Shoulder pain is a common complaint and shoulder hyperostosis a frequent radiological condition. However, little is known about the association between the clinical and radiological findings. To evaluate the clinical relevance of shoulder hyperostosis we performed a controlled, blind study of 99 hospitalized probands with and without thoracospinal hyperostosis on lateral chest X-rays. The study included grading of the shoulder hyperostosis on the basis of three bilateral standard radiographs, assessing shoulder pain in a standardized way by an interviewer and recording extraskeletal causes of shoulder pain. The prevalence of shoulder hyperostosis was doubled in probands with thoracospinal hyperostosis compared to controls (chi 2 = 5.90, P less than 0.025, n = 99). Shoulder hyperostosis, irrespective of thoracospinal hyperostosis, predisposed to shoulder pain (40% versus 18%, chi 2 = 4.06, P less than 0.05, n = 74). Shoulder hyperostosis in combination with thoracospinal hyperostosis (shoulder DISH) predisposed to shoulder pain to an even greater extent (46% versus 12%, chi 2 = 6.64, P less than 0.01, n = 47). We conclude that shoulder hyperostosis is a radiological finding of potential clinical relevance.