Parent-determined oral montelukast therapy for preschool wheeze with stratification for arachidonate-5-lipoxygenase (ALOX5) promoter genotype.

Primary objectives

1. To assess the efficacy of parent-initiated intermittent montelukast for the reduction of unscheduled medical attendances for preschool wheeze.

2. To explore the role of the ALOX5 promoter genotype in montelukast efficacy.

Secondary objectives

1. To assess the impact on intermittent montelukast on respiratory morbidity.

2. To assess the impact of montelukast on health service usage.

3. To assess the impact of intermittent montelukast on concomitant medication use.

4. To assess the impact of intermittent montelukast on adverse events (AEs).

5. To gather exploratory data on related LT pathway genes.

6. To gather exploratory data on urinary LT/eicosanoid output.

7. To assess the impact of intermittent montelukast on economic outcomes.

8. To assess qualitative outcomes related to parent-initiated intermittent therapy for preschool wheeze and participation in a genetically stratified interventional trial.

Eligibility criteria

Patients were eligible for the study if they fulfilled the following criteria:

• aged ≥ 10 months and ≤ 5 years on the day of the first dose of the investigational medicinal product (IMP)

• two or more attacks of parent-reported wheeze

• at least one attack with wheeze validated by a clinician (nursing or medical)

• the most recent attack within the last 3 months

• contactable by telephone and able to attend one face-to-face review

• parent or guardian able to give written informed consent for their child to participate in the study.

Exclusion criteria

The following characteristics rendered patients ineligible for the study:

• any other chronic respiratory condition diagnosed by a clinician, including structural airway abnormality (e.g. floppy larynx) and cystic fibrosis

• any chronic condition that increases vulnerability to respiratory tract infection, such as severe developmental delay with feeding difficulty or sickle cell disease

• history of neonatal chronic lung disease

• current continuous oral montelukast therapy

• in a trial using an IMP in the previous 3 months prior to recruitment.

Selection of study population

As indicated previously, wheezing is common in otherwise healthy preschool children; however, safe effective treatment options are limited. We therefore sought to conduct a pragmatic trial with the widest possible useful application. Thus, participants were not limited in terms of wheeze severity or concomitant medications, notwithstanding the prohibition of regular montelukast. We did not include children aged less than 10 months and greater than 5 years in order to exclude children with classical bronchiolitic or asthmatic phenotypes, for which treatment strategies differ.

Recruitment setting

Participants were identified in primary and secondary care centres. Recruitment was planned to encompass only three secondary care centres (the Royal London Hospital, University Hospital Leicester and the Royal Aberdeen Children’s Hospital), but increased to 41 secondary care centres in England and Scotland (see Acknowledgements) in response to observed recruitment rates.

Invitation of potential study participants to attend screening visit

A member of the child’s usual general practitioner (GP) care team or the hospital paediatric team (as appropriate) identified potentially eligible children based on age and history of wheeze from reviewing surgery and emergency department records. The parent/guardian was then approached in person or via a posted invitation letter and/or information sheet, to ask if they would like to be contacted about the study by a member of the research team. Individuals who agreed to be contacted about the study were then contacted by a research nurse or research assistant, who briefly described the study to them, and asked them if they would like to read a patient information sheet (PIS; see Appendix 1) if not already given. The research nurse or research assistant then provided a PIS to parents who expressed an interest in the study; those who subsequently confirmed their interest in participation were offered a screening appointment at a study site. A second invitation letter was posted to individuals who did not respond to the first invitation letter.

Screening visit 2 weeks prior to first investigational medicinal product dispensing

At the screening visit, an investigator, or a suitably trained person delegated by the investigator (a research nurse or a research assistant who had attended a UK good clinical practice training course), gave an adequate explanation of the aims, methods, anticipated benefits and potential hazards of the study. The eligibility of children to participate in the study was assessed in accordance with the criteria documented in Participants. The investigator then obtained written informed consent (see Appendix 2) from the parent or guardian prior to participation in this study. A period of at least 24 hours, or an overnight stay in hospital (for patients recruited during an acute admission), was required for consideration by the parent or guardian before they gave consent to enter the study. During the consent process it was made clear that parents or guardians were completely free to refuse to enter the study or to withdraw at any time during the study, for any reason. The parents of all eligible children were asked to complete baseline assessments of their child’s wheeze status including recording of baseline demographic and clinical data and details of concomitant medications (see Appendix 3). They also underwent measurement of weight and height, provided a salivary sample for genotyping (see Salivary sample) and gave a urine sample for LT analysis (see Urine sample and Appendix 4). A follow-up appointment was arranged for the issue of the IMP.

