Introduction We report the results of the first direct comparison of the once-daily fixed-dose long-acting muscarinic antagonist/long-acting β 2 -agonist (LAMA/LABA) combinations umeclidinium/vilanterol (UMEC/VI) and tiotropium/olodaterol (TIO/OLO) in patients with COPD. Methods This was a randomized, two-period crossover open-label study in symptomatic patients with COPD age 40 years or older, postbronchodilator forced expiratory volume in 1 s (FEV 1 ) of 70% or less and 50% or more of predicted normal values, and modified Medical Research Council Dyspnoea Scale score of 2 or greater not receiving inhaled corticosteroid therapy. Patients were randomized to receive UMEC/VI (62.5/25 µg once daily) via a multidose dry powder inhaler (ELLIPTA) followed by TIO/OLO (5/5 µg once daily) via a soft mist inhaler (Respimat), each for 8 weeks with an interim 3-week washout or vice versa. The primary end point was the change from baseline in trough FEV 1 at week 8 with a noninferiority margin of − 50 mL in the per-protocol (PP) population. The incidence of adverse events was also assessed. Results In total, 236 patients (mean age 64.4 years, 60% male) were included in the intent-to-treat population and 227 were included in the PP population. UMEC/VI treatment was noninferior in the PP population and superior in the intent-to-treat population to TIO/OLO treatment with regard to trough FEV 1 at week 8 FEV 1 change from baseline 180 mL vs 128 mL; difference 52 mL (95% confidence interval 28–77 mL); p  < 0.001. Patients receiving UMEC/VI had twofold increased odds of experiencing a clinically meaningful increase (100 mL or more) from baseline in trough FEV 1 at week 8 compared with patients receiving TIO/OLO (odds ratio 2.05; 95% confidence interval 1.34–3.14). Adverse events occurred in 25% of patients in the UMEC/VI group and in 31% of patients in the TIO/OLO group. Conclusion In this first direct comparison of two once-daily fixed-dose LAMA/LABA combinations, superiority was observed for the primary end point of trough FEV 1 at week 8 with UMEC/VI compared with TIO/OLO in patients with symptomatic COPD. Both treatments had similar safety profiles. These findings confirm the results of previous indirect LAMA/LABA comparisons, and show that an efficacy gradient exists within the LAMA/LABA class. Trial Registration ClinicalTrials.gov identifier NCT02799784. Funding GlaxoSmithKline.