Machine Learning-Driven Evaluation of Lipid Markers for Cardiovascular Risk in Chronic Kidney Disease Patients: An Analysis from NHANES

Background CKD is an important health issue for the public globally, greatly raising the prevalence and mortality of CVD. Dyslipidemia is a prevalent metabolic disease in people with CKD that is linked to atherosclerosis and cardiovascular events. Nevertheless, the association in various lipid markers and CVD is still uncertain. The study tries to analyze the relationship with various lipid markers and CVD in CKD patients using machine learning methods and identify the optimal lipid markers as predictors. Methods This study analyzed 2,696 CKD participants from the NHANES from 2005–2018, through multivariate logistic regression for analyzing the relationship with various lipid markers and CVD, employing restricted cubic splines to evaluate linear and nonlinear relationships between variables and outcomes, and evaluating the predictive value of various lipid markers for cardiovascular risk using machine learning models. Results In CKD patients, TC [OR: 0.679(0.614–0.751),P < 0.001], LDL-C [OR: 0.629(0.560–0.704), P < 0.001], and ApoB [OR: 0.986(0.982–0.991),P < 0.001] are independent predictors of cardiovascular disease risk, among which TC, LDL-C, and ApoB show a significant L-type relationship with cardiovascular disease risk, while RC and TG exhibit a U-type relationship. HDL-C has an almost linear relationship. Among lipid markers, LDL-C is the best indicator of CKD patients' risk for CVD. Conclusions In patients with CKD, lipid markers are significant predictors of CVD risk. The integration of these variables into predictive models significantly enhances the precision of risk stratification and facilitates timely identification of individuals at elevated risk within the studied population. To confirm these findings in external cohorts with chronic renal disease, further investigation is required.