Variability in rates of arthrodesis for patients with thoracolumbar spine fractures with and without associated neurologic injury

STUDY DESIGN. Retrospective cohort study using a large clinical database. OBJECTIVE. Assess hospital-based rates of thoracolumbar spine arthrodesis within the United States for patients with a thoracolumbar spine fracture. SUMMARY OF BACKGROUND DATA. Substantial variation has been documented in per capita rates of elective lumbar spinal arthrodesis. Similar data regarding rates of arthrodesis for traumatic thoracolumbar injuries have not been reported. METHODS. Data from the Health Care Cost and Utilization Project Nationwide Inpatient Sample from 1998 to 2002 was used. ICD-9-CM codes were used to identify patients with thoracolumbar vertebral fractures with and without associated neurologic injury. Hospitals were grouped by teaching status and volume of thoracolumbar fracture patients and compared for rates of arthrodesis. Arthrodesis rates for the 25 highest fracture volume hospitals in the database were compared individually. RESULTS. A total of 24,098 patients with thoracolumbar fracture were identified: 91.7% had a thoracolumbar fracture without neurologic injury, while 8.3% had a thoracolumbar fracture with neurologic injury. Overall, 9.1% of thoracolumbar fracture patients without neurologic injury underwent arthrodesis, while 61.4% of thoracolumbar fracture patients with neurologic injury underwent arthrodesis. For all patients, rates of arthrodesis at high-volume hospitals were significantly greater than arthrodesis rates at lower-volume hospitals. Similarly, urban teaching hospitals had higher rates of arthrodesis than those at urban nonteaching hospitals. Among the 25 highest-volume hospitals in our sample, spinal arthrodesis rates for patients without associated neurologic injury ranged from 4% to 23%, a 5.8-fold variation. Among fracture patients with neurologic injury, arthrodesis rates ranged from 50% to 91%, a 1.8-fold variation. CONCLUSION. Hospital teaching status and spine fracture volume affected rates of spine arthrodesis in thoracolumbar fracture patients with and without neurologic injury. Variability in fusion rate for thoracolumbar spine trauma appears to be lower than that reported for elective spine procedures, especially in the presence of a neurologic injury.