A pragmatic guide to choosing biologic therapies in severe asthma

Primary outcomes: exacerbation and OCS reduction

The phase 3 trials of omalizumab focused their inclusion criteria on patients who had evidence of sensitisation to aeroallergens, airway hyperresponsiveness and ongoing asthma symptoms. Omalizumab treatment resulted in a reduction in exacerbations as well as reductions in hospitalisations, emergency department visits and inhaled steroid dose [8–10]. A Cochrane review published in 2014 included data from 10 studies directly addressing annualised exacerbation rate (AER) and reported an overall risk ratio of 0.55 compared with placebo [11]. The reduction in AER is difficult to compare to later biologic trials in asthma, as most omalizumab studies did not specify a minimum pre-treatment AER, ICS dose was often weaned as part of the trial design (increasing the risk of exacerbation) and an exacerbation was not consistently defined [10]. Study populations are therefore heterogeneous. Only one study attempted to limit recruitment to severe asthma patients only by including a requirement for high-dose ICS/long-acting β₂-agonist (LABA) combination therapy [12]. However, even in this study the AER of 0.88 in the placebo arm suggests a less than severe cohort were recruited in the end. In the most severe subgroup who required mOCS at baseline, omalizumab was not effective at reducing the exacerbation rate.

Most of the available data on mOCS dose reduction in omalizumab comes from real-world, observational trials [13–17]. A systematic review of 42 observational or registry trials reported a mean OCS dose reduction of 68% at 12 months, however with a broad range from −78% to −12% [18]. The large “eXpeRience” registry reported on over 260 patients on baseline OCS and observed that 57% either reduced or stopped mOCS by 1 year of treatment with omalizumab. The mean prednisolone dose at the end of 1 year was still 7.7 mg (compared with 15.5 mg at baseline) [19]. A randomised trial reported more patients reduced or stopped mOCS with omalizumab than with optimal standard care; however, this was an open label study [20]. The true corticosteroid-sparing efficacy of omalizumab in severe asthma remains to be determined.

Secondary outcomes: quality of life, asthma control and FEV₁

The initial omalizumab trials used the St George's Respiratory Questionnaire (SGRQ) (validated in COPD) to assess the impact on quality of life, and there was a statistically and clinically significant improvement compared with placebo. Asthma diaries were generally used to assess patient reported asthma control, rather than a validated tool. The methodology used was variable; however, there were improvements in the treatment group compared with placebo and this was mirrored by a reduction in rescue short-acting β₂-agonist use in those receiving omalizumab. The impact on forced expiratory volume in 1 s (FEV₁) was variable, with only some studies reporting an improvement compared with placebo. An analysis conducted as part of the Cochrane review of omalizumab reported an overall improvement in FEV₁ of 2.15% compared with placebo [11]. The minimal clinically important difference (MCID) for FEV₁ in asthma is debated, but a recent expert consensus report suggested cut-offs of ≥20% in short trials, or ≥15% for trials ≥1 year duration [21].

Predictors of response

A post-hoc analysis of two large randomised controlled trials (RCTs) found those most likely to benefit from therapy were participants who had required emergency asthma treatment in the year before enrolment, those who were on high-dose ICS and those with a lower FEV₁ at baseline [22]; essentially patients with severe asthma. The licence-related criteria for selecting patients who might benefit from omalizumab focus on identifying those with an allergic phenotype; however, neither allergen-specific IgE nor total IgE predicts response to treatment [23] and there are data to suggest omalizumab can reduce the exacerbation rate in non-atopic patients as well [24]. In contrast, the level of the two T2 biomarkers that closely reflect active T2 inflammation, fraction of exhaled nitric oxide (F_ENO) and the blood eosinophil count, do both relate to efficacy of omalizumab. Specifically, Hanania et al. [25] demonstrated that omalizumab was ineffective in patients with a blood eosinophil count <260 cells·μL⁻¹ and a F_ENO <20 ppb. This has also been demonstrated in a single RCT [26]. A study by Djukanovic et al. [27] examined airway inflammation in mild-to-moderate asthma patients treated with omalizumab and observed a reduction in eosinophil counts, which may in part explain the observed relationship between eosinophilia and efficacy. Additionally, elevated T2 biomarkers at the time of omalizumab cessation was shown to be a predictor of future exacerbation [28]. This latter study is particularly helpful in determining whether a trial off omalizumab is likely to be successful or not.

