Advertisement

A network meta-analysis of the efficacy of inhaled antibiotics for chronic Pseudomonas infections in cystic fibrosis

Open AccessPublished:June 21, 2012DOI:https://doi.org/10.1016/j.jcf.2012.03.010

      Abstract

      Background

      Various inhaled antibiotics are currently used for treating chronic Pseudomonas aeruginosa lung infection in cystic fibrosis (CF) patients, however their relative efficacies are unclear. We compared the efficacy of the inhaled antibiotics tobramycin (TIP, TIS-T, TIS-B), colistimethate sodium (colistin) and aztreonam lysine for inhalation (AZLI) based on data from randomised controlled trials.

      Methods

      In the base case, efficacies of antibiotics were compared using a network meta-analysis of seven trials including change from baseline in forced expiratory volume in 1 second (FEV1) % predicted, P. aeruginosa sputum density and acute exacerbations.

      Results

      The tobramycin preparations, AZLI and colistin, showed comparable improvements in efficacy in terms of FEV1% predicted at 4 weeks; the difference in % change from baseline (95%CrI) for TIP was compared to TIS-T (-0.55, -3.5;2.4), TIS-B (-0.64, -7.1;5.7), AZLI (3.64, -1.0;8.3) and colistin (5.77, -1.2;12.8).

      Conclusion

      We conclude that all studied antibiotics have comparable efficacies for the treatment of chronic P. aeruginosa lung infection in CF.

      Keywords

      1. Introduction

      Cystic fibrosis (CF) is the most common fatal inherited disease among Caucasians, affecting approximately 80,000 individuals in Europe and North America [
      • Döring G.
      • Ratjen F.
      Immunology of cystic fibrosis.
      ]. CF is characterised by chronic lung infections with opportunistic pathogens [
      • Hoiby N.
      • Frederiksen B.
      Microbiology of cystic fibrosis.
      ] due to the consequences of mutations in the CF transmembrane conductance regulator (CFTR) gene [
      • Tsui L.-C.
      The spectrum of cystic fibrosis mutations.
      ] which compromise innate immune functions [
      • Döring G.
      • Gulbins E.
      Cystic fibrosis and innate immunity: how chloride channel mutations provoke lung disease.
      ]. Respiratory infections with P. aeruginosa are recognised as having the largest impact on morbidity and mortality. Improved antibiotic strategies against respiratory tract infections are considered the main reason for the increased life expectancy in CF patients in the last decades. Particularly, early eradication antibiotic therapy has successfully eradicated P. aeruginosa from CF airways for considerable time periods and delayed onset of chronic infection in many European CF centres in the last two decades [
      • Döring G.
      • Hoiby N.
      for the Consensus Study Group
      Early intervention and prevention of lung disease in cystic fibrosis: a European consensus.
      ]. However, once established, chronic P. aeruginosa infections are difficult to treat with antibiotics and the pathogen is virtually never eradicated due to biofilm formation [
      • Döring G.
      • Hoiby N.
      for the Consensus Study Group
      Early intervention and prevention of lung disease in cystic fibrosis: a European consensus.
      ]. Therefore, clinical trials of antibiotics for chronic P. aeruginosa lung infection in individuals with CF use endpoints such as relative increase in lung function and reduction of P. aeruginosa sputum density or acute exacerbations rather than pathogen eradication. However, while there are a number of antibiotics available which have been used in the past to treat individuals with CF with chronic P. aeruginosa infections, there is a paucity of data comparing the efficacy and safety of these treatments to each other which would guide the CF clinician.
      We compared the efficacy and safety of inhaled antibiotics tobramycin, colistimethate sodium and aztreonam lysine, using a network meta-analysis (NMA). In the absence of head to head trials against all comparators, NMA methods allow assessment of relative efficacy in a structured and objective way [
      • Lu G.
      • Ades A.E.
      Combination of direct and indirect evidence in mixed treatment comparisons.
      ,
      • Jansen J.P.
      • Fleurence R.
      • Devine B.
      • Itzler R.
      • Barrett A.
      • Hawkins N.
      • et al.
      Interpreting indirect treatment comparisons and network meta-analysis for health-care decision making: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: part 1.
      ]. We assessed the endpoints; % change from baseline (CFB) in forced expiratory volume in 1 second (FEV1) % predicted, CFB of P. aeruginosa sputum density and exacerbations. Previously, a Cochrane review [
      • Ryan G.
      • Singh M.
      • Dwan K.
      Inhaled antibiotics for long-term therapy in cystic fibrosis.
      ] was conducted on inhaled antibiotics for CF. However; a meta-analysis was, at that time, not conducted judging study designs and reporting of results too variable.

      2. Methods

      2.1 Antibiotic formulations

      The following approved antibiotic formulations for inhalation were compared in this study: (i) tobramycin inhalation powder capsule (TIP), containing 112 mg tobramycin BID (TOBI Podhaler®), (ii) tobramycin inhalation solution (TIS-T), containing 300 mg/5 ml tobramycin BID (TOBI®), (iii) tobramycin inhalation solution (TIS-B), containing 300 mg/4 ml tobramycin BID (Bramitob®), (iv) colistimethate sodium (colistin), containing 80 mg/3 ml colistimethate sodium BID (Colomycin®, Promixin®), and (v) aztreonam lysine inhalation solution (AZLI), containing 75 mg aztreonam lysine TID (Cayston®).

      2.2 Identification and selection of studies

      A systematic literature review was conducted for English language publications in Medline, Medline in Process, Embase and the Cochrane Library. Search terms included free-text and thesaurus terms for cystic fibrosis, Pseudomonas, the interventions of interest and randomised controlled study design. The search was conducted on October 2010 and two publications of the two TIP trials were added in January 2011. Conference data from 2009 and 2010 were screened in BIOSIS and from the European CF Conference, North American CF Conference, European Respiratory Society and American Thoracic Society. Two reviewers independently screened first the abstracts, then selected full papers against the following predefined inclusion criteria: (1) population: CF patients aged 6 years or older with chronic P. aeruginosa infection; (2) interventions: inhaled tobramycin, colistin, AZLI, amikacin or ciprofloxacin; (3) comparisons: interventions above (at licensed dose) to each other or placebo; (4) outcomes: percent CFB in FEV1% predicted at 4 weeks and 20 weeks, CFB in P. aeruginosa sputum density (measured by Log10 CFUs) at 4 weeks and 20 weeks, percent of patients with use of additional anti-P. aeruginosa antibiotics (intravenous, oral or inhaled) and percent of patients with at least one hospitalisation for acute respiratory events within 24 weeks; and (5) study design: randomised controlled trials in full publications or from conferences (if sufficient data were presented).

