OBJECTIVE: To compare the prevalence of the Metabolic Syndrome (MetS) defined by four definitions and to determine which definition best identifies those at high cardiovascular disease (CVD) risk and with insulin resistance. METHODS: AusDiab is a population-based survey of 11,247 Australians. Participants had anthropometry, blood pressure, and fasting biochemistry. Ten-year CVD risk was calculated. RESULTS: The prevalence of the MetS using the ATPIII, WHO, IDF, and EGIR definitions was 22.1% (95%Cl: 18.8, 25.4), 21.7% (19.0, 24.3), 30.7% (27.1, 34.3), and 13.4% (11.8, 14.9), respectively. Comparing those with to those without the MetS, the odds ratios (95%CI) for having a 10 year CVD risk > or =15% were 6.6 (5.4, 8.2), 5.5 (4.7, 6.5), 5.6 (4.8, 6.6), and 3.5 (3.0, 4.1), for the WHO, ATPIII, IDF, and EGIR definitions, respectively. The population attributable risk (PAR) of high CVD risk due to the MetS was highest for the IDF (23.4%). Insulin resistance was detected in 56.1, 69.7, 50.9, and 91.1% of those meeting the ATPIII, WHO, IDF, and EGIR definitions, respectively. CONCLUSION: The WHO definition was associated with the greatest CVD risk, but is not practical for clinical use. The higher PAR due to the IDF definition, with only slightly lower CVD risk than WHO, and clinical utility of the IDF definition, indicates that it may be a useful tool for CVD prevention.