BACKGROUND: Most studies linking obesity and metabolic syndrome (MS) have used body mass index (BMI) and waist circumference (WC) to measure obesity. While BMI is correlated with direct measures of total and central adiposity, it is influenced by lean body and bone mass. We hypothesize that direct measures of adiposity may help develop further insight into the link between obesity and MS, thus more accurately identifying individuals at high risk for MS. AIM OF THE STUDY: We examined how surrogate and direct measures of adiposity were associated with MS risk and if direct adiposity measures enhanced BMI and WC identification of MS risk. METHODS: 3,734 Chinese female twins aged 20-39 years were studied. Percent body fat (%BF) and proportion of trunk fat to total BF (%TF) were assessed by DEXA. Graphic plots and generalized estimating equations were used to examine the associations of adiposity measures with MS and its components. Concordance of adiposity measures and MS abnormalities between monozygotic (MZ) and dizygotic (DZ) twin pairs were compared. RESULTS: The prevalence of MS increased for high BMI (>or=23 kg/m(2)), %BF (>or=32), WC (>or=80 cm), and (to a lesser degree) %TF (>or=50). Below those thresholds, the prevalence of MS was low (0-5.3%). %TF was independently associated with higher risk of MS and its components even after adjusting for BMI and WC. As a result, among women with normal BMI and WC, high %TF was associated with 1.3-2.0-fold elevated risk of MS components. In contrast, women with high BMI but normal WC and %TF neither have significantly increased risk of MS, nor for any component other than high BP. MZ twins showed higher concordance for MS and its components than DZ twins. CONCLUSIONS: In this lean Chinese rural female sample, BMI >or= 23 and WC >or= 80 were associated with a markedly increased risk of MS, which was further enhanced by elevated %TF. Even in women with a normal BMI and WC, %TF was independently associated with MS and its components. Twin analysis findings suggest that adiposity measurements and MS risk are influenced by genetics.