Abstract
The Standardisation Committee of the European Cystic Fibrosis Society Clinical Trial Network has undertaken the evaluation of clinical end-points for therapeutic interventions regarding their use in multicentre clinical trials in cystic fibrosis (CF). This review of biomarkers in bronchoalveolar lavage (BAL) is part of the group’s work.
The aims of this project were: 1) to review the literature on reliability, validity and responsiveness of BAL in patients with CF; 2) to gain consensus of the group on the feasibility of BAL; and 3) to gain consensus on answers to key questions regarding the promotion of BAL to surrogate end-point status.
Assessment of BAL inflammatory markers in the literature indicates that their reliability, validity and responsiveness are adequate for clinical trials. After discussion of the practical characteristics it was concluded that BAL has an attractive validity profile, albeit with limited feasibility. It is particularly applicable to multicentre trials in preschool children with CF and early or mild lung disease. This is the first article to collate the literature in this manner. This provides a rationale to support the use of BAL in early clinical trials in preschool children with CF.
Abstract
Review of the literature provides rationale to support the use of BAL in well-designed early-phase clinical trials in CF http://ow.ly/qGInZ
Introduction
The European Cystic Fibrosis Society Clinical Trial Network (ECFS-CTN) Standardisation Committee is undertaking a rigorous evaluation of outcome measures used in clinical trials in cystic fibrosis (CF). The committee is composed of six groups consisting of researchers with expertise in specific outcome measures: CF transmembrane conductance regulator biomarkers [1], respiratory function, inflammatory markers, anthropometrics, microbiology definitions and chest imaging. This article summarises the work of the inflammatory markers group on inflammatory markers obtained via bronchoalveolar lavage (BAL) and is one of a series of documents from the six groups.
Outcome measures can be classified as clinical end-points, surrogate end-points or biomarkers. Clinical end-points reflect how a patient feels, functions or survives and detect a tangible benefit for the patient [2, 3]. A surrogate end-point is a laboratory measurement used to predict the efficacy of therapy [2, 3] when direct measurement of clinical effect is not feasible or practical. Surrogate end-points may shorten the period of follow-up required; however the link between the surrogate end-point and long-term prognosis must be proven. Forced expiratory volume in 1 s (FEV1) is still the only accepted surrogate outcome for the European Medicines Agency and the US Food and Drug Administration. With regard to their use as outcome measures, inflammatory markers obtained via BAL are currently considered to be “biomarkers”. A biomarker is defined as “a characteristic that is objectively measured and evaluated as an indicator of normal biologic processes, pathogenic processes or pharmacologic response to a therapeutic intervention” [2, 3]. Biomarkers are mainly used to explore proof-of-concept for a specific compound. Some are currently being considered for “promotion” to the status of surrogate end-point. A full description of the classification of outcome measures is provided in the first document in the series of articles from the ECFS-CTN Standardisation Committee [1].
BAL inflammatory markers that may be considered, depending on the clinical or research goal [4, 5], include total and differential cell counts (in particular neutrophils), elastase and its complexes with inhibitors (for example neutrophil elastase (NE)-α1-antitrypsin) [6, 7], pro- and anti-inflammatory cytokine/chemokine levels (interleukin (IL)-8 and tumour necrosis factor (TNF)-α), levels of eicosanoids and leukotrienes, biomarkers of oxidative stress (such as 3-chlorotyrosine) and the inflammatory state of epithelial cells recovered from the airways. For the purpose of the present report the working group focused on BAL polymorphonuclear (PMN) neutrophil count or percentage, NE, IL-6 and IL-8.
To gain acceptance by researchers and licensing bodies, an end-point must have a body of supporting evidence: acceptable clinimetric properties such as reliability, validity and responsiveness to treatment, and sufficient feasibility and safety (table 1). Clinimetric properties and feasibility are population- and situation-dependent, therefore data cannot readily be extrapolated to the CF population from other disease populations. It is necessary to provide an evidence base from CF specifically. It is also necessary to consider subgroups of patients, for example age range and clinical severity may be important. The intervention is also an important factor when considering the responsiveness of an outcome measure.
