Association between modified cardiometabolic index and cardiometabolic multimorbidity in middle-aged and older adults: evidence from two nationwide cohort studies

The Modified Cardiometabolic Index (MCMI) is an enhanced version of the Cardiometabolic Index (CMI) and a novel integrative biomarker. Its predictive value for cardiometabolic multimorbidity (CMM), defined as co-occurrence of multiple cardiometabolic conditions, has not been fully explored. We analyzed 7,203 participants from the China Health and Retirement Longitudinal Study (CHARLS, 2011 baseline) and 2,225 from the English Longitudinal Study of Ageing (ELSA, 2012 baseline), with follow-up until 2018 and 2019, respectively. MCMI was calculated as: MCMI = ln [Triglycerides × Fasting Glucose / High-Density Lipoprotein Cholesterol] × Waist Circumference / Height. CMM was defined based on self-reported physician diagnoses of ≥ 2 of the following: hypertension, diabetes, heart disease, or stroke. To assess the association between MCMI and incident CMM, we applied Cox proportional hazards models to estimate hazard ratios (HRs) for time-to-event relationships, restricted cubic spline (RCS) analyses to evaluate potential nonlinear dose–response patterns, and time-dependent receiver operating characteristic (ROC) curves to assess and compare the dynamic predictive performance of MCMI throughout follow-up. Over 7 years, higher MCMI levels were associated with increased CMM risk in both cohorts (CHARLS: HR 1.19, 95% CI 1.16–1.21; ELSA: HR 1.74, 95% CI 1.42–2.12), with risk increasing across quartiles. Participants in the highest quartile had the greatest risk (CHARLS: HR 3.81, 95% CI 3.18–4.56; ELSA: HR 2.77, 95% CI 1.81–4.23). RCS analysis indicated a nonlinear association in CHARLS (P < 0.001) and a linear trend in ELSA. Subgroup analyses showed consistent associations across all subgroups in CHARLS, with significantly higher risk observed in older participants, males, those with higher education, smokers, and drinkers (P for interaction < 0.05). In ELSA, associations were consistent except for education, with no significant interactions observed. Time-dependent ROC analysis showed higher area under the curve (AUC) values for MCMI than CMI at 3 and 5 years; DeLong’s test was significant in CHARLS (P < 0.05) but not in ELSA. MCMI was positively associated with CMM risk in both cohorts. Its predictive performance was superior to CMI in the CHARLS cohort, whereas no significant difference was observed in ELSA. MCMI may improve clinical risk assessment in the Chinese population, although additional evidence is required to verify its predictive value across different ethnic groups.