If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Open-label, parallel-group, international trial comparing aztreonam for inhalation solution (AZLI) and tobramycin nebulizer solution (TNS) for cystic fibrosis patients with airway Pseudomonas aeruginosa.
Methods
273 patients (≥6 years); randomized to three 28-day courses (AZLI 75 mg [three‐times/day] or TNS 300 mg [twice/day]); 28 off-days separated each course.
Results
268 patients were treated (AZLI/TNS: 136/132). Mean baseline FEV1 was 52% predicted. Mean relative changes after 1 course (AZLI: 8.35%; TNS: 0.55%; p<0.001) and mean actual changes across 3 courses (AZLI: 2.05%; TNS: −0.66%; p=0.002) indicated AZLI statistical superiority vs. TNS. AZLI-treated patients had fewer respiratory hospitalizations (p=0.044) and respiratory events requiring additional antipseudomonal antibiotics (p=0.004); both treatments were well tolerated. 133 patients received 1 to 3 courses of AZLI treatment in the open-label extension-period (28-day courses separated by 28 days off-treatment); lung function improvements were comparable regardless of whether patients had received TNS or AZLI in the preceding comparative period.
Conclusions
AZLI demonstrated statistical superiority in lung function and a reduction in acute pulmonary exacerbations compared to TNS over 3 treatment courses (ClinicalTrials.gov: NCT00757237).
]. Relatively high delivered drug dose maximizes airway concentrations, and low systemic absorption reduces side effects. Three inhaled antipseudomonal antibiotics with different mechanisms of action are available for treating CF patients: tobramycin nebulizer solution (TNS; an aminoglycoside), colistimethate sodium (colistin; a polymyxin; not licensed in the US), and aztreonam for inhalation solution (AZLI; a monobactam) [
]. TNS has proven efficacy for CF patients, and established the standard of care of every other month inhaled antibiotic treatment for suppression of chronic airway infection [
]. Since AZLI was recently approved, it is appropriate to evaluate it against standard of care, given the ongoing need for more effective and better tolerated inhaled antibiotic treatments.
AZLI is a lyophilized formulation of the monobactam antibiotic aztreonam, specifically designed for inhalation therapy with lysine as an excipient [
]. In previous placebo-controlled clinical trials, efficacy and safety of AZLI were demonstrated in patients with CF and PA airway infection who previously received TNS treatment and in patients with relatively little previous TNS exposure. Additionally, therapeutic benefits observed in the placebo-controlled trials were sustained across 18 months of open-label treatment [
An 18-month study, AIR-CF3, of the safety and improvement in pulmonary function and respiratory symptoms with repeated courses of aztreonam for inhalation solution in patients with cystic fibrosis and airway Pseudomonas aeruginosa.
The trial described herein compared the efficacy and safety of AZLI to TNS, across three 28‐day treatment courses. An optional open‐label extension period for eligible patients included 3 additional AZLI treatment courses.
2. Methods
2.1 Study design
This open-label, randomized, parallel-group, active-comparator study was conducted at 91 CF‐centers in Europe and the US (August 2008–May 2010). Eligible patients were randomized (1:1) to three 28-days on/28 days off treatment courses of AZLI (Cayston®, Gilead Sciences Inc.) 75 mg three‐times daily or TNS (TOBI®, Novartis Pharmaceuticals Corp.) 300 mg twice daily (Online Fig. 1) [
]. The 26‐week active-comparator period was followed by an optional, 24‐week, open‐label, single‐arm, extension period of 3 AZLI treatment-cycles for eligible patients at European study sites (completed: Nov 2010).
The study was conducted in accordance with principles of the Declaration of Helsinki (as amended in Edinburgh, Tokyo, Venice, Hong Kong, and South Africa), International Conference on Harmonisation guidelines, Good Clinical Practice principles, and laws/regulations of each study location. Institutional Review Boards/Ethics Committees approved the study for each site. Patients/parents provided written informed consent/assent prior to undergoing study procedures.
2.2 Participants
Eligible patients (≥6 years of age) had a documented CF diagnosis, PA-positive sputum culture within the previous 3 months, and FEV1 ≤75% predicted at screening. Additional TNS use was prohibited during the active-comparator period. Additional antipseudomonal antibiotics could be administered for symptoms consistent with the diagnosis of acute pulmonary exacerbation: decreased exercise tolerance, increased cough, increased sputum/chest congestion, or decreased appetite [
]. Patients receiving additional antipseudomonal antibiotics at any point after randomization could continue study treatments.
2.3 Study endpoints
Co-primary efficacy endpoints were designed to satisfy registrational requirements based on European Medicines Agency and the US Food and Drug Administration guidance: non‐inferiority of AZLI for relative change in FEV1% predicted at Day 28 and superiority of AZLI for actual change in FEV1% predicted across 3 treatment cycles. Selected secondary and tertiary endpoints were: time-to-need for IV antipseudomonal antibiotics for respiratory events, change from baseline on cystic fibrosis questionnaire—revised (CFQ-R) respiratory symptoms scale (RSS), number of respiratory hospitalizations, additional antipseudomonal antibiotic use, weight change from baseline, change in sputum PA density (colony forming units per gram sputum [CFU/g]), and responses to the treatment satisfaction questionnaire for medication (TSQM) [
Validation of a general measure of treatment satisfaction, the treatment satisfaction questionnaire for medication (TSQM), using a national panel study of chronic disease.
