Asthma Exacerbations in Severe Asthma: Why Systemic Corticosteroids May not Always Be the Best Treatment Option

Advances in the management of severe, eosinophilic asthma have improved, but asthma exacerbations continue to occur. This review aims to look at the evidence we have about why exacerbations may occur; their phenotype and why oral corticosteroids may not always be the best treatment option for all exacerbation of symptoms in individuals with severe asthma. Studies dating back to the 1990s showed that asthma exacerbations across the spectrum of asthma severity were of different inflammatory endotypes. In addition, there is a wealth of evidence suggesting that eosinophilic inflammation is very responsive to corticosteroid therapy, but that non-eosinophilic inflammation is less so. Two recent UK-based studies have undertaken systematic phenotyping of exacerbations in severe asthma and have shown that there are a significant minority of exacerbation events with an increase in asthma symptoms, fall in lung function, but without evidence of raised T2 biomarkers. The evidence to date would suggest that T2 biomarker low asthma exacerbations do not benefit from the administration of oral corticosteroids; in fact, the effect of the oral corticosteroids is harmful. However, there is a paucity of data to answer this question directly. Further research is needed to assess the evolution of non-T2 exacerbations not treated with OCS in a randomised, placebo-controlled, manner.


Introduction
Treatment strategies to achieve asthma control have developed with the evolving understanding of the pathobiology of asthma. Early descriptions of asthma were of paroxysmal breathlessness which was later recognised to be due to bronchial hyperresponsiveness causing variable airways obstruction; therefore, initial treatments focused on agents that result in bronchodilatation, including theophylline, ephedrine and isoprenaline, and then more selective beta2-agonists, some of which are still used today [1,2]. Improvement in asthma symptoms with the use of corticosteroids in those with evidence of airways eosinophilia [3] led to more widespread use of corticosteroid treatment, with the subsequent development of inhaled corticosteroid (ICS) therapy [4,5], and the identification of eosinophilic inflammation in the airways as the major pathological process. In parallel, asthma management guidelines evolved to incorporate the earlier introduction of anti-inflammatory therapy to treatment regimes in a step-wise [6,7] manner within this paradigm. In the past, patients with persisting poor control, despite good adherence to maximal doses of ICS and long-acting bronchodilator treatment, would often end up on low dose maintenance oral corticosteroids (OCS) or have frequent OCS bursts for exacerbations. However, the well-recognised burden of OCS-related toxicities within these OCStreated asthma populations [8][9][10] fuelled the drive for a greater understanding of the precise mechanisms of inflammation to allow for the development of targeted immune therapies. The development of efficacious, targeted anti-T2 agents has transformed the treatment of severe asthma (anti-IgE, anti-IL5/5R, anti-IL4R, and anti-TSLP), reducing asthma exacerbations by around 50% [11][12][13][14][15] and facilitating reduction of maintenance OCS (mepolizumab, benralizumab, and dupilumab) [11][12][13], thereby substantially reducing OCS exposure. However, despite these significant advances with treatments, which remove T2-eosinophilic inflammation, patients with severe asthma continue to have exacerbations, which has led to a focus on the additional mechanisms underlying these clinical events.

