About 40%–50% of cases of severe asthma are due to allergies. They begin in infancy, are clearly atopic and appear to progress from mild-moderate asthma, although some patients present the severe form from onset. The most severe cases within this cluster are those with a greater number of positive skin pricks or higher IgE levels in blood, those with a family history of asthma, and a longer time since disease onset. However, not all patients have eosinophil predominance; indeed, the most severe forms of the phenotype (in terms of lung function, frequency and intensity of symptoms and use of healthcare resources) are accompanied by increases in both eosinophils and neutrophils in sputum14 (evidence C-R2).

From a pathobiological point of view, severe allergic asthma is orchestrated by the activation of type 2 helper T-cells (Th2), the production of specific cytokines, IL-4, IL-5 and IL-13, and B-cell isotype switching to produce IgE, although it must be pointed out that not all atopic asthmas have a highly active Th2 component (“high” Th2 endotype). Periostine and fractionated exhaled nitric oxide (FeNO) have been shown to be good biomarkers in this “high” Th2 variant15 (evidence C-R2).

Over 25% of cases of severe asthma are thought to belong to this phenotype. They are characterized by persistent eosinophils in bronchial biopsies and induced sputum, despite high-dose IC or SC. Levels of eosinophils and cysteinyl leukotrienes are higher than in severe allergic asthma. In general, the disease occurs in patients older than 20 years, and may be preceded by symptoms of chronic rhinosinusitis and nasal polyps. A subgroup of these patients also develops intolerance to non-steroidal anti-inflammatory drugs (NSAIDs), designated as “aspirin-exacerbated respiratory disease” (AERD). Clinical symptoms are florid from onset, airway obstruction is considerable and exacerbations are frequent. These patients are generally less atopic than those in whom the disease appears at younger ages and they have little family history of asthma, but their IgE levels and FeNO measurements may be equally high. Its pathogenesis is related to changes in arachidonic acid metabolism (decreased production of prostaglandin E2 and increased cysteinyl leukotriene synthesis), associated with inflammation that appears to related with Th2 activation.15

This population mainly comprises women with a high body mass index and abundant clinical signs and symptoms. Onset occurs between 50 and 60 years of age, or even later. Sputum eosinophilia is not always detected; exacerbations are frequent and lung function is not overly compromised. The genetic basis and etiology of the disease are not well understood, nor is the role of hormonal factors, although onset after natural or surgical menopause has often been reported. The association between asthma and obesity raises numerous questions, and a number of mechanisms have been proposed as its cause: immunoinflammatory factors, mechanical factors, loss of corticosteroid efficacy, vitamin D deficiency, and the additional burden of other comorbidities, such as obstructive sleep apnea–hypopnea syndrome16 (evidence D-R2).

The natural history of adult-onset neutrophilic asthma is not well understood. Increased matrix metalloproteinase-9 levels have been found in bronchoalveolar lavage, and a history of smoking and chronic airflow limitation (CAFL) with significant air trapping have also been observed. SC provide little disease control. Neutrophil predominance in the airways may be due to modifications in the expression of genes responsible for their activation and mobilization, concomitant diseases (bronchiolitis obliterans), or residual airway inflammation after continuous use of SC, inhibiting neutrophil apoptosis.15

A diagnosis reached in the past by other clinicians and provided by the patient must not be accepted unless the objective test results necessary for confirming the diagnosis are available. This attitude has been supported in several studies which found that a previous diagnosis of asthma was overturned in almost one third of patients after a new systematic assessment20 (evidence D-R2).

To date, there is no generally accepted definition of what has become known as asthma phenotypes, but there does seem to be consensus with regard to certain clinical profiles of patients with distinct pathophysiological characteristics that cause them to respond better to certain medications. Attempting to define these phenotypes may be of use in optimizing treatment, and to this end, certain clinical characteristics may be of interest, e.g., age of onset, obesity, NSAID intolerance, in addition to laboratory parameters, such as eosinophils in induced sputum and blood, FeNO determinations, and skin prick tests (Fig. 2).

