Objective: Greater demand and limited resources for intensive care monitoring for patients with neurologic disease may change patterns of intensive care unit utilization. The necessity and duration of intensive care unit management for all neurosurgical patients after brain tumor resection are not clear. This study evaluates a) the preoperative and perioperative variables predictive of extended need for intensive care unit monitoring (>1 day); and b) the type and timing of intensive care unit resources in patients for whom less intensive postoperative monitoring may be feasible.
Design: Retrospective chart review.
Setting: A neurocritical care unit of a university teaching hospital.
Patients: Patients were 158 consecutive postoperative brain tumor resection patients admitted to a neurocritical care unit within a 1-yr period (1998-1999).
Measurements and main results: Twenty-three patients (15%) admitted to the neurocritical care unit for >24 hrs were compared with 135 (85%) patients admitted for <24 hrs. Predictors of >1-day stay in the neurocritical care unit in a logistic regression model were a tumor severity index comprising radiologic characteristics of tumor location, mass effect, and midline shift on the preoperative magnetic resonance imaging scan (odds ratio, 12.5; 95% confidence interval, 3.1-50.5); an intraoperative fluid score comprising estimated blood loss, total volume of crystalloid, and other colloid/hypertonic solutions administered (odds ratio, 1.8; 95% confidence interval, 1.2-2.6); and postoperative intubation (odds ratio, 67.5; 95% confidence interval, 6.5-702.0). Area under the receiver operating characteristic curve for the model of independent predictors for staying >1 day in the neurocritical care unit was 0.91. Neurocritical care unit resource use was reviewed in detail for 134 of 135 patients who stayed in the neurocritical care unit for <1 day. Sixty-five (49%) patients required no interventions beyond postanesthetic care and frequent neurologic exams. A total of 226 intensive care unit interventions were performed (mean +/- sd, 1.7 +/- 2.6) in 69 (51%) patients. Ninety (67%) patients had no further interventions after the first 4 hrs. Neurocritical care unit resource use beyond 4 hrs, largely consisting of intravenous analgesic use (72% of orders), was significantly associated with female gender, benign tumor on frozen section biopsy, and postoperative intubation (chi-square, p <.05).
Conclusions: A small fraction of patients require prolonged intensive care unit stay after craniotomy for tumor resection. A patient's risk of prolonged stay can be well predicted by certain radiologic findings, large intraoperative blood loss, fluid requirements, and the decision to keep the patient intubated at the end of surgery. Of those patients requiring intensive care unit resources beyond the first 4 hrs, the interventions may not be critical in nature. A prospective outcome study is required to determine feasibility, cost, and outcome of patients cared for in extended recovery and then transferred to a skilled nursing ward.