Lung microbiota and bacterial abundance in patients with bronchiectasis when clinically stable and during exacerbation

Abstract

Rationale: Characterization of bacterial populations in infectious respiratory diseases will provide improved understanding of the relationship between the lung microbiota, disease pathogenesis, and treatment outcomes.

Objectives: To comprehensively define lung microbiota composition during stable disease and exacerbation in patients with bronchiectasis.

Methods: Sputum was collected from patients when clinically stable and before and after completion of antibiotic treatment of exacerbations. Bacterial abundance and community composition were analyzed using anaerobic culture and 16S rDNA pyrosequencing.

Measurements and main results: In clinically stable patients, aerobic and anaerobic bacteria were detected in 40 of 40 (100%) and 33 of 40 (83%) sputum samples, respectively. The dominant organisms cultured were Pseudomonas aeruginosa (n = 10 patients), Haemophilus influenzae (n = 12), Prevotella (n = 18), and Veillonella (n = 13). Pyrosequencing generated more than 150,000 sequences, representing 113 distinct microbial taxa; the majority of observed community richness resulted from taxa present in low abundance with similar patterns of phyla distribution in clinically stable patients and patients at the onset of exacerbation. After treatment of exacerbation, there was no change in total (P = 0.925), aerobic (P = 0.917), or anaerobic (P = 0.683) load and only a limited shift in community composition. Agreement for detection of bacteria by culture and pyrosequencing was good for aerobic bacteria such as P. aeruginosa (κ = 0.84) but poorer for other genera including anaerobes. Lack of agreement was largely due to bacteria being detected by pyrosequencing but not by culture.

Conclusions: A complex microbiota is present in the lungs of patients with bronchiectasis and remains stable through treatment of exacerbations, suggesting that changes in microbiota composition do not account for exacerbations.

Figures

Figure 1.

Comparison of total viable counts per gram sputum of aerobic and anaerobic bacteria cultured from sputum samples collected from patients with clinically stable bronchiectasis receiving (n = 16) and not receiving (n = 24) chronic maintenance antibiotic therapy. Although not statistically significant, total (P = 0.064, Mann-Whitney test), aerobic (P = 0.060), and anaerobic (P = 0.070) counts were less in patients receiving chronic maintenance therapy.

Figure 2.

Total viable counts per gram sputum of aerobic and anaerobic bacteria cultured from sputum samples collected from patients with bronchiectasis at the start of treatment (n = 14) and end of treatment (n = 14) for an exacerbation and when clinically stable after exacerbation (n = 13). Aerobic bacteria were present in significantly greater numbers than anaerobic bacteria at the end of antibiotic treatment (median difference in total viable count, 1.9 × 107 cfu/g [95% CI, 1.41 × 107, 5.15 × 107]; P = 0.001, Wilcoxon signed-rank test) and when the same patients were clinically stable (median difference in total viable count, 6.0 × 107 cfu/g [95% CI, 3.2 × 107, 15.0 × 107]; P = 0.001) but less so at the start of antibiotic treatment (median difference in total viable count, 4.3 × 106 cfu/g [95% CI, 1.2 × 107, 2.0 × 107]; P = 0.064). *Comparing two groups using Wilcoxon signed-rank test; ^comparing three groups using Friedman test.

Figure 3.

(a) Comparison of the percent abundance of the major identified phyla in pooled samples collected from patients when clinically stable (cross-sectional study, n = 10) and at the start of treatment for an exacerbation (longitudinal study, n = 11). Similar patterns of phyla distribution were observed in both groups. OTUs = operational taxonomic units. (b) Box plot comparison of microbial diversity (Shannon-Wiener diversity index) in samples from patients when clinically stable (cross-sectional study, n = 10) and at the start of treatment for an exacerbation (longitudinal study, n = 11), where higher values correspond to higher diversity. The top and bottom boundaries of each box indicate 75th and 25th quartile values, respectively, with the black line inside each box representing the median (50th quartile). The ends of the whiskers indicate the range. No significant difference (P = 0.105, Mann-Whitney test) in microbial community diversity is apparent between the two groups.

Figure 4.

Pairwise comparison of Bray-Curtis distances for 29 sputum samples from 21 patients with bronchiectasis, using nonmetric multidimensional scaling (nMDS) analysis. Considerable overlap between communities in samples from patients in each group can be discerned, with differences in community composition not reflecting to which cohort an individual patient/sample belonged. A limited shift in community composition within individuals after antibiotic treatment was also apparent, with samples from individual patients clustering closely together. Clinically stable samples (C), n = 10; exacerbation samples, n = 19 (start of treatment [SOT], n = 11; end of treatment [EOT], n = 5; stable after exacerbation [ST], n = 3). Red boxes group together samples collected from the same patient at various time points.

Figure 5.

Comparison of the percent abundance of the major identified phyla in samples collected from patients at the start of treatment (SOT, n = 5) and end of treatment (EOT, n = 5) of an acute infective exacerbation and when clinically stable after exacerbation (ST, n = 3). Abundance of the predominant bacteria decreased slightly after treatment, with a similar small increase in the abundance of unclassifiable bacteria and obligate anaerobes such as Veillonella and Prevotella.

Figure 6.

Box plot comparison of microbial diversity (Shannon-Wiener diversity index) in samples dominated by the genera Haemophilus (n = 10), Pseudomonas (n = 8), and Streptococcus (n = 6), where higher values correspond to higher diversity. The top and bottom boundaries of each box indicate 75th and 25th quartile values, respectively, with the black line inside each box representing the median (50th quartile). The ends of the whiskers indicate the range. There was no significant difference in diversity between the three groups (median Shannon-Wiener diversity index [interquartile range]: Haemophilus, 0.63 [0.25, 0.90]; Pseudomonas, 1.17 [0.79, 1.40]; Streptococcus, 1.51 [0.19, 2.35]; P = 0.183, Kruskal-Wallis test).