Open Forum Infectious Diseases MAJOR ARTICLE The Role of Epstein-Barr Virus in Adults With Bronchiectasis: A Prospective Cohort Study Chun-Lan Chen,1,2 Yan Huang,1 Miguel Angel Martinez-Garcia,3 Jing-Jing Yuan,1 Hui-Min Li,1 David de la Rosa-Carrillo,4 Xiao-Rong Han,1 Rong-Chang Chen,5 Wei-Jie Guan,1 and Nan-Shan Zhong1,a 1 State Key Laboratory of Respiratory Disease, National Clinical Research Center for Respiratory Disease, Guangzhou Institute of Respiratory Health, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou, China, 2Guangdong General Hospital (Guangdong Academy of Medical Sciences), Guangzhou, China, 3Pneumology Department, University and Politechnic La Fe Hospital, Valencia, Spain, 4Pulmonology Service, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, and 5Shenzhen People’s Hospital, Shenzhen, China Background. Epstein-Barr virus (EBV) is implicated in the progression of chronic obstructive pulmonary disease. We aimed to Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 determine whether EBV correlates with bronchiectasis severity, exacerbations, and progression. Methods. We collected induced sputum in healthy controls and spontaneous sputum at 3–6-month intervals and onset of ex- acerbations in bronchiectasis patients between March 2017 and October 2018. EBV DNA was detected with quantitative polymerase chain reaction. Results. We collected 442 sputum samples from 108 bronchiectasis patients and 50 induced sputum samples from 50 healthy controls. When stable, bronchiectasis patients yielded higher detection rates of EBV DNA (48.1% vs 20.0%; P = .001), but not viral loads (mean log10 load, 4.45 vs 4.76; P = .266), compared with controls; 64.9% of patients yielded consistent detection status between 2 consecutive stable visits. Neither detection rate (40.8% vs 48.1%; P = .393) nor load (mean log10 load, 4.34 vs 4.45; P = .580) differed between the onset of exacerbations and stable visits, nor between exacerbations and convalescence. Neither detection status nor viral loads correlated with bron- chiectasis severity. EBV loads correlated negatively with sputum interleukin-1β (P = .002), CXC motif chemokine-8 (P = .008), and tumor necrosis factor–α levels (P = .005). Patients initially detected with, or repeatedly detected with, EBV DNA had significantly faster lung function decline and shorter time to next exacerbations (both P < .05) than those without. Detection of EBV DNA was unrelated to influ- enza virus and opportunistic bacteria (all P > .05). The EBV strains detected in bronchiectasis patients were phylogenetically homologous. Conclusions. Patients with detection of EBV DNA have a shorter time to bronchiectasis exacerbations. EBV may contribute to bronchiectasis progression. Keywords. airway inflammation; chronic airway disease; chronic viral infection; exacerbation; human herpes virus–4; lung function. Bronchiectasis is a debilitating chronic airway inflammatory suggesting that chronic viral infection correlates with chronic disease characterized by pathogen infections, chronic airway airway inflammation. inflammation, and airway destruction, which provide the niche Epstein-Barr virus (EBV) is a ubiquitous virus associated for bacterial, viral, and fungal infections [1–3]. Acute viral in- with chronic infection of B-memory lymphocytes and airway fections are responsible for aggravating airway inflammation epithelial cells [7, 8]. The lower respiratory tract is an important that frequently triggers exacerbations [4]. Consistent with reservoir of EBV [8], which could be readily detected in COPD. findings in chronic obstructive pulmonary disease (COPD), The considerable detection rate of EBV deoxynecluotide acid CD8+ lymphocyte infiltrates the bronchiectatic airways [5, 6], (DNA) in COPD patients (46%–48%) [9], coupled with the link between chronic EBV infection and disease progression [9, 10], have hinted at the role of EBV in bronchiectatic airways where host–defense responses are defective. Received 4 March 2020; editorial decision 8 June 2020; accepted 11 June 2020. a We hypothesized that EBV predisposes to the inflammatory Senior author for the study. Correspondence: Wei-jie Guan, PhD, State Key Laboratory of Respiratory Disease, National responses and progression in bronchiectasis. Our primary ob- Clinical Research Center for Respiratory Disease, Guangzhou Institute of Respiratory Health, jectives were to elucidate the association between EBV and the The First Affiliated Hospital of Guangzhou Medical University, 151 Yanjiang Road, Guangzhou, Guangdong, China (battery203@163.com). severity, exacerbations, and progression of bronchiectasis. We Open Forum Infectious Diseases® also determined the changes in viral loads during exacerbations © The Author(s) 2020. Published by Oxford University Press on behalf of Infectious Diseases and risk factors for the detection of EBV DNA. Furthermore, we Society of America. