Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance

+

MODEL

Respiratory Medicine (2014) xx, 1e11

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/rmed

Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance Melissa J. McDonnell a,1, Hannah R. Jary a,1, Audrey Perry b, James G. MacFarlane a, Katy L.M. Hester a, Therese Small a, Catherine Molyneux b, John D. Perry b, Katherine E. Walton b, Anthony De Soyza a,* a Institute of Cellular Medicine Newcastle University and Adult Bronchiectasis Service, Department of Respiratory Medicine, The Freeman Hospital, Newcastle upon Tyne, UK b Department of Medical Microbiology, The Freeman Hospital, High Heaton, Newcastle upon Tyne, NE7 7DN, England, UK

Received 17 September 2012; accepted 28 July 2014

KEYWORDS Bronchiectasis; Exacerbation; Longitudinal; Microbiology; Pseudomonas aeruginosa

Summary Background: The hallmark of non-cystic fibrosis bronchiectasis is recurrent bronchial infection, yet there are significant gaps in our understanding of pathogen persistence, resistance and exacerbation frequencies. Pseudomonas aeruginosa is a key pathogen thought to be a marker of disease severity and progression, yet little is known if the infection risk is seen in those with milder disease or if there is any potential for eradication. These data are important in determining risk stratification and follow up. Methods and patient cohort: A retrospective review of consecutive adult patients attending a specialist UK bronchiectasis clinic over a two-year recruitment period between July 2007 and June 2009 was performed. Analysis of our primary outcome, longitudinal microbiological status, was recorded based on routine clinical follow-up through to data capture point or date of death. Patients were stratified by lung function and infecting organism. Results: 155 patients (mean (SD) age 62.2 (12.4) years; 60.1% female) were identified from clinic records with microbiological data for a median (IQR) follow up duration of 46 (35e62) months. Baseline mean FEV1% predicted was 60.6% (24.8) with mean exacerbation frequency

* Corresponding author. Tel.: þ44 191 2137468; fax: þ44 1912231099. E-mail address: [email protected] (A. De Soyza). 1

Joint first authorship.

http://dx.doi.org/10.1016/j.rmed.2014.07.021 0954-6111/ª 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

2

M.J. McDonnell et al. of 4.42/year; 73.6% reported 3 or more exacerbations/year. Haemophilus influenzae was isolated in 90 (58.1%) patients and P. aeruginosa in 78 (50.3%) patients with persistent infection in 51 (56.7%) H. influenzae and 47 (60.3%) P. aeruginosa, respectively. Of the P. aeruginosa colonised patients, 16 (34%) became culture negative on follow-up with a mean of 5.2 negative sputum cultures/patient. P. aeruginosa was isolated from 5 out of 39 patients (12.8%) with minimal airflow limitation as compared to 18 out of 38 patients (47.4%) with severe airflow limitation. Although hospital admissions were significantly higher in the P. aeruginosa infected group (1.3 vs. 0.7 admissions per annum, p Z 0.035), overall exacerbation rates were the same (4.6 vs. 4.3, p Z 0.58). Independent predictors of P. aeruginosa colonisation were low FEV1% predicted (OR 2.46; 95% CI 1.27e4.77) and polymicrobial colonisation (OR 4.07; 95% CI 1.56 e10.58). 17 (11%) patients were infected with multi-resistant strains; however, none were pan-resistant. Conclusions: P. aeruginosa is associated with greater persistent infection rates and more hospital admissions than H. influenzae. Exacerbation rates, however, were similar; therefore H. influenzae causes significant out-patient morbidity. P. aeruginosa infection occurs across all strata of lung function impairment but is infrequently multi-resistant in bronchiectasis. Careful microbiology follow up is required even in those with well-preserved lung function. ª 2014 Elsevier Ltd. All rights reserved.

Introduction Non-cystic fibrosis (CF) bronchiectasis is characterised by irreversibly damaged and dilated bronchi with impaired mucociliary clearance that leads to recurrent bacterial infections. Frequent exacerbations are a significant cause of morbidity and mortality and may contribute to substantial socioeconomic costs manifest by increased hospitalisations and healthcare utilisation [1,2]. Recent European evidence suggests a steady increase in bronchiectasis-associated hospitalisations and subsequent mortality, particularly in older patients, females and those with associated chronic obstructive pulmonary disease (COPD) [3]. UK studies have demonstrated a 29% death rate in bronchiectasis patients over 13 years, twice that predicted in age matched patients [4]. The British Thoracic Society (BTS) bronchiectasis guidelines emphasise microbiology assessment to guide antimicrobial therapy [5]. Prior studies demonstrate a predominance of non-typeable Haemophilus influenzae and Pseudomonas aeruginosa, as two significant pathogens in adult bronchiectasis [6]. P. aeruginosa persistence has been shown to be associated with poorer lung function, more extensive disease and poorer quality of life [7,8]. It has traditionally been suggested that only patients with severely impaired lung function are considered at risk for P. aeruginosa infections [9]. Few studies have assessed the association between persistently isolated pathogens and exacerbations or the potential impact of P. aeruginosa on lung function impairment [7]. This is important as there is an increasing need to define which patients with bronchiectasis need specialist follow up and which can be discharged from specialist clinics for primary care follow up only. We therefore assessed the longitudinal microbiological profile in adult patients with non-cystic fibrosis bronchiectasis in a UK population and investigated associations

between markers of functional impairment, exacerbation frequency and microbiological status.

Methods Study design A retrospective observational cohort analysis of consecutive adult patients attending a specialist UK bronchiectasis clinic between July 2007 and June 2009 in the North East of England was performed. Rigorous aetiological screening investigations were undertaken to establish disease phenotype according to a standardised protocol [10]. The diagnosis of bronchiectasis was established in all cases by high-resolution computed tomography (HRCT). Retrospective data collection and analysis of our primary outcome, longitudinal microbiological status, was recorded based on routine clinical follow-up from initial recruitment visit through to data capture point (October 2013) or date of death. Data pertaining to secondary outcomes, exacerbation frequency and hospital admission rates, was collected over a pre-defined one year follow-up period (July 2010eJune 2011) by means of telephone or outpatient interview assessment. Patients had their individual case notes reviewed for collection of demographic details, aetiological investigations, smoking history, Medical Research Council (MRC) dyspnoea score, baseline and follow up lung function, nebulised antibiotic treatment and chronic macrolide treatment.

