Results

W Lenney; AJ McKay; C Tudur Smith; PR Williamson; M James; D Price; MASCOT Study Group

Chapter 3Results

Early trial closure

Recruitment rates were closely monitored throughout the study, which highlighted within 8 months of the first patient being registered that additional strategies and additional centres would be required to improve recruitment and achieve the target of 450 patients randomised. Site visits were undertaken from October 2009 and a number of new strategies were developed through discussion with sites (see Appendix 3). A funding extension application was submitted to the NIHR HTA programme in February 2010 requesting an extension of recruitment time together with funding for new sites. One of the differences noted about recruitment centres that were more successful than others was how well integration had occurred between staff within the newly formed Medicines for Children Research Network (MCRN), Primary Care Research Network (PCRN) and Comprehensive Local Research Network (CLRN). If the study was to succeed it would also have been necessary to open other recruitment centres and these had been carefully selected based on the knowledge we had accrued through discussions and particularly the ability to liaise and work effectively between secondary and primary care. It was recognised that some of the initial recruitment sites would have to be closed. Funding was not granted, however, and the study was closed down prematurely on 24 June 2010. For those patients who were already randomised, follow-up was to continue to T48 or the end of January 2011, whichever was the earliest, as January 2011 was the date that the medications expired. Data cleaning and site close-down visits took place between February and July 2011, with final analyses undertaken during August and September 2011.

Participant flow and recruitment

The first patient was registered on 27 January 2009, the first patient was randomised on 19 May 2009, the last patient was registered on 25 June 2010 and the last patient was randomised on 24 June 2010. Table 3 shows all 13 recruiting centres, the date each site was initiated, the original target recruitment figure, the number of participants registered, the number of participants randomised and the dates of the first and last randomisation. All centres registered at least one patient and 12 centres randomised at least one patient.

TABLE 3

Recruitment by centre

A total of 166 patients were registered at T–4 and 63 (38%) of these were eligible and consented to be randomised at T0. The percentage randomised at each centre ranged between 20% and 58% for those that randomised at least one patient.

Figure 1 provides a CONSORT flow diagram showing the numbers of participants recruited and randomly assigned to the three trial arms.

FIGURE 1

CONSORT flow diagram. SABA, short-acting beta-2 agonist. All patients who did not reach the minimum of 43 weeks because of early closure of the trial have primary outcome data up to at least T24.

Baseline comparability of randomised groups

A summary of the baseline characteristics for all randomised participants is given in Table 4. Overall there was a preponderance of males with the major comorbid condition being eczema, neither of which was unexpected. The male-to-female ratio in children with asthma is approximately 2 : 1 and there is a strong association between eczema and asthma in children. The age of the children ranged from 6.50 to 14.67 years (average 10.39 years).

TABLE 4

Baseline characteristics for all randomised participants

The distribution of characteristics is similar across treatment groups but a higher percentage of males were randomised to fluticasone alone and a higher percentage of children with respiratory and dermatological abnormalities were randomised to fluticasone plus salmeterol. Because of the small number of children recruited this probably occurred by chance but could suggest that patients randomised to fluticasone plus salmeterol had more severe asthma.

No ineligible patients were randomised.

Unblinding of randomised treatments

No patient required unblinding from randomised treatment other than at the end of the study. The process was described in Chapter 2, Randomisation and blinding.

Protocol deviations

Protocol deviations were classified as major or minor according to a preplanned classification system outlined in the SAP (see Appendix 2). There were 18 minor protocol deviations reported for 17 patients (three randomised to fluticasone, five randomised to fluticasone plus salmeterol, two randomised to fluticasone plus montelukast and seven who were not randomised) (Table 5). There were no major protocol deviations. Eleven of the protocol deviations were due to visits that occurred outside the predefined time interval, one of which led to missing data. One protocol deviation occurred because two different parents completed the quality of life questionnaire at two different visits. Six protocol deviations were related to the spirometry test undertaken at T0 or T48.

