Objective: To determine cutoff points of anthropometric variables for predicting incident cardiovascular disease (CVD) in Iranian adults. Design: It is a population-based longitudinal study. Subjects: A total of 1614 men and 2006 women, aged X40 years, free of CVD at baseline were included in the study. Measurements: Body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR), waist-to-height ratio (WHtR) and cardiovascular risks were assessed. Incident CVD was ascertained over a median of 7.6 years follow-up. The adjusted hazard ratios (HRs) for CVD were calculated for 1 s.d. change in all obesity variables using Cox proportional hazards regression analysis. Receiver operator characteristic (ROC) curve analysis was used as the method of defining the points of the maximum sum of sensitivity and specificity (MAXss) of each variable as a predictor of CVD. Results: We found 333 CVD events during follow-up. The risk-factor-adjusted HRs were significant for all anthropometric variables in males and WHR in females and were 1.19, 1.24, 1.21 and 1.24 for BMI, WC, WHR and WHtR in males and 1.27 for WHR in females, respectively (all Po0.05). ROC analysis showed the highest area under curve (AUC) for WHR, WHtR and WC, followed by BMI in males and both genders agedp60 years. In females, WHR and WHtR had the highest AUC, followed by WC and BMI. Among those 460 years old, all the anthropometric variables showed same CVD predicting power. The cutoff values (MAXss) for CVD prediction in males and females were BMIs 26.95 and 29.19 kg m À2 ,WCs 94.5 and 94.5 cm, WHRs 0.95 and 0.90, and WHtR 0.55 and 0.62, respectively. Conclusion: There was no difference between central obesity variables in predicting CVD in males, whereas in females WHR and WHtR were more appropriate. The cutoff values of anthropometric variables were higher in the Iranian than in other Asian populations.