The use of serial chest radiographs to assess disease activity in patients with sarcoidosis is controversial. However, reliance on the symptomatic clinical course to assess disease activity may be misleading. As many patients being treated with corticosteroids have an abrupt clinical deterioration when doses of those medications are decreased, we questioned whether the chest radiograph could depict alterations in disease activity as measured by spirometry in this subset of patients. We retrospectively reviewed the clinical course of all patients with pulmonary sarcoidosis in whom the corticosteroid dose was reduced during a 6-month period. The 15 patients without fever, chills, or purulent sputum during that time were then examined to determine the presence (n = 10) or absence (n = 5) of a symptomatic relapse. All patients who had a symptomatic relapse also had a fall in forced vital capacity of at least 10%, suggesting an increase in disease activity. Serial chest radiographs were evaluated during and after corticosteroid dose reductions and after clinical recovery on higher steroid doses in the patients who had had a relapse. In eight patients, the disease was in radiographic stage 2 (hilar adenopathy and parenchymal lung disease); in seven patients it was in radiographic stage 3 (parenchymal lung disease alone). The disease did not change stage in any patient during the study. Chest radiographs worsened more frequently in patients who had a clinical relapse (seven of 10) than in those who did not have a relapse (zero of five, p less than .05). An alveolar chest radiographic pattern (n = 4) or reticulonodular pattern (n = 3) was noted in the seven patients who had a relapse, with worsening on radiographs often occurring before detection of relapse by symptomatology (four of seven) or spirometry (three of seven). Spirometry and radiographs improved or stabilized after an increase in corticosteroid dose in all 10 patients who had a relapse. We conclude that serial chest radiographs can reflect clinical relapse in patients with sarcoidosis during corticosteroid dose reduction. Furthermore, worsening seen on chest radiographs may be the first evidence of relapse.