Patients and setting The inclusion criterion was a clinico-radiological diagnosis of non-CF bronchiectasis on HRCT. Consecutive patients attending clinic during the recruitment period were included to ensure a fully representative cohort

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

Non cystic fibrosis bronchiectasis irrespective of aetiological cause of bronchiectasis. Patients were excluded if microbiology data was unavailable during subsequent follow-up.

Aetiology Aetiological designation was based upon the protocol suggested by Pasteur et al. and others [5,6,10]. Cystic Fibrosis genotyping and/or sweat testing was undertaken as suggested by national British Thoracic Society (BTS) bronchiectasis guidelines [5].

Sputum microbiology Results were retrospectively reviewed for each patient from recruitment visit through to data capture point or date of death. Where available, prior microbiology was also recorded for separate sub-analysis of microbiology samples since diagnosis. All microbiology samples were processed in a UK Clinical Pathology Accreditation (CPA) accredited laboratory to routine diagnostic standards. Sensitivity testing was carried out using the agar disc diffusion method according to the prevailing British Society for Antimicrobial Chemotherapy (BSAC) methods from 2007 to June 2009, with the methods of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) having been established in our institution in July 2010 [11,12]. Isolates were tested against multiple antimicrobial agents including amikacin, ceftazidime, ciprofloxacin, colistin, gentamicin, meropenem, piperacillintazobactam, ticarcillin-clavulanic acid and tobramycin. Sputum was routinely cultured for Mycobacterium in all new patients, and when clinical features suggested mycobacterial disease (manifest by nodular changes on HRCT scan, progressive decline in lung function or significant weight loss and recurrent exacerbations unresponsive to standard antibiotics). Mycobacterial cultures were also sent annually as part of our routine work-up. ‘Colonisation’ was defined as the same bacterial species being isolated on 2 or more occasions, at least 3 months apart, within a 1 year period [6]. ‘Polymicrobial colonisation’ was defined as colonisation by  2 pathogens on follow-up. ‘Isolation’ was defined as the presence of the pathogen on a single occasion without colonisation [6].

Resistance definitions We studied P. aeruginosa resistance by reviewing all available antibiogram data. Definitions of multidrug resistance and pan resistance followed those of Falagas et al. [13] Isolates resistant to at least three drugs from different antimicrobial categories, including aminoglycosides, antipseudomonal penicillins, carbapenems, cephalosporins and quinolones, were classified as multidrug resistant [13]. An isolate was defined as pan-resistant only if it was resistant to agents from all anti-pseudomonal classes of antimicrobial agents, i.e. anti-pseudomonal penicillins, cephalosporins, carbapenems, monobactams, quinolones, aminoglycosides and polymyxins. [13.]

3

Lung function testing Lung function was measured according to ERS/ATS standards at each clinic visit [14,15]. The most recent FEV1 % recorded was used for analysis with a predicted value of 40% or less, 41e79% and 80% or more, classified as severe, moderate or minimal airway limitation, respectively.

Infective exacerbation and admission frequency Patients’ self-reported infective exacerbation frequencies and the number of hospital admissions for bronchiectasis over a pre-defined 12 month period (July 2010eJuly 2011) were noted. Data on exacerbations and hospitalisations was obtained from patient histories on follow-up and verified against electronic and hard copy hospital records and primary care prescribing records where possible. Individual case notes were reviewed to exclude any nonerespiratory related hospital admissions. The patients’ primary care teams were contacted for confirmation of the number of antibiotic treated exacerbations over the specified time period.

Relationship between P. aeruginosa colonisation and disease severity The Bronchiectasis Severity Index (BSI) consists of 8 commonly measured clinical parameters reflecting age, body mass index (BMI), FEV1% predicted, MRC dyspnoea score, exacerbation frequency, prior hospitalisation, chronic bacterial colonisation, and the number of lobes and/or presence of cystic bronchiectasis on HRCT, with subsequent classification into low, intermediate and high risk groups [16]. The BSI was calculated for patients with and without P. aeruginosa colonisation with the exception of radiology scoring in this patient cohort.

Statistical analysis Analyses were conducted using SPSS (V21) and Minitab (V16). Simple descriptive statistics of mean and standard deviation were used for continuous parametric data, median and interquartile range for continuous non-parametric data, and frequencies and percentages for categorical data. Subgroup comparisons were performed using the unpaired t-test, Mann-Witney U-Test or Chi-squared test, depending on data distribution. Univariate correlations between outcome measures were assessed using Pearson and Spearman Rank correlation according to data distribution. A p-value < 0.05 was considered to be statistically significant. Logistic regression models were fitted to determine factors associated with P. aeruginosa colonisation. Variables with large proportions of missing data were excluded from analysis due to convergence problems in the logistic models. The full binary logistic regression model with presenting characteristics positive on univariate analysis was performed to choose the most parsimonious set of predictors for the outcome of interest. The odds ratio (OR),

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

4 95% confidence interval and p-value were computed for each of the presenting predictors. The reporting of this observational study conforms to the recommendations of STROBE. [17.]

Results Patient population 174 patients were identified between July 2007 and June 2009; 19 patients were excluded due to lack of microbiology data throughout the study period and follow up (Fig. 1). Of the 155 patients analysed, mean age was 61.3 (13.9), with 108 (62.1%) below 65 years of age. For gender 94 (60.1%) were female and 61 (39.9%) male. 40 (25.8%) patients had never smoked, 51 (32.9%) were previous smokers, 10 (6.5%) were current smokers, 17 (10.9%) were abstinent but a history of prior smoking was not recorded in these patients. The majority of cases were attributed to idiopathic and postinfective (58.0%), with COPD-associated bronchiectasis in 10.3% and asthma and immune deficiency in 6.5% respectively (Table 1). During a mean follow up period of 58.6 (12.9) months, 22 patients (14.2%) died.

