Objective: This paper aims to differentially depict potential patterns of the loss of correction in surgically treated thoraco-lumbar burst fractures. These may eventually serve to foreseeing and even forestalling loss of correction. Methods: The study focused on 253 patients with surgically treated thoraco-lumbar fractures. This cohort of patients was clustered in four subgroups according to the fracture spine segment (T11-L1 or L1-L2) and surgery type (short segment fi xation or anterior approach). Relevant recorded and processed data were the fracture level, post-operative (Kpo) and last follow-up (Kf) kyphosis angle values. Correlation, regression and determination testing were performed for the last follow-up kyphosis angle and post-operative kyphosis angle, and regression equations were determined for each subgroup of patients.
Results:The patterns of loss of correction were described through the following equations: K f = 0.95*K po + 3.2° for the T11-L1 level fractured vertebrae treated by posterior short segment fi xation; K f = 0.98*K po + 3.4° for the L1-L2 level fractured vertebrae treated by posterior short segment fi xation; K f = 1.1*K po + 1.6° for the T11-L1 level fractured vertebrae treated by anterior approach; and K f = 0.7*K po + 2.8° for the L1-L2 level fracture vertebrae treated by anterior approach.
Conclusions:The loss of correction may be predicted, to a certain extent, for thoraco-lumbar fractured vertebrae treated surgically. The bestfi t equations depicted for both type of surgery (short segment fi xation and anterior approach) and both spinal segments (T11-L1 and L2-L3) are signifi cantly different than the equations delineated for the collapse of non-surgically treated fractures.