Multilevel analyses of effects of variation in body mass index on serum lipid concentrations in middle-aged Japanese men.

For the effective use of the annual workplace health checkup data, we tried to perform multilevel analyses to explore whether the year-to-year weight variation causes any concurrent effects on the lipid profiles among middle-aged Japanese workers. Subjects were 1939 healthy male workers 40-59 in age from whom serial data of total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and triglyceride (TG) were collected during health checkups conducted in the years 1997-2000. The effects of body mass index (BMI) on serum concentrations of those lipids were investigated by statistical analysis with multilevel modeling to distinguish multiple levels of information with individual repeated measurements within individuals. A significant increase of TC and TG, and decrease of HDL-C with BMI increase were confirmed. Subanalyses according to both the baseline BMI status (< 25 kg/m2 or > or =25 kg/m2) and smoking status (never, former, or current) yielded the same BMI-dependent changes of lipid profiles, but obese never smokers failed to show significant effects of BMI on HDL-C concentrations. Multilevel analyses of annual health checkup data linked at individual levels indicated that year-to-year weight variation, though usually in a much narrower range than the between-individual variation, had a strong impact on the corresponding changes of serum concentrations of TC, HDL-C, and TG. This result supports the public health significance of intervention into weight control to prevent the development of atherogenic risks among a healthy workplace population.