Real-world data and length of treatment

The benefit of omalizumab as an add-on therapy in severe asthma has been confirmed by numerous “real-life” observational trials, summarised in a meta-analysis by Alhossan et al. [29]. The largest real-world study of omalizumab showed modest efficacy: in a Japanese cohort of over 3000 patients with severe allergic asthma, 42% experienced two or more exacerbations despite omalizumab (compared with 63% prior to treatment) and 28% of the cohort continued to experience four or more exacerbations (compared with 48% prior to treatment). The ongoing exacerbation frequency highlights the limitations of this treatment [30].

While extension trials have confirmed the continued efficacy of omalizumab to 1 year, the optimal treatment length for omalizumab is not known. The Xolair Persistency of Response after Long-term Therapy (XPORT) study randomised patients who had been treated with omalizumab for at least 5 years to either continue the drug or to have placebo for a further year. 48% of those in the placebo arm had zero exacerbations in the year, suggesting omalizumab could be safely withdrawn. However, importantly there were no data available on AER prior to starting omalizumab and more patients were exacerbation-free in the treatment arm (67%) [28].

Comorbidities

The treatment of chronic rhinosinusitis with nasal polyposis (CRwNP) with omalizumab has been evaluated in two phase 3 trials and resulted in improved endoscopic nasal obstruction scores, nasal congestion scores (patient reported) and disease specific quality of life scores (SNOT-22) [31]. It has recently been licensed for this indication as an add-on therapy to intranasal steroids. It is also licensed for use in chronic spontaneous urticaria.

Practical considerations

Omalizumab is only licensed in allergic asthma, which means patients need a positive skin prick test or in vitro reactivity to a perennial aeroallergen to qualify for the drug. The EMA licensing also states the requirement for FEV₁ <80% predicted, ongoing symptoms and “multiple” exacerbations despite high-dose ICS/LABA. The UK National Institute for Health and Care Excellence (NICE), which considers the cost effectiveness as well as clinical effectiveness of treatments for the National Health Service, quantifies the number of exacerbations required as ≥4 severe exacerbations in the last year (or the need for OCS to maintain asthma control). There is no requirement to demonstrate an eosinophilic phenotype, despite data demonstrating a lack of efficacy without one.

Omalizumab is given by subcutaneous injection 2–4 weekly, with dose and frequency determined by weight and serum IgE. If the serum IgE is >1500 international units (IU) omalizumab is not licensed for use, and this level decreases as weight increases (e.g. a patient of ≥90 kg would need a serum IgE of ≤600 IU). It is notable that the major clinical trials only included patients with an IgE of 30–700 IU. Pre-filled syringes of omalizumab come as 75 mg and 150 mg, therefore those on the highest dose require four subcutaneous injections every 2 weeks, which may be burdensome for patients when compared to a single injection every 4 or 8 weeks with mepolizumab or benralizumab, for example. The main side-effects from the clinical trial programme were injection site reactions, arthralgia/pain, fatigue, and dizziness. Anaphylaxis is reported in ∼0.1% [32].

Primary outcomes: exacerbation and OCS reduction

The phase 3 MENSA trial demonstrated a reduction in exacerbation rate of 52% with mepolizumab treatment when compared with placebo [36]. Inclusion criteria included an eosinophilic phenotype (with blood eosinophils of ≥150 cells·μL⁻¹ at screening, or ≥300 cells·μL⁻¹ in the past 12 months) and frequent exacerbations (≥2 exacerbations in the year prior to commencing mepolizumab). As well as reducing exacerbations mepolizumab has demonstrated efficacy in reducing OCS dose in those dependent on daily OCS to maintain asthma control. The SIRIUS trial recruited subjects with severe eosinophilic asthma dependent on daily OCS and blood eosinophils of ≥300 cells·μL⁻¹ in the past year, or ≥150 cells·μL⁻¹ in the OCS dose optimisation phase. At 24 weeks, 14% of subjects treated with mepolizumab were able to completely stop prednisolone and overall, a median reduction of 50% was achieved. In the placebo arm, the median reduction was zero; however, a third of subjects were able to reduce their prednisolone dose by at least 50% despite only receiving placebo, highlighting a failure to optimise systemic corticosteroid exposure in these patients prior to randomisation [37]. Importantly, despite the overall reduction in OCS, exacerbations were reduced (AER 1.44 in the mepolizumab arm versus 2.12 in the placebo arm).

