Eleven patients with chronic obstructive pulmonary disease (age, 61 ± 2 yr; FEV1, 1.36 ± 0.24 L, 46 ± 7% predicted) were given 4 wk of treatment with either a conventional low dose of inhaled terbutaline (LDT), 500 μg four times a day, or a high dose of inhaled terbutaline (HDT), 2,000 μg four times a day, delivered by a spacer. A randomized double-blind crossover design was used with 2-wk run-in and washout periods, when ipratropium bromide was substituted for inhaled beta-agonists. Dose response curves (DRC) to cumulative doubling doses of inhaled terbutaline (125 to 4,000 μg) were constructed after each treatment period, and baseline spirometry, finger tremor (Tr), plasma potassium (K), plasma cAMP, and ECG (HR and T wave) were measured at each dose step of the DRC. Daily PEFR measurements (A.W. and P.M.) and Holter ECG were performed during run-in and treatment periods. Baseline values for FEV1 were not significantly different during run-in, treatment, or washout periods. There were dose-related increases in FEV1 (p < 0.0001) with no significant differences between DRC after treatment with HDT compared with those with LDT: Δ FEV1 max, 0.46 ± 0.14 L, 15.5 ± 3.7% predicted (HDT); 0.50 ± 0.11 L, 16.0 ± 3.1% predicted (LDT). There were also no differences between DRC for ΔFVC: 1.08 ± 0.22 L, 31.1 ± 5.4% predicted (HDT); 0.99 ± 0.14 L, 28.5 ± 3.8% predicted (LDT). There were no significant changes in K or HR in response to cumulative doses of terbutaline after either treatment. The other systemic responses to terbutaline were blunted after HDT compared with those after LDT: Δ Tr, 52 ± 10% versus 107 ± 16% (p < 0.001); ΔT wave, -0.04 ± 0.01 mV versus -0.09 ± 0.02 mV (p < 0.05); ΔcAMP, 4.70 ± 0.89 pmol/ml versus 10.34 ± 1.76 pmol/ml (p < 0.05). Holter analysis showed an increase in ventricular ectopy during treatment with HDT compared with that during treatment with LDT, although there were no significant arrhythmias. Thus, bronchodilator dose-response effects were not attenuated after treatment with HDT, in comparison with the effects of conventional low dose therapy. Systemic beta-responses were attenuated with both LDT and HDT, but to a significantly greater degree with HDT.