Analysis of Risk Factors and Construction of Nomogram Prediction Model for Hydrocephalus after Intracranial Hemorrhage in Children

Background: Intracranial hemorrhage (ICH) in children can cause serious complications like post‐hemorrhagic hydrocephalus (PHH). Early identification of PHH risk factors is crucial for timely intervention, so we aimed to identify independent predictors of PHH and develop a simplified nomogram for timely risk estimation. Methods: In this single‑center retrospective cohort study, 147 children (≤12 years) with radiologically confirmed ICH (September 2016-August 2024) were analyzed. Patients were classified as PHH (n=36) or non‐PHH (n=111) based on clinical and imaging criteria. Candidate predictors were pre‑specified from clinical relevance and prior literature: age group, admission Glasgow Coma Scale (GCS), CSF total protein, intraventricular hemorrhage (IVH), and concomitant subarachnoid hemorrhage (SAH). Multivariable logistic regression identified independent predictors, and a nomogram was constructed. Internal bootstrap validation assessed optimism-corrected calibration and discrimination. Results: PHH developed in 24.5% of patients. Five predictors were retained in the final model: younger age (OR = 2.81, 95% CI: 1.03-7.72), lower GCS score (OR = 2.76, 95% CI: 1.21-6.30), elevated CSF protein (OR = 5.54, 95% CI: 1.70-18.10), intraventricular hemorrhage (OR = 8.85, 95% CI: 2.36-33.18), and subarachnoid hemorrhage (OR = 5.10, 95% CI: 1.28-20.38). The nomogram demonstrated apparent discrimination (AUC = 0.94) and reasonable calibration within this dataset. Performance estimates should be interpreted cautiously given the limited events-per-variable (EPV) ratio and reliance on internal validation. Conclusion: A simplified five-variable nomogram offers an exploratory tool to estimate PHH risk in pediatric ICH. External, prospective multi-center validation is required before clinical implementation.