      2.3 Data extraction and validity assessment

      For included studies, data were extracted on study design, population, interventions, and outcomes described above. In cases where CFBs of FEV1% predicted were not reported, data were calculated from reported data, or else extracted from figures using Digitizit software version 1.5.8 (Share-it!, Cologne, Germany). The standard error of the difference in percent CFB was extracted where reported, or calculated (e.g. based on 95% confidence interval or standard deviation). If there were insufficient data for the calculations, an average standard deviation was calculated from included studies in each analysis and combined with study-specific sample sizes to estimate the study-specific standard error.

      2.4 Network meta-analysis (NMA)

      Bayesian NMA models were applied with non-informative priors [
      • Lu G.
      • Ades A.E.
      Combination of direct and indirect evidence in mixed treatment comparisons.
      ,
      • Jansen J.P.
      • Fleurence R.
      • Devine B.
      • Itzler R.
      • Barrett A.
      • Hawkins N.
      • et al.
      Interpreting indirect treatment comparisons and network meta-analysis for health-care decision making: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: part 1.
      ,
      • Caldwell D.M.
      • Ades A.E.
      • Higgins J.P.
      Simultaneous comparison of multiple treatments: combining direct and indirect evidence.
      ] to simultaneously synthesize the results of included studies and compare the efficacy and safety of antibiotics. The outcome was the relative treatment effect of each intervention versus placebo, and of TIP versus other comparators with a probability of being the better treatment. Both fixed and random effects models were evaluated. Fixed effects models assume one true treatment effect, while random effects models allow for different true treatment effects across studies, thus take account of heterogeneity in relative treatment effects. The choice between a fixed or random effects model for reported outcomes was based on model fit criteria (Deviance Information Criteria (DIC)) which penalise greater model complexity [
      • Caldwell D.M.
      • Ades A.E.
      • Higgins J.P.
      Simultaneous comparison of multiple treatments: combining direct and indirect evidence.
      ].
      The NMA can synthesize a network of randomised controlled trials (RCTs) and allow inferences about comparisons that have not been studied in a head-to-head fashion. Even when direct evidence is available for some treatments, combining this with indirect comparisons in a NMA may yield a more refined and precise estimate for the relative treatment effect [
      • Lu G.
      • Ades A.E.
      Combination of direct and indirect evidence in mixed treatment comparisons.
      ,
      • Jansen J.P.
      • Fleurence R.
      • Devine B.
      • Itzler R.
      • Barrett A.
      • Hawkins N.
      • et al.
      Interpreting indirect treatment comparisons and network meta-analysis for health-care decision making: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: part 1.
      ]. Data on the following outcomes were analysed from RCTs that were connected via common comparators; changes in (i) FEV1% predicted, (ii) P. aeruginosa sputum density and (iii) exacerbations (i.e. assessed by respiratory hospitalisations or need for additional anti-P. aeruginosa antibiotic use). Separate analyses for each outcome included models for all available study data, various subgroups of studies and/or inclusion of one of the following study-level covariates [
      • Cooper N.J.
      • Sutton A.J.
      • Morris D.
      • Ades A.E.
      • Welton N.J.
      Addressing between-study heterogeneity and inconsistency in mixed treatment comparisons: application to stroke prevention treatments in individuals with non-rheumatic atrial fibrillation.
      ] for possible effect modifiers that could bias results: mean baseline age or mean baseline FEV1% predicted and mean percent of patients with prior exposure to study drug. WinBUGS 1.4.1 statistical software was used for the analyses [
      • Lunn D.J.
      • Thomas A.
      • Best N.
      • Spiegelhalter D.
      WinBUGS — a Bayesian modelling framework: concepts, structure, and extensibility.
      ].

