A. Prevalent diabetes

Metabolic syndrome definition	Prevalence of metabolic syndrome (n/N,%)	Prevalence of diabetes (n/N,%)	Sensitivity	Specificity
WHO, WHR ³ 0.91	250/1005 (24.9%)	47/1005 (4.7%)	0.83	0.78
WHO, waist ³ 94 cm	212/1005 (21.1%)	47/1005 (4.7%)	0.68	0.81
NCEP, waist >102 cm	138/1005 (13.7%)	47/1005 (4.7%)	0.61	0.89
NCEP, waist >94 cm	206/1005 (20.5%)	47/1005 (4.7%)	0.72	0.82

B. Incident diabetes

Metabolic syndrome definition	Prevalence of the metabolic syndrome at baseline (n/N,%)	Incidence of diabetes during follow-up (n/N,%)	Sensitivity	Specificity
WHO, WHR ³ 0.91	211/958 (22.0%)	51/958 (5.3%)	0.67	0.80
WHO, waist ³ 94 cm	180/958 (18.8%)	51/958 (5.3%)	0.57	0.83
NCEP, waist >102 cm	109/958 (11.4%)	51/958 (5.3%)	0.41	0.90
NCEP, waist >94 cm	167/958 (18.0%)	51/958 (5.3%)	0.49	0.84

* WHO, World Health Organization. †WHR, waist-hip ratio. ‡NCEP, National Cholesterol Education Program. Differences in sensitivity of the WHO definition with WHR >0.90 vs. the WHO definition with waist ³94 cm, at baseline P=0.016, at follow-up, P=0.063; vs NCEP definition with waist >102 cm, at baseline P=0.006, at follow-up, P<0.001; vs. NCEP definition with waist >94 cm, at baseline P=0.18, at follow-up, P=0.022. Differences in specificity of the WHO definition with WHR >0.90 vs. all other definitions at baseline and follow-up, P<0.001. Statistical significance was calculated with the McNemar test. In corresponding analyses with impaired fasting glycemia excluded from the definitions, sensitivity for especially the NECP definitions was decreased (WHO definition with waist-hip ratio >0.90 or body-mass index ³30 kg·m^-2, sensitivity 0.55; WHO definition with waist ³94 cm, sensitivity 0.49; NECP definition with waist >102 cm, sensitivity 0.31; NECP definition with waist >94 cm, sensitivity 0.37), with almost no effect of specificity [not shown]).