]. An independent, blinded data adjudication committee determined respiratory hospitalizations and respiratory events requiring additional antipseudomonal antibiotics.
2.4 Statistical analysis
Statistical analyses were performed on the intent-to-treat (ITT) population: randomized patients receiving ≥1 dose of AZLI/TNS. The primary non‐inferiority endpoint was assessed with an analysis of covariance (ANCOVA) model with terms for treatment, baseline FEV1% predicted (continuous variable), and inhaled tobramycin use in previous year (≥84, <84 days). If the 95% confidence interval (CI; 2-sided) lower boundary for the treatment difference (AZLI–TNS) was >−4% (pre‐specified non‐inferiority margin), non-inferiority of AZLI to TNS was concluded. Superiority was concluded if the lower limit of the 95% CI was above zero (i.e., p<0.05).
The primary superiority endpoint was the average least‐square means from Weeks 4, 12, and 20 visits, based on a mixed-effect model repeated measures (MMRM) analysis method outlined by Siddiqui, which included terms for treatment, baseline FEV1% predicted (continuous variable), inhaled tobramycin use (≥84, <84 days), visit, and treatment/visit interaction [
Of 273 randomized patients, 268 received treatment with AZLI (n=136) or TNS (n=132); 233 patients (85.3%) completed the active-comparator period (Fig. 1). Mean use of distributed vials was 94.0% (AZLI) and 94.2% (TNS). Of 169 eligible patients, 133 (78.7%) entered the open‐label extension period (AZLI: 68; TNS: 65), and 118 patients completed 3 AZLI courses (88.7%).
Fig. 1Patient disposition. Efficacy and safety analyses for the active-comparator period were conducted on the intent to treat (ITT) population, which included 268 randomized and treated patients (AZLI: 136; TNS: 132).
Patient characteristics were balanced between treatment groups (Table 1). Mean age was 25.5 years, with 78.0% of patients ≥18 years of age (n=209/268). FEV1 was ≤50% predicted for 43.7% of patients (n=117). Inhaled tobramycin use ≥84 days in the previous year (equivalent to at least three 28-day courses) was reported for 85.1% of patients (n=228). The 40 patients (AZLI: 21, TNS: 19) with inhaled tobramycin use <84 days during the previous year were a heterogeneous group: 25 used no tobramycin in the previous year, some used inhaled tobramycin more than 1 year prior, and many had used inhaled colistin. The incidence of multi-drug resistant PA for cohorts with <84 or ≥84 days of inhaled tobramycin use in the previous year was comparable (34.4% vs. 30.6%, respectively), indicating significant prior antibiotic exposure in both groups. Characteristics of the 133 European patients who entered the extension period were similar to those of the overall group.
Table 1Baseline demographic and clinical characteristics.
Enrollment included ≤40 patients with <84days inhaled tobramycin use in the previous year. One patient was mistakenly stratified to ≥84days and 2 patients were mistakenly stratified to <84days.
Resistance assessed using CF Foundation definition: resistance to all antibiotics tested in 2 of the 3 drug classes (aminoglycosides, beta-lactams, quinolones) [18]. CFQ-R = cystic fibrosis questionnaire—revised; CFTR = cystic fibrosis transmembrane conductance regulator; CFU = colony forming units; FEV1 = forced expiratory volume in 1s; MIC = minimum inhibitory concentration; PA = Pseudomonas aeruginosa.
n (%)
35 (30.4)
35 (31.8)
70 (31.1)
a n=117, 111, 228 for 3 columns.
b Enrollment included ≤40 patients with <84 days inhaled tobramycin use in the previous year. One patient was mistakenly stratified to ≥84 days and 2 patients were mistakenly stratified to <84 days.
c Colistin is not licensed in the US.
d n=102, 103, and 205 for 3 columns.
e n=131, 131, and 262 for 3 columns.
f n=115, 110, and 225 for 3 columns.
g Resistance assessed using CF Foundation definition: resistance to all antibiotics tested in 2 of the 3 drug classes (aminoglycosides, beta-lactams, quinolones)
the CF Foundation Consensus Conference on Infection Control Participants
Infection control recommendations for patients with cystic fibrosis: microbiology, important pathogens, and infection control practices to prevent patient-to-patient transmission.
Mean relative changes from baseline FEV1% predicted at Day 28 were 8.35% (AZLI) and 0.55% (TNS; Fig. 2A ). The treatment difference (AZLI–TNS) was 7.80% (95% CI=3.86%, 11.73%; p<0.001). The lower CI boundary was >0, indicating AZLI was superior to TNS.
