Associations of the Hs-CRP/HDL-C ratio with stroke among US adults: Evidence from NHANES 2015–2018

BACKGROUND: The high-sensitivity C-reactive protein (Hs-CRP)-to-high-density lipoprotein cholesterol (HDL-C) ratio, which integrates insights into inflammation and lipid metabolism, serves as a comprehensive indicator. The association between this ratio and stroke prevalence is endeavored to be explored in this research. METHODS: Drawing on information gathered during the 2015-2018 cycles of the NHANES, the association between the Hs-CRP/HDL-C ratio and stroke was examined through multivariate logistic regression. Additionally, subgroup analysis, interaction test, and restricted cubic spline (RCS) were carried out. Multiple machine learning methods were used to identify the key factors affecting stroke and combined with Shap interpretable models to determine the degree of influence of the key factors. Finally, the results of the logistic regression analysis are used to construct a predictive model, which is represented using a nomogram. RESULTS: This research sample comprised 8,064 participants, yielding a stroke prevalence of 4.04%. A positive correlation was shown between the Hs-CRP/HDL-C ratio and stroke (OR: 1.17, 95% CI: 1.02, 1.35). Interaction tests demonstrated that younger participants were more sensitive to higher Hs-CRP/HDL-C ratios, with a significant interaction in stroke. The RCS analysis indicated a nonlinear association between the exposure variable and to outcome variable. The AUC > 0.8 for a random forest model and an XGBoost model demonstrated their strong predictive value. Ultimately, the generated predictive model is a visual nomogram with an AUC of 0.799. CONCLUSION: The results of the study showed a positive correlation between Hs-CRP/HDL and the prevalence of stroke, with higher Hs-CRP/HDL levels associated with a higher likelihood of stroke. As a stroke prediction model incorporating Hs-CRP/HDL, the nomogram may play a significant role in the early identification of high-risk populations.