The effect of high orally administered doses of prednisone for 1 week early in the course of an acute exacerbation of asthma incompletely responsive to bronchodilators was examined in 41 patients randomly assigned to receive either prednisone or an identical appearing placebo. All 22 of the patients who received prednisone improved during the week of treatment, although one had a subsequent exacerbation 5 days after discontinuing the study medication. Of the 19 who received placebo, eight required rescue intervention (P = less than 0.004) in association with continued symptoms, increased frequency of metered-dose inhaler use, and decreased pulmonary function; the other 11 improved at about the same rate as those who received prednisone. Although the mean initial FEV1 was suggestively lower among those who did not improve and required intervention, there was considerable overlap with those who improved spontaneously, and no reliable distinguishing characteristics were found at entry into the study that could serve as predictors of those who would or would not improve spontaneously. There were no clinically important adverse effects from the prednisone. Because continued symptoms of asthma often result in emergency care or hospitalization, these data support early intervention with orally administered prednisone for acute exacerbations that do not respond fully to bronchodilators, at least in those patients with a prior history of a protracted course or emergency care.