Fig. 2Change in FEV1% predicted. A. Adjusted mean relative change from baseline FEV1% predicted (MMRM analysis). The difference between AZLI and TNS after the first treatment course (Day 28) was the non‐inferiority co‐primary efficacy endpoint, shown by the arrow; results shown in the box were based on an ANCOVA analysis, as described in the Methods section. AZLI–TNS treatment differences were 6.77 at Week 2 (p<0.001); 7.88 at Week 4 (p<0.001); 6.04 at Week 12 (p=0.007), and 5.89 at Week 20 (p=0.004). B. Adjusted mean actual change from baseline FEV1% predicted (MMRM analysis). The difference between AZLI and TNS across 3 treatment courses was the superiority co-primary efficacy endpoint, as described in the box. AZLI–TNS treatment differences were 3.24 at Week 2 (p<0.001); 3.35 at Week 4 (p<0.001); 2.61 at Week 12 (p=0.02), and 2.15 at Week 20 (p=0.03). C. Adjusted mean relative change from baseline FEV1% predicted, with censoring for additional antibiotic use. The MMRM analysis shown in panel A was repeated with censoring of subsequent FEV1% predicted values after a course of IV/inhaled antipseudomonal antibiotics was prescribed for treatment of respiratory symptoms. AZLI–TNS treatment differences (trt. diff.) are shown at Week 2 (p<0.001), Week 4 (p<0.001), Week 12 (p=0.002) and Week 20 (p=0.005). Patients censored: AZLI-treated 52, TNS-treated 76. D. Adjusted mean relative change from baseline FEV1% predicted, for patients participating in the open-label AZLI extension period. The MMRM analysis shown in panel A was repeated for this patient population. TNS/AZLI patients received TNS during the active-comparator period and AZLI during the open-label extension period.
Mean actual changes from baseline FEV1% predicted across 3 treatment courses were 2.05% (AZLI) and −0.66% (TNS). The treatment difference (AZLI–TNS) was 2.70% (p=0.002; Fig. 2B), again demonstrating AZLI superiority over TNS. Both mean relative change and mean actual change from baseline FEV1% predicted were significantly larger for AZLI‐treated than for TNS‐treated patients after each on-treatment course (Fig. 2A, B).
Across 3 treatment cycles in the active-comparator period, mean relative changes in FEV1% predicted were 3.11% (AZLI) and −0.29% (TNS). The treatment difference (AZLI–TNS) was statistically significant (3.4%, p=0.02). The treatment difference for actual change in FEV1% predicted across all active-comparator study visits favored AZLI and approached statistical significance (AZLI: 0.85%, TNS: −0.5%; treatment difference 1.35%, p=0.066).
AZLI–TNS differences in mean relative change from baseline FEV1% predicted at Day 28 were statistically significant for several subgroups (Online Fig. 2).
Average change from baseline CFQ-R RSS across 3 treatment cycles was significantly larger for AZLI-treated than for TNS‐treated patients (Table 2). Decreases in sputum PA density after each course were similar for AZLI and TNS-treated patients. AZLI-treated patients demonstrated weight gains above baseline at each study visit, while TNS-treated patients' body weight remained at or below baseline. Differences favoring AZLI were statistically significant at Weeks 4, 12, and 16, but not at Week 24. Scores on the TSQM effectiveness and global satisfaction scales were significantly higher for AZLI‐treated than TNS-treated patients.
Table 2Selected secondary and tertiary efficacy endpoints.
Endpoint
AZLI-treated n=136
TNS-treated n=132
p-Value
Time to need for IV antipseudomonal antibiotics for respiratory events
Analyzed with ANCOVA models, with terms for treatment, baseline FEV1% predicted, and previous TNS use. CFQ-R respiratory symptoms, PA density, and weight models included the corresponding values at baseline (Week 0).
adjusted mean (SE)
Week 4 (after course 1; AZLI: n=131; TNS: n=131)
8.2 (1.7)
2.6 (1.7)
0.005
Average across 3 courses (Weeks 4, 12, 20; AZLI: n=131; TNS: n=131)
p-Value is for comparison of the number of events for each patient adjusted for time on study through Week 24 based on a negative binomial regression model.
40
58
0.044
Number of respiratory events requiring IV and/or additional inhaled antipseudomonal antibiotics
p-Value is for comparison of the number of events for each patient adjusted for time on study through Week 24 based on a negative binomial regression model.
84
121
0.004
Patients requiring IV and/or additional inhaled antipseudomonal antibiotics for respiratory events,
Analyzed with ANCOVA models, with terms for treatment, baseline FEV1% predicted, and previous TNS use. CFQ-R respiratory symptoms, PA density, and weight models included the corresponding values at baseline (Week 0).
adjusted mean (SE)
Week 4 (after course 1; AZLI: n=88; TNS: n=94)
−0.60 (0.23)
−0.34 (0.23)
0.330
Average across 3 courses (Weeks 4, 12, 20; AZLI: n=97; TNS: n=97)
−0.55 (0.19)
−0.32 (0.19)
0.295
Weight, relative change from baseline at Week 24 (end of active‐comparator period),
Analyzed with ANCOVA models, with terms for treatment, baseline FEV1% predicted, and previous TNS use. CFQ-R respiratory symptoms, PA density, and weight models included the corresponding values at baseline (Week 0).
Analyzed with ANCOVA models, with terms for treatment, baseline FEV1% predicted, and previous TNS use. CFQ-R respiratory symptoms, PA density, and weight models included the corresponding values at baseline (Week 0).
Analyzed with ANCOVA models, with terms for treatment, baseline FEV1% predicted, and previous TNS use. CFQ-R respiratory symptoms, PA density, and weight models included the corresponding values at baseline (Week 0).
