Study design: Retrospective study.
Objectives: This study is to measure and analyze the changes of three-dimensional (3D) distances of spinal column and spinal canal at the three-column osteotomy sites and address their clinical and neurologic significance.
Summary of background data: Three-column osteotomies were developed to treat severe and stiff spine deformities with insufficient understanding on the safe limit of spine shortening and the relationship between the shortening distance of the spinal column and that of the spinal canal.
Methods: Records of 52 continuous patients with severe and stiff scoliosis treated with three-column spine osteotomies at our institution from July 2013 to June 2015 were reviewed. The preoperative spinal cord function classification were type A in 31 cases, type B in 10 cases, and type C in 11 cases. The types of osteotomies carried out were extended pedicle subtraction osteotomy in nine patients and posterior vertebral column resection in 43 patients. Multimodality neuromonitoring strategies were adopted intraoperatively. 3D pre- and postoperative spine models were reconstructed from the computed tomography (CT) scans. The distances of convex and concave spinal column and the spinal canal shortening were measured and analyzed.
Results: The spinal column shortening distance (SCSD) measured on the 3D models (27.8 mm) were statistically shorter than those measured intraoperatively (32.8 mm) (P < 0.05); however, they were strongly correlated statistically (r = 0.82). The central spinal canal shortening distance (CCSD) was significantly shorter than the convex SCSD (P < 0.05). The convex SCSD and CCSD were significantly shorter in cases with anterior column strut graft than in those with bone-on-bone fusion (P < 0.05).
Conclusion: The shortening distance of the convex spinal column cannot represent that of the central spinal canal in patients with severe scoliosis. The spinal column shortening procedure in appropriately selected patient groups with bone-on-bone fusion is a viable option with the CCSD being significantly shorter than the convex SCSD.
Level of evidence: 4.