TABLE 5

Protocol deviations

Efficacy outcomes

Primary outcome

Number of asthma exacerbations requiring treatment with oral corticosteroids

In total, 38 (60%) randomised patients completed their 48-week visit and 54 (86%) patients completed their 24-week visit and could be included in the ITT analyses (Tables 6 and 7 respectively). Reasons for exclusions are provided in Appendix 6. For the primary analysis (ITT at 48 weeks), seven (18%) patients had at least one exacerbation requiring oral steroids, with most having only one episode. The effect of treatment overall (p = 0.98) and for each pair-wise comparison was not significant (Table 8). Because of limited data, the CI around the RR for each pair-wise comparison is extremely wide and includes values of RR that could indicate beneficial, harmful or equivalent treatment effects (see Table 8). The secondary ITT analysis at 24 weeks includes data for 54 patients (see Table 7) of whom 10 (18.5%) had experienced at least one exacerbation. Similarly to the 48-week analysis, all treatment effects are not significant with very wide CIs (see Table 8). The per-protocol analyses show similar results and uncertainty to the ITT analyses (see Table 8). Because of the limited number of patients and events available for parameter estimation, secondary analyses adjusting for prognostic factors were not undertaken as planned in the SAP.

TABLE 6

Exacerbations requiring course of oral steroids over 48 weeks (ITT analysis)

TABLE 7

Exacerbations requiring course of oral steroids over 24 weeks (ITT analysis)

TABLE 8

Relative treatment effects from Poisson regression model adjusted for overdispersion: number of exacerbations requiring oral corticosteroids

Secondary outcomes

Quality of life of children as measured by the Paediatric Asthma Quality of Life Questionnaire with Standardised Activities

A total of 37 children completed the quality of life questionnaire at baseline and at 48 weeks' follow-up. The results show an improvement in mean scores for activity limitations, symptoms and emotional function across all treatment groups at 48 weeks, with the largest improvement in mean score for fluticasone plus montelukast across all domains (Table 9). However, the pair-wise treatment comparison results from ANCOVA show that none of the mean differences adjusted for baseline score is statistically significant. The 24-week analysis results based on 50 patients show an improvement in mean scores for activity limitations, symptoms and emotional function across all treatment groups (Table 10). The largest improvement was seen for fluticasone plus salmeterol across all domains but none of the mean differences adjusted for baseline score is statistically significant.

TABLE 9

Quality of life of children measured using the PAQLQ(S) over 48 weeks

TABLE 10

Quality of life of children measured using the PAQLQ(S) over 24 weeks

Quality of life of caregivers as measured by the Paediatric Asthma Caregiver's Quality of Life Questionnaire

A total of 38 caregivers completed the quality of life questionnaire at baseline and at 48 weeks. The results show an improvement in mean scores for activity limitations and emotional function across all treatment groups at 48 weeks (Table 11). In support of the PAQLQ(S) analysis, the largest improvement in mean score at 48 weeks is seen for fluticasone plus montelukast across all domains. The pair-wise treatment comparison results from ANCOVA, however, show that none of the mean differences adjusted for baseline score is statistically significant. Similar results were obtained using data up to 24 weeks (Table 12).

TABLE 11

Quality of life of caregivers measured using the PACQLQ over 48 weeks

TABLE 12

Quality of life of caregivers measured using the PACQLQ over 24 weeks

Time from randomisation to first exacerbation requiring treatment with a short course of oral corticosteroids

In total, 63 randomised patients could be included in the analysis of time to first exacerbation (Table 13) [median length of follow-up 47 weeks (range 0 to 57 weeks, IQR 33 to 49 weeks) for the 49 censored patients]. Overall, there were 14 (22.2%) patients who had at least one exacerbation event, with most events in the fluticasone plus salmeterol group [seven events (30.4%)] and least events in the fluticasone plus montelukast group [three events (14.3%)]. No statistically significant difference in time to first exacerbation could be detected between the three treatment groups with the log-rank test (Figure 2, χ² = 1.90, 2df, p = 0.39). Because of the limited data, wide CIs for pair-wise HRs (Table 14) make it impossible to draw any meaningful conclusions.

TABLE 13

Time to first exacerbation

FIGURE 2

Kaplan–Meier survival curves for time to first exacerbation by treatment.

TABLE 14

Hazard ratios for time to first exacerbation

Number of school days missed due to respiratory problems

A total of 37 children with data available up to 48 weeks were included in the ITT analysis. Fewer children missed at least 1 day of school over 48 weeks on fluticasone plus montelukast (18.2%) than on fluticasone (63.6%) and fluticasone plus salmeterol (60%) (Table 15); however, this pattern is not supported by the 24-week data, based on 54 patients, which show that a similar percentage of patients missed at least 1 day across treatment groups (Table 16). The RRs for number of school days missed, estimated from Poisson regression, are not statistically significant (Table 17) and wide CIs prevent useful conclusions for the 48-week and 24-week analyses.