      3. Results

      3.1 Study identification

      The literature search identified 413 abstracts. Of these, 39 full papers were included for full-text screening from which 11 potentially relevant trials were assessed for the NMA (Fig. 1). Eight were placebo-controlled trials and three compared two active drugs (Table 1, Table 2). The base case analysis included a network of 7 trials (Fig. 2).
      Figure thumbnail gr1
      Fig. 1Study selection flow diagram. *Oermann data was only available in a poster.
      Table 1Study characteristics of 11 identified inhaled antibiotic studies in patients with CF.
      SourceType of RCTDuration (weeks)Treatments
      • Konstan M.W.
      • Geller D.E.
      • Minić P.
      • Brockhaus F.
      • Zhang J.
      • Angyalosi G.
      Tobramycin inhalation powder for P. aeruginosa infection in cystic fibrosis: the EVOLVE trial.
      *DB, PC, MC24: 3 cycles**TIP 122 mg BID vs. PLA
      • Konstan M.W.
      • Flume P.A.
      • Kappler M.
      • Chiron R.
      • Higgins M.
      • Brockhaus F.
      • et al.
      Safety, efficacy and convenience of tobramycin inhalation powder in cystic fibrosis patients: the EAGER trial.
      OL, MC24TIP 112 mg BID vs. TIS-T 300 mg BID
      • Ramsey B.W.
      • Pepe M.S.
      • Quan J.M.
      • Otto K.L.
      • Montgomery A.B.
      • Williams-Warren J.
      • et al.
      Intermittent administration of inhaled tobramycin in patients with cystic fibrosis. Cystic fibrosis inhaled tobramycin study group.
      DB, PC, MC24: 3 cyclesTIS-T 300 mg/5 mL BID vs PLA
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      DB, PC (pilot)24: 3 cyclesTIS-T 300 mg/5 mL BID vs PLA
      • Hodson M.E.
      • Gallagher C.G.
      • Govan J.R.
      A randomised clinical trial of nebulised tobramycin or colistin in cystic fibrosis.
      OL, MC4TIS-T 300 mg/5 mL BID vs. Colistin 80 mg/3 mL BID
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      DB, PC13Colistin 1 MU/3 ml BID vs. PLA
      • Lenoir G.
      • Antypkin Y.G.
      • Miano A.
      • Moretti P.
      • Zanda M.
      • Varoli G.
      • et al.
      Efficacy, safety, and local pharmacokinetics of highly concentrated nebulized tobramycin in patients with cystic fibrosis colonized with Pseudomonas aeruginosa.
      DB, PC, MC, PG8: 1 cycleTIS-B 300 mg/4 ml BID vs. PLA
      • Chuchalin A.
      • Csiszér E.
      • Gyurkovics K.
      • Bartnicka M.T.
      • Sands D.
      • Kapranov N.
      • et al.
      A formulation of aerosolized tobramycin (TIS-B) in the treatment of patients with cystic fibrosis and Pseudomonas aeruginosa infection: a double-blind, placebo-controlled, multicenter study.
      DB, PC, MC, PG24: 3 cyclesTIS-B 300 mg/4 ml BID vs PLA
      • McCoy K.S.
      • Quittner A.L.
      • Oermann C.M.
      • Gibson R.L.
      • Retsch-Bogart G.Z.
      • Montgomery A.B.
      Inhaled aztreonam lysine for chronic airway Pseudomonas aeruginosa in cystic fibrosis.
      DB, PC, MC4+8 follow upAZLI 75 mg BID or TID vs PLA BID or TID
      • Retsch-Bogart G.Z.
      • Quittner A.L.
      • Gibson R.L.
      • Oermann C.M.
      • McCoy K.S.
      • Montgomery A.B.
      • et al.
      Efficacy and safety of inhaled aztreonam lysine for airway Pseudomonas in cystic fibrosis.
      DB, PC, MC4AZLI 75 mg TID vs PLA TID
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      OL, PC, MC24: 3 cyclesAZLI 75 mg TID vs. TIS-T 300 mg BID
      *Double blind (DB), placebo-controlled (PC), multi-centre (MC), open-label (OL), parallel-group (PG), placebo (PLA), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300 mg/5 ml (TIS-T), tobramycin inhalation solution 300 mg/4 ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), twice a day (BID), three times a day (TID).**Cycle=4 weeks on/4 weeks off.
      Table 2Baseline characteristics of participants in the identified trials.
      SourceTreatmentRandomised populationGender (% male)Mean age (y)Mean FEV1% predictedExposure
      • Konstan M.W.
      • Geller D.E.
      • Minić P.
      • Brockhaus F.
      • Zhang J.
      • Angyalosi G.
      Tobramycin inhalation powder for P. aeruginosa infection in cystic fibrosis: the EVOLVE trial.
      Studies 2301 and 2302, which corresponded to [23,22] respectively.
      TIP46
      Randomised for safety. Patients received at least one capsule of study drug,
      41.3%13.454.7Naïve
      Placebo49
      Randomised for safety. Patients received at least one capsule of study drug,
      46.9%13.258.5Naïve
      • Konstan M.W.
      • Flume P.A.
      • Kappler M.
      • Chiron R.
      • Higgins M.
      • Brockhaus F.
      • et al.
      Safety, efficacy and convenience of tobramycin inhalation powder in cystic fibrosis patients: the EAGER trial.
      Studies 2301 and 2302, which corresponded to [23,22] respectively.
      TIP30855.5%25.952.9Exposed
      TIS-T20955.0%25.252.8Exposed
      • Ramsey B.W.
      • Pepe M.S.
      • Quan J.M.
      • Otto K.L.
      • Montgomery A.B.
      • Williams-Warren J.
      • et al.
      Intermittent administration of inhaled tobramycin in patients with cystic fibrosis. Cystic fibrosis inhaled tobramycin study group.
      TIS-T25858%20.849.9Naïve
      Placebo26250%20.651.2Naïve
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      TIS-T1637.5%11.895.7Unknown
      Placebo1637.5%15.983.7Unknown
      • Hodson M.E.
      • Gallagher C.G.
      • Govan J.R.
      A randomised clinical trial of nebulised tobramycin or colistin in cystic fibrosis.
      TIS-T5337.7%21.355.4Exposed
      Colistin6251.6%20.159.4Exposed
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      Colistin2050%13.671Unknown
      Placebo2055%14.779Unknown
      • Lenoir G.
      • Antypkin Y.G.
      • Miano A.
      • Moretti P.
      • Zanda M.
      • Varoli G.
      • et al.
      Efficacy, safety, and local pharmacokinetics of highly concentrated nebulized tobramycin in patients with cystic fibrosis colonized with Pseudomonas aeruginosa.
      TIS-B2951.7%11.057.7Naïve
      Placebo3056.7%14.259.8Naïve
      • Chuchalin A.
      • Csiszér E.
      • Gyurkovics K.
      • Bartnicka M.T.
      • Sands D.
      • Kapranov N.
      • et al.
      A formulation of aerosolized tobramycin (TIS-B) in the treatment of patients with cystic fibrosis and Pseudomonas aeruginosa infection: a double-blind, placebo-controlled, multicenter study.
      TIS-B161
      Intention to treat population. Cystic fibrosis (CF), P. aeruginosa (Pa), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300mg/5ml (TIS-T), tobramycin inhalation solution 300mg/4ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), forced expiratory volume in 1second percentage (FEV1%).
      55.3%14.860.7Naïve
      Placebo84
      Intention to treat population. Cystic fibrosis (CF), P. aeruginosa (Pa), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300mg/5ml (TIS-T), tobramycin inhalation solution 300mg/4ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), forced expiratory volume in 1second percentage (FEV1%).
      54.8%14.763.6Naïve
      • McCoy K.S.
      • Quittner A.L.
      • Oermann C.M.
      • Gibson R.L.
      • Retsch-Bogart G.Z.
      • Montgomery A.B.
      Inhaled aztreonam lysine for chronic airway Pseudomonas aeruginosa in cystic fibrosis.
      AZLI66
      Intention to treat population. Cystic fibrosis (CF), P. aeruginosa (Pa), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300mg/5ml (TIS-T), tobramycin inhalation solution 300mg/4ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), forced expiratory volume in 1second percentage (FEV1%).
      57.6%24.155.4Naïve
      Placebo76
      Intention to treat population. Cystic fibrosis (CF), P. aeruginosa (Pa), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300mg/5ml (TIS-T), tobramycin inhalation solution 300mg/4ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), forced expiratory volume in 1second percentage (FEV1%).
      59.2%27.953.9Naïve
      • Retsch-Bogart G.Z.
      • Quittner A.L.
      • Gibson R.L.
      • Oermann C.M.
      • McCoy K.S.
      • Montgomery A.B.
      • et al.
      Efficacy and safety of inhaled aztreonam lysine for airway Pseudomonas in cystic fibrosis.
      AZLI8060%27.454.4Naïve
      Placebo8453.6%31.754.8Naïve
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      AZLI136
      Intention to treat population. Cystic fibrosis (CF), P. aeruginosa (Pa), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300mg/5ml (TIS-T), tobramycin inhalation solution 300mg/4ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), forced expiratory volume in 1second percentage (FEV1%).
      50%25.852.3Naïve
      TIS-T132
      Intention to treat population. Cystic fibrosis (CF), P. aeruginosa (Pa), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300mg/5ml (TIS-T), tobramycin inhalation solution 300mg/4ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), forced expiratory volume in 1second percentage (FEV1%).
      50%25.152.2Exposed
      Study
      • Lenoir G.
      • Antypkin Y.G.
      • Miano A.
      • Moretti P.
      • Zanda M.
      • Varoli G.
      • et al.
      Efficacy, safety, and local pharmacokinetics of highly concentrated nebulized tobramycin in patients with cystic fibrosis colonized with Pseudomonas aeruginosa.
      included patients with a first or intermittent P. aeruginosa infection
      a Studies 2301 and 2302, which corresponded to
      • Konstan M.W.
      • Geller D.E.
      • Minić P.
      • Brockhaus F.
      • Zhang J.
      • Angyalosi G.
      Tobramycin inhalation powder for P. aeruginosa infection in cystic fibrosis: the EVOLVE trial.
      ,
      • Konstan M.W.
      • Flume P.A.
      • Kappler M.
      • Chiron R.
      • Higgins M.
      • Brockhaus F.
      • et al.
      Safety, efficacy and convenience of tobramycin inhalation powder in cystic fibrosis patients: the EAGER trial.
      respectively.
      b Randomised for safety. Patients received at least one capsule of study drug,
      c Intention to treat population. Cystic fibrosis (CF), P. aeruginosa (Pa), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300 mg/5 ml (TIS-T), tobramycin inhalation solution 300 mg/4 ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), forced expiratory volume in 1 second percentage (FEV1%).
      Figure thumbnail gr2
      Fig. 2The base case network of evidence. The network of studies that was included in the base case analysis is shown. Lines between boxes indicate which treatments were compared directly in a randomised controlled trial (RCT). Numbers in brackets reveal references. Tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300 mg/5 ml (TIS-T), tobramycin inhalation solution 300 mg/4 ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI).

