Abstract
Introduction The A allele of rs1042713 (Arg16 amino acid) in the β2-adrenoreceptor is associated with poor response to long-acting β2-agonist (LABA) in young people with asthma. Our aim was to assess whether the prescribing of second-line controller with LABA or a leukotriene receptor antagonist according to Arg16Gly genotype would result in improvements in Pediatric Asthma-Related Quality of Life Questionnaire (PAQLQ).
Methods We performed a pragmatic randomised controlled trial (RCT) via a primary care clinical research network covering England and Scotland. We enrolled participants aged 12–18 years with asthma taking inhaled corticosteroids. 241 participants (mean±sd age 14.7±1.91 years) were randomised (1:1) to receive personalised care (genotype directed prescribing) or standard guideline care. Following a 4-week run-in participants were followed for 12 months. The primary outcome measure was change in PAQLQ. Asthma control, asthma exacerbation frequency and healthcare utilisation were secondary outcomes.
Results Genotype-directed prescribing resulted in an improvement in PAQLQ compared to standard care (0.16, 95% CI 0.00–0.31; p=0.049), although this improvement was below the pre-determined clinical threshold of 0.25. The AA genotype was associated with a larger improvement in PAQLQ with personalised versus standard care (0.42, 95% CI 0.02–0.81; p=0.041).
Conclusion This is the first RCT demonstrating that genotype-driven asthma prescribing is associated with a significant improvement in a clinical outcome compared to standard care. Adolescents with the AA homozygous genotype benefited most. The potential role of such β2-adrenoceptor genotype directed therapy in younger and more severe childhood asthma warrants further exploration.
Abstract
Personalised prescribing in adolescents with asthma demonstrated that β2-adrenoreceptor genotype directed treatment results in a small but significant improvement in PAQLQ. β2-adrenoreceptor genotype guided treatment requires further investigation. https://bit.ly/3oDvP1N
Introduction
Asthma is the most common chronic condition affecting children [1]. It is associated with substantial health and quality-of-life burden for the patient as well as significant healthcare expenditure globally [2]. Childhood asthma is treated in a stepwise approach using controller medications, initially with inhaled corticosteroids (ICS) and if symptoms are not subsequently well controlled then addition of either inhaled long-acting β2-agonist (LABA) or a leukotriene receptor antagonist (LTRA), or by a further increase in ICS dose [3].
There is a wide degree of heterogeneity in response to treatment among young people with asthma, with estimates that 60–80% of the observed variance between individuals could be due to genetic differences [4]. There has been particular interest in a variation in the gene encoding for the β2-adrenergic receptor (ADRB2) at position 16 (rs1042713), resulting in an allelic substitution from glycine to arginine (Gly16Arg). The homozygous AA variant is found in ∼15% of people and has been associated with poor response to ICS-LABA controller therapy in young people [5–7]. These findings have not been widely replicated in adult studies aside from the demonstration of greater bronchoprotective subsensitivity with the A allele in response to LABA therapy [8–11]. It is hypothesised that, for young people with the A genotype, regular ICS-LABA use results in agonist-induced downregulation and associated uncoupling of the β2 receptor, thus impairing the efficacy of the medication [12, 13].
A meta-analysis from the Pharmacogenomics in Childhood Asthma consortium comprising 4226 children showed a 34% elevated risk of asthma exacerbation for each copy of the A allele in young people with ICS-LABA controller treatment, with at least one copy of A allele being present in 62.8% of people [7]. One prospective study showed that use of a LTRA instead of LABA in young people homozygous for AA reduced school absence and improved asthma symptom and quality-of-life scores [14]. Thus, it is important to test whether Arg16Gly genotype directed therapy (personalised medicine) in adolescents with asthma leads to improvement in quality of life.
The principal aim was to test our hypothesis that prescribing of second-line asthma controller medication (LABA or LTRA) according to Arg16Gly genotype compared to standard care provided according to the British Thoracic Society (BTS) guidelines would result in an improvement in quality of life determined by standardised Pediatric Asthma Quality of Life Questionnaire (PAQLQ) in 12–18-year-olds with asthma. Secondary aims included assessing the effect of genotype-directed prescribing on 1) asthma control (validated Asthma Control Questionnaire (ACQ)-6), 2) exacerbation frequency (requirement for oral steroids) and 3) healthcare utilisation (nonroutine primary-care review, emergency department attendance or hospital admission).
