ABSTRACT Background Cardiovascular–kidney–metabolic (CKM) syndrome reflects complex pathobiological interactions among metabolic disorders, kidney injury, and cardiovascular disease (CVD). Stages 2 and 3 represent critical phases of disease progression characterised by high pathological heterogeneity. This study aimed to develop a CVD protein risk score (PRS) for this population and evaluate its incremental predictive value over the PREVENT model. Methods This study included 24 017 participants with CKM Stages 2–3 from the UK Biobank. Using 2923 plasma proteins measured via the Olink platform, a PRS was developed in a training set ( n = 19 218) using the LASSO method. In the validation set ( n = 4799), the incremental predictive performance of this score over the PREVENT model was assessed using Harrell's C‐statistic, net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Results A risk score comprising 63 proteins was constructed, primarily reflecting inflammation, kidney injury and matrix remodelling. Key proteins included growth differentiation factor 15 (GDF15), hepatitis A virus cellular receptor 1 (HAVCR1), matrix metallopeptidase 12 (MMP12) and NT‐proBNP. In the validation set, after adjusting for PREVENT risk factors, individuals in the high PRS group had a 2.56‐fold higher risk of CVD compared to those in the low score group (HR: 2.56, 95% CI: 1.96–3.37). Integrating the score into the PREVENT model improved the C‐statistic by 0.034 (0.672–0.706) and achieved a 10‐year NRI of 15.8% (95% CI: 9.5%–20.9%) and an IDI of 2.2% (95% CI: 1.3%–3.3%). Conclusion Combining the PREVENT model with the PRS developed in this study enhances the prediction of future CVD events in the CKM Stages 2–3 population. This approach facilitates the capture of residual risk and supports precision risk stratification and management for this high‐risk group.
Han et al. (Mon,) studied this question.