The plain radiograph in ophthalmology: a wasteful and potentially dangerous anachronism.

The indications for 822 consecutive referrals for skull radiography were prospectively studied in a large eye hospital over a one year period. In 85.9% of patients the results were normal, and in 89% of the remainder they had no positive effects on management; all patients in whom a 'beneficial' effect could be identified would have been more appropriately investigated by other means. Fourteen of 25 patients whose skull radiographs were normal were shown by computed tomography or magnetic resonance imaging to have orbital or intracranial lesions. Views of the optic canals, orbits or paranasal sinuses were also requested in 336 patients. With appropriate use of alternative imaging methods, no patient's treatment would have been adversely affected if none of the skull radiographs had been obtained.