Secondary outcomes: quality of life, asthma control and FEV₁

In MENSA, asthma control as measured by the five-question asthma control questionnaire (ACQ5) improved by 0.4, just short of the MCID of 0.5. This trial also used the SGRQ and demonstrated an increase of 7 points versus placebo (meeting the MCID for mild–moderate COPD). There was a small improvement in FEV₁ of ∼100 mL.

Predictors of response

Additional analysis of data from MENSA demonstrated that baseline eosinophil count and AER were the only covariates that influenced efficacy. In those with high blood eosinophils of ≥500 cells·μL⁻¹ there was a 79% reduction in exacerbation rate versus placebo, as well as gains in pre-bronchodilator FEV₁ of 132 mL and ACQ5 of 0.75.

Response to mepolizumab does not appear related to IgE levels, the presence of atopy, or co-eligibility of patients for omalizumab [38, 39]. Real-world characteristics associated with improved response include nasal polyposis, lower baseline OCS dose and lower body mass index (BMI) [40]. In the context of eosinophilic asthma, F_ENO level does not appear to predict response [39, 41].

Real-world data

There is a substantial body of real-world data supporting the efficacy seen in the phase 3 programme [40, 42, 43]. These data are very important as the subjects recruited to the controlled studies appeared to be less severe than initially intended. For example, in the placebo arms of the phase 3 MENSA and phase 3b MUSCA studies of mepolizumab, only 13% and 6%, respectively, of subjects randomised to placebo had an AER >2.

Real-world data also suggest that response can be accurately assessed earlier than 1 year. A retrospective analysis of 99 patients treated with mepolizumab defined response as a reduction in exacerbation rate of at least 50% or, for those on mOCS, a reduction in prednisolone dose of at least 50%. At 16 weeks, 81% of patients had the same responder status as when assessed after 1 year of treatment, rising to 93% of patients by 24 weeks [40].

Length of treatment

In an extension study to early phase work, Haldar et al. [44] followed-up participants for 12 months after they discontinued mepolizumab. Exacerbations increased 3–6 months after discontinuation and after 12 months there was no significant difference between those who had been treated with mepolizumab and those who had placebo [44]. “COMET” randomised patients who had received mepolizumab for at least 3 years to either continue mepolizumab or to placebo for a further year. Those who stopped mepolizumab had a significantly shorter time to first exacerbation and a decrease in asthma control (measured by ACQ5) when compared with those who continued mepolizumab [45]. An open label extension study (“COLUMBIA”) has confirmed a sustained response to mepolizumab up to 4.5 years of treatment [46].

Comorbidities

Practical considerations

Mepolizumab is given as a fixed 100 mg dose every 4 weeks, subcutaneously. It has recently been made available in a pre-filled syringe for self-administration. The commonest side-effects with mepolizumab are headache, backache and injection site reactions; however, overall there are very good safety data available up to 4.5 years [46, 49].

Primary outcomes

Two large phase 3 trials in severe eosinophilic asthma (SIROCCO [59] and CALIMA [60]) recruited patients with two or more exacerbations in the past year stratified by blood eosinophil counts (≥300 cells·μL⁻¹ and <300 cells·μL⁻¹). In those with blood eosinophils <300 cells·μL⁻¹, there was a 17–40% reduction in exacerbation rate compared with placebo, compared to a 28–51% reduction in those with eosinophils ≥300 cells·μL⁻¹. The difference in exacerbation reduction versus placebo reflected a very large placebo response in CALIMA. However, the actual AER on treatment in both studies was ∼0.6. A phase 3b study, ANDHI, demonstrated an overall reduction of 49% in AER compared with placebo at 24 weeks, increasing to 59% in those with baseline blood eosinophils ≥300 cells·μL⁻¹ [61].

