Abstract / Summary
Cardiometabolic multimorbidity (CMM), encompassing the co-occurrence of major cardiometabolic conditions such as diabetes, heart disease, and stroke, is a growing clinical and public-health challenge associated with adverse functional outcomes and high long-term care needs. Insulin resistance and frailty are plausible, potentially interrelated determinants of CMM, including cerebrovascular complications, but their joint longitudinal contribution has not been fully characterized. We investigated whether estimated glucose disposal rate (eGDR), frailty index (FI), and FI-to-eGDR ratio (FI/eGDR), assessed as baseline values, time-weighted cumulative exposure, and two-wave repeated-exposure patterns, were associated with incident CMM. This prospective cohort study included 5171 participants from the China Health and Retirement Longitudinal Study who were free of CMM at baseline. Baseline eGDR, FI, and FI/eGDR were evaluated as standardized continuous variables and tertiles. Time-weighted cumulative exposures were calculated from Wave 1 and Wave 3 values, and two-wave repeated-exposure patterns were identified using K-means clustering. Cox regression models with sequential covariate adjustment were used to estimate associations with incident CMM, with diabetes, heart disease, and stroke incorporated as components of the cardiometabolic disease burden within the outcome framework. Restricted cubic splines, exploratory subgroup analyses, conventional receiver operating characteristic and decision curve analyses, and sensitivity analyses were performed. A Wave 3 landmark cohort included 5052 participants who remained free of CMM through Wave 3. During follow-up, 441 participants developed CMM. In fully adjusted models, per-SD increases in baseline eGDR, FI, and FI/eGDR were associated with HRs of 0.57 (95% CI, 0.51–0.64), 1.29 (95% CI, 1.19–1.40), and 1.39 (95% CI, 1.29–1.48), respectively; corresponding HRs for cumulative measures were 0.52 (95% CI, 0.46–0.59), 1.51 (95% CI, 1.40–1.64), and 1.52 (95% CI, 1.43–1.60). Compared with T1, the fully adjusted HR for T3 of cumulative FI/eGDR was 6.52 (95% CI, 4.55–9.35). Compared with FI/eGDR repeated-exposure pattern class 1, HRs were 3.87 (95% CI, 3.03–4.93) for class 2 and 9.45 (95% CI, 6.95–12.86) for class 3. In the Wave 3 landmark cohort, FI/eGDR pattern class 2 and class 3 remained associated with higher risk (HR, 3.45 [95% CI, 2.62–4.54] and 6.38 [95% CI, 4.39–9.27], respectively). Spline analyses supported nonlinear associations. Cumulative FI/eGDR had the highest observed AUC among individual measures (0.742; 95% CI, 0.718–0.766), whereas the cumulative FI plus eGDR model had an AUC of 0.758 (95% CI, 0.735–0.781); in the landmark analysis, the corresponding AUCs were 0.731 (95% CI, 0.704–0.759) and 0.740 (95% CI, 0.714–0.767), respectively. Lower eGDR, higher FI, and higher FI/eGDR were associated with incident CMM after multivariable adjustment. Cumulative and repeated-exposure pattern analyses provided complementary information beyond a single baseline measurement. Longitudinal integration of metabolic insulin-resistance and frailty measures may help characterize adults at increased risk of CMM, although the exploratory discrimination results require internal and external validation before any clinical application.