OBSERVATIONS Correspondence Between the International Diabetes Federation Criteria for Metabolic Syndrome and Insulin Resistance in a Cohort of Italian Nondiabetic Caucasians The GISIR database I n 2005, the International Diabetes Federation (IDF) released a consensus definition of the metabolic syndrome. The definition, intended to provide a fea- sible predictor of cardiovascular disease and type 2 diabetes, includes elevated waist circumference plus two of the fol- lowing factors: reduced HDL cholesterol or raised blood pressure, triglycerides, or fasting glucose (1). Insulin resistance is a major pathogenic factor for the metabolic syndrome and may independently con- tribute to the risk of cardiovascular dis- ease and type 2 diabetes (2). We assessed the diagnostic accuracy of the IDF defini- tion of metabolic syndrome in identifying subjects with insulin resistance, defined as being in the lower quartile of insulin- stimulated glucose disposal (M clamp ) deter- mined by a standardized hyperinsulinemic- euglycemic clamp (40 mU/min per m 2 body surface area), in a sample of 531 non- diabetic Caucasians from the GISIR (Group of Italian Scientists of Insulin Resistance) database; 28.2% of subjects met the IDF criteria. All components of the metabolic syndrome correlated significantly with M clamp with the closest association being with waist circumference (r =-0.63, P  0.0001) and the weakest with fasting glucose (r =-0.28, P  0.0001) and blood pressure (r =-0.23, P  0.0001). Stepwise regression analysis in a model, including the components of the meta- bolic syndrome, sex, and age, revealed that only three variables were indepen- dently associated with M clamp : waist cir- cumference accounted for 38.4% of its variation, triglycerides accounted for 2.3% of the variation, and HDL choles- terol accounted for 1.0% of the variation. In a logistic regression analysis with ad- justment for age and sex, the risk of IDF- defined metabolic syndrome increased according to the M clamp quartile, with the most insulin-resistant subjects having 15.6-fold higher risk (95% CI 7.7–33.8) than the most insulin-sensitive subjects. Sensitivity of the IDF criteria to identify insulin resistance was low (54.5%), but specificity was high (80.5%). Among the components of the metabolic syndrome, waist circumference showed high sensitiv- ity (90.9%) but low specificity (42.6%), and fasting glucose and triglycerides showed low sensitivity (25.0 and 34.8%, respectively) and relatively high specific- ity (85.0 and 79.7%), whereas blood pressure and HDL cholesterol showed both low sensitivity (53.0 and 56.1%, re- spectively) and low specificity (64.9 and 65.9%). The poor sensitivity of IDF crite- ria in identifying subjects with insulin re- sistance suggests that a significant number of subjects are insulin resistant, possibly at risk for cardiovascular disease and type 2 diabetes, but are not labeled as having metabolic syndrome. Indeed, 60 of 381 (15.7%) subjects who did not meet IDF criteria were insulin resistant (lowest M clamp quartile). Compared with the sub- group of insulin-sensitive subjects (M clamp quartiles 2– 4) who did not meet IDF crite- ria, these insulin-resistant subjects had a significantly worse cardiometabolic risk profile, including higher BMI, waist circum- ference, fasting glucose, triglycerides, and blood pressure. These results suggest that the IDF criteria have good specificity but low sensitivity in identifying insulin resis- tance in nondiabetic subjects but fail to recognize a significant number of insulin- resistant subjects who have an unfavour- able cardiometabolic risk profile. GIORGIO SESTI, MD 1 BRUNELLA CAPALDO, MD 2 PAOLO CAVALLO PERIN, MD 3 STEFANO DEL PRATO, MD 4 LUCIA FRITTITTA, MD 5 SIMONA FRONTONI, MD 6 MARTA LETIZIA HRIBAL, PHD 1 GIULIO MARCHESINI, MD 7 GIUSEPPE PAOLISSO, MD 8 PIER MARCO PIATTI, MD 9 ANNA SOLINI, MD 4 ENZO BONORA, MD 10 ON BEHALF OF THE GROUP OF ITALIAN SCIENTISTS OF INSULIN RESISTANCE (GISIR)* From the 1 University of Catanzaro, Catanzaro, Italy; the 2 University of Naples-Federico II, Naples, Italy; the 3 University of Turin, Turin, Italy; the 4 University of Pisa, Pisa, Italy; the 5 University of Catania, Ca- tania, Italy; the 6 University of Rome-Tor Vergata, Rome, Italy; the 7 University of Bologna, Bologna, Italy; the 8 University of Naples II, Naples, Italy; the 9 University of Milan-Vita e Salute, Milan, Italy; and the 10 University of Verona, Verona, Italy. Address correspondence to Giorgio Sesti, MD, Medicina Sperimentale e Clinica, Policlinico, Uni- versitario Mater Domini, Viale Europa, Campus Germaneto, 88100, Catanzaro, Italy. E-mail: sesti@unicz.it. DOI: 10.2337/dc06-2394 © 2007 by the American Diabetes Association. APPENDIX — *Other GISIR investi- gators: University of Catania: Roberto Baratta and Salvatore Graci; University of Catanzaro: Arturo Pujia and Francesco Andreozzi; University of Milan-Vita e Sa- lute: Livio Luzi, Lucilla Monti, and Eman- uela Setola; University of Padua: Saula Vigili de Kreutzenberg and Roberto Vet- tor; University of Pisa: Eleuterio Ferran- nini and Andrea Natali; University of Rome-La Sapienza: Giancarlo De Mattia and Frida Leonetti; University of Rome- Tor Vergata: Daniela Bracaglia and Maria Adelaide Marini; and University of Ve- rona: Riccardo Bonadonna. ●●●●●●●●●●●●●●●●●●●●●●● References 1. Alberti KGM, Zimmet P, Shaw J: The met- abolic syndrome: a new worldwide defi- nition. Lancet 366:1059 –1062, 2005 2. Bonora E: The metabolic syndrome and cardiovascular disease. Ann Med 38:64 – 80, 2006 O N L I N E L E T T E R S DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007 e33