      3.2 Study and patient characteristics

      All studies were published after 2006, except for studies by Jensen [
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      ], Ramsey [
      • Ramsey B.W.
      • Pepe M.S.
      • Quan J.M.
      • Otto K.L.
      • Montgomery A.B.
      • Williams-Warren J.
      • et al.
      Intermittent administration of inhaled tobramycin in patients with cystic fibrosis. Cystic fibrosis inhaled tobramycin study group.
      ] and Hodson [
      • Hodson M.E.
      • Gallagher C.G.
      • Govan J.R.
      A randomised clinical trial of nebulised tobramycin or colistin in cystic fibrosis.
      ] (Table 1). With the exclusion of Jensen [
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      ], the remaining studies were considered a homogenous group regarding study publication date. Five studies included a younger population (mean baseline age between 11 and 16 years), and six studies included an older population (mean baseline age between 20 and 32 years). Mean baseline FEV1% predicted was between 49.9 and 63.6 (i.e. moderate impairment (GOLD 2008) [
      • Global Initiative for Chronic Obstructive Lung Disease (GOLD)
      Global strategy for diagnosis, management, and prevention of chronic obstructive pulmonary disease.
      ]) in all studies except Nasr [
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      ] (mean baseline FEV1% predicted>80%) and Jensen [
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      ] (mean baseline FEV1% predicted>70%) where patients were considered to have mild impairment. The Lenoir study [
      • Lenoir G.
      • Antypkin Y.G.
      • Miano A.
      • Moretti P.
      • Zanda M.
      • Varoli G.
      • et al.
      Efficacy, safety, and local pharmacokinetics of highly concentrated nebulized tobramycin in patients with cystic fibrosis colonized with Pseudomonas aeruginosa.
      ] included a mixture of chronically- and intermittently-infected CF patients with P. aeruginosa infection; 24.1% and 16.7% of patients in the TIS-B and placebo arms, respectively, had a first or intermittent P. aeruginosa infection. Regarding prior exposure to study (active) drug, six studies had primarily treatment naïve populations (i.e. the majority of patients had never been exposed to the active drug before) and two studies had primarily exposed populations (i.e. most patients had previously used the active drug). In the Nasr [
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      ] and Jensen [
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      ] studies exposure status was unknown, and in the Oermann study [
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      ], one arm was exposed and one arm naïve to the study drugs. Patients who are maintained (previously exposed) on a nebulised antibiotic treatment appear to reach a plateau in FEV1% predicted after some time and do not show a great improvement if at all. Improvements in FEV1% predicted are therefore expected to be largest in patients who are naïve to the antibiotic. The population in each study arm was classified as naïve or exposed to the active ingredient tested, and within each trial there was consistency in exposure status. The Oermann data [
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      ], however, could not be included in the analysis as the study arm populations differed regarding naïve/exposed status.

