Abstract / Summary
ABSTRACT Aims To explore longitudinal and bidirectional associations between obesity and cardiometabolic diseases (CMDs) using a cross‐lagged panel network (CLPN) framework in three nationally representative Chinese cohorts. Materials and Methods We analysed participants aged 45 years and older from CHARLS (2011–2015; n = 12 491), CHNS (2004–2015; n = 8412) and CFPS (2010–2022; n = 22 193). A CLPN model was used to estimate bidirectional longitudinal associations between obesity and CMDs between participant‐specific first and last eligible surveys. CMDs included hypertension (HBP), diabetes (DM), ischaemic heart disease (IHD) and cerebrovascular disease (CVD). Network replicability was assessed across cohorts. Results Across cohorts, 3.43%–5.49% of participants transitioned from non‐obese to obese status, while 16.44%–26.03% transitioned from no CMD to CMD status. Three networks showed moderate structural similarity, with 12 directionally concordant obesity‐CMD edges and moderate edge‐weight correlations ( r = 0.59–0.64; all p < 0.01). The obesity‐to‐diabetes edge was the strongest directed association in CHARLS (OR 2.59, 95% CI:2.18–3.11) and CFPS (2.77, 2.31–3.29), and the second strongest in CHNS (3.43, 2.67–4.31). Among obesity‐originating edges, the same rank order (diabetes, hypertension, IHD and CVD) was observed across cohorts. Obesity showed the highest out‐expected influence and the lowest in‐expected influence. A positive hypertension‐to‐obesity association and bidirectional hypertension‐diabetes associations were consistently observed. Conclusions CMDs are highly interrelated, and obesity plays a central role in bridging CMDs in the longitudinal networks, highlighting its relevance to both individual CMDs and their clustering as cardiometabolic multimorbidity.