Complexity, temporal stability, and clinical correlates of airway bacterial community composition in primary ciliary dyskinesia.

Primary ciliary dyskinesia (PCD) is a genetic disease characterized by abnormalities in ciliary function, leading to compromised airway clearance and chronic bacterial infection of the upper and lower airways. The compositions of these infections and the relationships between their characteristics and disease presentation are poorly defined. We describe here the first systematic culture-independent evaluation of lower airway bacteriology in PCD. Thirty-three airway samples (26 from sputum, 7 from bronchoalveolar lavage [BAL] fluid) were collected from 24 PCD patients aged 4 to 73 years. 16S rRNA quantitative PCR and pyrosequencing were used to determine the bacterial loads and community compositions of the samples. Bacterial loads, which ranged from 1.3 × 10(4) to 5.2 × 10(9) CFU/ml, were positively correlated with age (P = 0.002) but not lung function. An analysis of ∼7,000 16S rRNA sequences per sample identified bacterial species belonging to 128 genera. The concurrently collected paired samples showed high bacterial community similarity. The mean relative abundance of the dominant genera was 64.5% (standard deviation [SD], 24.5), including taxa reported through standard diagnostic microbiology (members of the genera Pseudomonas, Haemophilus, and Streptococcus) and those requiring specific ex vivo growth conditions (members of the genera Prevotella and Porphyromonas). The significant correlations observed included a positive relationship between Pseudomonas aeruginosa relative abundance and age and a negative relationship between P. aeruginosa relative abundance and lung function. Members of the genus Ralstonia were also found to contribute substantially to the bacterial communities in a number of patients. Follow-up samples from a subset of patients revealed high levels of bacterial community temporal stability. The detailed microbiological characterization presented here provides a basis for the reassessment of the clinical management of PCD airway infections.

Standards of care for CFTR variant-specific therapy (including modulators) for people with cystic fibrosis.

Cystic fibrosis (CF) has entered the era of variant-specific therapy, tailored to the genetic variants in the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) gene. CFTR modulators, the first variant-specific therapy available, have transformed the management of CF. The latest standards of care from the European CF Society (2018) did not include guidance on variant-specific therapy, as CFTR modulators were becoming established as a novel therapy. We have produced interim standards to guide healthcare professionals in the provision of variant-specific therapy for people with CF. Here we provide evidence-based guidance covering the spectrum of care, established using evidence from systematic reviews and expert opinion. Statements were reviewed by key stakeholders using Delphi methodology, with agreement (≥80%) achieved for all statements after one round of consultation. Issues around accessibility are discussed and there is clear consensus that all eligible people with CF should have access to variant-specific therapy.

Long-term cultivation-independent microbial diversity analysis demonstrates that bacterial communities infecting the adult cystic fibrosis lung show stability and resilience.

BACKGROUND: Culture-independent analysis of the respiratory secretions of people with cystic fibrosis (CF) has identified many bacterial species not previously detected using culture in this context. However, little is known about their clinical significance or persistence in CF airways. METHODS: The authors characterised the viable bacterial communities in the sputum collected from 14 patients at monthly intervals over 1 year using a molecular community profiling technique-terminal restriction fragment length polymorphism. Clinical characteristics were also collected, including lung function and medications. Ecological community measures were determined for each sample. Microbial community change over time within subjects was defined using ecological analytical tools, and these measures were compared between subjects and to clinical features. RESULTS: Bacterial communities were stable within subjects over time but varied between subjects, despite similarities in clinical course. Antibiotic therapy temporarily perturbed these communities which generally returned to pretreatment configurations within 1 month. Species usually considered CF pathogens and those not previously regarded as such exhibited similar patterns of persistence. Less diverse sputum bacterial communities were correlated to lung disease severity and relative abundance of Pseudomonas aeruginosa. CONCLUSION: Whilst not true in all cases, the microbial communities that chronically infect the airways of patients with CF can vary little over a year despite antibiotic perturbation. The species present tended to vary more between than within subjects, suggesting that each CF airway infection is unique, with relatively stable and resilient bacterial communities. The inverse relationship between community richness and disease severity is similar to findings reported in other mucosal infections.

Using explainable machine learning to characterise data drift and detect emergent health risks for emergency department admissions during COVID-19.

A key task of emergency departments is to promptly identify patients who require hospital admission. Early identification ensures patient safety and aids organisational planning. Supervised machine learning algorithms can use data describing historical episodes to make ahead-of-time predictions of clinical outcomes. Despite this, clinical settings are dynamic environments and the underlying data distributions characterising episodes can change with time (data drift), and so can the relationship between episode characteristics and associated clinical outcomes (concept drift). Practically this means deployed algorithms must be monitored to ensure their safety. We demonstrate how explainable machine learning can be used to monitor data drift, using the COVID-19 pandemic as a severe example. We present a machine learning classifier trained using (pre-COVID-19) data, to identify patients at high risk of admission during an emergency department attendance. We then evaluate our model's performance on attendances occurring pre-pandemic (AUROC of 0.856 with 95%CI [0.852, 0.859]) and during the COVID-19 pandemic (AUROC of 0.826 with 95%CI [0.814, 0.837]). We demonstrate two benefits of explainable machine learning (SHAP) for models deployed in healthcare settings: (1) By tracking the variation in a feature's SHAP value relative to its global importance, a complimentary measure of data drift is found which highlights the need to retrain a predictive model. (2) By observing the relative changes in feature importance emergent health risks can be identified.

