Association between non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio (NHHR) and the characteristics of carotid artery plaques: a retrospective study and machine learning model construction.

BACKGROUND: Low-echo plaques are closely related to stroke. Non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio (NHHR) is a novel lipid metabolism indicator. This study aims to evaluate the predictive value of NHHR for identifying vulnerable plaques. METHODS: This study retrospectively collected data from 3,339 patients with carotid plaque. We used multivariable logistic regression and restricted cubic spline (RCS) method to evaluate the association between NHHR and high-risk plaques. The feature variables were determined through Boruta feature selection, and verified using the Shapley Additive Interpretation (SHAP) method. We constructed 9 machine learning model to further explore the ability of NHHR to predict high-risk plaques. RESULTS: The results of the multivariable logistic regression analysis showed a significant positive association between NHHR and high-risk plaques: In the three-category model (low/medium-high/high-risk plaques), patients with high NHHR had a 57.3% increased risk of developing high-risk plaques compared to those with low NHHR (OR = 1.573, 95% CI: 1.266-1.955; p < 0.001); while in the binary model (non-high-risk/high-risk plaques), the risk of developing high-risk plaques in the high NHHR group increased by 80.2% (OR = 1.802, 95% CI: 1.369-2.378; p < 0.001). RCS also revealed that NHHR was significantly and linearly associated with high-risk plaques. In machine learning, NHHR could be used as a new non-invasive screening indicator for plaque risk (RF model: AUC = 0.68). CONCLUSIONS: Higher NHHR is significantly associated with high-risk carotid plaques. NHHR has strong potential as a convenient and non-invasive auxiliary screening indicator, especially suitable for scenarios where ultrasound examination is not accessible or requires frequent monitoring.