A simple risk score effectively predicted type 2 diabetes in Iranian adult population: population-based cohort study

Background: Although several strategies to predict the development of diabetes have been developed the question is whether better scores can be developed without sacrificing simplicity. Methods: Data on 3242 participants of Tehran Lipid and Glucose Study aged epsilon 20 years, without diabetes at the baseline that completed a similar to 6-year follow-up were used to develop prediction models by running a series of logistic regression model. A simple score system was then developed based on the most important variables selected with forward stepwise approach. Results: During follow-up, 231 individuals developed diabetes. The area under the receiver operating characteristic curve for the score system based on the model including systolic blood pressure (SBP), family history of diabetes, waist-to-height ratio (WHtR), triglyceride-to-high-density lipoprotein cholesterol ratio (TG/HDL-C) epsilon 3.5 and fasting plasma glucose (FPG) levels epsilon 5mmol l(-1) was 0.83 (95% CI 0.80-0.86); the model discriminated subjects with substantial risk for diabetes, appreciably better than 2-h post-challenge plasma glucose (2h-PCPG) alone (0.78; 95% CI 0.75-0.82) (P < 0.001). Scoring epsilon 25 yielded a positive likelihood ratio of 3.27. FPG levels even in the presence of 2h -PCPG predicted incident diabetes. Conclusion: We presented a simple model based on SBP, family history of diabetes, WHtR, TG/HDL-C and FPG; concluding that this approach is superior to relying exclusively on the 2h-PCPG for identifying individuals at high risk for developing diabetes in a Middle Eastern adult population.