Development and temporal validation of the cardiovascular-kidney-metabolic syndrome four-biomarker score (CKM-BS4): a stage-stratified score for all-cause and cardiovascular mortality across CKM stages in a population-based cohort with linked mortali
Although CKM staging stratifies all-cause and cardiovascular mortality risk, substantial heterogeneity persists within individual stages. Existing risk tools typically focus on isolated lipid, glycemic, or renal domains, or are not designed to refine risk within the CKM staging framework. Therefore, we developed and temporally validated CKM-BS4, a four-biomarker score integrating lipids, HbA1c, eGFR and albuminuria, to improve prediction of all-cause and cardiovascular mortality beyond CKM stage. We pooled eight NHANES cycles from 2003–2018 with mortality linkage and incorporated cycle-pooled MEC weights and the complex survey design. Adults aged ≥20 years were split into a training sample (2003–2012) and a temporal validation sample (2013–2018). Non-HDL-C, HbA1c, UACR, and eGFR were log-transformed and standardized using training-set weighted means and standard deviations. Coefficients were learned in survey-weighted Cox models adjusted for age, sex, and race/ethnicity. After evaluation of discrimination, calibration, collinearity, bootstrap stability, and information loss after compression, the compressed CKM-BS4 score model was finalized and applied to temporal validation without refitting. We also compared CKM-BS4 with its four individual components and further examined its competing-risk association, incremental prognostic value beyond CKM stage, and interaction with CKM stage. Among 24,357 adults, 2,167 all-cause deaths and 558 cardiovascular deaths occurred. As a standalone score, the 5-year AUC of CKM-BS4 was 0.797 in training and 0.782 in temporal validation, exceeding each of its individual components. In competing-risk analyses, CKM-BS4 remained independently associated with cardiovascular mortality, with a cause-specific hazard ratio of 1.524 (95% CI 1.381–1.681) and a Fine–Gray subdistribution hazard ratio of 1.337 (95% CI 1.248–1.432). CKM-BS4 also increased across CKM stages and refined within-stage gradients for both all-cause and cardiovascular mortality. No significant CKM-BS4-by-stage interaction was observed. CKM-BS4 is a parsimonious four-biomarker score that remains independently associated with all-cause and cardiovascular mortality beyond CKM stage. It refines stage-specific risk stratification and serves as a pragmatic adjunct to CKM staging for prioritizing prevention and management.
Authors
- Meixian Chen (ORCID: https://orcid.org/0000-0003-4990-6420)
- Daqian Gu (ORCID: https://orcid.org/0000-0001-7237-5539)
- Zhurong Luo
- Na Zhao
- Ming Shen
- Yi Lin
- Wansheng Lin
Institutions
- Fujian University of Traditional Chinese Medicine (CN)
- Fujian Medical University (CN)
- Shanghai University of Traditional Chinese Medicine (CN)
- Yueyang Hospital (CN)
- Second Affiliated Hospital of Nanjing Medical University (CN)
- Nanjing Medical University (CN)
Publication Details
- Journal
- Diabetology & Metabolic Syndrome
- Published
- 2026-09-29
- DOI
- https://doi.org/10.1186/s13098-026-02316-8
- Primary Topic
- Diabetes, Cardiovascular Risks, and Lipoproteins
- Type
- article
- Field-Weighted Citation Impact
- 0.00