Stratification (–1 week)

Saliva samples were posted to the Blizard Institute, Queen Mary University of London, London, UK, where deoxyribonucleic acid (DNA) was extracted and children assigned either to the ALOX5 promoter polymorphism 5/5 stratum or to the [5/x + x/x] stratum, depending on the number of copies of the ALOX5 promoter polymorphism they had on each allele. Extracted DNA was stored in a freezer at –70 °C for later batch analysis of > 150 polymorphisms in 10 genes encoding components of the LT biosynthetic pathway and the LT receptors. The study pharmacist then randomised subjects within their strata, and the corresponding box of active or placebo medication was dispensed for issue at the time of first IMP dispensing (T0) visit (see Figures 1 and 2).

FIGURE 2.

Stratification and randomisation schematic. a, Randomisation key provided by manufacturer, held in sealed envelope by pharmacist and Data Monitoring Committee. T, telephone call; USMA, unscheduled medical attendance; V, visit.

Method of assigning patients to treatment groups: randomisation

Nova Laboratories Ltd (Novalabs, Leicester, UK) prepared the IMP for this trial. Preparation was intended to comprise 6-monthly batches tailored to recruitment rate, with an expectation that 1300 boxes of 50 sachets containing active montelukast and 1300 containing placebo would be produced at a minimum. However, a national shortage of montelukast necessitated a production of boxes containing between 20 and 50 sachets, so as to maintain supply and not compromise recruitment and subject retention. The change in box size received approval from the Medicines and Healthcare products Regulatory Agency (MHRA) prior to implementation. Boxes were allocated randomisation numbers in blocks of 10 using a computer-generated random sequence. Novalabs was responsible for generation of the random number sequence and labelling of boxes. Boxes bearing randomisation numbers were initially delivered to the pharmacy at participating sites. Subsequently, the expansion of site numbers prompted a move to central randomisation and distribution of IMP (from the sponsor pharmacy to participating sites). Novalabs produced additional boxes of IMP for those children whose IMP supply was lost, reached expiry or was exhausted such that they required additional boxes during the 1-year follow-up period. Clinicians remained blinded to allocation throughout.

Randomisation was stratified according to the ALOX5 promoter polymorphism status yielding two genotype groups:

• Group 1: children with the 5/5 ALOX5 promoter polymorphism genotype.

• Group 2: children with the [5/x + x/x] ALOX5 promoter polymorphism genotype, where x ≠ 5 SP1 repeats. (Groups were referred to as stratum A or B.)

Children in each of these two genotype groups (strata) were assigned consecutive randomisation numbers from randomised permuted blocks of 10 representing the randomisation numbers on the IMP boxes. Within each block, equal numbers of children were randomly allocated to placebo and active treatment. When all numbers from the first block had been assigned, a new block of randomisation numbers was allocated to that stratum until a total of 1300 children in the two strata combined had been assigned a randomisation number (see Figure 2).

Blinding

Novalabs produced a corresponding randomisation code denoting whether a given IMP box contained active medication or placebo. This was kept sealed and held only by the clinical trials pharmacist and a member of the Independent Data and Safety Monitoring Committee (DSMC); in this way all other clinical investigators and participants remained blinded to treatment allocation.

T0 visit (0 months)

The research nurse or research assistant met with parents, confirmed eligibility and issued parents a box containing IMP sachets. Parents were taught how to use the IMP. They were also provided with one study diary card (see Appendix 3) and one Freepost return envelope (addressed to the sponsor organisation) per 10 sachets. Parents were asked to return completed diary cards and empty sachets at completion of a course of IMP. Each diary card recorded clinical and IMP usage data for the 10 days of the IMP course.

Telephone calls at 2, 4, 6, 8, 10 and 12 months

At approximately 2-month intervals following the T0 visit, a research nurse or research assistant telephoned the subject’s carer to check if the parent had initiated the IMP, the numbers of days the IMP had been used, the use of health-care resources, any concomitant medications, any procedures, number of days lost from childcare and parent days lost from work. Any AEs experienced were also recorded.

