In an effort to determine the optimal emergency therapy for acute episodes of asthma, we randomly, assigned 102 acute ill patients to 60 minutes of treatment with inhaled isoproterenol alone, isoproterenol plus intravenous aminophylline or isoproterenol plus a single oral dose of an elixir of theophylline. Patients requiring treatment beyond this time were given an injectable sympathomimetic agent in addition. The combination of isoproterenol and a methylxanthine was not found to be better than isoproterenol alone, and the route of administration of methylxanthine was not an important determinant of either the serum theophylline level or the therapeutic response. A major variable that influenced the duration of therapy needed to produce a remission was the severity of the obstruction at presentation. Persons whose initial 1-second forced expiratory volumes were less than 30 percent of predicted and who did not improve 35 percent or more to at least 40 percent of predicted at the end of 60 minutes of intense treatment were those who ultimately required prolonged emergency room therapy and/or hospital admission for control of their symptoms. Thus, simple objective assessment of the degree of impairment at presentation coupled with the response to initial treatment will serve to identify early a high-risk group of asthmatic patients in whom the usual emergency room therapeutic modalities will often prove ineffective.