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (http://creativecommons.org/licenses/ evaluated the phylogenetic homogeneity of the EBV strains in by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any our bronchiectasis patient cohort. Our findings might indicate medium, provided the original work is not altered or transformed in any way, and that the the roles of EBV in bronchiectasis and highlight prophylactic work is properly cited. For commercial re-use, please contact journals.permissions@oup.com DOI: 10.1093/ofid/ofaa235 and therapeutic targets. EBV Infection in Bronchiectasis • ofid • 1 METHODS Details of sputum collection (including quality control) are Study Design and Subjects described in the Supplementary Data. We extracted viral nu- In this single-center prospective observational study, we en- cleic acids using the TaKaRa MiniBEST Viral RNA/DNA rolled symptomatic bronchiectasis patients (chronic cough, Extraction Kit (Takara Bio Inc., San Jose, USA) from sputum daily sputum production) from the First Affiliated Hospital homogenized with 0.1% dithiothreitol. TaqMan real-time quan- of Guangzhou Medical University between March 2017 and titative polymerase chain reaction (qPCR) was performed to October 2018. Eligible patients, aged 18–75 years, underwent detect 18 common respiratory viruses (including EBV) with chest high-resolution computed tomography (HRCT) within the ABI Prism 7500 Real-time qPCR System (Thermo Fisher the previous 12 months. Bronchiectasis was diagnosed based Scientific Inc., Waltham, USA). Viral detection kits were pur- on international guidelines [11]. Patients with active tubercu- chased from Guangzhou HuYanSuo Medical Technology Co. losis, traction bronchiectasis, COPD as the primary etiology of Ltd. (Guangzhou, China). The cycle threshold of 40 was con- bronchiectasis, malignancy, and severe systemic diseases were sidered positive. Sputum samples with positive EBV testing re- excluded. sults were chosen for the measurement of copy numbers. The Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 Healthy subjects included patients’ companions, medical target sequence of the gp58 gene with known concentration staff, and those who underwent health check-up, had normal (serial dilutions), which was extracted from the plasmid vector, chest x-ray and spirometry, and had no respiratory symptoms, was used for performing the standard curve evaluation. The severe systemic diseases, or antibiotic use within 4 weeks. mean computed tomography value was transformed to the copy The Ethics Committee of the First Affiliated Hospital of number according to the standard curve. For EBV, detection Guangzhou Medical University gave approval (medical ethics was defined as EBV DNA detected at any visit during follow-up, 2012 [the 29th]). All participants provided written informed whereas repeated detection denoted EBV DNA detected on at consent. least 2 occasions at least 3 months apart within 1 year, which was extrapolated from the definition of bacterial isolation or Procedures colonization [17]. We enrolled bronchiectasis patients who remained exacerbation- Sputum was centrifuged (20 000g) at 4°C for 120 minutes to free for >4 weeks. At initial visits, patients were asked for clinical prepare for the sputum sol phase (supernatant). Sputum CXC history, exacerbation frequency, and concomitant medications. motif chemokine–8 (CXCL-8), interleukin (IL)-1β, tumor ne- Bronchiectasis etiology, modified Reiff score, and spirometry crosis factor–alpha (TNF-α), and interferon-γ (IFN-γ) were were evaluated [12–14]. Disease severity was calculated with 2 detected with multiplex immunoassay kits (Biorad, Hercules, integrated metrics, including the Bronchiectasis Severity Index USA) [17]. (BSI) [15] and E-FACED score [16]. Spontaneous sputum was collected for assays of EBV and other respiratory viruses Statistics Analysis (Supplementary Data), bacterial culture, and inflammatory bi- The detection rate of EBV DNA in sputum in Chinese people omarker assays in bronchiectasis patients [17]. is unknown. Assuming a detection rate of 46% in stable bron- We routinely followed up patients every 3–6 months, during chiectasis [9] and 16% in healthy subjects (pilot study with 18 which patients contacted investigators in case of significant subjects), considering a dropout rate of 20%, we estimated that symptom worsening (within 2–5 days) and an exacerbation we needed to recruit 100 bronchiectasis patients and 50 healthy (AE) visit would be conducted within 48 hours. Following controls based on 2-sided significance of .05 and a power an international consensus, an AE was defined as continuous of 90%. (≥48 hours) significant deterioration of 3 or more symptoms The EBV load was presented as copy numbers and was sub- (increased cough frequency, sputum volume and/or consist- jected to logarithmic transformation. The EBV detection rate ency, sputum purulence, breathlessness and/or exercise toler- and loads were analyzed using the initial stable visit samples ance, fatigue and/or malaise, and hemoptysis) that required and samples during onset of AEs. Sputa collected at the initial changes in treatment [18]. Spontaneous sputum was sampled visits and follow-up were included as “any stable” visit sam- at stable visits, at AE visit (day 1, before antibiotic administra- ples, whereas sputa collected during AE visits were included as tion) and on-treatment days (including days 4, 7, and 14), and “any AE” samples. Sputa collected during stable visits and onset postexacerbation stable visit (days 40–80) for a subset of AE of AEs from the same bronchiectasis patient were defined as events (collectively termed AE convalescence follow-up visits) “paired stable” and “paired AE” samples. [19]. The study design is illustrated in Figure 1. Data were presented as mean ± SD or median (interquartile Healthy subjects attended a single visit, during which his- range [IQR]) for continuous