Objective: To examine relations between obesity and serum concentrations of lipoprotein cholesterol, apolipoproteins, triglycerides and insulin in American and Western Samoans. Associations are also described between these CHD risk factors and abdominal adiposity, and the potential mediating role of insulin in these relationships is examined.
Design: Cross-sectional, using a sub-sample from an observational epidemiological study of cardiovascular disease risk factors among Samoans.
Measurement: Obesity is estimated by the body mass index (BMI), and fat distribution by the abdomen-hip circumference ratio (AHR). All biochemical parameters were measured in the fasted stated.
Subjects: The sub-sample is 178 men and 147 women who were free from hypertension, diabetes and heart disease.
Results: In multivariate linear regression analyses in men the BMI was positively associated with levels of total cholesterol, the total-HDL cholesterol ratio, apolipoprotein B, and the log of triglyceride and insulin concentrations, and negatively associated with HDL and HDL2 cholesterol. The quadratic term for BMI was also found to be significantly predictive of all metabolic parameters in men, except for the log of serum insulin concentrations. Among the women, in contrast, BMI levels were significantly associated only with concentrations of HDL2 cholesterol, triglyceride and insulin. In men, the associations between the AHR and the metabolic parameters were similar to those described for the BMI, but showed no indication of non-linearity. Addition of the log of insulin to these models had little effect on the relations between the AHR and the lipid parameters, with the exceptions of total cholesterol and triglycerides. As with BMI, the AHR was much les predictive of metabolic parameters in women than in men, with a significant relation existing only with the log of insulin concentrations.
Conclusions: These cross sectional data indicate that overall and abdominal adiposity are important correlates of serum lipid parameters among Samoan men, though the associations with BMI are attenuated at higher levels. Neither anthropometric indicator has much relation with these CHD risk factors among the women, perhaps due to extremely high levels of obesity in this group.