Waist circumference does not predict circulating adiponectin levels in sub-Saharan women

BACKGROUND Because of previously reported ethnic differences in determinants and markers of obesity and related metabolic disorders, we sought to investigate circulating levels of adiponectin and their correlates in a sub-Saharan African (sSA) population. SUBJECTS AND METHODS We studied 70 non-diabetic volunteers (33M/37F) living in Yaoundé, Cameroon, aged 24-69 yr, with BMI 20-42 kg/m2. In all participants we measured waist circumference and total body fat by bioimpedance, and obtained a fasting venous blood sample for measurement of plasma glucose, serum insulin and adiponectin concentrations. We performed a euglycaemic hyperinsulinaemic clamp in 1/4 subjects, and HOMAIR was used as surrogate of fasting insulin sensitivity index since it best correlates to clamp measurements. RESULTS Males had lower adiponectin levels than females (8.8 +/- 4.3 vs. 11.8 +/- 5.5 microg/L). There was no significant correlation between adiponectin and total body fat (rs = -0.03; NS), whereas adiponectin was inversely correlated with waist circumference (rs = -0.39; p = 0.001). Adiponectin correlated negatively with insulin resistance (rs = -0.35; p = 0.01). In a regression analysis using fasting adiponectin concentration as the dependent variable, and age, HOMAIR, waist circumference, and fat mass as predictors, waist circumference (beta = -3.30; p = 0.002), fat mass (beta = -2.68; p = 0.01), and insulin resistance (beta = -2.38; p = 0.02) but not age (beta = 1.11; p = 0.27) were independent predictors of adiponectin. When considering gender, these relations persisted with the exception of waist circumference in females. CONCLUSION Adiponectin correlates in this study population are comparable to those observed in Caucasians with the exception of waist circumference in women. The metabolic significance of waist circumference is therefore questioned in sSA women.