Search

Chronic lung allograft dysfunction (CLAD) is the leading long-term contributor to mortality after lung transplant. The lung microbiome predicts subsequent development of CLAD, but whether lung bacteria differ between CLAD phenotypes is unknown. Using a biorepository of acellular bronchoalveolar lavage (BAL) fluid, we identified specimens which were collected within 90 days of CLAD onset. CLAD phenotype was assigned based on the presence of absence of obstruction, restriction, and computed tomography (CT) scan opacities, in accordance with ISHLT guidelines. Bacterial DNA burden was measured with BioRad QX200 Droplet Digital PCR and 16S rRNA gene sequencing was performed using the Illumina MiSeq platform. Bacterial burden was compared using Wilcoxon rank-sum. Community composition was compared using PERMANOVA and mvabund (a model-based approach to analysis of multivariable abundance data). Eighty (80) patients had a BAL specimen from near CLAD onset available for analysis. Of these, 41 (51%) had BOS, 13 (16%) had RAS, 6 (8%) had Mixed CLAD, and 20 (25%) had an undefined/unclassifiable phenotype. There were no differences in lung bacterial burden (p =0.56, Panel A) or overall community composition between CLAD phenotypes (p =0.53, Panel C). Shannon diversity index did differ between CLAD phenotypes, with lower average diversity observed in the patients with an unclassifiable CLAD phenotype (1.0 ± 0.4) vs. BOS (2.2 ± 0.8), RAS (2.1 ± 0.8), Mixed (2.1 ± 0.7), or undefined CLAD (2.1 ± 0.4, overall p =0.02, Panel B). This difference persisted even after accounting for clinical evidence of infection (BAL neutrophilia and bacterial culture results did not differ between groups). Lung microbiome characteristics are largely similar across CLAD phenotypes, although lower within-specimen diversity is seen in patients with unclassified CLAD. [ABSTRACT FROM AUTHOR]

Chronic lung allograft dysfunction (CLAD) is the leading long-term cause of mortality after lung transplant. The lung microbiome in healthy lung transplant recipients predicts subsequent development of CLAD, but whether lung bacteria differ between CLAD and healthy allografts is unknown. We performed a case control study using a biorepository of acellular bronchoalveolar lavage (BAL) fluid. Controls were specimens collected from asymptomatic patients undergoing surveillance bronchoscopy one-year post-transplant; cases were specimens collected within 90 days of CLAD onset. Bacterial DNA burden was measured using BioRad QX200 Droplet Digital PCR and 16S rRNA gene sequencing was performed using Illumina MiSeq. We identified 131 specimens which met criteria for use as CLAD-free controls and 78 specimens which met critieria for use as CLAD cases. The lung microbiome in specimens collected at CLAD vs. CLAD-free controls was characterized by higher bacterial burden (median [IQR]16S gene copies per mL 4991 [2889-13075] vs. 3561 [2426-8131], p =0.028, Panel A). There were significant overall community composition differences using both PERMANOVA (p =0.032) and mvabund , a model-based approach to analysis of multivariable abundance data (p =0.014, Panel B). As shown in biplot analysis, lung bacterial communities from patients with CLAD had relatively more oral flora than CLAD-free controls; although no individual taxa were responsible for these overall community composition differences. Twenty-two patients had BAL available for microbiome analysis at both time points. In paired comparisons, there was no intra-patient differences in bacterial burden, community diversity or community composition between CLAD-free and CLAD onset samples. The lung microbiome in patients with CLAD is characterized by higher bacterial burden and communities with increased predominance of taxa commonly found in the upper respiratory tract. [ABSTRACT FROM AUTHOR]