Qualitative interview visit (variable timing)

Qualitative interviews were conducted in a subgroup of families recruited at the sponsor site. The aim of these was to establish attitudes towards genetic testing to guide personalised therapy, acceptability of parent-initiated therapy for preschool wheeze, the expected advantages and disadvantages of using the IMP, and their views on the consent process and PIS. Interviews included either or both parents and, where possible, were conducted at the parental home. Interviewing, transcribing and analysis of interviews was performed by a researcher skilled in qualitative research, in the presence of a translator where necessary.

Withdrawal of patients from therapy or assessment

Patients were free to withdraw from the study at any time without giving a reason. Patients were advised that if they requested to withdraw from the study, at any time during the trial, then this would have no negative consequences. Investigators could also withdraw patients from the trial if they deemed it appropriate for safety or ethical reasons or if it was considered to be detrimental to the well-being of the patient. Where possible, patients who withdrew or were withdrawn underwent a final telephone or face-to-face evaluation. Those participants who withdrew and provided permission to use their data were included in the analysis up to the point of withdrawal. Full documentation was made of any withdrawals that occurred during the study in the case report form (CRF). The investigator documented the date and time of the withdrawal and results of any assessments made at this time. If the patient withdrew because of an AE or a serious adverse event (SAE), then details were forwarded to the Research Ethics Committee, as required, and to the sponsor, who forwarded details to the regulatory authorities as appropriate.

Safety measurements

Montelukast is an established drug with a good safety profile. Safety assessments were limited to standard AE reporting, with patterns monitored by the DSMC.

Other measurements

Subjects underwent the following assessments during the study (see Figure 1 for timings).

Appropriateness of measurements

The primary outcome measure was one that is of importance to patient/carers, clinicians and policy-makers and was deemed more robust to local variations in treatment practices than other measures. It has previously been used in similar studies in this population and is measurable without undue patient inconvenience.

Urine LTE₄ level reflects leukotriene metabolism and has been correlated with asthma severity and bronchoconstriction. 12 A significant correlation with montelukast efficacy would provide both a non-invasive and inexpensive marker to guide treatment.

The anthropomorphic and urine measurements are of minimal inconvenience, while the Oragene saliva kit (DNA Genotek Inc., ON, Canada) yields high concentrations of DNA and is well tolerated by patients.

Respiratory morbidity

• Number of admissions to hospital over the 12-month trial period.

• Duration of admissions to hospital over the 12-month trial period.

• Time to first attack of wheeze.

• Number of unscheduled GP consultations for wheeze.

• Duration of episodes as recorded in the diary card.

• Severity of episodes as recorded in the diary card.

• Parents’ overall impression of efficacy of the IMP.

Health service use

• Unscheduled GP consultation with exacerbation of wheeze, expressed as time from randomisation to first attendance and annual attendance rate.

• Accident and emergency attendance with wheeze exacerbation, expressed as time from randomisation to first attendance and annual attendance rate.

• Unscheduled hospital admission with wheeze exacerbation, expressed as time from randomisation to first admission and annual rate of admissions.

• Total duration of hospital admissions for exacerbation of wheeze.

Adverse events

• SAEs.

• Withdrawal from the trial.

• Mortality due to exacerbation of asthma.

• Mortality due to respiratory infection.

• All-cause mortality.

Medication use

• Use of oral corticosteroids, expressed as number of courses taken per year and proportion of children receiving at least one course of oral corticosteroids during the trial.

• Use of inhaled relief medication (salbutamol), expressed as mean use per wheeze episode as recorded in the diary card by a parent/guardian.

Inflammatory outcomes

Association between baseline urinary cLT level and:

• arachidonate 5-lipoxygenase status

• other polymorphisms of leukotriene genes

• previous history of virus-triggered episodic and multitrigger wheeze

• responsiveness to montelukast

• acute history of wheeze

• urinary cotinine.

Genetic parameters

• Differential responsiveness to montelukast for the primary outcome in the stratum with ALOX5 promoter polymorphism (5/5), compared with the stratum with the ALOX5 [5/x + x/x] genotype.

• Differential responsiveness to montelukast for the primary outcome resulting from other polymorphisms in genes influencing LT synthesis, metabolism and activity.

Qualitative outcomes (parental)

• Attitudes towards genetic testing in order to personalise therapy.

• Acceptability of parent-initiated therapy for preschool wheeze.

• Experience of using the trial medication.

