Polymyositis-dermatomyositis and malignant lesions: does an association exist?

To examine the purported association between polymyositis-dermatomyositis and malignant lesions, we identified all patients with histopathologically proven polymyositis-dermatomyositis examined at our institution from 1965 to 1974. For each patient, a control subject was selected, matched for age, sex, race, year and department of clinical evaluation, and geographic area of residence. Although the proportion of patients who ever had a diagnosis of malignant disease (25%) was greater than that for controls (17%), the difference was not statistically significant (P = 0.14). The slight excess of cancer seen among patients was contributed by the most distant referrals and may best be explained on the basis of referral bias. No clinical differences were found between patients with polymyositis-dermatomyositis who had a malignant neoplasm and those who did not, and intensive clinical investigation to exclude "occult" underlying malignant growths did not seem to be cost-effective. Although the present study does not prove that there is no association between polymyositis-dermatomyositis and malignant lesions, the existence of a clinically significant relationship seems unlikely.