The appropriateness of the metabolic syndrome criteria as an indicator of cardiovascular disease risk has been challenged in black Africans. Hence, the aims of this study were (1) to examine the level of agreement between the International Diabetes Federation (IDF) and the National Cholesterol Education Program Adult Treatment Panel III (ATP III) metabolic syndrome criteria, which differ in their emphasis on central obesity; (2) to investigate the degree to which these criteria predict insulin resistance, as estimated by the homeostasis model assessment of insulin resistance (HOMA-IR); and (3) to investigate the extent to which a diagnosis of the metabolic syndrome and insulin resistance may be explained by body fat and its distribution. In 103 normal-weight (body mass index <or=25 kg/m(2), mean: 22.0 +/- 1.8 kg/m(2)) and 119 obese (body mass index >or=30 kg/m(2), mean: 33.9 +/- 5.5 kg/m(2)) urbanized black South African women (27 +/- 7 years old), body composition (dual-energy x-ray absorptiometry), fat distribution (waist and computed tomography), blood pressure, fasting glucose, HOMA-IR, and lipid profiles were measured. Insulin resistance was defined as the upper tertile of HOMA-IR. The overall proportion of individuals who met the IDF and ATP III metabolic syndrome criteria were 13% and 10%, respectively. Agreement was high between the IDF and ATP III metabolic syndrome criteria (kappa = 0.87); however, neither criteria predicted HOMA-IR (kappa = 0.16, 95% confidence interval: 0.05-0.27 and 0.14, 95% confidence interval: 0.05-0.27, respectively). Visceral adipose tissue was the largest contributor to diagnosis of the metabolic syndrome, and waist alone (>80 cm or >88 cm) had an improved specificity (21% or 18% higher, respectively) and positive predictive value (64% or 57% higher, respectively) for identifying insulin resistance compared with the metabolic syndrome criteria. Waist circumference was a better predictor of HOMA-IR than the IDF or ATP III metabolic syndrome criteria in young black African women without known disease. The measurement of waist circumference, as an indicator of disease risk, should therefore be encouraged in the public health setting.