Some results have been reported previously in the form of an abstract [15].
Methods
Subjects
Our trial consisted of participants of either sex aged 12–18 years with 1) a documented physician diagnosis of asthma who were 2) taking ICS with or without the additional second-line controllers (LABA or LTRA). The target population was adolescents whose asthma was managed in primary care in England and Scotland. Exclusions were 1) known contraindication to LABA or LTRA; 2) on step 4 of BTS guidelines (e.g. use of theophylline-based controller medication such as Uniphyllin); 3) presence of other major airway or lung disease (other than asthma); 4) pregnancy or lactation; 5) participation in another clinical trial; and 6) inability to provide saliva/buccal cells for genotyping.
Study design
For this two-arm pragmatic randomised controlled trial (RCT), participants were recruited from throughout England and Scotland. The study duration was 13 months, consisting of a 4-week run-in period and 12 months of follow-up.
Participants were principally recruited through primary care via the Clinical Research Network across England and Scotland as well as the patient databases BREATHE and PAGES. Informed consent and assent were obtained online or by telephone with follow-up written consent. The study followed the Children's Research Network standard operating procedures, Health Research Agency guidelines and the Nuffield Council on Bioethics report [16] in obtaining informed consent. Participants aged 12–18 years consented independently, while parental consent was sought in addition for those aged ≤15 years. The trial was sponsored by the University of Sussex (approval December 2014) and ethical approval was obtained from the East of Scotland Research Ethics Committee (15/ES/0007; approval March 2015). The trial was registered on the UK Clinical Research Network website with details made available to the public before the recruitment of the first participant. This trial is registered with ClinicalTrials.gov NCT02758873.
Participants were randomised 1:1 to personalised care (rs1042713 single nucleotide polymorphism (SNP)-based prescribing) or standard care by a web-based system, TRuST (Tayside Randomisation SysTem). Participants were allocated as per block randomisation with no stratification or minimisation. The personalised care group were prescribed asthma controller medication on the basis of their Arg16Gly genotype, AA or AG receiving montelukast (LTRA) and GG receiving salmeterol (LABA). The standard care group were prescribed controller medication based on the current BTS guidelines. Neither group nor the study team was blinded to group allocation or prescribed medication.
Participants undertook a 4-week run-in period where they were asked to use only ICS as their controller medication at the previously prescribed dose. Reliever medication was used by each participant as required. Outcomes were measured at baseline, following completion of run-in and 3, 6, 9 and 12 months. Study questionnaires were completed online or over the phone. Medications were prescribed by the participant's general practitioner (GP).
Change to the ACQ score was used to determine the participant's controller treatment with a score ≥1.0 [17] or need for oral corticosteroid triggering escalation. The personalised care group were prescribed asthma controller medication on the basis of their Arg16Gly genotype, AA or AG genotypes receiving LTRA and GG receiving LABA. The standard care group were prescribed controller medication based on the BTS guidelines. A stable or decreased ACQ score resulted in continuation of current treatment.
DNA collection kits were posted to participants with instructions and a paid return envelope. Saliva samples were collected using a commercially available pot (GeneFiXDNA saliva collector; Isohelix.com, Harrietsham, UK). DNA was prepared with the Isohelix GeneFiX saliva prep DNA kit. DNA extraction and Arg16Gly genotype status was determined at the University of Dundee (Division of Population and Health Genomics, Dundee, UK) using TaqMan-based allelic discrimination assays on an ABI 7900 Sequence Detection System (Applied Biosystems, Foster City, CA, USA) as described previously [18].
Outcomes
The primary outcome was the change in PAQLQ [19] from completion of the run-in to completion of the study at 12 months. Secondary outcomes were change in ACQ score, healthcare utilisation for asthma management (nonroutine primary care review, emergency department attendance or hospital admission) and as exacerbation frequency (courses of oral corticosteroid). Adverse events were recorded as per Health Research Authority guidelines.
Analysis
A change of 0.5 units on the PAQLQ is considered to represent the minimal clinically important difference (MCID) [20]. We expected to see a 0.25-unit improvement in PAQLQ at 12 months in the standard care group with the improvements estimated on the basis of genotype frequency with a projected improvement of 0.5 in the GG (40%); 0.25 in the AG (35%) and 0 in the AA (15%) genotype. The calculated sample size, in order to detect a clinically relevant threshold of 0.25 units (0.5–0.25=0.25) in the primary outcome of PAQLQ score (sd 1.0, α=0.05; 90% power) was 100 participants in each group. To allow for a 15% attrition rate, the recruitment target was increased to 120 participants in each group.
