Articles. Funding Gilead Sciences.

Aztreonam for inhalation solution in patients with non-cystic fibrosis bronchiectasis (AIR-BX1 and AIR-BX2): two randomised double-blind, placebo-controlled phase 3 trials Alan F Barker, Anne E O Donnell, Patrick Flume, Philip J Thompson, Jonathan D Ruzi, Javier de Gracia, Wim G Boersma, Anthony De Soyza, Lixin Shao, Jenny Zhang, Laura Haas, Sandra A Lewis, Sheila Leitzinger, A Bruce Montgomery, Matthew T McKevitt, David Gossage, Alexandra L Quittner, Thomas G O Riordan Lancet Respir Med 2014; 2: 738 49 Published Online August 19, 2014 http://dx.doi.org/10.1016/ S2213-2600(14)70165-1 See Comment page 679 Department of Medicine, Division of Pulmonary and Critical Care, Oregon Health and Science University, Portland, OR, USA (Prof A F Barker MD); Division of Pulmonary, Critical Care & Sleep Medicine, Department of Medicine, Georgetown University, Washington DC, USA (Prof A E O Donnell MD); Departments of Medicine and Pediatrics, Medical University of South Carolina, Charleston SC, USA (Prof P Flume MD); Centre for Asthma, Allergy and Respiratory Research, University of Western Australia, Nedlands, WA, Australia (Prof P J Thompson FRACP); Scottsdale Healthcare Shea Medical Center, Scottsdale, AZ, USA (J D Ruzi MD); Department of Pneumology, Hospital Universitari Vall d Hebron, Barcelona, Spain (Prof J de Gracia MD); Department of Pulmonary Diseases, Medical Centre Alkmaar, Alkmaar, Netherlands (Prof W G Boersma MD); Freeman Hospital Bronchiectasis Service and Transplantation and Immunobiology Group, Institute of Cellular Medicine, Newcastle University, Newcastle-upon-Tyne, UK (A De Soyza MD); Gilead Sciences Inc., Seattle, WA, USA (L Shao MD, J Zhang PhD, L Haas BA, S A Lewis MS, S Leitzinger BS, M T McKevitt PhD, D Gossage MD, T G O Riordan MD); Cardeas Pharma Corp., Seattle, WA, USA(A B Montgomery MD); and Summary Background The clinical benefit of inhaled antibiotics in non-cystic fibrosis bronchiectasis has not been established in randomised controlled trials. We aimed to assess safety and efficacy of aztreonam for inhalation solution () in patients with non-cystic fibrosis bronchiectasis and Gram-negative bacterial colonisation. Methods AIR-BX1 and AIR-BX2 were two double-blind, multicentre, randomised, placebo-controlled phase 3 trials, which included patients aged 18 years or older who had bronchiectasis and history of positive sputum or bronchoscopic culture for target Gram-negative organisms. Patients were randomly assigned to receive either or placebo (1:1). Randomisation was done without stratification and the code was generated by a Gilead designee. In both studies, two 4-week courses of 75 mg or placebo (three-times daily; eflow nebulizer) were each followed by a 4-week off-treatment period. Primary endpoint was change from baseline Quality of Life-Bronchiectasis Respiratory Symptoms scores (QOL-B-RSS) at 4 weeks. These trials are registered with ClinicalTrials.gov, numbers are NCT01313624 for AIR- BX1 and NCT01314716 for AIR-BX2. Findings We recruited participants from 47 ambulatory clinics for AIR-BX1 and 65 ambulatory clinics for AIR-BX2; studies were done between April 25, 2011, and July 1, 2013. In AIR-BX1, of the 348 patients screened, 134 were randomly assigned to receive and 132 to receive placebo. In AIR-BX2, of the 404 patients screened, 136 were randomly assigned to receive and 138 to receive placebo. The difference between and placebo for adjusted mean change from baseline QOL-B-RSS was not significant at 4 weeks (0 8 [95% CI 3 1 to 4 7], p=0 68) in AIR- BX1, but was significant (4 6 [1 1 to 8 2], p=0 011) in AIR-BX2. The 4 6 point difference in QOL-B-RSS after 4 weeks in AIR-BX2 was not deemed clinically significant. In both studies, treatment-related adverse events were more common in the group than in the placebo group, as were discontinuations from adverse events. The most commonly reported treatment-emergent adverse events were dyspnea, cough, and increased sputum. Each was more common for -treated than for placebo-treated patients, but the incidences were more balanced in AIR-BX2. Interpretation treatment did not provide significant clinical benefit in non-cystic fibrosis bronchiectasis, as measured by QOL-B-RSS, suggesting a continued need for placebo-controlled studies to establish the clinical benefit of inhaled antibiotics in patients with this disorder. Funding Gilead Sciences. Introduction Bronchiectasis is a chronic lung disease in which small and medium-sized airways are permanently dilated and predisposed to persistent bacterial infections. Symptoms include chronic cough, daily mucopurulent sputum production, dyspnoea, and fatigue. Acute exacerbations are associated with increased cough, increased sputum production, pleuritic pain, and worsening systemic symptoms. 1,2 Bacteria are consistently isolated in sputum or bronchoscopic specimens from clinically stable patients with non-cystic fibrosis bronchiectasis, most commonly Haemophilus influenzae and other Gramnegative bacteria including Pseudomonas aeruginosa, with H influenzae infection associated with less severe disease. 3 In view of the similar pathophysiology, microbiology, and symptomatology, present non-cystic fibrosis bronchiectasis treatment is often extrapolated from efficacious cystic fibrosis treatments, including use of inhaled antibiotics. 4 However, clinical benefit has not been consistently recorded in non-cystic fibrosis bronchiectasis, 5 15 which might partly indicate the paucity of validated clinical trial endpoints. 16 Aztreonam for inhalation solution (; Cayston, Gilead Sciences) is an antipseudomonal antibiotic formulated for inhalation, which decreases respiratory symptoms, delays time to pulmonary exacerbation, and improves lung function in patients with cystic fibrosis with chronic P aeruginosa infection. 17 19 In AIR-BX1 and AIR-BX2, we assessed the safety and efficacy of in 738 www.thelancet.com/respiratory Vol 2 September 2014

patients with non-cystic fibrosis bronchiectasis and Gram-negative bacterial infection. Methods Study design and participants AIR-BX1 and AIR-BX2 were identical double-blind, multicentre, randomised, placebo-controlled, phase 3 trials. Eligible patients (aged 18 years) had bronchiectasis confirmed by CT chest scan, history of positive sputum or bronchoscopic culture for target Gram-negative organism or treatment of exacerbation (in the previous 5 years) with antibiotics with Gram-negative coverage, positive sputum culture for target Gram-negative bacteria (at screening), chronic sputum production ( 4 days/week; in the previous 4 weeks), forced expiratory volume in 1 s (FEV 1 ) of 20% predicted or higher after bronchodilator (at screening), and recent chest X-ray at screening or between screening and baseline without substantial acute findings (eg, no new infiltrate). Pre-defined target Gram-negative respiratory pathogens were species of Pseudomonas, as well as Achromobacter, Burkholderia, Citrobacter, Enterobacter, Escherichia, Klebsiella, Moraxella, Proteus, Serratia, and Stenotrophomonas. Presence of H influenzae alone did not meet study inclusion criteria. Key exclusion criteria included recent hospital admission (in the 2 weeks before screening); previous hospital