The aims of this study were: 1) to review the literature on clinimetric properties (e.g. reliability, validity and responsiveness) of BAL inflammatory markers in patients with CF; 2) to gain consensus of the group on the feasibility of BAL; and 3) to gain consensus on answers to key questions regarding the promotion of BAL inflammatory markers to surrogate end-point status.
Methods
An exhaustive literature search was conducted in MEDLINE, Allied and Complementary Medicine (AMED) and Embase using a combination of keywords: “bronchoalveolar lavage” or “pulmonary inflammation” or “airway inflammation” or “lung inflammation” or “neutrophil” or “polymorphonuclear” or “PMN” or “interleukin” or “IL-6” or “IL-8” or “cytokine” or “epithelial lining fluid” or “inflammatory marker”, and “cystic fibrosis”. The search was limited to full text articles in the English language, with no limits on year of publication. A bibliography search was also conducted of all included articles and other relevant literature such as reviews and editorials.
For clinimetric properties, data were extracted and tabulated for reliability, validity, correlation with other outcome measures, responsiveness and reference values. Definitions are given in table 1. To evaluate feasibility, data were extracted and tabulated on the proportion of attempts that were successful and reasons for excluding tests. An expert panel also discussed the following topics at several face-to-face meetings (Venice, Italy: November 17 and 18, 2010; Hamburg, Germany: June 9, 2011; Rotterdam, the Netherlands: March 23 and 24, 2012) and reached consensus on each: risk involved, cost, ease of performance, ease of administration, time to administer, equipment and space needed, and applicable age group. Specific advantages and limitations of BAL were also discussed.
Narrative answers to four key questions were prepared, discussed and agreed by the group. 1) Do BAL inflammatory markers have the potential to become surrogate outcome measures? 2) For what kind of therapeutic trial are BAL inflammatory markers appropriate? (Therapeutic aim, phase of trial, target population, number of patients involved and number of sites involved.) 3) Within what timeline can change be expected and what treatment effect can be considered clinically significant? 4) What are the most needed studies to further define BAL inflammatory markers in patients with CF and to explore their potential as a surrogate outcome measures? The consensus of the group is presented in the current article.
Results
Why use BAL inflammatory markers in clinical trials in CF?
Flexible bronchoscopy with BAL provides a regional measurement of lung inflammation and infection within the respiratory tract, including the small airways and alveoli [8]. Inflammation has been shown to be present early in CF lung disease before changes are detected by traditional pulmonary function techniques such as spirometry [9]. The ability to identify early airway inflammation and infection in these “silent years” is of great importance for investigating new therapies in infants and young children.
Being the most direct measurement of infection and inflammation of the lower airways and distal lung, BAL has greatly contributed to the understanding of the pathophysiological process in CF, especially in infants and young children. Lung secretions from patients with CF contain large concentrations of TNF-α and interleukins/chemokines (IL-8, IL-6, IL-1, etc.) [5]. All of these cytokines share a common characteristic: their synthesis is promoted by nuclear factor-κB, which is activated by cellular interaction with bacteria, bacterial products and proinflammatory cytokines. IL-8 is a chemoattractant of PMN cells and its mRNA increases with osmotic stress, which is a part of the pathogenesis of CF [10, 11]. Cell counts and mediators reflect endobronchial inflammation. PMN cell infiltration induces the release of oxidants and proteases, leading to lung tissue degradation. CF lungs may be primed for inflammation, and bacterial and/or viral infection may potentiate such inflammation [12]. Inflammation with increased numbers of PMN cells is seen in BAL from the first few weeks of life, independent of clinical signs and symptoms. In early disease, BAL changes represent a preclinical indicator of possible underlying (but clinically silent) inflammation [13–15]. A substantial proportion of infants diagnosed with CF after detection by newborn screening have active pulmonary inflammation, 30% have detectable NE activity, 20% have pulmonary infection and 80% have evidence of structural lung disease on chest computed tomography (CT) at 3 months of age [9].