Heterogeneity of Asthma Exacerbations
A number of studies in the 1990s assessed the inflammatory phenotype of asthma exacerbations using sputum expectorated at the time of exacerbation [16][17][18][19][20]. Although the definition of sputum eosinophilia and neutrophilia, the inclusion criteria, and the population sampled varied between these different studies, they all demonstrated heterogeneity of airway inflammation at times of increased asthma symptoms seen across the spectrum of asthma severity from mild to severe exaerbations. Furthermore, it was clear that the inflammatory response in the airways was generally reproducible with the same trigger of exacerbation, for example, neutrophilic airways inflammation with viral triggers, or eosinophilic airways inflammation with allergen exposure [21].
In parallel, more easily measured biomarkers (specifically fractional exhaled nitric oxide (FeNO) and peripheral blood eosinophil count) have emerged as robust proxy markers of T2-driven eosinophilic airways inflammation and as prognostic markers for exacerbation and predictive markers of therapeutic response (corticosteroids and novel T2-biologics). These biomarkers have also been shown to be of value in the phenotyping process in all severities of asthma to allow better treatment choices [22][23][24]25••, 26••] but outside of severe asthma, have not yet been widely adopted in clinical practice.
Recent evidence has further demonstrated substantial heterogeneity in asthma exacerbations in adult patients with severe asthma. Using FeNO and peripheral blood eosinophil counts, 71 exacerbations were phenotyped in a severe asthma population enriched for a T2-biomarker low (T2 low) population. An exacerbation visit was undertaken when the patient felt their asthma symptoms were outside of their normal daily variation, at the point where they would seek medical review or follow their personalised asthma action plan. T2 low exacerbations (FeNO ≤ 20 ppb and blood eos ≤ 150 cells/ µL) were compared to those with evidence of T2 inflammation, T2 high (FeNO > 20 ppb or blood eos > 150 cells/µL) at the time of increased symptoms. Nineteeen of 71 (26.7%) exacerbations were T2 low but notably, both T2 high and T2 low reported a similar increase in symptom burden (increase in ACQ7 T2 low 1.4 (0.8) versus T2 high 1.3 (0.8), p = 0.72) and had a similar fall in lung function (both groups had a mean FEV (L) drop of − 0.2L (T2 low -0.2 (0.4), T2 high − 0.2 (0.3), p = 0.93) from the clinically stable state to exacerbation suggesting that there was evidence of increased asthma symptoms and fall in lung function in the absence of T2 driven airways inflammation [25••].
In another study characterising exacerbations in a severe asthma population established on anti-IL5 therapy (mepolizumab), a sputum eosinophil count of ≥ 2% was used to delineate the presence of eosinophilic inflammation at exacerbation. In this study, half of the exacerbations on mepolizumab were non-eosinophilic (sputum eosinophils < 2%), and half had sputum eosinophils ≥ 2%. However, again exacerbations in both groups were associated with a similar fall in FEV1 from baseline to exacerbation (delta change in FEV1% from stable to exacerbation, sputum eosinophils < 2% group = − 10.6% (− 16.2, − 4.9), versus sputum eosinophils ≥ 2% group = − 15.1% (− 22.9, − 4.7), 95% CI − 12.9,5.0, p = 0.31). Of note, the group with low sputum eosinophil count had a greater increase in sputum neutrophils at exacerbation, a higher CRP, and was more likely to have an infectious agent identified on sputum PCR (pathogenic virus or bacteria). Again, ACQ-5 increase and fall in lung function in both groups was comparable, highlighting the fact that non-eosinophilic exacerbations are clinically indistinguishable in terms of symptoms or fall in lung function and supports the need for phenotyping at the time of exacerbation [26••]. In this study, FeNO was useful in phenotyping eosinophilic and non-eosinophilic exacerbations; a FeNO ≤ 20 ppb had a 100% negative predictive value (95% CI 80, 100) for a sputum eosinophil < 2% at exacerbation, while FeNO ≥ 50 ppb had a positive predictive value of 77% (95% CI 60-90%) for a sputum eosinophil count of ≥ 2% at exacerbation. Furthermore, a microbiome analysis from both aforementioned studies showed that a FeNO ≥ 50 ppb was associated with proteobacteria-low, alpha-diversity high microbial profile [27]. The inverse relationship between FeNO and the presence of proteobacteria has previously been reported [28]. Taken together, this reinforces the use of FeNO as a point of care test at exacerbation, to highlight the presence of T2 high inflammation likely to be OCS-responsive, and the reduced likelihood of an exacerbation driven by infectious microorganisms. This study of exacerbation phenotyping is being replicated in benralizumab-treated severe asthmatics; the results are awaited (https:// clini caltr ials. gov/ ct2/ show/ NCT04 102800).
Multiple studies have demonstrated the benefit of corticosteroids, to improve asthma control or reduce exacerbations, is only seen in those with evidence of T2-driven eosinophilic airways inflammation [3,18,[29][30][31][32]. In contrast [3,18,29,30], non-eosinophilic asthma has little response to corticosteroids [33,34] suggesting that corticosteroids will have little benefit in non-T2/non-eosinophilic exacerbations. This hypothesis needs to be formally tested in a clinical trial which phenotypes exacerbations using rapid diagnostic tests and withholds systemic corticosteroids where there is no evidence of T2-inflammtion. FeNO is already available for this type of study, but rapid turnaround of peripheral blood eosinophil count or use of a surrogate measure, e.g., point of care blood count tests [35], may also provide a mechanism to avoid unnecessary corticosteroid exposure. In parallel, given the prevalence of infective pathogens seen in exacerbations in T2-low severe asthma, this approach might be extended to use of rapid diagnostics in spontaneous sputum samples, where available, to determine if pathogenic bacterial and viral pathogens are present, which might also assist in directing appropriate use of antibiotics.