The ZONDA study enrolled OCS-dependent patients and demonstrated a 50% reduction in OCS dose compared with placebo. Over 50% of participants in the treatment arms achieved 100% reduction by 28 weeks and the reduction in exacerbations was an impressive 70% [62]. More recently, the open-label PONENTE study of 598 OCS-dependent patients treated with benralizumab reported that over 80% were able to eliminate use or achieved a dosage of 5 mg or less if the reason for stopping the reduction was adrenal insufficiency rather than asthma [63]. To date, PONENTE is the largest OCS sparing study of its kind.

Secondary outcomes

SIROCCO and CALIMA included ACQ and asthma quality of life questionnaire (AQLQ) as patient-reported outcome measures. There were statistically significant improvements, but these narrowly missed the MCID threshold of 0.5 compared with placebo. There were small improvements in FEV₁ of ∼100 mL. The ANDHI trial evaluated quality of life using the SGRQ rather than AQLQ and demonstrated an improvement of −8.1 (MCID of 4), with an improvement evident from week four. It also reported a slightly larger improvement in FEV₁ of 160 mL, with a noticeable separation between placebo and treatment arms by week four [61].

Predictors of response

A post-hoc analysis of pooled data from SIROCCO/CALIMA identified high baseline exacerbation rate and higher blood eosinophils as associated with an enhanced response to benralizumab [64]. The presence of nasal polyposis (itself associated with eosinophilic inflammation and severe disease), low baseline forced vital capacity (FVC) (<65% predicted) and dependence on OCS were also associated with a better response [65]. ANDHI also reported greater reductions in AER in those with blood eosinophils of ≥300 cells·μL⁻¹ at baseline, adult-onset disease and nasal polyposis [61]. Real-world data have similarly shown an association between enhanced response and high baseline eosinophils and the presence of nasal polyposis [66]. A differential response has not been shown to be related to baseline F_ENO level [41], IgE [67] or co-eligibility for omalizumab treatment [39].

Real-world data and length of treatment

Large real-world cohorts now confirm the clinical effectiveness of benralizumab in reducing AER and mOCS requirement as well as reducing symptom scores and improving quality of life metrics [66]. The BORA extension trial has demonstrated continued efficacy and safety of benralizumab out to 2 years [68]. In this study, more than 70% of patients in their second year of benralizumab treatment remained exacerbation free. A recent report of adolescents who had completed 3 years of benralizumab as part of the extension trial confirmed continued efficacy and safety [69], whilst the MELTEMI study has provided evidence of continued efficacy along with reassuring safety data out to 5 years [70].

Comorbidities

Benralizumab is not currently licensed as a treatment for CRwNP, although there is evidence to suggest efficacy for this indication. A post-hoc analysis of the ANDHI trial found significant improvement in SNOT-22 scores in those treated with benralizumab versus placebo [71], and this has also been reported in a real-world analysis [72]. A phase 3 trial examining the efficacy of benralizumab in CRwNP (OSTRO) has recently been published and reported a reduction in nasal blockage score (patient reported) and nasal polyp score (determined by nasoendoscopy). There was also an improvement in disease-specific quality of life (SNOT-22) versus placebo, but this did not reach statistical significance [73]. OSTRO did not require evidence of eosinophilic disease for enrolment, and a further trial of benralizumab in eosinophilic CRwNP (ORCHID, ClinicalTrials.gov identifier: NCT04157335) is currently recruiting.

The use of benralizumab to treat EGPA is currently being evaluated in a large RCT comparing the efficacy of benralizumab to mepolizumab (ClinicalTrials.gov identifier: NCT04157348). At present there is a promising pilot study (open label, prospective) [74], as well as real-world data [75] suggesting a significant steroid-sparing effect in EGPA.

