Metabolic dysfunction and central body geometry in cardiovascular disease: joint METS-IR–Conicity phenotypes in CHARLS and NHANES

Background Metabolic dysfunction and abdominal body geometry are related but non-interchangeable dimensions of cardiometabolic phenotype. Whether the Conicity index provides cardiovascular information beyond the metabolic score for insulin resistance (METS-IR) and conventional waist-to-height representations remains uncertain. Methods We analyzed adults aged ≥ 45 years in the China Health and Retirement Longitudinal Study (CHARLS) and 1999–2018 US National Health and Nutrition Examination Survey (NHANES). The primary CHARLS outcome was self-reported incident cardiovascular disease (CVD) recorded at the 2013 or 2015 follow-up; NHANES provided an external survey-weighted cross-sectional comparison using prevalent CVD. METS-IR and the Conicity index were analyzed continuously and as prespecified lower (Q1–Q2) versus higher (Q3–Q4) categories forming G00, G10, G01, and G11 phenotypes. Modified Poisson models estimated risk ratios (RRs) in CHARLS and prevalence ratios (PRs) in NHANES. Multiplicative and additive interaction, temporal consistency, and prespecified waist-to-height ratio (WHtR) and body roundness index (BRI) benchmarks were evaluated. Results CHARLS included 5,837 participants with 761 incident CVD events; NHANES included 10,903 participants with 1,881 prevalent CVD cases. In CHARLS, each 1 standard deviation (SD) increase in METS-IR was associated with higher CVD risk (RR 1.125, 95% confidence interval (CI) 1.053–1.202), whereas the continuous Conicity index association was compatible with the null (RR 1.016, 95% CI 0.950–1.085). Compared with G00, adjusted RRs were 1.377 (1.115–1.702) for G10, 1.244 (1.005–1.540) for G01, and 1.339 (1.112–1.612) for G11. G11/G10 was near null, and neither multiplicative nor additive interaction was supported. In NHANES, G01 was directionally similar but imprecise, whereas G11 was associated with prevalent CVD (PR 1.255, 95% CI 1.087–1.448). Paired G01 contrast differences between the Conicity index and WHtR/BRI classifications were supported in CHARLS but imprecise in NHANES. Conclusions METS-IR showed a consistent cardiovascular association, while the Conicity index contributed mainly at the phenotype level by identifying metabolic–body-geometry discordance rather than through a stronger continuous effect or statistical interaction. These associations do not establish incremental prediction or clinical usefulness; prospective validation is warranted. Clinical trial number Not applicable.