Perception of the efficacy of bronchodilators in relieving airflow obstruction is a likely determinant of compliance with treatment in patients prescribed these drugs on an 'as needed' basis. This study aimed to determine whether bronchodilator-induced improvements in lung function are associated with improvements in breathing difficulty in subjects with asthma or smokers with airflow limitation. Twenty smokers with airflow limitation and 16 subjects with previously physician-diagnosed asthma received salbutamol (200 micrograms) and ipratropium bromide (80 micrograms). Spirometry and lung volumes were measured before and 40 min after bronchodilator. Subjects recorded changes in 'difficult breathing' on a visual analogue scale (VAS). After bronchodilator, forced expiratory volume in 1 s (FEV1) increased by 23.0 +/- 6.4% of baseline (mean +/- 95% CI) in smokers, and by 25.2 +/- 8.5% in the asthmatics, while VAS improved by 31 +/- 23% in smokers and 45 +/- 25% in asthmatics. However, these changes were not significantly correlated in either smokers (r = -0.04) or asthmatics (r = 0.15). In the asthmatic subjects, good perceivers (> 25% improvement in VAS) had greater improvements in lung volumes, as percentage predicted, than did poor perceivers. In the smokers, changes in lung function did not differ significantly between good and poor perceivers. Improvement in FEV1, as percentage predicted, was significantly correlated with improvement in VAS in good perceivers (asthma: r = 0.78, P < 0.01; smokers: r = 0.68, P < 0.05), but not in poor perceivers. Asthmatic subjects had good perception of improvements in lung function. However, in smokers with airflow limitation there is little correlation between improvement in lung function and sensation of breathing difficulty. In these subjects symptoms appear to be an unreliable guide for 'as needed' use of bronchodilators.