admission for embolisation for treatment of haemoptysis; antibiotic use for respiratory symptoms (apart from chronic stable macrolide treatment) or haemoptysis of more than 30 ml (from 2 weeks before screening until baseline); serious adverse event (from screening to baseline); changes in other treatments (bronchodilator, corticosteroid, macrolide, or bronchial hygiene; 4 weeks before screening through to study completion); change in systemic corticosteroid treatment (4 weeks before screening through to baseline; after baseline, 14-day courses allowed for worsening respiratory signs or symptoms); current treatment for non-tuberculous mycobacteria infection; active Myco bacterium tuberculosis infection (previous year); previous treatment; history of cystic fibrosis; and pregnancy, lactation, or no acceptable birth control. Studies were done in accordance with principles of the Declaration of Helsinki, International Conference on Harmonisation guidelines (ICH-GCP), and good clinical practice principles. At every site, institutional review boards or ethics committees approved the study. Patients provided written informed consent before study participation. Randomisation and masking Patients were randomly assigned to either receive or placebo (1:1). Randomisation was done without stratification and the code was generated by a Gilead designee. Randomisation was done at baseline with an interactive voice and web response system. Treatment assignments were masked to patients, site personnel, study vendors, and the sponsor, apart from designated personnel reviewing randomisation and drug allocation; such personnel were independent of the data analyses. The data monitoring committee regularly monitored unmasked safety data and made recommendations to the sponsor about whether the study should proceed as planned. and placebo appeared identical. does not have a noticeable taste; therefore, placebo was not taste-masked. Procedures Each study included two 4-week courses of double-blind inhaled treatment with 75 mg or placebo given three times a day; followed by 4 weeks off-treatment (figure 1). Patients received a β-2 agonist bronchodilator before and placebo doses. Patients who had worsening respiratory signs or symptoms after baseline could receive non-study antibiotics without discontinuing the study. We assessed efficacy with the Quality of Life- Bronchiectasis (QOL-B) questionnaire (version [V] 3.0), a patient-reported outcome measure for non-cystic fibrosis bronchiectasis, which was self-administered before other study procedures at the beginning of all scheduled and non-scheduled visits (Quittner AL, personal communication). 20,21 Details of when other measures were administered is provided in the appendix. Outcomes The primary efficacy endpoint was change in QOL-B Respiratory Symptoms scores (QOL-B-RSS; baseline to week 4; scores are 0 100, with high scores representing few symptoms; Quittner AL, personal communication). 20,21 Secondary endpoints were change in QOL-B-RSS from baseline to week 12 and time to the first protocol-defined exacerbation through day 112 (week 16). The definition of an exacerbation was modified from Bilton and colleagues 9 and O Donnell and colleagues 22 in consultation with scientific advisors and investigators Study week Study day (±3) Visit 2 14 1 Randomisation Screening (14 days) 0 0 2 controlled (28 days) Follow-up 1 course 1 (28 days) 2 14 3 4 28 4 8 56 5 controlled (28 days) Follow-up 2 course 2 (28 days) 12 84 6 16 112 7 Department of Psychology and Pediatrics, University of Miami, and Behavioral Health Sciences Research, Coral Gables, FL, USA (Prof A L Quittner PhD) Correspondence to: Prof Alan F Barker, Department of Medicine, Division of Pulmonary and Critical Care, Oregon Health and Science University, Portland, OR 97239, USA barkera@ohsu.edu Open-label course (28 days) See Online for appendix 20 140 8 Open-label follow-up (56 days) 24 168 9 End of study Figure 1: AIR-BX1 and AIR-BX2 study design Baseline was week 0 (beginning of course 1). Results from the third 4-week treatment course (with open-label 75 mg) and the 56-day follow-up period are not presented in this report. =aztreonam for inhalation solution. 28 196 10 www.thelancet.com/respiratory Vol 2 September 2014 739

A AIR-BX1 348 patients screened 82 excluded* 50 did not meet inclusion criteria 20 met exclusion criteria 5 withdrew consent 8 other 266 randomised 134 assigned to receive 132 assigned to receive placebo 134 received 132 received placebo 38 did not complete study period 27 safety or tolerability reasons 9 withdrew consent 1 investigator s decision 1 other 10 did not complete study period 4 safety or tolerability reasons 4 withdrew consent 2 lost to follow-up 96 completed 16 week study period 122 completed 16 week study period B AIR-BX2 404 patients screened 130 excluded* 79 did not meet inclusion criteria 35 met exclusion criteria 10 withdrew consent 11 other 274 randomised 136 assigned to receive 138 assigned to receive placebo 1 not treated (other reasons) 1 not treated (withdrew consent) 135 received 137 received placebo 16 did not complete study period 10 safety or tolerability reasons 5 withdrew consent 1 other 14 did not complete study period 5 safety or tolerability reasons 6 withdrew consent 1 lost to follow-up 2 other 119 completed 16 week study period 123 completed 16 week study period Figure 2: Trial profile (A) AIR-BX1. (B) AIR-BX2. =aztreonam for inhalation solution. *Patients could fail screening for more than one reason; the appendix shows a detailed summary of patients who failed screening due to not meeting an inclusion criteria and/or meeting an exclusion criteria. 740 www.thelancet.com/respiratory Vol 2 September 2014

and required: acute worsening of respiratory disease meeting at least three major (or two major and at least two minor) criteria. Major criteria were increased sputum production, change in sputum colour, dyspnoea, and cough. Minor criteria were fever (>38 C) at clinic visit, increased malaise or fatigue, FEV 1 (L) or forced vital capacity decreased by more than 10% from baseline, and new or increased haemoptysis. The appendix describes additional endpoints. Study results assessing the psychometric properties of the QOL-B are reported elsewhere (Quittner AL, personal communication). Microbiology assessments included change in colony forming units (CFU) per g of sputum, presence or absence of respiratory pathogens, and changes in minimum inhibitory concentrations (MIC) of aztreonam. Statistical analysis A sample size of 266 patients per study provided 95% or higher power to detect a ten-point difference between groups for QOL-B-RSS (two-sided, 0 05-level test), assuming a common standard deviation of 20 and minimal missing data, and provided 80% power to detect a difference in time to first protocol-defined exacerbation (with 65 events expected over 112 days), assuming a hazard ratio (HR) of 0 5 and 15% dropout rate. Efficacy analyses included all patients randomly assigned to treatment (intention-to-treat population). Safety analyses included all treated patients. For the primary endpoint and week 12 QOL-B-RSS analyses, we used mixed-effect model repeated measures (two-sided; 0 05-level). We summarised time