TABLE 15

Number of school days missed over 48 weeks (ITT analysis)

TABLE 16

Number of school days missed over 24 weeks (ITT analysis)

TABLE 17

Relative treatment effects from Poisson regression model adjusted for overdispersion: number of school days missed

Number of hospital admissions due to respiratory problems

Tables 18 and 19 show the number of hospital admissions over 48 weeks and 24 weeks respectively. Two hospital admissions occurred in two patients on fluticasone plus salmeterol over 48 weeks (see Table 18). Analysis of the number of admissions with Poisson regression was not possible because of sparse data and no conclusions can be drawn (Table 20). For the patients with data up to 24 weeks, four (7.4%) patients required at least one hospital admission, with a similar frequency across treatment groups (see Table 19). The CIs for the 24-week RRs estimated from Poisson regression (see Table 20) are very wide and include relative effects in both directions.

TABLE 18

Number of hospital admissions over 48 weeks (ITT analysis)

TABLE 19

Number of hospital admissions over 24 weeks (ITT analysis)

TABLE 20

Relative treatment effects from Poisson regression model adjusted for overdispersion: number of hospital admissions

Amount of rescue beta-2 agonist therapy prescribed for asthma symptoms

A total of 23 out of 37 (62.2%) children were prescribed at least one beta-2 agonist over 48 weeks of follow-up (Table 21), with a median total dose per patient of 24,000 μg. In total, 28 out of 53 (52.8%) children were prescribed at least one beta-2 agonist over 24 weeks of follow-up (Table 22), with a median total dose per patient of 20,000 μg. There was no significant difference in the mean amount prescribed between the three groups from ANOVA at 48 weeks (F-value = 0.96, 2df, p = 0.39) or 24 weeks (F-value = 1.2, 2df, p = 0.31) (Table 23).

TABLE 21

Number of beta-2 agonists prescribed over 48 weeks

TABLE 22

Number of beta-2 agonists prescribed over 24 weeks

TABLE 23

Difference in mean amount of beta-2 agonists prescribed (ANOVA)

Time from randomisation to treatment withdrawal (because of lack of efficacy or side effects)

Although 13 patients were withdrawn from treatment or the study, only one (on fluticasone) withdrew from treatment because of poor asthma control (patient also recorded intercurrent illness as reason) during follow-up (see Table 34). The planned analysis of time to treatment withdrawal was therefore not possible.

TABLE 34

Withdrawals from the study and treatment

Lung function at 48 weeks assessed by spirometry

A total of 30 children had lung function assessed by spirometry at baseline and at 48 weeks and could be included in ANCOVA (Table 24). There were no statistically significant adjusted mean differences between the three treatment groups, with wide CIs that include clinically significant differences in both directions.

TABLE 24

Lung function over 48 weeks assessed by spirome

Safety outcomes

Adverse events were assessed at each visit between randomisation and the final visit. Adverse events were classified according to the Medical Dictionary for Regulatory Activities (MedRA) (www.meddramsso.com) System Organ Class by the MCRN CTU Data Manager, with all classifications approved by the chief investigator after discussion with the trial working group.

The number (and percentage) of patients experiencing each adverse event is shown for each treatment in Table 25; this is categorised further by severity in Table 26. For each patient who experienced more than one episode of the same adverse event, the maximum severity is shown. No formal statistical testing was undertaken on these data.

TABLE 25

All adverse events

TABLE 26

All adverse events according to severity

In total, 53 out of 63 (84%) randomised patients experienced at least one adverse event between randomisation and last follow-up, with a similar percentage of patients across treatment groups [fluticasone 17 (89%); fluticasone plus salmeterol 18 (78%); fluticasone plus montelukast 18 (86%)]. The most common events were respiratory and nervous system disorders, infections and infestations, and general and gastrointestinal disorders. The percentage of patients reporting each adverse event was similar across treatment groups (see Table 25) except for nervous system disorders. Fewer patients reported nervous system disorders on fluticasone plus salmeterol [one patient (4.3%)] than on fluticasone plus montelukast [seven patients (33.3%)] and fluticasone [five patients (26.3%)], with a much greater number of events on fluticasone plus montelukast (37 events compared with 13 on fluticasone and two on fluticasone plus salmeterol) because of one patient experiencing 20 events. All adverse events were graded as mild or moderate (see Table 26) with a similar distribution of severity across treatment groups except for the nervous system disorders described previously.

