Objective: To assess the correlation between the Denis classification and clinical outcomes.

Subjects and Design: We performed a retrospective study of 87 patients with spinal injuries in the thoracolumbar region. All patients were admitted to the Oswestry Spinal injuries unit between Jan 1990 and December 1998. Following a review of their notes, CT scans and radiographs, we attempted to classify their injuries according to the Denis (3 column) Classification of spinal injuries.

Outcome Measures: The patients were assessed both at the time of presentation and on subsequent follow up. Neurological function was assessed using the Frankel classification.

Results: The results of the study show that the correlation between Denis classification and clinical outcome is poor. In addition, the relative proportions of the two most common Major Injury types described by Denis were reversed in our study with Burst fractures forming the majority of injuries. This difference in out come was attributed primarily to the increased use of CT scanning in our study. It appears that Denis misdiagnosed a significant number of burst (two column) fractures as compression (anterior column) fractures.

Conclusions: Our findings showed no correlation between the degree of instability and the number of columns disrupted. We believe that only 3 column fracture dislocations are fundamentally unstable. In addition, our results support the practise of treating vertebral fractures by conservative means with no apparent correlation between treatment modality and neurological outcome at long term follow up.

We reviewed a series of 53 patients with closed traumatic complete injuries of the cervical spinal cord. They were admitted within two days to a spinal injuries centre, treated conservatively by six weeks of bedrest and skull traction, then mobilised in a neck support for six weeks. Eight patients had temporary neurological deterioration, four spontaneously and four after cervical manipulation; seven of these recovered to the initial neurological level without surgery. Of 40 patients followed for more than 12 months, 19 recovered useful motor power in local muscles which were initially paralysed (zonal recovery); one patient showed distal motor recovery. Zonal recovery did not correlate with the mechanism of skeletal injury or with the degree of residual canal stenosis. Sensory sparing and an initial neurological level higher than the level of skeletal injury were both good prognostic signs for zonal recovery.