Cumulative Cholesterol-HDL-Glucose Index Predicts Cardiometabolic Multimorbidity in MASLD

People with metabolic dysfunction-associated steatotic liver disease (MASLD) face an elevated likelihood of developing multiple cardiometabolic conditions at once, but identifying who is most at risk has remained difficult. A new nationwide cohort study from China reports that a single composite measure, the cholesterol, high-density lipoprotein, and glucose (CHG) index, may help. When the index is tracked cumulatively over repeated measurements, higher exposure is associated with a substantially increased risk of incident cardiometabolic multimorbidity (CMM), defined as the co-occurrence of multiple cardiometabolic diseases.

The study, published in Cardiovascular Diabetology, drew on the China Health and Retirement Longitudinal Study (CHARLS), a nationally representative survey running from 2011 to 2020. The researchers identified 2,058 participants aged 45 years or older who had MASLD but no cardiometabolic multimorbidity at baseline. To separate prior metabolic exposure from later outcomes, they used a 2015 landmark design: the cumulative CHG (cuCHG) index was calculated from repeated measurements taken in 2012 and 2015, and follow-up for new CMM cases began only at the 2015 landmark point.

Over a mean follow-up of 4.7 years, 246 participants, or 12.0 percent of the cohort, developed cardiometabolic multimorbidity. The central finding was a dose-like association: each one-unit increase in cuCHG corresponded to a 36 percent higher risk of CMM, with a hazard ratio of 1.36 (95 percent confidence interval 1.24 to 1.5). Comparing exposure levels directly, participants in the highest tertile of cuCHG had a 2.38-fold elevated risk relative to those in the lowest tertile (95 percent CI 1.68 to 3.37), and the trend across tertiles was statistically significant.

Because single measurements can miss how a person’s metabolic state evolves, the researchers also characterized trajectories. Using K-means clustering on the repeated CHG measurements, they identified three distinct change patterns. The pattern in which CHG remained persistently high carried the greatest risk: compared with the reference pattern, it was associated with a hazard ratio of 2.82 (95 percent CI 1.92 to 4.14) for developing CMM. This suggests that both the accumulated level of glucose-lipid dysfunction and its stability over time carry information about future risk.

The dose-response relationship itself was examined with restricted cubic splines. The analysis showed a linear pattern, with no evidence of non-linearity (P for non-linearity = 0.772). In other words, within the range studied, risk rose steadily with increasing cumulative exposure rather than appearing only above some threshold.

Beyond association, the team tested whether cuCHG improves prediction. Adding cuCHG to a basic statistical model raised the area under the receiver operating characteristic curve from 0.698 to 0.722 (P = 0.012), a modest but measurable gain in discrimination. Reclassification metrics also improved, with a net reclassification improvement of 0.259 and an integrated discrimination improvement of 0.02. In exploratory machine-learning analyses, Boruta feature selection and SHAP interpretability methods identified cuCHG as one of the most important predictors in the model.

The authors report that the findings held up across several sensitivity analyses. A complete-case analysis yielded a nearly identical hazard ratio for cuCHG (1.36, 95 percent CI 1.23 to 1.5), logistic regression models and Fine-Gray competing-risk models, which account for the possibility of death as a competing event, produced a similar estimate (hazard ratio 1.31, 95 percent CI 1.2 to 1.42), and subgroup analyses were consistent with the main result.

The CHG index is attractive as a risk marker because it is derived from routine biochemistry, combining cholesterol, high-density lipoprotein, and glucose values that are already collected in standard blood panels. The authors suggest that cumulative exposure and change patterns of this index could serve as readily available biomarkers for early identification of MASLD patients at elevated CMM risk, and that repeated CHG monitoring may improve cardiometabolic risk assessment in this population.

Several limitations frame these conclusions. The study is observational, so the association between cuCHG and incident CMM does not establish that lowering the index would reduce risk. The cohort consisted of middle-aged and older Chinese adults with MASLD, and generalization to other populations or to people without MASLD is not demonstrated by these data. The machine-learning component was explicitly exploratory, and the article was released as an accepted manuscript subject to further edits before the final version of record. Within those boundaries, the study adds a longitudinal dimension to CHG index research, which has previously relied largely on single-time-point measurements, and points to repeated metabolic assessment as a practical avenue for stratifying cardiometabolic risk in MASLD.

Subject of Research: Cumulative CHG index exposure and change patterns as predictors of incident cardiometabolic multimorbidity in Chinese adults with MASLD

Article Title: 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

Article References: Chen, Y., Chen, C., Fang, X., Liu, Y., & Shi, F. (2026). 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. Cardiovascular Diabetology. https://doi.org/10.1186/s12933-026-03418-3

Image Credits: AI Generated

DOI: 10.1186/s12933-026-03418-3

Keywords: MASLD, CHG index, cardiometabolic multimorbidity, cohort study, CHARLS, risk prediction, glucose-lipid metabolism, Cox regression, machine learning, China, Cumulative, exposure