Severe asthma is an ‘invisible’ condition not readily observable to others.1 This condition is not just ‘bad’ asthma, but a serious threat to life as it involves physical, mental and social impairment. Many patients with severe asthma live with fears, switch or leave their jobs, and have difficulties with friend and family relationships.23 Despite high-intensity medications and good adherence to them, the burden of severe asthma is not successfully reduced.

A major burden of severe asthma is the future risk of adverse health outcomes. Patients with severe asthma are prone to exacerbation and serious deterioration in lung functions, and may also experience side effects from medications such as systemic corticosteroid (SCS). Hence, the European Respiratory Society (ERS)/American Thoracic Society (ATS) guidelines4 define this disease not only as asthma which remains “uncontrolled” despite aggressive drug therapy (such as high-dose inhaled corticosteroid [ICS] plus second controller and/or SCS), but also one that requires aggressive therapy to prevent from becoming “uncontrolled.” In essence, therefore, “controlled” severe asthma is considered a severe form of asthma due to the future risk of serious adverse events.4

Future risk factors for severe asthma are not easily detected in daily clinical practice. In particular, currently available tools to measure asthma control and asthma-related quality of life incompletely predict the future risk of medication-related morbidity, which is a significant issue in asthma management.5 In our clinical experience, the future risks of severe asthma are poorly recognized by patients themselves and even physicians, especially when asthma control is relatively well maintained by high-intensity treatments including intermittent oral corticosteroid (OCS) bursts. Therefore, we review here future risk existing in patients with severe asthma. This review starts with the question – “What are the possible future risks in severe asthma, and why is it important to address?” As the risk is poorly perceived by controlled patients, our review will more focus on the risks in patients with ‘controlled’ severe asthma.

The concept of asthma severity has evolved during the last decade. The word ‘severity’ in the context of asthma had been used previously to describe the degree of airflow obstruction, symptoms, or exacerbations. In the Global Initiative for Asthma (GINA) guidelines 1995, asthma severity was used to classify the off-treatment status of a patient (such as intermittent, mild persistent, moderate persistent and severe persistent).6 This classification was used to determine initial treatment strategy. However, current clinical practice guidelines no longer classify asthma in terms of off-treatment severity, but are aimed at achieving disease control, which is now possible in most patients irrespective of their initial levels of off-treatment severity.7 Asthma severity is currently assessed by the intensity of treatment required to achieve good asthma control (symptoms and exacerbations) and, in very severe cases, by treatment refractoriness (Fig. 1).4 In this review, the definition of severe asthma followed the recommendations of the 2014 ERS/ATS guidelines.4

Asthma control refers to the extent to which the manifestations of asthma are reduced or removed by treatment.8 The assessment of asthma control incorporates 2 components - current symptoms and future risk of adverse health outcomes (such as exacerbations, fixed airway obstruction and drug side effects) (Fig. 1). Drug side effects can result from the use of a high-intensity treatment to achieve asthma control. However, it is important to note that current standard tools for asthma control, such as the asthma control test (ACT) or asthma control questionnaire (ACQ), do not include any items that address any future risk of drug side effect, even though this is a major burden on patients with severe asthma.5

Due to the evolving concepts of asthma severity, the definition and prevalence of severe asthma varies among published articles. In an observational study using the pharmacy database combined with a questionnaire survey in the Netherlands, it was estimated that 3.6% of adult asthmatics have severe refractory asthma as defined by a high-intensity treatment requirement, poor control, and good adherence and inhaler technique.9 When applying the ERS/ATS guideline definition,4 the prevalence of severe asthma has increased from 3.6% to 4.5%, suggesting that the prevalence of controlled severe asthmatics was about 0.9% among adult asthmatics in the community population.

