COPD exacerbations starting between first dose and one day after last treatment are included. COPD exacerbations that occurred within 7 days of each other are collapsed as one event. Estimates are from a generalized linear model assuming a negative binomial distribution with terms for treatment, baseline total symptom score, baseline COPD exacerbation history (i.e. number of COPD exacerbations during the past 12 months prior to study), smoking status at screening, ICS use at screening, airflow limitation severity, and region. As the offset variable log(exposure time in years) was used.

Pulmonary function assessments were performed using centralized spirometry according to international standards. Baseline FEV1 was defined as the average of the pre-dose FEV1 measured at -45 minutes (min) and -15 min at day 1. A mixed model for repeated measures (MMRM), used for this analysis, included terms of treatment, baseline FEV1 measurements, smoking status at baseline, baseline inhaled corticosteroid (ICS) use, region, baseline FEV1 \* visit interaction, and visit, treatment \* visit interaction. Missing values of FEV1 AUC0-12 at Day 1 and Week 12 will not imputed. The trapezoidal rule was used to calculate FEV1 AUC and then normalized to the length of time.

Pulmonary function assessments were performed using centralized spirometry according to international standards. Baseline FEV1 was defined as the average of the pre-dose FEV1 measured at -45 minutes (min) and -15 min at day 1. A mixed model for repeated measures (MMRM), used for this analysis, included terms of treatment, baseline FEV1 measurements, smoking status at baseline, baseline inhaled corticosteroid (ICS) use, region, baseline FEV1 \* visit interaction, and visit, treatment \* visit interaction. Missing values of FEV1 AUC0-12 at Day 1 and Week 12 will not imputed. The trapezoidal rule was used to calculate FEV1 AUC and then normalized to the length of time.

The overall rate of serious adverse events reported from initiation through 30 days post last dose.

The overall rate of adverse events reported from initiation through 30 days post last dose.

SGRQ is a health related quality of life questionnaire consisting of 40 items in three areas: symptoms (respiratory symptoms and severity), activity (activities that cause or are limited by breathlessness) and impacts (social functioning and psychological disturbances due to airway disease). The total score is 0 to 100 with a higher score indicating poorer health status. For patients who completed the study but with missing SGRQ-C during treatment, the missing SGRQ-C were replaced by the last observation carried forward (LOCF). Symptom scores were expected to improve over treatment, therefore the replacement of missing values with earlier measurements did not result in overoptimistic imputation and this procedure could be regarded as conservative.

An AE was the appearance or worsening of any undesirable sign, symptom, or medical condition occurring after starting the study drug even if the event was not considered to be related to study drug. Study drug includes the investigational drug under evaluation and the comparator drug or placebo that was given during any phase of the study. Adverse events starting on or after the time of the first inhalation of study drug were classified as a treatment emergent adverse event.