THE VALUE OF HEPARIN-BINDING PROTEIN IN PREDICTING EARLY PNEUMONIA IN NEUROSURGICAL INTENSIVE CARE UNIT PATIENTS.

Abstract: Background: The high incidence of pneumonia in neurosurgical intensive care unit (NICU) patients significantly impacts their prognosis. Early identification of high-risk individuals for pneumonia is crucial for timely intervention and personalized treatment. Heparin-binding protein (HBP), an early inflammatory marker, shows promise as a predictor for early-onset pneumonia. Methods: This study enrolled a prospective cohort of 389 NICU patients. Logistic regression analysis was used to identify risk factors for early pneumonia while accounting for the potential confounding effects of other variables on HBP. Restricted cubic splines (RCS) were employed to explore the potential nonlinear relationship between HBP and the risk of early pneumonia. Subgroup analyses were conducted to evaluate the sensitivity of HBP as a risk factor. A nomogram integrating HBP and four other independent risk factors was developed to predict early pneumonia. The performance of the model was assessed using receiver operating characteristic curves, calibration plots, and decision curve analysis. Results: A total of 300 NICU patients were included, among whom 201 developed early pneumonia. Multivariate logistic regression confirmed HBP as an independent risk factor for early pneumonia, with consistent results across all subgroups. The nomogram demonstrated excellent predictive performance, achieving high discrimination (AUC = 0.89) and calibration (Hosmer-Lemeshow test, P = 0.520). Additionally, the model showed significant clinical utility. Conclusions: Elevated HBP levels are independently associated with the risk of early pneumonia in NICU patients. The nomogram integrating HBP provides accurate predictions for early pneumonia.