OBJECTIVE The aim of the study was to evaluate, in a white population with chronic sarcoidosis, the rate and pattern of relapses, the correlated factors, and the course of disease after prednisone withdrawal. METHODS We have retrospectively examined the charts of 702 consecutive patients with histologically proven sarcoidosis, first seen in the Milan Sarcoidosis Clinic in the period October 1978-October 1994. 239 patients required corticosteroid therapy; in 82 it was possible to discontinue prednisone therapy and to have a follow-up of at least 18 months after withdrawal. RESULTS A relapse, requiring a new course of steroids, was observed in 30 (36.6%) of the 82 patients (R group). The other 52 patients (No-R group) did not relapse during a mean follow-up of 36.8 +/- 24.8 months (range 18-125). There were no relapses after 3 asymptomatic years of prednisone withdrawal. Extrapulmonary sarcoidosis was a reason for giving therapy in 46.6% of patients in the R group, vs 23.0% in the No-R group (P < 0.05). The first course of therapy lasted 22 months [median time; i.q. 11.5 to 34.5] in R group vs 26 months [i.q. 18 to 41] in No-R group (P > 0.05). The mean daily prednisone dose was higher in the R group: 17 mg [median value; i.q. 8.9 to 23.2] vs 10.6 mg [i.q. 8.1 to 13.8] in the No-R group (p < 0.05). Logistic regression confirmed the prognostic significance of mean daily prednisone dose and of extrapulmonary sarcoidosis at presentation (P < 0.01). A mild sarcoid activity at the time of withdrawal was still present in 51.9% of patients who did not relapse, and in 66.7% of patients who relapsed (p > 0.05). Relapse in the first year after withdrawal of prednisone therapy occurred in twenty-five of the 30 patients. The pattern of relapse was different from the initial manifestation in 5. Nine of the 30 patients could ultimately be weaned successfully from prednisone. CONCLUSIONS Relapses occurred in 36.6% of cases, and their pattern was the same as the initial manifestation in the majority of cases. A mild sarcoid activity at the time of withdrawal is not a reason for continuing steroids when the disease is abating. In our white population severe irreversible pulmonary impairment is rare, and even patients requiring chronic therapy need low prednisone dosage, usually around 10 mg daily, to control the disease in the late course.