Week 20 (after course 3), adjusted mean (SE)
Effectiveness (AZLI: n=118; TNS: n=110)
73.4 (2.3)
57.9 (2.5)
<0.001
Side-effects (AZLI: n=118; TNS: n=107)
94.3 (1.6)
92.5 (1.7)
0.328
Convenience (AZLI: n=118; TNS: n=107)
68.0 (2.2)
66.1 (2.4)
0.460
Global satisfaction (AZLI: n=118; TNS: n=109)
75.8 (4.6)
61.7 (4.9)
0.010
CFQ-R = cystic fibrosis questionnaire—revised; CFU = colony forming units; CI = confidence interval; FEV1 = forced expiratory volume in 1 s; IV = intravenous; NE = not estimable (median time exceeded study duration); PA = Pseudomonas aeruginosa; TSQM = treatment satisfaction questionnaire for medicine.
a Kaplan–Meier method; p-value based on log rank test.
b Analyzed with ANCOVA models, with terms for treatment, baseline FEV1% predicted, and previous TNS use. CFQ-R respiratory symptoms, PA density, and weight models included the corresponding values at baseline (Week 0).
c p-Value is for comparison of the number of events for each patient adjusted for time on study through Week 24 based on a negative binomial regression model.
Fewer AZLI‐treated patients received additional IV and/or inhaled antipseudomonal antibiotics for respiratory events (AZLI: 38% vs. TNS: 58%; p=0.002; Table 2). Significantly fewer respiratory events requiring IV and/or additional inhaled antipseudomonal antibiotics (AZLI: 84, TNS: 121; p=0.004) and fewer respiratory hospitalizations (AZLI: 40, TNS: 58; p=0.044) occurred in AZLI-treated patients.
Using Kaplan–Meier methods, median time-to-need for IV antipseudomonal antibiotics for respiratory events was significantly longer for AZLI‐treated than for TNS‐treated patients (p=0.003; Table 2; Fig. 3).
Fig. 3Kaplan–Meier estimates for the time to need for IV antipseudomonal antibiotics for a respiratory event. An independent, blinded committee reviewed IV antipseudomonal antibiotic use in the active-comparator period and identified those that were respiratory events. Patients without events were considered as right‐censored; time to censoring was the number of days from baseline (Day 0) to the date of completion of the active-comparator period or early withdrawal. Time to need was the number of days from baseline (Day 0) to the date of need. The log‐rank test was used to compare AZLI and TNS among all ITT patients.
To account for the confounding effect of additional antibiotic use on lung function, a post-hoc MMRM analysis of relative change from baseline FEV1% predicted was performed, censoring FEV1% predicted values for individual patients subsequent to their receiving a course of IV/inhaled antipseudomonal antibiotics (Fig. 2C). Compared with the primary analysis, FEV1% predicted responses were lower in TNS-treated patients at the end of each treatment cycle, and the magnitude of the treatment differences (AZLI–TNS) were better maintained over the 3 treatment courses.
3.4 Efficacy — extension period
Extension period patients (n=133; AZLI/AZLI: 68; TNS/AZLI: 65), demonstrated improvements in FEV1% predicted during each AZLI treatment cycle (Fig. 2D). For AZLI/AZLI-treated patients, active-comparator period treatment responses were sustained across the extension period. TNS-treated patients had minimal improvements in FEV1% predicted during the active-comparator period, but after switching to AZLI in the extension period, their lung function improvements were comparable to the AZLI/AZLI group.
Increases in body weight were observed for AZLI treated patients (Online Fig. 3). AZLI/AZLI-treated patients gained weight across the active-comparator period (mean relative change from study baseline: 0.55%, 0.86%, and 0.94%, after courses 1, 2 and 3) and the 3 extension-period courses (1.30%, 1.76%, and 1.93%, respectively). TNS/AZLI-treated patients lost weight during the 3 TNS treatment courses (mean relative change from study baseline: −0.83%, −0.82%, and −0.20%, respectively) and had increases in weight after switching to AZLI for the 3 extension-period courses (0.64%, 1.43%, and 1.43%, respectively).
3.5 Safety
Adverse event incidences in the active-comparator period were comparable between treatment groups, with cough and productive cough most commonly reported (Table 3). Severe adverse events were experienced by 22 AZLI-treated (16.2%) and 11 TNS-treated (8.3%) patients (p=0.063; Fisher's exact test), most commonly cough (AZLI: n=7/22; TNS: n=4/11) and productive cough (AZLI: n=4/22; TNS: n=5/11). Adverse events considered by investigators as treatment‐related were reported for 22.8% of AZLI-treated (n=31/136) and 12.9% of TNS‐treated (n=17/132) patients (p=0.039), most commonly respiratory incidents (AZLI: n=19/31; TNS: n=14/17).
Table 3Adverse events occurring during the active-comparator period for ≥10% patients in either treatment group.
Adverse events are coded with Medical Dictionary for Regulatory Activities (MedDRA) Version 11.1. Multiple occurrences per patient were counted once.