Analyses comparing personalised and standard care were completed as pre-specified in the statistical analysis plan. All analyses were performed on intention-to-treat population, i.e. by group randomised. Data for continuous outcome measures were assessed for normality prior to analysis. Transformations of the outcome variables were used where necessary if they were not normally distributed. If data were normally distributed, outcome measures were assessed using mixed model repeated measure analysis, adjusted for the corresponding baseline values and group allocation as fixed effects. Models used all available data from the end of the run-in period. Where data was not normally distributed and could not be transformed into a normal distribution, data was analysed using nonparametric methods. Subgroup analyses were performed on participants with the Arg16Gly status AA. A two-sided p-value of <0.05 was taken to be significant for all analyses. SAS software (version 9.4; SAS Institute, Cary, NC, USA) was used for all statistical analyses. The Personalised Medicine for Asthma Control (PACT) study conforms to Consolidated Standards of Reporting Trials 2010 guidelines on RCT reporting.
Results
Between 9 February 2016 and 25 April 2018, 247 participants consented and 241 were randomised before entering the 4-week run-in period (figure 1). Participants were randomised to either personalised care (n=121) or standard care (n=120) for their asthma controller therapy. Baseline demographics and clinical characteristics were broadly similar between those receiving personalised and standard care with a mean age of 14.7 years across the two groups (table 1). However, there are important differences between the groups with a greater proportion of adolescents receiving ICS, LABA and LTRA combination therapy at baseline in the standard care group (15.0% versus 6.6%) indicating a possible increased asthma severity in this group, and a lower prevalence of the AA genotype in the personalised care group (9.9% versus 15.0%), possibly impairing the overall clinical benefit of personalised medication prescribing in this group.
Completion of the run-in period (ICS only) resulted in 0.1 mean improvement in PAQLQ compared to baseline in both groups. A mean improvement in PAQLQ total score compared to end of run-in was observed in both the personalised and standard care groups (table 2).
Prescription of asthma controller medication as per Arg16Gly genotype SNP status (personalised care) resulted in an improvement in mean PAQLQ compared to standard care (0.16, 95% CI 0.00–0.31; p=0.049) (table 2, figures 2a and 3); however, the difference was below the pre-determined clinical threshold of 0.25. The PAQLQ domains of emotional function and activity limitation score had the greatest mean difference in change with personalised care.
Subgroup analysis of the children with the homozygous AA genotype who completed follow-up (n=27) demonstrated an improvement in mean PAQLQ that exceeded the clinical threshold of 0.25 in those receiving personalised care compared to standard care (0.42, 95% CI 0.02–0.81; p=0.04) (table 2; figure 2b). There were no adverse or serious adverse events reported throughout the duration of the trial.
A total of 28 (11.6%) participants experienced an asthma exacerbation (requirement for oral steroids) during 12-month follow-up, with numerically lower rates reported in the personalised care (8.3%) compared to the standard care group (15.0%) (p=0.10). In addition, there was a trend toward increased time to exacerbation in the personalised care group (225.7 days) compared to the standard care group (141.5 days) (p=0.10). There was a similar improvement in ACQ score from the end of run-in when mean change was compared in the personalised care (0.42) and standard care groups (0.47) (p=0.18). There was no association between personalised asthma controller directed prescribing and healthcare utilisation or the number of asthma medications prescribed with a small increase in the mean number prescribed in the personalised care (0.3) and standard care groups (0.2) (p=0.36) (table 3).
Discussion
The PACT study is the first prospective RCT assessing the efficacy of genotype-directed prescribing in adolescent asthma. Prescription of second-line controller medication with LABA or LTRA according to Arg16Gly genotype resulted in a statistically significant improvement in primary clinical outcome of PAQLQ, although the magnitude of improvement amounted to 0.16 and was below the a priori clinical threshold of 0.25. The quality-of-life benefit seen in adolescents with the homozygous AA genotype was 0.42, exceeding the clinically significant PAQLQ threshold.