Together with chest CT, BAL inflammatory markers have been used as clinical end-points in pathophysiological studies (Australian Respiratory Early Surveillance Team for CF) [9], in clinical trials of tobramycin inhaled solution [16–18] and in clinical trials of recombinant human (rh)DNase [19–21]. It has also been used to direct therapy against Pseudomonas aeruginosa infection (Australasian CF Bronchoalveolar Lavage) [22]. However, no soluble marker of inflammation has been studied in a comprehensive, longitudinal manner so as to be validated as a “gold standard” for assessing the inflammatory response [23, 24].
Clinimetric properties of BAL inflammatory markers
Reliability
To the best of our knowledge, short-term reproducibility has not been studied in CF. One study of 60 BAL samples from intubated non-CF children in the intensive care unit showed satisfactory short-term reproducibility of cell counts in blind nonbronchoscopic BAL, except for lymphocytes [25] (table E1). In noninfected infants with CF, some insight regarding long-term reproducibility has been provided by repeat BAL conducted 6–18 months apart, indicating a certain degree of variability related to an ongoing inflammatory process [26]. Similarly, Paul et al. [19] showed that the PMN cell percentage and NE in pooled BAL samples significantly increased over time (up to 36 months) in children with CF not treated by rhDNase.
Although considered the gold standard for sampling lower respiratory secretions, BAL may produce variable results due to regional heterogeneity of inflammation [23, 27]. PMN cell count and percentage and NE activity are generally greater in upper-lobe BAL fluid (BALF) in adults [27]. Upper- versus lower-lobe differences were more pronounced in subjects with better preservation of lung function [27].
Concurrent validity
Due to the localised nature of enhanced inflammation in CF, respiratory secretions from the lower airways are considered by many researchers to be the optimal source of inflammatory cells and soluble mediators [5]. BAL is considered the best method for obtaining samples from the lower respiratory tract (especially in very young children who are too young to successfully undergo sputum induction with hypertonic saline), thus yielding the most accurate measure of the infectious and/or inflammatory process in the CF airway lumen [24]. However, it has been shown that the inflammatory patterns and responses to infective stimuli in the airway lumen and the airway wall of children with CF are distinct and compartmentalised, and thus BAL and endobronchial biopsy provide different but complementary information [28]. In contrast to the neutrophil dominated inflammation present in the airway lumen, the bronchial mucosa is characterised by the recruitment and accumulation of lymphocytes [28].
Convergent validity
In young pre-adolescent children, there is a good correlation (r>0.55) between BAL PMN cell count/percentage and IL-8 (five out of five studies [13, 29–32]) or neutrophil elastase (one study) [13]. The correlation was not a constant finding in older patients and adults (three out of four studies) [27, 33–35] (table E2).
One study reported good correlation (r2=0.48, p=0.036) between the inflammatory response (IL-8 and PMN cells) in the upper airway, sampled by nasal washes, and the lower airways, sampled by BAL [36]. The correlation between inflammatory markers obtained via BAL and other sampling methods is variable. Two studies showed no correlation between BAL and induced sputum [6, 34], while one study showed excellent agreement between IL-8 recovered from BALF and spontaneously expectorated sputum (r=0.967, p=0.007) [37]. In contrast to induced sputum, which originates from the airways, BAL samples the bronchial or terminal airways and air spaces (alveoli), according to the fractions analysed (the latter aliquots more likely representing the alveolar region). Induced sputum and BAL thus provide different but complementary data.
Discriminate validity
Discriminate validity of BAL inflammatory markers is shown in Tables 2, 3 and 4.
The majority of studies (17 out of 19 studies [6, 26, 31, 33, 35, 37–43, 45–51]) indicate that all four biomarkers are present at significantly higher levels in CF versus non-CF patients, the latter consisting mainly of patients undergoing bronchoscopy for other causes such as recurrent lower respiratory tract infection, chronic cough and upper airway obstruction, over a wide age span (from 2 weeks to adult). However, there was no significant difference in BAL IL-8, NE and PMN cells between children with CF and primary ciliary dyskinesia (one study) [45].
