Evaluating indices of insulin resistance and estimating the prevalence of insulin resistance in a large biobank cohort

Abstract

Introduction: Insulin resistance (IR) is involved in the pathogenesis of various metabolic disorders. Several surrogate indices of IR have been proposed. We assessed the performance of seven clinically relevant indirect measures of IR and estimated the prevalence of IR in a large population-based cohort. Methods: The study was conducted on fasting individuals from the Qatar biobank (QBB) participants (n = 7,875). Individuals were considered insulin sensitive (IS) if lean, not diagnosed with diabetes, no hypertriglyceridemia, and not on lipid-lowering drugs, while individuals with Type 2 diabetes (T2D) were considered insulin resistant (IR). Cut-offs were determined as the top or lowest quartile values in the IS participants. The performance of IR indices was based on area under the curve (AUC), sensitivity and specificity. Results: The cut-off for HOMA-IR was determined at 1.878, HOMA2-IR (insulin); 1.128, HOMA2-IR (C-peptide); 1.307, QUICKI; 0.347, TyG; 8.281, McAi; 7.727 and 1.718 for TG/HDL. All IR indices analyzed yielded AUC values ranging from 0.83 to 0.92. TyG was the most robust measure for IR (AUC = 0.92, Sensitivity = 0.90, Specificity = 0.79). The overall prevalence of IR in Qatar was estimated at ~51 – 65%. Conclusions: TyG index was the most robust index for determining IR in the Qatari population. The proposed cut-offs could serve as a reference in Middle Eastern populations for IR screening.