Minimal Model Analysis of Intravenous Glucose Tolerance Test-Derived Insulin Sensitivity in Diabetic Subjects*

Abstract

Although minimal model analysis of frequently sampled iv glucose tolerance tests (FSIGTs) to measure insulin sensitivity is well recognized, application has been limited by the need for endogenous insulin secretion. In the present study we determined whether use of exogenous insulin could permit minimal model assessment of insulin sensitivity (SI) to be extended to diabetic subjects. Normal volunteers had separate FSIGT assessments supplemented with both tolbutamide and insulin to accelerate glucose disappearance, while diabetics had a FSIGT supplemented only with insulin. There was a strong and highly significant correlation between the two assessments in normal subjects (r = 0.87; P less than 0.001), and the rank order of SI generally was maintained with the two assessments over a 3-fold range of SI; however, insulin-determined SI was 16% lower (3.4 +/- 0.4 vs. 4.1 +/- 0.4 x 10(-4) min/microU.microL; P less than 0.01). Diabetic subjects had markedly lower insulin sensitivity than controls (SI = 0.61 +/- 0.16; P less than 0.0001). Across all subjects, the level of fasting serum glucose was correlated inversely with both insulin sensitivity (r = -0.62; P less than 0.05) and acute insulin responses (r = -0.72; P less than 0.02); however, insulin sensitivity in diabetic subjects with little insulin secretion (0.6 +/- 0.2) was comparable to insulin sensitivity in diabetic subjects with near-normal responses (0.6 +/- 0.3). In subjects with fasting hyperglycemia, there were significant correlations between insulin sensitivity and body mass index, percent fat mass, and waist/hip ratio (all P less than 0.03). Among all female subjects, there was also a strong correlation between insulin sensitivity and upper body obesity, as measured by waist/hip ratio (r = -0.68; P less than 0.02). Model parameters also permitted glucose uptake to be estimated in diabetic vs. normal subjects at comparable hyperglycemia (11.1 mmol/L). Total glucose uptake was decreased in diabetic subjects (5.2 +/- 0.8 vs. 12.7 +/- 1.7 mg/min.kg in normals; P less than 0.001), insulin-dependent glucose uptake was diminished to a greater extent (1.3 +/- 0.4 vs. 6.2 +/- 1.2) than noninsulin-independent glucose uptake (3.9 +/- 0.5 vs. 6.4 +/- 0.9; both P less than 0.02). Administration of insulin permits minimal model FSIGT analysis to be applied to diabetic as well as normal subjects, yielding information about both insulin- and noninsulin-mediated glucose uptake as well as insulin sensitivity and insulin secretion.