This ambispective longitudinal cohort study (NCT04803695) of patients with BE was conducted at a tertiary care hospital, and research laboratories. We included adult patients (>18 years) diagnosed with BE of any aetiology by high-resolution computerized tomography (HRCT) of the chest and with at least one previous PA isolation in sputum 1 year before inclusion. Patients were enrolled from May 2017 to November 2019. Written informed consent was obtained from all patients and the study was approved by the Internal Review Board of the Hospital (registry number HCB/2018/0236).

Exacerbation was defined as a deterioration in three or more of the following symptoms: cough, sputum volume and/or consistency, sputum purulence, dyspnoea and/or exercise tolerance, fatigue or general malaise, and hemoptysis.17Chronic PA infection was defined as two or more isolates separated by 3 months in a single year according to published guidelines.2 We included patients who met the definition of chronic infection at inclusion based on their clinical histories including microbiology (retrospective) and/or those whose cultures (by SOCT and/or ESCMID methods) at follow-up accomplished the definition of chronic infection (prospective). Patients who neither retrospectively nor prospectively did not met the definition of chronic infection were allocated in the intermittently infected group.

Demographics and clinical data were recorded at baseline. Patients attended the clinic once every 3 months during the stable phase throughout the study period. At each visit: (i) spontaneous sputum and blood samples were taken (also at exacerbations within the first 48h); (ii) lung function was assessed; and (ii) the Quality-of-Life Bronchiectasis questionnaire was completed. Patients were asked to call a 24/7 phone line if they experienced any of the symptoms of an exacerbation. The number of mild and hospitalized exacerbations was recorded prospectively and also 3 years retrospectively based on the clinical history (Table 1).

Each sputum sample was collected to assess the following: (1) SOCT and Gram Staining, (2) ESCMID recommended cultures (Recommended) for biofilm infections and Gram staining for detecting alginate presence, (3) Fluorescence in situ hybridization (FISH)+Confocal microscopy, (4) Microbiome analysis, and (5) Rheological properties of mucus. Only sputum samples of Murray-Washington classification degrees IV, V or VI, which are considered of good quality, were included in the analyses.18 SOCT and ESCMID recommended cultures are described in Fig. 1C and in the online data supplement.10,14P. aeruginosa in sputum smears was identified by FISH using the PA-specific PNA probe32 (AdvanDx, USA). Images of PA biofilms were obtained using a confocal microscope equipped with appropriate filters, as described elsewhere.10 For all imaging studies we analyzed≥2 sputum smears per patient obtained 3 months apart.

Fig. 1.

Comparison between microbial diagnosis by Standard-of-Care Tests (SOCT) and recommended cultures for biofilm infections. (A) Differences between SOCT and recommended methods in diagnosing P. aeruginosa (PA) presence; (B) Differences between SOCT and recommended methods in PA mucoid phenotype; (C) Microbiology following standard or recommended methods for sputum samples. SOCT+/Recommended+: samples that tested positive for PA by both SOCT and recommended methods; SOCT−/Recommended+: samples that tested positive for PA by recommended methods but negative for PA by SOCT; Recommended+/16S+: samples that tested positive for PA by recommended and 16S methods; Recommended−/16S+: samples that tested positive for PA by recommended methods but negative for PA by SOCT. DTT: dithiothreitol, FISH: fluorescent in situ hybridization, CLSM: confocal laser scanning microscopy, MALDI-TOF: matrix-assisted laser desorption/ionization time-of-flight mass spectrometry.

(0.34MB).