      3.3 Base case results using % CFB in FEV1% predicted at 4 weeks

      The base case NMA combined individual study results from 7 of the 11 trials assessing % CFB in FEV1% predicted (Table 3). The Lenoir study [
      • Lenoir G.
      • Antypkin Y.G.
      • Miano A.
      • Moretti P.
      • Zanda M.
      • Varoli G.
      • et al.
      Efficacy, safety, and local pharmacokinetics of highly concentrated nebulized tobramycin in patients with cystic fibrosis colonized with Pseudomonas aeruginosa.
      ] was excluded because not all patients were chronically-infected with P. aeruginosa [
      • Döring G.
      • Hoiby N.
      for the Consensus Study Group
      Early intervention and prevention of lung disease in cystic fibrosis: a European consensus.
      ]. The Oermann study [
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      ] was excluded because the study arm populations differed regarding naïve/exposed status. The studies of Nasr [
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      ] and Jensen [
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      ], where patients had mild FEV1 impairment, were excluded. Among the 7 included studies, there was variation in mean age (2 studies<18 y vs. 5 studies>18 y) and exposure status (2 studies exposed vs. 5 studies naïve to the study drug). As such, the base case included a covariate for exposure.
      Table 3Percent change from baseline in FEV1% predicted in cystic fibrosis patients treated with different antibiotics from data reported in trials and relative treatment effects.
      SourceMean ageExposure% CFB (SE)Difference in % CFB (SE)
      PlaceboTIS-TTIS-BAZLIColistinTIP
      • Konstan M.W.
      • Geller D.E.
      • Minić P.
      • Brockhaus F.
      • Zhang J.
      • Angyalosi G.
      Tobramycin inhalation powder for P. aeruginosa infection in cystic fibrosis: the EVOLVE trial.
      Randomised and treated population with no imputation of missing data.
      <18 yNaive0.68 (2.81)13.97 (2.81)13.29 (3.97)
      • Konstan M.W.
      • Flume P.A.
      • Kappler M.
      • Chiron R.
      • Higgins M.
      • Brockhaus F.
      • et al.
      Safety, efficacy and convenience of tobramycin inhalation powder in cystic fibrosis patients: the EAGER trial.
      Randomised and treated population with no imputation of missing data.
      >18 yExposed3.60 (1.03)2.80 (1.20)0.8 (1.58)
      • Ramsey B.W.
      • Pepe M.S.
      • Quan J.M.
      • Otto K.L.
      • Montgomery A.B.
      • Williams-Warren J.
      • et al.
      Intermittent administration of inhaled tobramycin in patients with cystic fibrosis. Cystic fibrosis inhaled tobramycin study group.
      Randomised and treated population with no imputation of missing data.
      >18 yNaive0.07 (1.02)11.85 (1.53)11.78 (1.84)
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      Per protocol population.
      <18 yUnknown0.90 (9.14)1.36 (6.24)0.46 (11.07)
      • Hodson M.E.
      • Gallagher C.G.
      • Govan J.R.
      A randomised clinical trial of nebulised tobramycin or colistin in cystic fibrosis.
      Randomised and treated population with no imputation of missing data.
      >18 yExposed6.70 (2.14)0.37 (2.44)6.33 (3.25)
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      Per protocol population.
      <18 yUnknown−13.92 (14.37)−5.63 (11.50)8.29 (18.41)
      • Lenoir G.
      • Antypkin Y.G.
      • Miano A.
      • Moretti P.
      • Zanda M.
      • Varoli G.
      • et al.
      Efficacy, safety, and local pharmacokinetics of highly concentrated nebulized tobramycin in patients with cystic fibrosis colonized with Pseudomonas aeruginosa.
      Randomised and treated population with no imputation of missing data.
      <18 yNaive4.23 (5.65)27.92 (4.34)23.69 (7.12)
      • Chuchalin A.
      • Csiszér E.
      • Gyurkovics K.
      • Bartnicka M.T.
      • Sands D.
      • Kapranov N.
      • et al.
      A formulation of aerosolized tobramycin (TIS-B) in the treatment of patients with cystic fibrosis and Pseudomonas aeruginosa infection: a double-blind, placebo-controlled, multicenter study.
      Randomised and treated population with imputation using LOCF.
      <18 yNaive0.77 (2.02)12.97 (1.46)12.2 (2.49)
      • McCoy K.S.
      • Quittner A.L.
      • Oermann C.M.
      • Gibson R.L.
      • Retsch-Bogart G.Z.
      • Montgomery A.B.
      Inhaled aztreonam lysine for chronic airway Pseudomonas aeruginosa in cystic fibrosis.
      Randomised and treated population with imputation using LOCF.
      >18 yNaive2.00 (1.00)4.50 (1.00)6.5 (1.41)
      • Retsch-Bogart G.Z.
      • Quittner A.L.
      • Gibson R.L.
      • Oermann C.M.
      • McCoy K.S.
      • Montgomery A.B.
      • et al.
      Efficacy and safety of inhaled aztreonam lysine for airway Pseudomonas in cystic fibrosis.
      Randomised and treated population with imputation using LOCF.
      >18 yNaive2.44 (1.33)8.03 (1.33)10.47 (1.88)
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      Randomised and treated population, unknown imputation status.
      >18 yNaive and exposed0.55 (1.42)8.35 (1.42)7.8 (2.01)
      Excluded studies:
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      had mild FEV1 impairment,
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      had mild FEV1 impairment and old study,
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      had inconsistency due to naive and exposed status, and
      • Lenoir G.
      • Antypkin Y.G.
      • Miano A.
      • Moretti P.
      • Zanda M.
      • Varoli G.
      • et al.
      Efficacy, safety, and local pharmacokinetics of highly concentrated nebulized tobramycin in patients with cystic fibrosis colonized with Pseudomonas aeruginosa.
      included a non-chronic Pa infected population. Calculated data in italics (from mean CFB data or mean FEV1 at 4 weeks and baseline FEV1, SE from SD and sample size).
      Tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300 mg/5 ml (TIS-T), tobramycin inhalation solution 300 mg/4 ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), percent change from baseline (%CFB), standard error (SE), P. aeruginosa (Pa), forced expiratory volume in 1 second percentage (FEV1%).
      a Randomised and treated population with no imputation of missing data.
      b Per protocol population.
      c Randomised and treated population with imputation using LOCF.
      d Randomised and treated population, unknown imputation status.
      TIP and TIS-T were more efficacious than placebo in improving % CFB in FEV1% predicted (Table 4). For TIS-B, AZLI and colistin, the point estimates suggested improvement in % CFB in FEV1% predicted over placebo, although there was a fair degree of uncertainty around the point estimates. Regarding the treatment effect of TIP, point estimates suggested that TIP was superior to colistin and AZLI and was comparable to TIS-T and TIS-B; the level of uncertainty around point estimates, however, suggested a comparable efficacy of all treatments. TIS-T and TIS-B, as well as AZLI and colistin seem to offer a comparable efficacy.
      Table 4Network meta-analysis results: Difference in % CFB in FEV1% predicted at 4 weeks in Cystic Fibrosis patients treated with different antibiotics.
      PlaceboTIS-TTIS-BAZLIColistinTIP
      % CFB in FEV1% predicted at 4 weeks, (95% CrL)
      Placebo0
      TIS-T13.47