variables depending on the nor- tory taking, 3% hypertonic saline sputum induction (in healthy mality (assessed with the Kolmogorov-Smirnov test) and ana- subjects only), chest x-ray, and spirometry were performed. lyzed using independent and paired t tests or Mann-Whitney 2 • ofid • Chen et al A History taking Day1 Days 4, 7, 14, and 40–80 Follow-up visit 1, 2, 3, etc. Sputum sampling Sputum sampling symptom assessment Exacerbation history taking * Bacterial culture * Bacterial culture sputum sampling sputum sampling * qPCR: virus * qPCR: virus * qPCR: virus * Bacterial culture Spirometry * qPCR: virus Etiologic assessment Spirometry Baseline stable Exacerbation Convalescence Follow-up period Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 B 130 patients screened 22 excluded 12 declined 6 had insufficient sputum 2 pregnancy 1 severe heart valve diseases 1 endometrial cancer 108 patients underwent initial visits 442 samples 10 discontinued 7 declined to follow-up 3 had insufficient sputum during follow-up 98 patients underwent follow-up visits 50 healthy controls 7 excluded 7 only provide AE sputum samples during follow-up 50 induced 108 patients: 108 49 patients: 91 patients: sputum initial stable 198 stable 68 AE sputum samples sputum samples sputum samples samples (Day1) during follow-up Convalescence samples: 23: Day4 samples 23: Day7 samples 22: Day14 samples 16: Day42–60 samples Figure 1. Flowchart of study participant recruitment and sampling schemes. The 18 sputum specimens that are underlined within the figure were collected during stable-state visits; hence, they were also included in the 198 stable samples of 91 bronchiectasis patients. Abbreviations: AE, acute exacerbation; qPCR, quantitative polymerase chain reaction. U tests and Wilcoxon tests where appropriate. Counts (per- the correlation between the load of EBV and clinical param- centages) were presented for categorical variables and ana- eters with Pearson’s correlation analysis. Changes in EBV load lyzed with the chi-square or Fisher exact test. We determined throughout the course of AE were analyzed with the Wilcoxon EBV Infection in Bronchiectasis • ofid • 3 test and stratified by the presence of common cold (frequently However, the detection rate did not differ between pooled stable caused by respiratory viruses). Generalized estimating equa- and pooled AE samples (P = .810) or between paired stable and tions (GEEs) with the logit link were used to explore the ef- paired AE samples (P = .761). fects of EBV DNA detection on the odds of experiencing AEs The EBV loads did not differ significantly among the whole compared with stable visits. A linear regression structure was cohort, AE cohort, and controls even after adjustment for age used for EBV load comparison between AE visits and stable (P = .393). Similar findings were found in both paired samples visits, taking into account the repeated observations from indi- and pooled samples (Supplementary Figure 1). EBV DNA was vidual participants. The trends of changes among the 3 groups detected more commonly in winter than in summer (39.0% vs (ie, lung function decline) were analyzed. Linear mixed-effect 61.5%; P = .036), whereas viral loads were comparable. EBV models were applied to compare lung function decline between loads did not differ significantly among healthy controls, in- patients detected and those not detected with EBV DNA and haled steroid users, and nonusers (P = .458). between patients with and without repeated detection, with the Furthermore, 36 (39.6%) patients yielded repeated detec- exacerbation frequency being treated as the covariate in the tion of EBV DNA and tended to have higher EBV loads than Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 model. The risks of AEs during follow-up were compared with those without (P = .073). Patients with EBV detected at stable the Kaplan-Meier survival model with the Gehan-Breslow- visits had increased risks (OR, 21.33; 95% CI, 6.44–70.63) for Wilcoxon test. The odds ratio (OR) of EBV DNA detection repeated detection in subsequent visits. was estimated with logistic regression models, and data with P < .20 were entered into multivariate analysis using a back- EBV Detection Status Remained Consistent Over Time ward selection algorithm. P < .05 was considered statistically For 91 patients with ≥2 stable visit samples, 64.9% yielded con- significant. We constructed the phylogenetic tree among 25 sistent EBV detection status in 2 consecutive visits (24.2% re- bronchiectasis patients whose EBV was successfully sequenced peatedly detected; 40.7% consistently negative). EBV detected based on the latent membrane protein–1 (LMP-1) sequences at initial stable visits was prone to repeated detection (relative by using MEGA 7.0 (http://www.megasoftware.net/) and cal- risk [RR], 2.94; 95% CI, 1.40–6.20). The RR for being consist- culated the amino acid similarity with BioEdit 7.0.1 (GenBank ently EBV negative was 1.58 (95% CI, 1.17–2.13). Similar EBV accession number: MK944375-944399). Statistical analyses transition patterns were found from the initial stable visit to the were performed using SPSS 18.0 (SPSS Inc., Chicago, IL, USA) post-AE stable visit (Supplementary Table 2). or Graphpad Prism, version 5.0 (Graphpad Inc., San Diego, Among 59 patients with ≥3 stable visit samples, EBV DNA CA, USA). detection status remained unchanged in 40.0% during 3 con- secutive