Evaluation of factors during OGTT to correlate insulin resistance in non-diabetic subjects

The correlation between various factors obtained during oral glucose tolerance tests (OGTT) and insulin sensitivity tests (steady state plasma glucose, SSPG method) were analyzed in non-diabetic subjects by Pearson's simple correlation. Similar data were obtained for normal as well as normal and impaired glucose tolerance (IGT) subjects. Three factors (AUCI(0)(120)AUCG(0)(120), AUCI(0)(120) and AUCI(60)(120), area of under curve (AUC), insulin (I), glucose (G)) were found to have significant correlation (r=0.47-0.54) with SSPG, thus indicating insulin resistance. Those with hyperinsulinemia were noted to have insulin resistance with high precision (98%), but insulin resistant subjects without hyperinsulinemia were all judged to have normal sensitivity. These three factors all had high specificity (98%), but low sensitivity (21-25%) with false negative results being obtained in 75-78%, and false positive results in less than 2%. It was difficult to distinguish positive and negative for the insulin resistance by receiver operating characteristic (ROC) curve of I(0)G(0). The sensitivity of which was 12.8% and specificity 93.8%. By ROC curve, the specificities and sensitivities for the proposed factors, including homeostatic model assessment (HOMA) product (I(0)G(0), I0-120, G(0-120), insulin and glucose at 0-120' during OGTT) ranged from 60 to 70%, which is far from an acceptable level of 90% or more. Therefore, these factors were useful to correlate insulin sensitivity with OGTT results in hyperinsulinemic subjects or in general non-diabetic subjects, but less useful when it came to identifying individual cases with insulin resistance. These parameters should be used only in non-diabetic subjects and their significance is thought to be restricted to cases with hyperinsulinemia.