Abstract
Advanced phenotyping using lipidomics enables personalized risk stratification in cardiometabolic disease. Integrating multi-omics approaches may improve detection of high-risk individuals in the Cardiovascular-Kidney-Metabolic (CKM) spectrum.
Objective: To identify lipid metabolites and immune signatures predictive of CKM Stage 4 and the need for advanced coronary interventions: PCI and CABG.
Methods: Patients undergoing coronary angiography provided blood samples for PBMC isolation (flow cytometry) and plasma analyses (immunoassay, targeted lipidomics via LC-MS). A total of 1,143 lipids across 20 classes were profiled. CKM staging followed AHA guidelines. Data were analyzed in R and GraphPad using LipidR and logistic regression models.
Results: Among 200 participants (median age 67, 21% female), 42.5% had obesity, 42% diabetes, and 31% ACS. CKM Stage 3 and 4 were identified in 39% and 20%, respectively; 31.5% underwent PCI, 13% CABG.
The model predicting CKM Stage 4 included NOX5 (PBMCs and monocytes), two lysophosphatidylcholines LPC(18:1), LPC(20:0) and two phosphatidylcholines PC(14:0_18:2) and PC(18:2_18:2), achieving AUC=0.90, NPV=91.0%, PPV=85.2%. PCI prediction included NOX5 and five lipids: phosphatidylinositol PI(20:0_18:1), triacylglycerols TG(54:3_FA(16:0)) and TG(56:6_FA(20:3)), PC(18:1_18:3), and the diacylglycerol DG(16:0_20:5), AUC=0.92, NPV=88.3%, PPV=85%. The CABG model featured NOX5 and PC(18:1_18:3) and PC(16:1_18:2), two TG(54:3_FA(16:0)), TG(56:6_FA(20:3)) and PI(18:0_20:3), with AUC=0.96, NPV=96.7%, PPV=90.9%. All models outperformed clinical variables alone (p<0.0001) and achieved perfect discrimination when clinical variables were integrated (AUC=1.0) Table 1, Figure 1.
Conclusion: Targeted lipidomics combined with immune markers, particularly NOX5, enables accurate classification of patients at advanced cardiometabolic risk. These signatures offer strong translational potential for early detection and intervention in CKM and coronary artery disease.