The learning curve for performing three-column osteotomies in adult spinal deformity patients: one surgeon's experience with 197 cases

BACKGROUND CONTEXT: Three-column osteotomy (3CO) is used to correct rigid adult spinal deformity. It presents risk of complications because it involves extensive osseous resection and spinal destabilization. PURPOSE: Our purpose was to characterize the learning curve for performing 3CO in adult spinal deformity patients. DESIGN: Retrospective review. PATIENT SAMPLE: A surgical registry at a tertiary care center was used to identify 238 cases of 3CO for correction of adult spinal deformity by 1 surgeon between 2005 and 2014. Patients with at least 1 year of clinical and radiographic follow-up were included (n=197; mean duration of follow-up, 43 months; range, 12–121). OUTCOME MEASURES: We quantified associations between surgeon experience and (1) estimated blood loss per vertebral level fused (EBL/VLF), (2) incidence of new neurologic deficits, (3) incidence of reoperation for instrumentation failure, (4) operative time in minutes, and (5) magnitude of correction at the level of the osteotomy. METHODS: The learning curve for binary outcomes was demonstrated using a LOWESS smoother plot of the probability of occurrence. Change in risk was calculated using a generalized linear model with link identity and binomial family. The learning curve for continuous variables was demonstrated using a scatter plot and a line of best fit based on linear regression analysis. Alpha=0.05. RESULTS: EBL/VLF decreased by a mean of 19.7 mL (95% confidence interval [CI]: 11.3–28.1) with each 10 cases (decrease of 388 mL/level fused by the end of the study period). The risk of a neurologic deficit declined by 7.98% (95% CI: 7.98%, 7.99%) with every 100 cases. The risk of reoperation declined by 1.99% (95% CI: 0.83%, 3.17%) with every 10 cases until the 100th case. After that point, there was no significant change in the probability of reoperation (p>.05). The magnitude of correction and operative time did not change with increasing surgeon experience (p>.05). CONCLUSION: Incidence of reoperation for instrumentation failure, incidence of new neurologic deficits, and estimated blood loss improved with increasing surgeon experience at performing 3CO. Most outcomes, except the risk of reoperation, improved through the last case.