Mycobacterium avium complex infection in non-cystic fibrosis bronchiectasis ZAID ZOUMOT,1,2 AFRODITI K. BOUTOU,1 SIMON S. GILL,3 MAFALDA VAN ZELLER,2 DAVID M. HANSELL,1,3 ATHOL U. WELLS,1,4 ROBERT WILSON1,2 AND MICHAEL R. LOEBINGER1,2 1 National Institute for Health Research Respiratory Biomedical Research Unit, Royal Brompton and Harefield Hospital Foundation Trust and Imperial College, 2Host Defence Unit, 3Department of Radiology and 4Interstitial Lung Disease Unit, Royal Brompton and Harefield NHS Foundation Trust, London, UK
ABSTRACT Background and objective: Reliable markers of disease progression or stability to assist in management decisions are lacking in patients with non-cystic fibrosis bronchiectasis and Mycobacterium avium complex (MAC) infection. Methods: Data from 52 adults with non-cystic fibrosis bronchiectasis and coexisting MAC infection managed at our institution over a 5-year period were retrospectively analysed. High-resolution computed tomography (HRCT) scans were scored using a scoring system that focused on findings associated with MAC infection. Results: Chronic pulmonary aspergillosis was independently associated with mortality (hazard ratio (HR) = 8.916, 95% confidence interval (CI) = 1.324– 60.027), as were nodules with cavitation (HR = 5.911, 95% CI = 1.095–25.911) and emphysema (HR = 1.027, 95% CI = 1.002–1.053) on HRCT. Anti-MAC chemotherapy was more likely to lead to MAC culture conversion (67% vs 27%, P = 0.005) but did not improve survival as compared with patients managed with observation. Longitudinally, patients who had improvements in HRCT scores were younger (60.2 ± 9.19 years vs 69.83 ± 12.43 years, P = 0.043), while the presence of cavitation within nodules predicted a deterioration in HRCT scores (0.5 (0–3) vs 0 (0–1), P = 0.033). No significant longitudinal differences were found in lung function in the cohort as a whole or within different groups. Conclusions: Chronic pulmonary aspergillosis in patients with bronchiectasis and coexisting MAC infection is a strong predictor of mortality. Cavitation within nodules and emphysema on HRCT at presentation were independently associated with mortality. Key words: aspergillus lung disease, atypical mycobacterial disease, bronchiectasis, respiratory infection (non-tuberculous). Correspondence: Michael Loebinger, Host Defence Unit, Royal Brompton Hospital, Sydney Street, London SW3 6NP, UK. Email: [email protected]
Received 6 November 2013; invited to revise 3 January 2014; revised 7 January 2014; accepted 4 February 2014 (Associate Editor: Yuanlin Song). Article first published online: 1 April 2014 © 2014 Asian Pacific Society of Respirology
SUMMARY AT A GLANCE In patients with non-cystic fibrosis bronchiectasis and Mycobacterium avium complex infection, chronic pulmonary aspergillosis is a strong independent predictor of mortality, as are nodules with cavitation and emphysema on high-resolution computerized tomography (HRCT) scans at presentation. Longitudinally, the presence of cavitation within nodules predicted deterioration in HRCT scores.
Abbreviations: ABPA, allergic bronchopulmonary aspergillosis; AFB, acid-fast bacilli; CF, cystic fibrosis; CI, confidence interval; CT, computed tomography; CPA, chronic pulmonary aspergillosis; FEV1, forced expiratory volume in 1 s; HR, hazard ratio; HRCT, high-resolution computed tomography; MAC, Mycobacterium avium complex; NTM, non-tuberculous mycobacteria; OR, odds ratio.
INTRODUCTION Non-tuberculous mycobacteria (NTM) pulmonary disease prevalence is increasing worldwide.1–8 These ubiquitous environmental organisms are often cultured from respiratory samples of patients with underlying lung disease, and it can be difficult to differentiate airway colonization from active infection due to non-specific clinical and radiological presentation. This is especially the case in patients with underlying non-cystic fibrosis (CF) bronchiectasis, with its diverse aetiologies (which include primary NTM infection). In a UK-based, unselected cohort of adults with non-CF bronchiectasis, prevalence of NTM was reported to be 2%, and Mycobacterium avium complex (MAC) was the predominant species. Seventytwo per cent of patients had multiple NTM isolates cultured MAC.9 The natural history and prognosis of patients with non-CF adult onset bronchiectasis with Respirology (2014) 19, 714–722 doi: 10.1111/resp.12287
MAC infection in non-CF bronchiectasis
coexisting MAC infection are unclear. A proportion of patients are asymptomatic with some achieving spontaneous clearance, and therefore, clinically stable patients may not be treated but simply observed in routine clinical practice. Also, treatment is lengthy and often poorly tolerated due to significant side-effects and poor cure rate.10 Hence, the decision to embark on treatment can be a difficult one. This needs to be determined on an individual basis giving careful consideration to the potential risks and benefits. There is a clinical need for prognostic information at initial MAC presentation in patients with underlying bronchiectasis that can identify those at highest risk of disease progression and mortality, as well as those likely to achieve spontaneous MAC clearance (or culture conversion). We conducted a retrospective study of patients with underlying bronchiectasis and coexisting MAC infection managed at our institution over a 5-year period to identify high-resolution computed tomography (HRCT) characteristics, lung function variables, and clinical features that can act as predictors of (i) mortality, (ii) MAC culture conversion, (iii) disease progression and longitudinal outcome, and (iv) determinants of treatment with observation or antiMAC chemotherapy.