stable visits (23.0% repeatedly detected; 17.0% con- tinuously negative). EBV was detected at 2 subsequent visits in RESULTS 35.1% of patients with EBV detected at the initial stable visit, Subject Recruitment and Clinical Characteristics whereas 45.5% with EBV-negative sputum yielded negative Of 130 bronchiectasis patients screened, 108 completed the findings at 2 subsequent visits (Supplementary Figure 2). The initial stable visits. Ninety-eight patients completed follow-up EBV load was constant during 3 (P = .268) and 4 consecutive until October 2018, of whom 91 patients with >1 stable visit stable visits (P = .610) (Supplementary Figure 3). sample were included for EBV DNA detection analysis. Forty- nine patients provided 68 AE samples, of whom 19 completed Association Between EBV and AEs 23 AE convalescence follow-up visits. Detection of EBV DNA was not associated with the onset of AE Overall, we collected a median (IQR) of 3 (3–5) sputum sam- (OR, 0.91; 95% CI, 0.54–1.54), based on the pooled stable visit and ples per patient (Figure 1). The mean age was 46.8 years, and AE visit samples at symptom onset (within 2.0 [IQR, 2.0–3.3] days), 60.2% were females among bronchiectasis patients. The median which yielded no substantial difference in viral loads (P = .498). follow-up (IQR) was 13.0 (9.3–15.0) months. Baseline charac- Nineteen patients yielded 23 AE convalescence pairs of sam- teristics of the AE cohort were similar to the whole bronchiec- ples (120 sputum samples), with 6.0 (IQR, 5.0–6.0) samples tasis patient cohort. Bronchiectasis patients tended to be older per patient. The detection rates were 33.3%, 39.1%, 50%, and than controls. Bronchiectasis patients also had significantly 52.9% for pre–stable visit, onset of AE, day 14, and post-AE lower body mass index, there were fewer current smokers, stable visit (median [IQR], 45 [40–54] days after day 1), re- and they had poorer lung function compared with controls spectively. The corresponding mean log10 load of EBV was 5.08, (Table 1). 4.29, 4.77, and 4.84, respectively. The viral loads on day 1 were significantly lower than those on day 4 (P = .011) and day 7 The Detection Rate of EBV DNA and Viral Load (P = .002). The EBV detection rate and load varied insignifi- At initial stable visits, EBV DNA was detected more frequently cantly throughout the natural course of AE, irrespective of the in bronchiectasis than in controls (48.1% vs 20.0%; P = .001). presence of common cold (Figure 2). 4 • ofid • Chen et al Table 1. Demographic and Clinical Characteristics of Participants AE Cohort Healthy Subjects Parameters Whole Cohort (n = 108) (n = 49) (n = 50) P Value Age, median (IQR), y 49.5 (36.0–59.0) 50.0 (36.0–59.5) 41.5 (29.0–55.5) .165 Body mass index, mean (SD), kg/m2 20.4 (3.3) 19.6 (3.3) 22.9 (3.4) <.001 Sex, female, No. (%) 65 (60.2) 34 (69.5) 33 (66.0) .502 Smoking status, No. (%) Never-smokers 100 (92.6) 46 (93.9) 44 (88.0) <.001 Ex-smokers 8 (7.4) 3 (6.1) 0 (0.0) Current smokers 0 (0.0) 0 (0.0) 6 (12.0) FEV1 % predicted, median (IQR) 52.9 52.5 94.3 <.001 (41.0–70.1) (40.0–69.2) (87.7–102.6) Exacerbation frequency within 1 y, median (IQR) 2.0 (1.0–2.5) 2.0 (2.0–3.0)a NA NA Bronchiectasis Severity Index 7 (4–9) 8 (4–10) NA NA Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 Mild, No. (%) 32 (29.6) 14 (28.6) Moderate, No. (%) 50 (46.3) 20 (40.8) NA NA Severe, No. (%) 26 (24.1) 15 (30.6) E-FACED score, median (IQR) 2.5 (1.0–4.0) 2.0 (1.0–4.0) NA NA Mild, No. (%) 73 (67.6) 33 (67.3) Moderate, No. (%) 34 (31.5) 16 (32.7) NA NA Severe, No. (%) 1 (0.9) 0 (0) Etiology, No. (%) Postinfective 27 (25.0) 15 (30.6) Idiopathic 26 (24.1) 11 (22.4) NA NA Post-tuberculous 17 (15.7) 7 (14.3) Primary immunodeficiency 11 (10.2) 5 (10.2) Othersb 27 (25.0) 11 (22.4) Medications, No. (%) Inhaled corticosteroids 28 (25.9) 14 (28.6) NA NA Low-dose macrolides 13 (12.0) 8 (16.3) Vaccination status, No. (%) Influenza vaccination within the last 12 mo 7 (6.5) 5 (10.2) 0 (0.0) .383 Pneumococcal vaccination within the last 5 y 4 (3.7) 3 (6.1) 0 (0.0) .723 Data were presented as mean (SD) or median (IQR) for continuous variables depending on normality (assessed with the Kolmogorov-Smirnov test) and analyzed using independent and paired t tests or Mann-Whitney U tests and Wilcoxon tests where appropriate. P values in the last column denote the tests that compare 3 groups of the whole cohort, AE cohort, and con- trols. Patients with the primary diagnosis of COPD as the underlying etiology were excluded from this study. The high rates of never smokers are consistent with our previously published studies. No test for nontuberculous Mycobacteria was conducted during the study. Abbreviations: AE, adverse event; COPD, chronic obstructive pulmonary disease; FEV1, forced expiratory volume in 1 second; IQR, interquartile range. a P < .05: AE cohort compared with those without an AE. b Other known causes, including Kartagener’s syndrome (n = 8, 7.4%), asthma (n = 8, 7.4%), gastroesophageal reflux (n = 3, 2.8%), diffuse panbronchiolitis (n = 3, 2.8%), connective tissue disease (n = 2, 1.9%), cystic fibrosis transmembrane regulator–related disease (n = 1, 0.9%), congenital airway defects (n = 1, 0.9%). Association Between EBV and Bronchiectasis Severity EBV DNA detection was not associated with heightened airway Next, we analyzed the association between EBV and disease inflammation during AE. Interestingly, EBV loads correlated