Abstract
BACKGROUND
Hypertension is closely related to the development and mortality of cardiovascular disease (CVD), insulin resistance (IR) plays a key role in hypertension-related metabolic dysfunction, but there is no universally accepted IR marker for hypertensive populations. Recently, the cholesterol-HDL-glucose (CHG) index and its obesity indicators (CHG-BMl, CHG-WC, and CHG-WHtR) have emerged as reliable IR markers. In this research, we aimed to assess their associations with all-cause and cardiovascular mortality among hypertensive patients with machine learning analysis.
METHODS
Data from 4661 hypertensive participants in the the China Health and Retirement Longitudinal Study (CHARLS) were analysed. Survival machine learning models (Cox, Cox-LASSO, Survival XGBoost) were applied. Model discrimination was assessed by Harrell's C-index and time-dependent ROC (tdAUC) at 3, 5, 7 years. Incremental predictive value was evaluated using continuous net reclassification index (NRI), integrated discrimination improvement (IDI), and decision curve analysis (DCA). Competing-risk (Fine-Gray) regression was performed for cardiovascular mortality.
RESULTS
The CHG index and its obesity indicators were independent predictors of both all-cause and cardiovascular mortality among hypertensive patients. The CHG-WHtR showed strong associations with both all-cause (HR = 2.060, 95% CI: 1.852-2.292, P < 0.001) and cardiovascular mortality (HR = 2.020, 95% CI 1.816-2.247, P < 0.001) per 1-SD increase. The test C-index of Cox-LASSO was 0.920 (95% CI 0.878-0.962), with tdAUC ranging 0.851-0.916 across 3, 5 and 7 years. Adding CHG indices substantially improved reclassification (continuous NRI ranging: 0.349-0.496, IDI ranging: 0.024-0.040). L-shaped relationships were observed with both all-cause (non-linear P < 0.001) and cardiovascular mortality (non-linear P < 0.001). The combination of the CHG index or its obesity indicators into predictive models substantially improved the prediction performance for mortality outcomes.
CONCLUSIONS
The CHG index and its obesity indicators are significant predictors of all-cause and cardiovascular mortality in hypertensive patients, and CHG-WHtR exhibits better prognostic performance in terms of mortality risk. The inclusion of obesity indicators enhances risk stratification and may benefit from tailored interventions among hypertension population.