Microbiology Of the 155 patients with longitudinal sputum sample cultures, a total of 2287 sputum cultures were analysed over

M.J. McDonnell et al. a median (IQR) follow up duration of 46 (35e62) months. This equates to a mean (SD) of 14.8 (11.7) sputum cultures per patient. Over 20 different species of bacteria or fungi were isolated from these 155 patients (Table 2). The two most commonly identified bacterial species were H. influenzae, in 89 patients (57.4%) and P. aeruginosa in 76 (49.0%) patients (Table 2). 9 (5.8%) individuals had no pathogenic organisms isolated from any of their sputum samples during follow up (mean number of 3.9 (3.4) sputum samples). Less common organisms seen in less than 1% of patients not listed in Table 2 included diverse bacteria and fungi; Pasteurella multicoda, Ralstonia picketti, Aeromonas, Elizabethkingia meningosepta, Bergeyella zoolelcum, Sphingobacterium spiritovorum, Delftia acidovorans and Extended-Spectrum Beta-lactamase (ESBL).

Pseudomonas aeruginosa 76 patients (49.0%) of patients had P. aeruginosa cultured from their sputum on at least one occasion (total 608 cultures throughout study follow up). 47 of these (30.3% of the total cohort) fulfilled colonisation criteria (Fig. 2). Two thirds of these patients, 31 (66.0%) remained colonised at the data capture point. 16 (34%) patients subsequently became culture negative for this pathogen on follow-up with a mean of 5.5 cultures without further P. aeruginosa isolates (median 4 (2e8)).

Figure 1 Overview of patient inclusion and exclusion criteria. This shows the patient flow chart including how many patients were analysed for each key aspect of the study. As sputum microbiology was the primary outcome of interest, only patients with sputum culture results on follow up were included in the study. The above flow chart depicts the number of patients included in each analysis and depicts reasons for patient exclusion. 90 patients had pre-existing microbiology results available which were reviewed in a separate sub-analysis of longer follow-up duration. Deceased patients up to data capture point for assessment of exacerbations and hospital admissions were excluded from these analyses due to incomplete data.

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

Non cystic fibrosis bronchiectasis Table 1

5

Aetiological causes of Bronchiectasis.

Aetiology Idiopathic Post e infectious Chronic obstructive pulmonary disease Immunodeficiency (including common variable immunodeficiency, specific antibody deficiency) Asthma Rheumatoid arthritis Other connective tissue disease Inflammatory bowel disease Pink’s disease (Mercury Poisoning) Reflux disease/Hiatus Hernia Wegener’s granulomatosis Young’s syndrome Familial Primary ciliary dyskinesia ABPA Lymphoma Total

n

Study population (%)

57 33 16

36.7 21.3 10.3

10

6.5

10 7 1 4 4 3 2 1 2 1 2 2 155

6.5 4.5 0.7 2.6 2.6 1.9 1.3 1.3 1.3 0.7 1.3 1.3 100

Patients with one or more isolates of P. aeruginosa throughout the study period had significantly worse FEV1% predicted and increased hospital admission rates (p < 0.001, Spearman test) compared to those without positivity for P. aeruginosa. There was no statistically significantly increased MRC dyspnoea scores nor increased exacerbation rates (p Z 0.083 and p Z 0.060, respectively) in these patients. When stratified according to P. aeruginosa colonisation, these patients were noted to be older (mean 65.1 vs. 60.8 years, p Z 0.047), have a lower FEV1 % predicted (49.8 vs. 65.4, p < 0.001), more hospital admissions (1.3 vs. 0.7 admissions per annum, p Z 0.035) and a higher incidence of polymicrobial colonisation (p < 0.001) than their non- P. aeruginosa colonised counterparts (Table 3). The mean BSI in P. aeruginosa colonised was 12.6 (4.0) compared with 8.4 (4.2) in P. aeruginosa negative patients (p < 0.0001) (Fig. 3). Negative correlations between P. aeruginosa colonisation with FEV1% predicted (p < 0.001) and H. influenzae colonisation (p < 0.001), and positive correlations with hospital admissions (p Z 0.010) and polymicrobial colonisation (p Z 0.001) were noted on univariate analysis (Spearman test). Trends towards associations with MRC dyspnoea score and total number of micro-organisms isolated were also noted at p Z 0.086 and 0.056 respectively. Binary logistic regression analysis of variables positive on univariate analysis showed low FEV1% predicted to be the main outcome associated with P. aeruginosa colonisation (p Z 0.001, OR 2.46 (95% CI 1.27e4.77)) followed by polymicrobial colonisation (p Z 0.004, OR 4.07 (95% CI 1.56e10.58)), suggesting that patients with P. aeruginosa colonisation are up to 4 times more likely to have reduced FEV1% and up to 10 times more likely to have polymicrobial colonisation (Table 4(a)).

Table 2 Bacterial and Fungal species isolated from patients. The sputum culture results were categorised by organism cultured. If more than one species was isolated they were counted for each species. Isolates with less than 1% frequency are not listed. Species

Number of individual patients culture positive for this organism (% cohort)

1. 2. 3. 4. 5.

89 76 51 38 31

(57.4) (49.0) (32.9) (24.5) (20.0)

20 16 25 15 22 17 10 8 8 11 5 4

(12.9) (10.3) (16.3) (9.7) (14.2) (11.0) (6.5) (5.2) (5.2) (7.1) (3.2) (2.6)

4 3 2 2 8

(2.6) (1.9) (1.3) (1.3) (5.2)

Haemophilus influenzae Pseudomonas aeruginosa Streptococcus pneumoniae Moraxella catarrhalis Methicillin sensitive Staphylococcus aureus (MSSA) 6. Escherichia coli 7. Aspergillus fumigatus 8. Stenotrophomonas maltophilia 9. Serratia spp. 10. Klebsiella spp. 11. Candida 12. Acinetobacter species 13. Enterobacter cloacae 14. Proteus mirabilis 15. Achromobacter xylosoxidans 16. Mycobacterium 17. Meticillin resistant Staphylococcus aureus (MRSA) 18. Morganella morganinii 19. Citrobacter koseri 20. Comamonas testosteroni 21. Pseudomonas otitidis 22. Others

Other species include: Pasteurella multocida, Ralstonia picketti, Aeromonas spp, Elizabethkingia meningoseptica, Bergeyella zoohelcum, Sphingobacterium spiritivorum, Delftia acidovorans and an Extended-Spectrum Beta-lactamase (ESBL) Gram negative.

