The clinical characteristics and pathology of sarcoidosis are well defined; however, the optimal therapy for this disorder remains unclear. Although patients respond, acutely, to corticosteroid therapy, it is not clear that these agents ultimately alter the natural history of this disease. These observations and that corticosteroids have significant side effects suggest that only patients who will clearly benefit from corticosteroid therapy should be treated. In a prospective study of patients' with sarcoidosis (n = 98), we limited our use of corticosteroids to those patients who had objective evidence of recent deterioration in lung function or serious extrapulmonary disease. All patients with sarcoidosis fulfilling these criteria were treated with corticosteroids. Patients were tapered off corticosteroids after they were treated for 1 yr. Of the 98 study subjects, 91 had not received therapy for the disease and 7 were on therapy before entry into the study. Of the 91 previously untreated patients, 55 were observed without therapy and 36 were treated with corticosteroids. Of those who were observed off therapy, only eight deteriorated. Of these latter patients, six responded and stabilized with the administration of corticosteroids for treatment of the underlying disease, to antibiotics for an associated bronchiectasis, or to diuretics for treatment for congestive heart failure; two were lost to follow-up. None of these six patients deteriorated while receiving corticosteroids. Of the 36 patients who deteriorated and were treated with corticosteroids, 20 remained stable and 16 improved clinically. Of the 37 patients who were eventually tapered off corticosteroids, five deteriorated and required reinitiation of corticosteroid therapy.(ABSTRACT TRUNCATED AT 250 WORDS)