Serious adverse events and suspected unexpected serious adverse reactions

Between randomisation and trial closure, six patients reported a total of seven serious adverse events (SAEs) (Table 27), with similar frequency across treatment groups (two patients reported three events on fluticasone; three patients reported three events on fluticasone plus salmeterol; one patient reported one event on fluticasone plus montelukast). All SAEs were mild or moderate in severity and all were classified as unrelated to treatment. There were no suspected unexpected serious adverse reactions (SUSARs) and no deaths reported during the trial.

TABLE 27

All SAEs from randomisation

Health economics

The sample size is extremely small in each group; therefore, although every endeavour has been made to fully analyse and report the economic results and to take necessary statistical safeguards, the results should be interpreted with extreme caution.

It can be seen from Table 28 that there was some loss of patients completing the health economics diary over each time period but completion of the health economics booklet was very similar across the groups. This included patients whose 48-week visit took place within 48 weeks because of the early closure of the trial. We removed patients from the 48-week visit analysis when the final visit took place inside 43 weeks from baseline.

TABLE 28

Completion of the health economics booklet at each time period

Table 29 indicates the total events of care (resource use) for each of the three treatment groups during the study period.

TABLE 29

Total resource use (events of care)

Table 30 shows the main cost drivers in the treatment of childhood asthma. The main cost driver, not surprisingly, is the cost of the intervention itself, followed by the costs of GP visits, prescribed inhalers and prescribed medications. With larger patient numbers these results could change.

TABLE 30

Total resource costs

Given that the cost of the trial intervention is the greatest cost driver it is not surprising that the fluticasone plus salmeterol group and the fluticasone plus montelukast group have the greatest overall costs. That said, even though the intervention cost is higher in the fluticasone plus montelukast group, once the overall costs have been factored in it is the fluticasone plus salmeterol group that has the greatest overall cost, although the difference is small. This shows the importance of calculating the full economic cost of an intervention, not just the cost of the intervention itself.

Outcome

The QALY gains shown in Table 31 are based upon independent bootstrapped samples of n = 2000.

TABLE 31

Quality-adjusted life-year gains

Bootstrapped t-tests and Wilcoxon rank-sum tests did not reveal any statistically significant differences in mean QALYs between baseline and 48 weeks when comparing any of the treatment groups in two-way tests. Figure 3 provides a graphical representation of quality of life over time.

FIGURE 3

Quality-adjusted life-years over time in the trial. A, fluticasone; B, fluticasone plus salmeterol; C, fluticasone plus montelukast.

Cost-effectiveness analysis

In terms of economics the variables of interest between the groups are the differences in costs and effects. Table 32 shows the mean differences in incremental costs and QALY gains between the three groups.

TABLE 32

Mean differences in costs and QALYs by group using bootstrapped data

The measure of interest from these data is incremental cost-effectiveness, captured by the ICER, which shows the difference in the sum of the costs over effects for each individual patient between each of the treatment arms. Table 31 shows that there is very little difference in QALYs before and after the interventions. Similarly, cost differences are not great between the groups, with fluticasone plus salmeterol and fluticasone plus montelukast incurring approximately three times the cost per person as fluticasone. The difference in outcomes was greater for fluticasone plus salmeterol and fluticasone plus montelukast even though the cost was greater; hence, it can be seen that a positive ICER is generated for fluticasone plus salmeterol compared with fluticasone and fluticasone plus montelukast compared with fluticasone (Table 33).

TABLE 33

Incremental cost-effectiveness ratios

It can be seen from the ICER results that the treatment of choice is inhaled fluticasone propionate 100 μg twice daily plus montelukast 5-mg tablet once daily, as this produces 1 additional QALY for £6827, whereas it would cost £12,054 to produce an additional QALY from inhaled fluticasone propionate 100 μg and salmeterol 50 μg twice daily (combination inhaler) plus placebo tablet once daily. Both are by comparison with inhaled fluticasone propionate 100 μg. To further aid interpretation of the results the CEACs are presented for fluticasone plus salmeterol and fluticasone plus montelukast compared with fluticasone, and fluticasone plus montelukast compared with fluticasone plus salmeterol (Figure 4).