Although SC are very effective in controlling inflammation in asthma, the spectrum of response to these agents among asthmatics is broad, and a small percentage of subjects fail to respond at all, despite the administration of high doses. Total resistance to SC is uncommon; partial resistance or “lack of sensitivity” is more common among severe asthmatics, making high SC doses necessary, although not even with these is full control achieved. Some patients do improve notably with the use of SC, but when they are discontinued, the asthma worsens (corticosteroid-dependent asthma). In a study of 102 SUCA children, only 11% did not respond to an intramuscular (IM) dose of triamcinolone, suggesting that 89% had some degree of response to SC28 (evidence D-R2). Accordingly, the recent ATS/ERS consensus1 uses the expression “SC insensitivity” instead of “resistance”, and we have adhered to this nomenclature in these guidelines.

From a clinical point of view, SC-insensitive asthma is defined by forced expiratory volume in the first second (FEV1) less than 75% of the predicted value and a response <15% and 200ml after the administration of a 2-week cycle of 40mg/day prednisone or prednisolone29 (evidence D-R2). These patients have normal plasma cortisol and adrenal suppression response to exogenous cortisol, so they are prone to the side effects of these agents. Factors contributing to asthma insensitivity to SC are not well defined. Some studies report weak associations between certain genetic changes and environmental factors (continuous exposure to allergens, smoking, NSAID intolerance, low vitamin D levels, and chlamydia, mycoplasma or viral infections), but these cannot be considered as clearly established risk factors. The most significant environmental factor in SC insensitivity is exposure to tobacco smoke. This may act by various mechanisms: by altering the inflammatory pattern (increased neutrophils and CD8 lymphocytes and reduced eosinophil levels), by impairing mucociliary function, thus allowing excessive mucous deposition in the airways, or by tobacco-induced oxidative stress, which inactivates histone deacetylase, reducing nuclear translocation and the number of corticosteroid receptors and their affinity30 (evidence D-R2).

SC act at different levels, but their greatest anti-inflammatory effect is produced by inhibition of the genetic transcription of numerous genes that code for proinflammatory proteins (cytokines, chemokines, adhesion molecules) and by increased transcription of anti-inflammatory mediators. They act by binding to a specific intracytoplasmic receptor with 2 molecular isoforms (α and β). The β receptor cannot bind to the hormone, and, as such, is inactive31 (evidence D-R2). Lack of sensitivity to SC cannot be explained by pharmacokinetic changes or malabsorption, and is probably due to a range of different mechanisms32 (evidence C-R2) (Table 5).

Response to IC varies notably among individuals11 (evidence C-R2). However, there is some evidence to suggest that SUCA patients may respond to higher doses than normally recommended33 (evidence A-R1). Although some studies support the greater therapeutic efficacy of fine particle IC (related with their effect on the peripheral airways), there is no evidence of their superiority in SUCA34 (evidence C-R2).

Among the new glucocorticoids, ciclesonide has fewer local and systemic side effects since it is a prodrug converted to its active form in the lung parenchyma35 (evidence B-R2). New steroids, known as “dissociated” compounds (mapracorat), that aim to separate the anti-inflammatory mechanisms from the side effects, are currently under development36 (evidence D-R2).

The best time for introducing maintenance treatment with SC is not well defined, nor is there evidence that continuous treatment with low-dose SC is more effective than cycles of SC in reducing the number of exacerbations. IM triamcinolone administration (Trigon® depot 40mg) in asthmatics with corticosteroid insensitivity improves control, reduces eosinophils in sputum, increases FEV1 and prevents exacerbations. Reasons for its efficacy may include reinforced compliance or the higher dosage of triamcinolone compared to other corticosteroids used in the clinic37 (C-R2).

Adding a LABA to an IC has been shown to be more effective than doubling the dose of the IC or adding an antileukotriene, although there may be notable variability in response, which needs to be monitored38 (evidence A-R1). Formoterol as a complete agonist has greater intrinsic efficacy and causes a greater number of adverse effects39 (evidence D-R2), the most common of which are tachycardia and hypokalemia, which can be more pronounced in individuals homozygous for arginine in position 16 of β2-AR.

