OBJECTIVE To determine indications for thoracolumbar radiography. METHODS Case series with prospective data collection. METHODS Level I trauma center. METHODS Blunt-trauma victims more than 12 years old who underwent routine thoracic and lumbar radiography according to institutional protocol. Patients were classified as group 1, not able to be evaluated clinically (Glasgow Coma Scale score of less than 13, intoxication, intubation, or cervical neurologic deficit); and group 2, able to be evaluated clinically. RESULTS Twenty-four of 319 patients sustained 25 thoracic or lumbar fractures. Seven of 136 group 1 patients and 17 of 183 group 2 patients had fractures. Eight of 17 patients with pain and 9 of 17 with tenderness had fractures (P = .001). No group 2 patients without pain, tenderness, thoracic or lumbar neuro-deficit, or major distracting injury, including cervical fracture, had fractures. The negative predictive value of pain and tenderness was 95%. Five of 46 patients with spinal fractures at any level had multiple fractures. CONCLUSIONS Blunt-trauma victims who cannot be evaluated clinically should undergo thoracolumbar radiography. Routine radiography may be unnecessary in asymptomatic patients who can be evaluated clinically and who do not have neurologic deficits or distracting injuries. Spinal fracture at any level mandates complete spinal radiography.