All subjects provided written informed consent approved by an Independent Ethics Committee or Institutional Review Board. All subjects had a diagnosis of COPD based on currently accepted criteria 14, and were current smokers who failed a mandatory smoking cessation program or self-reported ex-smokers who had stopped smoking ≥ 12 months before the study. Eligible subjects had a prebronchodilator FEV₁/FVC (forced vital capacity) ratio ≤ 70%, postbronchodilator FEV₁ between 30% and 70% predicted, and low postbronchodilator FEV₁ reversibility (< 10% of predicted normal). Per protocol, subjects did not receive inhaled, oral, or parenteral corticosteroids for 6 weeks prior to screening. During the study, ipratropium bromide, theophylline, short- and long-acting β₂-adrenergic agonists (with appropriate washout before study visits) were allowed.

Subjects with a clinical history of asthma or any other clinically significant medical illness other than COPD were excluded. Other exclusion criteria included a COPD exacerbation within 3 months before the baseline visit; ventilator support for respiratory failure within the past year; lobectomy, pneumonectomy, or lung volume reduction surgery; lung cancer within the past 5 years; nasal continuous positive airway pressure or oxygen use > 2 L/min or for >2 hours per day; initiation of pulmonary rehabilitation within the past 3 months; treatment with chronic or prophylactic antibiotics; inability to use the MF-DPI inhaler; and < 80% adherence in recording diary data between screening and baseline.

This was a randomized double-blind, placebo-controlled, parallel-group study in males and females of any race, ≥ 40 years of age, with a clinical history and spirometry diagnostic of COPD. The study was conducted at 95 sites in 11 countries. Subjects underwent a 2-week run-in before randomization, in which spirometry results, exacerbations, symptom scores, and health status were recorded to ensure clinical stability. Subjects who had an exacerbation during the run-in were rescreened 2 or 6 weeks after completion of antibiotic or oral corticosteroid therapy, respectively. These measurements were made at each subsequent clinic visit (weeks 1, 4, 13, 26, 39, and 52). Telephone contacts occurred at weeks 8, 17, 21, 30, 34, 43, and 47 to reinforce adherence to study procedures and monitor adverse events, exacerbations, and concomitant medication use. Eligible subjects were randomized via computer-generated code in a ratio of 2:2:1:1 to 52 weeks of treatment with MF-DPI 800 μg QD PM, MF-DPI 400 μg BID, placebo QD PM, or placebo BID. Dosing regimens (QD or BID) were not blinded. Spirometry was performed before, and 30 minutes after, inhalation of albuterol 400 μg. Testing was conducted to American Thoracic Society standards 21, and the reference values of Crapo et al 22 were used to determine the % predicted FEV₁. Subjects maintained twice-daily diaries documenting symptom scores, daily use of rescue medication, and cigarette consumption. Exacerbations requiring treatment with antibiotics or oral corticosteroids were recorded throughout the study. Subjects who discontinued during the treatment period continued to maintain their diaries, which were reviewed at follow-up visits. Health-related quality of life was evaluated at baseline and every 3 months.

The primary efficacy variable was the change from baseline in postbronchodilator FEV₁. Other prespecified efficacy variables were the percentage of subjects with 1 or more exacerbations during the study and the change from baseline in total COPD symptom scores. An exacerbation was defined as a clinically significant worsening of COPD symptoms requiring treatment with antibiotics and/or systemic steroids. Subjects who experienced 3 COPD exacerbations or needed more than 3 weeks of treatment for an exacerbation were discontinued. Total symptom scores were the average of daytime and nighttime scores for difficulty breathing, coughing, and wheezing. Subjects recorded scores for each of these symptoms in daily diaries. Difficulty breathing was rated on separate scales for daytime and nighttime scores of 0 (none) to 4 (severe), while coughing and wheezing were rated on a scale of 0 (none) to 3 (very uncomfortable). Secondary efficacy evaluations were changes from baseline in St. George's Respiratory Questionnaire (SGRQ) and 36-item Short Form (SF-36) scores, prebronchodilator FEV₁, prebronchodilator and postbronchodilator FVC and forced expiratory flow between 25% and 75% of vital capacity (FEF_25%–75%), and individual daytime and nighttime symptom scores.

Safety assessments included monitoring of adverse events, with specific oropharyngeal and forearm examinations, and assessment of vital signs at all study visits. Physical examinations and laboratory tests were done at screening and the final visit. Plasma cortisol levels were assessed at the baseline and final visits in subjects at approximately 15 centers; samples were taken at 4 AM, 5 AM, 6 AM, 7 AM, 8 AM, 9 AM, 10 AM, 12 PM, 4 PM, 8 PM, and 11 PM. In other selected centers, bone mineral density (BMD) in the lumbar spine and proximal femur, using dual-energy X-ray absorptiometry (DXA) was assessed. The bone scans were performed by local radiologists and the results were reviewed by Synarc, Inc. (Portland, OR).

