Abstract: BACKGROUND: When percutaneous vertebral augmentation is performed for osteoporotic vertebral compression fractures, the four main types of cement dispersion type on the reinforced vertebrae are central separation, central integration, anterior separation, and uneven separation. At present, no relevant biomechanical studies have been found to compare the biomechanical effects of four cement dispersion types on the reinforced vertebrae. 
OBJECTIVE: To analyze the effects of cement dispersion types on biomechanical characteristics of the reinforced vertebrae using the three-dimensional finite element analysis method. 
METHODS: A three-dimensional finite element model of the thoracolumbar segment of thoracic 12 osteoporotic vertebral compression fracture was constructed, and then four models were obtained by simulating four cement dispersion types: central separation, central integration, anterior separation and uneven separation. The material property parameters, and boundary conditions were set, and loads (a vertical downward load of 260 N continuously on the T10 upper-end plate, and simultaneously, 10 N • m moments of flexion, extension, left and right bending, and left and right rotation) were applied, and data were imported to LS-DYNA software for a solution.
RESULTS AND CONCLUSION: (1) Under all loading conditions, the stress of the uneven separation model was the largest. The maximum stress of bone cement in the uneven separation model was 4.4, 8.6 and 4.8 times higher than that in the central separation model, the central integration model and the anterior separation model respectively. The maximum Von Mises stress of bone cement was the largest in the left flexion model with central separation, anterior separation and uneven separation. The maximum Von Mises stress of bone cement in the right flexion of the central integration model was the largest. The maximum Von Mises stress of bone cement was the smallest when all models were rotated left and right. (2) The maximum Von Mises stress at the cancellous bone cement interface of the uneven separation model was smaller than that of the other models when rotated from left to right. The maximum Von Mises stress at the cancellous bone cement interface of each model was similar under other working conditions. The maximum Von Mises stress at the cancellous bone cement interface was the largest during the left and right flexion of the four models. (3) The maximum Von Mises stress trends of upper endplates and lower endplates of T12 models were similar, and there was no significant difference under the same working conditions. Under all loading conditions, there was no significant difference in the maximum Von Mises stress of the T12 upper endplate and lower endplate between models. Under all loading conditions, there was no significant difference in the maximum Von Mises stress of T11/12 and T12/L1 intervertebral discs among the models. (4) When the displacement was 3.5 mm, the load of the central separation model was the minimum, which was 40 N. The load of the central integration and anterior separation models was 50 N. The load of the uneven separation model was the largest, which was twice that of the central separation model. (5) The results show that the uneven cement separation significantly increases cement stress, so percutaneous vertebral augmentation should be performed with the adequate and symmetrical distribution of cement in the vertebral fracture area and appropriately increased in the area of contact between the cement and cancellous bone interface. 

Key words: bone cement, osteoporotic, vertebral compression fracture, percutaneous vertebral augmentation, finite element analysis, bone cement dispersion type