Several studies have demonstrated the ability of BAL inflammatory markers to discriminate between groups of patients with CF who have different degrees of lung disease severity. BAL PMN cell percentage (two out of two studies [9, 30]) and NE (one study) [9] are significantly more elevated in symptomatic versus asymptomatic children with CF (but not adults (one study) [53]).
BAL PMN cell counts (seven out of eight studies) and percentage (seven out of seven studies), NE (four out of five studies) and IL-8 (eight out of 11 studies) have been shown to increase in children infected with P. aeruginosa, Staphylococcus aureus and/or viruses versus no pathogens [9, 13, 26, 30, 39, 43, 52–54, 56]. However, BAL biomarkers are similar in the presence of P. aeruginosa versus other pathogens [41, 52, 57]. In one study, PMN cell count and percentage were increased in the presence of a mucoid-type versus nonmucoid P. aeruginosa [57]. Clinical presentation was also shown to differ between patients grouped according to the level of BAL inflammatory markers [31, 56, 58] (table 4).
Taken together, among the five biomarkers evaluated, PMN cell count and percentage appear to be the most consistently discriminant. BAL IL-8 is less consistently increased and NE and IL-6 appear to be of a much lesser utility in older children and adults.
Predictive validity
To the best of our knowledge, no study has looked specifically at the predictive value of BAL biomarkers in CF.
Correlation with other outcome measures
In older children and adults, BAL NE is significantly inversely correlated to FEV1 (two out of two studies [45, 59]). BAL PMN cell percentage is most constantly (but moderately) associated to the other measures of lung function, such as respiratory system compliance [54], tissue elastance [30], forced expiratory flow at 25–75% of forced vital capacity [32] or lung distension [32]. Lung function impairment is not a good marker of BAL IL-8 levels (significant correlation in only two out of five studies [30, 32, 52, 54, 59]).
BAL biomarkers have been compared with radiological images obtained from CT scanning, deemed the best means of evaluating early structural lung damage. In 57 infants (median age 3.6 months), radiological evidence of structural lung disease was common, with 46 (80.7%) having an abnormal CT: 11 (18.6%) had bronchial dilatation, 27 (45.0%) had bronchial wall thickening and 40 (66.7%) had gas trapping [9]. As shown in table E3, the inflammatory marker most associated with structural lung disease is free NE activity, which has been significantly associated with presence and extent of bronchial dilatation (OR 10.9 and 10.8, respectively), presence and extent of bronchial wall thickening (OR 9.4 and 4.3, respectively) and extent of airway trapping (OR 3.9) [9]. No significant association was found between free NE activity and presence of air trapping.
BAL biomarkers are not consistently correlated with clinical status (respiratory rate and arterial oxygen saturation (SaO2)), anthropometric parameters and Shwachman score (six studies [27, 29, 31, 37, 52, 60]).
Responsiveness
BAL inflammatory markers have been demonstrated to be responsive to a number of interventions in patients with CF (table 5).
Studies have demonstrated the responsiveness of BAL inflammatory markers to acquisition of infection and to treatment with antibiotics. In long-term longitudinal studies in infants, the acquisition of bacterial infection is accompanied by a greater than two-fold increase in BAL PMN cell count, NE and IL-8 [13, 26]. A decrease in NE (one study in adults [61]) was observed 2–6 weeks after i.v. antibiotics. Long-term studies show that inhaled antibiotics (tobramycin inhalation solution) inconsistently decrease PMN cell percentages or counts (one out of three studies [16–18]), but do not influence IL-6 or IL-8 (none out of three studies [16–18]). Repeat BAL performed up to 18 months after eradication of infection in young infants was accompanied by a sustained reduction in NE and IL-8 [13, 26].