Consequences of Exacerbations
Asthma exacerbations are critically important events, implicated in airway remodelling, loss of lung function, and worse quality of life with economic consequences both to the individual patient and the health economy [36][37][38][39]. Asthma exacerbations are particularly important in those with severe asthma, although only ~ 3.7% of the whole asthma population [40]; they consume over 60% of the costs associated with asthma with significant unscheduled care utilisation [40]. The problem of recurrent exacerbations in severe asthma is a global one. A recent publication from the International Severe Asthma Registry (ISAR) included 4990 severe asthma patients and reported that 25% had ≥ 4 exacerbations per year, with a mean exacerbation rate of 1.7 (2.7)/year [41]. The repeated cycle of poor asthma control and exacerbations treated with OCS is associated with synergistic physiological consequences to the patient.
• Firstly, uncontrolled asthma/recurrent exacerbations are associated with airway remodelling and loss of lung function. The rate of lung function decline has been shown to be inversely proportional to age, further underlining the cruciality of early intervention in the case of recurrent exacerbations in young patients, and the need for timely referral to specialist care [42]. To date lung function decline been shown to be largely non-reversible, [43] even when disease control is achieved in these patients, e.g., on starting a biologic, lung function may improve a little, but does not return to the pre-disease state as targeted biological therapies treat inflammation rather than structural change [12][13][14]44]. Exacerbations and uncontrolled disease therefore result in structural change and lung function impairment.
• Secondly, the consequence of OCS use is by no means insignificant. A previous study of 2 large UK-based registries has shown that even as little as one course of OCS/year on top of background high dose ICS resulted in a statistically significant increase in corticosteroid-induced morbidity. [45] The toxicities associated with OCS use are ubiquitous and well documented; in a study that quantified toxicity associated with OCS use in 101 severe asthmatics about to commence biologics, every single patient had evidence of OCS toxicity, with toxicity burden varying between individuals [8]. OCS use to treat exacerbations leads to the development of OCS-related toxicities and can make disease control more difficult, e.g., weight gain, increased risk of infection, impaired glucose control, increased anxiety, and depression [39,[45][46][47].
There is ample evidence for heterogeneity in airways inflammation at the time of exacerbation; there is robust evidence that in biologics as well as non-biologics treated severe asthma; there are increases in symptom burden and fall in lung function both in the presence and the absence of T2 biology. The question is how can we phenotype these patients, do they all need OCS treatment, and do we have the evidence base to comfortably address these questions?

