Abstract: BACKGROUND: Vertebral compression fracture is a common disease in the current orthopedic field. However, the occurrence of re-fracture in neighboring vertebrae after surgery is a problem that cannot be ignored, which has a serious impact on the normal life of patients. 
OBJECTIVE: The aim of this study is to establish four postoperative models with different recovery heights using computed tomography images. By using finite element analysis, we derived the stresses on the neighboring vertebrae at different recovery heights and further explored the importance of postoperative recovery of the height of the injured vertebrae.
METHODS: A finite element model of the thoracolumbar spine (T11-L3) was established and validated, on the basis of which four postoperative finite element models of L1 with different recovery heights of 100%, 80%, 60%, and 40% were constructed, in which the cement capacity varied with the recovery height. The specific models are as follows: Model 1 was the postoperative model with normal recovery height, and the cement capacity was 8.3 mL. Model 2 was the postoperative model in which 20% of the anterior height of the L1 was removed and the posterior convexity angle became 10.41°, and the cement capacity was 6.9 mL. Model 3 was the postoperative model in which 40% of the anterior height of the L1 was removed and the posterior convexity angle became 20.17°, and the cement capacity was 4.7 mL. Model 4 was a postoperative model with 60% of the L1 anterior height removed and the posterior convexity angle changed to 28.85°, with a cement capacity of 3.6 mL. For evaluation of the postoperative model, we applied a moment of 7 Nm and an axial force of 500 N. The followings were recorded and analyzed: peak stresses in the L2 upper endplate and T12 lower endplate; peak stresses in the L2 and T12 cancellous bone. 
RESULTS AND CONCLUSION: (1) The highest peak stresses for each condition of the L2 upper endplate, T12 lower endplate, L2 cancellous bone, and T12 cancellous bone occurred in Model 1 and Model 4. In particular, the T12 lower endplate, except for the posterior extension condition, the anterior flexion, left and right lateral bending, and left and right rotation conditions all reached their highest peak stresses in Model 4, with stresses of 50.3, 33.1, 44.9, 34.3, and 31.9 MPa. (2) Based on the peak stresses in the adjacent vertebral endplates and cancellous bone, after excluding Model 1 and Model 4, the minimum peak stresses for most of the conditions appeared in the Model 2, and the minimum peak stresses appeared in the Model 2 in 66.6% of the cases, especially in the upper endplates of the L2 and cancellous bone except for the posterior extension condition, the minimum peak stresses all appeared on the Model 2. (3) Therefore, controlling the recovery height at about 100% and 40% of the original height was a dangerous recovery height, which had a greater impact on the neighboring vertebrae. Controlling the recovery height at about 80% of the original height may be a more ideal choice. With a recovery height of about 80% of the original height, the adjacent vertebrae are subjected to less stress, thus reducing the risk of re-fracture of the adjacent vertebrae in the patient.

Key words: osteoporotic vertebral compression fracture, thoracic spine, lumbar spine, vertebral body strengthening, vertebral body recovery height, finite element analysis