Treatment with inhaled rhDNase has been shown to attenuate lung inflammation over a 3-year study period, with significant decreases in PMN cell percentage, NE and IL-8 in the treated group, contrasting with no change in the placebo group [19]. Nonsignificant data observed in some smaller studies may be due to small sample size [20, 21]. In adults, NE is significantly decreased 12 h following recombinant secretory leukoprotease inhibitor therapy [62].
The range of detectable change is variable, depending on the parameters studied. A floor effect may be present for certain markers, e.g. nondetectable levels of IL-8 [9]. However, there is no ceiling effect regarding absolute number of inflammatory cells or soluble markers (multi-log increases may be observed), although it may be possible to observe almost 100% PMN cells in BALF, i.e. PMN cells being the sole BAL cell subtype.
Reference values
BAL inflammatory biomarker levels in infants, older children and adults with and without CF are shown in table E4. A review including studies conducted in Australia, USA and Europe indicated “normal” values of BAL PMN cells, IL-8 and NE in young children and infants with CF [63]. The mean reported values for BAL PMN cells are 17.3% [30] or 134×103 cells·mL−1 [13] in noninfected, nonsymptomatic infants. In such infants, NE is undetectable and IL-8 levels are very low (24 pg·mL−1) [13]. Values vary according to whether the first versus pooled BALF aliquots are analysed. Neutrophil dominated airway inflammation is more pronounced in the first (mean±sd 52±30%), mainly bronchial, BAL sample versus the pooled second and third BALF aliquots (mean±sd 29±23%), suggesting that sequential analysis of BALF may have a higher sensitivity to detect early inflammatory changes in CF patients [48]. In stable adults with CF and chronic bacterial infections, PMN cell count is ≥200×103 cells·mL−1, PMN cells exceed 37.5% of total cell counts, with detectable levels of NE and IL-8 [27].
Feasibility and safety
Two studies included in the review reported success rates [13, 19] (table E5). One study reported a 78% success rate for a one-time BAL with loss of data due to insufficient BALF recovery [13]. One longitudinal study reported a dropout rate of 48–58% due to an unwillingness to repeat bronchoscopy [19].
Eight studies out of those included in this review were found to describe short-term (hours or days) adverse events related to BAL procedures [6, 8, 22, 50, 52, 54, 64, 65] (table E6). Reported adverse events were mainly mild, for example mild fever, worsening of cough, mild respiratory distress or stridor. There is a small risk of severe adverse events due to BAL, for example severe bronchoconstriction has been reported [52]. Bronchoscopy also increases the risk of contamination of the lower airways. In adults, the rate of complications may be greater (three out of 11 patients experienced serious adverse events following BAL in one study [6, 64]).
The large long-term (years) lavage studies have been conducted in CF patients only and no normal controls were studied in parallel [19, 21, 22]. The Australian CF BAL study indicates that the long-term tolerance of repeated BAL (initiated prior to the age of 6 months until the age of 5 years) is satisfactory. In this study, CF controls were treated similarly but without lavage. At the age of 5 years, eight (10%) out of 79 children in the BAL-directed therapy group and nine (12%) out of 76 in the standard therapy group had P. aeruginosa in final BAL cultures (p=0.73). Mean total CF-CT scores for the BAL-directed therapy and standard therapy groups were 3.0% and 2.8%, respectively (p=0.74). No long-term side-effects were reported in the paper.
Group consensus on feasibility
The cost of a bronchoscopy/BAL procedure varies across Europe. Self-reported costs by ECFS investigators are in the region of €100–213 per procedure, excluding the cost of the hospitalisation. Costs may also differ according to mediator(s) analysed and analysis technique used (estimated costs range from €18 for differential counts (including preparation in the UK), €8 (Australia) to €38 (including preparation, UK) for a cytokine assay such as IL-8, and €4 (Australia) per sample for NE). The cost may be doubled if the values are very high and the assay has to be repeated on diluted samples (in ∼10–20% of samples). Moreover, analysing the different fractions of the lavages will increase cost and effort.
