Biomechanical evaluation of posterior instrumentation for lumbar burst fracture: comparison of two internal devices

Purpose of the study Burst fractures generally occur due to trauma to the thoracolumbar spine. Surgery is indicated for unstable fractures. Posterior instrumentation with pedicular screws is generally proposed. In certain circumstances, hooks may be preferred due to excessive risk of insertion of the pedicular screw. The purpose of this study was to compare two posterior instrumentations, one using pedicular screws on either side of the fracture each protected by hoods and a second composed of the same pedicular screws inserted under the fracture hooks above. Material and methods Twelve spinal specimens from human cadavers composed of segments T10 to L2 were used. Range of flexion, extension, lateral inclination, and rotation were noted on T10 up to application of 7 Nm. Spinal segments were tested first intact, then in four configurations: 1) instrumented without lesion, 2) lesion simulating burst fracture of L1 without section of the interspinous ligament, 3) and with section of the interspinous ligament, and 4) with L1 corporectomy. Finally a test to rupture was performed by applying a flexion moment up to fracture. Results Mean flexion-extension of the instrumented spine was limited compared with the intact spine for both instrumentation configurations and irrespective of the lesion. The same behavior was observed for lateral inclination with less pronounced motion with the first instrumentation. For rotation, the range of motion increased clearly with the second instrumentation and this with the first lesion while with the first instrumentation, rotation amplitude remained below that of the intact spine. There was however an increase in the vertical displacement during flexion-extension for both instrumentations. For the rupture test, the mean flexion moment at rupture was 14.4 Nm (10.6-22 Nm) with no difference between the two instrumentations. Discussion This mode simulating burst fractures of the spine appears to be reproducible and more realistic than corporectomy. Attention should be taken concerning the limits of this type of study since fractures can occur for forces as small as 10.6 Nm. Thus we observed that pedicle screw configurations and also fractures produced mean ranges of motion greater than intact segments irrespective of the type of lesion simulated. However, the net increase in motion was observed during rotation movements when hooks were used, even when they were placed only below the fracture. Putting pressure on the hooks does not prevent them from slipping along the lamina. But neither of these two configurations controls the fracture gap. A vertebral reinforcement might be necessary.