Abstract:

Objective To explore the risk factors of hydrocephalus in children with cerebral hemorrhage and to create a prediction model. Methods A retrospective analysis was conducted on the clinical data of 120 children with cerebral hemorrhage admitted from April 2016 to December 2023. The incidence of concurrent hydrocephalus was statistically analyzed. Univariate and multivariate logistic regression analyses were used to clarify the risk factors for its occurrence. Results Among 120 children with cerebral hemorrhage, 22/120 (18.33%) developed hydrocephalus. The proportion of patients<8 years old, cerebrospinal fluid protein>3 g/L, intraventricular hemorrhage, and hospitalization complications in the hydrocephalus group was significantly higher than that of the non-hydrocephalus group (P<0.05). Multivariate logistic regression analysis showed that <8 years old (OR=4.593), cerebrospinal fluid protein>3 g/L (OR=6.525), and intraventricular hemorrhage (OR=14.500) were independent risk factors affecting the occurrence of hydrocephalus (P<0.05). The area under the curve (AUC) of the prediction model in identifying hydrocephalus in children with cerebral hemorrhage was 0.841, with a 95% confidence interval (CI) of 0.760-0.937, a sensitivity of 0.928, and a specificity of 0.783. In the validation set, the AUC was 0.822, with a 95%CI of 0.706-0.937. Conduct Hosmer Lemeshow test showed an acceptable goodness of fit (χ²=1.635, P=0.201), indicating a good credibility. Conclusion Children with cerebral hemorrhage may develop hydrocephalus, and age, cerebrospinal fluid protein level, bleeding site are all risk factors for hydrocephalus. Establishing a prediction model can effectively evaluate the incidence of hydrocephalus.

Key words: cerebral hemorrhage, hydrocephalus, risk factors, prediction model