In experienced hands, BAL is a relatively easy procedure. The actual time the patient will have to bear the bronchoscope in the lower airways may be limited to a few minutes. In children, general anaesthesia increases the patient’s tolerance of the procedure.
Extensive training is mandatory, with clear protocols, in order to reduce the risk of lower airway contamination and to reduce variability in results. North American Therapeutics Development Network standard operating procedures are available for sampling and processing of BAL. There is a need for specialised laboratories and trained technologists to process the samples. Some methodological issues can arise, for example concentrations of IL-8 are higher with cytokine bead analyses [66] compared to ELISA.
In infants and children, it is recommended that BAL be performed under general anaesthesia in circumstances where full anaesthetic monitoring and recovery are available (fully equipped operating theatres). In adults, bronchoscopy is generally conducted under local anaesthesia with or without sedation. A total of 30–45 min is generally required to conduct the whole procedure. Paediatric bronchoscopes come in different sizes. Beyond the neonatal period, a flexible bronchoscope with an external diameter of 3.5–3.7 mm is recommended up to the age of 9 years. Thereafter, 4.6–4.9-mm diameter bronchoscopes are more appropriate [67].
The requirement for a combination of well-adapted equipment, experienced personnel (pulmonologists, anaesthetists, nurses, and laboratory technicians and practitioners) and fully functional operating theatre limits availability of this procedure to tertiary care medical centres.
Advantages and limitations of BAL inflammatory markers
BAL offers the advantages of being able to be performed in any age group, and for its ability to directly sample the same area of the lung for studies that require repeat sampling (as is commonly done with intervention studies). Of note, BAL has been extensively studied in children (>333 procedures in 107 infants in one study [65]), but not adults, in particular in the presence of very altered lung function (FEV1 <40%) [13, 22, 26, 52, 65]. In this latter subgroup of patients, the volume recovered may be decreased and the risk of the procedure increased. It is noteworthy that inflammatory markers may change with age (e.g. IL-8 increases with age), requiring placebo controls for long-term studies.
The invasive character of BAL limits its repeatability in clinical practice and research. Lavage of multiple lobes [8] or all lobes may be appropriate in some studies, and in studies involving serial lavages over time, these lavages should be performed in the same lobe. In addition, BAL techniques may vary to some degree between centres. Sources of variability include type of anaesthesia/sedation, introduction of bronchoscope via nose versus laryngeal mask airway, bronchoscope size (which determines the extent of the lavaged lung areas), lavage aliquot size and number, and lobe(s) lavaged [25]. In order to decrease such variability, a European BAL standard operating procedure has been developed, and is available on request from the ECFS-CTN (ECFS-CTN@uzleuven.be).
One major factor that has the potential to cause difficulty for combining data from different studies is the dilution of the lung specimen. There is general agreement that correction for dilution is extremely difficult, if not impossible. Unlike urine, where markers can be corrected for dilution using urinary creatinine concentrations, no such marker of BAL dilution exists. Both urea and albumin have been used in attempts to correct for dilution but neither is suitable. As discussed in the European Respiratory Society Task Force report [67], there are major drawbacks with both, especially the fact that the concentrations of both urea and albumin increase in lung disease. For surveillance of inflammation in CF reporting the following is recommended: 1) the volume returned as a percentage of that instilled [63, 67]; 2) total cell count [63, 67]; and 3) differential cell counts, i.e. macrophages, neutrophils, lymphocytes and other cell types, as the number of cells [63] and percentage of cell types [67] per millilitre of BALF.
Noncellular components, such as cytokines, inflammatory markers and NE should be reported as the concentration per millilitre of BALF. In addition, where the marker is only detectable once the body’s defence has been overwhelmed, e.g. unbound NE detected in an activity assay, then this should be reported both as a binary variable (detectable or not) and as the concentration per millilitre of BALF [63].
Group Consensus: The “Four Key Questions”
Question 1: does this outcome have the potential to become a surrogate outcome?