      (0.67, 26.29)
      0
      TIS-B13.55

      (-0.27, 27.3)
      0.09

      (-5.82, 6.06)
      0
      AZLI9.28

      (-3.76, 22.14)
      -4.19

      (-8.14, -0.21)
      -4.28

      (-9.68, 1.07)
      0
      Colistin7.15

      (-7.19, 21.49)
      -6.32

      (-12.61, -0.02)
      -6.40

      (-15.11, 2.18)
      -2.13

      (-9.58, 5.36)
      0
      TIP12.92

      (0.18, 25.73)
      -0.55

      (-3.50, 2.4)
      -0.64

      (-7.05, 5.74)
      3.64

      (-1.04, 8.26)
      5.77

      (-1.20, 12.75)
      0
      Tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300 mg/5 ml (TIS-T), tobramycin inhalation solution 300 mg/4 ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI), percent change from baseline (%CFB), forced expiratory volume in 1 second percentage (FEV1%), credible limit (CrL).

      3.4 Scenario analysis results

      Several scenario analyses were run with or without a covariate for baseline age, exposure or baseline FEV1, and in naïve or exposed subgroups or in different age subgroups: (1) The ‘No EXP covariate’ scenario included all 7 studies without a covariate for exposure; (2) The naïve subgroup included the 5 studies with no prior exposure to active drug; (3) The exposed subgroup included the 2 studies with previously exposed populations; (4) Scenario mild FEV1 included the Nasr [
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      ] and Jensen [
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      ] studies with baseline FEV1% predicted >70% and a covariate for baseline FEV1; (5) ‘All ages — cov AGE’ scenario included all 7 studies with a covariate for age rather than for exposure status; and (6) The subgroups for age either included the 2 studies with a mean age<18 years, or 5 studies with mean age>18 years.
      The effects of each intervention versus placebo were fairly consistent across each scenario compared to the base case (Fig. 3). The largest effects were seen for inhaled tobramycin (TIP, TIS-T and TIS-B), followed by AZLI and colistin (Fig. 4). The base case with no covariate scenario resulted in less uncertainty around results. It was not possible to compare the age subgroups, due to differences in comparators included in these subgroups, except for TIP; the effect of TIP versus placebo was higher in the subgroup with mean age<18 years compared to >18 years. Due to the lack of data, it was not possible to assess whether this difference was driven by the age differences or the naïve/exposed status differences.
      Figure thumbnail gr3
      Fig. 3Overview of network meta-analysis scenario results for change in lung function of different antibiotics versus placebo. Exposed subgroup network did not include placebo. Fixed effects model (FEM), random effects model (REM), exposure status (EXP), forced expiratory volume (FEV), covariate (Cov), percent change from baseline (%CFB), forced expiratory volume in 1 second percentage (FEV1%), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300 mg/5 ml (TIS-T), tobramycin inhalation solution 300 mg/4 ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI).
      Figure thumbnail gr4
      Fig. 4Overview of network meta-analysis scenario results for change in lung function of TIP versus alternative antibiotics. Fixed effects model (FEM), Random effects model (REM), exposure status (EXP), forced expiratory volume (FEV), covariate (Cov), percent change from baseline (%CFB), forced expiratory volume in 1 second percentage (FEV1%), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300 mg/5 ml (TIS-T), tobramycin inhalation solution 300 mg/4 ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI).

      3.5 Scenarios on impact of exposure status (Scenarios 1–3)

      The naïve subgroup did not allow a comparison to colistin and produced slightly greater differences in efficacy than the base case results, which is in line with the larger % CFB in FEV1% predicted observed in these studies compared to the exposed status studies. The comparison of TIP to colistin was possible via the indirect route TIP–TIS-T/TIS-T–colistin. Given the lack of comparators in the exposed subgroup, it was difficult to compare it to naïve subgroups. There was, however, a consistent trend across all scenario analyses in the effect of TIP relative to alternatives; with the largest benefits compared to placebo, colistin and AZLI, and with comparable benefits to TIS-T and TIS-B (Fig. 4).

      3.6 Scenario on impact of baseline FEV1% predicted (Scenario 4)

      A consistent trend to the base case was observed across all FEV1 analyses for the effect of TIP relative to alternatives. As Nasr [
      • Nasr S.Z.
      • Sakmar E.
      • Christodoulou E.
      • Eckhardt B.P.
      • Streetman D.S.
      • Strouse P.J.
      The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
      ] and Jensen [
      • Jensen T.
      • Pedersen S.S.
      • Garne S.
      • Heilmann C.
      • Høiby N.
      • Koch C.
      Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
      ] studies were small (≤40 patients) these had less weighting in the network, and therefore a minor impact on results.

      3.7 Scenarios on impact of mean baseline age (Scenarios 5 and 6)

      Compared to the base case analysis, adding a covariate for age rather than exposure produced similar results but greater uncertainty. The subgroup with mean age<18 years included all naïve patients, and only allowed a comparison between TIP and TIS-B. The subgroup with mean age>18 years included a mix of naïve and exposed status studies but did not include TIS-B in the network. Therefore comparisons between the <18 year and >18 year populations were limited by lack of data. Overall, a consistent trend to the base case was observed across all scenario analyses for the effect of TIP relative to alternatives.

