltose of beer” favored the disease onset; and unequivocal believers (Raphaël Lépine of France was one), who would also note that vegetarian, beer-drinking Trappist monks frequently became diabetic, as did laborers in sugar factories.

Those diabetologists who believed that a connection existed argued that the glucose resulting from the digestion of sugar and refined carbohydrates passed with exceptional ease into the blood, and so it was easy to imagine that it might tax the body’s ability to use it. Add sugar to the diet of someone whose ability to assimilate carbohydrates is already borderline or damaged in some way, and that person might pass from an apparently healthy condition to one that is pathological. In such cases, explained Allen, “in the absence of any radical difference between diabetes and nondiabetic conditions, the assumption of a possible production by sugar is logical…. A sufficiently excessive indulgence may presumably weaken the assimilative power of individuals in whom this power is normal or slightly reduced.”