Multivariate joint modeling to identify markers of growth and lung function decline that predict cystic fibrosis pulmonary exacerbation onset

Mathematical models of cystic fibrosis as a systemic disease

Cystic fibrosis (CF) is widely known as a disease of the lung, even though it is in truth a systemic disease, whose symptoms typically manifest in gastrointestinal dysfunction first. CF ultimately impairs not only the pancreas and intestine but also the lungs, gonads, liver, kidneys, bones, and the cardiovascular system. It is caused by one of several mutations in the gene of the epithelial ion channel protein CFTR. Intense research and improved antimicrobial treatments during the past eight decades have steadily increased the predicted life expectancy of a person with CF (pwCF) from a few weeks to over 50 years. Moreover, several drugs ameliorating the sequelae of the disease have become available in recent years, and notable treatments of the root cause of the disease have recently generated substantial improvements in health for some but not all pwCF. Yet, numerous fundamental questions remain unanswered. Complicating CF, for instance in the lung, is the fact that the associated insufficient chloride secretion typically perturbs the electrochemical balance across epithelia and, in the airways, leads to the accumulation of thick, viscous mucus and mucus plaques that cannot be cleared effectively and provide a rich breeding ground for a spectrum of bacterial and fungal communities. The subsequent infections often become chronic and respond poorly to antibiotic treatments, with outcomes sometimes only weakly correlated with the drug susceptibility of the target pathogen. Furthermore, in contrast to rapidly resolved acute infections with a single target pathogen, chronic infections commonly involve multi-species bacterial communities, called "infection microbiomes," that develop their own ecological and evolutionary dynamics. It is presently impossible to devise mathematical models of CF in its entirety, but it is feasible to design models for many of the distinct drivers of the disease. Building upon these growing yet isolated modeling efforts, we discuss in the following the feasibility of a multi-scale modeling framework, known as template-and-anchor modeling, that allows the gradual integration of refined sub-models with different granularity. The article first reviews the most important biomedical aspects of CF and subsequently describes mathematical modeling approaches that already exist or have the potential to deepen our understanding of the multitude aspects of the disease and their interrelationships. The conceptual ideas behind the approaches proposed here do not only pertain to CF but are translatable to other systemic diseases. This article is categorized under: Congenital Diseases > Computational Models.

Statistical Evaluation of Metaproteomics and 16S rRNA Amplicon Sequencing Techniques for Study of Gut Microbiota Establishment in Infants with Cystic Fibrosis

Newborn screening for cystic fibrosis (CF) can identify affected but asymptomatic infants. The selection of omic technique for gut microbiota study is crucial due to both the small amount of feces available and the low microorganism load. Our aims were to compare the agreement between 16S rRNA amplicon sequencing and metaproteomics by a robust statistical analysis, including both presence and abundance of taxa, to describe the sequential establishment of the gut microbiota during the first year of life in a small size sample (8 infants and 28 fecal samples). The taxonomic assignations by the two techniques were similar, whereas certain discrepancies were observed in the abundance detection, mostly the lower predicted relative abundance of Bifidobacterium and the higher predicted relative abundance of certain Firmicutes and Proteobacteria by amplicon sequencing. During the first months of life, the CF gut microbiota is characterized by a significant enrichment of Ruminococcus gnavus, the expression of certain virulent bacterial traits, and the detection of human inflammation-related proteins. Metaproteomics provides information on composition and functionality, as well as data on host-microbiome interactions. Its strength is the identification and quantification of Actinobacteria and certain classes of Firmicutes, but alpha diversity indices are not comparable to those of amplicon sequencing. Both techniques detected an aberrant microbiota in our small cohort of infants with CF during their first year of life, dominated by the enrichment of R. gnavus within a human inflammatory environment. IMPORTANCE In recent years, some techniques have been incorporated for the study of microbial ecosystems, being 16S rRNA gene sequencing being the most widely used. Metaproteomics provides the advantage of identifying the interaction between microorganisms and human cells, but the available databases are less extensive as well as imprecise. Few studies compare the statistical differences between the two techniques to define the composition of an ecosystem. Our work shows that the two methods are comparable in terms of microorganism identification but provide different results in alpha diversity analysis. On the other hand, we have studied newborns with cystic fibrosis, for whom we have described the establishment of an intestinal ecosystem marked by the inflammatory response of the host and the enrichment of Ruminococcus gnavus.

Sweat metabolomics before and after intravenous antibiotics for pulmonary exacerbation in people with cystic fibrosis

BACKGROUND

People with cystic fibrosis (PWCF) suffer from acute unpredictable reductions in pulmonary function associated with a pulmonary exacerbation (PEx) that may require hospitalization. PEx symptoms vary between PWCF without universal diagnostic criteria for diagnosis and response to treatment.

RESEARCH QUESTION

We characterized sweat metabolomes before and after intravenous (IV) antibiotics in PWCF hospitalized for PEx to determine feasibility and define biological alterations by IV antibiotics for PEx.

STUDY DESIGN AND METHODS

PWCF with PEx requiring hospitalization for IV antibiotics were recruited from clinic. Sweat samples were collected using the Macroduct® Sweat Collection System at admission prior to initiation of IV antibiotics and after completion prior to discharge. Samples were analyzed for metabolite changes using ultra-high-performance liquid chromatography/tandem accurate mass spectrometry.

