The prognosis of sarcoidosis is difficult to establish and it depends mainly on the persistence of activity over time and the degree of functional impairment of the involved organs. The aim of this study was to identify factors predicting persistence of disease activity at diagnosis. In a 14-year period (1974-1987), 209 patients were diagnosed with sarcoidosis at Bellvitge Hospital, a 1,000-bed teaching institution in Barcelona, Spain. One hundred ninety-three patients were followed up and included in the study. Clinical and radiological data were collected at diagnosis and a definition of disease activity was established. A Cox proportional-hazards regression model identified the following variables as independently influencing the persistence of activity: absence of erythema nodosum (risk ratio, RR = 2.37; 95% confidence interval, CI: 1.54-3.66), pulmonary infiltrates in chest x-ray (RR = 1.89, 95% CI: 1.28-2.8), splenomegaly (RR = 3.67, 95% CI: 1.46-9.23), age > or = 40 years (RR = 1.01, 95% CI: 1.006-1.03), and absence of lymphadenopathy in chest x-ray (RR = 2.26, 95% CI: 1.08-4.77). We suggest that the identification of factors predicting persistence of sarcoidosis activity at diagnosis may help to establish the prognosis of the disease and therefore improve the therapeutic approach.