Phenotyping Responses to Systemic Corticosteroids in the Management of Asthma Attacks

RESEARCH BACKGROUND AND RATIONALE

Asthma attacks are loosely defined as symptom deterioration and/or lung function from baseline. In contrast, the more stringent classification of attack severity is based on the decision to treat: a severe episode requires ≥3 days of oral corticosteroids and/or hospitalization. Severe asthma attacks cause substantial morbidity, healthcare utilization, and avoidable deaths. Despite growing evidence of heterogeneity of mechanisms driving asthma attacks, the standard of care in acute asthma has not changed for 30 years. It consists of a 'one size fits all' treatment with oral corticosteroids and often antibiotics.

Data in a cross-sectional analysis of stable-state severe asthma showed the complementary and potentially additive value of blood eosinophils and FeNO. In this study of a cohort of patients with severe asthma proven to be highly adherent to high-intensity corticosteroid therapy, it was showed that blood eosinophils and FeNO non-suppression provide mechanistic information on two different immune compartments: blood eosinophils reflect the systemic pool of effector cells and circulating IL-5; whereas FeNO correlates with type-2 cytokine, chemokine, alarmin and eosinophilic inflammation in the airways.

In addition, epidemiological work investigating the annualized severe asthma attack rates in the control arm of randomized clinical trials showed additive value for baseline blood eosinophils and FeNO to predict asthma attacks. In those trial populations, a raised baseline FeNO (≥50 vs. <25 ppb) was associated with double the severe asthma attack rates for patients with similar blood eosinophilia (≥0.30×109/L). Further analyses revealed that the excess risk of asthma attacks conferred by both biomarkers was removed by type-2 targeted anti-inflammatory therapy.

In acute asthma, two studies have documented the heterogeneity of attacks. Still, treatment responses to acute systemic corticosteroid and antibiotic courses have not been related to the type-2 inflammatory phenotype and the presence of airway infection, respectively. Furthermore, evidence supporting acute point-of-care biomarker assessments is lacking, even though asthma attacks often present to general community practitioners and disproportionally affect isolated communities.

STUDY HYPOTHESIS

Based on previous knowledge from chronic stable asthma and reports that asthma attacks are heterogeneous events, it is hypothesize that treatment responses to systemic corticosteroids in acute asthma vary according to the underlying inflammatory phenotype. Specifically, it is speculated that the type-2 inflammatory events (identified by raised blood EOS and FeNO) benefit most from corticosteroids and that events presenting with low type-2 biomarkers have a lesser objective clinical and biological anti-inflammatory response but perhaps higher rates of airway infection or dysfunction.

A critical secondary hypothesis of this study is that POC inflammatory and microbiological assessments in acute asthma are possible, acceptable, and reliable. The ramifications of such findings would be extending biomarker-guided asthma management to primary care and isolated areas, thus assisting corticosteroid- and antibiotic-stewardship efforts.

STUDY DESIGN

The study will be conducted at the Centre Hospitalier Universitaire de Sherbrooke (CHUS)'s research center. This is not a clinical trial; systemic corticosteroids are the standard treatment for asthma attacks and will be administered as required.

PARTICIPANTS

Patients included in the primary analysis and counting toward the target sample size will need to have objective proof of asthma as defined by international guidelines: either based on the medical record, the variability in respiratory function between study visits 1 and 2, or with a clinically required follow-up test ordered by the study respirologist and performed after study participation. Unconfirmed asthmatics data will be reported but not used in the primary analysis.

All participants (± their parent or legal tutor) must provide informed consent for the primary study protocol. Patients who cannot or refuse to participate in the diary card data collection for the home monitoring period will still be able to participate in all other study procedures.

STUDY VISITS

Visit one will be on day 0 of the asthma attack, with oral prednisone issued by the community pharmacy (40mg x 7 days); visit two will occur on day 7; a web-based survey will be sent 90 days later.

Concomitant antibiotic use will only be considered for patients with ≥1 of the following criteria: C-reactive protein >20 mg/L, procalcitonin >0.25 mcg/L, or positive sputum culture or multiplex POC test for the significant bacterial pathogen.

Healthy controls will attend one visit.

