Background Inhaled corticosteroids have been used in patients with chronic obstructive pulmonary disease (COPD), but the potential benefits of their use in triple therapy are not well known. We aimed to compare the efficacy of a triple therapy with corresponding dual therapies in symptomatic patients with moderate to very severe COPD, without a requirement for a history of exacerbations. Methods In this double-blind, parallel-group, multicentre phase 3 randomised controlled trial, we recruited patients from hospitals and care centres in Canada, China, Japan, and the USA. Eligible patients were 40-80 years of age, were current or former smokers (with a smoking history of >= 10 pack-years), had an established clinical history of COPD, and were symptomatic for COPD, despite receiving two or more inhaled maintenance therapies for at least 6 weeks before screening. We randomly assigned patients (2:2:1:1) using an interactive web response system to receive budesonide/glycopyrrolate/formoterol fumarate metered-dose inhaler 320/18/9.6 mu g (BGF MDI), glycopyrrolate/formoterol fumarate metered-dose inhaler 18/9.6 mu g (GFF MDI), budesonide/formoterol fumarate metered-dose inhaler 320/9.6 mu g (BFF MDI), or open-label budesonide/formoterol fumarate dry-powder inhaler 400/12 pg (BUD/ FORM DPI). Primary endpoints for the Europe/Canada statistical analysis approach were FEV1 area under the curve from 0-4 h (AUC(0-4)) for BGF MDI versus BFF MDI and BGF MDI versus BUD/FORM DPI over 24 weeks; and change from baseline in morning pre-dose trough FEV, for BGF MDI versus GFF MDI and non-inferiority of BFF MDI versus BUD/FORM DPI (margin of -50 mL from lower bound of 95% CI) over 24 weeks. Comparisons with BUD/FORM DPI were made for the Europe/Canada statistical analysis approach only. This study is registered with ClinicalTrials.gov, number NCT02497001. Findings Between Aug 20, 2015, and Jan 5, 2018, 3047 patients were screened from 215 sites, and 1902 were randomly assigned to receive BGF MDI (n=640), GFF MDI (n=627), BFF MDI (n=316), or BUD/FORM DPI (n=319). Over 24 weeks, BGF MDI significantly improved FEV(1)AUC(0-4) versus BFF MDI (least squares mean difference 104 mL, 95% CI 77 to 131; p<0.0001) and BUD/FORM DPI (91 mL, 64 to 117; p<0.0001). BGF MDI also significantly improved pre-dose trough FEV1 versus GFF MDI (22 mL, 4 to 39; p=0 . 0139) and was non-inferior to BUD/FORM DPI (-10 mL, -36 to 16; p=0 . 4390). At week 24, patients in the BGF MDI group had a significantly improved FEV1 AUC(0-4) compared with patients receiving BFF MDI (116 mL, 95% CI 80 to 152; p<0 . 0001); there was a non-significant improvement in the change from baseline in morning pre-dose trough FEV1 at week 24 versus GFF MDI (13 mL, -9 to 36 mL; p=0.2375). The most common treatment-emergent adverse events were nasopharyngitis (n=49[8%) in the BGF MDI group; n=41 [7%] in the GFF MDI group; n=26 [8%] in the BFF MDI group; and n=30[9%] in the BUD/FORM DPI group) and upper respiratory tract infection (n=65 [10%]; n=38 [6%]; n=18 [6%]; and n=22 [7%]). Pneumonia incidence was low (<2%) and similar across treatments. There were two treatment-related deaths, both in the GFF MDI group. Interpretation BGF MDI was efficacious, well tolerated, and could be a more appropriate treatment than the corresponding dual therapies for symptomatic patients with moderate to very severe COPD, irrespective of exacerbation history. Copyright (C) 2018 Elsevier Ltd. All rights reserved.