FIGURE 4

Cost-effectiveness acceptability curves. A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

Figure 4 shows the probabilities that fluticasone plus salmeterol and fluticasone plus montelukast are under the £30,000 per QALY cost-effectiveness threshold as required by NICE. The probability of cost-effectiveness at this level is around 80% for fluticasone plus montelukast and 60% for fluticasone plus salmeterol. Moreover, the comparison between fluticasone plus montelukast and fluticasone plus salmeterol shows that fluticasone plus montelukast dominates fluticasone plus salmeterol; however, the CEAC for fluticasone plus montelukast compared with fluticasone plus salmeterol declines after the threshold ICER because of increasing uncertainty, reflecting the lack of statistical power in the incremental QALYs. In fact, there is very little evidence supporting fluticasone plus salmeterol or fluticasone plus montelukast over the other.

Figure 5 illustrates the relative spread of incremental costs and QALYs for fluticasone plus salmeterol and fluticasone plus montelukast compared with fluticasone on a cost-effectiveness plane. This is useful because it shows that, although fluticasone plus montelukast has a lower average incremental cost and higher average incremental QALY gain, it also has greater uncertainty, which is why fluticasone plus salmeterol and fluticasone plus montelukast cannot easily be compared directly.

FIGURE 5

The cost-effectiveness plane.

Withdrawals

In total, 13 (20.6%) patients withdrew from the study before the 48-week visit or before early study closure. The numbers of withdrawals and reasons for withdrawal are similar across the three treatment groups (Table 34). Six of the 13 patients had indicated a withdrawal from treatment but no further follow-up data were available. There were no crossovers to another treatment arm.

Compliance

The number of doses missed (inhaler and tablets) since the previous visit was captured on the follow-up CRF. The expected number of tablet doses for each visit was calculated as the number of days since the last visit (i.e. one per day) and the expected number of inhaler doses was calculated as twice the number of days since the last visit (i.e. two per day).

For each visit, each patient had the proportion of doses missed (number of doses missed/number of expected doses) calculated separately for inhalers and tablets. These proportions were averaged over all patients per treatment group separately for inhalers and tablets and are displayed as summary measures in Table 35. The proportion of treatments missed is similar across the groups and time points, except for fluticasone plus montelukast at the 48-week visit.

TABLE 35

Summary of adherence to therapy

Outcome of non-randomised patients

The GPs of all 103 children who were registered at T–4 but who were not randomised at T0 were contacted towards the end of the study and asked to complete a questionnaire about each of their patients, as described in the protocol. As the exposure time since date of registration varied, data are summarised according to the time intervals for which the data were collected. The data presented for ‘up to 1 year’ are likely to be an underestimate of events because only the cumulative data over the period from registration until completion of the form were collected. Some patients with data reported over periods > 1 year may have had a number of their events during the first year from registration.

The questionnaire was kept very simple and short to try and maximise response. Nine (14.5%) children had at least one exacerbation that required a course of oral corticosteroids (Table 36). A total of 50 (80.6%) required at least one prescription of beta2 agonist after registration (Table 37). Unfortunately it was not possible to summarise the amount of beta2 agonist prescribed as the reporting of type of inhaler and amount prescribed was inconsistent and incomplete. In total, 18 (29.0%) patients were prescribed at least one additional treatment for asthma control (Table 38), with the majority (61%) of these being combination therapy (ICS plus long-acting beta2 agonist). Four (6.5%) children required an A&E visit (Table 39) and one (1.6%) child needed a hospital admission for their asthma (Table 40).

TABLE 36

Number of exacerbations requiring a course of oral corticosteroids since T–4 and time to first exacerbation

TABLE 37

Number of beta-2 agonists prescribed since T–4

TABLE 38

Prescriptions since T–4

TABLE 39

Number of A&E admissions due to respiratory problems since T–4

TABLE 40

Number of inpatient admissions due to respiratory problems since T–4

Publication Details

Copyright

Copyright © Queen's Printer and Controller of HMSO 2013. This work was produced by Lenney et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Included under terms of UK Non-commercial Government License.

Publisher

NIHR Journals Library, Southampton (UK)

NLM Citation

Lenney W, McKay AJ, Tudur Smith C, et al.; MASCOT Study Group. Management of Asthma in School age Children On Therapy (MASCOT): a randomised, double-blind, placebo-controlled, parallel study of efficacy and safety. Southampton (UK): NIHR Journals Library; 2013 Feb. (Health Technology Assessment, No. 17.4.) Chapter 3, Results.

