The patterns of loss of correction after posterior wedge osteotomy in ankylosing spondylitis-related thoracolumbar kyphosis: a minimum of five-year follow-up

BACKGROUND Short-term studies have demonstrated good surgical outcomes after pedicle subtraction osteotomy (PSO) in ankylosing spondylitis (AS) patients, but there is a paucity of literature focused on middle-term results, especially regarding patterns of loss of correction. The objective of this study is to assess the durability of surgical outcomes and the patterns of loss of correction in thoracolumbar kyphosis secondary to AS following lumbar PSO with over 5-year follow-up. METHODS We performed a retrospective review of 155 consecutive AS patients undergoing lumbar PSO from January 2001 to December 2011. Twenty-four patients were included with an average follow-up of 6.9 years (range, 5-15 years). Radiographical evaluations included global kyphosis (GK), lumbar lordosis (LL), sagittal vertical axis, spinal-sacral angle, kyphosis of proximal non-fused segments (KPNS), angle of fused segments (AFS), osteotomized vertebra angle (OVA), distal intervertebral disc wedging (DIDW) and proximal junctional angle. Meanwhile, clinical outcomes were assessed by the Oswestry disability index (ODI) and Numerical rating scale (NRS). RESULTS The average correction per PSO segment was 34.9°. Significant improvement in sagittal parameters were found postoperatively, and no obvious deterioration was noticed during the follow-up. Mild loss of correction in GK (2.82°) and LL (3.77°) were observed at the final follow-up (P < 0.05). The KPNS and DIDW increased from 26° and -5.0° postoperatively to 30° and -2.2° at the final follow-up (P < 0.05), respectively. In contrast, no significant diminishment was identified in OVA and AFS (P > 0.05). The ODI and NRS improved significantly from 20.6 and 6.6 preoperatively to 5.9 and 2.3 at the final follow-up (P < 0.05). CONCLUSIONS PSO is an effective procedure for treating AS-related thoracolumbar kyphosis and can maintain sustained surgical outcomes during the middle-term follow-up. The loss of correction was mainly attributable to non-instrumented segments without fully ossified bridging syndesmophyte in the thoracolumbar region instead of instrumented levels.