to protocol-defined exacerbation through week 16 (Kaplan Meier statistics) and analysed it with log-rank test. We tested additional endpoints for descriptive purposes (0 05-level). A gatekeeper multiple testing procedure was prespecified to control the familywise error rate of 0 05 across primary and secondary endpoints. We stopped the formal sequential testing when the null hypothesis for the primary endpoint was not rejected. We made no other adjustments for multiple testing; only nominal significance is cited. Separate AIR- BX1 and AIR-BX2 (not pooled) analyses were prespecified. We explored baseline predictive factors using classification and regression tree and analysis of variance model selection methods. Statistical analyses used SAS V9 2 (SAS Institute, Cary, NC). These trials are registered with ClinicalTrials.gov, numbers are NCT01313624 for AIR-BX1 and NCT01314716 for AIR- BX2. Role of the funding source The funder of these studies was involved in study design, in collection, analysis, and interpretation of data; and in writing of the report. AFB, AEOD, LS, JZ, LH, SAL, SL, MTM, DG, and TGOR had access to raw data. The corresponding author had full access to all data and the final responsibility to submit for publication. AIR-BX1 (n=134) (n=132) Total (n=266) AIR-BX2 (n=136) (n=138) Total (n=274) Age, years Mean (SD) 64 2(12 9) 64 9 (12 1) 64 6 (12 5) 63 3 (14 2) 62 7 (13 3) 63 0 (13 8) Minimum 23 83 20 88 20 88 22 85 18 87 18 87 maximum 65 years 81 (60%) 78 (59%) 159 (60%) 83 (61%) 72 (52%) 155 (57%) Female sex 84 (63%) 97 (73%) 181 (68%) 89 (65%) 101 (73%) 190 (69%) Race Asian 5 (4%) 3 (2%) 8 (3%) 1 (1%) 0 1 (<1%) Black or 3 (2%) 6 (5%) 9 (3%) 2 (1%) 4 (3%) 6 (2%) African heritage White 121 (90%) 119 (90%) 240 (90%) 119 (88%) 128 (93%) 247 (90%) Other or not 5 (4%) 4 (3%) 9 (3%) 14 (10%) 6 (4%) 20 (7%) provided Ethnicity, 12 (9%) 11 (8%) 23 (9%) 5 (4%) 7 (5%) 12 (4%) Hispanic Mean body-mass index (SD) kg/m² 25 0 (5 1) 24 7 (4 9) 24 9 (5 0) 23 9 (5 0) 24 7 (6 0) 24 3 (5 5) Cause of bronchiectasis ABPA 4 (3%) 3 (2%) 7 (3%) 0 3 (2%) 3 (1%) Aspiration/ 5 (4%) 2 (2%) 7 (3%) 5 (4%) 3 (2%) 8 (3%) GERD Ciliary 6 (4%) 5 (4%) 11 (4%) 7 (5%) 4 (3%) 11 (4%) dysfunction Idiopathic 41 (31%) 41 (31%) 82 (31%) 43 (32%) 44 (32%) 87 (32%) Immune 2 (1%) 7 (5%) 9 (3%) 5 (4%) 4 (3%) 9 (3%) defect Post infection 47 (35%) 49 (37%) 96 (36%) 47 (35%) 41 (30%) 88 (32%) Rheumatoid 1 (<1%) 2 (2%) 3 (1%) 2 (1%) 1 (<1%) 3 (1%) arthritis Other 27 (20%) 23 (17%) 50 (19%) 27 (20%) 37 (27%) 64 (23%) Missing 1 (<1%) 0 1 (<1%) 0 1 (1%) 1 (<1%) Region North America 103 (77%) 90 (68%) 193 (73%) 58 (43%) 54 (39%) 112 (41%) Australia 31 (23%) 42 (32%) 73 (27%) 3 (2%) 2 (1%) 5 (2%) Europe 0 0 0 75 (55%) 82 (59%) 157 (57%) Mean QOL-B-RSS 55 0 (19 3) 55 5 (19 3) 55 2 (19 3) 56 2 (18 0) 57 4 (18 1) 56 8 (18 0) (SD) Mean EQ-5D 66 1 (18 1) 69 9 (16 5) 68 0 (17 3) 65 7 (16 0) 68 0 (14 9) 66 8 (15 5) score* (SD) Mean FEV 1 % 60 4 (22 6) 64 5 (18 7) 62 4 (20 8) 63 8 (19 5) 63 4 (21 6) 63 6 (20 5) predicted (SD) FEV 1 % predicted <50% 52 (39%) 33 (25%) 85 (32%) 37 (27%) 42 (30%) 79 (29%) 50% to <80% 49 (37%) 70 (53%) 119 (45%) 72 (53%) 61 (44%) 133 (49%) 80% 33 (25%) 29 (22%) 62 (23%) 27 (20%) 35 (25%) 62 (23%) Mean 6MWT score (SD), m 421 (119 3) 426 (118 5) 423 (118 7) 423 (127 5) 428 (120 8) 426 (124 0) Number of bronchiectasis exacerbations in previous year 0 53 (40%) 45 (34%) 98 (37%) 54 (40%) 48 (35%) 102 (37%) 1 26 (19%) 44 (33%) 70 (26%) 32 (24%) 34 (25%) 66 (24%) 2 27 (20%) 25 (19%) 52 (20%) 20 (15%) 28 (20%) 48 (18%) 3 28 (21%) 18 (14%) 46 (17%) 30 (22%) 28 (20%) 58 (21%) (Table 1 continues on next page) www.thelancet.com/respiratory Vol 2 September 2014 741

AIR-BX1 (n=134) (n=132) Total (n=266) AIR-BX2 (n=136) (n=138) Total (n=274) (Continued from previous page) P aeruginosa 112 (84%) 105 (80%) 217 (82%) 116 (85%) 103 (75%) 219 (80%) present History of COPD 42 (31%) 25 (19%) 67 (25%) 33 (24%) 44 (32%) 77 (28%) History of 16 (12%) 14 (11%) 30 (11%) 8 (6%) 12 (9%) 20 (7%) mycobacterium infection History of 63 (47%) 40 (30%) 103 (39%) 44 (32%) 57 (41%) 101 (37%) smoking Chronic macrolide use 34 (25%) 23 (17%) 57 (21%) 41 (30%) 35 (25%) 76 (28%) Use of non-antibiotic inhaled medications Any LAMA 41 (31%) 31 (23%) 72 (27%) 35 (26%) 44 (32%) 79 (29%) LABA and ICS 46 (34%) 55 (42%) 101 (38%) 53 (39%) 50 (36%) 103 (38%) only LAMA, LABA, and ICS 37 (28%) 22 (17%) 59 (22%) 28 (21%) 32 (23%) 60 (22%) Data are number of patients (%) or mean (SD), unless otherwise stated. =aztreonam for inhalation solution. ABPA=allergic bronchopulmonary aspergillosis. GERD=gastroesophageal reflux disease. QOL-B-RSS=Quality of Life- Bronchiectasis Respiratory Symptoms scores. EQ-5D=European Quality of Life-5 Dimensions. FEV 1 =forced expiratory volume in 1 s. 6MWT=6-minute walk test. COPD=chronic obstructive pulmonary disease. LAMA=long-acting muscarinic antagonist. LABA=long-acting β 2 -agonist. ICS=inhaled corticosteroid. *Data available for AIR-BX1: 131 patients ( group), 131 patients (placebo group); for AIR-BX2: 133 ( group), 134 (placebo group). We defined the presence of Pseudomonas aeruginosa as a positive culture at screening or at baseline, or both. Table 1: Demographics and patient characteristics Results For AIR-BX1, we recruited participants from 47 ambulatory clinics in Australia, Canada, and USA; the study extended from April 25, 2011, to June 4, 2013. For AIR-BX2, we recruited participants from 65 ambulatory clinics in Australia, Belgium, Canada, France, Germany, Italy, the Netherlands, Spain, UK, and USA; the study extended from April 25, 2011, to July 1, 2013. In AIR-BX1, of the 348 patients screened, 266 patients were randomly assigned to receive either (n=134) or placebo (n=132; figure 2). 218 treated patients completed double-blind treatment. Discontinuations for safety or tolerability were higher in the group (27 [20%] patients) than in the placebo group (four [3%) patients; figure 2). In AIR-BX2, of the 404 patients screened, 274 patients were randomly assigned to receive either (n=136) or placebo (n=138; figure 2). 242 treated patients completed double-blind treatment, and discontinuations for safety or tolerability were higher in the group (ten [7%] patients) than in the placebo group (five [4%] patients), but were of lower magnitude than in AIR-BX1 (figure 2). We assessed treatment adherence (expected vial use during time on study) by counting returned vials. Mean adherence was 89 8% (SD 16 6) for the group and 93 5% (10 4) for the placebo group in AIR-BX1, and 94 4% (10 1) for the group and 90 7% (17 6) for the placebo group in AIR-BX2. Patient characteristics were generally well balanced between groups (table 1). However, some characteristics were more common in patients assigned than in those assigned placebo, such as FEV 1 lower than 50% predicted, history of chronic obstructive pulmonary disease (COPD), history of smoking, and combined use of inhaled corticosteroids and long-acting beta-2 agonists and muscarinic antagonists in AIR-BX1, and P aeruginosapositive sputum samples in AIR-BX2. QOL-B-RSS numerically increased in all groups in both studies at weeks 4 and 12, with increases indicating fewer symptoms. However, differences between and placebo for adjusted mean