Variability in length of stay following neuromuscular spinal fusion.

Background: Patients with neuromuscular scoliosis (NMS) who undergo posterior spinal fusion (PSF) often have long, protracted hospital stays because of numerous comorbidities. Coordinated perioperative pathways can reduce length of hospitalization (LOH) without increasing complications; however, a subset of patients may not be suited to rapid mobilization and early discharge.
Methods: 197 patients with NMS underwent PSF at a single hospital by two surgeons with a post-operative care pathway emphasizing early mobilization, rapid transition to enteral feeds, and discharge prior to first bowel movement. Average LOH was 4.9 days for all patients. Patients were divided into quartiles (< 3 days, 3-5 days, 5-7 days, > 7 days) based on their LOH, and their charts were retrospectively reviewed for preoperative, intraoperative, and postoperative factors associated with their LOH.
Results: Age at surgery, gender, the need for tube feeds, and specific underlying neuromuscular disorder were not significant predictors of LOH; however, severely involved cerebral palsy (CP) patients (GMFCS 4/5) were more likely to have extended stays than GMFCS 1-3 patients (p = 0.02). Radiographic predictors of LOH included major coronal Cobb angle (p = 0.002) and pelvic obliquity (p = 0.02). Intraoperative predictors included longer surgical times, greater numbers of levels fused and need for intraoperative or postoperative blood transfusion (p < 0.05). The need for ICU admission and development of a pulmonary complication were significantly more likely to fall into the extended LOH group (p < 0.05).
Conclusions: Several variables have been identified as significant predictors of LOH after PSF for NMS in the setting of a standardized discharge pathway. Patients with smaller curves and less complex surgeries were more amenable to accelerated discharge. Conversely, patients with severe CP with large curves and pelvic obliquity requiring longer surgeries with more blood loss may not be ideal candidates. These data can be used to inform providers' and families' post-operative expectations.
Level of Evidence: Therapeutic Level III.