The reported results of compression fractures are poor. These results are not influenced by the severity of compression, the fracture site or the residual deformity. Otherwise, the factors that determine a patient's recovery are unknown. This study wants to identify the factors determining a patient's recovery after surgical treatment of compression fractures of the thoracolumbar spine. Therefore, in 31 surgically treated patients the pre-injury versus the 12-month follow-up differences in back pain, in global outcome and in participation were prospectively recorded. For this, the visual analogue scale for pain (VAS scale) and the Greenough and Fraser low back outcome scale were used. Of the latter scale, the 3 questions pertaining to participation were combined to create a participation subscale. For these differences and for time lost from work multiple linear regressions with combinations of 16 possible predictors were performed.

At one year patients who smoke report a 25% less favorable global outcome and return 2.8 points (out of 10) less closely to their pre-injury pain level than patients who do not smoke. Patients with a fracture at the thoraco-lumbar junction return 3.3 points less closely to their pre-injury level on the VAS scale than those with a lumbar fracture. For each decrease in 1 of the 3 education levels, the patients stay away from work 15 weeks longer. Per degree of sagittal index at follow-up, patients stay 9 days longer at home. For each increase in level of occupation the return to the pre-injury participation level is 10% less favorable. The variability of time lost from work and of recovery of global outcome, pain and participation level explained by our models is 73%, 37%, 25% and 13% respectively.

Smoking, localization of the fracture at the thoraco-lumbar junction and a high pre-existent level of occupation are strong negative predictors for recovery. A lower education level and sagittal fracture deformity at follow-up are negative predictors for sick leave but might also reflect the concerns of the physician when deciding about return to work.

Aims: To throw a new light on the fragmentary information from litera-ture, to add information to the mechanism of this injury, to clarify the cloudy treatment indications in the group between 4 and 10 years and to report the outcome of conservative treatment after a mean of 80 months. Materials: After reviewing the þles and X-rays, 11 children were included in this series, according to strict criteria. The decision for closed reduction depended on the severity of the deformity, on the associated lesions and on the age of the patient. At follow-up, they were subjected to a thorough anamnesis and clinical evaluation with speciþc concern about pain, function and cosmesis. Comparative X-rays to evaluate the remodeling were made in a standardized way. Results: The mean age at the trauma was 7 years. (4 to 12) In 3 patients, the fracture was caused by a transverse force. The diagnosis was missed 3 times. Five patients were simply put in a plaster cast, in 6 others closed reduction was performed þrst. In 8 patients a residual curve was accepted. After the age of 6, spontaneous remodeling was poor, with a bad cosmetic result in 1 case (residual curve of 11û) and a functional problem in at least 1 other case. Conclusions: 1. Contrary to literature, these injuries can be caused by a transverse force as well. 2. Spontaneous remodeling is far less than generally accepted. 3. Curves > 10û need reduction with an adapted technique from as early as 7 years of age on.