Epidemiology
Adding Lp(a) to standard risk factors raises coronary calcium prediction accuracy from 0.741 to 0.755 in elderly diabetics, 486-patient study finds (Rev Cardiovasc Med 2025)
Original title: The Association between Lipoprotein(a) and Coronary Artery Calcification in Elderly Patients with Diabetes: A Cross-Sectional Study
This cross-sectional study included 486 elderly patients with diabetes, stratified into three Lp(a) tertiles, to assess the relationship between Lp(a) and coronary artery calcification (CAC) using logistic regression, stratified analysis, ROC and restricted cubic spline analysis. The highest Lp(a) tertile had significantly higher CAC prevalence than the lower two, and multivariate logistic regression confirmed elevated Lp(a), including as a continuous log-transformed or per-SD variable, independently predicted higher CAC risk across all models. This association held across multiple subgroups, including by age, sex, smoking, hypertension, and hyperlipidaemia status. Adding Lp(a) to the baseline risk model improved the area under the ROC curve from 0.741 to 0.755, and restricted cubic spline analysis showed an approximately linear association between log-Lp(a) and CAC risk (P nonlinear = 0.115). The authors conclude elevated Lp(a) is strongly linked to CAC risk in elderly diabetic patients, and that integrating Lp(a) with conventional risk factors improves predictive accuracy.
Original abstract
Background: Lipoprotein(a) [Lp(a)] is associated with the development of coronary artery calcification (CAC), yet its exact function is not fully understood. This study sought to assess the relationship between Lp(a) levels and the risk of CAC in elderly diabetic patients.
Methods: This cross-sectional study included 486 elderly diabetic patients. The exposure factor was Lp(a) levels, categorized into three groups (T1, T2, T3). The outcome was the presence of CAC. The relationship between Lp(a) levels and CAC was evaluated using several statistical methods, including univariate and multivariate logistic regression, multivariable stratified analysis, receiver operating characteristic (ROC) curve analysis, and restricted cubic spline (RCS) analysis.
Results: The highest Lp(a) group (T3) showed significantly higher prevalence of CAC compared to the T1 and T2 groups. Univariate logistic regression indicated a significant link between Lp(a) and CAC. Furthermore, multivariate logistic regression supported the finding that elevated Lp(a) levels correlated with a heightened risk of CAC in all models. Specifically, each unit rise in Lp(a) was associated with a notable increase in CAC risk, and Log10Lp(a) and each 1 standard deviation increase in Lp(a) also significantly elevated CAC risk. Multivariable stratified analysis demonstrated significant differences in CAC risk across various subgroups, including age ≤70 years, males, females, smokers, hypertensive, non-hypertensive, hyperlipidemic, non-hyperlipidemic, non-stroke, and non-chronic kidney disease patients. ROC curve analysis showed that adding Lp(a) to the baseline model improved the area under the curve from 0.741 to 0.755. RCS analysis indicated a significant, approximately linear association between Log10Lp(a) and CAC risk (p nonlinear = 0.115).
Conclusions: In an elderly diabetic population, elevated levels of Lp(a) were strongly linked to a greater risk of CAC. Integrating Lp(a) measurements with conventional risk factors improves the predictive accuracy for CAC.
Summary written by lp-a.org from the published abstract; figures as published. Page updated 18 August 2026. Methods.