Abstract
BACKGROUND: Cardiovascular disease (CVD) is the leading cause of death in patients with chronic kidney disease (CKD). However, there is still a lack of reliable biomarkers to predict cardiovascular events (CVEs) in this population. METHODS: This study aimed to develop a protein risk score (ProRS) model to predict CVEs in CKD patients. From the UK Biobank Pharma Proteomics Project (UKB-PPP), a total of 1,799 patients with CKD and no prior history of CVD were enrolled. Participants were randomly divided into a training set (70%) and an evaluation set (30%). We analyzed 2,920 plasma proteins to identify associations with CVEs, including coronary heart disease, heart failure, and ischemic stroke. RESULTS: After adjusting for significant clinical factors, 38 proteins remained consistently significant in both the training and evaluation sets. Using an elastic net model, we selected 34 to construct the ProRS. The area under the receiver operating characteristics curve for annual CVEs prediction using the ProRS ranged from 0.67 to 0.74, compared to 0.60 to 0.69 for a clinical risk model, and 0.58 to 0.63 for a polygenic risk score. The 10-year incidence of CVEs among individuals in the top 5% of the ProRS distribution was 44.4%, significantly higher than 29.6% observed in the top 5% of the clinical risk model. Conversely, the bottom 5% of the ProRS group showed a 0% incidence rate, compared to 3.7% in the bottom 5% of the clinical risk model, demonstrating superior performance in both risk identification and exclusion. Notably, among patients classified as low risk by the clinical risk model, those with a high ProRS showed an increased risk of CVEs. In contrast, when the ProRS was low, the influence of the clinical risk model on event prediction was minimal. Mendelian randomization analysis identified 25 proteins whose levels were causally influenced by CKD, 10 of which were also associated with CVD. CONCLUSIONS: We demonstrated that plasma proteomics holds promise as a predictive biomarker for CVEs in patients with CKD. By enabling early identification of high-risk individuals, this approach may facilitate timely preventive interventions and ultimately reduce cardiovascular mortality in this vulnerable population.