Does solid fusion eliminate rod fracture after pedicle subtraction osteotomy in ankylosing spondylitis-related thoracolumbar kyphosis?

Abstract Background Context Rod fracture (RF) has a negative impact on the surgical outcome of patients with ankylosing spondylitis (AS) after lumbar pedicle subtraction osteotomy (PSO). However, there is a paucity of published studies analyzing the risk factors for RF in PSO-treated patients with AS with thoracolumbar kyphosis. Purpose The objective of this study was to investigate the risk factors for RF after PSO for thoracolumbar kyphosis secondary to AS. Study Design/Setting This is a retrospective single-center study. Patient Sample Patients with AS who underwent PSO for thoracolumbar kyphosis between January 2002 and December 2016 were included. Outcome Measures Demographic data, including age, sex, body mass index, and smoking status, were summarized. The surgical data analyzed included the levels of osteotomy, the fusion levels, the upper instrumented vertebra, the lower instrumented vertebra, the osteotomy site, the rod material, the rod diameter, and the rod contour angle (RCA). Radiographic parameters included the sagittal vertical axis, thoracic kyphosis, lumbar lordosis, sacral slope, pelvic tilt, and pelvic incidence. Radiographic parameters were measured at baseline, immediately after the operation, and at the final follow-up. Adequate ossification of the anterior longitudinal ligament (ALL) at the PSO level was defined by a total bony bridge. Adequate ossification of the ALL was also measured at baseline, immediately after the operation, and at the final follow-up. Methods Patients with a minimum of 2 years' follow-up or patients who developed RF were enrolled in the study. Recruited patients were divided into the RF group and the no-RF group based on whether they developed RF. Patient demographics, operative data, radiographic parameters, and adequate ossification of the ALL were analyzed to determine the risk factors for RF. For patients with RF, the fusion status at the PSO level, the time course to the development of RF, the site of RF, and the corresponding solution were also recorded. Results Rod fracture occurred in 11 (8.9%) of the 123 recruited patients. Solid fusion at the PSO level was found in all patients in the RF group. The average duration to the onset of RF was 31.4 months (range, 12–68 months). All RFs occurred at or immediately adjacent to the PSO level. The RCA was greater in the RF group than in the no-RF group (27.8° vs 22.9°, p=.031). A greater proportion of patients with a rod diameter of 5.50 mm were found in the RF group than in the no-RF group (100.0% vs 68.8%, p=.033). There was a larger proportion of patients with adequate ossification of the ALL at the final follow-up visit in the no-RF group than in the RF group (67.0% vs 27.3%, p=.018). Multivariate analyses demonstrated that the RCA (odds ratio, 1.174; 95% confidence interval, 1.018–1.354; p=.028) and adequate ossification of the ALL at the final follow-up visit (odds ratio, 0.079; 95% confidence interval, 0.014–0.465; p=.005) were independent factors for RF. Notably, revision surgery was performed among six patients, whereas conservative treatment was used for the remaining five patients. Conclusions In patients with AS after PSO for thoracolumbar kyphosis with solid fusion at the PSO level, the incidence of RF was 8.9%. Rod diameter was identified as a risk factor for RF. Furthermore, the RCA was identified as an independent risk factor for RF. In contrast, adequate ossification of the ALL around the PSO level at the final follow-up visit was identified as an independent protective factor for RF.