OBJECTIVES: To determine complications (exacerbations, pneumonia) and health resource use in patients in whom inhaled corticosteroids (IC) are replaced by a long-acting anticholinergic (LAMA), in dual therapy, for the treatment of COPD in usual clinical practice. METHODS: Retrospective, observational study (before/after with control group) made by review of medical records. In patients aged ≥40 years with ≥ 1year of treatment with CI/ long-acting beta-adrenergic agonists (LABA), IC were replaced by a LAMA (index date). Study groups were: a) CI/LABA (without replacement, dual therapy) and LABA/LAMA (with replacement of IC, dual bronchodilation). From the index date, a 12-month reference period and a subsequent 12-month follow-up period was observed. Each patient in the LABA/LAMA group was matched with a patient in the CI/LABA group using propensity score matching (PSM). The statistical analysis was made using multivariate models. RESULTS: 2,334 patients were recruited. In 19.9% (N=464) IC were replaced with a LAMA. The study groups after matching were: a) CI/LABA N=462 and b) LABA/LAMA N=462. The mean age was 71.8 years, 80.3% were male and the mean Charlson index was 0.9. The mortality rate was 10.4%, 35.5% had severe COPD, 13.9% high daily CI doses, FEV1 was 54.9% and 8.5% received home oxygen therapy. Between-group results were very similar. 22.4% had exacerbations (CI/LABA: 21.4% vs. LABA/LAMA: 23.4%; p=0.478). Severe exacerbations (CI/LABA: 5.6% vs. 6.3%, p=0.675), and total exacerbations prior to the study period (22.7% vs. 22.9%, p=0.938) were similar. 5.3% had pneumonia (CI/LABA: 4.5% vs. LABA/LAMA: 6.1%; p=0.554). Resource use did not differ significantly between groups. CONCLUSIONS: Relatively few patients had an IC replaced by a LAMA. The comparison between CI/LABA (dual therapy) and LABA/LAMA (dual bronchodilation) showed similar rates of exacerbation, pneumonia and use of health resources. However, further studies are needed to reinforce the consistency of the results.