Factors Associated with Use of Bone Morphogenetic Protein During Pediatric Spinal Fusion Surgery

Background Our goal was to investigate whether the use of recombinant human bone morphogenetic protein (rhBMP, or BMP) during pediatric spinal fusion surgery has been increasing and how patient, surgical, and hospital characteristics influence BMP use. Methods Using the Nationwide Inpatient Sample database, we identified 4817 children eighteen years old or younger who had undergone spinal fusion surgery with the use of BMP from 2003 through 2009. A multivariate logistic regression model, the Z-test of proportions, and simple linear regression were used for statistical analysis (significance, p < 0.05). Results There was a 3.4-fold increase in BMP use, from 2.7% in 2003 to 9.3% in 2009—an average 16% per year increase (p < 0.01). For each additional year of age, BMP use increased 1.09-fold (odds ratio [OR]: 1.05 to 1.13, p < 0.01). Compared with BMP use for adolescent idiopathic scoliosis, the adjusted odds of BMP use were increased 1.3-fold for congenital scoliosis (OR: 1.02 to 1.76, p = 0.04), 2.8-fold for thoracolumbar fractures (OR: 2.1 to 3.8, p < 0.01), and 5.0-fold for spondylolisthesis (OR: 3.9 to 6.3, p < 0.01). Patients with private insurance were 1.5-fold more likely to receive BMP (OR: 1.2 to 1.9, p < 0.01). Patients in whom autograft bone was used intraoperatively were 0.63-fold less likely to receive BMP (OR: 0.52 to 0.77, p < 0.01). The rate of BMP use was 0.38-fold lower in teaching hospitals (OR: 0.31 to 0.48, p < 0.01) and 1.7-fold higher in hospitals with a large bed capacity (OR: 1.3 to 2.2, p < 0.01). Compared with hospitals located in the Northeast, those in the West had a 1.7-fold increase (OR: 1.3 to 2.4, p < 0.01) and those in the South had a 2.0-fold increase (OR: 1.5 to 2.7, p < 0.01) in the odds of BMP use. Conclusions Use of BMP during pediatric spinal fusion has increased significantly. Patient factors (age, diagnosis, and insurance), surgical factors (autograft use), and hospital factors (teaching status, bed capacity, and location) influenced the variation in BMP use.