Management of Asthma Exacerbations
Although there has been considerable improvement in treatment options for disease control in severe asthma and a significant reduction in exacerbations, asthma exacerbations still occur, and unfortunately, the mainstay of treatment for increased asthma symptoms remains OCS.
The assumption that asthma exacerbations are driven by corticosteroidresponsive T2 airways inflammation, along with the knowledge that lack of corticosteroid treatment at exacerbation has been linked to asthma deaths [48, 49], has driven a reflex response to prescribe OCS for all increase in asthma symptoms often without judicious clinical review and in the absence of phenotyping exacerbations prior to treatment.
The potential for avoidance of OCS at exacerbation is suggested by two recent studies of severe asthma exacerbations; one in a non-biologics setting [25••] and the other, exacerbations in the context of biologics-treated severe asthmatics [26••]. In the first, 97.4% (265/272) of reported episodes of increased asthma symptoms received OCS in non-specialist settings, in comparison to 71.2% (84/118, p < 0.001) reviewed by asthma specialists [25••]. In the second biologics-treated cohort, 91% (69/76) of those who sought non-specialist review or followed their personalised asthma action plan for increased symptoms received OCS in comparison to 63.5% (61/96, p < 0.001) of those who had specialist review prior to commencing treatment [26••].
This cultural generalisation of exacerbations being homogenous, corticosteroid-responsive inflammation is also reflected within the world of specialist asthma within our evidence-base. It has become apparent through phase III clinical trials of anti-T2 monoclonal antibodies that phenotyping patients is crucial for assessing the efficacy of targeted biological therapies, with monoclonal antibodies generally being more beneficial in cohorts with a high exacerbation rate and evidence of elevated biomarkers of T2 inflammation [50,51]. However, the inclusion criteria of phase III clinical trials state an exacerbation count which participants must meet, which is more often than not, a count of OCS bursts for asthma exacerbations and assumes that each OCS burst was given for an episode of T2 high asthma exacerbation. Similarly, OCS burst counts also form part of the access criteria for biologics in the UK as a proxy for identifying frequent T2 exacerbators likely to respond to anti-T2 biologics [52][53][54][55]. However, this does not consider that episodes of increased respiratory symptoms can be due to infections, exacerbations of sinus or nasal disease, dysfunctional breathing, or vocal cord dysfunction for example. Unfortunately, these are often managed ubiquitously with courses of OCS and labelled as exacerbations of severe asthma which loses the important delineation of the cause of worsening symptoms by not recognising the underlying pathology of the exacerbation event. It is unsurprising that infectious events are common in patients receiving local immunosuppressive agents (ICS); sometimes, systemic immunosuppressive agents (OCS, now less common) and the effects of biologics on airways immune defence are not yet entirely defined.
Ultimately, part of the problem in exacerbation management is that we do not have an adequate definition or nomenclature when discussing exacerbations. Defining exacerbations using oral steroids assumes that the steroids were appropriate and that the pathology is steroid responsive. Ideally, an ethos of phenotyping each exacerbation would enable more direct and appropriate therapy and the avoidance of inappropriate OCS exposure. Figure 1 outlines an approach to assessing a patient with severe asthma and an increase in symptoms. Firstly, to those who have increased symptoms but no significant fall in lung function. FeNO is useful to assess if the patient has uncontrolled symptoms due to T2 airways inflammation which needs further delineation as to whether this is due to poor inhaler technique or non-adherence meaning a lack of treatment delivery to the airways, or if the patient is not getting enough ICS and their maintenance treatment needs to be escalated. If a patient has increased symptoms in the context of stable lung function and a stable FeNO, other causes of the increased symptom burden should be considered and investigated as appropriate, and oral corticosteroids are not indicated for an acute asthma exacerbation (see Fig. 1).