TABLE 3

Recruitment by centre

Centre code	Hospital	Date site initiated	Original target recruitment	Date of first registration (T-4)	Date of last registration	Number registered	Date of first randomisation (T0)	Date of last randomisation	Date of last follow-up visit	Number randomised
002	Royal Devon & Exeter NHS Foundation Trust	19 November 2008	23	16 February 2009	25 May 2010	10	13 October 2009	9 March 2010	7 February 2011	3
009	Arrowe Park Hospital, Wirral	20 January 2009	13	27 January 2009	23 June 2010	26	28 July 2009	15 June 2010	26 January 2011	12
031	Leicester Royal Infrmary	11 February 2009	45	28 April 2009	8 June 2010	21	2 June 2009	24 June 2010	28 January 2011	7
036	Norfolk and Norwich University Hospitals^a	01 April 2009	45	2 July 2009	28 April 2010	8	4 November 2009	26 May 2010	24 November 2010	3
116	Bristol Royal Hospital for Children	08 December 2008	23	26 June 2009	28 May 2010	7	24 July 2009	18 November 2009	22 October 2010	2
134	Royal London Hospital	23 December 2008	45	9 February 2009	17 May 2010	17	26 May 2009	25 May 2010	18 October 2010	8
168	Derbyshire Children's Hospital	18 February 2009	23	21 April 2009	3 June 2010	10	19 May 2009	3 June 2010	18 November 2010	3
213	Queen's Medical Centre, Nottingham	06 April 2009	23	4 June 2009	21 June 2010	10	29 June 2009	11 May 2010	8 February 2011	4
243	Alder Hey Children's NHS Foundation Trust	26 February 2009	32	23 June 2009	21 June 2010	14	21 July 2009	22 June 2010	27 January 2011	6
246	Royal Manchester Children's Hospital	23 December 2008	45	5 June 2009	25 June 2010	7	9 December 2009	29 January 2010	12 January 2011	2
451	Royal Aberdeen Children's Hospital	13 January 2009	45	20 August 2009	9 June 2010	10	24 February 2010	24 February 2010	24 January 2011	2
511	University Hospitals Coventry and Warwickshire NHS Trust	26 January 2009	45	4 February 2009	21 June 2010	7	N/A	N/A	N/A	0
522	University Hospital of North Staffordshire	06 April 2009	45	26 June 2009	21 May 2010	19	24 July 2009	16 June 2010	11 January 2011	11
Total			452			166				63

: N/A, not applicable.

a: Includes Norfolk GP practices.

FIGURE 1

CONSORT flow diagram. SABA, short-acting beta-2 agonist. All patients who did not reach the minimum of 43 weeks because of early closure of the trial have primary outcome data up to at least T24.