change from baseline QOL-B-RSS were not statistically significant at weeks 4 or 12 in AIR-BX1 or week 12 in AIR-BX2 (table 2). We noted a significant difference at week 4 in AIR-BX2; however the magnitude of the difference (difference in least squares means of 4 6 points [95% CI 1 1 8 2]) was smaller than the minimal important difference (MID) of 8 0 points (which was calculated with data from this study and is reported elsewhere; AL Quittner personal communication). Thus we did not deem this change to be a clinically significant improvement. Differences in individual responses were less pronounced at 12 weeks (figure 3). As estimated by Kaplan-Meier analyses, the median time to first protocol-defined exacerbation was only reached in one group (placebo in AIR-BX1; table 2; appendix). The estimated rates of protocol-defined exacerbation by week 16 were numerically higher in the group than in the placebo group. However, the risk of development of the first protocol-defined exacerbation was not statistically different between treatment groups in both studies (AIR-BX1: HR 1 26, p=0 33; AIR-BX2: HR 1 23, p=0 35). In AIR-BX1, 38 (28%) of 134 patients who were randomly allocated to and 35 (27%) of 132 patients who were randomly allocated to placebo had a protocol-defined exacerbation and were treated with antibiotics. In AIR-BX2, 43 (32%) of 136 patients who were randomly allocated to and 38 (28%) of the 138 patients who were randomly allocated to placebo had a protocol-defined exacerbation and were treated with antibiotics. Rates of protocol-defined exacerbations per patient-year were 1 32 ( group) and 1 08 (placebo group) in AIR-BX1 (p=0 35; negative binomial regression method) and 1 20 ( group) and 1 14 (placebo group) in AIR-BX2 (p=0 81). We did not note any clinical benefit after treatment for other efficacy endpoints (appendix). No patient subgroups consistently showed a clinically meaningful response across studies and endpoints (table 3). -treated AIR-BX2 patients with baseline FEV 1 lower than 50% predicted was the only subgroup with a statistically larger (p=0 010) and clinically significant (>8 0 points) response compared with placebo; the difference in QOL-B-RSS at week 4 between and placebo was 9 6 points (95% CI 2 4 16 9). 742 www.thelancet.com/respiratory Vol 2 September 2014

AIR-BX1 (n=134) (n=132) -placebo difference, HR (95% CI) Change from baseline QOL-B-RSS At week 4, adjusted mean (SE) 6 4 (1 4) 5 6 (1 4) 0 8 ( 3 1 to 4 7) At week 12, adjusted mean (SE) 5 7 (1 6) 4 4 (1 5) 1 3 (-3 0 to 5 6) Time to first protocol-defined exacerbation Kaplan-Meier estimates of patients with protocol-defined exacerbation at week 16 0 33 0 26 HR 1 26 (0 79 to 1 99) Estimated median days (95% CI) NE 120 (117 to NE) p value AIR-BX2 (n=136) (n=138) -placebo difference, HR (95% CI) 0 68* 7 9 (1 3) 3 3 (1 3) 4 6 (1 1 to 8 2) 0 56* 5 2 (1 4) 4 1 (1 4) 1 1 (-2 7 to 5 0) 0 33 0 33 0 27 HR 1 23 (0 80 to 1 91) p value 0 011* 0 56* 0 35 NE NE Data are least squares mean (SE) unless otherwise stated. =aztreonam for inhalation solution. HR=hazard ratio. NE=not estimable. QOL-B-RSS=Quality of Life- Bronchiectasis Respiratory Symptoms scores. *p value is based on t-test from mixed effect model repeat measurement model, which included baseline QOL-B-RSS, treatment, visit, and treatment and visit interaction. p value is based on log-rank test. Table 2: Efficacy data A 100 90 B Cumulative proportion of patients (%) 80 70 60 50 40 30 20 10 MID +8 0 points MID 8 0 points MID +8 0 points MID 8 0 points 0 Cumulative proportion of patients (%) C D 100 90 80 70 60 50 40 30 MID 8 0 points MID 8 0 points 20 10 MID +8 0 points MID +8 0 points 0 80 70 60 50 40 30 20 10 0 10 20 30 40 50 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 Change in QOL-B RSS from baseline Change in QOL-B RSS from baseline Figure 3: Cumulative distribution curves for AIR-BX1 (A, B) and AIR-BX2 (C, D) The cumulative percentages of patients are shown for changes in QOL-B-RSS from baseline for (A) AIR-BX1 at week 4, (B) AIR-BX1 at week 12, (C) AIR-BX2 at week 4, and (D) AIR-BX2 at week 12. Dotted lines show the 8 0 point change from baseline scores that correspond to the MID; +8 0 points equals an improvement from baseline values and 8 0 points equals a worsening from baseline values. =aztreonam for inhalation solution. MID=minimal important difference. QOL-B-RSS=Quality of Life- Bronchiectasis Respiratory Symptom scores. The week 12 response was even larger (13 4 points [7 0 19 7]; p<0 0001); however, time to first protocoldefined exacerbation was similar for - and placebo-treated patients in this subgroup (data not shown; p=0 36). Significant differences in QOL-B-RSS or time to first exacerbation were not noted for patients www.thelancet.com/respiratory Vol 2 September 2014 743

AIR-BX1 vs placebo, n, n Difference in leastsquares means (95% CI) AIR-BX2 vs placebo p value*, n, n Difference in leastsquares means (95% CI) All patients 117 124 0 8 ( 3 1 to 4 7) 0 68 128 132 4 6 (1 1 to 8 2) 0 011 Age 65 years 69 73 0 4 ( 5 4 to 4 6) 0 88 78 68 4 7 (0 1 to 9 3) 0 04 Age <65 years 48 51 2 8 ( 3 5 to 9 0) 0 38 50 64 4 5 ( 1 2 to 10 1) 0 12 Australia (+Europe, AIR-BX2) 25 41 0 2 ( 7 4 to 7 0) 0 95 72 82 4 2 ( 0 4 to 8 7) 0 07 North America 92 83 0 9 ( 3 7 to 5 6) 0 70 56 50 5 5 ( 0 3 to 11 3) 0 06 FEV 1 80% predicted 31 28 3 6 ( 12 2 to 5 0) 0 41 24 35 3 2 ( 4 0 to 10 5) 0 37 FEV 1 50% to <80% predicted 44 64 0 4 ( 5 9 to 6 7) 0 90 69 61 2 6 ( 2 3 to 7 6) 0 30 FEV 1 <50% predicted 42 32 4 9 ( 1 0 to 10 8) 0 10 35 36 9 6 (2 4 to 16 9) 0 01 Women 73 90 0 8 ( 4 1 to 5 7) 0 74 83 96 5 7 (1 4 to 10 0) 0 01 Men 44 34 0 8 ( 5 7 to 7 3) 0 80 45 36 3 5 ( 2 7 to 9 6) 0 27 Chronic macrolide use: no 87 101 0 6 ( 3 8 to 5 0) 0 79 87 98 3 9 ( 0 1 to 7 9) 0 06 Chronic macrolide use: yes 30 23 0 5 ( 7 9 to 8 9) 0 91 41 34 5 9 ( 1 5 to 13 2) 0 12 Use of any LAMA: no 84 96 0 3 ( 5 0 to 4 5) 0 91 94 91 5 6 (1 4 to 9 8) 0 01 Use of any LAMA: yes 33 28 3 3 ( 3 5 to 10 0) 0 34 34 41 1 4 ( 5 0 to 7 8) 0 67 Use of LABA and ICS only: no 77 72 1 1 ( 3 7 to 5 9) 0 66 78 83 5.4 (1 1 to 9 7) 0 015 Use of LABA and ICS only: yes 40 52 0 1 ( 6 6 to 6 8) 0 97 50 49 3 3 ( 2 9, 9 5) 0 29 Use of LAMA, LABA, and ICS: no 87 105 0 5 ( 5 0 to 4 1) 0 84 100 103 5 3 (1 3 to 9 2) 0 009 Use of LAMA, LABA, and ICS: yes 30 19 4 7 ( 3 0 to 12 4) 0 23 28 29 1 8 ( 6 2 to 9 7) 0 66 Pseudomonas spp: absent 18 25 3 6 ( 13 5 to 6 3) 0 47 18 33 1 4 ( 9 0 to 6 2) 0 72 Pseudomonas spp: present 99 99 1 5 ( 2 7 to 5 7) 0 47 110 99 5.8 (1 8 to 9 9) 0 005 Smoking status: never 60 86 3 1 ( 7 7 to 1 6) 0 20 88 77 7.1 (2 9 to 11 3) 0 001 Smoking status: ever 57 38 6 6 ( 0 4 to 13 5) 0 06 40 55 0 1 ( 6 5 to 6 2) 0 97 All patients, excluding data reported after antibiotic use 83 108 0 7 ( 3 6 to 4 9) 0 76 110 111 5 8 (2 1 to 9 6) 0 002 =aztreonam for inhalation solution. ICS=inhaled corticosteroid. LABA=long-acting beta 2 -agonist. LAMA=long-acting muscarinic antagonist. MMRM=mixed model repeated measures. QOL-B-RSS=Quality of Life-Bronchiectasis Respiratory Symptoms scores. FEV 1 =forced expiratory volume in 1 s. *p value is based on t-test from MMRM model, which included baseline QOL-B-RSS, treatment, visit, and treatment and visit interaction. Table 3: Change