Abstract 10457: Improved 10-Year Cardiovascular Risk Prediction Combining Serum Metabolites With Traditional Cardiovascular Risk Factors in People With Type 2 Diabetes: Edinburgh Type 2 Diabetes Study

Abstract

Introduction: Cardiovascular disease (CVD) is the leading cause of death for people with type 2 diabetes (T2D). The predictive performance of existing CVD risk scores in T2D populations is suboptimal. Metabolomics is a promising method of identifying novel biomarkers which might improve risk prediction. We aimed to identify a group of metabolites associated with incident CVD in people with T2D and assess its predictive performance over-and-above a current CVD risk score (QRISK3). Methods: In 1,066 individuals with T2D (Edinburgh Type 2 Diabetes Study), a panel of 228 serum metabolites was measured at baseline and incident CVD events were identified over the subsequent 10 years. We applied 100 repeats of Cox LASSO (least absolute shrinkage and selection operator) to select metabolites with frequency >90% as candidate components for a metabolites-based risk score (MRS). The MRS was calculated using the linear predictor in an unpenalized Cox regression model where only candidate metabolites were included. The predictive performance of the MRS was assessed in relation to a reference score which refitted components of QRISK3 plus prevalent CVD and statin use at baseline. Predictive metrics were internally validated using 500-repeat bootstrapping. Results: In 1,021 available individuals (mean age 67.9 years, 51.7% male), 255 people developed CVD (25.0%) during a median of 10.6 years of follow-up. Twelve metabolites relating to fluid balance, ketone bodies, amino acids, fatty acids, glycolysis and lipoproteins were selected to construct the MRS. C-statistics were 0.673 (95%CI 0.642, 0.704) for the MRS alone and 0.718 (95%CI 0.689, 0.747) for the reference score, increasing slightly to 0.736 (95%CI 0.707, 0.764) for the combination of the two. The improved prediction by combining the MRS with the reference score was internally validated in bootstrapping samples, where the C-statistics for the reference score and the combination were 0.679 (95%CI 0.650, 0.708) vs. 0.699 (95%CI 0.671, 0.728) separately. Conclusions: Metabolomics data might improve predictive performance of current CVD risk scores based on traditional risk factors in people with T2D. External validation is warranted to assess the generalizability of improved CVD risk prediction using the MRS.