The purpose of this study is to identify predictive factors for successful external ventricular drain (EVD) wean and removal in nontraumatic subarachnoid hemorrhage (SAH) patients.

Nontraumatic SAH patients often require EVD placement for acute hydrocephalus, but predicting which patients will require long-term CSF diversion has been challenging. Predicting successful weaning parameters could mitigate EVD-related complications from longer use times.

Charts of 275 nontraumatic SAH patients (2016-2022) were retrospectively reviewed. EVD weaning involves a stepwise decrement in EVD settings intending on removal, considered successful when removed without development of increased intracranial pressure or worsened neurologic exam.

Fisher’s exact test assessed the association between categorical variables and first wean success. Variables assessed: age, sex, Hunt Hess Score, Modified Fisher Grade, aneurysm location, treatment modality, opening pressure, EVD outputs (days 3, 5, 7, and at 10, 15, 20mmHg), presence of fever, vasospasm, need for shunt, and length of stay. Logistic regression was utilized to model probability of success of the first wean as a function of multiple continuous variables. A classification tree and cross-validation evaluated model accuracy.

105 patients with at least one EVD wean were included. Success rates of first and second EVD wean were 56.2% and 70.6%, respectively. Significant predictors included age (OR = 0.955; p = 0.022), EVD output at day 7 (OR = 0.996; p = 0.048), day 7 to day 3 output ratio (OR = 0.459; p = 0.014), and output ratio at 20mmHg to 10mmHg (OR = 0.238; p = 0.041). The multivariate model (age, day 7 to day 3 output ratio, last output) had 65% accuracy.

Predictors of successful first EVD wean included younger age, CSF output at day 7, day 7 to day 3 output ratio, and the 20mmHg to 10mmHg output ratio. Predictors of successful EVD weaning can help mitigate EVD-related complications.