Research Square · Nov 16, 2026
Continuous CKM Severity Score: Longitudinal Associations and Treatment-Related Interpretability
By Liu Y, Xin W
Background Initial evaluation of the continuous cardiovascular-kidney-metabolic syndrome severity score (cCKMS-S), including analyses in NHANES, linked the score to mortality. Its outcome associations in an independent longitudinal setting and the interpretation of within-person change remain uncertain because treatment indicators contribute positively to the score. Methods Published sex-specific equations were applied without recalibration. In CHARLS, associations of the 2015 score and 2011–2015 change with incident cardiovascular disease were evaluated using weighted complementary log-log models. NHANES served as a survey-aware benchmark for associations with all-cause and heart-disease mortality. Covariates were multiply imputed with outcomes and follow-up included as auxiliary predictors. Secondary analyses examined dose–response, medication and nonmedication change, ordered-indicator one-factor structure, and prespecified robustness checks. Results Among 2,807 CHARLS participants, 534 developed cardiovascular disease. Each 1-SD higher 2015 cCKMS-S was associated with incident cardiovascular disease (hazard-ratio analogue 1.39, 95% confidence interval [CI] 1.29–1.50); each SD greater score change was also associated with the outcome (1.29, 95% CI 1.20–1.39). Among 18,592 NHANES participants, 2,385 all-cause and 605 heart-disease deaths occurred; corresponding hazard ratios per 1-SD were 1.19 (95% CI 1.11–1.27) and 1.50 (95% CI 1.35–1.68). Mutually adjusted nonmedication and medication-score changes were each associated with incident cardiovascular disease (1.19 for both), but this decomposition was noncausal. Ordered-indicator loading patterns were similar, whereas global one-factor fit was mixed. Alternative score equations attenuated the CHARLS estimates, but positive directions were retained across sensitivity analyses. Conclusions In CHARLS, cCKMS-S level and two-time-point increase were associated with subsequent cardiovascular disease; NHANES provided a consistent survey-aware mortality benchmark. Medication-related scoring and incomplete one-factor fit constrain interpretation of change and cross-population comparability. Any clinical application would require independent calibration and utility-focused validation.