severity in the whole cohort (n = 108) for the initial stable negatively with sputum IL-1β (r = –.33; P = .002), CXCL-8 visit samples. Higher HRCT scores (≥13 points) were associ- (r = –.30; P = .008), and TNF-α levels (r = –.31; P = .005), re- ated with higher EBV loads (4.80 vs 4.21; P = .023) compared gardless of stable visits or AE visits (Supplementary Figure 4). with lower HRCT scores (7–12 points). However, the detection rate of EBV DNA and viral load differed unremarkably when EBV and Disease Progression stratified by lung function impairment and disease severity, in- Patients followed up for >6 months (n = 91, only 9.3% had con- cluding BSI and E-FACED scores. Repeated detection of EBV comitant asthma) were included. Overall, the median frequency DNA was not associated with bronchiectasis severity (Figure 3). of AEs (IQR) was 1.5 (0–2.4) per patient-year. This figure was 0.9 Bacterial isolation and colonization (especially Pseudomonas (0.2–2.3), 1.6 (0–3.0), and 1.6 (0.2–2.5), respectively (P = .587), for aeruginosa) reportedly augment airway inflammation. One patients with consistently negative EBV, detection of EBV DNA, hundred seventy-three stable visit samples and 55 AE samples and repeated detection of EBV DNA. The mean forced expiratory were available for sputum cytokine detection. EBV DNA de- volume in a 1-second (FEV1) decline was 49.3 mL/y among 70 tected was not associated with higher levels of sputum IL-1β, patients who completed spirometry. The mean FEV1 decline was CXCL-8, or IFN-γ in stable visits, except for TNF-α. Moreover, 19.9 mL/y, 28.7 mL/y, and 81.4 mL/y (age-adjusted P = .493) for EBV Infection in Bronchiectasis • ofid • 5 Detection rate (P = .038). However, FEV1 decline was comparable between EBV load patients detected and those not detected with EBV DNA A 60 57 57 6 (P = .240). There was no significant interaction between the ex- 53 5.08 acerbation frequency and the detection status of EBV (P = .529 Log10 (EBV copy number) 50 EBV detection rate, % 50 4.29 5.02 4.85 4.77 4.84 and .114, respectively) for the above comparisons. 4 40 EBV DNA detected at initial visits was associated with a sig- 39 nificantly shorter time to the next AE compared with negative 30 33 2 EBV (median, 160 vs 230 days; P = .024), as was repeated de- 20 tection compared with no repeated detection (median, 139 vs 227 days; P = .025). Meanwhile, FEV1 decline from baseline to 10 0 follow-up was significantly faster in patients with detection and Prestable Day1 Day4 Day7 Day14 Poststable repeated detection of EBV DNA (Figure 4). EBV loads did not B correlate with FEV1 decline (r = .066; P = .696) (Supplementary 8 Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 Figure 5). Log10, EBV copy number 6 Risk Factors Associated With EBV DNA Detection Multivariate analysis (among 108 patients with initial stable 4 visit samples) revealed that having greater daily sputum volume (OR, 2.03; 95% CI, 1.20–3.42), more bronchiectatic lobes (OR, 1.35; 95% CI, 1.03–1.78), and treatment with long-acting mus- 2 carinic antagonists (OR, 4.16; 95% CI, 1.24–13.92) were risk fac- tors for EBV DNA detection. Nonetheless, low-dose macrolides 0 Prestable Day1 Day4 Day7 Day14 Poststable were protective for EBV DNA detection (OR, 0.19; 95% CI, 0.05–0.77) (Supplementary Table 3). See the Supplementary C Cold Data regarding the independent variables finally included in 8 6.46 Not cold the regression model. Gender, age, body mass index, detection of respiratory viruses, and bacterial infection or colonization Log10, EBV copy number 5.42 5.37 6 5.24 were not associated with EBV DNA detection (Supplementary 4.74 4.99 Table 4). P. aeruginosa infection was also unrelated to EBV loads 4 4.93 4.73 (Supplementary Figure 6). 4.40 4.38 4.50 3.45 2 EBV Strains Were Phylogenetically Homologous Finally, we sequenced EBV strains based on nucleic acid se- 0 quence of LMP-1, a canonical component of EBV. The EBV Prestable Day1 Day4 Day7 Day14 Poststable strains were highly homologous within our bronchiectasis patient cohort, with amino acid similarity of 75.4%–100%. Figure 2. Changes in the detection rate and load of EBV at stable visits and during the natural course of AEs and convalescence. A, Changes in the EBV detection rate Moreover, these EBV strains were homologous (similarity, and EBV load (log10 load) during the natural course of AEs and convalescence. B, 74.5%–100%) compared with other EBV strains detected in Change in the EBV load at stable visits and during the natural course of AEs and Chinese patients (with diseases other than bronchiectasis) at convalescence within the same bronchiectasis patient. C, Change in the EBV load regions outside Guangdong province (Supplementary Figure 7, at stable visits and during the natural course of AEs and convalescence in patients with and without cold symptoms. Data are presented as percentage and mean (SD). Supplementary Table 5). The numbers appearing at different time points indicate the log10 load for EBV in sputum. The colors in Figure 2B represent different individual bronchiectasis pa- DISCUSSION tients. Day 1: within 2 (IQR, 2.0–3.3) days of the onset of AE symptoms; Poststable: median 45 (IQR, 40–54) days from the onset of AE symptoms. Abbreviations: AE, This is the first prospective study to explore the association be- adverse event; EBV, Epstein-Barr virus; IQR, interquartile range. tween EBV and bronchiectasis. EBV DNA was