摘要Thoracolumbar fractures are most common in spine fractures. Most thoracolumbar fractures could cause kyphotic deformity. Proper treatment is necessary even though the surgical methods remain controversial. Aims of surgery are to decompress the spinal canal, restore vertebral body height, correct the kyphotic deformity and accomplish firm and stable fixation for the rehabilitation. Posterior short-segment pedicle instrumentation ( SSPI ) is widely used in the treatment of thoracolumbar fractures. However, kyphosis recurrence after the surgical treatment is not uncommon. It has been reported that the loss of correction is primarily attributed to the intervertebral space collapse, especially the upper intervertebral space. The additional insertion of two screws in the fractured vertebra increases the rate of being stiff and reduces the failure rate of SSPI. It will be of great value if kyphosis recurrence after SSPI can be preoperatively predicted by plain radiographs. Therefore, the purpose of this review is to investigate the factors predicting the kyphosis recurrence after SSPI in treating thoracolumbar fractures. Based on the studies discussed in this review, we conclude that SSPI with screw insertion into the fractured vertebra is effective in treating thoracolumbar fractures with lower kyphosis recurrence rate compared with SSPI. This review also suggests that preoperative A / P ( anteroposterior ratio ) ratio, preoperative AVH ( anterior vertebra height ) < 50%, age and gender ( female ) are associated with kyphosis recurrence after SSPI. However, the true relations between UIVA ( upper intervertebral angle ), preoperative LSC ( load-sharing fracture classification score ), preoperative BMI ( body mass index ) and kyphosis recurrence after SSPI are still controversial. Preoperative VWA ( vertebral wedge angle ), preoperative VAS ( visual analogue scale ), preoperative TLICS ( thoracolumbar injury classification and severity score ) may have few effects on kyphosis recurrence after SSPI. More multi-center, prospective randomized studies are required to confirm whether these or any other factors are associated with kyphosis recurrence.