Correspondence between the adult treatment panel III criteria for metabolic syndrome and insulin resistance.

OBJECTIVE The aim of the present study was to assess the diagnostic accuracy of the Adult Treatment Panel III (ATP-III) definition of the metabolic syndrome in identifying insulin-resistant individuals and to explore alternative approaches to improve identification of insulin-resistant individuals among asymptomatic adults from the general population. RESEARCH DESIGN AND METHODS The sample consisted of 256 non-Hispanic white subjects without treated hypertension or diabetes, from the Rochester (Minnesota) Heart Family Study (123 men and 133 women; aged 20-60 years). Frequently sampled intravenous glucose tolerance tests were performed in all subjects. The reference standard for insulin resistance was determined by Bergman's minimal model; insulin resistance was defined as an insulin sensitivity index <2 x 10 min(-1) . microU(-1) . ml(-1). Component metabolic syndrome measures included blood pressure determined by sphygmomanometer; fasting serum triglycerides, HDL cholesterol, and glucose concentrations determined enzymatically; and waist circumference determined by tape measure. RESULTS By ATP-III criteria, the prevalence of metabolic syndrome was 15.6% (16.3% in men and 15.1% in women; P = 0.465). The presence of metabolic syndrome had low sensitivity to identify insulin resistance (45% in men and 39% in women; sex difference, P = 0.137) but high specificity (93% in men and 95% in women; sex difference, P = 0.345). Based on the area under the receiver operating characteristic curve (AUC) constructed by counting metabolic syndrome components as recommended by ATP-III, diagnostic accuracy was fair (AUC = 0.797 in men and 0.747 in women). When component metabolic syndrome measures were considered as quantitative traits rather than dichotomized, use of waist circumference alone, rather than counting metabolic syndrome components, improved diagnostic accuracy for insulin resistance (in men, AUC = 0.906, P = 0.001; in women, AUC = 0.822, P = 0.10). CONCLUSIONS Application of the ATP-III metabolic syndrome criteria provides good specificity but low sensitivity to screen asymptomatic white adults for insulin resistance. Measuring just waist circumference is simpler and may provide greater accuracy for identifying insulin resistance.