• Difficulties/advantages of the parent-initiated approach.

• Views on the PIS.

Need for unscheduled medical attention

This was defined as an episode requiring an unscheduled attendance at either a general practice or an accident and emergency department, or a combination of both, where wheeze is diagnosed by a clinician.

Time from randomisation to first attack of severe wheeze

This was defined as the number of days from the date of administration of first dose of the IMP to the first date on which a wheeze exacerbation meets the criteria for severity stated in Need for unscheduled medical attention.

Number of days with parent-reported wheeze

This was defined as the number of days with wheeze over the 12-month trial period obtained by telephone contact with the researcher and recorded in the diary card.

Use of inhaled relief medication

This was expressed as the total number of occasions inhaled relief medication was used over the 12-month trial period. The mean number per wheeze episode was obtained from the number of actuations calculated from records in the diary card.

Statistical analysis plan

The statistical analysis plan is available in Appendix 5. The analysis was based on intention-to-treat (ITT) principles.

Determination of sample size

This trial was powered to detect a clinically significant difference in the number of attacks of wheeze between the intervention and control arms. The trial also had power to detect large differential responsiveness (in terms of the primary outcome) to montelukast in the stratum with the ALOX5 promoter polymorphism 5/5 compared with the stratum with the ALOX5 [5/x + x/x] genotype.

Prior to the start of the trial, data on the mean (0.76) and standard deviation (SD) (1.22) number of attacks of wheeze came from the UK GP Research Database on courses of oral steroids (a proxy for number of episodes). These data followed an overdispersed Poisson distribution. To take account of this, a Markov chain Monte Carlo simulation in WinBUGS™ (version 1.4.3; MRC Biostatistics Unit, Cambridge, UK) was used to estimate the sample sizes required to detect a 33% drop in attack rate requiring medical attention, with a power of 90%, a significance level of 5% and a 6% loss to follow-up. In total, 1050 children were required. A 33% drop in attack rates equates to an attack rate of 0.51 for the treatment group. The clinical significance of these changes is that approximately four children will need to be treated to prevent one clinically severe attack. A sample size of 1200 also gave just over 80% power at the 5% significance level to detect an interaction between treatment and genotype if the effect was a 60% reduction in the [5/x + x/x] genotype and a 20% reduction in the 5/5 stratum. In addition, assuming a 6% dropout, 1300 children needed to be recruited.

Analysis of primary end points

Initial analyses were performed according to ITT for all participants with outcome data. Per-protocol efficacy analyses were also performed, excluding data collected after discontinuation of the IMP for those participants who discontinued using the IMP. We used Poisson regression with a random effect representing individuals to account for overdispersion. Fixed effects represent the stratification factor (ALOX5 promoter) and treatment centre. The incidence rate ratio (IRR) (relative risk) and 95% confidence interval (CI) were calculated. The analysis was to be conducted in Stata V12 (StataCorp, College Station, TX, USA). To test for a differential effect by stratum an interaction between stratum and treatment was fitted to this model as described in Genetic analysis.

Analysis of secondary end points

A Poisson regression analysis with a random effect for individuals to allow for overdispersion was applied to determine the influence of treatment allocation on number of days with parent-reported wheeze, number of hospital attendances and number of admissions to hospital. An IRR for each factor is presented with 95% CIs.

Time to first attack of wheeze was analysed using a log-rank test with adjustment for clustering and (where hazards are proportional) Cox’s proportional hazards models adjusting for clustering. In the Cox model, stratum and centre were included as covariates.

Other continuous variables were analysed with analysis of covariance. Dichotomous variables were analysed with logistic regression analysis. AEs are analysed with descriptive statistics.

Genetic analysis

To assess the difference in responsiveness to montelukast in the two ALOX5 strata, an interaction term was fitted for each treatment arm to test for the interaction between montelukast and stratum in our main model. The associations between genotype and clinical phenotype, urinary LT level and clinical outcome were reported. To test the polymorphisms in each gene in combination, a composite likelihood approach was used, which combines the regression coefficients for all polymorphisms at each locus. Analysis of clinical effectiveness (utility) of stratification of ALOX5 status utilised in vitro diagnostic multivariate index assays. The benefits of a multivariate index assay were estimated based on the data in both clinical and economic terms (e.g. days off school, days off work for parents, costs of attendance at GP and hospital, and costs of treatment).