In a Swedish general population-based study of 744 subjects with “active asthma” (as defined by asthma symptoms or medication in the prior year), 36.2% of the asthmatics had at least one sign of asthma severity, i.e., 1) multiple symptoms during the day despite the ongoing use of asthma medications; 2) an FEV1 of less than 70% of the predicted value; 3) daily, or almost daily, use of rescue medications; 4) night-time symptoms once a week or more; and 5) OCS use/emergency department (ED) visits.10 Importantly, the prevalence of OCS use/ED visits was 9% among active asthmatics, and about 13% reported at least 2 signs of severity,10 suggesting a potential burden of future risks related to severe asthma existing in asthmatic patients even in the general population.

In a systematic review of the epidemiology and burden of uncontrolled severe asthma, the percentage of patients with uncontrolled severe asthma among all asthmatics varied widely from 0.7% to 49.2%.11 There was a substantial heterogeneity in the tools used to define asthma control in the studies included, such as ACT, ACQ, and international guideline- or exacerbation history-based definitions. Of note, the percentage of uncontrolled severe asthmatics among overall severe patients in that study ranged from 14.2% to 79%, suggesting that the frequency of controlled severe asthma was also substantial.11

There are no solid longitudinal data on natural disease courses in patients with severe asthma or at the control status. The available evidence suggests that the severity of asthma may be an inherent characteristic of the affected patient. In a 12-month longitudinal follow-up study, a loss of control is common in severe asthmatics, but poor control is likely to be persistent over time.12 Several host factors, such as comorbidities and genetics, may interact with socio-environmental factors in determining the likelihood of persistence in asthma severity and these traits are not easily treatable or modifiable.

Patients with severe asthma frequently experience poor control.13141516171819 In the Epidemiology and Natural History of Asthma: Outcomes and Treatment Regimens (TENOR) II cohort study,19 a follow-up survey of 341 patients with severe or difficult-to-treat asthma more than 10 years after enrolment in the TENOR I study, more than half of the subjects (58.1%) were still very poorly controlled, and only 7.7% were well controlled.19

The relevance of current asthma control to future clinical outcomes has been consistently demonstrated in the literature. In the TENOR I study (severe persistent asthma cases, 54.1%), the asthma control score showed significant correlations with the health-related quality of life measure at baseline and during longitudinal follow-ups.20 In addition, the control status of asthma was significantly correlated with the number of OCS courses, unscheduled clinic visits, ED visits and hospitalizations.21 These correlations remained significant when adjusted for the baseline asthma severity (off-treatment).2021 In a prior post hoc analysis of a randomized clinical trial (RCT) with an interleukin (IL)-4 receptor α antagonist in patients with moderate to severe atopic asthma, an increase in the ACQ score by 1 increased the future risk of exacerbations by 50%.22 Moreover, in a retrospective analysis of 5 budesonide/formoterol maintenance and reliever therapy trials, the current control status predicted the control status in the following week with at least 80% probability and was also significantly associated with future exacerbation risk.23 In a 24-month longitudinal follow-up in the TENOR I study, very poor control was significantly associated with future risk of exacerbations.24 These findings collectively indicate that the current asthma control status is highly relevant to the future risk of adverse clinical outcomes and that expected future risk may differ between controlled and uncontrolled severe asthmatics (Fig. 2).

Exacerbation is a defining feature of severe asthma, which is defined by a worsening of symptoms requiring the use of SCS or an increase in the SCS doses in patients on a stable maintenance dose, to prevent serious outcomes.25 In a cross-sectional assessment of the TENOR II cohort participants with severe or difficult-to-treat asthma, 48.2% of the cases experienced exacerbation in the recent 12 months and 21.3% had severe exacerbations (defined as hospitalization or ED visit due to asthma requiring SCS).19 In the Australian Severe Asthma Web-Based Database, severe asthmatic patients experienced a median of 2 exacerbations during the 2-year follow-up.26