Could We Do Things Differently to Reduce Inappropriate Oral Corticosteroid Exposure?
Moving to the left side of the algorithm, if there is a fall in lung function, assess T2 biomarkers available including blood eosinophils and FeNO. If markers of T2 biology are raised in the context of increased symptoms and a fall in lung function, corticosteroid treatment (usually OCS in severe asthma) is indicated to reduce T2 inflammation, and the opportunity should be taken to access inhaler technique, check adherence (which may include prescription counts of inhalers or digitalised adherence monitors such as FeNO suppression [56] or connected inhaler systems [57]), and lastly, to consider if the patient meets criteria for biological therapy, particularly if they already have good inhaler technique and are adherent to prescribed treatment. Our only treatment arsenal for acute exacerbations driven by T2 inflammation is with corticosteroids; going forward, our collective aim should be to either develop targeted, non-corticosteroid treatments for episodes of acute T2 airways inflammation, or to have asthma management strategies that achieve remission of asthma disease so negating the regular use of OCS for treatment exacerbations [58].
The last group to discuss is those who have increased asthma symptoms, a fall in lung function with suppressed markers of T2 biology. This is likely to be the smallest group of patients in this algorithm, but as described in the exacerbation phenotyping studies above, this group of patients is present in severe asthma cohorts receiving anti-IL5 treatment, and in non-biologics treated severe asthmatics. Although evidence shows that OCS are not effective in the absence of T2 biology, and that OCS use induces morbidity, to date, there has not been a trial that assesses outcomes in non-T2 exacerbations not treated with OCS.
In the study, phenotyping exacerbations in a mepolizumab-treated subgroup (the MEX study [26••], T2 low exacerbations were statistically more likely to be associated with a viral or bacterial pathogen on PCR than T2 high exacerbations, but there is a paucity of information on the mechanisms of asthma exacerbations in the absence of T2 biology. With current knowledge, it is worth considering whether infection is the driving factor, treating infection as appropriate and consider down titrating corticosteroid treatment or the addition of azithromycin in the case of recurrent exacerbations [59].  Fig. 1 Proposed approach to assessing acutely increased symptoms in an asthmatic patient. ^T2 biomarkers including FeNO and peripheral blood eosinophil count. *CRP, consolidative changes on chest Xray, positive sputum culture or viral swab. **This list is not exhaustive but serves to illustrate some differential diagnosis of increased symptoms with preserved lung function.

Stability of exacerbation phenotype
Multiple trials, including the two described previously, [25••, 26••] have shown that the phenotype of exacerbations in an individual is not entirely stable across multiple exacerbations in that individual. Approximately, 75% of exacerbations will have the same phenotype at first and second exacerbations, but making treatment decisions based on phenotyping would require phenotyping at each exacerbation.

Conclusion
To summarise, increase in respiratory symptoms in a patient known to have asthma can occur for many reasons, and is not always synonymous with asthma exacerbations. When they do occur, asthma exacerbations are heterogenous in nature with evidence of eosinophilic and non-eosinophilic inflammation underlying. These can be differentiated at the time of increased symptoms using biomarkers including blood eosinophils, Feno, and C-reactive protein. There is evidence that eosinophilic inflammation is very responsive to corticosteroid treatment, and non-eosinophilic inflammation less so. An urgent unmet research need is to assess the evolution of non-T2 asthma exacerbations not treated with OCS in a placebo-controlled trial. However, the holy grail of asthma treatment is to aim for disease remission where exacerbations of disease do not occur.

Conflict of interest
Dr McDowell has received speaker's honoraria from GSK and support for attending meetings from Cheisi, outside the submitted work. Dr. Busby reports grants from Astrazeneca, personal fees from Nuvoair, outside the submitted work. Prof Heaney has received payment honoraria from LGH has received non-financial support from GlaxoSmithKline during the conduct of the study; has received grants from Amgen, AstraZeneca, Medimmune, Janssen, Novartis, Roche/Genentech, GlaxoSmithKline, Boehringer Ingelheim, Aerocrine, and Vitalograph; is the lead for the UK MRC Consortium for Stratified Medicine in Severe Asthma in partnership with Novartis, Hoffman la Roche/Genentech, Evelo Biosciences, Sanofi, GlaxoSmithKline, AstraZeneca, Teva, Theravance, and Circassia; has received travel funding support for international respiratory meetings (institution remunerated) from AstraZeneca, Boehringer Ingelheim, Chiesi, GlaxoSmithKline, and Napp Pharmaceutical; has received project grant funding from Medimmune, Novartis UK, Roche/Genentech, and GlaxoSmithKline; and has received funding for clinical trials (institution remunerated) from AstraZeneca, GlaxoSmithKline, Schering Plough, Synairgen, Novartis, and Roche/Genentech. These are outside of the submitted work.

Human and animal rights and informed consent
This article does not contain new study data with human or animal studies performed by any of the authors.

Open Access
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.