TABLE 4

Baseline characteristics for all randomised participants

Baseline characteristic	Fluticasone (n = 19)	Fluticasone plus salmeterol (n = 23)	Fluticasone plus montelukast (n = 21)	Total (n = 63)
General
Age (years), mean (SD), range (n)	10.37 (1.82), 7.33 to 13.58 (19)	10.46 (2.33), 6.50 to 13.75 (23)	10.33 (2.37), 6.50 to 14.67 (21)	10.39 (2.17), 6.50 to 14.67 (63)
Sex, n/N (%)
Male	17/19 (89.5)	13/23 (56.5)	10/21 (47.6)	40/63 (63.5)
Female	2/19 (10.5)	10/23 (43.5)	11/21 (52.4)	23/63 (36.5)
Physical examination
Height (cm), mean (SD), range (n)	143.18 (15.63), 121.00 to 184.40 (19)	144.60 (12.11), 128.00 to 166.00 (23)	141.56 (15.98), 117.60 to 167.30 (21)	143.16 (14.38), 117.60 to 184.40 (63)
Weight (kg), mean (SD), range (n)	39.88 (12.44), 22.00 to 59.50 (19)	42.28 (11.49), 22.90 to 64.80 (23)	38.56 (12.41), 20.50 to 64.80 (21)	40.31 (12.00), 20.50 to 64.80 (63)
General appearance, n/N (%)
Normal	19/19 (100.0)	23/23 (100.0)	21/21 (100.0)	63/63 (100.0)
Abnormal	0/19 (0.0)	0/23 (0.0)	0/21 (0.0)	0/63 (0.0)
Ear, nose, throat, n/N (%)
Normal	17/19 (89.5)	19/23 (82.6)	19/20 (95.0)	55/62 (88.7)
Abnormal^a	2/19 (10.5)	4/23 (17.4)	1/20 (5.0)	7/62 (11.3)
Cardiovascular, n/N (%)
Normal	19/19 (100.0)	23/23 (100.0)	20/20 (100.0)	62/62 (100.0)
Abnormal	0/19 (0.0)	0/23 (0.0)	0/20 (0.0)	0/62 (0.0)
Respiratory, n/N (%)
Normal	18/19 (94.7)	17/23 (73.9)	18/20 (90.0)	53/62 (85.5)
Abnormal^b	1/19 (5.3)	6/23 (26.1)	2/20 (10.0)	9/62 (14.5)
Gastrointestinal, n/N (%)
Normal	19/19 (100.0)	22/23 (95.7)	18/20 (90.0)	59/62 (95.2)
Abnormal^c	0/19 (0.0)	1/23 (4.3)	2/20 (10.0)	3/62 (4.8)
Endocrine, n/N (%)
Normal	19/19 (100.0)	23/23 (100.0)	20/20 (100.0)	62/62 (100.0)
Abnormal	0/19 (0.0)	0/23 (0.0)	0/20 (0.0)	0/62 (0.0)
Dermatological, n/N (%)
Normal	18/19 (94.7)	18/23 (78.3)	17/20 (85.0)	53/62 (85.5)
Abnormal^d	1/19 (5.3)	5/23 (21.7)	3/20 (15.0)	9/62 (14.5)
Musculoskeletal, n/N (%)
Normal	18/19 (94.7)	23/23 (100.0)	20/20 (100.0)	61/62 (98.4)
Abnormal^e	1/19 (5.3)	0/23 (0.0)	0/20 (0.0)	1/62 (1.6)
Other site,^f n/N (%)
Normal	10/10 (100.0)	7/7 (100.0)	10/10 (100.0)	27/27 (100.0)
Abnormal	0/10 (0.0)	0/7 (0.0)	0/10 (0.0)	0/27 (0.0)
Systolic blood pressure (mmHg), mean (SD), range (n)	103.89 (13.11), 80.00 to 135.00 (19)	106.96 (10.12), 90.00 to 134.00 (23)	102.24 (10.86), 84.00 to 119.00 (21)	104.46 (11.33), 80.00 to 135.00 (63)
Diastolic blood pressure (mmHg), mean (SD), range (n)	62.84 (12.11), 46.00 to 89.00 (19)	63.87 (8.38), 49.00 to 82.00 (23)	60.00 (8.63), 44.00 to 77.00 (21)	2.27 (9.71), 44.00 to 89.00 (63)
Pulse rate (beats/minute), mean (SD), range (n)	80.58 (17.81), 57.00 to 118.00 (19)	85.65 (14.96), 54.00 to 125.00 (23)	85.48 (15.35), 55.00 to 114.00 (21)	84.06 (15.90), 54.00 to 125.00 (63)
Respiratory rate (breaths/ minute), mean (SD), range (n)	21.89 (4.84), 16.00 to 36.00 (19)	20.48 (2.85), 16.00 to 25.00 (23)	21.33 (3.37), 14.00 to 30.00 (21)	21.19 (3.70), 14.00 to 36.00 (63)
Spirometry^g
FEV₁ (l), mean (SD), range (n)	1.98 (0.89), 0.90 to 4.70 (18)	1.83 (0.68), 0.80 to 3.10 (19)	1.82 (0.77), 0.20 to 3.20 (19)	1.88 (0.77), 0.20 to 4.70 (56)
FVC (l), mean (SD), range (n)	2.29 (0.92), 1.20 to 4.70 (18)	2.20 (0.60), 1.30 to 3.20 (19)	2.35 (0.77), 0.80 to 3.50 (19)	2.28 (0.76), 0.80 to 4.70 (56)
FEV₁ (%), mean (SD), range (n)	88.29 (17.55), 48.0 to 116.0 (17)	79.79 (19.90), 47.0 to 106.0 (19)	86.47 (13.32), 56.0 to 104.0 (19)	84.73 (17.21), 47.0 to 116.0 (55)
FVC (%), mean (SD), range (n)	92.06 (15.09), 57.0 to 112.0 (17)	87.00 (16.91), 49.0 to 112.0 (19)	94.05 (15.32), 53.0 to 128.0 (19)	91.00 (15.82), 49.0 to 128.0 (55)
FEV₁ (l) after 400 µg salbutamol (GSK) administered, mean (SD), range (n)	2.19 (0.84), 0.80 to 4.30 (15)	1.99 (0.67), 0.80 to 3.10 (20)	2.04 (0.68), 0.80 to 3.20 (18)	2.06 (0.72), 0.80 to 4.30 (53)
FVC (l) after 400 µg salbutamol administered, mean (SD), range (n)	2.60 (0.94), 1.30 to 4.80 (15)	2.38 (0.62), 1.30 to 3.70 (20)	2.34 (0.80), 0.90 to 3.60 (18)	2.43 (0.78), 0.90 to 4.80 (53)
FEV₁ (%)after 400 µg salbutamol administered, mean (SD), range (n)	90.64 (17.95), 46.0 to 123.0 (14)	86.30 (19.75), 49.0 to 111.0 (20)	92.28 (13.94), 56.0 to 110.0 (18)	89.54 (17.30), 46.0 to 123.0 (52)
FVC (%) after 400 µg salbutamol administered, mean (SD), range (n)	94.21 (17.17), 58.0 to 124.0 (14)	91.50 (15.83), 55.0 to 117.0 (20)	95.06 (15.51), 54.0 to 118.0 (18)	93.46 (15.85), 54.0 to 124.0 (52)

