Previous studies have described that there are different types of disease in patients with established chronic obstructive pulmonary disease (COPD) with different clinical course and functional responses. The aim of this study was to evaluate if the presence of low transfer factor (LTF) values can predict the effectiveness of bronchodilator therapy, and to assess whether this group has different risk factors that may be related with the responses. Eighty patients with COPD were evaluated on three occasions. Initial assessment included a standard respiratory questionnaire, blood analysis, skin prick test and baseline lung function, all performed on the first visit. Bronchodilator response was evaluated after low (0.2 mg) and high (1 mg) doses of salbutamol, and after 2 weeks of oral prednisone. In patients with normal TLCO/VA % (NTF), a higher proportion of subjects with previous history of atopy was the only statistically significant difference compared to those with LTF (odds ratio 4.33; 95% confidence interval 1.06-25.15). Although the mean response in forced expiratory volume in 1 s (FEV1) to treatment was analogous in both groups, when bronchodilation was expressed as percent of predicted, there was a clear trend to a lower response in patients with LTF (0.2 mg salbutamol: 6.99 +/- 5.64 vs. 8.94 +/- 6. 61, p = 0.15; 1 mg salbutamol: 10.18 +/- 6.37 vs. 13.45 +/- 7.90, p < 0.05; oral prednisone: 5.51 +/- 6.94 vs. 8.74 +/- 10.81, p = 0.06). The percentage of patients who had >12% improvement from that predicted in FEV1 was also lower in this group (42 vs. 72%; p < 0. 05). Moreover, TLCO/VA% was significantly lower in those subjects with a negative bronchodilator trial with salbutamol (68 +/- 25 vs. 81 +/- 26; p < 0.05) and prednisone (69 +/- 26 vs. 90 +/- 22; p < 0. 01). In patients with LTF and NTF, airway responsiveness was only significantly related with basal airflow limitation (LTF, r = 0.44; NTF, r = 0.38). All other interaction terms were not statistically significant. These results indicate that in patiens with similar serverity of COPD, the presence of LTF indicates a decreased probability of a positive bronchodilator response, probably reflecting different pathological lesions. We suggest that transfer factor should be taken into consideration when bronchial response is evaluated in large clinical trials.