Abstract

Background/Objectives: Metabolic-obesity phenotypes are dynamic, yet little is known about how their longitudinal transitions affect cardiovascular disease (CVD) risk. We aimed to characterize these transitions and examine their associations with incident CVD in middle-aged and older Chinese adults. Methods: We included 4516 participants from the China Health and Retirement Longitudinal Study (CHARLS) who were free of CVD at the follow-up baseline (wave 3). Metabolic-obesity phenotypes (metabolically healthy non-obese [MHNO], metabolically unhealthy non-obese [MUNO], metabolically healthy obese [MHO], and metabolically unhealthy obese [MUO]) were assessed in 2011 and 2015, and six trajectory patterns were defined. Incident CVD (heart disease or stroke) was ascertained from 2015 to 2020. Cox proportional hazards models estimated associations between trajectory groups and incident CVD, and Kaplan–Meier curves compared cumulative incidence across groups. Results: Stable MHNO accounted for 35.96% of the cohort, whereas the stable non-MHNO group accounted for 28.06%. After multivariable adjustment, the small obesity recovery group (n = 89) showed an elevated CVD risk estimate compared with the stable MHNO (HR 2.32, 95% CI 1.33–4.05), while stable non-MHNO group were associated with higher CVD risk (HR 1.72, 95% CI 1.36–2.18). Incident obesity showed elevated but non-significant risk (HR 1.71, 95% CI 0.94–3.10), and metabolic decline showed borderline significance (HR 1.36, 95% CI 1.00–1.85). Sex-stratified analyses showed heterogeneous risk patterns. Conclusions: Metabolic-obesity phenotypes are dynamic, but unfavorable phenotypes often persist. The obesity recovery group, defined using BMI, may represent a heterogeneous group and does not necessarily indicate true cardiometabolic recovery; the observed association should not be interpreted as evidence that weight loss itself increases CVD risk. These findings highlight the importance of jointly considering longitudinal changes in metabolic health and obesity when examining their associations with incident CVD.

Keywords