ObjectiveTo compare the effectiveness of short segmental pedicle screw fixation with and without fusion in the treatment of thoracolumbar burst fracture. MethodsA retrospective analysis was made on the clinical data of 57 patients with single segment thoracolumbar burst fractures, who accorded with the inclusion criteria between February 2012 and February 2014. The patients underwent posterior short segmental pedicle screw fixation with fusion in 27 cases (fusion group) and without fusion in 30 cases (non-fusion group). There was no significant difference in gender, age, cause of injury, time between injury and admission, fracture segment and classification, and neurologic function America Spinal Injury Association (ASIA) classification between 2 groups, which had the comparability (P > 0.05). The operative time, blood loss, and hospitalization days were compared between 2 groups. The height of the injured vertebra, the kyphotic angle, and the range of motion (ROM) were measured on the X-ray film. The functional outcomes were evaluated by using the Greenough low-back outcome score and the visual analogue scale (VAS) for back pain. The neurologic functional recovery was assessed by ASIA grade. ResultsThe operative time was significantly shortened and the blood loss was significantly reduced in the non-fusion group when compared with the fusion group (P < 0.05), but no significant difference was found in hospitalization days between 2 groups (P > 0.05). The patients were followed up for 2.0-3.5 years (mean, 3.17 years) in the fusion group and for 2-4 years (mean, 3.23 years) in the non-fusion group. X-ray films showed that 2 cases failed bone graft fusion, the fusion time was 12-17 weeks (mean, 15.6 weeks) in the other 25 cases. Complication occurred in 2 cases of the fusion group (1 case of incision deep infection and 1 case of hematoma at iliac bone donor site) and in 1 case of the non-fusion group (fat liquefaction); primary healing of incision was obtained in the others. The Cobb angle, the height of injured vertebrae showed no significant difference between 2 groups at pre-operation, immediate after operation, and last follow-up (P > 0.05). The ROM of injured vertebrae showed no significant difference between 2 groups at 1 year after operation (before implants were removed) (P > 0.05). The implants were removed at 1 year after operation in all cases of the non-fusion group, and in 11 cases of the fusion group. At last follow-up, the ROM of injured vertebrae in the non-fusion group was significantly higher than that in the fusion group (P < 0.05), but no significant difference was found in Greenough low-back outcome score, VAS score, and ASIA grade between 2 groups (P > 0.05). ConclusionFusion is not necessary when thoracolumbar burst fracture is treated by posterior short segmental pedicle screw fixation, which can preserve regional segmental motion, shorten the operative time, decrease blood loss, and eliminate bone graft donor site complications.