Longitudinal patterns and determinants of lipoprotein(a) variability: a 15-year multi-center study
D Y Chen, C Y Chang, M L TsaiAbstract
Background/Introduction
Lipoprotein(a) [Lp(a)] concentrations are mainly determined by genetic factors, yet evidence on their long-term variability and determinants remains limited, particularly among Asian populations.
Purpose
This study examined temporal changes in Lp(a) concentrations and identified clinical predictors of substantial intra-individual elevation.
Methods
We conducted a retrospective cohort study using the Chang Gung Research Database from January 2004 through December 2019. Adults with baseline and follow-up Lp(a) measurements at least one year apart were included. Participants were categorized according to temporal changes: increase ≥10 mg/dL, increase <10 mg/dL, or decrease. Multivariable logistic regression was used to identify predictors of Lp(a) increase ≥10 mg/dL.
Results
Among 17,363 participants (mean age 42.1 years; 73.3% male) with a mean follow-up of 5.2 years, median Lp(a) levels declined from 9.3 mg/dL (interquartile range [IQR] 4.1–17.3) to 7.4 mg/dL (IQR 3.6–16.3). Only 479 participants (2.8%) exhibited an increase of ≥10 mg/dL. Independent predictors of such elevation included higher baseline Lp(a), new statin initiation, diabetes mellitus, female sex, and higher triglyceride levels. By modeling Lp(a) using restricted cubic splines, we identified a baseline threshold of 17.1 mg/dL for predicting an increase in Lp(a) ≥10 mg/dL during follow-up. We constructed a simple score by summing 5 risk factors and evaluated its association with subsequent Lp(a) elevation. The model demonstrated good discrimination (AUC, 86.4%; 95% CI, 84.8%–88.0%). The risk of Lp(a) increase ≥10 mg/dL rose progressively with a greater number of risk factors.
Conclusion(s)
In this large longitudinal Taiwanese cohort, Lp(a) concentrations were generally stable, with only a small proportion showing substantial increases. However, specific clinical factors predicted Lp(a) elevation, suggesting that repeat Lp(a) measurements may be warranted in selected high-risk individuals to enhance cardiovascular risk stratification.Fig 1.Flowchart for the study participants.Fig 2.Predictive model for Lp(a) change