RESULTS

Twenty-six of 29 hospitalized PWCF completed the entire study. A total of 326 compounds of known identity were detected in sweat samples. Of detected metabolites, 147 were significantly different between pre-initiation and post-completion of IV antibiotics for PEx (average treatment 14 days). Global sweat metabolomes changed from before and after IV antibiotic treatment. We discovered specific metabolite profiles predictive of PEx status as well as enriched biologic pathways associated with PEx. However, metabolomic changes were similar in PWCF who failed to return to baseline pulmonary function and those who did not.

INTERPRETATION

Our findings demonstrate the feasibility of non-invasive sweat metabolomic profiling in PWCF and the potential for sweat metabolomics as a prospective diagnostic and research tool to further advance our understanding of PEx in PWCF.

Joint modelling of multivariate longitudinal clinical laboratory safety outcomes, concomitant medication and clinical adverse events: application to artemisinin-based treatment during pregnancy clinical trial

Background

In drug trials, clinical adverse events (AEs), concomitant medication and laboratory safety outcomes are repeatedly collected to support drug safety evidence. Despite the potential correlation of these outcomes, they are typically analysed separately, potentially leading to misinformation and inefficient estimates due to partial assessment of safety data. Using joint modelling, we investigated whether clinical AEs vary by treatment and how laboratory outcomes (alanine amino-transferase, total bilirubin) and concomitant medication are associated with clinical AEs over time following artemisinin-based antimalarial therapy.

Methods

We used data from a trial of artemisinin-based treatments for malaria during pregnancy that randomized 870 women to receive artemether–lumefantrine (AL), amodiaquine–artesunate (ASAQ) and dihydroartemisinin–piperaquine (DHAPQ). We fitted a joint model containing four sub-models from four outcomes: longitudinal sub-model for alanine aminotransferase, longitudinal sub-model for total bilirubin, Poisson sub-model for concomitant medication and Poisson sub-model for clinical AEs. Since the clinical AEs was our primary outcome, the longitudinal sub-models and concomitant medication sub-model were linked to the clinical AEs sub-model via current value and random effects association structures respectively. We fitted a conventional Poisson model for clinical AEs to assess if the effect of treatment on clinical AEs (i.e. incidence rate ratio (IRR)) estimates differed between the conventional Poisson and the joint models, where AL was reference treatment.

Results

Out of the 870 women, 564 (65%) experienced at least one AE. Using joint model, AEs were associated with the concomitant medication (log IRR 1.7487; 95% CI: 1.5471, 1.9503; p < 0.001) but not the total bilirubin (log IRR: -0.0288; 95% CI: − 0.5045, 0.4469; p = 0.906) and alanine aminotransferase (log IRR: 0.1153; 95% CI: − 0.0889, 0.3194; p = 0.269). The Poisson model underestimated the effects of treatment on AE incidence such that log IRR for ASAQ was 0.2118 (95% CI: 0.0082, 0.4154; p = 0.041) for joint model compared to 0.1838 (95% CI: 0.0574, 0.3102; p = 0.004) for Poisson model.

Conclusion

We demonstrated that although the AEs did not vary across the treatments, the joint model yielded efficient AE incidence estimates compared to the Poisson model. The joint model showed a positive relationship between the AEs and concomitant medication but not with laboratory outcomes.

Trial registration

ClinicalTrials.gov: NCT00852423

New Insights in Microbial Species Predicting Lung Function Decline in CF: Lessons from the MucoFong Project

Several predictive models have been proposed to understand the microbial risk factors associated with cystic fibrosis (CF) progression. Very few have integrated fungal airways colonisation, which is increasingly recognized as a key player regarding CF progression. To assess the association between the percent predicted forced expiratory volume in 1 s (ppFEV1) change and the fungi or bacteria identified in the sputum, 299 CF patients from the “MucoFong” project were included and followed-up with over two years. The relationship between the microorganisms identified in the sputum and ppFEV1 course of patients was longitudinally analysed. An adjusted linear mixed model analysis was performed to evaluate the effect of a transient or chronic bacterial and/or fungal colonisation at inclusion on the ppFEV1 change over a two-year period. Pseudomonas aeruginosa, Achromobacter xylosoxidans, Stenotrophomonas maltophilia, and Candida albicans were associated with a significant ppFEV1 decrease. No significant association was found with other fungal colonisations. In addition, the ppFEV1 outcome in our model was 11.26% lower in patients presenting with a transient colonisation with non-pneumoniae Streptococcus species compared to other patients. These results confirm recently published data and provide new insights into bacterial and fungal colonisation as key factors for the assessment of lung function decline in CF patients.

An Animated Functional Data Analysis Interface to Cluster Rapid Lung Function Decline and Enhance Center-Level Care in Cystic Fibrosis

Identifying disease progression through enhanced decision support tools is key to chronic management in cystic fibrosis at both the patient and care center level. Rapid decline in lung function relative to patient level and center norms is an important predictor of outcomes. Our objectives were to construct and utilize center-level classification of rapid decliners to develop an animated dashboard for comparisons within patients over time, multiple patients within centers, or between centers. A functional data analysis technique known as functional principal components analysis was applied to lung function trajectories from 18,387 patients across 247 accredited centers followed through the United States Cystic Fibrosis Foundation Patient Registry, in order to cluster patients into rapid decline phenotypes. Smaller centers (<30 patients) had older patients with lower baseline lung function and less severe rates of decline and had maximal decline later, compared to medium (30-150 patients) or large (>150 patients) centers. Small centers also had the lowest prevalence of early rapid decliners (17.7%, versus 24% and 25.7% for medium and large centers, resp.). The animated functional data analysis dashboard illustrated clustering and center-specific summaries of the rapid decline phenotypes. Clinical scenarios and utility of the center-level functional principal components analysis (FPCA) approach are considered and discussed.