Our findings are consistent with the only previous trial conducted examining the effects of LABA in relation to Arg16Gly genotype in children. That proof-of-concept study focused on previously genotyped asthmatic children who were all homozygous for the AA variant. A total of 62 children were randomised to receive either ICS and LABA or ICS and LTRA and were followed-up for 12 months. Children treated with a LTRA had a clinically relevant mean improvement in PAQLQ scores amounting to 0.53 as well as reduced school absences and use of rescue medication in comparison to those treated with LABA [14]. In keeping with these findings, adolescents with the AA homozygous genotype benefitted most from genotype-directed prescribing in our current study.
Both RCTs are underpinned by a large meta-analysis of observational studies comprising five childhood asthma cohorts demonstrating an increased risk of exacerbations with each copy of the A allele amounting to a 34% (95% CI 15–50%) difference for asthmatic children treated with ICS-LABA therapy [7]. In this study, we report a nonsignificant numerical trend toward decreased exacerbation frequency in the genotype-directed treatment group (8.3% versus 15.3%, p=0.10). The low exacerbation frequency (11.6%) within our well-controlled population recruited for this study is a likely confounding factor as to why this well-described association did not reach significance. A 2018 systematic review further assessed the role of Arg16Gly genotype variation on LABA response in asthma, confirming that the contribution of genetics to LABA response is more consistently shown in children compared to adults [21]. It is hypothesised that this could relate to the altered phenotype of children's asthma, with a greater emphasis of atopy in its pathogenesis as well as reduced duration of chronic airway inflammation and airway wall rigidity [22]. It is thus possible that a greater mean improvement in PAQLQ could be achieved with this genotype-directed intervention in younger children, especially in those with more severe disease.
A conspicuous observation from the study was that both the personalised and standard care groups had improved PAQLQ and ACQ at final follow-up, this notable “trial effect” is well described [23]. A component of this known as the “care effect” could result from participants in both arms of the PACT trial having more regular face-to-face contact with primary care healthcare practitioners. The frequent telephone or e-mail communication with the study team as per the study protocol may have altered health behaviour (the Hawthorne effect) [24]. It is also conceivable that by virtue of regularly completing the PAQLQ and ACQ questionnaires, adolescents may have developed improved asthma symptom awareness, promoting self-education about asthma triggers and the benefits of controller treatment and improved adherence. This, potential “trial effect”, may explain how the mean improvement in PAQLQ exceeded the MCID in both groups, which may in turn have contributed to a smaller than anticipated response to genotype-directed therapy.
Our study has some limitations. The lower than expected effect of personalised medicine on PAQLQ score could partly be explained by the excellent symptom control at baseline with very good PAQLQ and ACQ scores demonstrated, leaving limited room for further improvement. This explanation is supported by the large number of adolescents who did not experience an exacerbation (88.4%) and did not require any nonroutine healthcare for their asthma (71.4%) during 12-month follow-up. The good baseline control on ICS treatment following the run-in period resulted in 65% of adolescents in the personalised medicine group not actually experiencing specific genotype-based prescribing, e.g. add-on of LABA or LTRA, thus diminishing the possibility of proving the benefit of personalised care. Future studies aiming to identify the potential role of Arg16Gly genotype directed therapy may require selective recruitment of children with poorly controlled asthma. In addition, differences between the groups could explain the lower than anticipated response to personalised treatment. The greater proportion of adolescents receiving ICS, LABA and LTRA combination therapy at enrolment in the standard care group (15.0% versus 6.6%) could indicate potentially increased disease severity at baseline. A further contributing factor could be the lower than anticipated number of adolescents with the significant AA genotype in the personalised medicine group (9.9%) compared to the standard care group (15%) and in a previous large population meta-analysis (16.1%) [7].
Ideally, the criteria for asthma diagnosis should be defined in accordance with the Global Initiative for Asthma guidelines and documented for each patient involved in the study. By recruiting patients with a physician diagnosis of asthma from primary care, where there may not be routine access to spirometry or the capability to perform bronchodilator reversibility, the accuracy of asthma diagnosis is impaired, which is a notable limitation. However, as the vast majority of adolescents with asthma are managed in this setting this approach assesses the pragmatic real-world efficacy of personalised asthma prescribing. The lack of ethnic diversity within the study participants (88.8% Caucasian) is a potential limitation. However, there is reason to believe our results could be generalisable to other populations given the finding of increased rate of exacerbation with LABA use with presence of the A allele among cohorts of young people of differing ethnicities [7].
