The degree to which an individual's glycemic response to a meal is determined by the glycemic index (GI) and other components of the meal remains unclear, especially when meals are not consumed in a highly controlled research setting. To address this question, we analyzed data collected during the run-in period of a clinical trial. Free-living, nondiabetic adults (n = 57) aged 53.9 ± 9.8 y (mean ± SD) with a BMI of 33.9 ± 5.3 kg/m(2) and waist circumference of 109 ± 11 cm underwent a 75-g oral glucose tolerance test (OGTT) and, on a separate day, wore a continuous glucose-monitoring system (CGMS) for 24 h during which time they recorded all foods consumed. The protein, fat, and available carbohydrate (avCHO) content and GI of the breakfast meals were calculated from the food records and the incremental areas under the glycemic response curves (iAUC) for 2 h after breakfast (iAUC(breakfast)) were calculated from CGMS data. Values for iAUC(breakfast), avCHO, fat, fiber, and BMI were normalized by log-transformation. The ability of participant characteristics and breakfast composition to predict individual iAUC(breakfast) responses was determined using step-wise multiple linear regression. A total of 56% of the variation in iAUC(breakfast) was explained by GI (30%; P < 0.001), iAUC after the OGTT (11%; P < 0.001), avCHO (11%; P < 0.001), and waist circumference (3%; P = 0.049); the effects of fat, protein, dietary fiber, age, sex, and BMI were not significant. We concluded that, in free-living, abdominally obese adults, GI is a significant determinant of individual glycemic responses elicited by self-selected breakfast meals. In this study, GI was a more important determinant of glycemic response than carbohydrate intake.
Trial registration: ClinicalTrials.gov NCT00147264.