All study visits will be conducted in the morning to avoid the confounding effect of circadian variation of type-2 inflammatory biomarkers.

• Visit 1 Multimodal assessments will be performed at the study visit lasting ~2 hours.
• Visit 2 It will additionally include a patient satisfaction score and permission to contact the patient's primary care provider for a physician survey about POC testing.
• Visit 3 Ninety days after visit 2, patients will be contacted by email to complete a quick web-based survey form.

RESPIRATORY RESEARCH LABORATORY ANALYSES

The study will be closely associated with the local Quebec Respiratory Health Research Network biobank. It will be done in collaboration with Sherbrooke's Lévesque-Martin microbiology research laboratory and the Université de Sherbrooke RNOmics platform.

Nasosorption strips will be spun with NELF eluted through a plastic mesh and then frozen (-80°C).

Sputum will be processed with Dulbecco's phosphate-buffered saline (DPBS)-eluted sputum supernatant frozen before differential cell counting, as previously described.

Whole blood will be aliquoted into serum, plasma, buffy coat, and complete blood samples and frozen.

The nasal swab be analyzed immediately using the ultrarapid ID NOW molecular test to rule out SaRS-CoV2.

One nasopharyngeal swab on visit one will be analyzed using a point-of-care multiplex PCR for 19 respiratory viruses and bacteria (using the BIOFIRE® Respiratory panel - not repeated if RP2.1-EZ is made available) to identify other infectious contributors to the attack.

The other nasopharyngeal brushes plus sputum plugs will have RNA stabilized and then frozen (2x in RNA-Protect and 1x flash frozen).

If sufficient sputum is produced at visit 1, a sputum plug will be sent for conventional bacterial culture, and up to three other plugs will be frozen (2x flash frozen, 1x in 10% glycerol). Urine samples will be frozen.

STATISTICAL ANALYSES

Descriptive statistics will be generated for the complete sample and separated by groups (i.e., frequencies and percentages for categorical, mean, and standard deviation (SD) for normally distributed continuous, median, and interquartile range for skewed continuous). The normality of variables will be assessed visually using Q-Q plots and histograms.

Demographics will be compared between asthmatics and healthy controls using unpaired t-tests (Mann-Whitney) for continuous variables and Chi-squared (Fisher's exact test) for categorical variables. No imputation methods are planned for missing data. Frequencies and percentages of missing will be presented, and listwise deletion will be used for concerned analyses.

The primary analysis conducted in asthmatics will evaluate the change in FEV1.

To consider the influence of potentially confounding variables, simple and multiple linear regression models will be created for both independent variables (blood eosinophils and FeNO). Confounding factors considered are age, sex at birth, body mass index, atopic status, serum IgE, Asian race, nasal polyposis, smoking history, Charlson index, treatment intensity at baseline, exacerbation history in the past 12 months, and ACQ-5.

Only confounders associated (p<0.1) with FEV1 will be included in the final multivariable model. Model assumptions (Linearity, normality of residuals, and homoscedasticity) will be assessed, and models will be adjusted in case of violation.

Coefficients will be presented with their respective transformations and 95% confidence intervals.

Secondary analyses will comprise:

• Change in other spirometry values (FVC, FEV1/FVC)
• Change in symptom scores (VAS dyspnea rating, ACQ-5)
• Change in oscillometry (R5-20)

All statistics will be analyzed with a two-sided α of 0.05 and, when multiple inflammatory mediators are tested, by controlling for a false discovery rate <0.05 [54]. Statistical analyses will be performed with SPSS (SPSS Inc, USA), RStudio 2021.09.01 build 372 (RStudio, USA) with R v4.1.2 (R Foundation), and GraphPad Prism (GraphPad Software, USA).

SAMPLE SIZE AND FEASIBILITY

For the primary analysis (aim 1), 47 asthma attacks must be assessed to identify correlation coefficients r≥0.40 with a two-sided p-value <0.05 and 80% power. This conservative approach will allow us to reduce the presence of overfitting when considering a maximum of 5 independent variables in the multivariable model, following the rule of thumb of 10 observations per predictor.

The secondary analyses are exploratory.