For microbiome analyses we followed methods previously reported.19 16S profiling was compared between chronically infected patients with the mucoid PA versus intermittently infected patients with the non-mucoid PA. Rheological properties of mucus were measured using a 35-mm serrated parallel-plate rheometer (Haake Rheostress1, ThermoFisher, MA, US) as reported elsewhere.20 IgG against the O-Antigen of P. aeruginosa (Anti-PA IgG) were analyzed in serum samples using ELISA at the Department of Microbiology, Rigshospitalet, University of Copenhagen, Denmark, as described previously.10 Lung function was evaluated using the EasyOne™ World Spirometer (NDD Medical Technologies, Zurich, Switzerland), and classified according to the American Thoracic Society/European Respiratory Society Guidelines. Differences in Health-Related Quality of Life (HRQoL) (assessed with the Quality of Life Questionnaire-Bronchiectasis (QOL-B) v3.1) and the minimal important difference (MID) scores were contrasted in both the first visit and the last visit after 1 year follow-up by infection status (intermittent vs. chronic).

Categorical and continuous variables were reported as percentages or as mean±standard deviation (SD) or median (IQR), depending on data distribution. Group comparisons were conducted using the Chi-squared test for categorical variables and the Student's t-test or Mann–Whitney U test for continuous variables. Paired samples were analyzed using the paired t-test or Wilcoxon signed-rank test. A multivariable logistic regression model was used to identify independent factors associated with chronic PA status, with results expressed as odds ratios (ORs) and 95% confidence intervals (CIs). Model performance was assessed using the Hosmer–Lemeshow test and the area under the ROC curve (AUC). Statistical significance was set at p<0.05. Analyses were conducted using SPSS Statistics 25.0. Complete statistical methods are provided in the Supplemental Material.

During enrolment a total of 666 BE patients were screened for PA-positive sputum. PA isolates were found in the samples of 80 of these patients, who formed our study population. Of these 80, 19 (24%) met the definition of intermittently infected and 61 (76%) were chronically infected. Of the patients with chronic infection, 34 (56%) had been colonized by PA for a median [interquartile range (IQR)] of 7.00 [3.75–12.00] years and 27 (44%) were newly diagnosed during the course of the study, being colonized for 0.5 [0.5–0.5] years (Fig. 1S). The baseline clinical characteristics of the study population are summarized in Table 1. At study inclusion, patients with intermittent or chronic infection status did not differ in terms of age, gender, aetiology of BE, predicted FEV1%, FACED score or Bronchiectasis Severity Index (BSI), nor the number of previous exacerbations. In contrast, higher PA load, AntiPA-IgG and presence of alginate and mucoid phenotype (biofilms) were significantly associated to patients with chronic infection. In intermittently infected patients, mucoid PA was barely found (Table 1).

Two hundred and thirty-five sputum samples of good quality, with a median of 3.00 [2.00–5.00] per patient, were analyzed using both Standard-of-Care Tests (SOCT) and ESCMID (recommended) cultures immediately after sample collection (Fig. 1C).

Overall, PA isolation was higher when using the ESCMID recommended methods than with SOCT (74% vs. 44%, p<0.001). Furthermore, the ESCMID recommended methods for biofilm infections allowed detection of PA in 11 (38%) of the 29 exacerbations that gave false-negative results using SOCT, and diagnosis of 18 (66%) of the 27 chronic infections that emerged during the study period, compared to nine (33%) that would have been diagnosed using SOCT alone (Fig. 1A).

The frequency of mucoid PA detected in SOCT was significantly lower than that detected in the recommended cultures (41% vs. 71%, Fig. 1B). Mucoids compared to non-mucoids needed two days or more to growth (2.00 [2.00–4.00] vs. 2.00 [1.00–2.00], p<0.001, respectively), but SOCT culture plates are typically discarded after two days. The sensitivity of ESCMID Recommended cultures and Gram staining methods was superior compared to SOCT (Fig. 3S).

In a subgroup analysis (n=51) using 16S rRNA gene sequencing as a culture-independent technique control, in terms of PA isolation high and similar sensitivity was found between ESCMID recommended cultures and 16S but not between SOCT and 16S (Figs. 3S A and 3S B).

