Loss of bronchodilator effectiveness or tolerance has been observed with inhaled beta-agonists but not with inhaled anticholinergic medications. Initially, tolerance is reflected in loss of bronchial protection against stimuli followed by loss of bronchodilator properties. However, generally such observations have been reported in asthma. A 6-month randomized, double-dummy placebo-controlled trial comparing tiotropium to salmeterol provided the opportunity to examine spirometric tolerance to long-acting beta-agonists in patients with COPD. Spirometry was measured over 12h at baseline and at days 15, 57, 116 and 169. Changes over time from baseline were compared relative to changes observed with placebo. A total of 623 patients participated (tiotropium = 209, salmeterol = 213, placebo = 201). The groups were similar in age (mean = 65 years), gender (75% men), and baseline FEV1 (mean = 1.08 +/- 0.37l [40 +/- 12% predicted]). Relative to placebo, both active drugs improved morning pre-drug, peak and average FEV1 and FVC throughout the trial. However, from day 1 to 169, salmeterol was associated with a higher decline in average FEV1 and FVC (0-12h) (difference from placebo: -36 and -115 ml, P < 0.05), which was most prominent over the 8-12 h period (difference from placebo: -45 and -138 ml, P < 0.01). Significant declines in peak FVC relative to placebo (-83 ml, P < 0.05) but not FEV1 (-12ml) were observed with salmeterol. Tiotropium was associated with further improvements in spirometry from days 1 to 15 and no evidence of tolerance from day 15 to the end of the trial. In conclusion, tolerance to pharmacologic bronchodilation occurs with long-acting beta-agonists such as salmeterol and not with inhaled anticholinergics.