Perioperative risk factors associated with ICU intervention following select neurosurgical procedures

Background/Objective: Following cranial neurosurgical procedures, intensive care unit (ICU) admission is routine; however, our institution's growing referral network has led to more frequent bed shortages. Consequently, there are increased requests to transfer our postoperative patients out of the ICU early in the monitoring window. We aimed to find risk factors to prioritize which postoperative neurosurgical patients that should remain in the unit. Patients and Methods: An unmatched case-control study was conducted following retrospective chart review of patients who underwent common cranial procedures between August 2015 and June 2016 at our institution. Patients receiving postoperative ICU intervention were defined as cases. Several perioperative events were investigated for association with postoperative ICU level care. Individual risk factors were analyzed using Chi-squared tests for categorical variables (reported as odds ratio) and independent sample two tailed t-tests for continuous variables. Regression models were used for multivariate analysis. Results: We identified 282 patients who met inclusion criteria, with 148 cases and 134 controls and no statistically significant differences between group demographics. Elective cases carried an odds ratio (OR 0.12, 95 % CI 0.05-0.26, p < 0.001), suggesting decreased likelihood of postoperative intensivist intervention. Single variable analysis showed ICU level of care was more more likely with general anesthesia (OR 3.72, 95 % CI 1.90–7.25, p < 0.001) and American Society of Anesthesiologists (ASA) class IV patients (OR 3.28, 95 % CI 1.59–6.78, p < 0.001). Continuous variables (blood loss and operative time) both demonstrated statistically significant differences (p < 0.001) between case and control groups with higher blood loss (100 ± 167 mL) and operative times (245 ± 119 min) seen in the ICU intervention group. Our regression model identified non-elective cases, operative time, and blood loss having associations with postoperative intensivist intervention. Conclusion: Growing demand for ICU beds at our institution has us looking for more objective data guiding decisions on lower-risk patients who could transfer early out of the ICU in times of overcapacity. Elective endovascular aneurysm treatment and DBS are cranial procedures least likely to receive postoperative ICU level intervention. Consideration to procedural blood loss of 100 cc or more and operative time greater than 4 h should also be given as these risk factors were associated with more likely needing postoperative ICU intervention. These results should not spur drastic changes in ICU protocols, but continued quality improvement projects should investigate these correlations to add more objective data for ICU utilization.