• BACKGROUND
    • The management of thoracolumbar (TL) burst fractures is still controversial. The thoracolumbar injury classification and severity score (TLICS) algorithm is now widely used to guide clinical decision making, however, in clinical practice, we come to realize that TLICS also has its limitations for treating patients with total scores less than 4, for which conservative treatment may not be optimal in all cases.
  • PURPOSE
    • The aim of this study is to identify several risk factors for the failure of conservative treatment of TL burst fractures according to TLICS algorithm.
  • METHODS
    • From June 2008 to December 2013, a cohort of 129 patients with T10-l2 TL burst fractures with a TLISC score ≤3 treated non-operatively were identified and included into this retrospective study. Age, sex, pain intensity, interpedicular distance (IPD), canal compromise, loss of vertebral body height and kyphotic angle (KA) were selected as potential risk factors and compared between the non-operative success group and the non-operative failure group.
  • RESULTS
    • One hundred and four patients successfully completed non-operative treatment, the other 25 patients were converted to surgical treatment because of persistent local back pain or progressive neurological deficits during follow-up. Our results showed that age, visual analogue scale (VAS) score and IPD, KA were significantly different between the two groups. Furthermore, regression analysis indicated that VAS score and IPD could be considered as significant predictors for the failure of conservative treatment.
  • CONCLUSION
    • The recommendation of non-operative treatment for TLICS score ≤3 has limitations in some patients, and VAS score and IPD could be considered as risk factors for the failure of conservative treatment. Thus, conservative treatment should be decided with caution in patients with greater VAS scores or IPD. If non-operative management is decided, a close follow-up is necessary.