Purpose of study: Because of the smaller size of the thoracic pedicle, primarily in the coronal plane, most surgeons have opted to avoid the used of transpedicular screws in the thoracic spine. A new technique for placement of screws through thoracic pedicles with and without lateral wall violation is described.Methods used: From April 1992 through February 1999, 1,064 pedicle screws were placed within the thoracic spine of 137 patients. Primary diagnoses included scoliosis (59), fracture (41), kyphosis (16), tumor (12), pseudarthrosis (6), ankylosing spondylitis (2) and infection (1). Small posterior laminotomies were made at each level. A ball probe was used to identify the medial and inferior wall of the pedicle at each location. In general, 5-mm diameter screws of 30 to 35 mm in length were placed using a two-person technique. Screws were placed in a lateral to medial direction with the entry site being just lateral to the boundaries of the pedicle over the posterior laminar ridge. It was not unusual for the lateral wall of the pedicle to be violated. However, because of articulation of the rib to the transverse process, pedicle and vertebral body, the screws that were placed remained within the bony articulation without violating surrounding soft tissues. Through the laminotomies performed at each level, all walls of the pedicle with the exception of the lateral walls were directly inspected and ensured not to be violated by instrumentation. After placement of all screws, anteroposterior and lateral radiographs were taken to confirm appropriate placement and alignment.of findings: Mean follow-up was 64 months (range, 24 to 106 months). Construct failure included four misplaced screws, five loose screws, three broken rods and two loose nuts for a total implant complication rate of 1.3% (12 of 1,064). There were no vascular injuries or intrathoracic/pulmonary injuries. There was one transient thoracic radiculopathy with complete resolution, one transient posterior cord syndrome with subsequent resolution and retention of hardware and one transient partial cord injury after attempted reduction of complex scoliosis resulting from probable traction and vascular insult. With the exception of one transient thoracic radiculopathy, there were no identified neurologic deficits directly resulting from screw placement within the thoracic spine. The transient neurologic complication rate was 2.2% (3 of 137).Relationship between findings and existing knowledge: A new technique for placement of transpedicular screws through a posterior approach in the thoracic spine is described. It is a safe and reliable technique with low morbidity. Screws can be placed larger than the size of the width of the pedicle without neurologic or vascular injury while generally maintaining adequate purchase and construct stability.Overall significance of findings: Use of screws placed posteriorly in the thoracic spine is attractive in that improved segmental control can be obtained over that of the standard hook or wire construct.Disclosures: Device or drug: pedicle screws. Status: approved.Conflict of interest: No conflicts.