The metabolic syndrome and incident diabetes: Assessment of alternative definitions of the metabolic syndrome in an Iranian urban population

Aims: To compare the ability of definitions of Metabolic Syndrome (MetS) in the prediction of type 2 diabetes. Methods: We examined 4756 subjects in an Iranian population who were non-diabetic at baseline. After 3.6 years, 188 individuals developed diabetes. Results: Impaired glucose tolerance (IGT) and MetS definitions predicted type 2 diabetes with odds ratios ranging from 3.7 to 11.9 (all P < 0.05) although IGT had the highest area under the receiver operator characteristic (aROC) curve than all the MetS definitions. Reduction of glucose in the National Cholesterol Education Program (NCEP) definition of the MetS increased diabetes prediction, but adding the family history of diabetes did not change aROC curves. The International Diabetes Federation (IDF) definition had the highest sensitivity and false positive rate (72.2 and 35.1%, respectively) and the WHO definition had the lowest ones (54.8 and 9.2%, respectively) for predicting diabetes. The positive predictive values of all definitions were low (8.6-19.7%) but their negative predictive values were around 98%. Conclusions: In Iranian population, the MetS was inferior to IGT for predicting type 2 diabetes. The NCEP definition of the MetS with reduced level of glucose (not including the family history of diabetes) and IDF definition predicted type 2 diabetes at least as well as WHO definition. (C) 2008 Elsevier Ireland Ltd. All rights reserved.