from baseline QOL-B-RSS at week 4 p value* with baseline FEV 1 lower than 50% predicted in AIR- BX1. Exploratory analyses did not identify patient characteristics predictive of efficacy responses. Decreases in CFU per g of sputum for target Gramnegative organisms were larger for -treated patients than for placebo-treated patients at week 4 and week 12, increasing towards baseline values during off-treatment periods (figure 4). treatment decreased the proportion of patients who were culture-positive for species of Pseudomonas, Klebsiella, Escherichia, and Enterobacter, at weeks 4 and 16, compared with placebo (appendix). After baseline in both studies, the proportion of patients for whom Gram-negative organisms were isolated for susceptibility testing was lower in the group than in the placebo group (except for week 16 in AIR-BX2; data not shown); for example, at week 4 in AIR-BX1, susceptibility testing was conducted on organisms isolated from 36% of -treated patients (47 of 131) vs 66% of placebo-treated patients (85 of 128). For patients with detectable organisms in AIR-BX1, 15% (seven of 47) of -treated patients and 6% (five of 85) of placebo-treated patients had increases of four-fold or higher in the MIC of aztreonam for target Gram-negative organisms after 4 weeks; per centages were 35% (18 of 52; group) and 11% (nine of 79; placebo group) after 12 weeks, and were 23% (14 of 62; group) and 14% (11 of 76; placebo group) after 4 weeks off-treatment (week 16). For patients with detectable organisms in AIR-BX2, increases of fourfold or higher in the MIC of aztreonam for target Gramnegative organisms were noted after 4 weeks in 23% of -treated patients (12 of 53) and 7% of placebo-treated patients (six of 86); percentages were 34% (19 of 56) and 11% (eight of 76) after 12 weeks, and 20% (13 of 64) and 6% (four of 64) after 4 weeks off-treatment (week 16). Incidences of adverse events, treatment-related adverse events, serious adverse events, as well as adverse events leading to study drug discontinuation were higher for -treated patients than for placebo-treated patients, with a larger difference between groups generally noted in AIR-BX1 (table 4). The most commonly reported treatment-emergent adverse events were dyspnea, cough, and increased sputum. In AIR-BX1, each of these events was reported with higher incidence for -treated 744 www.thelancet.com/respiratory Vol 2 September 2014

than placebo-treated patients. In AIR-BX2, the incidences were more balanced. Dyspnea was the most common treatment-related event. A difference in the incidence of pneumonia occurred in AIR-BX1 (: six [4%] of 134 patients; placebo: one [1%] of 132 patients) but not in AIR-BX2 (: three [2%] of 135 patients; placebo: three [2%] of 137 patients). Bronchospasm ( 15% decline in FEV 1 (L) 30 min after first study treatment) was not noted for more than three patients in any group. In AIR-BX1, severe or life-threatening adverse events (grade 3 or 4) were reported for 27 (20%) of 134 treated patients and 14 (11%) of 132 placebo-treated patients (table 4), with respiratory events and infections reported most commonly. Two of these 27 patients reported three grade 4 adverse events (placebo: intracranial haemor rhage, cardiac arrest, and bronchopneumonia; all deemed unrelated to treatment). Three patients died (bronchiectasis progression []; severe sepsis syndrome []; cardiac arrest secondary to broncho pneumonia [placebo]). In AIR-BX2, adverse events of grade 3 severity were reported for 16 (12%) of 135 -treated and 16 (12%) of 137 placebo-treated patients, with respiratory events and infections reported most commonly. No Grade 4 events were reported. One patient died (respiratory failure and septic shock [placebo]). In AIR-BX1 and AIR-BX2, more patients in the group had adverse events leading to study drug discontinuation than in the placebo group (table 4). Dyspnea, pyrexia, and cough were the most common events leading to study drug discontinuation in AIR-BX1, cough and dyspnea were most common in AIR-BX2. For some patients, the reasons for discontinuing study medication differed from the reasons for discontinuing from the study (shown in figure 2). Reasons for discontinuing study medication for patients in the group of AIR-BX1 were: safety or tolerability (29 patients); withdrew consent (six); investigator s discretion (two); and other (one). Reasons for the group of AIR-BX2 were safety or tolerability (11); withdrew consent (four); and other (one), and reasons for the placebo group of AIR-BX2 were safety or tolerability (four); withdrew consent (six); investigator s discretion (one); lost to follow-up (one); and other (two). Reasons for discontinuing study medication were the same as for discontinuing the double-blind portion of the study for the placebo group of AIR-BX1. Discussion To our knowledge, AIR-BX1 and AIR-BX2 are the largest placebo-controlled inhaled antibiotic trials done in noncystic fibrosis bronchiectasis and the first to use change on the QOL-B-RSS as the primary endpoint (panel). did not improve respiratory symptoms or delay time to exacerbation in patients with non-cystic fibrosis bronchiectasis. Numerical improvements in QOL-B-RSS were noted for both -treated patients and Target Gram-negative organisms: change from baseline Log 10 CFU/g sputum (SE) Target Gram-negative organisms: change from baseline Log 10 CFU/g sputum (SE) A AIR-BX1 1 0 0 5 0 0 0 5 1 0 1 5 2 0 2 5 3 0 3 5 Baseline Week 4 B AIR-BX2 1 0 0 5 0 0 0 5 1 0 1 5 2 0 2 5 (n=102) (n=97) Treatment course 1 (n=104) (n=90) (n=97) (n=99) (n=91) (n=101) (n=76) (n=100) (n=95) 3 0 (n=101) (n=101) 3 5 Baseline Week 4 Week 8 Week 12 Week 16 placebo-treated patients; however, the difference between groups was only significant in AIR-BX2 after one course of, and this difference was less than the MID determined for this study (8 0 points; Quittner AL, personal communication) and thus not deemed clinically significant. Additionally, clinical benefit from treatment was not consistently noted in any patient subpopulation studied. The incidences of treatmentrelated adverse events were 1 4 times greater for (n=78) (n=94) (n=85) (n=94) (n=70) Week 8 Week 12 Week 16 Treatment course 2 Mean (SD) change from baseline Log 10 CFU/g sputum for target Gram-negative pathogens 2 6 (2 8) 2 5 (2 9) 0 4 (2 3) 1 8 (3 0) 0 2 (1 8) 0 0 (1 9) 0 0 (1 8) 0 1 (2 1) Treatment course 1 Treatment course 2 Mean (SD) change from baseline Log 10 CFU/g sputum for target Gram-negative pathogens 2 7 (2 8) 2 7 (3 2) 0 5 (2 7) 2 0 (2 9) 0 2 (1 9) 0 0 (2 2) 0 4 (2 2) 0 3 (2 5) 0 7 (2 6) 0 3 (1 7) (n=80) (n=90) 0 5 (2 7) 0 9 (2 4) Figure 4: Mean change from baseline Log 10 CFU per g of sputum for target Gram-negative organisms CFU per g of sputum was imputed as 0 for patients for whom no pathogens were cultured. N=number of patients with available data for change calculations. =aztreonam for inhalation solution. CFU=colony forming units. www.thelancet.com/respiratory Vol 2 September 2014 745

