Purpose
Pull-out of pedicle screw in posterior pedicle fixation for thoracic and lumbar burst fractures causes delayed rehabilitation, persistant pain, and imblance of sagittal plane. In this study we try to analyse the factors that cause the pull-out of pedicle screw.
Materials and Methods
From March 01, 2006 to December 31, 2009, we assorted into two group; Group I for pullout pedicle, Group II for control. Plane lateral x-ray view film of thoracolumbar spine was taken on preoperation, postoperation, the first time when screw was pulled out and last follow up. we measure inserted angle for the upper endplate of screw, convergency angle and change of body height loss and kyphotic angle. We analysed corelation between these measuring values and pedicle screw pull-out by Mann-Whitney test and T-test.
Results
Pull-out of pedicle screw was found at mean 5weeks among nine cases. For inserted pedicle screws, which place in upper and lower vertebral body of fractured one, Value of inserted angle for upper end plate and convergency angle was found non-significant(p>0.05, Mann-Whitney test). Restoration of height loss and kyphotic angle of fractured vertebral body was statically significant(p<0.05, T-test).
Conclusion
In posterior pedicle fixation for thoracic and lumbar burst fractures, sufficient restoration of height loss and kyphotic angle is important factor for prevention of screw pull-out than inserted angle for upper end plate and convergency angle at a short period of time. Therefore we think that sufficient anterior fixation of vertebral body and restoration of kyphotic angle have a decisive effect on prognosis of patients.

Introducing a 61-year old woman who was suffering from complicated traumatic thoracolumbar spine fractures, we contemplated the appropriate management algorithm. The Thoracolumbar Injury Classification and Severity score (TLICS) system is the latest and widely used scoring system by spine surgeons for thoracolumbar injuries (TLI). The originator of the system claims for easy application, high reproducibility, and direct link to a clinical decision-making algorithm. However, because of its simple and narrow boundaries, there are many limitations to apply the system in complicated situations. Besides, a fair number of TLI are caused by high velocity traumas, which mostly lead to complicated fractures and other medical conditions. For these reasons, practically, we also consider traditional and former concepts of TLI classifications. Furthermore, new algorithm should be suggested which includes not only the spine morphology and neurological manifestation but also comprehensive medical considerations of the patient.