Not All Flares Are Equal: Stratifying Risk During Asthma/COPD Overlap Exacerbations

Asthma/chronic obstructive pulmonary disease overlap (ACO) defines a distinct clinical cohort characterised by persistent airflow limitation with overlapping features of both asthma and chronic obstructive pulmonary disease (COPD). In its most recent update, the Global Initiative for Asthma (GINA) emphasised the importance of considering ACO not as a single disease entity, but instead a spectrum of overlapping clinical phenotypes with varying underlying biological mechanisms [1]. Despite its prevalence, reported to affect up to 30% of patients with COPD and 27% of patients with asthma based on pooled epidemiological data, ACO remains one of the most poorly characterised entities in obstructive airways disease [2, 3]. Compared to either condition alone, ACO is associated with more frequent exacerbations, greater symptom burden and worse clinical outcomes, yet it continues to occupy an uncomfortable diagnostic grey zone, particularly in the acute phase, often mislabelled as pure COPD [2].

This misclassification has important clinical consequences. Treatment paradigms for COPD exacerbations, centred on bronchodilators, systemic corticosteroids and empiric antibiotics, were not designed with the variable biology of ACO in mind. With the expanding therapeutic options for treatment of acute eosinophilic airway disease, including targeted biologic therapies, contextualising ACO exacerbations becomes increasingly important [4]. Treating a neutrophil‐dominant exacerbation as though it were eosinophilic, and vice versa, risks both undertreatment and potential harm.

In a recent publication in Respirology, Yuan et al. sought to characterise the inflammatory heterogeneity of acute exacerbations of ACO, identifying distinct phenotypes within a hospitalised cohort and exploring their association with clinical outcomes [5]. Employing a prospective cohort design across two Chinese centres, the authors enrolled 2444 patients with confirmed diagnosis of ACO and applied unsupervised clustering analysis to routine admission blood parameters, including total white cell count (WBCs), absolute neutrophil, eosinophil and lymphocyte counts, neutrophil‐to‐lymphocyte ratio (NLR), C‐reactive protein (CRP), procalcitonin and interleukin‐6 (IL‐6), to identify three reproducible inflammatory phenotypes:

• –Cluster T1: NLR elevated, hypoeosinophilic phenotype
• –Cluster T2: eosinophilic phenotype
• –Cluster T3: eosinophil‐neutrophil balanced phenotype

Notably, phenotype assignment relied exclusively on peripheral blood parameters alone, with no incorporation of airway compartment measures such as sputum eosinophilia or fractional exhaled nitric oxide (FeNO). In‐hospital outcome measures, including ICU admission, requirement for invasive ventilation, and in‐hospital mortality, were prospectively documented, with patients followed to discharge or in‐hospital death. External validation in an independent centre supported the reproducibility of the proposed clustering methods.

Yuan et al. nicely demonstrate that inflammatory phenotype at admission predicts in‐hospital outcomes in ACO, with cluster T1, characterised by elevated NLR and hypoeosinophilia, carrying a composite adverse outcome rate of 24.2%, more than three times that of the eosinophilic T2 cluster (7.2%) and the balanced T3 cluster (5.6%). The T1 cluster was significantly associated with elevation of inflammatory markers including WBCs, CRP, procalcitonin and IL‐6 (all p < 0.001), reflecting a greater burden of systemic inflammation. Conversely, higher eosinophil counts at the time of an exacerbation were associated with more favourable in‐hospital outcomes, potentially reflective of greater corticosteroid responsiveness in the acute phase.

The strengths of this study are notable. To date, ACO has been poorly characterised with little known about its acute presentation pattern or in‐hospital clinical course, making large prospective studies in this space particularly valuable. The authors should be commended for their prospective design, incorporating a large sample size and ensuring external validation of the clustering solution which strengthens confidence in the reproducibility of the identified phenotypes. The unsupervised clustering approach utilised to classify patients is particularly noteworthy, allowing natural trends to emerge from the data itself. The decision to anchor the model in routine admission parameters which are universally available and inexpensive contributes to the generalisability of the findings allowing for real world implementation.

The inclusion of outcome correlation particularly strengthens this study. Biologically, worse outcomes in cluster T1, a group with higher systemic inflammatory burden, are intuitive and clinically coherent in the acute setting, reflecting poor prognosis in the neutrophil predominant phenotype. Unfortunately, routine microbiological testing was not included in the algorithm so infectious vs. non‐infectious drivers of inflammation could not be elucidated, perhaps over‐stating the true neutrophilic burden of disease. Additionally, patients with systemic corticosteroid or biologic use within the 12 weeks prior to admission were excluded from the study, ensuring the cohort is effectively naïve to biologics and steroids, allowing for accurate phenotyping.

Some important questions remain which, while not fully addressed in this study, open the doorway for meaningful conversation around this understudied topic. As the authors themselves acknowledge, reliance on peripheral blood parameters alone for phenotyping may not fully capture the true inflammatory biology at the airway level. Whilst systemic markers alone are pragmatic and justified in the acute setting, sputum eosinophilia and FeNO may diverge from blood‐based inflammatory indices during exacerbations, and their inclusion in future phenotyping frameworks may provide further insight into the compartmental divergence that exists in ACO. In addition, the follow‐up period for this study opens discussion for future exploration. Outcome assessment was confined to the point of discharge or in‐hospital death, leaving the early post‐discharge period, a well‐documented window of vulnerability in COPD and asthma, entirely uncharacterised [6, 7]. Whether inflammatory phenotype can accurately predict those at higher risk of re‐admission represents a compelling next step in this story.

Yuan et al. have addressed a clinically understudied, important sub‐phenotype of obstructive airways disease that requires accurate phenotyping in order to ensure optimal treatment. This is particularly timely in the era of advancing biologic therapies for the treatment of acute exacerbations of COPD and asthma. The phenotyping framework proposed here, if validated appropriately and extended to address limitations in compartmental diversity and post‐discharge outcomes, has the potential to modify how we triage, treat, and follow‐up this complex patient population. Highlighting this important topic should be commended, with further work needed by all to continue to deepen our knowledge base.

Author Contributions

Alessandro N. Franciosi: writing – review and editing. Helen O'Brien: writing – original draft.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.