Improved sex-specific cardiovascular risk prediction with multi-omics data in people with type 2 diabetes

Abstract

Background: To evaluate whether integrating proteomics, metabolomics, and a cardiovascular disease specific polygenic risk score (CVD-PRS) in the SCORE2-Diabetes model improves sex-specific 10-year prediction of major adverse cardiovascular events (MACE) in individuals with type 2 diabetes (T2D). Methods: Genome-wide association study (GWAS), plasma proteomics (with the Olink Explore 3072 platform), and metabolomics (with nuclear magnetic resonance spectroscopy by Nightingale Health) data were measured in the UK Biobank. A novel sex-specific protein algorithm was developed using bootstrap-LASSO (Least absolute shrinkage and selection operator) regression. The CVD-PRS and sex-specific metabolite algorithms were used from previous UK Biobank projects. In a subset of 990 participants with T2D, age 40–69 years, with no prior MACE, and complete multi-omics data, we evaluated, which omics data improved the SCORE2-Diabetes model performance using Harrell’s C-index. Results: Overall 9 proteins were selected for males and 7 for females and adding them to the SCORE2-Diabetes model significantly improved discrimination in the total population (C-index increase from 0.766 to 0.835 (P < 0.001)). Further adding of metabolites significantly improved model performance (C-index, 0.846, P = 0.035), which was mostly attributable to model improvement among males (∆C-index, 0.012, P = 0.078) but not among females (∆C-index, 0.004, P = 0.723). Further adding the CVD-PRS did not statistically significantly improve the SCORE2-Diabetes + proteomics + metabolomics model further in the total population (C-index, 0.848 (P = 0.070)). Conclusions: Sex-specific proteomic signatures markedly improved 10-year MACE risk prediction in individuals with T2D. In men but not in women, further integration of metabolomics may enhance model performance whereas adding the CVD-PRS is not needed. External validation is warranted.