AZLI-treated n=136
TNS-treated n=132
Cough
96 (70.6)
104 (78.8)
Productive cough
70 (51.5)
79 (59.8)
Pyrexia
43 (31.6)
40 (30.3)
Oropharyngeal pain
36 (26.5)
37 (28.0)
Dyspnea
32 (23.5)
36 (27.3)
Hemoptysis
31 (22.8)
21 (15.9)
Rales
30 (22.1)
35 (26.5)
Headache
29 (21.3 )
27 (20.5)
Nasal congestion
29 (21.3)
26 (19.7)
Rhinorrhea
25 (18.4 )
33 (25.0)
Exercise tolerance decreased
25 (18.4)
27 (20.5)
Fatigue
24 (17.6)
25 (18.9)
Decreased appetite
19 (14.0)
19 (14.4)
Abdominal pain
18 (13.2)
8 (6.1)
Respiratory tract congestion
16 (11.8)
19 (14.4)
Wheezing
16 (11.8)
20 (15.2)
Chest discomfort
14 (10.3)
13 (9.8)
Nausea
14 (10.3)
10 (7.6)
Vomiting
14 (10.3)
14 (10.6)
Pulmonary function test decreased
11 (8.1)
17 (12.9)
Breath sounds abnormal
8 (5.9)
15 (11.4)
a Adverse events are coded with Medical Dictionary for Regulatory Activities (MedDRA) Version 11.1. Multiple occurrences per patient were counted once.
Fifteen patients discontinued treatment due to adverse events (AZLI: 9; TNS: 1) or unspecified safety and/or tolerability reasons (TNS: 5), most commonly productive cough (AZLI: 3) and hemoptysis (AZLI: 3). Fourteen of these 15 patients subsequently discontinued the study for safety or tolerability reasons (AZLI: 3; TNS: 5), consent withdrawn (AZLI: 3, TNS: 1), or investigator decision (AZLI: 2; Fig. 1).
As was observed for the active-comparator period, cough (n=92/133; 69.2%) and productive cough (n=62/133; 46.2%) were the most commonly reported adverse events in the extension period. Severe adverse events were experienced by 15/133 (11.3%) patients, most commonly cough (n=7/133), productive cough (n=5/133), and pyrexia (n=5/133). One life-threatening event was reported (lung infiltration unrelated to treatment). Two patients died due to complications of CF disease; both deaths were considered unrelated to treatment.
Adverse events considered by investigators as treatment‐related were reported for 6.8% of patients (n=9/133), most commonly respiratory incidents (n=6/133).
3.6 Microbiology
In the active-comparator period, no changes (≥4-fold) were observed in MIC50 values of any antibiotics for all PA isolates in both treatment groups and the proportion of AZLI-treated patients with PA isolates with an aztreonam MIC>8 μg/mL (parenteral breakpoint [
]) increased from 33.9% to 49.1%. The percentage of patients with multi‐drug resistant PA was comparable at baseline and Week 24. No clinically concerning changes were observed in PA susceptibility to other aminoglycoside, quinolone, or beta‐lactam antibiotics. The prevalence of other respiratory pathogens remained similar for both treatment groups.
During the extension period, AZLI/AZLI patients had no changes (≥4-fold) observed in MIC50 values of any antibiotics for all PA isolates, except for 3 intermittent increases in aztreonam and 1 intermittent increase in ticarcillin/clavulanate. TNS/AZLI-treated patients had no changes (≥4‐fold) during either period in MIC50 values of any antibiotics for all PA isolates.
The proportion of AZLI/AZLI-treated patients with PA isolates with an aztreonam MIC>8 μg/mL increased from 35.2% to 50.0% during the study. The proportion of TNS/AZLI‐treated patients with PA isolates with a tobramycin MIC>4 μg/mL increased from 37.3% to 46.0% during the active-comparator phase, while the proportion with PA isolates with an aztreonam MIC>8 μg/mL increased from 38.3% to 49.0% during the extension phase. The percentage of patients with multi‐drug resistant PA was comparable at study baseline and study end. The prevalence of other respiratory pathogens remained similar for both treatment groups. Burkholderia cepacia complex was not isolated during the study.
4. Discussion
Comparative efficacy trials can demonstrate the value of newly developed therapies compared with existing therapies. For CF patients, inhaled antibiotics comparisons can be assessed by lung function, measurements of pulmonary exacerbations, respiratory symptoms, surrogate markers of health (e.g., weight gain/loss), adverse events, and changes in microbial antibiotic sensitivity. Crossover designs, such as the open-label extension in this study, can also show incremental benefits of a new therapy.
This study demonstrated superiority of AZLI to TNS in lung function improvement after 28 days and superiority across 3 treatment courses. Compared with TNS, AZLI also reduced the total number of pulmonary exacerbations, delayed time to need for additional antibiotic therapies, improved respiratory symptoms, and increased weight in a contemporary CF population that was considered stable at baseline. Importantly, patients with prior TNS experience responded to AZLI treatment during the active-comparator portion of the study, and patients treated with TNS in the active-comparator portion of the study showed incremental improvement in the cross-over period, achieving the same benefit in lung function responses as patients initially randomized to AZLI.
Sputum PA density decreased equally for TNS and AZLI groups in the current study, suggesting a similar antibacterial effect for the 2 treatment groups. Given confounding factors (e.g., patient ability to produce sputum throughout a clinical trial; regional intra-pulmonary variability of bacterial density) these data must be interpreted with caution, in light of the marked differences in clinical benefit favoring AZLI treatment.
