This is a prospective, randomized study to compare the efficacy of two similar "long-segment" Texas Scottish Rite Hospital instrumentations with the use of hooks in the thoracic spine and pedicle screws versus laminar hook claw in the lumbar spine for thoracolumbar A3, B, and C injuries. Forty consecutive patients with such thoracolumbar fractures (T11-L1) associated with spinal canal encroachment underwent early operative postural reduction and stabilization. The patients were randomly sampled into two groups: Twenty patients received hooks in "claw configuration" in both the thoracic and the lumbar spine (group A), and 20 patients received hooks in the thoracic vertebrae and pedicle screws in the lumbar vertebrae (group B). Pre- and postoperative plain roentgenograms and computed tomography scans were used to evaluate any changes in Gardner post-traumatic kyphotic deformity, anterior and posterior vertebral body height at the fracture level, and spinal canal clearance (SCC). All patients were followed for an average period of 52 months (range 42-71 months). The correction of anterior vertebral body height was significantly more (P < 0.01) in the spines of group B (33%) than in group A (16%), with a subsequent 11% loss of correction at the latest evaluation in group A and no loss of correction in group B. There were no significant differences in the changes of posterior vertebral body height and Gardner angle between the two groups. The SCC was significantly more (P < 0.05) immediately postoperatively in the spine of group B (32%) than in group A (19%). In the latest evaluation, there was a 9% loss of the immediately postoperatively achieved SCC in group A, while SCC was furthermore increased at 10.5% in group B. All patients with incomplete neurologic lesions in groups A and B were postoperatively improved at 1.1 and 1.7 levels, respectively. There were two hook dislodgements in the thoracic spine, one in each group, while there was no screw failure in group B. There was neither pseudarthrosis nor neurologic deterioration following surgery. Visual Analog Pain Scale and Short Form-36 scores were equally improved and did not differ between the two groups. The use of pedicle screws in the lumbar spine to stabilize the lowermost end of a long rigid construct applied for A3, B, and C thoracolumbar injuries was advantageous when compared with that using hook claws in the lumbar spine because the constructs with screws restored and maintained the fractured anterior vertebral body height better than the hooks without subsequent loss of correction and safeguarded postoperatively a continuous SCC at the injury level.
