The non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio (NHHR) as predictors of hypertensive patients: Analyses of NHANES data with machine learning

Elevated values of the non-HDL/HDL cholesterol ratio (NHHR) have been associated with increased hypertension risk, indicating its potential as a pathogenic factor, but its assessment remains challenging. We analyzed data from 22,562 hypertensive participants in the National Health and Nutrition Examination Survey (NHANES) 2009 to 2018, employing predictive algorithms to evaluate the NHHR index's ability to forecast hypertension outcomes. We found that the risk of hypertension was higher in the highest than in the lowest NHHR tertile. Weighted logistic regression showed revealed a statistically significant positive correlation of NHHR with hypertension prevalence in the fully adjusted model. Restricted cubic-spline analysis showed a linear association in the fully adjusted model. Subgroup analysis indicated that significant interactions between NHHR and hypertension were observed in the subgroups of race, smoking, and educational level. Boruta, algorithm corroborated that NHHR is an important predictor of hypertension. Among the 8 machine-learning models evaluated for predictive capabilities, CatBoost methods are used to construct the models, and their performance is evaluated, with an area under the curve of 0.804. Therefore, NHHR is a significant predictor of hypertensive patients. Incorporating these factors into risk prediction algorithms enhances classification accuracy and facilitates earlier detection of at-risk subjects in this cohort.