AZLI was well tolerated with an adverse-event profile consistent with previous clinical trials. Drug‐related adverse events were reported for significantly more AZLI‐treated than TNS-treated patients. In this unblinded study, investigators may have attributed more adverse events to a less familiar antibiotic (AZLI) than to a drug used for over a decade (TNS). Also, patients with a history of intolerance to TNS were excluded from study enrollment. Concerning changes in PA susceptibilities were not observed for either treatment group.
The AZLI–TNS treatment effect for FEV1% predicted at Day 28 in this study (+7.80%) was comparable to the treatment effects observed in previous phase 3 AZLI‐placebo studies (+10.2%; +6.6%) [
]. The 0.55% improvement in FEV1% predicted observed at Day 28 of TNS treatment in this study was much smaller than the 9.7% TNS‐placebo treatment effect observed at Day 28 in early TNS studies, but is comparable to the effect seen in a more recent study with TNS-experienced patients [
]. Pulmonary exacerbations are associated with step-wise loss of lung function and deleterious effects on quality of life. These events also incur significant costs, whether treated in-hospital or with home IV-antibiotics [
Hospitalization risk of current standard of care (SOC) vs. aztreonam for inhalation solution (AZLI) in patients with cystic fibrosis (CF) [abstract 806].
]. AZLI‐treated patients experienced a prolonged time-to-need for IV antipseudomonal antibiotics for a respiratory event compared with TNS-treated patients, and had significantly fewer respiratory hospitalizations and respiratory events requiring IV and/or additional inhaled antipseudomonal antibiotics. Patients receiving additional antipseudomonal antibiotics for exacerbations were allowed to continue in the study; these additional treatments likely supplemented their FEV1 responses over the course of the trial. While FEV1 responses at the end of each off-treatment cycle were similar for both treatment groups, significantly more IV/inhaled antibiotics were used for TNS‐treated patients in order to maintain lung function; 38% of AZLI‐treated patients vs. 58% of TNS-treated patients received additional therapies during the active‐comparator portion of the study. After adjusting for the use of additional antipseudomonal antibiotics, the AZLI–TNS treatment differences were better maintained across 3 treatment cycles. Although FEV1 was below baseline for both groups at Week 24, the AZLI treatment effects in each of the extension phase courses were comparable in magnitude to those observed for the first 3 cycles. Similarly, favorable treatment responses to repeated AZLI courses were observed during a prior 18‐month study [
An 18-month study, AIR-CF3, of the safety and improvement in pulmonary function and respiratory symptoms with repeated courses of aztreonam for inhalation solution in patients with cystic fibrosis and airway Pseudomonas aeruginosa.
]. Additional evidence of improved patient well-being associated with AZLI treatment was the increase above baseline values for body weight observed during AZLI treatment compared with the decrease during TNS treatment.
The open-label design is a limitation of this study; differences in taste, administration frequency, and approved nebulizers did not allow blinding the study. A double‐dummy design would impose an excessive treatment burden on study participants. In this open-label study, AZLI‐treated patients with TNS experience may have been more likely to report improvements in respiratory symptoms or treatment satisfaction with AZLI than with TNS. A further limitation is that only 4.9% of the patients enrolled in this study were 6 to 12 years of age (n=13/268). This reflects the contemporary CF population; children with CF are generally not chronically infected with PA and they do not tend to have diminished lung function. Results from an ongoing clinical trial assessing the safety of AZLI in children with CF and chronic PA infection will provide more data for the pediatric population (Clinicaltrials.gov: NCT01404234).
Another perceived limitation is the inclusion of TNS-experienced patients. Since TNS is standard of care for CF patients with chronic PA infection, enrolling only TNS-naïve patients would not be feasible. Further, in a comparative efficacy study, it is necessary to enroll the group of patients receiving existing standard of care therapy, to allow generalization of the results to the overall patient population. Although treatment paradigms vary by both country and center, several lines of evidence suggest that the population in this study was representative of the general CF population with chronic PA infection. During 2010, 68.3% patients with chronic PA infection in the US registry received inhaled tobramycin [
In the current study, AZLI (75 mg; 3 times a day) and TNS (300 mg; 2 times a day) doses were used in accordance with the respective EMA and FDA approved labels. Dosing regimens have been established during the development of each drug on the basis of efficacy results in clinical trials [
]. The difference in the approved dosing frequencies, in part, reflects the difference in the mechanisms of action for these 2 antibiotics. The antimicrobial activity of aztreonam is time dependent, relating to the elapsed time above a specific drug concentration in the airway. The activity of tobramycin is concentration dependent, thus its antibiotic effects will depend on the peak drug concentration achieved in the airway. The availability of 2 inhaled antibiotics approved for use in CF patients with chronic PA infection now provides the opportunity to prospectively study the benefit of a continuous alternating therapy regimen of 2 antibiotics compared to the standard of care alternate-month monotherapy. Eliminating off-treatment periods has the potential to better preserve lung function, control respiratory symptoms, and reduce the rates of acute pulmonary exacerbations and hospitalizations.
In conclusion, AZLI demonstrated statistical superiority to TNS over the first comparative treatment course, and maintained superiority to TNS across 3 treatment courses in a patient cohort that was representative of the current CF population receiving standard of care treatment for chronic PA infection. AZLI-treated patients had fewer respiratory hospitalizations than TNS‐treated patients and less frequently used IV or additional inhaled antipseudomonal antibiotics for respiratory events. This study demonstrates that AZLI provides an effective and well tolerated treatment option for CF patients with chronic airway PA infection.