In a 5-year prospective longitudinal follow-up study of 177 patients with “problematic asthma” (as defined according to the Innovative Medicines Initiative27 which encompasses both difficult-to-treat asthma and severe asthma) from Singapore, there were 3 distinct trajectories described in terms of severe exacerbation rates: “infrequent exacerbators” (58.5%; stable disease with few intermittent severe exacerbations), “non-persistently frequent exacerbators” (32.0%; frequent severe exacerbations at baseline, but improved subsequently), and “persistently frequent exacerbators” (9.5%).17 The most unfavourable trajectory group had an annual mean incidence of severe exacerbation of around 4 person-years, whereas the other 2 groups showed a less than 1 person-year incidence at 5 years. The unfavourable group characteristics included a higher body mass index, more frequent history of near-fatal asthma or severe exacerbation as well as more frequent gastroesophageal reflux disease (GERD), obstructive sleep apnoea, and depression than the other 2 trajectory groups.17 However, no longitudinal correlations with asthma control status were presented.

It is expected that the future risk of exacerbations will be lower in severe controlled than uncontrolled asthmatics, but the precise incidence of this adverse event in controlled severe asthma has not been reported.

Overall, a history of exacerbation is one of the strongest predictors of the future risk of exacerbation events in different asthmatic populations.2628293031323334 Various host factors, such as smoking history, poor lung function, nasal polyps, obesity, autoimmune disorders, GERD/vocal cord dysfunction, and comorbid depression, are also associated with the future risk of exacerbation in patients at various ages and with different asthma severity.263435363738394041 In patients with severe asthma, uncontrolled type 2 airway inflammation is also significantly associated with high exacerbation rates.42434445 These findings collectively indicate that systematic assessments and interventions are required to achieve proper asthma control and to reduce the future risk of exacerbations in severe asthma cases.

Only a few studies to date have examined the longitudinal changes in lung function in severe asthmatics. In a 6-year follow-up study of 97 patients with severe asthma in the Glenfield cohort, the recorded annual mean decline in the post-bronchodilator FEV1 was −25.7 mL/year.46 Of note, there were 3 clusters in that study differing in terms of eosinophilic airway inflammation and the rate of post-bronchodilator FEV1 decline, of which the cluster with low-grade but positive sputum eosinophilia (geometric mean eosinophils%, 4.28% [95% confidence interval {CI}, 1.07%–7.49%]) and large variations over time showed the highest rate of lung function decline (−40.9 mL/year).46

The longitudinal course of lung function was also examined over a period of 10 years in a prior retrospective study of 54 Japanese patients with severe asthma (as defined by the ATS criteria for refractory asthma47).48 The mean daily ICS dose was 1,830 mcg of beclomethasone dipropionate equivalent, the mean OCS use duration was 67.3 ± 62.9 days/year and annual exacerbation rate was 0.48 ± 0.51. Two phenotypes emerged in that study according to the rapidity of lung function decline, rapid (n = 17) and slow (n = 37), based on a forced vital capacity (FVC) decline rate of −20 mL/year. However, the mean annual rate of change in the FEV1 was +0.6 mL/year. Annual exacerbation rate, OCS use and age were positively associated with the rapid FVC decline in that Japanese cohort.48 However, the possible relationship between lung function decline and asthma control status was also not examined in detail.

In an 8-year longitudinal observation study of 234 patients with severe asthma from an under-privileged area in Brazil, the mean annual decrease in post-bronchodilator FEV1 measured −27.1 (95% CI, −51.1–1.4) mL/year.49 The subjects in this study were receiving appropriate treatment from specialists, and the daily ICS dose was 1,295.5 mcg (of beclomethasone dipropionate equivalent). OCS was required in 62% of these patients during the last year of observation. The percentage of patients with an ACQ score of ≥ 1.5 decreased from 60% at the initial presentation to 37%, and that with an ACQ score of ≤ 0.75 increased from 19% to 42%. The authors noted that lung function decline was not completely prevented by currently available and best-practice treatments, although asthma control was improved by these interventions.49

These longitudinal findings suggest that lung function may continue to decline in severe asthmatics despite high-intensity treatment and that improved asthma control is not sufficient to prevent such progressive deterioration.