Inclusion of prior data Prior microbiology data was available in 90 (58.1%) patients over a mean follow-up period of 110.9 (52.9) months, i.e. approximately 9 years. Of these, 60 patients had tested positive for P. aeruginosa, either in isolation or colonisation. Review of this data demonstrated that 27 (45.0%) of these patients had no further positive P. aeruginosa cultures throughout the study period follow up. Using all the available data to assess correlations between P. aeruginosa isolates with other potential predictors of outcome, univariate analyses showed negative correlations between P. aeruginosa isolation with FEV1% predicted (p < 0.001) and H. influenzae colonisation (p < 0.001), and positive correlations with hospital admissions (p Z 0.021), polymicrobial colonisation (p Z 0.006), total number of micro-organisms isolated (p Z 0.028) and mortality (p Z 0.038). Trends towards associations with MRC dyspnoea score and Fatigue Impact Score (FIS) were also noted at p Z 0.061 and 0.057 respectively. Binary logistic regression analysis of variables positive on univariate analysis showed low FEV1% predicted

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

6

M.J. McDonnell et al.

Figure 3 The Bronchiectasis Severity Index (BSI) in a cohort of non-cystic fibrosis bronchiectasis patients with (PAþ) and without (PA) Pseudomonas aeruginosa colonisation. Bars indicate median values. Box plots indicate median, and 25the75th percentiles; whiskers indicate 5the95th percentile. Figure 2 Pseudomonas isolation and persistence rates throughout study follow up.

to be the main outcome associated with P. aeruginosa isolation (p Z 0.005, OR 2.29 (95% CI 1.28e4.09)) followed by polymicrobial colonisation (p Z 0.033, OR 2.78 (95% CI 1.09e7.13) and mortality (p Z 0.046, OR 3.55 (95% CI 1.15e12.35)), suggesting that patients with P. aeruginosa isolation are up to 4 times more likely to have reduced FEV1%, up to 7 times more likely to be colonised by multiple organisms on follow up, and have an up to 12 times higher risk of mortality than those who have never cultured P. aeruginosa (Table 4(b)).

Table 3

Pseudomonas resistance patterns Susceptibility testing was performed on 908 strains of P. aeruginosa isolated throughout the stated study period. Slightly different panels of antimicrobial agents have been tested over the years, reflecting changes in antimicrobial regimens favoured over time (Table 5). We found relatively low levels of resistance in vitro with the highest resistance rate to ciprofloxacin in 16% of isolates; no isolates were resistant to colistin. None were classified as pan-resistant but 77 isolates from 17 patients were multidrug resistant; the majority of these were resistant to only 2 organisms with a small subset of patients demonstrating resistance to 3 antibiotics. Persistence of a multi-resistant phenotype

Non-cystic fibrosis bronchiectasis cohort stratified by presence of Pseudomonas aeruginosa colonisation.

n

Female gender, n (%) Age in years, mean (SD) Smokers (current or ex-smokers), n (%) (n Z 112) MRCD 4-5, n (%) (n Z 152) FEV1% predicted (n Z 146), mean (SD) Lung function severity, n (%) Mild Moderate Severe No. exacerbations, mean (SD) (n Z 143) 3 exacerbations, n (%) No. patients requiring hospital admission, n (%) (n Z 143) Polymicrobial colonisation, n (%) No. microbes isolated, mean (SD) No. deaths, n (%) Follow-up, months, median (IQR)

PAþ

PA-

Total

47

108

155

24 (51.1) 65.1 (8.9) 17/33 (51.5) 17 (36.2) 49.8 (20.9) n Z 46 5 (10.9) 23 (48.9) 18 (38.3) 4.6 (2.6) 34 (72.3) 21 (44.6) 21 (44.7) 3.5 (1.7) 9 (19.1) 46 (40e65)

70 (64.8) 60.8 (13.5) 51/79 (64.5) 29 (26.9) 65.4 (24.9) n Z 100 29 (29.0) 52 (52.0) 19 (19.0) 4.3 (3.3) 75 (69.4) 32 (29.6) 18 (16.7) 3.0 (2.0) 13 (12.0) 47 (31e59)

94 (60.6) 62.1 (12.4) 68 (607) 46 (29.7) 60.6 (24.7) n Z 146 34 (22.8) 75 (51.4) 37 (25.3) 4.4 (3.1) 109 (73.6) 53 (37.1) 39 (25.2) 3.1 (1.9) 22 (14.2) 46 (35e62)

p-value

0.107 0.047 0.898 0.236 <0.001 <0.001

0.581 0.836 0.035 <0.001 0.317 0.316 0.936

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

Non cystic fibrosis bronchiectasis Table 4

7

Results of logistic regression analyses for factors associated with the primary Pseudomonas aeruginosa.

Presenting characteristics

Radiological progression p-value

OR

95% CI

Significance

(a) Results of binary logistic regression analysis of factors associated with the primary Pseudomonas aeruginosa colonisation FEV1% predicted 0.017 2.00 1.27e4.77 S Hospital admissions 0.177 1.20 0.92e1.55 NS No. isolated microorganisms 0.353 0.89 0.70e1.13 NS Polymicrobial colonisation 0.004 4.07 1.56e10.58 S (b) Results of binary logistic regression analysis of factors associated with the primary Pseudomonas aeruginosa isolation 0.005 2.29 1.28e4.09 S FEV1% predicted Hospital admissions 0.760 1.04 0.81e1.34 NS No. isolated microorganisms 0.634 1.05 0.85e1.31 NS Polymicrobial colonisation 0.033 2.78 1.09e7.13 S Mortality 0.046 3.55 1.15e12.35 S

occurred very infrequently. In 52% of patients colonised with Pseudomonas, the isolates were sensitive to all agents tested (203 isolates).

term azithromycin therapy grew an NTM; Mycobacterium simiae was isolated on four occasions from this individual.

