The glycemic index: meat and potatoes or just gravy?

In 1674, Sir Thomas Willis attributed the polyuria of diabetes to thin blood and proposed the therapeutic use of dietary starch and salt as a means of thickening the blood, as one would thicken gravy. Since then, recommendations for the inclusion of carbohydrate in diabetic diets have been predicated on the understanding (and misunderstanding) of the pathophysiology of diabetes. Thus, dietary recommendations vacillated between low (1) and high (2) carbohydrate content depending on the prevailing attitude regarding the effect of carbohydrate on glucose control. Seminal observations by Himsworth (3) in the 1930s exonerated carbohydrate as a cause of worsening metabolic control and led to the current widespread acceptance of carbohydrate as an integral component of diabetic diets. The most recent diet controversy revolves around what type of carbohydrate is appropriate for the diabetic person rather than how much carbohydrate. This controversy was spawned by the results of studies by Crapo et al. (4,5) that demonstrated that all carbohydrates are not equal in their effect on glucose profiles. The initial studies used individual food substances, not in the context of mixed meals, and showed that I) similar quantities of different simple sugars (e.g., sucrose vs. fructose) had variable effects on the post-prandial glucose levels (5), 2) the form of a food (e.g., glucose drink vs. glucose meal) influenced the resultant glucose profile (4), and 3) different carbohydrate food sources (e.g., potato, rice, and pasta) containing similar carbohydrate and caloric content appeared to have very different effects on postprandial glucose levels (4). Differences in digestibility, fiber, and fat content, the presence of antinutrients, and other factors were invoked to explain the differences in the glycemic effect of the foodstuffs. Jenkins and co-workers employed these observations to generate a "glycemic index," a means of comparing carbohydrate containing foods on the basis of their effects on post-prandial (2-h) glucose profiles relative to a similar gram quantity of glucose (6) or white bread (7). This functional measurement was used to classify various carbohydrate-containing foods in nondiabetics (6) and subsequently in type II (8) and type I (9) diabetic patients. The glycemic index challenged the validity of exchange lists that were based on equivalent grams of carbohydrate. In fact, the original exchange list (1950) proposed by the American Dietetic Association , American Diabetes Association, and the U.S. Public Health Service (10) noted that the fiber content and food form (e.g., raw vs. cooked) might affect the …