Studying bacteria in respiratory specimens by using conventional and molecular microbiological approaches.

BACKGROUND: Drawing from previous studies, the traditional routine diagnostic microbiology evaluation of samples from chronic respiratory conditions may provide an incomplete picture of the bacteria present in airways disease. Here, the aim was to determine the extent to which routine diagnostic microbiology gave a different assessment of the species present in sputa when analysed by using culture-independent assessment. METHODS: Six different media used in routine diagnostic microbiology were inoculated with sputum from twelve patients. Bacterial growth on these plates was harvested and both RNA and DNA extracted. DNA and RNA were also extracted directly from the same sample of sputum. All nucleic acids served as templates for PCR and reverse transcriptase-PCR amplification of "broad range" bacterial 16S rRNA gene regions. The regions amplified were separated by Terminal Restriction Fragment Length Polymorphism (T-RFLP) profiling and compared to assess the degree of overlap between approaches. RESULTS: A mean of 16.3 (SD 10.0) separate T-RF band lengths in the profiles from each sputum sample by Direct Molecular Analysis, with a mean of 8.8 (SD 5.8) resolved by DNA profiling and 13.3 (SD 8.0) resolved by RNA profiling. In comparison, 8.8 (SD 4.4) T-RF bands were resolved in profiles generated by Culture-derived Molecular Analysis. There were a total of 184 instances of T-RF bands detected in the direct sputum profiles but not in the corresponding culture-derived profiles, representing 83 different T-RF band lengths. Amongst these were fifteen instances where the T-RF band represented more than 10% of the total band volume (with a mean value of 23.6%). Eight different T-RF band lengths were resolved as the dominant band in profiles generated directly from sputum. Of these, only three were detected in profiles generated from the corresponding set of cultures. CONCLUSION: Due to their focus on isolation of a small group of recognised pathogens, the use of culture-dependent methods to analyse samples from chronic respiratory infections can provide a restricted understanding of the bacterial species present. The use of a culture-independent molecular approach here identifies that there are many bacterial species in samples from CF and COPD patients that may be clinically relevant.

Respiratory microbiota resistance and resilience to pulmonary exacerbation and subsequent antimicrobial intervention.

Pulmonary symptoms in cystic fibrosis (CF) begin in early life with chronic lung infections and concomitant airway inflammation leading to progressive loss of lung function. Gradual pulmonary function decline is interspersed with periods of acute worsening of respiratory symptoms known as CF pulmonary exacerbations (CFPEs). Cumulatively, CFPEs are associated with more rapid disease progression. In this study multiple sputum samples were collected from adult CF patients over the course of CFPEs to better understand how changes in microbiota are associated with CFPE onset and management. Data were divided into five clinical periods: pre-CFPE baseline, CFPE, antibiotic treatment, recovery, and post-CFPE baseline. Samples were treated with propidium monoazide prior to DNA extraction, to remove the impact of bacterial cell death artefacts following antibiotic treatment, and then characterised by 16S rRNA gene-targeted high-throughput sequencing. Partitioning CF microbiota into core and rare groups revealed compositional resistance to CFPE and resilience to antibiotics interventions. Mixed effects modelling of core microbiota members revealed no significant negative impact on the relative abundance of Pseudomonas aeruginosa across the exacerbation cycle. Our findings have implications for current CFPE management strategies, supporting reassessment of existing antimicrobial treatment regimens, as antimicrobial resistance by pathogens and other members of the microbiota may be significant contributing factors.

Does bacterial density in cystic fibrosis sputum increase prior to pulmonary exacerbation?

BACKGROUND: Cystic Fibrosis (CF) lung disease is characterised by an inexorable decline in lung function, punctuated by periods of symptomatic worsening known as pulmonary exacerbations (referred to here as CFPE). Despite their clinical significance, the cause of CFPE remains undetermined. It has been suggested that an increase in bacterial density may be a trigger, although this has not been shown empirically. METHODS: Here, a previously validated quantitative PCR-based approach was used to assess numbers of Pseudomonas aeruginosa and of total bacteria in respiratory secretions from patients during the period leading up to CFPE. Sputum samples collected from 12 adult CF patients were selected retrospectively to fall approximately 21, 14, 7 and 0 days prior to CFPE diagnosis. In addition, the relationships between clinical parameters (FEV(1), temperature and patient reported outcome measures) and microbiological data were investigated. RESULTS: No significant changes either in total bacterial or P. aeruginosa numbers were identified prior to CFPE. Of all the correlations tested, only temperature showed a significant correlation with total bacterial numbers in the period leading to CFPE. CONCLUSIONS: These findings strongly suggest that CFPE do not generally result from increased bacterial density within the airways. Instead, data presented here are consistent with alternative models of pulmonary exacerbation.

Revealing the dynamics of polymicrobial infections: implications for antibiotic therapy.

As a new generation of culture-independent analytical strategies emerge, the amount of data on polymicrobial infections will increase dramatically. For these data to inform clinical thinking, and in turn to maximise benefits for patients, an appropriate framework for their interpretation is required. Here, we use cystic fibrosis (CF) lower airway infections as a model system to examine how conceptual and technological advances can address two clinical questions that are central to improved management of CF respiratory disease. Firstly, can markers of the microbial community be identified that predict a change in infection dynamics and clinical outcomes? Secondly, can these new strategies directly characterize the impact of antimicrobial therapies, allowing treatment efficacy to be both assessed and optimized?