Study design: Retrospective study. Objective: To report the incidence of revision surgery due to mechanical failure following pedicle subtraction osteotomy (PSO) in adult spinal deformity (ASD) patients. Summary of background data: PSO allow major surgical correction of ASD, although; the risk of mechanical complications remains considerable. Previous reports have been based on smaller cohorts or multicenter databases and none have utilized competing risk (CR) survival analysis. Methods: All ASD patients undergoing PSO surgery from 2010 to 2015 at a single, tertiary institution were included. Demographics, long standing radiographs as well as intra- and postoperative complications were registered for all. A CR-model was used to estimate the incidence of revision surgery due to mechanical failure and two predefined multivariable models were used to assess radiographic prediction of failure and reported as odds ratios (OR) with 95% confidence intervals (95% CI). Results: A total of 171 patients were included with 2-year follow-up available for 91% (mean [IQR]: 35 [24–50] months). Mechanical failure occurred in 111 cases (65%) at any time in follow-up, the most frequent being rod breakage affecting 81 patients (47%). Cumulative incidence of revision surgery due to mechanical failure was estimated to 34% at 2 years and 58% at 5 years. A multivariable proportional odds model with death as competing risk showed significantly increased odds of revision with fusion to the sacrum (OR: 5.42; 95% CI 1.89–15.49) and preoperative pelvic tilt (PT) > 20° (OR: 2.41; 95% CI 1.13–5.16). History of previous surgery, number of instrumented vertebra, as well as postoperative SRS-Schwab modifiers and Global Alignment and Proportion score were not associated with significant effects on odds of revision. Conclusions: In a consecutive single-center cohort of patients undergoing PSO for ASD, we found an estimated incidence of revision surgery due to mechanical failure of 34% 2 years postoperatively. Fusion to the sacrum and preoperative PT > 20° were associated with elevated risks of revision. Level of evidence: Prognostic III.