Differences in the relationship between lipid CHD risk factors and body composition in Caucasians and Japanese

To examine differences in the relationship between fat distribution and lipid coronary risk factors in Caucasian and Japanese population and further to determine whether the cut-points for body mass index (BMI) and waist circumference (WC) proposed by WHO and NHLBI are applicable to Japanese population as a predictor of a lipid risk factor abnormality or not.

Subjects were 895 participants of the Baltimore Longitudinal Study of Aging in the US (BLSA) and 1705 participants of the Longitudinal Study of Aging by the National Institutes for Longevity Science in Japan (NILS-LSA). Subjects were divided into four demographic groups as younger (age<65 y) men and women, and older (age> or =65 y) men and women. Blood total cholesterol, triglycerides, LDL- and HDL-cholesterol and anthropometry were measured. Regression coefficients of BMI and WC on risk factors, sensitivity and specificity of the BMI and WC cut-points for blood lipid abnormality, and mean values of blood lipids at BMI or WC cut-points were computed in both populations.

Height, weight, WC and BMI were significantly greater in the BLSA than those in the NILS-LSA subjects. Total cholesterol, HDL- and LDL-cholesterol were significantly greater in the NILS-LSA than in the BLSA subjects. Sensitivities of BMI and WC cut-points were much lower in the NILS-LSA than in the BLSA subjects. Specificities of BMI and WC cut-points were higher in the NILS-LSA than in the BLSA subjects. Mean values of triglycerides, total cholesterol, HDL- and LDL-cholesterol at BMI=25 were significantly greater in the NILS-LSA than in the BLSA subjects. At the WC cut-point (94 cm for men, 80 cm for women), mean values of all lipids were significantly greater in the NILS-LSA than in the BLSA subjects with the exception of triglycerides in younger women.

The Japanese subjects have smaller BMI and WC, worse total and LDL-cholesterol levels and better HDL-cholesterol levels compared to Caucasians. Sensitivities of BMI and WC for predicting lipid risk factor abnormality are much lower in Japanese. The cut-points for BMI and WC proposed by WHO and NHLBI may be too high for predicting an abnormality in triglycerides, total and LDL-cholesterol in Japanese. For detecting an abnormal HDL-cholesterol level, the BMI and WC cut-points may not be as beneficial for the Japanese population as for Caucasians.