5. Study investigators
In addition to the authors, the other members of the AZLI Active Comparator Study Team were:
Austria:
Helmuth Ellemunter and Isidor Huttegger
Belgium:
Frans de Baets, Kristin Desager, Lieven DuPont and Anne Malfroot
France:
Marc Albertini, Chantal Belleguic, Christophe Delacourt, Philippe Domblides, Jean Francois Duhamel, Marcel Guillot, Gerard Lenoir, Sylvie Leroy, Jean Claude Pautard and Natascha Remus
Germany:
Thomas Biedermann, Eckard Hamelmann, Hans-Eberhardt Heuer, Uwe Mellies, Jens Schreiber, Doris Staab, Jens-Oliver Steiss, Helmut Teschler and Hubert Wirtz
Ireland:
Gerard Canny, Charles Gallagher, Peter Greally and Mary Herzig
Carmen Antelo, Jose Villa Asensi, Gloria Garcia Hernandez, Luis Maiz and Francisco Peres‐Frias
Switzerland:
Annette Boehler and Alexander Moeller
United Kingdom:
Mary Carroll, Steven Conway, Stuart Elborn, Charles Haworth, Ian Ketchell, Christopher Taylor and Martin Walshaw
United States:
Frank Accurso, Steven Boas, Aaron Chidekel, Rubin Cohen, Paul Comber, Cori Daines, Robert Fink, David Geller, Gavin Graff, Denis Hadjiliadis, Jeffrey Hoag, Theodore Liou, Craig Nakamura, Edward Naureckas, David Nichols, Bruce Nickerson, Peggy Radford, Santiago Reyes, Dion Roberts, Mark Rolfe, Howard Schmidt, Pedro Sepulveda, Terry Spencer, Arvey Stone, Bruce Trapnell, Pierre Vauthy and Isabel Virella‐Lowell
Sub-investigators for this study included:
Austria:
Johannes Eder, Watraud Eder, Katalin Kovacs and Udo Langenhorst
Belgium:
Elke Dewachter, Valerie Dufresne, Michel Dumonceaux, Monique Lequesne, Alain Michils, Pascal Van Bleyenbergh and Sabine Van Daele
Cordula Conrad-Kabbe, Norbert Grammann, Jorg Grobe-Onnebrink, Andrea Guttler, Andrea Jobst, Swante Klahr, Cordula Koerner-Rettberg, Sabine Matena, Susanne Nahrig, Hans Christoph Runge, Carsten Schwarz, Wolfgang Sextro, Albrecht Tacke, Kathrin Christina Weber, Diane Wieczorek and Monika Wolf
Ireland:
Peter Barry, Sanjay Chotirmall, Basil Elnazir, Cedric Gunaratnam, Edina Moylett and Trevor Nicolson
Italy:
Federico Alghisi, Giuseppe Cimino, Natalia Cirilli, Marco Cipolli, Laura Claut, Diana Costantini, Annarita Costantino, Luisa De Cristofaro, Fabiola De Gregorio, Violetta Di Pierantonio, Valeria Di Stefano, Nadia Faelli, Francesco Longo, Lucia Spicuzza, Gabriella Traverso, Maria Candida Tripodi, Patrizia Troiani and Giovanna Vitaliti
Netherlands:
Geert Jan Wesseling
Portugal:
Pilar Azeved, Joao Bento, José Cavaco and Luisa Pereira
Spain:
Maria Isabel Gonzalez Alvarez, Pilar Caro Auvilena, Isabel Barrio, Maria Castro Codesal, Lucrecia Suarez Cortina, Marta Ruiz de Valbuena, Adelaida Lamas Ferreiro, Adolfo Sequeiros Gonzalez, Carmen Martinez, Maria Teresa Martinez, Estela Perez-Ruiz, Antonio Salcedo Posadas and Elena Urgelles
Switzerland:
Klaus Haemmarle, Markus Hoefer, Demet Inci and Renate Spinas
United Kingdom:
Ina Aldag, Helen Barker, Gary Connett, Jane Davies, Christine Etherington, Andres Floto, Ben Green, Khin Ma Gyi, Mark Ledson, Anand Palamarthy, Felicity Perrin, Jacqueline Rendall and Dennis Wat
United States:
James Acton, Gabriel Aljadeff, Lisa Allwein, Raouf Amin, Bruce Barnett, Sandy Bartosi, Michael Biggin, Debra Boyer, Mark Brown, Terry Byars, Holly Carveth, Eduardo Cepeda-Davila, Barbara Chatfield, Michael Daines, Ruth DeVoogd, Karim Djekidel, Carolyn Donovan, Henry Dorkin, Lorrie Duan, Dawn Ericson, Monica Federico, Peter Fornos, Fadel Fuiz, James Goodwin, Gerald Gong, Roni Grad, Tarik Haddad, Anas Hadeh, Robert Heinle, Laura Herpel, Peter Hiatt, David Hicks, Douglas Kyle Hogarth, Douglas Holsclaw, Angelica Honsberg, Ashley Jones, Marion Jones, Diane Kitch, Lucille Lester, Susan Lester, Jeffrey Lewis, Floyd Livingston, Thomas Martin, Fernando Martinez, Martin Martinez, Margo Moore, Gary Mueller, Angala Narasimhan, Imre Noth, Adrian O'Hagan, Gary Onady, Raj Padman, Reynold Panettieri, Kathleen Peeke, Terri Phillips, Diane Rhodes, Jonathan Rosen, Frederick Royce, Milene Saavedra, Scott Sagel, Connie St. Clair, Gregory Sawicki, Scott Schroeder, Edward S. Schulman, Michael Schwartz, Amrapali Shah, Michael Sherman, Thomas C. Smith, Patrick Sobande, James Stark, Antine Stenbit, Mary Strek, Alumet Uluer, Robert Vender, William Walsh, Ronald Williams, Elizabeth Woods, James Woodward, Jamie Wooldridge, Nozomi Yagi, Robert Young and Anchalee Yuengsrigul
Conflict of interest statement
DB had consultancies with Gilead Sciences, Insmed and Aradigm, and received a research grant from Novartis. MF had a consultancy with Gilead Sciences. RF had consultancies with Gilead Sciences, MPEX, Gruenenthal, and Novartis, and participated in speaker's bureaus for Gilead Sciences, Astra Zeneca, Astellas, and Roche. ML participated in speaker's bureaus for Chiesi Farmaceutici and Mediolanum. CMO had a consultancy with Gilead Sciences and received a research grant from Gilead Sciences. SAL, MB, and ABM are employees and/or shareholders of Gilead Sciences. All other authors have no perceived conflicts of interests.