Recent studies have shown that long-acting anticholinergics (LAMA) may be useful in patients with severe asthma and concomitant CAFL40 (evidence B-R2), in cases of ACOS, in severe asthma with a non-eosinophilic inflammatory profile41 (evidence D-R2), and in asthmatics with the ArgGly variant in codon 16 of the β2-receptor42 (evidence B-R2). There is an increasing trend in the use of LAMA as a treatment for bronchial asthma, and some CPG suggest their use in the higher stages of severity, when control cannot be achieved.

Levels of 25-hydroxyvitamin D<30ng/ml are related with asthma severity, increased risk of exacerbations and poor lung function, and are also implicated in the mechanisms of steroid insensitivity. However, no definitive results are available to support supplementation with this vitamin43 (evidence D-R2).

Treatment with macrolides for 3 weeks or more in severe asthma is not associated with a significant improvement in FEV1, although improvement has been reported in morning PEF, symptom control, bronchial hyperreactivity, and health-related quality of life (HRQoL)44 (evidence D-R2). There is insufficient evidence on generalized use for reducing the number of exacerbations, although some studies recommend macrolides in the severe neutrophilic asthma phenotype45 (evidence B-R2).

Patients with AERD generally have excessive basal leukotriene production, so would appear to be ideal candidates for antileukotriene treatment46 (evidence C-R2). Some recent studies report findings in air trapping and radiological changes on CT that suggest montelukast may be useful, in general, in patients with severe asthma, so it may tried as an add-on therapy in these patients47 (evidence C-R2).

Anti-IgE treatment has shown clinical efficacy in inflammatory response markers, frequency of exacerbations, visits to emergency departments, severity of symptoms, use of IC, and HRQoL. Omalizumab is indicated for improving asthma control when administered as add-on treatment in adults and older children (over 6 years of age) with persistent severe uncontrolled allergic asthma with perennial allergies, reduced lung function and documented severe exacerbations, despite appropriate treatment for their level of severity48 (evidence B-R2). Data from non-atopic patients receiving omalizumab have been published, which could open a new avenue for treatment in this patient group49 (evidence D-R2).

Although early clinical studies performed in asthmatics classified as moderate and severe have not shown particularly favorable results, due to adverse events and the lack of effect on bronchial hyperreactivity50 (evidence B-R2), it is not yet clear which subgroup of patients might benefit from this technique. A recent study, with a 5-year follow-up, showed a sustained improvement in disease control (reduction in number of severe exacerbations and visits to the emergency room) and confirmed the safety of the procedure51 (evidence B-R2). It is currently recommended only in experienced units and in the research setting1,52 (evidence D-R2)53 (evidence B-R2).

Although no controlled trials have been performed in severe asthma, theophylline as a single agent has relatively weak anti-inflammatory activity, but at low doses it can markedly enhance the action of corticosteroids on the expression of inflammatory genes54 (evidence D-R2).

Data are also available on the utility of the new phosphodiesterase-4 inhibitors (roflumilast) in asthmatics55 (evidence B-R2) and the 2014 Spanish COPD Guidelines (GesEPOC) update continues to suggest these drugs as add-on treatment in ACOS patients with uncontrolled symptoms56 (evidence D-R2CPG).

Studies have confirmed the unfavorable risk/benefit profile of etanercept, methotrexate and cyclosporin.

An anti-IL5 monoclonal antibody, mepolizumab, has been recently investigated in patients with eosinophilic asthma, showing significant improvements in HRQoL, exacerbations and symptoms57 (evidence A-R1)58 (evidence B-R2) and reduced use of SC59 (evidence B-R2). Studies with reslizumab and benralizumab (anti-IL-5 receptor) are opening up new prospects60 (evidence B-R2).

Another immune activation pathway in asthma is IL-13. Molecules are already under development with this target in mind: tralokinumab and lebrikizumab with anti-IL-13 effect, and dupilumab61 (evidence B-R2) which is directed against the alpha subunit of the IL-4 receptor, a target shared by both IL-4 and IL-13. Lebrikizumab showed a greater positive effect in patients with elevated pre-treatment levels of circulating periostine62 (evidence B-R2). Trials have been conducted with pitrakinra, an IL-4 mutein, that works as an antagonist by inhibiting IL-4 and IL-13 binding to the shared IL-4Rα/IL-13Rα1 complex63 (evidence B-R2). Daclizumab, a humanized IgG1 monoclonal antibody against the IL-2R-α chain of activated T-lymphocytes, has been shown to be useful in moderate and severe asthma inadequately controlled on IC64 (evidence B-R2).

