Serum occludin combined with clinical features for predicting early neurological deterioration in intracerebral hemorrhage

Introduction/Objective. To evaluate the incremental predictive value of 24-hour serum occludin beyond clinical features and to develop and internally validate a clinically applicable risk prediction model. Methods. This was a single-centre prospective cohort study. Patients with spontaneous intracerebral haemorrhage (ICH) presenting within 24 hours of symptom onset were enrolled. The primary outcome, early neurological deterioration (END), was assessed at 72 hours post-admission. Variables for model development were selected using least absolute shrinkage and selection operator regression. A logistic regression model was constructed incorporating clinical and imaging factors and serum occludin levels measured at 24 hours post-admission. Model performance was evaluated using the area under the curve (AUC). Decision curve analysis was used to assess net clinical benefit across different risk thresholds. Results. Among the 600 enrolled patients, 210 (35.0%) developed END. The base model incorporating age, admission Glasgow Coma Scale score, ICH volume, intraventricular haemorrhage, location, surgical intervention and systolic blood pressure achieved an optimism-corrected AUC of 0.78 (95% CI: 0.74-0.82). The extended model with added 24-hour serum occludin significantly improved discrimination (AUC = 0.84, 95% CI: 0.81-0.87; ΔAUC = 0.06, p < 0.001), with greater clinical net benefit across threshold probabilities of 10%-40%. Category-free net reclassification improvement was 0.42 (95% CI: 0.28-0.56), and integrated discrimination improvement was 0.08 (95% CI: 0.05-0.11). Risk stratification at 10% and 20% thresholds demonstrated high sensitivity (87%) and specificity (79%) for the identification of the high-risk group. Conclusion. Twenty-four-hour serum occludin significantly enhances predictive performance for END and holds potential as an improved biomarker. However, external validation is necessary before widespread implementation.