Association of External Ventricular Drain Wean Strategy with Shunt Placement and Length of Stay in Subarachnoid Hemorrhage: A Prospective Multicenter Study

Abstract

Background: Survivors of aneurysmal subarachnoid hemorrhage (SAH) face a protracted intensive care unit (ICU) course and are at risk for developing refractory hydrocephalus with the need for a permanent ventriculoperitoneal shunt (VPS). Management of the external ventricular drain (EVD) used to provide temporary cerebrospinal fluid diversion may influence the need for a VPS, ICU length of stay (LOS), and drain complications, but the optimal EVD management approach is unknown. Therefore, we sought to determine the effect of EVD discontinuation strategy on VPS rate.

Methods: This was a prospective multicenter observational study at six neurocritical care units in the United States. The target population included adults with suspected aneurysmal SAH who required an EVD. Patients were preassigned to rapid or gradual EVD weans based on their treating center. The primary outcome was the rate of VPS placement. Secondary outcomes were EVD duration, ICU LOS, hospital LOS, and drain complications.

Results: A rapid EVD wean protocol was associated with a lower rate of VPS placement, including a delayed posthospitalization shunt, in an adjusted Cox proportional analysis (hazard ratio 0.52 [p = 0.041]) and adjusted logistic regression model (odds ratio 0.43 [95% confidence interval 0.18-1.03], p = 0.057). A rapid wean was also associated with 2.1 fewer EVD days (p = 0.007) and saved an estimated 2.5 ICU days (p = 0.049), as compared with a gradual wean protocol. There were fewer nonfunctioning EVDs in the rapid group (odds ratio 0.32 [95% confidence interval 0.11-0.92]). Furthermore, we found that the time to first wean and the number of weaning attempts were important independent covariates that affected the likelihood of receiving a VPS and the duration of ICU admission.

Conclusions: A rapid EVD wean was associated with decreased rates of VPS placement, decreased ICU LOS, and decreased drain complications in survivors of aneurysmal SAH. These findings suggest that a randomized multicentered controlled study comparing rapid vs. gradual EVD weaning protocols is justified.

Keywords: Brain aneurysm; External ventricular drain; Hydrocephalus; Intracranial pressure; Neurocritical care; Neurosurgery; Ventriculoperitoneal shunt.

Fig. 1

Flowchart of enrolled rapid versus gradual wean protocol patients. Italics indicate patients who were excluded from analysis. Patients in the gradual and rapid wean groups were analyzed in an intention-to-treat like manner. VPS were placed acutely in patients who received a VPS during their initial hospitalization. A VPS was placed in a delayed manner in patients who were discharged from the ICU and required an unplanned VPS after the EVD was discontinued. *There were four patients categorized as CMO or who died while in the ICU and who were angio negative such that a symptomatic aneurysm was never found. ^†There was one patient excluded for having an EVD-associated epidural hematoma requiring a craniotomy. angio, angiogram, CMO, comfort measures only, EVD, external ventricular drain, ICU, intensive care unit, VPS, ventriculoperitoneal shunt

Fig. 2

Rapid EVD wean trials are associated with fewer VPS. Adjusted Cox proportional hazards model for VPS placement in patients undergoing rapid vs. gradual weans. EVD, external ventricular drain, VPS, ventriculoperitoneal shunts

Fig. 3

Rapid, less frequent, and early EVD wean trials are associated with shorter EVD duration and shorter ICU LOS. Adjusted estimates for number of days saved in EVD duration and ICU or hospital LOS in relation to rapid vs. gradual wean approaches a, frequency of wean trials b, and the timing to the first trial of EVD wean c. Statistics are for the null hypothesis being no difference between wean protocol, fewer or greater number of weans, and earlier or later weans. EVD, external ventricular drain, ICU, intensive care unit, LOS, length of stay. *p < 0.05, **p < 0.01, ****p < 0.0001

Fig. 4

Individual EVD height and closure status trajectories. EVD status is organized by preassigned wean protocol and whether a VPS was placed. EVD, external ventricular drain, VPS, ventriculoperitoneal shunts