Three-dimensional translations following posterior three-column spinal osteotomies for the correction of severe and stiff kyphoscoliosis

Abstract Background Context Posterior three-column spinal osteotomies were shown to be effective to treat severe and stiff kyphoscoliosis. Translations at the site of osteotomy after deformity correction were commonly seen intraoperatively, which might cause potential neurologic deficits. However, this phenomenon was not thoroughly discussed in the current literature. Purpose This study aimed to evaluate the three-dimensional (3D) translations at the three-column osteotomy site and their effects on neurologic outcome in the surgical correction of severe and stiff kyphoscoliosis. Study Design/Setting A retrospective study was carried out. Patient Sample Sixty-nine patients treated by posterior three-column spinal osteotomy for severe kyphoscoliosis of idiopathic, congenital, neuromuscular, neurofibromatosis, and tuberculosis origin were included. Outcome Measures General, coronal, and sagittal translations were graded three-dimensionally according to the theory of Meyerding. Methods The charts of 69 clinical patients with severe and stiff kyphoscoliosis treated by posterior three-column osteotomy from January 2013 to June 2015 were reviewed. There were 35 male patients with an average age of 21.5 years and 34 female patients with an average age of 22.5 years. The etiologies of these spinal deformities were idiopathic, congenital, neuromuscular, neurofibromatosis, and tuberculosis. According to our classification system of spinal cord neurologic function, there were 41 type A, 13 type B, and 15 type C cases. The 3D spine models were reconstructed from thin-sliced computed tomography (CT) scan, and the 3D translations at the three-column osteotomy site were graded and analyzed. Results The incidences of general translation (GT), frontal translation (FT), and sagittal translation (ST) were 62.3%, 52.2%, and 26.1%. The incidence of evoked potential (EP) change in cases with GT/FT being or more than grade II (GT, 42.9%; FT, 50.0%) was significantly higher than that with GT/FT being less than grade II (GT, 16.7%; FT, 18.2%), whereas the incidence of EP change in cases with ST being or more than grade I (33.3%) was significantly higher than that with ST being less than grade I (9.8%). No linear correlations were found between spine shortening distance, deformity correction rate, and the degree of translation. Conclusions The 3D translations are common in posterior three-column spinal osteotomies regardless of anterior strut graft placement. The increase of translation will increase neurologic risks, with GT or FT less than grade II and ST less than grade I being relatively safe.