AIR-BX1 (n=134) (n=132) AIR-BX2 (n=135) (n=137) Any adverse event* 128 (96%) 117 (89%) 124 (92%) 118 (86%) Adverse events reported for 10% of patients in any group Dyspnea** 74 (55%) 47 (36%) 50 (37%) 47 (34%) Cough 70 (52%) 57 (43%) 66 (49%) 65 (47%) Sputum increased 63 (47%) 48 (36%) 59 (44%) 50 (36%) Fatigue*** 56 (42%) 29 (22%) 25 (19%) 25 (18%) Sputum discoloured 42 (31%) 33 (25%) 48 (36%) 38 (28%) Pyrexia* 32 (24%) 16 (12%) 30 (22%) 21 (15%) Headache 21 (16%) 23 (17%) 17 (13%) 14 (10%) Wheezing 21 (16%) 14 (11%) 18 (13%) 13 (9%) Non-cardiac chest pain 20 (15%) 14 (11%) 9 (7%) 13 (9%) Chills* 18 (13%) 6 (5%) 10 (7%) 3 (2%) Oropharyngeal pain 14 (10%) 17 (13%) 12 (9%) 14 (10%) Respiratory tract congestion* 14 (10%) 3 (2%) 3 (2%) 4 (3%) Malaise 11 (8%) 14 (11%) 14 (10%) 19 (14%) Grade 3 or 4 severity adverse event(s)* 27 (20%) 14 (11%) 16 (12%) 16 (12%) Treatment-related adverse event(s) 53 (40%) 37 (28%) 45 (33%) 25 (18%) Treatment-related adverse events reported for 10% of patients in any group Dyspnea** 17 (13%) 3 (2%) 9 (7%) 4 (3%) Cough 14 (10%) 13 (10%) 12 (9%) 11 (8%) Grade 3 or 4 severity treatment-related 9 (7%) 0 0 0 adverse event(s)** Serious adverse event(s)* 28 (21%) 13 (10%) 18 (13%) 16 (12%) Treatment-related serious adverse 12 (9%) 0 1 (1%) 0 event(s)*** Adverse event(s) leading to study drug 30 (22%) 8 (6%) 13 (10%) 7 (5%) discontinuation*** Serious adverse event(s) leading to study 12 (9%) 3 (2%) 3 (2%) 3 (2%) drug discontinuation* Hospitalization for protocol-defined 2 (1%) 3 (2%) 11 (8%) 7 (5%) exacerbation 15% decline in FEV 1 (L), 30 minutes after 1 (1%) 0 3 (2%) 2 (1%) first dose of study drug Grade 3 or 4 treatment-emergent laboratory abnormalities ll 1 (<1%) 3 (2%) 2 (2%) 5 (4%) Data are number of patients (%). =aztreonam for inhalation solution. FEV 1 =forced expiratory volume in 1 s. Adverse events were coded with Medical Dictionary for Regulatory Activities (MedDRA) V15.1 preferred terms and terms are presented by decreasing incidence in the AIR-BX1 group. We assessed differences in incidences of adverse events between and placebo by Fisher s exact test. * vs placebo p <0 05 for AIR-BX1 only; ** vs placebo p<0 01 for AIR-BX1 only, with the exception of treatment-related adverse events; only AIR-BX2 had p<0 01 for treatment-related adverse events. *** vs placebo p <0 001 for AIR-BX1 only. Only treatment-emergent adverse events are presented, which had onset dates on or after the start of treatment and up to 30 days after the last dose of study drug during course 2, or up to the beginning of week 16, whichever came first. Graded by the investigators as Grade 1 (mild), Grade 2 (moderate), Grade 3 (severe), or Grade 4 (life threatening) according to the Common Terminology Criteria for Adverse Events (CTCAE), V4.0. Treatment-related adverse events were considered by the investigator to be related to treatment with or placebo. For some patients, the reasons for discontinuing study medication differed from the reasons for discontinuing from the study (as shown in figure 2). lltreatment-emergent clinical laboratory abnormalities were defined as 1 toxicity grade worsening from baseline; for AIR-BX1, n=118 in the group and n=126 for the placebo group; in AIR-BX2, n=129 for the group, and n=130 for the placebo group. Table 4: Summary of treatment-emergent adverse events and other safety variables -treated patients than for placebo-treated patients in AIR-BX1 and 1 8 times greater in AIR-BX2, and the incidences of study discontinuations due to safety or tolerability were 6 7-times greater for -treated patients than placebo-treated patients in AIR-BX1 and 2 1 times greater in AIR-BX2. These differences between and placebo have not been explained. However, the most common adverse events were worsening respiratory symptoms in the first course of treatment, suggesting that local intolerance to inhaled medication contributed to the increased incidence of adverse events in treated patients. The decreases in sputum bacterial density noted in both studies are consistent with several previous studies; 5 15,23 however, such an effect has not corresponded consistently with improvement in clinical endpoints. The changes in preliminary (V2 0) QOL-B-RSS in an open-label study in non-cystic fibrosis bronchiectasis patients were larger than those described herein, but did not correlate with changes in sputum bacterial density. 20 In a small placebo-controlled study 5 of stable patients with non-cystic fibrosis bronchiectasis, inhaled tobramycin treatment significantly decreased sputum bacterial density but provided no evidence of clinical efficacy, and more adverse events were noted in the actively treated group than in the placebo group. Similarly, disparities were noted for tobramycin plus oral ciprofloxacin 9 and inhaled colistin, although significantly increased time to exacerbation was noted in exploratory analyses of treatment-adherent patients in the colistin study. 13 Inhaled ciprofloxacin treatment increased time to first exacerbation in a small study; 14 a phase 3 study is planned (ClinicalTrials.gov: NCT01515007). Oral macrolides (eg, azithromycin) have shown clinical efficacy in non-cystic fibrosis bronchiectasis; however, they also have complex mucus-modulating and anti-inflammatory properties so their antimicrobial activity might not be solely responsible for their clinical benefit. 23 27 Effects of chronic macrolide treatment were not noted on primary efficacy outcomes in the two current studies (21% [AIR-BX1] and 28% [AIR-BX2] of patients received them). We cannot exclude the presence of a subset of responders that was not identified by our analyses. However, the absence of significant improvement in QOL-B-RSS was consistent with the absence of clinical efficacy in changing time to next exacerbation. It is difficult to reconcile the disappointing clinical outcomes of inhaled antibiotics in non-cystic fibrosis bronchiectasis with the consistent benefit recorded for bronchiectasis associated with cystic fibrosis. Four factors might play a part in explaining this discrepancy. First, some unsuccessful aerosolised antibiotic trials in non-cystic fibrosis bronchiectasis used doses that had been optimised for cystic fibrosis, 5,9,13 assuming dose responses would be similar in both populations. Airway intolerance to inhaled antibiotics might be more problematic in patients with non-cystic fibrosis bronchiectasis and could potentially obscure clinical benefits; thus, future studies should consider exploring optimum dosing in non-cystic fibrosis bronchiectasis. Second, stratification of patients with cystic fibrosis by 746 www.thelancet.com/respiratory Vol 2 September 2014