Role of the Funding Source
This study was sponsored by Gilead Sciences.
BMA, TP, MB, and ABM participated in study design. BMA, TP, DB, MF, RF, RC, ML, CK, NM, and CMO were clinical investigators for the study. SAL oversaw statistical analyses. SAL, MB, ABM, and CMO wrote and edited the draft manuscript. All authors revised the manuscript and approved the final version for submission. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
Acknowledgments
We thank the patients and their families. Study conduct was managed by Laura Haas and Sheila Leitzinger (Gilead Sciences, Inc.). Statistical analyses were performed by Lixin Shao, (Gilead Sciences, Inc.). Writing assistance was provided by Kate Loughney, under the sponsorship of Gilead Sciences. Independent Data Review Committee for Respiratory Events: Marcia Katz (Baylor University, Houston TX) and Patrick Flume (Medical University of South Carolina, Charleston SC). Cystic Fibrosis Foundation Therapeutics Data and Safety Monitoring Board: Lynne M Quittell (Children's Hospital of New York; Columbia University, New York NY), Richard A Kronmal (University of Washington, Seattle WA), Natalie Neu (Columbia University, New York, NY), and Margaret (Lou) Guill (Dartmouth–Hitchcock Medical Center, Lebanon NH).
Relative change from baseline FEV1% predicted at Day 28 for patient subgroups. The adjusted means are from an ANCOVA model that included terms for treatment, baseline FEV1% predicted, and previous inhaled tobramycin use. Missing data were imputed with the last observation carried forward method. The US patient subgroup did not include any patients with inhaled tobramycin use <84 days in the previous year. At the bottom of the figure, the SE bars for the TNS and AZLI data sets overlap for patients <18 years of age and for patients with prior inhaled tobramycin use <84 days. Moderate disease severity=FEV1>50% to ≤75% predicted at baseline and severe disease severity=FEV1≤50% predicted at baseline.
Adjusted mean relative change in body weight (%) from study baseline by study visit for patients participating in the AZLI open-label extension period. Significant differences between the AZLI and TNS treatment groups were observed at Week 4 (p=0.004), Week 12 (p=0.003), and Week 16 (p=0.009).
References
Flume P.A.
O'Sullivan B.P.
Robinson K.A.
Goss C.H.
Mogayzel Jr., P.J.
Willey-Courand D.B.
et al.
Cystic fibrosis pulmonary guidelines. Chronic medications for maintenance of lung health.
An 18-month study, AIR-CF3, of the safety and improvement in pulmonary function and respiratory symptoms with repeated courses of aztreonam for inhalation solution in patients with cystic fibrosis and airway Pseudomonas aeruginosa.
Validation of a general measure of treatment satisfaction, the treatment satisfaction questionnaire for medication (TSQM), using a national panel study of chronic disease.
the CF Foundation Consensus Conference on Infection Control Participants
Infection control recommendations for patients with cystic fibrosis: microbiology, important pathogens, and infection control practices to prevent patient-to-patient transmission.
Hospitalization risk of current standard of care (SOC) vs. aztreonam for inhalation solution (AZLI) in patients with cystic fibrosis (CF) [abstract 806].
☆Study results were reported at the 24th Annual North American Cystic Fibrosis Conference, Baltimore Convention Center, Baltimore, Maryland, USA, 21–23 October 2010; the American Thoracic Society 2011 International Conference, Denver, Colorado, USA, 13–18 May 2011; the 34th European Cystic Fibrosis Conference, Hamburg, Germany, 8–11 June 2011; and the 25th Annual North American Cystic Fibrosis Conference, Anaheim, California, USA, 3–5 November 2011.
00136-1&id=gr1.jpg){kind=link}
00136-1&id=gr2.jpg){kind=link}
00136-1&id=gr3.jpg){kind=link}