(524) Striking a BAL-NS : combining metagenomics with culture methods shows nasal swabs to be a good proxy for broncho-alveolar lavage in microbial / pathogen surveillance

Background: Airway surveillance is standard practice in clinical cystic fibrosis (CF) management, where sampling of the respiratory tract combined with microbial culture-work provide an early-warning system against incumbent pathogens, allowing appropriate prophylactic or therapeutic measures to be taken.

There is no agreement on the most appropriate sampling method in young children with CF. Broncho-Alveolar Lavage (BAL) is the gold standard sampling method to survey the CF lung microbiome but is invasive and requires the child to undergo general anaesthesia. Common non-invasive alternatives to BAL include nasal swabs (NS) or throat swabs (TS). Although easier to employ in a clinical setting, these methods lack comparative evaluation for their efficacy with respect to BAL in monitoring the respiratory bioburden in CF.

Methods: Using a healthy Irish paediatric CF cohort, we compared BAL, NS, and TS samples using both culture characterisation (CC) and high-throughput metagenomic sequencing (MGX; taxonomic assignment: Kaiju). We evaluated not only the respiratory microbiome, but also the efficacy with which these different sampling (BAL, NS, TS) and surveillance methods (CC, MGX) serve to monitor the microbiology of the lower respiratory tract. Finally, we evaluated how well CC matches MGX in surveying pathogens of clinical significance.

Results: Across the study cohort (samples=86, participants=30, age 1-6yrs), we showed that MGX-NS is highly similar to MGX-BAL, with both MGX-BAL and MGX-NS corresponding equally well to CC-BAL. However, we also found that direct agreement between CC and MGX pathogen surveillance is low for all sampling types, and accuracy varied both with the pathogen under consideration and with the MGX detection cut-offs applied. MGX-TS presented a distinct and significantly more diverse microbiome relative to MGX-BAL/NS. Major pathogens were ubiquitous and abundant in the lung and nares (MGX-BAL/ NS: Staphylococcus, Escherichia) as well as in the throat (MGX-TS: Streptococcus) across the dataset, leading to an artificially high positive predictive value (PPV) and low negative predictive value (NPV) in culture surveillance. Other pathogens appeared rarely (Achromobacter, Chryseobacterium, Haemophilus, Moraxella), leading to a low PPV and high NPV.

Conclusions: Here we used metagenomic sequencing to show that BAL and NS samples are highly similar in their total microbial composition, and as a result performed equally well in proxying for culture-based surveillance. This advocates for the use of NS in routine microbial monitoring of paediatric CF if employing metagenomic surveillance, and suggests that matter collected by NS is a strong proxy for the more invasive BAL-based microbial culture surveillance. There is nevertheless a clear disjunct between culture-based and metagenomic-based surveillance, which remains to be resolved. Until this is possible, the two methods must be considered complementary.

Acknowledgements: We would like to thank all of our study participants and parents for their time and dedication to this study, the entire SAINTs team, and our collaborators at Teagasc and clinical sites. Funding for this study was provided by the National Children’s Research Centre within the Children’s Clinical Research Unit at Children’s Health Ireland Crumlin, and the MicrobiomeSupport program (EU H2020, grant 818116).

References: