There have been two approaches to examining the effects of carbohydrate on postprandial glycemia: 1) studies designed to determine the glycemic response of foods containing carbohydrate, regardless of the source, i.e., starch, sugar, or fiber (glycemic index [GI] of carbohydrates), and 2) studies comparing the glycemic response to equivalent amounts of starches or sugars. A number of food factors determine the glycemic response to carbohydrates: food form, digestibility, cooking, sugars, type of starch, presence of antinutrients, and second meal or lente effect. In people with diabetes, the severity of glucose intolerance and premeal glucose concentrations also influence the glycemic response. The GI attempts to classify individual foods (50-g portions) by the extent to which they raise blood glucose levels compared with a standard (reference carbohydrate), initially glucose and in later studies bread. Acute glycemic responses differ, however, when meals containing low GI foods are compared to meals containing high GI foods long term (measured by fructosamine or hemoglobin A1c), the outcomes are mixed. If there is an effect from the GI of foods on glycemia, it is modest at best. In other studies, when sucrose is substituted for a variety of starches--in meals or snacks and both acutely and for up to 6 weeks--the glycemic response is similar if the total amount of carbohydrate is similar. Therefore, the recommendation for persons with diabetes in regard to the glycemic effect of carbohydrates is that the total amount of carbohydrate in meals or snacks is more important than the source or type and is the first priority in the planning of meals or snacks. This has led to the implementation of carbohydrate counting, in which foods are listed as carbohydrate choices based on the amount and not the source of the carbohydrate.