Abstract
CONTEXT: Small dense low-density lipoprotein cholesterol (sdLDL-C) refers to an atherogenic subclass of lipoproteins. Owing to limited access to direct measurement in routine clinical settings, several estimation formulas have been developed, with the equation proposed by Sampson et al. being the most widely adopted. Nevertheless, the clinical relevance of estimated sdLDL-C (E-sdLDL-C), particularly in predicting cardiovascular risk, remains to be validated in large-scale, diverse populations. METHODS: The international cohort analysis used two population-based datasets: the China Health and Retirement Longitudinal Study (CHARLS, n = 8,112; median follow-up: 7 years) and the UK Biobank (UKB, n = 321,310; median follow-up: 14.26 years). E-sdLDL-C was estimated using Sampson’s equation. Kaplan–Meier curves were employed to evaluate the incidence of cumulative event. Multivariable Cox models were applied to explore the predictive value of E-sdLDL-C for incident cardiovascular disease while adjusting for conventional lipid parameters and cardiovascular risk factors. Model performance was assessed using Harrell’s C-index and continuous net reclassification improvement to evaluate the incremental predictive value of E-sdLDL-C. RESULTS: An increase in E-sdLDL-C levels was strongly associated with a greater incidence of cardiovascular disease across both cohorts, independent of traditional lipid biomarkers, including low-density lipoprotein cholesterol, non-high-density lipoprotein cholesterol, triglyceride-rich lipoprotein cholesterol, and apolipoprotein B. In the UK Biobank cohort, both ischaemic heart disease and stroke showed a significant association with E-sdLDL-C (adjusted P < 0.05), whereas in the Chinese cohort, the association was evident for stroke but not for cardiac outcomes. Compared with conventional lipid measures, E-sdLDL-C showed stronger and more consistent associations. Notably, in sex-stratified analyses, the association between E-sdLDL-C and cardiovascular risk was stronger in males than in females, suggesting potential sex-specific differences in pathophysiology. Furthermore, E-sdLDL-C demonstrated significantly higher C-index and NRI in predicting cardiovascular outcomes compared with conventional lipid parameters. The findings were further validated through a series of sensitivity checks. CONCLUSION: This international study demonstrated that E-sdLDL-C independently predicts cardiovascular risk and outperforms conventional lipid markers in terms of its prognostic value. In clinical settings where direct sdLDL-C measurement is not available, its estimation from routine lipid panels provides a practical and accessible alternative for cardiovascular risk stratification, potentially improving patient management and outcomes. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12944-025-02717-0.