Polymicrobial infection Haemophilus influenzae Although isolated from 89 (57.4%) patients, only 51 out of the 155 patients (32.9%) met H. influenzae colonisation criteria, 14 of whom were still colonised at the data capture point (27.5%), suggesting that 37 patients (74.5%) had cleared this organism as compared to 34.0% who cleared P. aeruginosa colonisation. 33 of 90 (36.7%) patients where data was available had previous single or intermittent isolation with H. influenzae.

Non-tuberculous mycobacteria (NTM) NTM were isolated from 6 patients, three of whom had no record of azithromycin treatment. These comprised Mycobacterium avium complex (MAC), Mycobacterium kansasii, Mycobacterium fortuitum, Mycobacterium gordonae, Mycobacterium simiae, and Mycobacterim chelonae. MAC was isolated on three occasions from one patient before azithromycin was started. Only one patient taking long

Table 5 isolates.

Resistance patterns in Pseudomonas aeruginosa

Antibiotic

Number of isolates tested

Ciprofloxacin Ticarcillin- clavulanic acid Gentamicin Ceftazidime Piperacillin-tazobactam Amikacin Tobramycin Colistin

n n n n n n n n

Z Z Z Z Z Z Z Z

975 746 975 972 972 821 816 972

Number of isolates resistant (%) 129 63 69 74 68 36 31 0

(13) (8) (7) (7) (7) (4) (4) (0)

The isolation of multiple species during the clinical course was common: in 155 patients (94.2%) had at least one positive culture result; the mean number of organisms isolated longitudinally was 3.1 (2.0). Polymicrobial colonisation was noted in 39 (25.2%) throughout the study follow up.

Airflow limitation severity 146 (94.2%) patients had repeat spirometry during the study follow up. Mean FEV1 % predicted was 65.4 (24.9); 34 (22.8%) patients had minimal airflow limitation, 75 (51.4%) had moderate and 37 (25.3%) had severe airflow limitation respectively. In the 137 patients where both FEV and FVC data were recorded, 89 (64.9%) had obstructive spirometry. FEV1% predicted was negatively correlated with smoking (p Z 0.005), MRC dyspnoea score (p < 0.001), hospital admission rate (p < 0.001), mortality (p < 0.001), P. aeruginosa colonisation (p Z 0.001), and FIS (p Z 0.002, Spearman test). Patients with severe airflow limitation were more likely to have had P. aeruginosa detected in their sputum than those with minimal airflow limitation (p < 0.001). P. aeruginosa was isolated from 5 out of 39 patients (12.8%) with minimal airflow limitation as compared to 18 out of 38 patients (47.4%) with severe airflow limitation.

Functional limitation The median (IQR) MRC dyspnoea score was 3 (2e4) (n Z 152) patients. Significant positive correlations between MRC dyspnoea score and smoking (p Z 0.009), exacerbation frequency (p Z 0.022), hospital admission rates (p < 0.001), and mortality (p < 0.001) but not P. aeruginosa colonisation (p Z 0.061) were noted along with

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

8 a significant negative correlation between MRC dyspnoea scores and FEV1% predicted (p < 0.001, Spearman test).

Infective exacerbation and admission frequency Mean self-reported 12 month exacerbation frequency in the 143 patients where data was available at this time period was 4.42 (3.1, range 0e12). 9 (6.3%) of patients reported no infective exacerbations over this 12 month period. Of the 134 patients who had at least one exacerbation, the mean number of exacerbations over the 12 month period was 4.71 (3.0). 109 patients (73.6%) reported having 3 or more exacerbations with 53 patients out of 143 (35.1%) reporting a bronchiectasis related admission. Self-reported exacerbation and hospital admission data was unavailable for 12 (7.7%) due to mortality. Mean 12 month hospital admission frequency in 143 patients was 0.9 (1.6, range 0e10). For the 53 patients who had at least one admission, the mean number of admissions over the previous 12 months was 2.51 (1.6). The selfreported total number of admissions for these 143 patients was 131. During this time, we identified 119 (90.8%) admissions due to bronchiectasis from within our cohort. Significant correlations were noted between selfreported exacerbation frequency and admission frequency (p < 0.001), age (p Z 0.009), total number of microorganisms isolated (p Z 0.021), MRC dyspnoea score (0.022) and FIS (0.049). There were also significant correlations between hospital admission frequency and FEV1% predicted (p < 0.001), MRC dyspnoea score (p < 0.001), P. aeruginosa colonisation (p Z 0.010), total number of microorganisms isolated (p Z 0.002), FIS (p Z 0.003) and mortality (p Z 0.017, Spearman correlation). In contrast, selfreported exacerbation frequency was not significantly correlated with FEV1% predicted. At data capture, 49 and 33 of (70.0% and 47.1% of n Z 70 patients where data was available) were noted to be taking or have previously been prescribed long term azithromycin (250 mg thrice weekly) or nebulised gentamicin (2 mega units twice daily) with only 5 (7.1%) prescribed both. All patients taking nebulised gentamicin were P. aeruginosa colonised. Discontinuation of these medications was attributed primarily to intolerance or lack of perceived benefit by the patient or clinician. On univariate analyses, both azithromycin and nebulised gentamicin were significantly correlated with P. aeruginosa colonisation (p Z 0.027 and p < 0.001, respectively). Other associations noted with azithromycin included polymicrobial colonisation (p Z 0.030), and with nebulised gentamicin, an increased rate of hospital admissions (p Z 0.014). These variables were excluded from the multivariate model to prevent convergence due to a significant proportion of missing data.

