Prediction of adult post-hemorrhagic hydrocephalus: a risk score based on clinical data

Abstract

AbstractThere is lacking research on risk factors and prediction models associated with Post-hemorrhagic hydrocephalus (PHH). Thus, this present study aimed to analyze the risk factors of PHH and establish a risk-scoring system through a large-scale study. A retrospective study of 382 patients with intracranial hemorrhage assessed age, history and diagnosis, Glasgow coma score (GCS), and fever time. After univariate and logistic regression analysis, a risk scoring system was established according to independent risk factors and evaluated using the area under the curve (AUC). Of the 382 patients, 133 (34.8%) had PHH, 43 (11.3%) received surgical treatment. Factor classification showed that age > 60 years old [odds ratio (OR): 0.347, II = 5 points], GCS < 5 (OR: 0.09, IV = 10 points), GCS 6‒8 (OR = 0.232, III = 6 points), fever time > 9 (OR: 0.202, III = 7 points), fever time 5–9 (OR: 0.341, II = 5 points), CSF-TP x time > 14,4000 group (OR: 0.267, IV = 6 points), and CSF-TP x time 9,601‒14,400 group (OR: 0.502, III = 3 points) were independent risk factors. The result of the receiver operating characteristic (ROC) prediction showed that AUC = 0.790 (0.744‒0.836). Low-risk (IV-VII), moderate (VIII-X), and high-risk group (XI-XIII) incidence of PHH were 11.76%, 50.55%, and 70.00% (p < 0.001), respectively. The coincidence rates in the validation cohort were 26.00%, 74.07%, and 100.0% (p < 0.001), respectively. AUC value was 0.860 (0.780‒0.941). The predictive model was conducive to determining the occurrence of PHH and facilitating early intervention.