Risk factors of kyphosis recurrence after implant removal in thoracolumbar burst fractures following posterior short-segment fixation

Purpose Our aim was to evaluate the results of short-segment pedicle instrumentation with screw insertion in the fracture level and find factors predicting kyphosis recurrence in thoracolumbar burst fractures. Methods We retrospectively analysed 122 patients with thoracolumbar burst fracture who were divided into two groups: kyphosis recurrence and no kyphosis recurrence. Pre-operative radiographic data comprising Cobb angle (CA), regional angle, anterior vertebra height (AVH), upper intervertebral angle, vertebral wedge angle (VWA), pre-anteroposterior A/P approach, superior endplate fracture, load-sharing classification (LSC) score and clinical data including age, visual analogue scale (VAS) score, thoracolumbar injury classification and severity score were compared between groups. T test, Pearson's chi-square and multivariate logistic regression were calculated for variables. Results CA, VWA and AVH were significantly corrected after surgery. CA changed from 23.7 to 3.0 (p < 0.001), VWA from 38.7 to 9.6 (p < 0.001) and AVH from 48.8 % to 91.2 % (p < 0.001). These parameters were well maintained during the follow-up period with a mild, tolerant loss of correction. Neurological function and pain were significantly improved without deterioration. Age, pre-A/P and pre-AVH < 50 % influenced kyphosis recurrence (p = 0.032, 0.026, 0.011, respectively). Conclusions Short-segment pedicle instrumentation including the fractured vertebra was effective in treating thoracolumbar burst fractures. The loss of correction at follow-up after implant removal was associated with age, A/P ratio and anterior vertebral height < 50 %.