Often chronic obstructive pulmonary disease (COPD) patients treated for acute exacerbations receive intravenous (IV) aminophylline in addition to inhaled bronchodilators that may raise serum levels of theophylline into the toxic range. A double-blind, randomized study of 52 men with COPD who came to the emergency department for treatment of exacerbations was initiated to establish the efficacy and safety of this common practice. After history and physical examination, patients were treated with 28% oxygen by Venturi mask and 0.3 cc metaproterenol sulfate in 2.5 cc saline by nebulizer; an IV line was started and patients received either aminophylline or D5W. Measurements included baseline and two-hour serum theophylline levels, pulmonary function tests, and symptom questionnaires. Mean values from the entire group showed decreases in respiratory rate, cardiac rate, and pulsus paradoxus, and increases in forced expiratory volume in one second (FEV1) and vital capacity (VC) over a two-hour treatment period (P less than .01). Despite the increase in serum theophylline in the treatment group, the demographic, clinical, pulmonary function, and outcome data were found to have no statistically significant differences when compared to control patients. The data were then analyzed according to serum theophylline levels. Theophylline level greater than 20 micrograms/mL occurred in 15 patients with no untoward effects; premature ventricular contractions (PVCs) were no more frequent in this group than in those with lower serum theophylline levels. A theophylline level greater than 10 micrograms/mL after two hours of treatment resulted in the following differences, which were not statistically significant: mean FEV1 response less than or equal to 10 micrograms/mL vs greater than 10 micrograms/mL, 20% vs 28%; mean VC change, 17% vs 30%; or mean emergency department returns in one week, 0.1 vs 0.26. In our experience, oxygen and inhaled metaproterenol are effective treatment for exacerbations of COPD.