Do All Patients With Cerebral Palsy Require Postoperative Intensive Care Admission After Spinal Fusion?

Study Design Retrospective review of a prospective cohort. Objective To identify patient and surgical factors that alter the length of postoperative intensive care unit (ICU) stays after spinal fusion/instrumentation in patients with neuromuscular scoliosis secondary to cerebral palsy (CP). Summary of Background Data High perioperative complication rates in patients with CP contribute to the practice of utilizing the ICU postoperatively for monitoring. However, this is costly and little is known regarding which patients truly need this increased acuity of care. Methods A prospective, multicenter database was queried for patients with CP who underwent spinal fusion and instrumentation. Patients with an ICU length of stay (LOS) ≤1 day were assumed to not have required postoperative ICU admission. Demographic and surgical characteristics were compared between those with ICU LOS of ≤1 day versus >1 day. A classification and regression tree (CART) analysis was utilized to create a decision algorithm for postoperative ICU admission. Results Three hundred twenty-four patients were identified with a mean ICU LOS of 4.7 days (range 0-47). Sixty-eight patients (21%) had an ICU LOS ≤1 day and 256 patients (79%) had an ICU LOS >1 day. CART analysis demonstrated that the institution where the surgery was performed was the primary predictor with two groups: sites that almost routinely had ICU stay >1 day (92%) and those that were split (50.5% >1 day). In the latter group, an operative time greater than 4 hours was a risk factor for a longer ICU stay. Conclusion Because of their heterogeneous makeup, CP patients should be evaluated individually and their postoperative disposition should not be based on institutional tradition but instead on objective surgical factors. For those patients with surgical times less than 4 hours, discussions should be held regarding the safety of a postoperative disposition to a regular floor. Level of Evidence Level III.