      3.8 Other endpoints

      There were insufficient data to produce reliable results for % CFB in FEV1% predicted at 20 weeks, P. aeruginosa sputum density at 20 weeks, respiratory hospitalisation and anti-P. aeruginosa antibiotic use at 24 weeks. Only 3 studies reported these outcomes, with differences across studies. i.e., mix of naïve/exposed status and age groups, and limited comparisons possible. Regarding P. aeruginosa sputum density at 4 weeks; 7 studies reported this outcome and comparisons were possible between TIP, TIS-T, TIS-B, colistin and AZLI. For this outcome, TIP was expected to have better outcomes than placebo, and comparable outcomes to other active treatments. As this analysis included the same studies as for FEV1 at 4 weeks, there was variability across the studies due to age and exposure (Fig. 5). This resulted in a lot of uncertainty around results, and the limited data available made it difficult to draw conclusions. Subgroup analyses were limited because they did not allow comparisons to all treatments of interest, and the inclusion of a covariate to address the underlying variability increased the uncertainty of the results.
      Figure thumbnail gr5
      Fig. 5Overview of network meta-analysis scenario results for change in sputum density of TIP versus alternative antibiotics. Fixed effects model (FEM), random effects model (REM), exposure status (EXP), forced expiratory volume (FEV), covariate (Cov), percent change from baseline (%CFB), forced expiratory volume in 1 second percentage (FEV1%), tobramycin inhalation powder capsule (TIP), tobramycin inhalation solution 300 mg/5 ml (TIS-T), tobramycin inhalation solution 300 mg/4 ml (TIS-B), colistimethate sodium (colistin), aztreonam lysine for inhalation (AZLI).

      4. Discussion

      CF is a rare disease with few pivotal studies of inhaled antibiotics, and limited data on direct comparisons of currently available treatment options. NMA helps overcome the lack of head to head trials against all comparators. NMA is a valid accepted method to assess relative efficacy and provides relevant information for healthcare decisions [
      • Jansen J.P.
      • Fleurence R.
      • Devine B.
      • Itzler R.
      • Barrett A.
      • Hawkins N.
      • et al.
      Interpreting indirect treatment comparisons and network meta-analysis for health-care decision making: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: part 1.
      ]. However, NMA is limited by the quality of the data included. As randomisation only holds within a trial but not across trials, it is possible to have bias in indirect comparisons. In this study, we made every effort to avoid bias by ensuring only comparable trials, i.e. similar populations and study designs, were combined. To further control for potential bias, we included the following covariates to control effect modifiers in our models: 1) exposure status; 2) baseline FEV1% predicted; and 3) age.
      Five antibiotics were presented in 11 trials, with a large amount of heterogeneity across studies (e.g. naïve/exposed status, mild/moderate FEV1% predicted, age differences, publication date) which hampered the analyses and resulted in relatively low precision of the estimates. Controlling for both potential effect modifiers (mean age and previous exposure to the study drug) was not possible simultaneously in this small dataset. Therefore scenarios on exposure were not able to control for age differences, and vice versa. Despite these drawbacks, the analyses, based on point estimates, suggested that all treatments provided an improvement compared to placebo, and that tobramycin formulations provided an improvement over AZLI or colistin. However, credible limits were too large to make this statistically significant. The difference in % CFB for TIP compared to placebo, colistin and AZLI can be considered clinically meaningful, using a cut-off of 3.5% CFB in FEV1% predicted. Our results are comparable to the conclusions of a previous Cochrane review [
      • Ryan G.
      • Singh M.
      • Dwan K.
      Inhaled antibiotics for long-term therapy in cystic fibrosis.
      ]. However; a meta-analysis was, at that time, not conducted judging study designs and reporting of results too variable.
      Prior exposure to active drug was identified as a key factor affecting outcomes, yet, this was not typically reported in trials as a predictive factor. Most trials involved the first use of the active drug, and therefore had a population who was naïve to the active drug. The head to head studies were conducted after approval of colistin and tobramycin, and therefore included a population that was already exposed to these drugs. From the data reported in trials, there was generally a larger gain in % CFB in FEV1% predicted for naïve populations than for exposed populations. More recent trials, such as the Oermann study [
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      ], are more likely to include exposed populations as the majority of chronically-infected patients are now regularly treated with inhaled tobramycin or colistin. The Oermann study [
      • Oermann C.M.
      • Assael B.
      • Nakamura C.
      • Boas S.R.
      • Bresnik M.
      • Montgomery A.B.
      • et al.
      Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
      ], however, included one arm with prior exposure to the active drug (TIS-T) and one arm naïve to the active drug (AZLI). Therefore the study design presented an inconsistency in the network by not comparing ‘like with like’. Objective comparison between the treatments in this study was therefore challenging.
      Future trials and comparisons will increasingly need to consider study design and population differences. Additional research is needed on potential effect modifiers to understand their impact on outcomes and their relative importance. One such example is chest computed tomography to characterize structural lung abnormalities at baseline [
      • Tiddens H.A.W.M.
      • De Jong P.A.
      Imaging and clinical trials in cystic fibrosis.
      ,
      • Loeve M.
      • van Hal P.T.W.
      • Robinson P.
      • de Jong P.A.
      • Lequin M.H.
      • Hop W.C.
      • et al.
      The spectrum of structural abnormalities on CT scans from CF patients with severe advanced lung disease.
      ]. Future studies should include a more homogenous population regarding age, baseline FEV1% predicted and inclusion of either chronically- or non-chronically-infected patients. The major study design difficulty is in controlling for exposure status as most patients today will have prior exposure to several antibiotics. This field needs clinicians and researchers to align the approach to study design, population selection and the most relevant comparative outcomes.

      Acknowledgements

      The authors wish to thank Gert Bergman (Mapi Consultancy) for his review of the analysis and manuscript.
      This study was sponsored by Novartis Pharma AG.

