REPRODUCIBILITY OF INSULIN SENSITIVITY MEASURED BY THE MINIMAL MODEL METHOD

Insulin resistance is a critical component underlying the altered glucose homeostatis in a variety of metabolic and non-metabolic disorders. Aging, body fat distribution, obesity, diabetes mellitus or hypertension are well recognized conditions associated with an impaired tissue sensitivity to insulin action. Apart from such constant factors, insulin sensitivity can be acutely modified by independent variables such as physical exercise, dietary factors, alcohol intake or harmless drugs. To evaluate the day-to-day intra-individual variation in insulin sensitivity, glucose homeostasis and lipid profiles, we investigated the insulin sensitivity index (S(I)) (determined by the minimal model method of Bergman), basal and post-glucose-load insulin and glucose levels, serum total triglyceride and lipoprotein cholesterol fractions in 15 healthy young men (24 +/- 1 year, mean +/- SEM), on two different occasions at an interval of 3 weeks (days 1 and 21), after 3 days of a standard dietary regimen and after an overnight fast. Blood pressure, heart rate, body weight and 24 h urinary sodium excretion were almost identical in the two phases. S(I(day 1)) varied from 4.2 to 15.8.10(-4).min-1 pro mu-U/ml (mean: 10.2 +/- 0.9) and correlated with S(I(day 21)) (11.2 +/- 1.2.10(-4).min-1 pro mu-U/ml, r = 0.78, p < 0.0007). The slope of the relationship did not differ from 1 (1.01, p > 0.90), the intercept was close to the origin (0.8, p > 0.73) and the coefficient of variation was 14.4%. Other variables of carbohydrate and lipid metabolism that were strongly correlated to each other on the two different days were: total serum lipids (p < 0.0009 to 0.0002) and the fasting and peak plasma insulin (p < 0.001 to 0.0001) and glucose levels (p < 0.005 to 0.001). These findings indicate that the assessment of in vivo insulin sensitivity using the minimal model method is practical and can be reproduced with accetable intra-individual variation in young healthy subjects.