The efficacy and safety analyses were based on all randomized subjects (intent-to-treat population). A confirmatory analysis of subjects who met key eligibility and evaluability criteria was also performed. Results are expressed as least squares means (± SD). A longitudinal analysis-random coefficient model was used to evaluate treatment effects on postbronchodilator FEV₁ and COPD symptom scores. Longitudinal analysis of results for postbronchodilator FEV₁ extracted sources of variability due to smoking status, treatment, number of days on treatment, and treatment-by-time interaction, with a random slope and intercept for each subject. An unstructured covariance matrix, with variance of random intercepts and slopes and covariance between intercepts and slopes, was chosen to allow full flexibility of the model. The Cochran-Mantel-Haenzsel test was used to analyze exacerbation frequency.

The primary hypothesis was that 800 μg QD PM would be superior to placebo with respect to changes from baseline in postbronchodilator FEV₁ and either total COPD symptom scores or proportion of subjects with one or more exacerbations during the study. The study design required that 780 subjects meet the criteria for evaluation of the 3 prespecified efficacy variables. To control for type I error, the pooled MF-DPI groups were to be compared with the pooled placebo groups. If pooled MF-DPI was significantly superior to pooled placebo for FEV₁ and at least 1 of the other prespecified criteria, then the following comparisons were made: MF-DPI 800 μg QD PM versus pooled placebo, MF-DPI 400 μg BID versus placebo. If at least 1 of the MF-DPI treatments was superior to placebo, then the 2 MF-DPI treatments were compared with each other. Mean changes from baseline were compared at a 2-sided α = 0.05, providing at least 90% power to detect between the treatment means a difference of 50 mL in postbronchodilator FEV₁, a difference of 0.2 in total symptom scores, and a 16% difference in the proportion of subjects having at least 1 exacerbation. This power was maintained throughout the stepwise comparisons of MF-DPI treatments with placebo. A post hoc analysis was performed to test whether the MF-DPI 800 μg QD PM and 400 μg BID were equivalent in terms of their effects on the co-primary endpoints of FEV₁, total symptom scores, and exacerbations. The equivalence margin chosen was 50% of the difference between placebo and MF-DPI 400 μg BID. The post hoc analysis had 78% power to detect equivalence with regard to FEV₁, 59% power to detect equivalence with regard to total symptom scores, and 21% power to detect equivalence with regard to exacerbations.

To protect against inflation of the error rate for multiple primary endpoints, comparisons of exacerbations and symptom scores were adjusted using a modified Bonferroni correction, with a 2-sided α = 0.025 for the more significant comparisons and α = 0.05 for the less significant comparisons. This was determined by the size of the results from the symptom score and exacerbation comparisons, the smaller of which was considered more significant as it required α = 0.025 for statistical significance, and the results of the larger value considered less significant, as it required α = 0.05 for statistical significance.

Significant improvements in postbronchodilator FEV₁ were observed in both MF-DPI groups over the 1-year treatment period (Table 2 and Figure 2). In addition, both MF-DPI regimens were superior to placebo (P ≤ 0.012) for changes from baseline in prebronchodilator FEV₁ and prebronchodilator and postbronchodilator FVC and FEF_25%–75% (Table 2). The MF-DPI treatments were consistently superior to placebo in LABA users and nonusers. No significant differences were observed between MF-DPI treatments in any of the pulmonary function variables, and there were no differences in treatment effect based on subjects' sex, age, prebronchodilator FEV₁% predicted, or prior medical history. However, the response to MF-DPI treatment was greater in ex-smokers (those who quit smoking ≥ 10 months before baseline) than in those who continued smoking. Mean changes in cigarette use during treatment were low (< 2 cigarettes/day) in all treatment groups. In ex-smokers, postbronchodilator FEV₁ increased approximately 50 mL with MF-DPI compared with a decrease (-11 mL) with placebo. In current smokers, postbronchodilator FEV₁ increased (29 mL) in the MF-DPI 800 μg QD PM group, but decreased in the MF-DPI 400 μg BID and placebo groups (-9 mL and -41 mL, respectively).

A total of 334 randomized subjects (37%) had 1 or more COPD exacerbations during treatment: 105 (34%) in the MF-DPI 800 μg QD PM group, 107 (35%) in the MF-DPI 400 μg BID group, and 122 (41%) in the placebo group. The difference between the pooled MF-DPI groups and placebo was significant (P = 0.043). Analysis by the log-rank test showed that each MF-DPI treatment significantly (P < 0.019) prolonged the time to first exacerbation compared with placebo. In each MF-DPI group, 46% of those with a history of ≥ 3 exacerbations exacerbated during the treatment period versus 30% of those with a history < 3 exacerbations. In the placebo group, 61% of those with a history of ≥ 3 exacerbations exacerbated during the treatment period, whereas 34% of patients with < 3 previous exacerbations exacerbated. Treatment efficacy with respect to second and third exacerbations was evaluated as the total number of events divided by the total follow-up time. With this measurement the exacerbation rates for MF-DPI 800 μg QD PM, MF-DPI 400 μg BID, and placebo were 0.62, 0.65, and 0.96, respectively.