The presence of PA alginate and biofilms in sputum was associated with the mucoid PA phenotype (Fig. 4S) and chronic PA (Table 1). The presence of PA biofilms in sputum from BE patients was demonstrated by FISH and confocal laser scanning microscopy, and resembled to that of the PA biofilms shown in CF patients (Fig. 2).10

Fig. 2.

Demonstration of P. aeruginosa (PA) biofilms by Gram staining and confocal laser scanning microscopy (CLSM) and FISH and its added diagnostic value. First demonstration of PA biofilms by Gram staining microscopy (left lung) and by CLSM+FISH in patients with non-Cystic Fibrosis bronchiectasis. Pseudomonas aeruginosa biofilms are tolerant of the host immune response: see polymorphonuclear leukocytes surrounding PA biofilms. The sputum in the Gram staining image (63× oil immersion objective) was obtained from a patient with bronchiectasis and chronic PA infection for 12 years (mean PA load: 5.38LogCFU/mL, mean precipitating PA antibodies: 21.25μ/mL and relative abundance of 42.21% of PA). In the CLSM+FISH (63× oil immersion objective) image, a specific Pseudomonas aeruginosa probe (red fluorescence) shows PA in biofilm aggregates and polymorphonuclear leukocytes (blue fluorescence) surrounding PA biofilms. The sputum of the CLSM+FISH picture was obtained from a patient with bronchiectasis and chronic PA infection for 12 years (mean PA load: 5.61LogCFU/mL, mean precipitating PA antibodies: 31.86μ/mL and relative abundance of 13.35% of PA).

(0.45MB).

Chronically infected patients with the mucoid PA had a significant reduction in lung alpha and beta biodiversity compared to intermittently infected patients with the non-mucoid PA (Fig. 3A). Community compositional dissimilarity among Chronically infected with the mucoid (red dots) and Intermittendly infected with the non-mucoid (blue dots) showed significant compositional differences between them (R2=0.21, pAdonis=0.001) (Fig. 3B).

Fig. 3.

Diversity of respiratory microbiome and linear discriminant analysis effect size (LefSE) clustered by mucoid or non-mucoid P. aeruginosa (PA). (A) Alpha diversity, measured by Chao1, Fisher and Shannon's and Simpson's diversity index is plotted for patients with the mucoid phenotype/chronic PA (red) and non-mucoid phenotype/intermittent PA (blue). The line inside the boxplot represents the median. The lowest and highest values within the 1.5 interquartile range are represented by the whiskers. The dots show the outliers and individual sample values. (B) Principal coordinate analysis (PCA) based on Bray–Curtis dissimilarity. (C) The LDA finds taxa that are significantly more abundant in one group. Negative (red bars) LDA scores represent the most abundant bacterial groups in mucoid samples, while positive (green bars) represent those in non-mucoid samples. The bar size represents the effect size of the taxa within each group. Each sample is represented by a dot: blue dots represent the non-mucoid phenotype found in sputum, and red dots represent the mucoid phenotype. Samples with similar compositions appear in clusters; the two groups appear to present compositional differences. (D) Comparisons between Mucoid pathobiome No/Yes in terms of PALoad (CFU/mL), AntiPA-IgG and Years of chronic infection. The mucoid pathobiome was significantly associated with higher PA Load, higher AntiPA-IgG and more years of chronic PA infection.

(0.4MB).

Additionally, using linear discriminant analysis effect size (LEfSe) we determined the respiratory microbiota composition of chronically infected with the mucoid and Intermittendly infected with the non-mucoid. Samples with mucoid PA were dominated by Proteobacteria phyla, with a marked abundance of Pseudomonas and Neisseria at the genus level. In contrast, microbial patterns in the non-mucoid group were more varied, including genera such as Actinobacteria (Actinomyces, Rothia), Bacteroidetes (Prevotella) and Firmicutes (Streptococcus, Veillonella, Lactobacillus, Granulicatella) (Fig. 3C). The mucoid pathobiome of chronically infected patients was associated with higher PA load, higher systemic AntiPA-IgG and more years of chronic infection (Fig. 3D).

