STUDY DESIGN: Retrospective matched cohort analysis. SUMMARY OF BACKGROUND DATA: Posterior-only vertebral column resection (P-VCR) is a potential alternative to combined anterior/posterior vertebrectomy (A-P/VCR) for the treatment of severe spinal deformity. OBJECTIVE: To examine a matched cohort of adult and pediatric patients with severe spinal deformity treated with A/P-VCR versus P-VCR. METHODS: Databases of 2 spine surgeons at 1 institution from 1994 to 2007 were reviewed. Patients were matched based on age at surgery (within 10 years), diagnosis, curve pattern, vertebrae resected (within 1), levels of vertebrae resected (within 2), levels fused (within 5), and minimum 2-year follow-up. A total of 34 P-VCR patients were identified who appropriately matched 34 A/P-VCR patients. The etiology of the deformity and type of curve were matched directly so that they were identical for each matched pair. The remainder of the inclusion parameters was matched as closely as possible between the 2 groups according to the criteria listed above. RESULTS: Final coronal Cobb correction P-VCR versus A/P-VCR showed that 52.6% versus 53.9% (p = .8) was similar, whereas P-VCR final sagittal Cobb correction was superior: 53.0% versus 40.0% (p = .017). The P-VCR group had a significantly shorter total operative time (p = .002) and total length of stay (p = .003). Complications rates were similar and relatively infrequent for both P-VCR and A/P-VCR, including wound infections requiring operative intervention, subsequent revision surgery, and transient motor deficits. Total Scoliosis Research Society scores improved from preoperative to final follow-up for both P-VCR (p = .007) and A/P-VCR (p = .07) groups. CONCLUSIONS: Posterior-only vertebral column resection is a challenging yet safe and effective means of treating severe scoliosis and/or kyphosis. Compared with an A/P-VCR for severe spinal deformity, P-VCR demonstrated shorter operative time and hospital stay, as well as improved sagittal correction and Scoliosis Research Society scores.