It is estimated that about 1% of the general population receives SCS treatments at any time point and respiratory conditions, including asthma, are reported to account for a major proportion (about 25%–40%) of these prescriptions.505152 It is well known that the use of SCS in patients with different health conditions is positively associated with higher risk of psychiatric events, infection, gastric problems and fractures.5354 Until recently, however, morbidities from SCS exposure had not been described in asthmatic populations. This knowledge gap was a significant issue, as susceptibility to developing drug-related morbidities may differ according to age, sex, underlying disease, and doses or durations of drug exposure.

Current findings on SCS-induced morbidity in asthmatic populations are summarized in Table. A major finding from these studies is that the risk of complications may increase even at low-dose SCS exposures. The study by Lefebvre et al.55 was the first to investigate the risk of SCS-induced morbidity in patients with severe asthma who had more than 6 months of continuous SCS use. By examining the Medicaid health insurance claims database (1997–2013) in the United States, the authors investigated the positive dose relationships between SCS exposure (≤ 6 mg/day, 6–12 mg/day, and > 12 mg/day of prednisolone equivalent) and the incidence of related complications. Patients with higher exposures to these drugs (> 6 mg/day) had significantly higher risk of complications, including cardiovascular, metabolic, psychiatric, ocular, gastrointestinal and bone disorders as well as infections (odds ratios [ORs] ranging between 1.23 and 2.12 by complication).55

In another study that utilized US administrative claims database analysis of severe asthmatic patients and matched controls, the ORs of developing SCS-related complications were found to increase significantly in a dose-dependent manner. Importantly, even the patients with low-dose exposures (< 5 mg/day of prednisolone equivalent) showed higher risks of complications compared to those without any SCS exposure (OR, 2.50; 95% CI, 1.22–5.10).56 In a further report that analyzed the US MarketScan claims database (2000–2014) for asthmatic patients with a recent history of severe exacerbation (asthma-related ED visit or hospitalization), the number of OCS prescriptions was positively associated with greater odds of osteoporosis, hypertension, obesity, type 2 diabetes mellitus (DM), gastrointestinal ulcers/bleeding, fractures and cataracts (ORs, 1.21–1.44 by complication).57

In the UK, primary care databases have been utilized to assess the risk of SCS-induced morbidity in asthmatic patients. Analyses of the Optimum Patient Care Research Database (OPCRD) has shown that patients with severe asthma and with a median cumulative OCS dose of 3,920 mg over the 2-year study period (interquartile range [IQR], 2,395-6,500 mg) had more comorbidities related to SCS exposure such as type 2 DM, osteoporosis, dyspeptic disorders or cataracts, than those with mild/moderate asthma or non-asthmatic controls.58 Similar findings were observed in another UK study of patients in the British Thoracic Society (BTS) Difficult Asthma Registry. It was notable that 57.4% of the patients in the BTS registry had corticosteroid-dependent severe asthma and had significantly greater occurrence of type 2 DM, hypertension, hypercholesterolemia, obesity, obstructive sleep apnoea, dyspeptic disorders and depression/anxiety/low mood than non-corticosteroid-dependent severe asthmatics (42.6% of the registry who did not require regular maintenance of OCS, but required frequent rescue courses of corticosteroids).58 However, the latter group also frequently had treatment-related complications: 75% and 20% of these patients had ≥1 and ≥3 corticosteroid-related morbidities, respectively.58 In another report involving extended analyses of the OPCRD for age and sex associations, high-dose OCS-related morbidity was more frequent in younger patients (e.g., the OR for type 2 DM was 4.7 [95% CI, 1.5–15.1] in patients ≤45 years, but was 1.2 [95% CI, 0.7–2.0] in those aged 61–70 years).59 Other retrospective cohort analyses utilizing the Clinical Practice Research Datalink or OPCRD demonstrated that the risk of complications was positively associated with cumulative doses, daily average doses, and frequencies of OCS use.606162 Of note, increased risk of a few adverse outcomes (such as cerebrovascular accident, type 2 DM, or depression/anxiety) was observed at a very low threshold of cumulative SCS exposure (0.5–1 g of prednisolone equivalent; vs. 0–0.5 g as the reference), which is equivalent to just 4 lifetime SCS courses.62