METHODS Approval from our institution’s audit board was obtained to access patients’ records for use in this study, and patient confidentiality was strictly maintained. The Royal Brompton Hospital Microbiology Department registry database was used to retrospectively identify all adult patients with non-CF bronchiectasis with two or more positive MAC cultures from respiratory specimens within 12 months at our institution over a 5-year period between 1 July 2004 and 31 June 2009. The Royal Brompton Hospital is a tertiary referral centre covering all respiratory specialities including thoracic surgery. There are approximately 2000 adults with non-CF bronchiectasis under the care of the Host Defence Unit. Sputum is routinely examined for acid-fast bacilli (AFB) culture on an annual basis in all bronchiectasis patients and more frequently if there are clinical or radiological features to suggest NTM infection. Mycobacterial culture is performed using the mycobacteria growth indicator tube liquid-based culture system. Identification of mycobacterial species was performed by GenoType Mycobacterium CM/AS testing (Hain Lifescience GmbH, Nehren, Germany). A minimum inhibitory concentration of >32 mcg/mL was considered to represent macrolide resistance. Data were collected retrospectively on demographics, age, sex, clinical features, aetiology of bronchiectasis, coexistent sputum pathogens, sputum culture conversion, MAC macrolide resistance, and clinical and serological features of aspergillus lung disease by reviewing case notes and electronic records for all patients. Chronic pulmonary aspergillosis (CPA) encompassed aspergilloma, chronic cavitary pulmo© 2014 Asian Pacific Society of Respirology
715 nary aspergillosis, chronic fibrosing pulmonary aspergillosis and chronic necrotizing aspergillosis of longer than 3 months duration as previously defined,11 but not allergic bronchopulmonary aspergillosis (ABPA). The index date was defined as the date of the first positive MAC culture. MAC culture conversion was defined as three consecutive negative cultures within 6 months (as previously described12). The conversion date was the first negative culture within this period. Stable patients without evidence of clinical disease were reviewed every 6 months or more frequently as needed. Subjects initiated on or being considered for treatment with anti-MAC chemotherapy were reviewed as per clinical need. Pulmonary function tests including whole body plethysmography and gas transfer measurement were performed according to American Thoracic Society and European Respiratory Society guidelines and using European Community for Coal and Steel workers’ cohort normal values.13 Baseline lung function results closest to the index date (within 6 months) were used for analysis. For patients who had serial lung function testing, measurements closest to the end of the follow-up period were also retrieved. HRCT (Sensation 64, Siemens Medical, Erlangen, Germany) was acquired using volumetric spiral acquisition with contiguous slices (0.6-mm slice thickness on 0.5 mm for lung windows, and 10 mm on 10 mm mediastinal windows) with the participant in the supine position at full inspiration. HRCTs obtained within 6 months of first MAC diagnosis were scored by a thoracic radiologist (S.G.) blinded to clinical information using a scoring system devised for the purpose of this study with a focus on features encountered in MAC infection (Table 1). The following were scored in each lobe: bronchiectasis extent, bronchiectasis severity, tree-in-bud, nodules, mosaicism, consolidation, cavitation within nodules, severe cavitation, percentage of lobe with emphysema, features of aspergilloma (a fungal ball seen as a mass of softtissue attenuation within a pulmonary cavity) or semi-invasive aspergillosis (localized areas of consolidation often with irregular margins and satellite (acinar) nodules surrounding eccentric cavitation with intracavitary material, often associated with thickened pleura). The sum of the lobar scores for each HRCT feature was used in the analyses, except the emphysema score where the average lobar emphysema score was used. Follow-up HRCT was scored using the same scoring system at the same setting to allow side-by-side comparison.