Discussion Bronchiectasis is a recognised important cause of morbidity and mortality in the current era with prevalence estimates suggesting it is approximately three times more common than cystic fibrosis in the U.S [18]. Major healthcare costs relate to infective exacerbations that mandate hospital

M.J. McDonnell et al. admission. This cohort is one of the largest reported to date with extensive longitudinal microbiological follow up allowing novel insights into our understanding of P. aeruginosa persistence and resistance and the potential effects on markers of disease severity and progression. Our first key finding was the high period prevalence of P. aeruginosa isolation (47%) and colonisation (30%). Rates of 31% for P. aeruginosa isolation and 24% for colonisation were reported in point prevalence studies by Pasteur et al. in 150 patients referred to a specialist centre with varying chronic colonisation rates in other UK series of 9%e22 [4,6,8] Other series report lower rates of P. aeruginosa infection with 13% in Spain and 12% in Australia, whilst a Hong Kong and smaller Korean series have reported similar rates of 33% and 31%, respectively [19e22]. More severe underlying bronchiectasis seems an unlikely casual factor for our higher rates of P. aeruginosa as the FEV1% and MRC dyspnoea scale characteristics in our cohort were similar to those previously reported [20]. Our rates of P. aeruginosa may therefore reflect the longer period of follow-up, or cross-infection, as has been reported in CF clinics [23,24]. As yet, we have not instituted segregation in outpatients with a Pseudomonas-specific clinic. For in-patients with non-cystic fibrosis bronchiectasis, we prefer to use cubicle based management, but patients are managed in 6-bedded bays when cubicles are unavailable. Systematic crossinfection studies in this area have not been conducted and are certainly needed. [25.] This is the first study to explore modifiable and nonmodifiable factors associated with P. aeruginosa isolation and colonisation in a clinical cohort of patients with stable non-cystic fibrosis (CF) bronchiectasis. Independent factors associated with P. aeruginosa colonisation were: low FEV1% predicted (OR 2.46 (95% CI 1.27e4.77) and polymicrobial colonisation (OR 4.07 (95% CI 1.56e10.58), suggesting that patients with P. aeruginosa colonisation are up to 4 times more likely to have reduced FEV1% and are up to 10 times more likely to be colonised by multiple organisms throughout follow-up. Independent factors associated with P. aeruginosa isolation on extended follow-up again included low FEV1% (OR 2.29 (95% CI 1.28e4.09) and polymicrobial colonisation (OR 2.78 (95% CI 1.09e7.13), but also mortality (OR 3.55 (95% CI 1.15e12.35), suggesting that patients who have cultured P. aeruginosa have a higher risk of mortality than those who have always been culturenegative for this organism. We also report for the first time an extensive study of resistance rates in P. aeruginosa isolates from bronchiectasis. No isolates were pan-resistant with multi-resistant strains relatively uncommon. The highest resistance rate for P. aeruginosa isolates was to ciprofloxacin, perhaps reflecting the widespread use of this as an oral antibiotic, often as a single agent, in both primary and secondary care settings local to our centre. Although there was a change in the laboratory method of sensitivity testing during the period of review, BSAC and EUCAST methods are harmonised and should therefore generate equivalent data [11,12]. It is accepted that variation in resistance patterns may occur between strains isolated from the same patient and that variation in reporting may also be a factor [26,27]. These factors are however unlikely to explain the much lower resistance rates herein as compared to the major

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

Non cystic fibrosis bronchiectasis challenges seen in CF associated P. aeruginosa infection [28]. These data support the widely held contention that CF and non CF-bronchiectasis have different pathobiology and should be viewed as distinct conditions. Importantly, we identified that P. aeruginosa infection occurred in w10% of patients with normal FEV1 values and w50% of patients with moderate airflow obstruction, suggesting that this may be an important potential pathogen across all stages of airflow limitation. Our data add to a 1996 UK study where the association between P. aeruginosa infection and more advanced bronchiectasis led the authors to suggest that P. aeruginosa was a marker of disease severity [29]. This data corroborates that of the previous study and adds further information relating to the longitudinal effects of P. aeruginosa infection and colonisation. Our data highlights that ongoing sputum surveillance of bronchiectasis should be considered in all patients irrespective of airflow limitation severity. This mandates rigorous surveillance protocols with primary care for patients regarded as “too well” for specialist follow up [30]. We also demonstrate that 34% of P. aeruginosa colonised patients subsequently became culture negative for this pathogen. It is hoped that P. aeruginosa colonisation has been truly eradicated although we cannot exclude biofilm mode of growth in vivo preventing successful in vitro culture. A further 45% of patients, with previous positive isolates or colonisation with P. aeruginosa prior to the study period, cultured no further P. aeruginosa throughout the duration of follow-up. Disappointingly, as per prior studies, we found that nebulised antibiotic therapy was often stopped due to intolerance and/or perceived lack of efficacy [31]. Hence there still significant challenges in designing eradication strategies and, where this fails, improved therapies for longer term management are needed [32]. The observed greater rate of hospital admissions and poorer lung function in patients from whom P. aeruginosa was isolated underlines the importance of P. aeruginosa as a pathogen. Notably the overall exacerbation rates were similar to those without P. aeruginosa confirming prior data that P. aeruginosa infection was associated with more hospital admissions than other pathogens but not with greater exacerbation rates [8]. Secondly we noted many (milder non-hospitalised) exacerbations occurring in patients of working age without P. aeruginosa infection suggesting exacerbations may have an important societal cost through lost productivity. To date there are few comprehensive studies of the economic impact of bronchiectasis with most available data limited to direct healthcare costs [2]. New therapies to reduce exacerbations including nebulised and inhaled antibiotics are in development [32]; focussing trials only on the P. aeruginosa infected population is to the detriment of a large sector of the current patient population. Notably 72% of patients had 3 or more exacerbations over the 12 month period of data collection (a threshold suggested in guidelines for intensifying treatment) [5]. There are surprisingly scarce data on the exacerbation frequency in bronchiectasis outwith intervention trials; these range from yearly exacerbation rates of 1.2e7.0 per year [7,33,34]. A study from Spain noted a mean exacerbation frequency of 2.4 vs. 1.5 in severe vs. mild disease; in