Panel: Research in context Systematic review This trial was conceived after review of other studies of nebulised antibiotics in bronchiectasis (tobramycin, colistin). A phase 2 study was completed and published validating the Quality of Life-Bronchiectasis (QOL-B) as a measure of patient-reported outcomes in bronchiectasis. The current reported trial was patterned after successful phase 3 trials of aztreonam nebulisation in cystic fibrosis. We prepared the discussion section after scrutiny of published reviews of the management of bronchiectasis and ClinicalTrials.gov, and discussion among the authors and sponsor. Research in context One of the main pillars of the management of bronchiectasis is to reduce respiratory infections. Strategies include prompt attention to acute exacerbations and reducing the burden of potentially pathogenic bacteria such as Pseudomonas aeruginosa. This study used an aerosol antibiotic, aztreonam, in an attempt to improve symptoms and reduce exacerbations in patients. Almost all aerosol antibiotics reduce the microbial load (including tobramycin, gentamicin, amikacin, and colistin), yet no double-blind randomized trials have met their primary clinical endpoint. This report represents two large identical phase 3 studies of inhaled aztreonam in bronchiectasis. Compared with placebo, aztreonam did not significantly improve scores on a validated questionnaire, QOL-B, nor reduce time to first exacerbations. Aztreonam did significantly reduce the microbial load of Gram-negative organisms including Pseudomonas. More side-effects were noted in aztreonam-treated patients than in placebo-treated patients. Although tobramycin and aztreonam by nebulisation are approved for use in cystic fibrosis, no nebulised antibiotic has been approved for bronchiectasis by regulatory agencies in any country. baseline FEV 1 enables selection of patients more likely to have exacerbations and facilitates an even distribution of severity between study groups. By contrast, baseline spirometry alone has limited predictive value in noncystic fibrosis bronchiectasis, although FEV 1 lower than 30% predicted is one of six predictors in a recently developed algorithm to identify patients at higher risk of exacerbation, hospital admission, and mortality. 28 Third, clinical benefits of inhaled antibiotics might be synergistic with or require airway clearance techniques. Airway clearance is different in non-cystic fibrosis bronchiectasis compared with cystic fibrosis. The disease is mainly lower lobe, an area potentially more difficult to clear mucus and organisms from after killing. Further, old age and comorbidities might produce less efficient cough and clearing, and patients with non-cystic fibrosis bronchiectasis use supportive airway clearance techniques less often than patients with cystic fibrosis. These differences might explain some of the disconnect between decreased bacterial density and the absence of clinical efficacy. Fourth, in elderly patients with noncystic fibrosis bronchiectasis, there is a spectrum of clinical overlap with other lung diseases, most notably COPD, with no standardised method to objectively assess contributions of each disease to an individual s symptomatology or response to treatment. Further, COPD was reported with higher incidence in the AIR-BX1 group, compared with placebo, and the high incidence of use of long acting β-2 agonists and muscarinic antagonists and inhaled corticosteroids might have had a confounding effect on respiratory symptoms or exacerbations. In conclusion, treatment was not associated with significant clinical benefit in non-cystic fibrosis bronchiectasis, as measured by changes in QOL-B-RSS and time to first exacerbation, and the groups had a higher incidence of treatment-related adverse events and discontinuations due to adverse events, compared with placebo. The discordance between microbiological and clinical endpoints, recorded in this and other placebocontrolled trials, shows a continued need for placebocontrolled and dose-ranging studies to establish the clinical benefit of inhaled antibiotics in stable patients with non-cystic fibrosis bronchiectasis and shows the challenges inherent in the definition of which patients with non-cystic fibrosis bronchiectasis might benefit from such treatments. Contributors AFB was the lead investigator for AR-BX1 and contributed to the design of the study, and reviewed the report. AEOD was the lead investigator for AIR-BX2 and contributed to the design of the study, and reviewed the report. PF was an investigator in the trials, contributed to the interpretation of the data, and editing of the report. PJT was the investigator in the trials, contributed to the design of the study and reviewed the report. JDR was an investigator in the trials and reviewed the report. JdG contributed to study design, patient recruitment, review of data analysis and results, and review of drafts and final report. WGB contributed to data collection, data interpretation, and writing of the paper. ADS was an investigator in the trials and reviewed the report. LS and JZ performed the statistical analysis, contributed to study design, and reviewed the report. LH contributed to the design of the study and reviewed the report. SAL performed the statistical analysis, contributed to study design, and reviewed the report. SL, ABM, and MTM contributed to the design of the study and reviewed the report. DG served as Medical Monitor, contributed to the design of the study analyses, and reviewed the report. ALQ designed the QOL-B, contributed to the design of the study and reviewed the report. TGOR served as Medical Monitor, contributed to the conduct and design of the study, reviewed the report, and assisted with collation of comments from the other authors. Study investigators AIR-BX1: Daniel Chambers, Chermside, QLD, Australia; Michael (Mon Ming) Chia, Toorak Gardens, SA, Australia; Huw Davies, Daw Park, SA, Australia; Chien-Li Homes-Liew, Adelaide, SA, Australia; Paul King, Clayton, VIC, Australia; Steven Lindstrom, Kogarah, NSW, Australia; Peter Middleton, Westmead, NSW, Australia; Michael Musk, Perth, NSW, Australia; Matthew Peters, Concord, NSW, Australia; David Serisier, South Brisbane, QLD, Australia; Philip Thompson, Nedlands, WA, Australia; Celine Bergero, Montreal, QC, Canada; Nasreen Khalil, Vancouver, BC, Canada; Irvin Mayers, Edmonton, AB, Canada; Christopher Mody, Calgary, AB, Canada; Douglass Rolf, Kelowna, BC, Canada; Elizabeth Tullis, Toronto, ON, Canada; www.thelancet.com/respiratory Vol 2 September 2014 747

Pearce Wilcox, Vancouver, BC, Canada; Timothy Aksamit, Rochester, MN, USA; Devendra Amin, Clearwater, FL, USA; Alan Barker, Portland, OR, USA; Anne Brown, Falls Church, VA, USA; Jeff Cary, Seattle, WA, USA; John Cohn, Philadelphia, PA, USA; Charles Daley, Denver, CO, USA; Kevin De Boer, Orlando, FL, USA; Edward Eden, New York, NY, USA; Neil Ettinger, Chesterfield, MO, USA; Jeremy P.Feldman, Phoenix, AZ, USA: Patrick Flume, Charleston, SC, USA; Peter Fornos, San Antonio, TX, USA; Mark Gotfried, Phoenix, AZ, USA; Anas Hadeh, Weston, FL, USA; Jonathan Ilowite, Mineola, NY, USA; Patricia Joseph, Cincinnatti, OH, USA; Shahrukh Kureishy, McKinney, TX, USA; Chrçandon, Ventura, CA, USA; Mark Metersky, Farmington, CT, USA; Brian Morrissey, Sacramento, CA, USA; Jonathan Ruzi, Scottsdale, AZ, USA; Matthias Salathe, Miami, FL, USA; Varsha Taskar, Tyler, TX, USA; James Taylor, Tacoma, WA, USA; Craig Thurm, Jamaica, NY, USA; Gregory Tino, Philadelphia, PA, USA; Jose Urdaneta, Kissimmee, FL, USA; Rajat Walia, Phoenix, AZ, USA. AIR-BX2: Chien-Li Holmes-Liew, Adelaide, SA, Australia; Philip Thompson, Nedlands, WA, Australia; Lieven Dupont, Leuven, Belgium; Christiane Knoop, Bruxelles, Belgium; Pascal Chanez, Marseille, France; Raphael Chiron, Montpellier, France; Charles Hugo Marquette, Nice, France; Marlène Murris, Toulouse, France; Christian Gessner, Leipzig, Germany; Rudolf Huber, Munchen, Germany; Rainald Fischer, Munchen, Germany; Harmut Lode, Berlin-Charlottenburg, Germany; Mathias