Practical considerations

Benralizumab is a fixed dose subcutaneous injection, given 4-weekly for the first three doses, and then 8-weekly thereafter. This is the least frequent dosing of any of the current biologic options for severe asthma. Benralizumab is available as a pre-filled auto-inject syringe for self-administration. The main side-effects reported in the phase 3 clinical trials were injection site reactions (2–3%) and nasopharyngitis, although these occurred at similar rates in the treatment and placebo groups.

Concerns about the health implications of the effective “knock-out” of a highly evolutionarily conserved cell like the eosinophil have been raised. This most commonly relates to a possible homeostatic role for the eosinophil in host-defence against both helminth and other infections, as well as in tumour surveillance [76]. However, despite over 8000 patients being exposed to benralizumab across clinical development programmes in asthma, COPD and other conditions, there have, so far, been no signals of harm including no increased infection or malignancy rates [77]. MELTEMI confirmed ongoing safety at 5 years following initiation of benralizumab with rates of malignancy and serious infections at less than 2% [70].

Exacerbation and OCS reduction

The LIBERTY ASTHMA QUEST trial enrolled 1902 asthma patients irrespective of their blood eosinophils or F_ENO with one or more exacerbations in the past year. As the subjects could be on either medium or high-dose ICS, it was a less severe cohort than the anti-IL-5/IL-5R-α biologic trials [80]. Compared with placebo, there was an overall reduction in AER of 48%. A statistically significant benefit was only seen in those with a blood eosinophil count of ≥300 cells·μL⁻¹, in whom the AER fell by almost two-thirds. Clinical efficacy was also seen in those with a F_ENO of ≥25 ppb. Although there is reported efficacy in subjects with a blood eosinophil count >150 cells·μL⁻¹ this effect size was driven by the patients with an eosinophil count of ≥300 cells·μL⁻¹ and there was no statistically significant benefit in subjects with a blood eosinophil count <300 cells·μL⁻¹.

Dupilumab has also demonstrated efficacy in reducing mOCS use in asthma. The LIBERTY ASTHMA VENTURE trial [81] showed an overall median prednisolone dose reduction of 50% versus placebo and 69% reduced their dose below 5 mg·day⁻¹, compared with 33% of the placebo group. This trial suffered from a very marked placebo response, but did nonetheless show excellent corticosteroid-sparing efficacy. There was a significant reduction in corticosteroid dose in both low (<300 cells·μL⁻¹) and high (≥300 cells·μL⁻¹) blood eosinophil groups, although the greatest benefit was again seen in those subjects with higher baseline blood eosinophil counts.

FEV₁ and patient-reported outcome measures

In LIBERTY QUEST [80], dupilumab treatment resulted in an overall improvement of 130–140 mL in pre-bronchodilator FEV₁ versus placebo at 12 weeks; however, much of this result was driven by the T2 high groups. For example, in those with baseline F_ENO ≥50 ppb there was a 300–390 mL improvement versus placebo, compared to 120–190 mL in the F_ENO ≥25 ppb to <50 ppb group, and only 30–50 mL in the F_ENO <25 ppb group. Although there was an initial rise in FEV₁ in the placebo group (presumably due to improved ICS/LABA adherence), there was then a decline over time of ∼40 mL per year, which did not occur in the treatment group. AQLQ and ACQ both improved, more so in the high blood eosinophil subgroup, but still less than the MCID when compared with placebo.

Predictors of response

As outlined above, the biggest reductions in exacerbation rate and improvements in FEV₁ were seen in those with blood eosinophils ≥300 cells·μL⁻¹ or F_ENO ≥25 ppb. A post-hoc analysis of QUEST stratified patients according to baseline exacerbation rate and the biggest improvements in exacerbation rate, FEV₁ and ACQ were in those with more exacerbations at baseline [82]. Improvements were seen irrespective of whether participants were on medium-dose or high-dose ICS at baseline [83].

Real-world and long-term data

To date, there are no substantial real-world data on the use of dupilumab in asthma. The publication of an extension trial to 96 weeks (“LIBERTY ASTHMA TRAVERSE”, ClinicalTrials.gov identifier: NCT02134028) is awaited.