In terms of mucus viscoelasticity, the mucoid group (n=47, 76%) presented worse viscoelastic properties at 1rad/s than the non-mucoid group (n=15, 24%) (Table 1S).

We analyzed 262 BE patient serum samples (an average of 4.00 [2.00–4.50] independent samples per patient) collected at follow-up visits. Higher levels of IgG against the O-Antigen of P. aeruginosa (Anti-PA IgG) were detected in chronic than in intermittently infected BE patients (Fig. 4A). Indeed, PA antibody levels increased in a time-dependent manner, being significantly higher in patients with ≥3 years since the diagnosis of chronic infection than in those with <3 years (Fig. 4B). We found a positive correlation between the mean level of systemic Anti-PAIgG and the mean PA load per patient (r:0.409; p<0.001) (Fig. 4C). AntiPA-IgG positively correlated with years of chronic PA infection (r:0.559; p<0.001) (Fig. 4D). Additionally, the systemic levels antiPA-IgG displayed high sensitivity for diagnosing chronic PA infection (Fig. 3S).

Fig. 4.

AntiPA-IgG in serum samples from bronchiectasis (BE) patients. (A) AntiPA-IgG from BE patients with intermittent vs. chronic P. aeruginosa (PA) infection showed significant differences, being higher in the chronic group (median [IQR] 20.23 [9.89–38.33] vs. 3.25 [0.00–14.53], p<0.001 respectively). (B) AntiPA-IgG levels were higher in patients with chronic infection of ≥3 years than in those with chronic infection of <3 years (median [IQR] 30.63 [13.58–49.82] vs. 9.55 [3.18–21.66], p<0.001 respectively]. (C) Simple scatter plot (r:0.409; p<0.001) charting the mean level of AntiPA-IgG and mean PA load. (D) Simple scatter plot (r:0.559; p<0.001) charting the level of AntiPA-IgG up by years of chronic infection.

(0.23MB).

After 1-year follow-up, chronic PA infection was associated with a decrease in hospitalized exacerbations (Fig. 5A). Similarly, higher levels of AntiPA-IgG were significantly associated with lower number of hospitalized exacerbations (Fig. 5B, E), No differences in predicted FEV1% were found between chronically and intermittently colonized groups, either at inclusion (Table 1) or after 1-year follow-up (58.33±27.09 vs. 62.87±19.43, p=0.481 respectively). After 1-year follow-up patients with intermittent PA infection improved in vitality, respiratory symptoms and emotional function domains of QoL. In contrast, chronically infected patients deteriorated in the emotional function and social domains of QoL (Fig. 5F).

Fig. 5.