      References

        • Döring G.
        • Ratjen F.
        Immunology of cystic fibrosis.
        in: Hodson M.E. Geddes D. Bush A. Cystic fibrosis. Arnold Hammer, London, England2007: 69-80
        • Hoiby N.
        • Frederiksen B.
        Microbiology of cystic fibrosis.
        in: Hodson M.E. Geddes D. Cystic fibrosis. Arnold, London, England2000: 83-108
        • Tsui L.-C.
        The spectrum of cystic fibrosis mutations.
        Trends Genet. 1992; 8: 392-398
        • Döring G.
        • Gulbins E.
        Cystic fibrosis and innate immunity: how chloride channel mutations provoke lung disease.
        Cell Microbiol. 2009; 11: 208-216
        • Döring G.
        • Hoiby N.
        • for the Consensus Study Group
        Early intervention and prevention of lung disease in cystic fibrosis: a European consensus.
        J Cyst Fibros. 2004; 3: 67-91
        • Lu G.
        • Ades A.E.
        Combination of direct and indirect evidence in mixed treatment comparisons.
        Stat Med. 2004; 23: 3105-3124
        • Jansen J.P.
        • Fleurence R.
        • Devine B.
        • Itzler R.
        • Barrett A.
        • Hawkins N.
        • et al.
        Interpreting indirect treatment comparisons and network meta-analysis for health-care decision making: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: part 1.
        Value Health. 2011; 14: 417-428
        • Caldwell D.M.
        • Ades A.E.
        • Higgins J.P.
        Simultaneous comparison of multiple treatments: combining direct and indirect evidence.
        BMJ. 2005; 331: 897-900
        • Cooper N.J.
        • Sutton A.J.
        • Morris D.
        • Ades A.E.
        • Welton N.J.
        Addressing between-study heterogeneity and inconsistency in mixed treatment comparisons: application to stroke prevention treatments in individuals with non-rheumatic atrial fibrillation.
        Stat Med. 2009; 28: 1861-1881
        • Lunn D.J.
        • Thomas A.
        • Best N.
        • Spiegelhalter D.
        WinBUGS — a Bayesian modelling framework: concepts, structure, and extensibility.
        Stat Comput. 2000; 10: 325-337
        • Jensen T.
        • Pedersen S.S.
        • Garne S.
        • Heilmann C.
        • Høiby N.
        • Koch C.
        Colistin inhalation therapy in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung infection.
        J Antimicrob Chemother. Jun 1987; 19: 831-838
        • Ramsey B.W.
        • Pepe M.S.
        • Quan J.M.
        • Otto K.L.
        • Montgomery A.B.
        • Williams-Warren J.
        • et al.
        Intermittent administration of inhaled tobramycin in patients with cystic fibrosis. Cystic fibrosis inhaled tobramycin study group.
        N Engl J Med. Jan 7 1999; 340: 23-30
        • Hodson M.E.
        • Gallagher C.G.
        • Govan J.R.
        A randomised clinical trial of nebulised tobramycin or colistin in cystic fibrosis.
        Eur Respir J. Sep 2002; 20: 658-664
        • Global Initiative for Chronic Obstructive Lung Disease (GOLD)
        Global strategy for diagnosis, management, and prevention of chronic obstructive pulmonary disease.
        2008
        • Nasr S.Z.
        • Sakmar E.
        • Christodoulou E.
        • Eckhardt B.P.
        • Streetman D.S.
        • Strouse P.J.
        The use of high resolution computerized tomography (HRCT) of the chest in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis lung disease.
        Pediatr Pulmonol. May 2010; 45: 440-449
        • Lenoir G.
        • Antypkin Y.G.
        • Miano A.
        • Moretti P.
        • Zanda M.
        • Varoli G.
        • et al.
        Efficacy, safety, and local pharmacokinetics of highly concentrated nebulized tobramycin in patients with cystic fibrosis colonized with Pseudomonas aeruginosa.
        Paediatr Drugs. 2007; 9: 11-20
        • Oermann C.M.
        • Assael B.
        • Nakamura C.
        • Boas S.R.
        • Bresnik M.
        • Montgomery A.B.
        • et al.
        Aztreonam for inhalation solution (AZLI) vs. tobramycin inhalation solution (TIS): a 6-month comparative trial in cystic fibrosis (CF) patients with Pseudomonas aeruginosa (PA).
        in: 24th Annual North American Cystic Fibrosis Conference. Oct 2010 (Poster 305)
        • Ryan G.
        • Singh M.
        • Dwan K.
        Inhaled antibiotics for long-term therapy in cystic fibrosis.
        Cochrane Database Syst Rev Mar. 2011; 16 (CD001021)
        • Tiddens H.A.W.M.
        • De Jong P.A.
        Imaging and clinical trials in cystic fibrosis.
        Proc Am Thorac Soc. 2007; 4: 343-346
        • Loeve M.
        • van Hal P.T.W.
        • Robinson P.
        • de Jong P.A.
        • Lequin M.H.
        • Hop W.C.
        • et al.
        The spectrum of structural abnormalities on CT scans from CF patients with severe advanced lung disease.
        Thorax. 2009; 64: 876-882
        • Chuchalin A.
        • Csiszér E.
        • Gyurkovics K.
        • Bartnicka M.T.
        • Sands D.
        • Kapranov N.
        • et al.
        A formulation of aerosolized tobramycin (TIS-B) in the treatment of patients with cystic fibrosis and Pseudomonas aeruginosa infection: a double-blind, placebo-controlled, multicenter study.
        Paediatr Drugs. 2007; 9: 21-31
        • Konstan M.W.
        • Flume P.A.
        • Kappler M.
        • Chiron R.
        • Higgins M.
        • Brockhaus F.
        • et al.
        Safety, efficacy and convenience of tobramycin inhalation powder in cystic fibrosis patients: the EAGER trial.
        J Cyst Fibros. Jan 2011; 10: 54-61
        • Konstan M.W.
        • Geller D.E.
        • Minić P.
        • Brockhaus F.
        • Zhang J.
        • Angyalosi G.
        Tobramycin inhalation powder for P. aeruginosa infection in cystic fibrosis: the EVOLVE trial.
        Pediatr Pulmonol. 2011; 46: 230-238
        • McCoy K.S.
        • Quittner A.L.
        • Oermann C.M.
        • Gibson R.L.
        • Retsch-Bogart G.Z.
        • Montgomery A.B.
        Inhaled aztreonam lysine for chronic airway Pseudomonas aeruginosa in cystic fibrosis.
        Am J Respir Crit Care Med. Nov 1 2008; 178: 921-928
        • Retsch-Bogart G.Z.
        • Quittner A.L.
        • Gibson R.L.
        • Oermann C.M.
        • McCoy K.S.
        • Montgomery A.B.
        • et al.
        Efficacy and safety of inhaled aztreonam lysine for airway Pseudomonas in cystic fibrosis.
        Chest. May 2009; 135: 1223-1232