A greater proportion of subjects in the placebo group had severe exacerbations (resulting in hospitalization, use of both oral steroids and antibiotics, or additional oral steroids) than did those in either of the MF-DPI groups. A greater proportion of exacerbations in the MF-DPI groups were treated with antibiotics alone. These differences were not statistically significant (Table 3). Also, subjects with baseline FEV₁ < 50% predicted (ie, GOLD Stages III-IV) had more exacerbations than those with FEV₁ > 50% predicted (ie, GOLD Stages I-II). For subjects in GOLD Stages I-II, exacerbations were reported for 18% in the MF-DPI 800 μg QD PM group, 27% in the MF-DPI 400 μg BID group, and 35% in the placebo group. For subjects in GOLD Stages III-IV, exacerbations were reported for 43% in the MF-DPI 800 μg QD PM group, 41% in the MF-DPI 400 μg BID group, and 48% in the placebo group.

Of the 258 patients who currently smoked, 93 exacerbated 1 or more times during the study, with 28 (32%) in the MF-DPI 800 μg QD PM group, 34 (36%) in the MF-DPI 400 μg BID group, and 31 (40%) in the placebo group. Of the 653 ex-smokers, 241 subjects exacerbated 1 or more times, with 77 (35%) in the MF-DPI 800 μg QD PM group, 73 (34%) in the MF-DPI 400 μg BID group, and 91 (42%) in the placebo group.

Total COPD symptom scores (average of daytime and nighttime scores), improved significantly (P < 0.05) from baseline over the 12-month treatment period with both MF-DPI 400 μg BID (-0.53) and MF-DPI 800 μg QD PM (-0.34) compared with placebo (-0.12). A confirmatory analysis based on the efficacy evaluable data set and all randomized subjects while on treatment indicated a significant improvement from baseline in total COPD symptom scores for each active treatment group compared with placebo (P ≤ 0.021). For individual symptom scores (Table 4), significant improvements (P < 0.025) in daytime difficulty breathing scores were observed for each active treatment compared with placebo and for most other scores for the MF-DPI 400 μg BID group.

Scores for the SGRQ Total, Activity, Impacts, and Symptoms domains (Table 5) improved significantly in the pooled MF-DPI groups compared with the pooled placebo groups (P ≤ 0.031). A significant improvement in the SF-36 physical component summary score was also observed in the pooled MF-DPI groups compared with placebo (P < 0.05). For the individual groups, treatment with MF-DPI 800 μg QD PM significantly improved SGRQ Symptoms scores (-5.77, P = 0.050). Compared with placebo, treatment with MF-DPI 400 μg BID significantly improved SGRQ Total scores (-3.99, P = 0.008), Impacts scores (-3.41, P = 0.042), and Symptoms scores (-6.85, P = 0.009).

The incidence rates of any treatment-emergent adverse event were comparable in the 3 treatment groups (Table 6). Treatment-related adverse events were reported by 27% of subjects taking MF-DPI 800 μg QD PM, 28% of subjects taking MF-DPI 400 μg BID, and 20% of placebo-treated subjects. Most adverse events reported were mild to moderate in severity; no life-threatening events were considered to be related to study medication. Oral candidiasis was the most frequent treatment-related adverse event; 10%–11% for MF and 3% for placebo. The incidence of bruising was 14% in each MF-DPI group and 11% in the placebo group. In both groups, the incidence rates of fracture, osteoporosis, and cataracts were ≤ 1%.

Ten subjects died during treatment or within 30 days of the last dose of study treatment: 2 in the MF-DPI 800 μg QD PM group, 5 in the MF-DPI 400 μg BID group, and 3 in the placebo group. None of the deaths was due to respiratory disease or lung cancer or considered to be related to treatment. Serious adverse events were reported by 142 subjects during the treatment period: 44 (14%) in the MF-DPI 800 μg QD PM group, 47 (15%) in the MF-DPI 400 μg BID group, and 51 (17%) in the placebo group. The only serious events reported by > 1% of any treatment group were pneumonia (2% in each MF-DPI group and 1% in the placebo group) and COPD aggravated (4% in each MF-DPI group and 5% in the placebo group).

The total lumbar spine BMD increased slightly (0.857%) at endpoint in the MF-DPI 800 μg QD PM group and decreased slightly in the MF-DPI 400 μg BID and placebo groups (-0.944% and -0.068%, respectively). Likewise, total femoral BMD increased slightly (0.347%) at endpoint in the MF-DPI 800 μg QD PM group and decreased slightly in the MF-DPI 400 μg BID and placebo groups (-2.002% and -0.677%, respectively). Femoral neck results were similar in direction and inference between treatment groups to the total femoral BMD results. No significant differences were observed between any of the treatment groups.

At endpoint, there was a significant decrease of 22.9% in the plasma cortisol level for subjects in the MF-DPI 400 μg BID treatment group compared with an increase of 3.7% for subjects in the MF-DPI 800 μg QD PM group (P = 0.040) and 5.3% for subjects in the placebo group (P = 0.007).