Abstract
BACKGROUND: The non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio (NHHR) is associated with cardiovascular risk, but its relationship with cognitive impairment has not been well studied. METHODS: In this cross-sectional study, 1,103 adults (median age, 68 years; 54.7% women) from Liuyang were included. Cognitive function was assessed by Mini-Mental State Examination. Restricted cubic splines were used to evaluate associations of NHHR with cognitive impairment. Breakpoint regression identified inflection points. Discrimination was compared using area under the curve (AUC). Machine learning with SHapley Additive exPlanations (SHAP) was applied to assess the relative importance of NHHR and to further explore its relationship with cognitive impairment. RESULTS: Overall, 241 participants (21.9%) had cognitive impairment. NHHR demonstrated a significant U-shaped association with cognitive impairment (overall and nonlinearity p < 0.001). Breakpoint regression identified an inflection point at 2.772; NHHR ≥2.772 was associated with increased risk (odds ratio, 3.36; 95% CI, 2.23-5.05; p < 0.001). Compared with LDL, HDL, and non-HDL, NHHR had the greatest AUC for discriminating cognitive impairment. SHAP analysis confirmed the U-shaped relationship and identified NHHR as the most influential lipid-related predictor. CONCLUSION: In this cross-sectional analysis, NHHR was associated with cognitive impairment in a U-shaped pattern and demonstrated better discrimination than individual lipid measures. These findings suggest that NHHR may serve as an alternative lipid-related index in studies of cognitive health, although longitudinal studies are needed to clarify its predictive value.