High-Grade 3-Column Osteotomies Are Not Interchangeable

BACKGROUND: Pedicle subtraction osteotomy (PSO) and vertebral column resection (VCR) are powerful techniques for correcting severe spinal deformities. Although PSO has been proposed as a viable alternative to VCR, their comparative efficacies and safety profiles require further elucidation. METHODS: This single-center retrospective study analyzed 169 patients (mean age, 22 years; 84 male; 169 ethnic Chinese) with severe kyphoscoliosis who underwent primary corrective surgery via PSO (n = 85) or VCR (n = 84). Radiographic parameters, surgical data, intraoperative neuromonitoring (IOM) changes, Scoliosis Research Society (SRS)-22 scores, and complications were compared between groups. RESULTS: Both techniques significantly improved all radiographic parameters and SRS-22 scores (p < 0.001). The VCR group demonstrated superior correction of the major curve (65.5% versus 56.9%, p = 0.003), segmental kyphosis (68.1% versus 61.5%, p = 0.03), and apical vertebral rotation (48.5% versus 34.4%, p = 0.001). At the critical osteotomy stage, 105 (62.1%) of 169 patients experienced IOM signal decline. The neurological complications rate was significantly higher in the VCR group (13 of 84 versus 7 of 85, p = 0.038), as was the overall complication rate (43 of 84 versus 29 of 85, p = 0.008). Each 1° increase in correction achieved with VCR was associated with a 1.6% higher risk of complications (OR = 1.016, p = 0.045). CONCLUSIONS: Although both PSO and VCR were highly effective for major deformity correction, VCR provided a greater magnitude of correction in the coronal, sagittal, and axial planes. However, this advantage was counterbalanced by a significantly higher risk of complications, including neurological deficits. Despite the frequent reversibility of IOM signal declines, VCR retained a higher risk profile due to its higher overall complication rate. LEVEL OF EVIDENCE: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.