Overtly obesity is relatively rare among the Japanese despite the high prevalence of metabolic disorders, which suggests the need to develop simple and effective methods for assessing metabolic risks among the non-obese individuals as part of public health education. We compared body mass index (BMI), waist circumference, and waist-to-height ratio (W/Ht) as indices for evaluation of clustering of coronary risk factors (hypertension, hyperglycemia, hypertriglyceridemia, hypercholesterolemia, and low HDL cholesterol) in 4,668 men and 1,853 women with BMI < 25 kg/m(2). The sum of coronary risk factors correlated positively with all anthropometric indices, with the closest correlation found for W/Ht. Multiple regression analysis showed that height was a negative independent predictor of the sum of coronary risk factors, while age and waist circumference were positive independent predictors. Among the various proposed anthropometric indices for the evaluation of metabolic risk, the sensitivities for identification of clustering of >/=2 and >/=3 coronary risk factors were highest for a waist-to-height ratio >/=0.5 in both genders. Waist-to-height ratio is more sensitive than BMI or waist circumference alone to evaluate clustering of coronary risk factors among non-obese men and women.