STUDY DESIGN.: A retrospective clinical study. OBJECTIVE.: To find the corrective capacity of a thoracic pedicle subtraction osteotomy (PSO), determine if segmental correction is dependent on level, and to compute the impact of thoracic PSO on regional and global spinal balance. SUMMARY OF BACKGROUND DATA.: PSO is a technique popularized in the lumbar spine primarily for the correction of fixed sagittal imbalance. Despite several studies describing the clinical and radiographic outcome of lumbar PSO, there is no study in literature reporting its application in the thoracic spine. METHODS.: We retrospectively analyzed patients with fixed thoracic kyphosis who underwent thoracic PSOs for sagittal realignment. Segmental pedicle screw instrumentation and intraoperative neurophysiologic monitoring was used in all patients. Data acquisition was performed by reviewing medical charts and radiographs to determine sagittal correction (segmental/regional/global) and complications. Clinical outcome using the Scoliosis Research Society-22 (SRS-22) instrument was determined by interview. RESULTS.: A total of 25 thoracic PSOs were performed (mean: 1.7 PSOs/patient, range: 1-3) in 15 patients (9 M/6 F). The study population had an average age of 56 years (range, 36-81 years) and was followed up after surgery for a mean of 3.5 years (range, 24-75 months). The osteotomies were carried out in the proximal thoracic spine (T2-T4, n = 6), midthoracic spine (T5-T8, n = 12), and distal thoracic spine (T9-T12, n = 7). Mean correction at the PSO for all 25 levels was 16.3 degrees +/- 9.6 degrees . Stratified by region of the spine, thoracic PSO correction was as follows: T2-T4 = 10.7 degrees +/- 15.8 degrees , T5-T8 = 14.7 degrees +/- 4.6 degrees , and T9-T12 = 23.9 degrees +/- 4.1 degrees . Mean thoracic kyphosis (T2-T12 Cobb angle) was improved from 75.7 degrees +/- 30.9 degrees to 54.3 degrees +/- 21.4 degrees resulting in a significant regional sagittal correction of 21.4 degrees +/- 13.7 degrees (P < 0.005). Global sagittal balance was improved from 106.1 +/- 56.6 to 38.8 +/- 37.0 mm yielding a mean correction of 67.3 +/- 54.7 mm (P < 0.005). One patient, in whom there was segmental translation during osteotomy closure, had a decline in intraoperative somatosensory-evoked potentials. No patient sustained a temporary or permanent neurologic deficit after surgery. The mean SRS-22 Questionnaire score at final follow-up was 82.4 +/- 10.2. CONCLUSION.: Thoracic PSO can be performed safely. Segmental sagittal correction appears to vary based on the region of the thoracic spine the PSO is performed. The distal thoracic segments, which more closely resemble lumbar segments in morphology, rendered the greatest sagittal correction after PSO, approximately 24 degrees . There was no case of neurologic injury associated with thoracic PSO, and clinical outcomes according to the SRS-22 instrument were generally favorable.