Changes in the conduct of the study or planned analyses

No scientifically significant protocol changes occurred during the study. All amendments were approved by the ethics committee unless the sponsor deemed them to be minor amendments. A list of changes is included in Appendix 6. No changes in planned analysis occurred after the database was locked.

Study duration

The study was intended to recruit for 24 months. Slower than predicted early recruitment necessitated an increase in recruitment period to 26 months and an expansion of recruitment sites. This extension was approved by the Research Ethics Committee, the regulatory authority and also by the funding body. Thus, recruitment spanned from October 2010 to December 2012 and follow-up was completed in December 2013, with data cleaning, verification and database locking completed by January 2014.

Primary outcome

There was no difference between the montelukast and placebo group for the primary outcome (Table 6).

Secondary outcomes

A possible effect was seen within the 5/5 genotype stratum (Figure 4), with a suggestion of increased responsiveness in this group.

FIGURE 4.

Forest plot of unscheduled medical attendances by genotype stratum. USMA, unscheduled medical attendance.

Platform

• High-performance liquid chromatography–mass spectrometry.

• Gas chromatography–mass spectrometry.

Sample requirements:

• Frozen urine, two aliquots of 1 ml (Eppendorf tubes).

Sample preparation

Urine (from the first morning micturition or sampled using a schedule of the clinical protocol) immediately transferred to the laboratory in a 50-ml disposable jar.

Preprocessing

If clear, aliquot into two 1-ml Eppendorf tubes and label. Store frozen at –70 °C. If not clear, centrifuge at 5000 g for 10 minutes in a swinging bucket rotor, then aliquot.

Stability

Tested for 2 years’ storage; no decay of eicosanoids.

Urinary creatinine measurement

Use one aliquot, thaw on ice or in the fridge (4 °C, 3–5 hours). Measure 200 µl using the standard protocol and Vitros 350 Chemistry System (Ortho Clinical Diagnostics, High Wycombe, UK).

Organic phase extraction

Thaw on ice or in the fridge (4 °C, 3–5 hours; batch up to 20 samples). Adjust pH to 3.5 with 1 N HCl (30–80 µl), check pH using the narrow-range pH stick. Add internal deuterated standards mix containing: LTE₄-d₃ (2 ng), tetranor-PGE-M-d₆ (10 ng), tetranor-PGD-M-d₆ (10 ng), 13,14-dehydro,15-keto PGE₂-d₄ (1 ng), 13,14-dehydro,15-keto PGD₂-d₄ (1 ng), 13,14-dehydro,15-keto-tetranor-PGE₂-d₄ (1 ng), 13,14-dehydro,15-keto-tetranor-PGD₂-d₄ (1 ng), 9α11β-PGF₂-d₄ (1 ng), 15-deoxy,delta-12,14-PGJ₂-d₄ (1 ng) in methanol – 10 µl of the mix. If uric acid precipitate is present, spin for 10 minutes at 10,000 g at 4 °C (in a microcentrifuge) and transfer the supernatant to a fresh tube. Mix in a conical 10-ml tube with 1 ml of tertiary-butylmethyl-ether (TBM), vortex for 2 minutes, spin as before. Collect upper organic phase to fresh tube, repeat extraction with another 1 ml of TBM, combine organic phases. Dry at room temperature under nitrogen flow (1 l/minute) for 30 minutes. Dissolve in 60 ml of methanol and immediately proceed with analysis.

High-performance liquid chromatography–tandem mass spectrometry equipment

Autosampler (Shimadzu Sil-2-AC; Shimadzu Europa GmbH, Duisburg, Germany), reverse phase column (Zorbax Eclipse XDB C-18; Agilent Technologies, Santa Clara, CA, USA) stabilised thermally at 37 °C, multiple reaction monitoring mode (MRM) tandem mass spectrometry (Qtrap 4000; Applied Biosystems, MA, USA) equipped with electrospray ion source negative ionisation mode, use batch protocol for urinary eicosanoids.

Test: inject 10 µl of internal standard mix. Check for area under the peak > 20,000.

Injection

10 µl of methanol extract

Elution

Gradient consisting of two mobile phases: (1) acetonitril/water/acetic acid (20/80/0.01) and (2) acetonitril/isopropanol/acetic acid (55/45/0.01) using the flow rate of 0.11 ml/minute.

Gas chromatography–mass spectrometry