2235-P: Comparison of Surrogate Indexes of Insulin Resistance as Predictors of Diabetes: Analysis of Five U.S. Cohorts

Introduction and Objective: Identifying simple surrogate indexes of insulin resistance may improve prevention of type 2 diabetes (T2D). We compared the ability of 17 indexes to predict T2D across 5 diverse US cohorts, well-characterized for T2D incidence. Methods: Data came from 5 cohorts with data from oral glucose tolerance tests (OGTTs): Blacks (n=643) and Whites (n=3508) from ARIC, Blacks (n=897) and Whites (n=1277) from CARDIA, and Hispanics (n=1220) from SAMAFS. We calculated indexes from OGTTs, using fasting and 2-hr glucose and insulin levels, and we calculated indexes that do not require insulin measures. Predictive performance was assessed by hazard ratios (HR) per SD, by change in area under the receiver operating characteristic curve (ΔAUC), and by net reclassification information (NRI). Results: There were 1628 T2D cases in 118K years of follow-up. Table 1 shows HRs for T2D risk, ΔAUCs and NRIs for all indexes. Indexes calculated from OGTTs performed best; Gutt had the highest predictive properties. In indexes that only needed fasting insulin, QUICKI and HOMA-IR had nearly equal NRI and ΔAUC. Among indexes that do not require insulin measurement SPISE performed best. Conclusion: In these large US cohorts, surrogate measures of insulin resistance based on OGTTs are best for predicting T2D, while measures based on fasting insulin perform better than those that do not require insulin measures.