OBJECTIVE: Fat distribution as measured by waist-to-hip ratio has been shown to be an important independent predictor of glucose intolerance. Few studies, however, have considered the contributions of the waist and hip circumferences independently. The aim of this study was to investigate the independent associations of waist and hip circumference with diabetes in a large population-based study, and to investigate whether they also apply to other major components of the metabolic syndrome (hypertension and dyslipidemia). In addition, as previous studies were performed in older persons, we investigated whether these associations were present across adult age groups. METHODS: Weight, height, waist and hip circumferences were measured in 11 247 participants of the nationally representative Australian Diabetes, Obesity and Lifestyle (AusDiab) Study. HDL-cholesterol, triglycerides, fasting and 2-h postload glucose were determined, and diastolic and systolic blood pressure was measured. After exclusion of persons already known to have diabetes, hypertension or dyslipidemia, logistic and linear regression were used to study cross-sectional associations of anthropometric variables with newly diagnosed diabetes, hypertension and dyslipidemia, and with continuous metabolic measures, all separately for men (n=3818) and women (n=4582). Analyses were repeated in the same population stratified for age. RESULTS: After adjustment for age, body mass index and waist, a larger hip circumference was associated with a lower prevalence of undiagnosed diabetes (odds ratio (OR) per one s.d. increase in hip circumference 0.55 (95% CI 0.41-0.73) in men and 0.42 (0.27-0.65) in women) and undiagnosed dyslipidemia (OR 0.58 (0.50-0.67) in men and 0.37 (0.30-0.45) in women). Associations with undiagnosed hypertension were weaker (OR 0.80 (0.69-0.93) in men and 0.88 (0.70-1.11) in women). As expected, larger waist circumference was associated with higher prevalence of these conditions. Similar associations were found using continuous metabolic variables as outcomes in linear regression analyses. Height partly explained the negative associations with hip circumference. When these analyses were performed stratified for age, associations became weaker or disappeared in the oldest age groups (age > or =75 y in particular), except for HDL-cholesterol. CONCLUSION: We found independent and opposite associations of waist and hip circumference with diabetes, dyslipidemia and less strongly with hypertension in a large population-based survey. These results emphasize that waist and hip circumference are important predictors for the metabolic syndrome and should both be considered in epidemiological studies. The associations were consistent in all age groups, except in age > or =75 y. Further research should be aimed at verifying hypotheses explaining the 'protective' effect of larger hips.