It would appear that in carefully preselected circumstances, dual Harrington distraction rods may be successfully utilized in a large percentage of thoracolumbar fractures. Mechanistic classification of these fractures offers insight into the extent of soft-tissue disruption. As the circumferential soft-tissue damage increases, the degree of spinal stability obtained with distraction fixation decreases. This is especially true when considering the anterior longitudinal ligament. In compression flexion and vertical compression, the anterior structures are predominantly intact, and thus distraction fixation offers excellent stability. Utmost care must be offered to hook placement and stability, particularly in the proximal area. In the torsional flexion and lateral flexion groups, complex mechanisms of soft-tissue disruption require special consideration when distraction rods are to be employed. Likewise, analysis of soft-tissue and bony disruption in the distractive extension, distractive flexion, and translation injury groups suggests that serious consideration be given to spinal stability prior to the routine use of distraction rods in these patients. Surgically, the posterior approach does not preclude removal of bony fragments from the canal, either by rod distraction and reduction alone or via canal exploration. Below L1, laminectomy with direct access to the fragments is possible. Above L1, pedicle resection and costotransversectomy have been eminently successful procedures in fractures less than 3 weeks old. None of these surgical procedures preclude a subsequent anterior approach if necessary. In no instance was a patient neurally damaged in our series by this surgical protocol. We are loath to allow only internal support during the fusion maturation period and strongly urge external support during this phase. Removal of the rods prior to solid fusion is to be discouraged, as this may predispose to persistent pain and return of deformity. CT scan appears to represent an essential tool preoperatively and postoperatively, particularly in complete neural lesions. We have gleaned some presumptive evidence relative to posterior superior body fragment position, pedicle integrity and position, and possible posterior longitudinal ligament integrity. With the information available through radiographs and CT, we are in the process of attempting to determine prospectively where posterior distraction rods alone may suffice. Thus, we would hope to circumvent the concomitant need for laminotomy, pedicle resection and fragment removal, or reduction or anterior corpectomy. It is hoped that this information will be available soon.