ABSTRACT
Objective:To investigate the related factors of vertebral body height reloss after pedicle screw fixation of thoracolumbar fracture and to determe the optimum prediction point.Methods:A retrospective case control study was made on 215 patients with thoracolumbar fracture admitted to Second Affiliated Hospital of Soochow University from January 2010 to December 2017. There were 155 males and 60 females，aged 21-80 years［（48.6±10.4）years］. According to Denis fracture classification，there were 73 patients with compression fractures（type A in 15 patients，type B in 51，type C in 7），135 burst fractures（type A in 28 patients，type B in 87，type C in 20）and flexion distraction fractures（type A in 4，type B in 2，type C in 1）. All patients were treated by pedicle screw fixation. Follow-up lasted for 12- 48 months［（23.8±8.2）months］. Vertebral body height loss occurred in 86 patients（loss group），but did not in 129 patients（non-loss group）. The two groups were compared concerning sex，age，osteoporosis self-assessment tool for Asians（OSTA），body mass index（BMI），fracture types，number of fractured vertebrae，preoperative sagittal Cobb angle，preoperative degree of vertebral compression，number of screws placed in injured vertebrae，extent of vertebral reset and other related factors. Univariate analysis was used to identify the correlation of those factors with vertebral body height reloss. Multivariate Logistic regression analysis was performed to identify the independent factors for the height reloss with the receiver operating characteristic curve（ROC）and area under the curve（AUC）calculated to evaluate the optimum point in prediction of vertebral height reloss.Results:The two groups showed no significant differences in sex，age，BMI，fracture types，number of injured vertebrae，preoperative sagittal Cobb angle and number of screws placed in injured vertebrae（ P>0.05），but the differences were statistically significant in OSTA，preoperative degree of vertebral compression and extent of vertebral reset（ P<0.05）. According to the univariate analysis，OSTA，preoperative degree of vertebral compression and extent of vertebral reset were significantly correlated with the occurrence of vertebral body height reloss（ P<0.05）. According to the multivariate Logistic regression，OSTA（ OR=1.109，95% CI 0.527-0.685， P<0.05）and preoperative degree of vertebral compression（ OR =0.038，95% CI 0.539-0.689， P<0.05）were significantly related to vertebral body height reloss. The AUC relating OSTA and preoperative degree of vertebral compression to vertebral body height reloss was 0.604 and 0.614，respectively. The optimum prediction point of OSTA and preoperative degree of vertebral compression for vertebral body height reloss was 1.9 and 31.3%，respectively. Conclusions:OSTA and the preoperative degree of vertebral compression are independent risk factors for vertebral body height reloss. OSTA≤1.9 or preoperative degree of vertebral compression ≥31.3% indicates a significantly higher risk of postoperative vertebral body height reloss.