NEW evidence suggests cardiovascular risk prediction can be significantly improved by combining clinical biomarkers, metabolomics and genetics, potentially allowing clinicians to identify high risk individuals earlier and prevent more heart attacks and strokes.
Why Cardiovascular Risk Prediction Needs an Upgrade
Cardiovascular disease remains a leading cause of death worldwide, despite widespread use of risk calculators in primary care. Current European guidelines recommend the SCORE2 model to estimate 10-year risk of fatal and non-fatal cardiovascular events, but cardiovascular risk prediction based solely on traditional factors may miss a substantial proportion of individuals who later develop disease. As screening expands to younger populations with longer life expectancy, there is growing interest in whether additional biological information could refine decision making and better target preventive therapies.
Biomarkers And Genetics in Cardiovascular Risk Prediction
Researchers analysed data from 297 463 UK Biobank participants aged 40 to 69 years, all free of cardiovascular disease, diabetes and lipid lowering treatment at baseline. Over follow up, 8919 incident cardiovascular events occurred. Risk discrimination using SCORE2 alone achieved a C index of 0.719. Adding 11 clinical biomarkers improved prediction with a ΔC index of 0.014 [0.012–0.015], while nuclear magnetic resonance metabolomic biomarker scores improved risk prediction by 0.010 [0.009–0.012]. Polygenic risk scores added a further ΔC index of 0.009 [0.008–0.011]. When all three components were combined, cardiovascular risk prediction improved most, with a ΔC index of 0.024 (0.022–0.027). Risk stratification also improved substantially, with net case reclassification of 16.66% (15.50%–17.81%). Population modelling suggested that events prevented per 100 000 screened could increase from 229 to 413, while largely maintaining the number of statins prescribed per event prevented.
What This Means for Clinical Practice
These findings suggest cardiovascular risk prediction could move beyond cholesterol alone toward a more integrated biological approach. Clinical biomarkers such as cystatin C, Lp(a), CRP and vitamin D, alongside metabolomic inflammation markers and genetic risk, provide complementary information. If validated in routine care, such models could help clinicians personalise prevention strategies, focus treatment on those most likely to benefit, and reduce avoidable cardiovascular events. Future work will need to address cost, accessibility and implementation before widespread adoption becomes feasible.
Reference
Ritchie SC et al. Combined clinical, metabolomic, and polygenic scores for cardiovascular risk prediction. European Heart Journal. 2025;ehaf947.
