Impaired beta cell glucose sensitivity rather than inadequate compensation for insulin resistance is the dominant defect in glucose intolerance

Abstract

Aims/hypothesis It is commonly thought that hyperglycaemia results from insufficient compensation of insulin secretion for insulin resistance. To verify this hypothesis, we assessed beta cell function and insulin sensitivity (IS) in a large cohort of volunteers with normal glucose tolerance (NGT) or impaired glucose regulation (IGR), i.e. impaired glucose tolerance or impaired fasting glucose. Methods In men and women with NGT (n=1,123) or IGR (n=156) (age 44±8 years, BMI 25±4 kg/m 2 , mean ± SD) we measured: (1) IS by clamp; (2) insulin secretion rates (ISR) and beta cell glucose sensitivity (=slope of the insulin secretion/plasma glucose dose-response) by C-peptide deconvolution and OGTT modelling; and (3) acute insulin response to intravenous glucose. Results After controlling for centre, sex, age and BMI, fasting and total ISR were inversely related to IS in both groups, whereas beta cell glucose sensitivity was not. Acute insulin response was reciprocally related to IS in both groups, but the relationships were incompatible with inadequate compensation and significance was lost after controlling for fasting ISR. In IGR vs NGT, IS was impaired (92 [75] , p<0.0001), whereas fasting and total ISR were increased (35% and 25%, respectively, p< 0.0001). In fully adjusted models, beta cell glucose sensitivity was the strongest determinant of OGTT glucose levels. Conclusions/interpretation Insulin resistance normally upregulates the secretory tone, with no evidence of defective compensation in IGR. In contrast, beta cell glucose sensitivity is independent of insulin resistance, but a key determinant of glucose tolerance. This suggests that hyperglycaemia results from an intrinsic beta cell defect rather than from inadequate compensation for insulin resistance.

Insulin secretion and impaired glucose tolerance

“…We suggest that this is the reason for the lack of fit to a hyperbola shown in Fig. 1c of Mari et al [1].…”

“…To the Editor: The recent publication in Diabetologia of baseline results from two large-scale longitudinal studies provides welcome confirmation of the association of impaired beta cell function and glucose intolerance outside the diabetic range and promises useful information when follow-up studies are completed regarding factors related to progression of glucose intolerance [1,2]. While both studies conclude that beta cell function is the dominant determinant of glucose tolerance, they differ in one significant detail.…”

“…While both studies conclude that beta cell function is the dominant determinant of glucose tolerance, they differ in one significant detail. Mari et al [1] conclude that variations in their primary index of insulin secretion (glucose sensitivity) are not associated with variations in wholebody insulin sensitivity and therefore are not related to any feedback relationship between beta cell function and insulin sensitivity. On the contrary, DeFronzo et al [2] conclude from their primary index of insulin secretion (a disposition index [DI]), that impairments do represent a failure of feedback between insulin sensitivity and insulin secretion.…”

“…The model is therefore sufficient (or saturated) in the sense that there is no room for improvement in the test data. Whether it is sufficient in different data sets is not known at present, but could be readily tested in the data of Mari et al [1]. It cannot be rigorously tested in the data of DeFronzo et al [2], which include a surrogate index of S i with only moderate correlation (r=0.73) with measured S i [7].…”

“…as glucose sensitivity of insulin secretion. As Mari et al show [1], the particular measure of insulin secretion used makes a difference to the form of relationship between measured insulin secretion and insulin sensitivity. Their carefully derived measure of glucose sensitivity obtained from OGTT responses showed no significant relationship with S i for the reasons discussed above, whereas the measure of total insulin output after an i.v.…”