Impacts of resection level and baseline deformity features on upper lumbar pedicle subtraction osteotomy closure ratio in ankylosing spondylitis-related thoracolumbar kyphosis

Abstract

Abstract Background: Pedicle subtraction osteotomy (PSO) is an unique technique widely applied in the correction of thoracolumbar kyphosis arise from ankylosing spondylitis (AS). During the osteotomy closure procedure, both the upper part of the osteotomized vertebra and the lower part of that contributed to the spinal realignment jointly. However, the relative proportion of truncal and pelvic sagittal correction which occurs with the closure of PSO has not been detailed described in AS. This study aims to investigate whether the upper lumbar osteotomy level and baseline deformity characteristics impact the distribution of global correction. Methods: 73 AS patients who underwent single-level lumbar PSO were reviewed. Radiographic parameters included sagittal vertical axis (SVA), spino-sacral angle (SSA), global kyphosis (GK), spinal tilt (ST) and T1 pelvic angle (TPA), thoracic kyphosis (TK), lumbar lordosis (LL), pedicle subtraction angle (PSA), osteotomized vertebra angle (OVA), pelvic incidence (PI), pelvic tilt (PT), and sacral slope (SS). The truncal and pelvic closures were calculated by the change of superior or inferior endplate angle of the osteotomized vertebra. Besides, in accordance with Diebo’s project, all patients were divided into three groups based on the ratio of osteotomy closure (ROC). The impact of preoperative radiographic parameters and upper lumbar PSO level on the truncal versus pelvic closures were also investigated. Results: No significant difference between ROC and osteotomy level was observed (P=0.155>0.05). Furthermore, there were significant differences in SVA (P<0.001), PI (P<0.001), PT (P<0.001) and ST (P<0.001) among three groups. Preoperative SVA (r=0.581, P<0.001) was positively correlated with ROC while the PI (r=-0.510, P<0.001), PT (r=-0.547, P<0.001) and ST (r=-0.517, P<0.001) were negatively correlated with ROC. Significant improvement of all the spinopelvic parameters except TK and PI were noted postoperatively. The correction of SVA (r=0.692, P<0.001) and ST (r=0.629, P<0.001) was positively correlated with ROC, whereas the variation of PT (r=-0.570, P<0.001) and SS (r=-0.461, P<0.001) was negatively correlated with ROC. Conclusion: The osteotomy level is not the driver in the ratio of osteotomy closure but the baseline deformity features. Specifically, AS patients with preoperative larger PT realized a proportionally greater pelvic closure (improvement of pelvic retroversion) and less correction of SVA. Additionally, it’s important to get more pelvic closure as much as possible when closing the osteotomy gap for AS patients with higher PI.