9 this study, exacerbation frequency and P. aeruginosa persistent infection were both independently associated with accelerated lung function decline [19]. Our data and that reported recently from 4 other centres confirm there are high exacerbation rates and these affect the nonPseudomonas population [16]. Collectively these data suggest an exacerbation frequency equivalent to or much higher than that reported for most COPD patients and mirroring that of “frequent COPD exacerbators” [35,36]. Limitations of our study reflect the retrospective design and the heterogeneity of the study population. The cohort is a real life cohort and whilst the aetiologies frequencies are similar to prior UK series, our findings may not be more widely generalisable. Further limitations reflect variation in reporting of sensitivity patterns that can be observed between isolates from the same patient. Furthermore we have not exhaustively studied the medical records comparing sputum microbiology categorising data by “stable state” and by “exacerbation”, nor have we prospectively collected viral samples which may contribute to exacerbations. Indeed the current lack of a consensus definition on what contributes an exacerbation in bronchiectasis is problematic. We captured self-reported exacerbations arguing that patients’ use of acute antibiotic therapy was the primary area of interest to define the current symptom burden. A further limitation may include the recall bias for self-reported exacerbations. Our rates of P. aeruginosa are higher than previously reported and are likely due to the longitudinal follow up design. They may also indicate an unexpected bias in the referral patterns to our specialist bronchiectasis clinics. Notably Pasteur et al. reported rates of P. aeruginosa isolation of 31% in a UK a specialist referral centre [6]. Longitudinal lung function data would be very useful to further explore the relationship between P. aeruginosa infection and accelerated lung function decline. We are unable to report these data at present as they are not centralised at our centre. Other limitations include that microbiological handling of specimens whilst to prevailing national clinical standards would not be as consistent as with a prospective study and variations. Future prospective studies that include HRCT scores and longitudinal lung function to define the frequency of and risk factors for Pseudomonas acquisition and persistence will be useful. In summary we note the importance of P. aeruginosa as a pathogen in bronchiectasis with a high frequency of isolation and colonisation in patients with severe airflow limitation. It is not infrequent, however, in those with milder airflow limitation and careful ongoing screening is therefore required. Effective surveillance programs for P. aeruginosa, working across primary and secondary care need to be developed, tested and rigorously followed. Multicentre collaboration is required for the development and testing of novel P. aeruginosa detection methods, trials of eradication therapies and home intravenous antibiotic interventions.

Authors contributions ADS, MJM, HJ, KLMH, JMacF, AN, JP and KW designed the study and contributed to analysis and draft manuscripts.

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

10 MJM, HJ and ADS contributed to data extraction, analysis and final corrections of manuscript.

Conflicts of interest The authors report no known conflicts of interest relevant to this manuscript. Dr De Soyza has been on advisory boards or participated in clinical trials of inhaled antibiotics or anti-inflammatory agents in bronchiectasis for Aradigm, AstraZeneca, Bayer, Forest Labs, Gilead. He has received travel bursaries to attend symposia from Chiesi, GSK and Boehringer Ingelheim.

Acknowledgements ADS is a Higher Education Funding Council for England Senior lecturer and has received funding from the National Institute for Health Research (NIHR) Biomedical Research Centre for Aging Newcastle. MJM acknowledges funding from the European Respiratory Society/European Lung Foundation research fellowship and Health Research Board, Ireland towards research in bronchiectasis. KLMH acknowledges Northumbria Tyne and Wear NIHR Comprehensive Local Research Network funding for the Clinical Research associate post. We acknowledge the help of Mr P McAlinden and Ms K Sands for assisting with data collection. We also acknowledge the advice and support of the members of the Medical Research Council Bronch-UK project a funded multi-site partnership, registry and biobank in Bronchiectasis.

References [1] Seitz AE, Olivier KN, Adjemian J, Holland SM, Prevots DR. Trends in bronchiectasis among medicare beneficiaries in the United States, 2000e2007. Chest 2012. [2] Seitz AE, Olivier KN, Steiner CA, Montes de Oca R, Holland SM, Prevots DR. Trends and burden of bronchiectasis-associated hospitalizations in the United States, 1993e2006. Chest 2010 Oct;138(4):944e9. [3] Ringshausen FC, de Roux A, Pletz MW, Ha ¨ma ¨la ¨inen N, Welte T, Rademacher J. Bronchiectasis-associated hospitalizations in Germany, 2005e2011: a population-based study of disease burden and trends. PLoS One 2013 Aug 1;8(8):e71109. http: //dx.doi.org/10.1371/journal.pone.0071109. [4] Loebinger MR, Wells AU, Hansell DM, Chinyanganya N, Devaraj A, Meister M, et al. Mortality in bronchiectasis: a longterm study assessing the factors influencing survival. Eur Respir J 2009;34(4):843e9. [5] Pasteur MC, Bilton D, Hill AT. British Thoracic Society guideline for non-CF bronchiectasis. Thorax 2010;65(Suppl. 1):i1e58. [6] Pasteur MC, Helliwell SM, Houghton SJ, Webb SC, Foweraker JE, Coulden RA, et al. An investigation into causative factors in patients with bronchiectasis. Am J Respir Critical Care Med 2000;162(4 Pt 1):1277e84. [7] Wilson CB, Jones PW, O’Leary CJ, Hansell DM, Cole PJ, Wilson R. Effect of sputum bacteriology on the quality of life of patients with bronchiectasis. Eur Respir J 1997;10(8): 1754e60. [8] Davies G, Wells AU, Doffman S, Watanabe S, Wilson R. The effect of Pseudomonas aeruginosa on pulmonary function in patients with bronchiectasis. Eur Respir J 2006;28(5):974e9.

