Introduction: Finding simple tools for prediction of metabolic syndrome (MetS) is of high interest. Role and predictive value of lipid profile in MetS prediction is still controversial. The current study aimed to compare ability of lipid ratios in prediction of MetS and also to determine cut points using longitudinal data framework in Iranian population. Materials and Methods: We studied 3212 participants of age > 35 years followed for 12 years with 3 visits at baseline, year 6 and year 12 in Iranian adults. Predictive power and optimal cut points of lipid ratios were evaluated using receiver operating characteristic (ROC) curve analysis. We also considered missing data modelling to account for cohort attrition. Results: Of total sample, 1015 (38.2%) individuals had MetS in year 6 that increased to 1288 (40.1%) in year 12. TG/HDL-C was the best predictor of MetS with AUC of 0.88 and 0.87 for 6 and 12 years of follow up. Optimal cut point in 6 and 12 years of follow-up were, respectively, 2.83 and 3.35 for TG/HDL-C, 4.22 and 4.66 for TC/HDL-C and 2.88 and 2.73 for LDL-C/HDL-C. Conclusion: lipid ratios are simple and commonly measured in primary care and could help in identifying individuals with high risk of MetS and reduce related complications and costs. Large-scale multi-ethnic studies are needed to assess generalizability of the results.