Biomarkers of inflammation are particularly clinically and biologically relevant in CF. However, although it may be a useful surrogate marker, BAL is limited by its invasive nature.
Question 2: for what kind of therapeutic trial is this outcome appropriate? (Therapeutic aim, phase of trial, target population, trial duration, number of patients involved, number of sites involved)
Because of its invasive nature and expense, BAL is best suited for early phase clinical trials (phases 1 and 2), conducted in very specialised centres. BAL cultures have been shown to be an informative primary end-point for trials of antimicrobial therapies as well as anti-inflammatory therapies. BAL may be applied in short-term (weeks) as well as long-term (years) studies in preschool children.
Question 3: within what timeline can change be expected? What treatment effect can be considered clinically significant?
A minimum of 1 month may be sufficient for specific anti-inflammatory biomarkers and therapeutic agents. However, it is to be expected that generally biomarkers of inflammation will change over months rather than weeks. There are insufficient data and appropriate clinical outcome studies (e.g. using sensitive markers such as CT scans or sufficient duration) to confidently determine clinically significant levels of improvement in inflammatory markers.
Question 4: what are the most needed studies to further define this outcome parameter in CF patients and its potential to be a surrogate marker?
Further studies on the standardisation and reproducibility of this technique are warranted. Importantly, the original BAL routine of pooling second and third aliquots for inflammatory markers (initially established to investigate interstitial lung disease) should be investigated in CF. Furthermore, it may be important to determine which marker(s) are most informative, to show that changes in BAL characteristics are accompanied by modifications in the natural history of the disease (inflammation and/or infection) (i.e. longitudinal studies correlating change in biomarkers to a change in clinical status), to evaluate the validity of biomarkers at different stages of disease, among larger numbers of carefully phenotyped subjects with CF of varying ages, genotype, colonisation status and to perform “multimodal” measurement studies and explore the influence of multimodal anti-inflammatory therapies.
Conclusion
This document provides an overview of the work of the ECFS-CTN Standardisation Committee on BAL inflammatory markers. Assessments of BAL inflammatory markers in the literature indicate that their reliability, validity and responsiveness are adequate for clinical trials. The practical characteristics were discussed and it was concluded that the use of BAL in clinical research should be limited due to its invasive nature. It is particularly applicable to early-phase clinical trials (phases 1 and 2), conducted in very specialised centres. It is also applicable to multicentre trials in preschool children with early or mild lung disease. This is the first article to collate the literature in this manner and provides a rationale to support the use of BAL in well-designed early-phase clinical trials in CF.
Acknowledgments
This study was completed with the participation of the Working Group on Inflammatory Markers in CF: J. Efthimiou (London, UK); M. Rosewich, U. Zissler and O. Eickmeier (Frankfurt, Germany); E. Rietschel (Köln, Germany); I. Durieu (Lyon, France); R. Chiron (Montpellier, France); A. Magnan (Nantes, France); I. Vachier (Montpellier, France); P-R. Burgel (Paris, France); A. De Alessandri (Genova, Italy); C. Pereira (Lisbon, Portugal); M. van der Eerden (Rotterdam, the Netherlands); J. Beekman and S. Terheggen (Utrecht, the Netherlands)&semi J. Whitehouse ( J. Whitehouse (Birmingham, UK); A. Fogarty and D. Forrester (Nottingham, UK); S. Elborn and D. Downey (Belfast, UK); and J. Davies (London, UK). Other contributors were M. Konstan and J. Hilliard (Cleveland, OH, USA); S. Davis (Chapel Hill, NC, USA), C. Wainwright (Brisbane, Australia), P. Sly (Perth, Australia), S. Stick (Melbourne, Australia), P. Pohunek (Prague, Czech Republic) and F. Ratjen (Toronto, Canada).
Footnotes
This article has supplementary material available from www.erj.ersjournals.com
Conflict of interest: Disclosures can be found alongside the online version of this article at www.erj.ersjournals.com
- Received January 29, 2013.
- Accepted June 19, 2013.
- ©ERS 2014