M.J. McDonnell et al. [9] Eller J, Ede A, Schaberg T, Niederman MS, Mauch H, Lode H. Infective exacerbations of chronic bronchitis: relation between bacteriologic etiology and lung function. Chest 1998; 113:1542e8. [10] Anwar GA, McDonnell MJ, Worthy SA, Bourke SC, Afolabi G, Lordan J, Corris PA, Desoyza A, Middleton P, Ward C, Rutherford RM. Phenotyping adults with non-cystic fibrosis bronchiectasis: a prospective observational cohort study. Respir Med 2013;107(7):1001e7. [11] Anonymous British Society for Antimicrobial Chemotherapy. BSAC methods for antimicrobial susceptibility testing, version 9. British Society for Antimicrobial Chemotherapy; 2010. http://www.bsac.org.uk/SusceptibilityþTesting/ GUIDELINESþStandardizedþDiscþSusceptibilityþTesting þMethod. [12] Anonymous European Committee on Antimicrobial Susceptibility Testing. Antimicrobial susceptibility testing; EUCAST disk diffusion method. European Committee on Antimicrobial Susceptibility Testing; 2012. http://www.eucast.org/ fileadmin/src/media/PDFs/EUCAST_files/Disk_test_ documents/Manual_v_2.1_EUCAST_Disk_Test.pdf. [13] Falagas ME, Koletsi PK, Bliziotis IA. The diversity of definitions of multidrug-resistant (MDR) and pandrug-resistant (PDR) Acinetobacter baumannii and Pseudomonas aeruginosa. J Med Microbiol 2006;55(Pt 12):1619e29. [14] Brusasco V, Crapo R, Viegi G. Coming together: the ATS/ERS consensus on clinical pulmonary function testing. Eur Respir J 2005;26(1):1e2. [15] Miller MR, Crapo R, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A, Enright P, Grinten CPMvd, Gustafsson P, Jensen R, Johnson DC, MacIntyre N, McKay R, Navajas D, Pedersen OF, Pellegrino R, Viegi G, Wanger J. General considerations for lung function testing. Eur Respir J 2005;26(1): 153e61. [16] Chalmers JD, Goeminne P, Aliberti S, McDonnell MJ, Lonni S, Davidson J, Poppelwell L, Salih W, Pesci A, Dupont LJ, Fardon TC, De Soyza A, Hill AT. The bronchiectasis severity index. An international derivation and validation study. Am J Respir Critical Care Med 2014;189(5):576e85. [17] Malta M, Cardoso LO, Bastos FI, Magnanini MM, Silva CM. STROBE initiative: guidelines on reporting observational studies. Rev saude Publica 2010;44(3):559e65. [18] Weycker D, Edelsberg J, Oster G, Tino G. Prevalence and economic burden of bronchiectasis. Clin Pulm Med 2005;12:205. [19] Martinez-Garcia MA, Soler-Cataluna JJ, Perpina-Tordera M, Roman-Sanchez P, Soriano J. Factors associated with lung function decline in adult patients with stable non-cystic fibrosis bronchiectasis. Chest 2007;132(5):1565e72. [20] King PT, Holdsworth SR, Freezer NJ, Villanueva E, Holmes PW. Microbiologic follow-up study in adult bronchiectasis. Respir Med 2007;101(8):1633e8. [21] Ho PL, Chan KN, Ip MS, Lam WK, Ho CS, Yuen KY, et al. The effect of Pseudomonas aeruginosa infection on clinical parameters in steady-state bronchiectasis. Chest 1998;114(6):1594e8. [22] Lee JH, Kim YK, Kwag HJ, Chang JH. Relationships between high-resolution computed tomography, lung function and bacteriology in stable bronchiectasis. J Korean Med Sci 2004; 19(1):62e8. [23] Cheng K, Smyth RL, Govan JR, Doherty C, Winstanley C, Denning N, et al. Spread of beta-lactam-resistant Pseudomonas aeruginosa in a cystic fibrosis clinic. Lancet 1996; 348(9028):639e42. [24] Fothergill JL, Walshaw MJ, Winstanley C. Transmissible strains of Pseudomonas aeruginosa in cystic fibrosis lung infections. Eur Respir J 2012;40(1):227e38. [25] De Soyza A, Brown JS, Loebinger MR. Bronchiectasis research academic network. Research priorities in bronchiectasis. Thorax 2013;68:695e6.

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021

+

MODEL

Non cystic fibrosis bronchiectasis [26] Foweraker JE, Laughton CR, Brown DF, Bilton D. Phenotypic variability of Pseudomonas aeruginosa in sputa from patients with acute infective exacerbation of cystic fibrosis and its impact on the validity of antimicrobial susceptibility testing. J Antimicrob Chemother 2005;55(6):921e7. [27] Gillham MI, Sundaram S, Laughton CR, Haworth CS, Bilton D, Foweraker JE. Variable antibiotic susceptibility in populations of Pseudomonas aeruginosa infecting patients with bronchiectasis. J Antimicrob Chemother 2009;63(4):728e32. [28] Hadjiliadis D, Steele MP, Chaparro C, Singer LG, Waddell TK, Hutcheon MA, et al. Survival of lung transplant patients with cystic fibrosis harboring panresistant bacteria other than Burkholderia cepacia, compared with patients harboring sensitive bacteria. J Heart Lung Transpl 2007; 26(8):834e8. [29] Evans SA, Turner SM, Bosch BJ, Hardy CC, Woodhead MA. Lung function in bronchiectasis: the influence of Pseudomonas aeruginosa. Eur Respir J 1996;9(8):1601e4. [30] Hill AT, Pasteur M, Cornford C, Welham S, Bilton D. Primary care summary of the British Thoracic Society Guideline on the management of non-cystic fibrosis bronchiectasis. Prim Care Respir J 2010;20(2):135e40.

11 [31] Dhar R, Anwar GA, Bourke SC, Doherty L, Middleton P, Ward C, et al. Efficacy of nebulised colomycin in patients with noncystic fibrosis bronchiectasis colonised with Pseudomonas aeruginosa. Thorax. 65(6):553. [32] Metersky ML. New treatment options for bronchiectasis. Ther Adv Respir Dis. 4(2):93e99. [33] Murray MP, Turnbull K, Macquarrie S, Hill AT. Assessing response to treatment of exacerbations of bronchiectasis in adults. Eur Respir J 2009;33(2):312e8. [34] Murray MP, Govan JR, Doherty CJ, Simpson AJ, Wilkinson TS, Chalmers JD, et al. A randomized controlled trial of nebulized gentamicin in non-cystic fibrosis bronchiectasis. Am J Respir Crit Care Med 2010;183(4):491e9. [35] Donaldson GC, Seemungal TA, Bhowmik A, Wedzicha JA. Relationship between exacerbation frequency and lung function decline in chronic obstructive pulmonary disease. Thorax 2002;57(10):847e52. [36] Seemungal TA, Donaldson GC, Paul EA, Bestall JC, Jeffries DJ, Wedzicha JA. Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998;157(5 Pt 1):1418e22.

Please cite this article in press as: McDonnell MJ, et al., Non cystic fibrosis bronchiectasis: A longitudinal retrospective observational cohort study of Pseudomonas persistence and resistance, Respiratory Medicine (2014), http://dx.doi.org/10.1016/j.rmed.2014.07.021