ICS is generally considered safe but can cause systemic side effects at chronic high-dose exposures, such as hypothalamic-pituitary-adrenal axis suppression, osteoporosis, DM and ophthalmic complications.63 In a prior longitudinal follow-up study of 196 adults with asthma (aged 20–40 years), a higher cumulative ICS dose (median duration of exposure: 6 years [range, 0.5–24 years]) was inversely correlated with the bone-mineral density.64 In a recent retrospective cohort study of severe asthmatic patients in Korea, cumulative doses of SCS and ICS were significantly correlated with trabecular bone score.65 In another systematic review of case-control studies of non-vertebral fracture risk in adult asthmatics, the relative risk of fracture increased by 12% for each 1,000 µg per day administration of beclomethasone dipropionate or equivalent.66 The cumulative dose of ICS was also significantly correlated with the risk of cataract extraction in a further study.67

Asthma mortality has decreased in recent decades, which is quite possibly related to the introduction of ICS therapy. However, the risk of mortality may still exist in patients with severe asthma. In a recent 20-year follow-up study of 52 patients with severe asthma (baseline age: 49.9 ± 18.9 years) at an expert clinic in France, the mortality rate was 50% (19% attributed to fatal asthma) and the relative survival rate (compared to the expected survival rate adjusted for location) was 0.778 (95% CI, 0.57–0.99), indicating an excess risk of death of 22% in this severe asthmatic population.68 In another study from France involving a 3-year longitudinal follow-up of severe asthmatics in a representative claims database, the 3-year cumulative mortality rate was significantly higher in severe asthmatics than in matched non-asthmatic controls (7.1% vs. 4.5%; P = 0.007).69 In a nested case-control study of severe asthmatic patients from Brazil (58 mortality and 232 non-mortality cases, at a 1:4 ratio), most of the deaths were from respiratory causes (60.3%). In the multivariate analyses in that study, male sex, poor asthma control and a lower pre-bronchodilator FEV1 < 60% of predicted were significantly associated with higher mortality risk.70

As reviewed above, the future risks of adverse outcomes in patients with controlled severe asthma have not been completely studied. These patients may belong to a “grey zone,” because their future health risk is not readily measurable but is also poorly perceived by this group. For this reason, there is an important question as to whether controlled severe asthma is completely being controlled. At the very least, the answer is ‘No’ in terms of the future risk of drug side effects. The risk of drug side effects increases even at low-dose SCS exposure. In addition, it is likely that chronic corticosteroid therapy will not prevent future asthma progression. Lung function decline was not prevented in severe asthmatics despite improvements in asthma control with higher-intensity treatments.49

