Estimation of Glucose Absorption, Insulin Sensitivity, and Glucose Effectiveness From the Oral Glucose Tolerance Test

• Published in: The journal of clinical endocrinology and metabolism Vol. 110; no. 4; p. e1108
• Main Authors: Stefanovski, Darko, Smiley, Dawn D, Punjabi, Naresh M, Umpierrez, Guillermo E, Vellanki, Priyathama
• Format: Journal Article
• Language: English
• Published: United States 17.03.2025
• Subjects: Adult; Aged; Blood Glucose - analysis; Blood Glucose - metabolism; Cross-Sectional Studies; Diabetes Mellitus, Type 2 - metabolism; Female; Glucose - metabolism; Glucose Intolerance - diagnosis; Glucose Intolerance - metabolism; Glucose Tolerance Test; Humans; Insulin Resistance - physiology; Male; Middle Aged; Models, Biological; insulin sensitivity; minimal model; OGTT; glucose absorption; glucose effectiveness
• ISSN: 1945-7197
• DOI: 10.1210/clinem/dgae308

Glucose tolerance during an oral glucose tolerance test (OGTT) is affected by variations in glucose effectiveness (GE) and glucose absorption and thus affects minimal model calculations of insulin sensitivity (SI). The widely used OGTT SI by Dalla Man et al does not account for variances in GE and glucose absorption. To develop a novel model that concurrently assesses SI, GE, and glucose absorption. In this cross-sectional study conducted at an academic medical center, 17 subjects without abnormalities on OGTT (controls) and 88 subjects with diabetes underwent a 75-gram 120-minute 6-timepoint OGTT. The SI from the Dalla Man model was validated with the novel model SI using Bland-Altman limits of agreement methodology. Comparisons of SI, GE, and gastrointestinal glucose half-life (GIGt1/2), a surrogate measure for glucose absorption, were made between subjects with diabetes and controls. In controls and diabetes, the novel model SI was higher than the current OGTT model. The SI from both controls (ƿ=0.90, P < .001) and diabetes (ƿ=0.77, P < .001) has high agreement between models. GE was higher in diabetes (median: 0.021 1/min, interquartile range [IQR]: 0.020-0.022) compared to controls (median: 0.016 1/min, IQR: 0.015-0.017), P = .02. GIGt1/2 was shorter in diabetes (median: 48.404 min, IQR: 54.424-39.426) than in controls (median: 55.086 min, IQR: 61.368-48.502) without statistical difference. Our novel model SI has a good correlation with SI from the widely used Dalla Man's model while concurrently calculating GE and GIGt1/2. Thus, besides estimating SI, our novel model can quantify differences in insulin-independent glucose disposal mechanisms important for diabetes pathophysiology.