Statistical analysis All analyses were performed using the PASW (Predictive Analytics Software, SPSS, Inc., Chicago, IL, USA) version 18 for Windows 8. Data are presented as percentages (%) for categorical variables, as mean ± 1 standard deviation for normally distributed continuous variables, and as median with minimummaximum range for non-normally distributed continuous variables. The normality test applied was the Shapiro–Wilk test. The differences between Respirology (2014) 19, 714–722
716 Table 1
Z Zoumot et al. Mycobacterium avium complex infection high-resolution computerized tomography scoring system
Subject number _____________ Date of HRCT _____________
Total (or average) score
Bronchiectasis extent 0 = none; 1 = 1 bronchopulmonary segment (BPS); 2 = >1 BPS;3 = >2 BPS Bronchiectasis severity 0 = none; 1 = mild (1–2 times diameter of artery); 2 = severe (>double artery diameter) Tree-in-bud 0 = none, 1 = mild; 2 = moderate; 3 = severe Nodules (>5 mm) 0 = none; 1 = 1nodule; 2 = 2–4 nodules; 3 = 5 or more modules Mosaicism 0 = none, 1 = mild; 2 = moderate; 3 = severe Consolidation 0 = none, 1 = mild; 2 = moderate; 3 = severe Cavitation within nodules 0 = no; 1 = yes Severe cavitation (e.g. MTB like) 0 = no; 1 = yes Emphysema
Percentage of lobe with emphysema (nearest 5%) Aspergilloma 0 = no; 1 = yes Semi-invasive aspergillosis 0 = no; 1 = yes LUL, left upper lobe; LLL, left lower lobe; MTB, mycobacterium tuberculosis; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe.
groups for continuous variables were studied using either the independent sample’s t-test or the Mann– Whitney U-test, depending on the normality of their distribution. The differences between groups for categorical variables were tested using the chi-square test. For the survival analysis, each parameter was entered in a univariate Cox proportional hazard analysis model; all parameters found to be univariately associated with mortality were subsequently entered in a multivariate Cox proportional hazard regression analysis model. The proportionality hazard assumption was individually tested for every variable that constructed the final model using the partial residual plots (Schoenberg residuals proportional hazard test) and was fulfilled for all independent predictors of mortality. Corresponding hazard ratios (HR) and 95% confidence intervals (CI) were calculated for each independent predictor, and the median survival was estimated using the Kaplan– Meier method. Level of P < 0.05 was considered significant. All variables found to be univariately associated with culture conversion were entered in a Respirology (2014) 19, 714–722
stepwise binary logistic regression analysis model to identify independent predictors of culture conversion; corresponding odds ratios (OR) and 95% CI were calculated for each predictor. For the longitudinal analyses, the paired sample’s t-test or the Wilcoxon signed-ranks test were used depending on the normality of the variables’ distribution.
RESULTS Patient population There were 1506 positive NTM cultures between 1 July 2004 and 31 June 2009, of which 894 were from 212 patients who had two or more positive cultures of the same NTM organism (single isolates were excluded). Of these, 117 patients cultured MAC on two or more occasions, 59 of whom were adults with non-CF bronchiectasis. Data from one patient were insufficient to allow addition to the database, and six were on long term anti-MAC chemotherapy (>48 months) with failed eradication at the index point and had missing © 2014 Asian Pacific Society of Respirology
MAC infection in non-CF bronchiectasis Study period: 1 July 2004 to 31 June 2009
Table 2 Baseline characteristics of the total study population (n = 52) n (%), mean ± SD or median (minimummaximum)
1506 posi ve NTM cultures
894 cultures in 212 pa ents culturing the same NTM at least twice
-29 Cys c fibrosis -8 Pulmonary alveolar proteinosis -8 Emphysema -4 Inters al lung disease -3 Asthma -6 Others Exclusions: - Insuﬃcient data: 1 pa ent. - Already on long-term (>48 months) an -MAC chemotherapy for refractory disease: 6 pa ents
Two or more MAC cultures: 117 pa ents
59 adults with non-CF bronchiectasis
Study popula on: 52 pa ents Figure 1 Sample numbers and patient recruitment. CF, cystic fibrosis; MAC, Mycobacterium avium complex; NTM, nontuberculous mycobacteria.
baseline information, thus giving a final study population of 52 patients (Fig. 1). Median follow-up time, defined as time from the index date to 1 July 2013 or death, was 88.4 months (95% CI 79.4–97.6). Forty-eight patients had baseline computed tomography (CT) scans, and 49 patients had baseline spirometry. Thirty-five patients had body plethysmography and 33 gas transfer measurements within 6 months of the index date. The baseline values for the cohort are presented in Table 2.