readily detected in sputum and was repeatedly detected in some bronchiectasis patients. Patients with detection and repeated detection of EBV patients with consistently negative EBV, detection of EBV DNA, DNA yielded significantly shorter time to the next AE and and repeated detection of EBV DNA, respectively. greater lung function decline than patients without, suggesting After taking into account the exacerbation frequency during that EBV might be associated with bronchiectasis progression. longitudinal follow-up, the annual FEV1 decline differed signif- EBV can be readily detected in blood in >90% of the world’s icantly between patients with and without repeated detection population [7, 20, 21]. Hence, serological assessment was not 6 • ofid • Chen et al A B C D EBV(+): P = .239 80 EBV(+): P = .304 EBV(+): P = .659 80 EBV(+): P = .494 80 EBV repeated detection: P = .660 EBV repeated detection: P = .631 80 EBV repeated detection: P = .252 EBV repeated detection: P = .536 66.7% 63.6% 57.7% 55.9% 60.0% n = 6 n = 14 60 n = 15 60 n = 19 46.9% 45.2% 60 50.0% 51.2% n=6 60 46.2% 47.6% n = 23 45.5% n = 21 Percentage 42.4% 45.2% Percentage n = 20 n = 33 n = 10 41.2% 39.3% n = 24 n = 10 Percentage Percentage n = 14 34.8% n = 24 35.5% 42.5% 40.5% 33.3% 37.5% n = 14 n = 11 35.7% 40 n=8 40 n = 11 n=6 n = 14 n = 15 40 n = 15 n=8 40 30.3% n = 10 20 20 20 20 0 0 0 0 Mild Moderate Severe Mild Moderate ≥80% 50–80% 30–50% <30% 1–6 7–12 13–18 BSI grades E-FACED grades FEV1 % pred HRCT reiff scores E F G H P = .587 P = .543 P = .584 P = .023 8 8 8 8 Log10, EBV copy number Log10, EBV copy number Log10, EBV copy number Log10, EBV copy number 6 6 6 6 4 4 4 4 Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 2 2 2 2 0 0 0 0 Mild Moderate Severe Mild Moderate ≥80% 50–80% 30–50% <30% 1–6 7–12 13–18 BSI grades E-FACED grades FEV1 % pred HRCT reiff scores Figure 3. The rates of detection and repeated detection of EBV DNA (A–D) and EBV loads (E–H) among different severities of bronchiectasis. “ ” indicates the P value for the comparison among 3 groups; “ ” indicates the P value for the comparison between 2 groups. The sputum samples collected during the first stable visits, but not all of the pooled samples, were included for the analyses presented herein. Only 1 patient was graded as having severe bronchiectasis based on E-FACED score. Hence, only patients with mild and moderate bronchiectasis were analyzed and presented. Abbreviation: EBV, Epstein-Barr virus. conducted further herein. Compared with blood, EBV DNA and epithelial cells. Importantly, in some bronchiectasis pa- detection in sputum would more directly reflect the pathogen- tients, EBV DNA could be detected in sputum but not in saliva esis of bronchiectasis because EBV mainly infects lymphocytes or detected with a substantially higher load in sputum than in A Probability of remaining AE-free No RD Probability of remaining AE-free 100 EBV (–) 100 6 P = .549 EBV (+) RD Exacerbations per year 80 P = .024 80 P = .037 4 60 60 40 40 2 20 20 0 0 0 Negative Detection RD 0 60 120 180 240 300 0 60 120 180 240 300 Time of follow-up (days) Time of follow-up (days) Number at risk Number at risk EBV (–) 50 45 36 25 17 12 No RD 55 50 41 32 24 15 EBV (+) 48 39 26 18 11 3 RD 36 29 22 15 11 7 B EBV (–) No RD 1700 EBV (+) 1700 RD 150 FEV1 fall in one year, mL P = .146 1588 1580 1568 1600 1600 1521 1515 1505 FEV1, mL FEV1, mL 100 1500 1500 1508 1497 1400 1454 *P = .014 1400 50 1424 1387 1300 1300 1358 *P = .008 0 1200 1200 Negative Detection RD Baseline 6 months 12 months Baseline 6 months 12 months Figure 4. The association between EBV detection status and longitudinal clinical outomes. Negative: EBV not detected during follow-up visits; Detection: EBV detected at 1 visit during follow-up visits; Repeated detection: EBV detected in at least 2 visits during follow-up within 1 year. “ ” indicates the P value for the comparison among the 3 groups; “ ” indicates the P value for the comparison between 2 groups. EBV (-): EBV not detected at the initial stable visit; EBV (+): EBV detected at the initial stable visit. a Within-group P value with respect to the baseline level. Abbreviation: EBV, Epstein-Barr virus; RD, Repeated detection. EBV Infection in Bronchiectasis • ofid • 7 saliva (Supplementary Table 1). Our observations and previous Macrolides suppress inflammation by strengthening epithelial findings [9, 20] suggest that the high loads of EBV in sputum defense, reducing airway hypersecretion, and accelerating lym- were not due to salivary contamination. phocyte apoptosis [26]. Moreover, macrolides reportedly inhib- We have demonstrated a higher detection rate of EBV DNA ited the growth of EBV-transformed B lymphocytes, providing (48.1%) but not viral loads (log10 load, 4.47) in bronchiectasis further hints on the protective effects of EBV in bronchiectasis compared with healthy subjects. Compared with COPD pa- [27]. Alternatively, patients with more prominent airway in- tients, smokers without COPD yielded a substantially lower de- flammation were less likely to have EBV detection, probably tection rate but similar loads of EBV [9]. Similarly, EBV DNA because of macrolide therapy. In fact, long-acting muscarinic was detected in the sputum of 46% of COPD patients (mean antagonists reportedly increased the risk of EBV detection by log10 load, 4.71) [9] and in the lung tissues of >60% of patients attenuating airway inflammation [28]. with idiopathic pulmonary fibrosis (IPF) [10]. EBV can either Our phylogenetic analyses suggested that EBV was largely be eliminated in immunocompetent individuals or persist (la- homologous among the bronchiectasis population, at least in tent infections) owing to a delicate balance with host immunity the Chinese population. Our findings call for greater awareness Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 [7, 22]. Patients with bronchiectasis, COPD, and IPF frequently toward the universal latent viral infection that drives bronchi- have immune dysfunction, which provides a favorable environ- ectasis progression. ment for EBV [3, 23]. Our findings are clinically relevant because EBV might Detection of EBV was not associated with the onset of AE. drive the progression of bronchiectasis. Inclusion of EBV de- Neither the detection rate of EBV DNA nor the viral load de- tection in clinical assessment may help indicate host immu- creased significantly during AE. Overall, the EBV loads were nity. Nevertheless, our study recruited patients from a single constant despite a temporary nonsignificant reduction at AE center. Comparison of EBV loads in spontaneous vs induced onset. This was true when viewing the difference between sputum is not ideal because shedding of epithelial cells differs stable-state and AE onset. Mechanisms underlying the greater between bronchiectasis and health. The EBV loads might also PCR positivity at AE onset are unclear given the small sample have been confounded by socioeconomic status, which was not size. Consistently, the log10 load was lower during AEs (from thoroughly taken into account for adjustment. The duration 4.71 to 4.08) in COPD [9]. Moreover, the EBV load pro- of our study was relatively short, and hence an extended fol- gressively recovered during convalescence and, around day low-up investigation would be needed to fully unravel the asso- 4 postexacerbation, approached pre-AE stable visit levels. ciation between EBV DNA detection and the clinical outcomes Hence, EBV was not eradicated from airways throughout AE. of bronchiectasis. Furthermore, no sputum lymphocyte count Interestingly, EBV load was negatively correlated with airway was conducted to provide further evidence of EBV infection. inflammation (possibly because of competition with other We cannot address why EBV detection status switched during pathogenic microorganisms), partly explaining why EBV loads follow-up, nor differentiate latent from active EBV infection. decreased during AE when inflammation was heightened. The EBV loads might have been affected by inhaled steroid use. Patients with detection and repeated detection of EBV DNA It remains unclear if the higher EBV detection rates and viral yielded faster lung function decline and a shorter time to the loads were surrogates of epithelial cell turnover. Mechanistic next AE. The precise mechanism remains unknown; however, studies revealing how EBV is acquired are needed. the greater number of bronchiectatic lobes might have ac- In conclusion, the results of this study indicate that patients counted for this finding. EBV LMP-1 positivity was associated with detection of EBV DNA have a shorter time to bronchiec- with more rapid disease progression in IPF [10, 24], possibly tasis exacerbations and that EBV may contribute to the progres- because EBV LMP-1 expression could elicit an altered inflam- sion of bronchiectasis. matory response and impairment of tissue repair that resulted in lung fibrosis [24]. Supplementary Data Supplementary materials are available at Open Forum Infectious Diseases We noted that there was a trade-off between the EBV loads online. Consisting of data provided by the authors to benefit the reader, and airway inflammatory responses. The precise mechanisms the posted materials are not copyedited and are the sole responsibility remain unclear, but the defective host defense and heightened of the authors, so questions or comments should be addressed to the airway inflammation as a consequence of airway surface secre- corresponding author. tory IgA deficiency have been associated with latent EBV infec- Acknowledgments tion. Latent EBV infection has also been associated with CD4+ Financial support. This work was supported by the National and CD8+ lymphocyte infiltration and airway remodeling [25]. Natural Science Foundation (No. 81870003), Guangdong Natural Treatment with low-dose macrolides protected from, whereas Science Foundation (No. 2019A1515011634), and Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme 2017 to inhaled corticosteroids predisposed to, EBV detection. Inhaled Prof. Guan; as well as the Impact and Mechanisms of Physical, Chemical corticosteroids may attenuate airway inflammation, which may and Biological Interventions on the Development and Outcome of Acute help partially elucidate the increased risk of EBV detection. Lung Injury (No. 81490534), National Key Technology R&D Program (No. 8 • ofid • Chen et al 2018YFC1311902), and Guangdong Science and Technology Foundation 11. Polverino E, Goeminne PC, McDonnell MJ, et al. European Respiratory Society (No. 2019B030316028) to Prof. Zhong. guidelines for the management of adult bronchiectasis. Eur Respir J 2017; Potential conflicts of interest. All authors have no potential conflicts of 50:1700629. 