The authors instituted a prospective, randomized study of patients presenting with acute burst fractures of the thoracolumbar and lumbar spine. Patients were alternately treated by posterior distraction using pedicle instrumentation (AO fixateur interne) or anterior decompression and instrumentation (Kostuik-Harrington device). Forty patients are presented with a mean follow-up of 20 months. Preoperatively, there was significant canal compromise in 39 patients. This measured 44.5% in those patients undergoing posterior surgery and 58% in those patients undergoing anterior surgery. Postoperatively, this was reduced to 16.5% and 4%, respectively. There is a statistically significant difference between these two groups (P less than 0.0001). The mean preoperative kyphotic deformity was 18.7 degrees in those patients treated by anterior surgery and 18.2 degrees in the group treated by posterior surgery. At last follow-up, the mean improvement in kyphotic deformity was 9.3 degrees in the anterior group and 11.3 degrees in the posterior group. There is no statistically significant difference between these two groups. There were two implant failures of the anterior Kostuik-Harrington construct and two implant failures of the AO fixateur interne. Blood loss was significantly higher in the patients undergoing anterior surgery, but there were no complications from thoracotomy and anterior decompression of the dural sac. This study supports the hypothesis that posterior distraction instrumentation can effectively decompress the canal and correct kyphosis in patients sustaining burst-type injuries. Anterior surgery, however, results in a more complete and reliable decompression of the canal.