Failure patterns and related risk factors of sagittal reconstruction following pedicle subtraction osteotomy in patients with ankylosing spondylitis and thoracolumbar kyphosis

OBJECTIVE The aim of this study was to analyze the specific patterns and risk factors of sagittal reconstruction failure in ankylosing spondylitis (AS)-related thoracolumbar kyphosis after pedicle subtraction osteotomy (PSO). METHODS A retrospective study was performed in patients with AS and thoracolumbar kyphosis after lumbar PSO with a minimum follow-up of 2 years. Patients were classified as having successful realignment (group A), inadequate correction immediately postoperatively (group B), and sagittal decompensation during follow-up (group C) according to the immediately postoperative and latest follow-up sagittal vertical axis (SVA). Radiographic parameters and clinical outcomes were collected. Pelvic tilt (PT) was used to assess the magnitude of pelvic backward rotation. Hip structural damage and ossification of the anterior longitudinal ligament (ALL) at the proximal junction, PSO level, and distal junction were also evaluated on radiographs. RESULTS Overall, 109 patients with a mean age of 35.3 years were included. Patients in both group B (n = 16) and group C (n = 13) were older than those in group A (n = 80) (mean ages 43.6 vs 32.9 years, p < 0.011; and 39.2 vs 32.9 years, p = 0.018; respectively). Age (OR 1.102, p = 0.011), and preoperative PT (OR 1.171, p = 0.041) and SVA (OR 1.041, p = 0.016) were identified as independent risk factors of inadequate correction. Additionally, a higher distribution of patients with adequate ALL ossification at the PSO level was found in group B than in group A (37.5% vs 22.5%, p = 0.003). Age (OR 1.101, p = 0.011) and preoperative SVA (OR 1.013, p = 0.020) were identified as independent risk factors of sagittal decompensation. Furthermore, compared with group A, group C showed a higher distribution of patients with severe hip structural damage (15.4% vs 0, p = 0.018) and higher incidences of rod fracture (RF) (38.5% vs 8.8%, p = 0.011) and pseudarthrosis (15.4% vs 0, p = 0.018). Additionally, the incidence of RF (19.6% vs 6.9%, p = 0.045) and changes in the proximal junctional angle (0.5 degrees vs 2.2 degrees, p = 0.027) and the distal junctional angle (0.3 degrees vs 2.2 degrees, p = 0.019) were lower during follow-up in patients with adequate ALL ossification than in those without adequate ossification. CONCLUSIONS Sagittal reconstruction failure in patients with AS could be attributed to inadequate correction immediately after surgery (14.7%) and sagittal decompensation during follow-up (11.9%). Adequate ALL ossification was a risk factor of inadequate correction. However, adequate ALL ossification could decrease the development of RF and relieve the junctional kyphotic change during follow-up. Older age and greater baseline SVA were independent risk factors for both inadequate correction and sagittal decompensation.