12. Guan WJ, Gao YH, Xu G, et al. Aetiology of bronchiectasis in Guangzhou, interest with any companies/organizations whose products or services may Southern China. Respirology 2015; 20:739–48. be discussed in this article. All authors have submitted the ICMJE Form for 13. Reiff DB, Wells AU, Carr DH, et al. CT findings in bronchiectasis: limited value Disclosure of Potential Conflicts of Interest. Conflicts that the editors con- in distinguishing between idiopathic and specific types. Am J Roentgenol 1995; sider relevant to the content of the manuscript have been disclosed. 165:261–7. Author contributions. Chen CL, Guan WJ, and Zhong NS participated 14. Miller MR, Hankinson J, Brusasco V, et al; ATS/ERS Task Force. Standardisation in study design; Chen CL performed laboratory experiments, data analysis, of spirometry. Eur Respir J 2005; 26:319–38 and interpretation; Chen CL, Li HM, Yuan JJ, Huang Y, Guan WJ, Han XR, 15. Chalmers JD, Goeminne P, Aliberti S, et al. The bronchiectasis severity index. An Chen RC, and Zhong NS recruited patients; Chen CL, Li HM, Yuan JJ, and international derivation and validation study. Am J Respir Crit Care Med 2014; 189:576–85. Huang Y performed follow-up; Chen CL, Guan WJ, de la Rosa-Carrillo D, 16. Martinez-Garcia MA, Athanazio RA, Girón R, et al. Predicting high risk of ex- and Martinez-Garcia MA drafted the manuscript; de la Rosa-Carrillo D, acerbations in bronchiectasis: the E-FACED score. Int J Chron Obstruct Pulmon Martinez-Garcia MA, Guan WJ, and Zhong NS were responsible for study Dis 2017; 12:275–84. conception and provided critical review of the manuscript. de la Rosa- 17. Guan WJ, Gao YH, Xu G, et al. Sputum bacteriology in steady-state bronchiec- Carrillo D, Martinez-Garcia MA, Chen CL, Li HM, Yuan JJ, Huang Y, Guan tasis in Guangzhou, China. Int J Tuberc Lung Dis 2015; 19:610–9. WJ, Han XR, Chen RC, and Zhong NS approved the final draft for publica- 18. Hill AT, Haworth CS, Aliberti S, et al; EMBARC/BRR definitions working group. Downloaded from https://academic.oup.com/ofid/article/7/8/ofaa235/5859613 by guest on 18 October 2023 tion. Guan WJ was the guarantor of the study. Pulmonary exacerbation in adults with bronchiectasis: a consensus definition for clinical research. Eur Respir J 2017; 49:1700051. 19. George SN, Garcha DS, Mackay AJ, et al. Human rhinovirus infection during nat- References urally occurring COPD exacerbations. Eur Respir J 2014; 44:87–96. 1. Amalakuhan B, Maselli DJ, Martinez-Garcia MA. Update in bronchiectasis 2014. 20. Grant WB. A possible role for Epstein-Barr virus infection in COPD? Eur Respir Am J Respir Crit Care Med 2015; 192:1155–61. J 2008; 32:1412–3; author reply 1413. 2. Nicotra MB, Rivera M, Dale AM, et al. Clinical, pathophysiologic, and micro- 21. Smatti MK, Al-Sadeq DW, Ali NH, et al. Epstein-Barr virus epidemiology, se- biologic characterization of bronchiectasis in an aging cohort. Chest 1995; rology, and genetic variability of LMP-1 oncogene among healthy population: an 108:955–61. update. Front Oncol 2018; 8:211. 3. Boyton RJ, Altmann DM. Bronchiectasis: current concepts in pathogenesis, im- 22. Ressing ME, van Gent M, Gram AM, et al. Immune evasion by Epstein-Barr virus. munology, and microbiology. Annu Rev Pathol 2016; 11:523–54. Curr Top Microbiol Immunol 2015; 391:355–81. 4. Gao YH, Guan WJ, Xu G, et al. The role of viral infection in pulmonary exacerba- 23. Chalmers JD, Hill AT. Mechanisms of immune dysfunction and bacterial persist- tions of bronchiectasis in adults: a prospective study. Chest 2015; 147:1635–43. ence in non-cystic fibrosis bronchiectasis. Mol Immunol 2013; 55:27–34. 5. Saetta M, Baraldo S, Corbino L, et al. CD8+ve cells in the lungs of smokers 24. Heukels P, Moor CC, von der Thüsen JH, et al. Inflammation and immunity in with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999; IPF pathogenesis and treatment. Respir Med 2019; 147:79–91. 160:711–7. 25. Polosukhin VV, Cates JM, Lawson WE, et al. Bronchial secretory immunoglob- 6. Silva JR, Jones JA, Cole PJ, Poulter LW. The immunological component of the ulin a deficiency correlates with airway inflammation and progression of chronic cellular inflammatory infiltrate in bronchiectasis. Thorax 1989; 44:668–73. obstructive pulmonary disease. Am J Respir Crit Care Med 2011; 184:317–27. 7. Thorley-Lawson DA. Epstein-Barr virus: exploiting the immune system. Nat Rev 26. Kanoh S, Rubin BK. Mechanisms of action and clinical application of macrolides Immunol 2001; 1:75–82. as immunomodulatory medications. Clin Microbiol Rev 2010; 23:590–615. 8. Lung ML, Lam WK, So SY, et al. Evidence that respiratory tract is major reservoir 27. Majewski M, Korecka M, Kossev P, et al. The immunosuppressive macrolide RAD for Epstein-Barr virus. Lancet 1985; 1:889–92. inhibits growth of human Epstein-Barr virus-transformed B lymphocytes in vitro 9. McManus TE, Marley AM, Baxter N, et al. High levels of Epstein-Barr virus in and in vivo: a potential approach to prevention and treatment of posttransplant COPD. Eur Respir J 2008; 31:1221–6. lymphoproliferative disorders. Proc Natl Acad Sci U S A 2000; 97:4285–90. 10. Kelly BG, Lok SS, Hasleton PS, et al. A rearranged form of Epstein-Barr virus 28. Trevethick M, Clarke N, Strawbridge M, Yeadon M. Inhaled muscarinic antagon- DNA is associated with idiopathic pulmonary fibrosis. Am J Respir Crit Care Med ists for COPD—does an anti-inflammatory mechanism really play a role? Curr 2002; 166:510–3. Opin Pharmacol 2009; 9:250–5. EBV Infection in Bronchiectasis • ofid • 9