Dynamic associations of BMI-metabolic phenotype transitions with incident cardiovascular disease in middle-aged and elderly adults: a longitudinal cohort study

BACKGROUND: The cardiovascular risk conferred by BMI-metabolic phenotypes is dynamic. However, a comprehensive understanding of how longitudinal transitions influence distinct cardiovascular disease (CVD) subtypes remains limited. We aimed to systematically investigate the associations of BMI-metabolic phenotype transitions with the specific risks of incident heart disease and stroke, and to explore possible sex differences. METHODS: This longitudinal study included 6,274 participants aged ≥ 50 years and free of CVD from two prospective cohorts (CHARLS and ELSA). We defined 16 phenotype transitions based on changes in BMI-metabolic status between a baseline (T0) and a four-year follow-up assessment (T1). Using a landmark approach with follow-up commencing from T1, we used adjusted, stratified Cox proportional hazards models to estimate hazard ratios (HRs) for incident composite CVD, heart disease, and stroke over a median of 5.0 years. The stable metabolically healthy normal weight (MHNW→MHNW) group was the reference. RESULTS: Phenotype transitions were frequent. In the pooled six-group analysis, compared with the stable MHNW group, persistent metabolic unhealthiness and deterioration were the most consistent predictors of elevated composite CVD risk (HR 2.04, 95% CI 1.72-2.41; and HR 1.57, 95% CI 1.29-1.90, respectively), whereas improvement was associated with attenuation but not complete elimination of excess risk (HR 1.33, 95% CI 1.07-1.65). In the more granular 16-transition analysis, associations were endpoint-specific: the transition from MHNW to MHOO was not associated with heart disease risk (HR 0.82, 95% CI 0.52-1.28) but was associated with increased stroke risk (HR 2.59, 95% CI 1.44-4.63), and a similar excess stroke risk was observed for transition from MHOO to MHNW (HR 2.80, 95% CI 1.31-6.00). CONCLUSIONS: Transitions in BMI-metabolic phenotypes are associated with future cardiovascular risk in an endpoint-specific manner. Persistent metabolic unhealthiness and deterioration were the most consistent predictors of elevated composite CVD risk, whereas improvement did not always fully eliminate residual risk. More detailed endpoint-specific and sex-specific patterns should be interpreted cautiously, particularly where event numbers were limited. These findings underscore the importance of dynamic risk monitoring and suggest the need for prevention strategies tailored to specific cardiovascular endpoints and patient subgroups.