Does the load-sharing classification predict ligamentous injury, neurological injury, and the need for surgery in patients with thoracolumbar burst fractures?

Object The load-sharing score (LSS) of vertebral body comminution is predictive of results after short-segment posterior instrumentation of thoracolumbar burst fractures. Some authors have posited that an LSS > 6 is predictive of neurological injury, ligamentous injury, and the need for surgical intervention. However, the authors of the present study hypothesized that the LSS does not predict ligamentous or neurological injury. Methods The prospectively collected spinal cord injury database from a single institution was queried for thoracolumbar burst fractures. Study inclusion criteria were acute (< 24 hours) burst fractures between T-10 and L-2 with preoperative CT and MRI. Flexion-distraction injuries and pathological fractures were excluded. Four experienced spine surgeons determined the LSS and posterior ligamentous complex (PLC) integrity. Neurological status was assessed from a review of the medical records. Results Forty-four patients were included in the study. There were 4 patients for whom all observers assigned an LSS > 6, recommending operative treatment. Eleven patients had LSSs ≤ 6 across all observers, suggesting that nonoperative treatment would be appropriate. There was moderate interobserver agreement (0.43) for the overall LSS and fair agreement (0.24) for an LSS > 6. Correlations between the LSS and the PLC score averaged 0.18 across all observers (range −0.02 to 0.34, p value range 0.02–0.89). Correlations between the LSS and the American Spinal Injury Association motor score averaged −0.12 across all observers (range −0.25 to −0.03, p value range 0.1–0.87). Correlations describing the relationship between an LSS > 6 and the treating physician's decision to operate averaged 0.17 across all observers (range 0.11–0.24, p value range 0.12–0.47). Conclusions The LSS does not uniformly correlate with the PLC injury, neurological status, or empirical clinical decision making. The LSSs of only one observer correlated significantly with PLC injury. There were no significant correlations between the LSS as determined by any observer and neurological status or clinical decision making.