Purpose: To determine the results and safety of patients undergoing|spinal cord level (SCL) pedicle subtraction osteotomy (PSO) for thetreatment of thoracic kyphosis.

Methods: Retrospective chart and radiographic review of 25 patients with severe thoracic kyphosis.

Results: The underlying diagnoses were: tumour (8), Scheuermann’s Kyphosis (4), degenerative/osteoporosis (3), fracture (3), inflammatory|(2), neurofibromatosis (2), congenital kyphosis (1), tuberculosis (1), and infected tumour (1). The osteotomy was combined with a lumbar PSO in|five patients. Three patients were treated with double thoracic|osteotomies. Two PSOs were extended transdiscally to debride the|infected disc. The mean focal PSO correction was 33.6° (range 9°–73°). The overall thoracic kyphosis measured from T5 to T12 improved from a|mean of 58.3° preoperatively to 37.1° postoperatively. Estimated blood|loss ranged from 400cc to 12500cc. All patients presenting with spinal|cord dysfunction neurologically improved postoperatively. There were 2 major neurological complications. One patient developed postoperative | progressive paraplegia following a prolonged period of intra- and | postoperative severe hypotension and coagulopathy. The other developed a| pseudoarthrosis five months postoperatively and suffered an incomplete|spinal cord injury during the subsequent revision. Other complications included: T3 radicular pain (1) -resolved; dural tears (2); respiratory failure -prolonged ICU admission(1); fractures proximal to the thoracic (2) and distal to lumbar (1) instrumentation; incomplete corrections of the sagittal alignment despite double osteotomies (2); wound breakdown associated with preoperative radiation (1).|

Conclusions: SCL-PSO is a feasible option for severe thoracic kyphosis. This procedure eliminates the need for anterior surgery; however, it does not reduce the potential for significant morbidity.