Predictors of the Need for Ventriculoperitoneal Shunt in Adults with Post-Hemorrhagic Hydrocephalus : A Single-Center Retrospective Cohort Study.

Abstract

Objective The contemporary definition of hydrocephalus is a disorder of cerebrospinal fluid (CSF) physiology producing active ventricular distension and neurological dysfunction, which is distinct from radiological ventriculomegaly alone. Once post-hemorrhagic ventriculomegaly (PHV) is established, the clinically decisive question is whether the patient will progress to shunt-dependent post-hemorrhagic hydrocephalus (PHH) requiring permanent CSF diversion. We sought to identify independent predictors of ventriculoperitoneal shunt (VPS) placement among adults in whom PHV was already established, and to derive a simple bedside risk score for this specific decision point. Methods We retrospectively studied 243 adults with radiologically established PHV (Evans index [EI] ≥0.30) after spontaneous or traumatic intracranial hemorrhage at a single tertiary center between December 2009 and May 2023 (176 shunted, 67 not shunted). The outcome was permanent VPS placement, used as the operational marker of shunt-dependent PHH. Predictors were examined by univariable and multivariable logistic regression. Cutoffs for continuous predictors were derived by receiver operating characteristic (ROC) analysis. The model was internally validated by bootstrap resampling and 10-fold cross-validation, and a simple additive score was constructed from the retained predictors (TRIPOD type 2a). Results On multivariable analysis, predisposing subarachnoid hemorrhage (SAH), a history of hypertension, peak C-reactive protein (CRP) ≥24.4 mg/dL, and EI ≥0.315 were independently associated with VPS placement (all p <0.001; Nagelkerke R² = 0.408). A four-point score combining these variables separated shunt risk from 0% (score 0) to 100% (score 4), with an area under the curve (AUC) of 0.822 and good calibration (Hosmer-Lemeshow p = 0.906); The apparent and optimism-corrected AUCs of the full model were 0.832 and 0.828. The VPS rate did not change significantly across treatment eras (2009- 2015, 63.0%; 2016-2019, 78.8%; 2020-2023, 75.0%; trend p = 0.086). 3. Conclusion In adults with established PHV, SAH, hypertension, elevated CRP, and greater ventricular size were independently associated with progression to shunt-dependent PHH. CRP is a nonspecific systemic inflammatory marker and should be interpreted as an associated, not causal, factor. The four-point score is an exploratory, hypothesis-generating tool derived and internally validated in a single-center retrospective cohort; external validation is required before any clinical application.