Clinical and biological outcomes of BE patients with chronic vs intermittent P.aeruginosa colonization. (A) Number of hospitalized exacerbations during follow-up in BE patients with intermittent vs. chronic PA respiratory infections. Chronic PA infection (n=61) was associated with a lower number of hospitalized exacerbations compared to those with intermittent colonization (n=19) (0.00 [0.00–1.00] vs 1.00 [0.00–1.00], p=0.040, respectively). The number of mild exacerbations did not differ between chronic vs. intermittent PA groups (1.00 [1.00–2.50] vs. 1.50 [0.00–3.00], p=0.947). (B) Number of hospitalized exacerbations during follow-up in BE patients with low (<5) vs. high (≥5) systemic AntiPA-IgG. Patients with high AntiPA-IgG (n=61) presented fewer hospitalized exacerbations than those with low AntiPA-IgG (n=16) (0.00 [0.00–1.00] vs 1.00 [0.00–1.75], p=0.045). The number of mild exacerbations did not differ between patients with high vs. low AntiPA-IgG (1.50 [1.00–3.00] vs. 1.50 [1.00–2.50], p=0.685 respectively). C and D) Principal Component Analysis (PCA) analysis plot to assess the dissimilarity in relevant variables (years of chronic infection, number of exacerbations (previous and follow-up, mild and hospitalized), FEV1, mean PA load, Anti-PA-IgG) between chronic (green) and intermittent (red) patients (C) or mucoid (green) and non-mucoid (red) P. aeruginosa isolation in sputa (D) or between females and males (Fig. 3S). NB: Circled Patients with chronic PA status but with non-mucoid PA in sputa are circle. Shaded ellipses indicate the confidence intervals. (E) Principal Component Analysis (PCA) plot showing the multivariate variation among 80 patients in terms of clinical variables. The plot shows the interplay between all clinical variables included in this multivariate analysis. Vectors indicate the direction and contribution of each clinical variable to the overall distribution. Coloured symbols correspond to the five most contributing variables identified in this study. The two principal axes explained the variance, in Dim1 (38.1%) and Dim2 (28.6%), respectively. Positively correlated variables are grouped together when the angle between them is less than 90°. In contrast, variables are non-correlated or negatively correlated when positioned on opposing quadrants of the plot and the angle between them is above 90° (contiguous or opposing quadrants). (F) Quality of life domains with significant changes after the 12-month follow-up period by colonization status, where the mean is marked with a red diamond. Sixty-nine patients were included in the QoL assessment: 49 (71%) with chronic PA colonization vs. 20 (29%) intermittent. When assessing the minimal important difference, we observed improvements in vitality (mean differences between visits 1 and 4 were 19.70, greater than the threshold of 10 points), emotional function (mean differences=12.12, threshold of 7 points), and respiratory symptoms (mean differences=9.09, threshold 8 points) in patients with intermittent colonization, whereas patients with chronic colonization recorded a loss of QoL in emotional (mean differences=−7.84, threshold 7 points) and social function (mean differences=−11.76, threshold 9 points). Additionally, the chronic vs. intermittent emotional function domain showed significant impairment at the 1-year follow-up (median 0.00 [0.00–33.33] vs. 0.00 [−16.67–0.00], p=0.036). (G) Intermittent versus chronic Pseudomonas aeruginosa (PA) colonization. This figure illustrates the biological changes during the transition from intermittent to chronic PA colonization status. Intermittent PA colonization (left) is characterized by the non-mucoid phenotype growing in planktonic form indicated by the PA flagellum, microbiota biodiversity in lungs, low systemic levels of AntiPA-IgG and normal viscoelastic propierties of mucus. As time with intermittent colonization progresses without successful PA eradication, the colonization becames chronic (early stage) in which non-mucoid PA load increases although few mucoid PA colonies may be present and a trend to higher AntiPA-IgG is observed than in intermittent colonization. Finally, chronic PA colonization (right) is characterized by increased PA load and increased presence of the mucoid phenotype growing in biofilms (without flagellum). When chronic PA has been established the low biodiversity lung microbiota, with a predominance of Proteobacteria, and the increased mucus viscoelasticity complicates mucus clearance. In this context, systemic AntiPA-IgG levels are higher than in intermittent colonization.

Further, multivariable analyses showed that PA load and the mucoid phenotype were factors independently associated to chronic infection, but not the number of previous exacerbations (Table 2).

Despite two main clusters were found according to infection status (Fig. 5C, chronic or intermittent) or to PA phenotype (Fig. 5D, non-mucoid or mucoid), highlighted in the circle are the 25 patients with chronic/non-mucoid PA (Fig. 5D) who, interestingly, presented lower PA load (LogCFU/mL) (4.20 [0.00–5.67] vs 5.38 [4.84–6.10], p=0.015 respectively), fewer years of chronic infection (1.0 [0.5–4.5] vs. 5.0 [0.5–12.0], p=0.014 respectively), and a trend towards lower levels of antiPA-IgG (13.80 [8.28–25.95] vs 26.00 [11.48–42.90], p=0.184 respectively) compared to those with chronic/mucoid PA. Overall, these findings elucidate important factors involved in the transition from intermittent to chronic infection status and displayed in Fig. 5G.