Predicting stroke risk through the neutrophil-to-albumin ratio: A NHANES-based study with clinical validation

Stroke is a major global health issue, and inflammation plays a key role in its pathogenesis. The neutrophil-to-albumin ratio (NPAR) is a novel inflammatory biomarker reflecting inflammation and oxidative stress. However, its association with stroke remains unclear. This cross-sectional study utilized data from the National Health and Nutrition Examination Survey (NHANES) 1999-2020. The association between NPAR and stroke was assessed by determining NPAR and employing various analytical methods, including weighted univariate logistic regression, restricted cubic spline (RCS) regression, and Least Absolute Shrinkage and Selection Operator (LASSO) regression models. Furthermore, a risk prediction nomogram was developed, and its efficacy was validated through the use of receiver operating characteristic (ROC) curve analysis. A total of 393 clinical cases were selected, comprising 193 stroke patients and 200 non-stroke patients. The disease data were validated using bootstrap methods, while clinical data underwent logistic regression analysis and T test to confirm compliance with the established NPAR prediction model. The median NPAR of stroke patients was significantly higher than that of non-stroke patients (14.53 vs 13.72, P < .01). RCS analysis demonstrated a positive association between NPAR and stroke risk, especially when NPAR > 13.75. Subgroup analyses showed that this association remained stable in subgroups such as gender, ethnicity, hypertension, coronary heart disease, etc. The LASSO regression model further identified 13 factors closely related to stroke, such as age, ethnicity, high-density lipoprotein, and hypertension, and constructed a nomogram model with high predictive performance (area under the curve [AUC] = 84%). The T test analysis of clinical data revealed a statistically significant difference in NPAR between stroke and non-stroke groups, as well as between low-risk and high-risk stroke classifications (P < .0001). Findings from the clinical cohort were consistent with those of the NHANES analysis. This suggests that NPAR can not only predict the risk of stroke but also reflect its severity. A nonlinear positive association has been observed between NPAR and stroke, a connection that persists regardless of multiple confounding variables. NPAR may function as a crucial indicator for stroke risk, particularly among men. Given the cross-sectional design, further longitudinal studies are needed to establish causality and explore the underlying mechanisms linking inflammation to stroke.