OBJECTIVE: To determine whether the positive statistical associations between measures of total and regional adiposity and measures of glucose, insulin and triacylglycerol (TAG) metabolism reported in Caucasian men, are also observed in UK Sikhs. DESIGN: A matched cross-sectional study in which each volunteer provided a blood sample after a 12-h overnight fast and had anthropometric measurements taken. SUBJECTS: A total of 55 healthy Caucasian and 55 healthy UK Sikh men were recruited. The Caucasian and Sikh men were matched for age (48.7710.9 and 48.3710.0 y, respectively) and body mass index (BMI) (26.172.8 and 26.373.2 kg/m 2 , respectively). MEASUREMENTS: Anthropometric measurements were performed to assess total and regional fat depots. The concentrations of plasma total cholesterol, high-density cholesterol (HDL-C), low-density cholesterol (LDL-C) and small dense LDL (LDL3), TAG, glucose, fasting insulin (ins) and nonesterified fatty acids (NEFA) were analysed in fasted plasma. Surrogate measures of insulin resistance (HOMA-IR) and insulin sensitivity (RQUICKI) were calculated from insulin and glucose (HOMA-IR) and insulin, glucose and NEFA (RQUICKI) measurements. RESULTS: The Sikh men had significantly higher body fat, with the sum of the four skinfold measurements (Ssk) (P ¼ 0.0001) and subscapular skinfold value (P ¼ 0.009) higher compared with the Caucasian men. The Sikh volunteers also had characteristics of the metabolic syndrome: lower HDL-C (P ¼ 0.07), higher TAG (P ¼ 0.004), higher % LDL3 (P ¼ 0.0001) and insulin resistance (P ¼ 0.05 ). Both ethnic groups demonstrated positive correlations between insulin and waist circumference (Caucasian: r ¼ 0.661, P ¼ 0.0001; Sikh: r ¼ 0.477, P ¼ 0.0001). The Caucasian men also demonstrated significant positive correlations between central adiposity (r ¼ 0.275, P ¼ 0.04), other measures of adiposity (BMI and suprailiac skinfold) and plasma TAG, whereas the Sikh men showed no correlation for central adiposity (r ¼ 0.019, ns) and TAG with a trend to a negative relationship between other measures (Ssk and suprailiac) which reached near significance for subscapular skinfold and TAG (r ¼ À0.246, P ¼ 0.007). The expected positive association between insulin and TAG was observed in the Caucasian men (r ¼ 0.318, P ¼ 0.04) but not in the Sikh men (r ¼ 0.011, ns). CONCLUSIONS: In the Caucasian men, the expected positive association between plasma TAG and centralized body fat was observed. However, a lack of association between centralized, or any other measure of adiposity, and plasma TAG was observed in the matched Sikh men, although both ethnic groups showed the positive association between centralized body fat and insulin resistance, which was less strong for Sikhs. These findings in the Sikh men were not consistent with the hypothesis that there is a clear causal relationship between body fat and its distribution, insulin resistance, and lipid abnormalities associated with the metabolic syndrome, in this ethnic group.