Construction and validation of prediction model of intracranial infection after external ventricular drainage

Abstract

Objective To explore the influencing factors of intracranial infection in patients after ventricular drainage, and to build and verify the established risk prediction model. Methods A retrospective analysis was carried out on 757 patients who underwent external ventricular drainage (EVD) in our department from January 2010 to December 2018. According to the occurrence of postoperative intracranial infection, they were divided into intracranial infection group (n=79) and non-intracranial infection group (n=678). Their clinical data were collected and compared. Univariate analysis and multivariate logistic regression analysis were used to explore the influencing factors of intracranial infection after ventricular drainage. A risk prediction model was established and anomogram was drawn. Another 148 patients with EVD in our hospital from 2019 to 2020 were selected as the validation group, including 8 patients with intracranial infection after surgery and 140 patients without. The area under receiver operating characteristic (ROC) curve and Hosmer-Lemeshow test were employed to evaluate the prediction efficacy of the model. Results Hypoproteinemia, lumbar cisternae drainage, EVD time >7 d, pulmonary infection, tracheotomy, open craniocerebral trauma, surgical incision infection, and complicated with bacteremia/sepsis were influencing factors of intracranial infection after EVD. The validation results for the prediction model showed that the area under ROC curve for the risk of postoperative intracranial infection was 0.923 in the modeling group and 0.920 in the verification sample (P < 0.01). As to the Hosmer-Lemeshow test, the P value was 0.450 in the modeling patients and 0.877 in those from the verification group. Conclusion There are 8 high-risk factors for incidence of intracranial infection in patients after EVD. Our established model can effectively predict the occurrence of intracranial infection in these patients.

Keywords

• external ventricular drainage
• intracranial infection
• influencing factors
• prediction model
• nursing care