Radiological morphology variances of osteotomized vertebra-disc complex following pedicle subtraction osteotomy for ankylosing spondylitis with thoracolumbar kyphosis: the incidence, mechanisms, and prognosis.

Abstract

Inaccurate osteotomy cut along with incomplete or even subluxated bone-on-bone closure of osteotomy gap following pedicle subtraction osteotomy (PSO) may be disastrous, hampering the lordosing effect and increasing the likelihood of complications. The inelastic yet osteoporotic spine in ankylosing spondylitis (AS) is specially predisposed to such suboptimal osteotomy, while the relevant data concerning this issue are scarce. This study aimed to analyze the incidence of radiological morphology variances (RMV) of osteotomized vertebra-disc complex (OVDC) following PSO in patients with kyphotic AS, conceptualize the mechanisms of the deviated morphology, and investigate the prognosis. This is a retrospective radiological data analysis. The sample being screened comprises 71 patients with thoracolumbar kyphotic AS who underwent single-level PSO at our hospital between March 2006 and February 2014. They were stratified by the presence of bridging syndesmophytes (BS) locating within the OVDC. Any irregular radiological configuration of OVDC other than the wedge morphology would be considered as RMV and were studied with care to fully describe and classify the spectrum of deviated morphologic features. Multiple spinopelvic sagittal parameters were measured to assess both the regional lordosing effect and the global realignment of sagittal spinal profile. For each selected patient with confirmed RMV, the radiological morphology was assessed, defined, and categorized. The prognosis involving surgical corrections and maintenance of spinopelvic sagittal parameters, as well as the remodeling in disordered osteotomized vertebral shape over time, were also investigated. The incidence of RMV was 21.9% in positive BS group (PG) and 30.8% in negative BS group (NG). Inappropriate angle and range of osteotomy accounted for the largest share (1 pts for PG and 10 pts for NG, 57.9%) of mechanisms responsible for RMV, followed by vertebral subluxation (VS) (5 pts for PG and 2 pts for NG, 36.8%) and failed osteotomy gap closure (1 pts for PG, 5.3%). For these patients, the mean bony lordosing effect per PSO segment was 36.0°±8.9° postoperatively, and decreased to 34.7°±8.7° by a mean follow-up of 3 years (p=.076). The magnitude of neighboring disc opening was significantly higher in NG (10.2°±6.5° vs. 2.4°±3.2°, p=.009). The global kyphosis and sagittal vertical axis were significantly corrected (77.0°±21.2° vs. 24.4°±18.8°; 160.6°±72.4° vs. 48.2°±38.6 mm, all p<.001) and remained stable by the ultimate follow-up (p>.05). No devastating neurologic deficits were noticed. Patients with VS and failed osteotomy gap closure exclusively showed solid bone healing and adaptive remodeling without rod breakage at final follow-up. Radiological morphology variances of OVDC were a high occurrence following PSO in AS, being mainly attributed to inaccurate osteotomy cut and VS. Neighboring disc opening and rotational or translational subluxation were major available remedial mechanisms strengthening the lordosing effect when that of vertebral wedging was impaired and insufficient. The bone fusion and remodeling concerning the subluxated or dislocated osteotomized vertebra was utterly favorable, maintaining the kyphosis correction and preventing instrumentation failure.