Comorbidities

Practicalities

Dupilumab has a somewhat broader indication than the anti-IL-5 biologics, as its European licence only requires evidence of T2-high inflammation (raised blood eosinophils and/or raised F_ENO) [85]. In the USA, it is also licensed for use in OCS-dependent asthma regardless of baseline blood eosinophils. Dupilumab is given 2-weekly by subcutaneous injection at a fixed dose and is available as a pre-filled auto-inject syringe.

The most common side-effects in clinical trials were injection site reactions (16.8%). This is the highest rate amongst biologic therapies. Akin to the theoretical concern about eosinophil depletion with benralizumab has been a concern regarding hypereosinophilia with dupilumab. A rise in peripheral blood eosinophils ≥3000 cells·μL⁻¹ was observed in 4% of those treated in LIBERTY ASTHMA QUEST. This was associated with clinical symptoms in four patients (out of 1264 in the combined treatment arms). The average blood eosinophil count returned to baseline by 24 weeks of therapy. Although this is a relatively new therapy in severe asthma, there are longer-term safety data available following its use in the treatment of atopic dermatitis. In patients with atopic dermatitis treated with dupilumab, a high incidence of conjunctivitis has been reported [86].

Primary outcomes

The phase 3 trial for tezepelumab, “NAVIGATOR”, has recently been published [90]. Inclusion criteria included two or more severe exacerbations in the past 12 months and at least moderate dose ICS plus an additional controller medication. There was no requirement for raised blood eosinophils; however, the trial population was monitored to recruit approximately 50% with blood eosinophils of ≥300 cells·μL⁻¹ and 40% with ≥3 exacerbations in the previous 12 months. Pre-planned subgroup analyses were carried out according to blood eosinophils and F_ENO. Overall, treatment with tezepelumab led to a reduction in exacerbation rate of 56% versus placebo (rate ratio (RR) 0.44). Unsurprisingly, the greatest reductions were seen in those with the strongest T2 signal (higher blood eosinophils and higher F_ENO). Those with blood eosinophils of <150 cells·μL⁻¹ still had a substantial benefit (RR of 0.47) if they had a F_ENO ≥25 ppb. The only biomarker subgroup with no efficacy was the dual biomarker low group (blood eosinophils <150 cells·μL⁻¹ and low F_ENO <25 ppb). A recent post-hoc analysis of the phase 2 trial PATHWAY confirmed response was unrelated to atopic status [91].

The SOURCE trial enrolled patients with OCS-dependent asthma and randomised them to tezepelumab or placebo [92]. The results have yet to be formally published; however, the investigators have announced the trial did not meet the primary endpoint of reducing daily OCS whilst maintaining asthma control [93].

Secondary outcomes

There were statistically significant improvements in FEV₁, ACQ and AQLQ with tezepelumab treatment, which were seen as early as week 2; however, these results were mainly driven by those with blood eosinophils of ≥300 cells·μL⁻¹. In this subgroup, there were significant improvements compared with placebo, meeting the MCIDs for FEV₁ (+230 mL), ACQ6 (−0.5) and AQLQ (+0.5). There are no extension, long-term, or real-world data yet available for tezepelumab.

Practicalities

Tezepelumab is not yet licensed for use; however, in the clinical trials it was given at a set dose of 210 mg, 2-weekly by subcutaneous injection. There have been no significant safety issues to date, with only a small increase in injection site reactions compared with placebo; however, there are no published data beyond 52 weeks. An extension trial, “DESTINATION”, is in progress (ClinicalTrials.gov identifier: NCT03706079).

Pregnancy and breastfeeding

mAbs, such as asthma biologics, are known to cross the placenta, increasing proportionally as the pregnancy progresses, so it is hypothesised that the greatest risk would be in the second and third trimester [94]. At present there are only limited data to help physicians and women to make an informed choice around whether to start or continue biologic therapy if they are planning to, or become, pregnant. The only substantive human data published concern omalizumab. Data from a registry of 191 pregnant women taking omalizumab (“EXPECT”) were published in 2019, and more recently these data have been compared with a diseased matched cohort, which did not suggest any increased risk of spontaneous abortion or fetal death [95]. Although registries are in place for the other asthma biologics, no data from these have yet been published.