Considering frequent drug side effects in severe asthmatic patients, novel treatments are highly desirable that will enable the tapering of OCS use without any loss of asthma control.71 T2-targeted biological therapies, such as monoclonal antibodies targeting IL-5 and IL-4 pathways, have been demonstrated to reduce OCS maintenance doses in patients with OCS-dependent severe asthmatics.727374 However, although the OCS-sparing benefits of these biologics were superior to those of a placebo, they were still partial in terms of absolute responses. In a 24-week placebo-controlled trial of 135 patients with OCS-dependent severe eosinophilic asthma (median daily OCS dose at baseline: 12.5 mg of prednisone equivalent in the treatment group), the median reduction in OCS dose was 50% from baseline (range: 20%–75% reduction) in the mepolizumab treatment group, compared to 0% (range: 20%–33.3% reduction) in the placebo group (P = 0.007).72 The reduction in daily OCS dose to a level of ≤ 5 mg occurred in 54% of the mepolizumab group and in 32% of the placebo group (OR, 2.45; 95% CI, 1.12–5.37; P = 0.02).72 In a 28-week RCT of 220 patients with OCS-dependent severe eosinophilic asthma (mean daily OCS dose at baseline: 10 mg of prednisolone or equivalent [range: 7.5–40 mg]), the OR of OCS dose reduction was 4.12 (95% CI, 2.22–7.63) with benralizumab compared to placebo. However, the mean OCS dose at the final visit was 5 mg (range: 0–30 mg) in the every-8-week benralizumab treatment group and 10 mg (range: 0–40 mg) in the placebo group.73 The percentages of patients with a reduction in the daily OCS dose to a level ≤ 5 mg was 59% in the benralizumab treatment group and 33% in the placebo group (OR, 2.74; 95% CI, 1.41–5.31; P = 0.002).73 In a 24-week RCT of OCS-dependent severe asthma, dupilumab treatment reduced the OCS dose by 70.1% from baseline, which was significantly higher compared to placebo treatment (41.9% reduction; OR, 3.25; 95% CI, 1.90–5.55; P < 0.001).74

Risk reduction by T2-targeted biologics in terms of annual severe exacerbation rates (compared to placebo) reportedly ranged from 32% to 70% (all statistically significant) in these trials.727374 However, exacerbations still occurred. The annual severe exacerbation rates were 1.44 per year in the mepolizumab treatment group (vs. 2.12 in the placebo group),72 0.54 per year in the benralizumab treatment group (vs. 1.83 in the placebo group),73 and 0.649 in the dupilumab treatment group (vs. 1.597 in the placebo group).74 The mean improvement in the ACQ score by the use of these biologics ranged between −0.52 and −0.47 (compared to placebo; all statistically significant).727374 However, their benefits on pre-bronchodilator FEV1 were weaker (mean improvement range: +112 to +220 mL from baseline ; statistical significance not consistent).727374

While the effects of these new biologics have only been documented after 1 year of treatment, their long-term effects are unclear. Mepolizumab maintained its beneficial effects on reduction in exacerbations over a 4-year period, but did not prevent yearly decline in FEV1,75 but there was no placebo-controlled group to determine whether the decline in FEV1 observed over the 4 years of investigation is attributed to the inability of mepolizumab to interfere with the future risk of lung function decline in severe asthmatics.

Overall, new T2-targeted biologics are promising in reducing the risk of exacerbations especially for corticosteroid-dependent asthmatics where there is a significant reduction in maintenance OCS doses. Further progress in reducing future risks will come from research into T2-low-associated mechanisms,76 which can reduce the future risks listed in this article, particularly those associated with chronic airflow obstruction and decline in lung function.

Severe asthma is a life-threatening condition. In addition to its direct impact on daily life, it may also pose a substantial burden on future health outcomes. However, the future risk is not readily measurable in daily practice and may be poorly perceived by many patients and their physicians. Patients with severe asthma are at very high risk of developing drug side effects due to high-intensity treatment requirements for controlling asthma activity. Importantly, the increased risk of OCS-induced complications is observed even at very low-dose exposures to these agents. In addition, it is not clear whether long-term OCS therapy prevents future asthma progression. Thus, controlled severe asthma is not completely being well controlled in terms of future risk. Newly developed T2-targeted biologics significantly reduce the OCS maintenance dose and exacerbation risk, and also improve asthma control status in patients with OCS-dependent severe asthma. However, their benefits are still only partial in terms of the response rate. These findings indicate that significant but unmet clinical needs remain in patients with severe asthma, even when it appears to be under control with high-intensity treatment. Further studies are warranted to develop novel treatments in order to further reduce the future risks, particularly chronic airflow obstruction and lung function decline. A practical tool to address the future risks in these patients should also be developed.