Cumulative exposure and change patterns of cholesterol, high-density lipoprotein, and glucose index and incident cardiometabolic multimorbidity in middle-aged and older Chinese adults with metabolic dysfunction-associated steatotic liver disease: a nationwide cohort study

The cholesterol, high-density lipoprotein, and glucose (CHG) index captures integrated glucose-lipid metabolic dysfunction. Whether cumulative CHG(cuCHG) and change patterns correlate with and forecast incident cardiometabolic multimorbidity (CMM) in patients with metabolic dysfunction-associated steatotic liver disease (MASLD) has yet to be established. This nationwide cohort study included 2,058 MASLD participants without baseline CMM, aged ≥ 45 years, from the China Health and Retirement Longitudinal Study (2011–2020). Using a 2015 landmark design, cuCHG was calculated from repeated measurements in 2012 and 2015, with follow-up beginning at the 2015 landmark. K-means clustering identified three CHG change patterns. Cox regression examined associations with incident CMM. Restricted cubic splines assessed dose-response relationships. Predictive value was evaluated by receiver operating characteristic curve analysis, net reclassification improvement, and integrated discrimination improvement. Exploratory machine learning–based Boruta feature selection and SHAP interpretability analyses were performed to identify and quantify the relative importance of variables. Over a mean follow-up of 4.7 years, 246 participants (12.0%) developed CMM. Each one-unit increment in cuCHG conferred a 36% higher CMM risk (HR 1.36, 95% CI 1.24–1.5). The highest versus lowest tertile was associated with a 2.38-fold elevated risk (95% CI 1.68–3.37; P for trend < 0.001). The persistently high change pattern exhibited the greatest risk (HR 2.82, 95% CI 1.92–4.14). A linear dose-response relationship was observed (P for non-linearity = 0.772). Adding cuCHG to the basic model improved the area under the curve from 0.698 to 0.722 (P = 0.012), with improvements in net reclassification (0.259) and integrated discrimination (0.02). Findings were robust in complete-case (cuCHG HR 1.36, 95% CI 1.23–1.5), logistic regression, Fine–Gray competing-risk (cuCHG HR 1.31, 95% CI 1.2–1.42), and subgroup analyses. Boruta and SHAP identified cuCHG among the most important predictors. In middle-aged and older Chinese adults with MASLD, both higher cuCHG exposure and persistently high CHG change pattern were independently linked to an increased risk of CMM. Combining cuCHG with CHG change patterns may enhance early detection of individuals at increased CMM risk in this population. •The CHG index captures integrated glucose-lipid metabolic dysfunction and has been linked to adverse cardiometabolic outcomes. •MASLD is a systemic metabolic disorder predisposing individuals to cardiometabolic diseases. •Prior CHG evidence has predominantly relied on single-time-point measurements, leaving the prognostic value of cumulative exposure and repeated-measurement change patterns unexplored. •Are cuCHG exposure and CHG change patterns associated with incident CMM in middle-aged and older Chinese adults with MASLD? •In this national MASLD cohort, we quantified cumulative CHG exposure and identified its change patterns as predictors of incident CMM. •Higher cuCHG and persistently high change pattern are independently associated with increased CMM risk, with a linear dose-response relationship. •Incremental predictive value was corroborated by improvements in AUC, NRI, and IDI. In an exploratory machine-learning analysis, Boruta and SHAP further identified cuCHG as an important predictor. •CuCHG exposure and change patterns may serve as readily available biomarkers for early identification of MASLD patients at elevated CMM risk. •Derived from routine biochemistry, the CHG index offers a cost-effective tool for risk stratification. •Repeated CHG monitoring may improve cardiometabolic risk assessment.