An important element of the PACT study is in relation to the novel study methodology employed, utilising telephone and online contact for consent and data collection along with study directed prescribing through the patient's GP without direct face-to-face contact with study team members at any point. This is unique in the investigation of young people's asthma. Replication of this methodology may help investigators to conduct trials in a safe manner during the coronavirus disease 2019 pandemic. The significant associated cost reduction has important implications for future research studies, with the cost per patient of USD 1495 for the PACT study substantially lower than a mean cost of USD 4630 per patient for RCTs funded by the National Institute for Health Research Health Technology Assessment programme between 2000 and 2005 [25].
Further areas need to be explored more fully to understand the potential value of Arg16Gly genotype directed prescribing in the treatment of children's asthma especially those with more severe disease. On the basis of our findings, it is important to re-evaluate our study question in a less well-controlled sample, including younger children, with suggestions that the benefits of Arg16Gly genotype based prescribing are more pronounced in this group. It is hoped that we will have a greater understanding following publication of the PUFFIN trial, an ongoing multicentre Dutch study exploring Arg16Gly genotype based prescribing in 6–17-year-olds with asthma utilising the Asthma Control Test as the primary outcome measure [26].
Conclusion
In this 12-month trial, asthma controller prescribing on the basis of Arg16Gly β2-receptor genotype resulted in a small, but significant, improvement in PAQLQ in adolescents compared to standard care; however, this was below the expected clinical threshold. A clear benefit was demonstrated for those with the AA homozygous genotype (15% of the population). We recommend further prospective randomised studies to help identify the potential clinical utility of personalised prescribing in young people's asthma.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material ERJ-04107-2020.Supplement
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-04107-2020.Shareable
Footnotes
Author contributions: Principal investigator: S. Mukhopadhyay. Study concept and design: C. Palmer, S. Turner, F. Hogarth, H. Smith, B. Lipworth and S. Mukhopadhyay. Patient recruitment, acquisition of data and database management: K. Pilvinyte and F. Hogarth. Genotyping: R. Tavendale. Statistical analysis: P. Rauchhaus. Drafting of manuscript and critical revision: T. Ruffles, C. Palmer, S. Turner, B. Lipworth and S. Mukhopadhyay. Supervision: B. Lipworth and S. Mukhopadhyay. All authors contributed to interpretation of the data, participated in the writing of the manuscript and have approved the final version for submission.
This article has supplementary material available from erj.ersjournals.com
The trial was registered on the UK Clinical Research Network (UKCRN) website with details made available to the public before the recruitment of the first participant. This trial is registered with ClinicalTrials.gov NCT02758873. Individual participant data that underlie the results reported in this article, after deidentification (text, tables, figures and appendices) will be made available, along with study protocol, statistical analysis plan and the analytical code, beginning 3 months and ending 5 years following article publication, to researchers who provide a methodologically sound proposal. Proposals should be directed to the corresponding author. To gain access, data requestors will need to sign a data access agreement.
Conflict of interest: All authors report study funding from The Henry Smith Charity and Action Medical Research. J. Grigg reports personal fees from AstraZeneca, GSK, Medimmune and BV Pharma, during the conduct of the study. F. Hogarth reports grants from University of Sussex, during the conduct of the study. B. Lipworth reports other (equipment) from GSK, grants, personal fees for advisory board work, consultancy and lectures, and non-financial support for meeting attendance from AstraZeneca, Chiesi and Teva, personal fees for advisory board work from Novartis, grants, personal fees for consultancy and lectures, and non-financial support for meeting attendance from Boehringer Ingelheim, personal fees for consultancy from Dr Reddys, Sandoz, Cipla and Glenmark, during the conduct of the study; personal fees for consultancy from Lupin and Vectura, grants and personal fees for consultancy from Sanofi Regeneron, outside the submitted work; and the author's son is an employee of AstraZeneca.
Support statement: The study was funded by The Henry Smith Charity and Action Medical Research (grant number GN2203). The funders had no role in the design and conduct of the study; collection, management, analysis an interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received November 6, 2020.
- Accepted January 1, 2021.
- Copyright ©The authors 2021. For reproduction rights and permissions contact permissions{at}ersnet.org