Molecules that act on the IL-17 (brodalumab) and the IL-17A receptor (secukinumab) in neutrophilic asthma are currently under investigation, since IL-17 induces IL-8 expression, the main neutrophil-attracting factor, but results to date have been disappointing65 (evidence B-R2).

If the inflammatory phenotype is identified, treatment response can be predicted. In this respect, there is sufficient data to confirm that eosinophilic inflammation has a high predictive value for response to IC in patients with this type of airway disease, irrespective of their diagnosis. There is evidence for adjusting treatment in SUCA patients on the basis of induced sputum testing (in experienced clinics), although this could not be shown for FeNO66 (evidence A-R1).

Triamcinolone could be an alternative in patients with refractory eosinophilia,28 and its effect may be enhanced with low-dose theophylline.54 Therapeutic options, such as omalizumab, are available, and mepolizumab, reslizumab, lebrikizumab or dupilumab could be used in this patient group in the near future.

Some selected patients with neutrophilic asthma may be candidates for macrolide treatment.44

Tiotropium is another alternative for patients with CAFL.41 Finally, thermoplasty can be considered in carefully selected patients in experienced units (Fig. 3).

Fig. 3.

Treatment proposal for severe, uncontrolled asthma according to phenotype. ACOS: asthma/COPD overlap syndrome; AERD: aspirin-exacerbated respiratory disease; CAFL: chronic airflow limitation.

(0.22MB).

Other diagnostic possibilities must be considered, and extra tests performed, according to the degree of suspicion (Table 2).

SUCA was ruled out in 75% of children after a review of inhalation techniques and treatment compliance69 (evidence D-R2).

Both repeated exposure to airborne allergens to which the child is sensitized and exposure to tobacco smoke or atmospheric pollution contribute to poor asthma control, due to increased bronchial reactivity and reduced response to SC. If these factors are combined with a respiratory viral infection, severe asthma exacerbations may be easily triggered.

Rhinosinusitis, obesity, gastroesophageal reflux and psychosocial factors worsen asthma, although their degree of influence and mechanism of action are disputed. Food allergies behave more as a marker for severity, while breathing dysfunction syndromes may present concomitantly and lead to overtreatment if they are not correctly identified.

After analyzing the above-mentioned factors, an attempt should be made to determine if the patient has corticosteroid-sensitive SUCA (the most common pediatric presentation), or not (even if the patient has an eosinophilic inflammatory pattern in the airway), or if they have a neutrophilic or mixed pattern, or if there is CAFL.

To this end, the various tests listed in Table 6 should be scheduled in a series of visits. In the first visit, allergy tests should be repeated, to determine if the patient has become sensitized to other airborne allergens, and cotinine in saliva should be analyzed to detect exposure to tobacco smoke70 (evidence D-R2CPG).

To confirm corticosteroid response, and since there is no agreement as regards the optimal dose, duration and route of administration in children, we suggest using triamcinolone, 40–80mg IM, depending on age and weight71 (evidence D-R2). Response should be confirmed 2 weeks later (Table 7). Responders are patients who improve in all 3 domains, non-responders are those who do not improve in any domains, and partial responders are those who improve in 1 or 2 domains. The latter category is the most common.37

Patients who respond to SC may benefit from omalizumab, which has been shown to be safe and effective in comparative trials in children over the age of 6.48 Any other treatment should also be maintained, including SC, until symptoms are controlled.

Non-responders who have persistent neutrophilic airway inflammation may benefit from macrolides, although there is insufficient evidence. Use of azithromycin has been suggested with the same regimen as for cystic fibrosis: 250mg/day for children weighing <40kg and 500mg/day for those >40kg, 3 times a week for 6 months, followed by a reevaluation of efficacy72 (evidence A-R1). Low-dose oral theophylline may also be tried: target blood concentrations are 5–10mg/l.