Survival analysis All-cause mortality was 19% (n = 10/52). Nonsurvivors had significantly higher HRCT scores for markers of active MAC infection at the time of MAC diagnosis (nodules with cavitation, lobes with severe cavitation and consolidation) and higher emphysema scores (Table 3). They were also significantly more likely to have coexisting CPA and have AFB smearpositive sputum of the first MAC isolate compared with survivors (Table 3). These variables were associated with mortality in the univariate Cox proportional hazard analysis and were then entered in a stepwise multivariate Cox proportional hazard regression analysis model (Table 4). CPA was independently associated with mortality (HR = 8.916, 95% CI = 1.324–60.027), as were the number of nodules with cavitation (HR = 5.911, 95% CI = 1.095–25.911) and © 2014 Asian Pacific Society of Respirology
Gender, n (%) Female Male Age (years) Body mass index (kg/m2) FEV1% predicted FVC% predicted TLCOc% predicted KCOc% predicted TLC% predicted Bronchiectasis aetiology, n (%) Idiopathic Postinfectious (including MTB) Primary MAC Allergic bronchopulmonary aspergillosis Immunodeficiency Other Chronic pulmonary aspergillosis, n (%)† Yes No MAC Culture conversion, n (%)‡ Yes No Smear at first MAC isolation, n (%) Positive Negative Macrolide resistance, n (%) Sputum bacteria other than MAC, n (%) None Pseudomonas aeruginosa Other bacteria HRCT parameter scores (maximum scores)† Extent of bronchiectasis (18) Severity of bronchiectasis (12) Tree in bud (18) Nodules (18) Cavitation within nodules (14) Lobes with severe cavitation(14) Mosaicism (18) Consolidation (18) Emphysema score (100) Semi-invasive aspergillosis (14) Aspergilloma (14)
36 (69.2) 16 (30.8) 63.1 ± 12.7 19.5 ± 2.9 64.1 ± 24.8 89 ± 21.2 64.2 ± 21.6 85.8 ± 25.2 108.8 ± 23.5 11 (21.2) 19 (36.5) 6 (11.5) 5 (9.6) 2 (3.9) 9 (17.3) 8 (16.7) 40 (83.3) 22 (43.1) 29 (55.8) 20 (38.5) 32 (61.5) 7 (13.5) 12 (23.1) 27 (51.9) 13 (25.0) 6 (0–16) 5 (0–10) 2 (0–11) 1.5 (0–14) 0 (0–3) 0 (0–3) 0 (0–14) 0 (0–5) 0 (0–97.5) 0 (0–2) 0 (0–2)
Data available for 48 patients. Data available for 51 patients. FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; HRCT, high-resolution computed tomography; KCOc, gas transfer (corrected); MAC, Mycobacterium avium complex; MTB, mycobacterium tuberculosis; SD, standard deviation; TLC, total lung capacity; TLCOc, diffusion capacity for carbon monoxide (corrected). ‡
Respirology (2014) 19, 714–722
718 Table 3
Z Zoumot et al. Baseline differences between survivors and non-survivors
Gender, n (%) Female Male Age (years) Body mass index (kg/m2) Lung function FEV1% predicted FVC% predicted KCOc% predicted TLC% predicted Bronchiectasis aetiology, n (%) Idiopathic Postinfection Primary MAC Allergic bronchopulmonary aspergillosis Immunodeficiency Other Chronic pulmonary aspergillosis, n (%)† Yes no MAC culture conversion, n (%)‡ Yes No Smear at first MAC isolation, n (%) Positive Negative MAC macrolide resistance, n (%) Yes No Sputum bacteria other than MAC, n (%) None Pseudomonas aeruginosa Other bacteria HRCT parameter scores† Extent of bronchiectasis Severity of bronchiectasis Tree in bud Nodules Nodules with cavitation Lobes with severe cavitation Mosaicism Consolidation Emphysema score Semi invasive aspergillosis Aspergilloma
Survivors (n = 42)
Non-survivors (n = 10)
31 (86.1) 11 (68.8) 62.1 ± 11.3 19.6 ± 2.7
5 (13.9) 5 (31.3) 63.8 ± 16.5 19.2 ± 3.9
0.149 0.701 0.773
69.1 ± 23.4 90.7 ± 22.1 88.2 ± 22.7 107.9 ± 23.7
51.5 ± 30.2 87.0 ± 16.5 66.3 ± 32.1 120.7 ± 24.8
0.069 0.625 0.058 0.273
9 (81.8) 16 (84.2) 5 (83.3) 4 (80) 2 (100) 6 (66.7)
2 (18.2) 3 (15.8) 1 (16.7) 1 (20) 0 (0) 3 (33.3)
1 (12.5) 37 (92.5)
7 (87.5) 3 (7.5)