BACKGROUND: According to widely used criteria (Bohan and Peter criteria, 1975), dermatomyositis (DM) is differentiated from polymyositis (PM) only by skin changes. More recent criteria also include histopathologic characteristics enabling the distinction between PM and DM and the differentiation of sporadic inclusion body myositis (s-IBM) from PM. The authors investigated the applicability of diagnostic features for diagnosing PM and DM. METHODS: The authors performed a retrospective follow-up study of 165 patients with 1) a previous diagnosis of myositis; 2) subacute onset of symmetric, proximal weakness; and 3) an evaluation between 1977 and 1998 excluding other neuromuscular disorders. RESULTS: The diagnoses at initial evaluation based on clinical, laboratory, and histopathologic criteria were PM, 9 (5%); DM, 59 (36%; 54 isolated, 3 with associated connective tissue disease [CTD], 2 with associated malignancy); unspecified myositis (perimysial/perivascular infiltrates, no PM or DM), 65 (39%; 38 isolated myositis, 26 with associated CTD, 1 with malignancy); and possible myositis (necrotizing myopathy, no inflammatory infiltrates), 32 (19%; 29 isolated myositis, 3 with associated CTD). At follow-up evaluation, five of the nine patients with PM had typical s-IBM features. None of the remaining four patients complied with the assumed typical signs of PM. Ten of the 38 patients with isolated unspecified myositis had been diagnosed with a CTD. CONCLUSIONS: Polymyositis is an overdiagnosed entity. At evaluation, more than half the patients with autoimmune myositis cannot be specifically diagnosed with polymyositis or dermatomyositis. A quarter of patients with isolated unspecified myositis subsequently developed connective tissue disease.