Pletz, Jena, Germany; Carsten Schwarz, Berlin, Germany; Stephan Sorichter, Freiburg, Germany; Tobias Welte, Hannover, Germany; Heinrike Wilkens, Homburg, Germany; Hubert Wirtz, Leipzig, Germany; Francesco Blasi, Milan, Italy; Leonardo Fabbri, Moderna, Italy; Pierluigi Paggiaro, Pisa, Italy; Alberto Papi, Ferrara, Italy; Francesca Santamaria, Naples, Italy; Wim G. Boersma, Alkmaar, Netherlands; Menno van der Eerden, Rotterdam, Netherlands; Javier de Gracia, Barcelona, Spain; Jordi Dorca, Barcelona, Spain; Eva Polverino, Barcelona, Spain; Concepcion Prados, Madrid, Spain; Jover Sole, Valencia, Spain; Montserrat Vendrell, Girona, Spain; Paul Dawkins, Woverhampton, UK; Anthony De Soyza, Newcastle upon Tyne, UK; Charles Haworth; Cambridgeshire, UK; Omar Pirzada, Sheffield, UK; Nicholas Withers, Exeter, Devon, UK; Nasreen Khalil, Vancouver, BC, Canada; Douglass Rolf, Kelowna, BC, Canada; Mohamed I. Ali, St. Petersburg, FL, USA; Jorge Alvarez, Council Bluffs, IA; USA; Abubakr Bajwa, Jacksonville, FL, USA; Julie Biller, Milwaukee, WI, USA; Holly Carveth, Salt Lake City, UT, USA; Thomas DeMarini, Decatur, GA, USA; Angela DiMango, New York, NY, USA; Faisal Fakih, Winter Park, FL, USA; Mark Gotfried, Phoenix, AZ, USA; Tarik Haddad, Tampa, FL, USA; Todd Ken Horiuchi, Sarasota, FL, USA; Sabiha Hussain, New Brunswick, NJ, USA; Richard Kahlstrom, Tacoma, WA, USA; William Krimsky, Rosedale, MD, USA; Douglas Mapel, Flagstaff, AZ, USA; John McConnell, Louisville, KY, USA; Keith Meyer, Madison, WI, USA; Melvin Morganroth, Portland, OR, USA; Felix Morris, Florence, AL, USA; Peadar Noone, Chapel Hill, NC, USA; Anne O Donnell, Washington DC, USA; Robert Orr, Peoria, AZ, USA; Jonathan Ruzi, Scottsdale, AZ, USA; Jack Stewart, Orange, CA, USA; Mary Strek, Chicago, IL, USA; Robert Sussman, Summit, NJ, USA; Craig Thurm, Jamaica, NY, USA; Gordon Yung, La Jolla, CA, USA. Declaration of interests AFB reports grants from Gilead, during the conduct of the study. AEOD reports grants from Gilead, during the conduct of the study; and grants and personal fees from Bayer, personal fees from Novartis, grants from COPD Foundation, grants from University of Miami, grants and personal fees from Insmed, and grants from Aradigm, outside the submitted work. PF reports grants from Gilead, during the conduct of the study; and grants and personal fees from Gilead, outside the submitted work. JdG reports grants and non-financial support from Gilead Sciences, during the conduct of the study; and reports personal fees from Gilead Sciences, grants from Novartis, personal fees from Novartis, grants from Vertex, and personal fees from Vertex, outside the submitted work. ADS reports grants and personal fees from Bayer, grants from Forest labs, grants from Gilead, non-financial support from Novartis, personal fees and non-financial support from Almirall, GSK, and Boehringer Ingelheim, outside the submitted work. LS, JZ, LH, SAL, SL, MTM, and TGOR are employees of and own stock in Gilead Sciences. ABM was a previous employee of Gilead Sciences. DG is an employee of and owns stock in Gilead Sciences and a former employee of MedImmune/AstraZeneca. ALQ reports funding from Gilead Sciences, during the conduct of the study; and grants from Gilead Sciences and grants from Insmed Inc, outside the submitted work; additionally, ALQ has a patent copyright to QOL-B Version 3.1 issued. JDR, PJT, and WGB declare no competing interests. Acknowledgments We thank the patients who participated in these studies and their families. We also thank the research coordinators at each clinical site. Medical writing assistance was provided by Kate Loughney, under the sponsorship of Gilead Sciences. The external data monitoring committee included Michael Knowles, MD (University of North Carolina, Chapel Hill, NC, USA), Tom Schaberg, MD (Diakoniekrankenhaus Rotenburg (Wümme) ggmbh, Rotenburg, Gemany), and Charles Davis, PhD (CSD Biostatistics, San Diego, CA, USA). References 1 Barker AF. Bronchiectasis. N Engl J Med 2002; 346: 1383 93. 2 O Donnell AE. Bronchiectasis. Chest 2008; 134: 815 23. 3 King PT, Holdsworth SR, Freezer NJ, Villanueva E, Holmes PW. Microbiologic follow-up study in adult bronchiectasis. Respir Med 2007; 101: 1633 38. 4 Döring G, Flume P, Heijerman H, Elborn JS; Consensus Study Group. Treatment of lung infection in patients with cystic fibrosis: current and future strategies. J Cyst Fibros 2012; 11: 461 79. 5 Barker AF, Couch L, Fiel SB, et al. Tobramycin solution for inhalation reduces sputum Pseudomonas aeruginosa density in bronchiectasis. Am J Respir Crit Care Med 2000; 162: 481 85. 6 Evans DJ, Bara AI, Greenstone M. Prolonged antibiotics for purulent bronchiectasis in children and adults. Cochrane Database Syst Rev 2007; 2: CD001392. 7 Drobnic ME, Suñé P, Montoro JB, Ferrer A, Orriols R. Inhaled tobramycin in non-cystic fibrosis patients with bronchiectasis and chronic bronchial infection with Pseudomonas aeruginosa. Ann Pharmacother 2005; 39: 39 44. 8 Scheinberg P, Shore E. A pilot study of the safety and efficacy of tobramycin solution for inhalation in patients with severe bronchiectasis. Chest 2005; 127: 1420 26. 9 Bilton D, Henig N, Morrissey B, Gotfried M. Addition of inhaled tobramycin to ciprofloxacin for acute exacerbations of Pseudomonas aeruginosa infection in adult bronchiectasis. Chest 2006; 130: 1503 10. 10 Steinfort DP, Steinfort C. Effect of long-term nebulized colistin on lung function and quality of life in patients with chronic bronchial sepsis. Intern Med J 2007; 37: 495 98. 11 Murray MP, Govan JR, Doherty CJ, et al. A randomized controlled trial of nebulized gentamicin in non-cystic fibrosis bronchiectasis. Am J Respir Crit Care Med 2011; 183: 491 99. 12 Serisier DJ, Bowler SD. Randomized controlled trial of nebulized gentamicin in non-cystic fibrosis bronchiectasis without patient blinding. Am J Respir Crit Care Med 2012; 186: 461. 13 Haworth C, Foweraker J, Wilkinson P, Kenyon R, Bilton D. Inhaled colistin in patients with bronchiectasis and chronic Pseudomonas aeruginosa infection. Am J Respir Crit Care Med 2014; 189: 975 82. 14 Serisier DJ, Bilton D, De Soyza A, et al. Inhaled, dual release liposomal ciprofloxacin in non-cystic fibrosis bronchiectasis (ORBIT-2): a randomised, double-blind, placebo-controlled trial. Thorax 2013; 68: 812 17. 15 Wilson R, Welte T, Polverino E, et al. Ciprofloxacin dry powder for inhalation in non-cystic fibrosis bronchiectasis: a phase II randomised study. Eur Respir J 2013; 41: 1107 15. 16 Smith MP, Hill AT. Evaluating success of therapy for bronchiectasis: what end points to use? Clin Chest Med 2012; 33: 329 49. 17 McCoy KS, Quittner AL, Oermann CM, Gibson RL, Retsch-Bogart GZ, Montgomery AB. Inhaled aztreonam lysine for chronic airway Pseudomonas aeruginosa in cystic fibrosis. Am J Respir Crit Care Med 2008; 178: 921 28. 18 Retsch-Bogart GZ, Quittner AL, Gibson RL, et al. Efficacy and safety of inhaled aztreonam lysine for airway Pseudomonas in cystic fibrosis. Chest 2009; 135: 1223 32. 19 Oermann CM, Retsch-Bogart GZ, Quittner AL, et al. An 18-month study of the safety and efficacy of repeated courses of inhaled aztreonam lysine in cystic fibrosis. Pediatr Pulmonol 2010; 45: 1121 34. 748 www.thelancet.com/respiratory Vol 2 September 2014