It should be noted that pregnant women with severe asthma have an increased risk of adverse perinatal outcomes, including an increased risk of prematurity and decreased fetal growth [96], and OCS are also known to increase the risk of pre-eclampsia. The known risk of poorly controlled asthma must therefore be weighed against the relatively unknown risk of biologic therapy.

Adrenal insufficiency

Effective mOCS reduction and withdrawal reveals underlying adrenal insufficiency in patients treated with anti-IL-5 mAbs in around 75% of cases [97]. It is, therefore, important to monitor early-morning cortisol as OCS is weaned towards physiological levels. Patients who are suspected to have adrenal axis suppression that does not normalise on slowly weaning the OCS dose should be referred for a more formal review and short synacthen test by endocrinology services.

Which biologic first?

The availability of biologic therapies, whether limited by health insurance, prescribing guidelines or other factors, will be the first factor dictating choice. With that consideration, for the patient on mOCS we would limit the choice to those with proven steroid-sparing efficacy in controlled trials: mepolizumab, benralizumab or dupilumab. If there remains more than one option, we then discuss with the patient their priorities for treatment and consider existing characteristics that predict response. Do they have a comorbidity, such as CRwNP or atopic dermatitis, that is impairing their quality of life? Would frequent dosing be difficult for them? Similar considerations are warranted for patients not on mOCS, but in this context omalizumab is also a consideration, particularly in younger patients with child-onset disease and those with clinically significant and treatment-resistant allergic comorbidities.

A summary of information to inform this two-step process is given in tables 1 and 2. Good practice is to review patients throughout the first 6 months of treatment and to make a more formal assessment of response at 6 months, to ensure a switch to an alternate biologic is not inappropriately delayed.

What if first-line therapy fails?

Some patients continue to exacerbate or are unable to wean mOCS despite treatment with a biologic therapy. The MEX trial investigated exacerbations whilst on mepolizumab treatment and found approximately half of the exacerbations were associated with eosinophilic sputum [98]. The phase 2b dosing study of mepolizumab, DREAM, demonstrated that there was a dose related reduction in sputum eosinophils, with no statistically significant difference in sputum eosinophil count in the lowest dose mepolizumab arm compared with placebo [99]. This was not associated with worse clinical outcomes; however, it is recognised that the sputum eosinophil count is associated with exacerbation risk, and this raises the question as to whether some patients might benefit from more potent suppression of airway eosinophilia than the licensed 100 mg dose of mepolizumab. In this case, a switch to either reslizumab or benralizumab, which offers greater suppression of eosinophilia, may give a superior response [100, 101].

Decreased adherence to ICS can also play a role in a suboptimal response to biologic therapy. In a real-world analysis of 91 mOCS-dependent severe asthma patients treated with mepolizumab those who had poor adherence to ICS experienced significantly more exacerbations and less reduction in prednisolone dose than those with good adherence [102], although this was not the case with benralizumab [103], potentially due to its more complete eosinophil depletion.

The other half of exacerbations in the MEX trial (mepolizumab) were non-eosinophilic and likely to be infection related, i.e. they were likely to have neutrophilic sputum, a raised C-reactive protein and be treated with antibiotics. In these instances, appropriate antibiotic therapy followed by consideration of a maintenance antibiotic (if the infections are frequent), such as azithromycin, is likely to be of more benefit than a switch in biologic.

In the phase 3 studies of benralizumab, ∼10% of subjects developed serum neutralising anti-drug antibodies during treatment with benralizumab [59, 60]. The authors reported that they could not detect any change in efficacy in this subgroup. There is no commercially available assay for detecting benralizumab; however, a detectable blood eosinophil count can be used as a surrogate marker. In a real-world analysis of 130 patients treated with benralizumab, 18 patients had a suboptimal response and of these five had detectable eosinophils, suggesting that neutralising anti-drug antibodies could be a mechanism of drug failure in a minority of patients [66].