: A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

a: Abnormal ear, nose, throat: A: infamed nasal mucosa 1, glue ear – left 1; B: tonsils moderate size: slightly infected 1, blocked left nostril 1, very swollen turbinates on left and blocking of nasal passage 1, slight upper respiratory tract infection – runny nose (sore throat 2 days ago – now better) 1; C: red tympanic membranes, red throat 1.

b: Abnormal respiratory: A: occasional end expiratory wheeze 1; B: wheezy 1, occasional expiratory wheeze 1, slight wheeze 1, crackles left base 1, few high-pitched wheezes 1, Harrison's sulcus – mild 1; C: crackles left base and is clear after cough 1, mild wheeze throughout 1.

c: Abnormal gastrointestinal: A: 0; B: complaining of abdominal pain hunger/diet related, has appointment with dermatology/allergy team 1; C: constipation and abdominal cramps, investigated, advice to change diet, start on medication, now getting better/going to see paediatrician 1, slightly tender in right iliac fossa and left iliac fossa 1.

d: Abnormal dermatological: A: mild eczema 1; B: hives – on chlorphenamine maleate (Piriton, GSK Consumer Healthcare) (mild) 1, eczema – hands, elbow fexures, perioral 1, eczema in cubital fossa right and left 1, mild eczema right hand 1, fexural eczema, mainly face 1; C: eczema right cubital fossa 1, mild lichenifcation fexural creases 1, mild eczema 1.

e: Abnormal musculoskeletal: A: hypermobile joints 1; B: 0; C: 0.

f: Optional question that only 27 out of 63 answered (10, 7 and 10 in treatment groups A, B and C respectively); no abnormal results.

g: Reasons spirometry not undertaken: A: technician not available 1; B: patient unable to perform correct technique 1, no time to allow for post-reversibility test 1, blue inhaler was not brought to appointment, patient was reminded prior to appointment 1; C: technician not available 1, equipment not available 1.

TABLE 5

Protocol deviations

Patient registration number	Date of deviation	Nature of deviation	Impact (minor/major)
10009	2 November 2009	The T0 visit occurred on the 31st day following the T–4 visit. The family were unable to make the T0 visit within the 30-day time frame because of other commitments^a	Minor
11006	13 November 2009	Patient's T0 visit date occurred 32 days following the T–4 visit. Site are unable to explain the reason for this as the member of staff has now left the trust^a	Minor
13010	18 October 2010	Patient was unable to attend the scheduled T48 visit (22 November 2010) because of religious commitments and holiday out of the country. Visit was carried out 5 weeks after the T36 visit on 18 October 2010 (42 weeks after randomisation)	Minor
22001	10 August 2009	Time period between the T–4 and T0 visits exceeded 30 days^a	Minor
30004	11 February 2010	The patient was seen for their T0 visit 22 days after the T–4 visit. This was checked with the CTU prior to the visit as Dr Turner was unable to see the patient the following week^a	Minor
30007	7 April 2010	Site provided a file note to explain that the patient was due to come in for her T0 visit on 1 April 2010; however, the PI was on annual leave that week. An attempt was made to sign up a co-investigator to see the patient; however, this was unsuccessful because of their Good Clinical Practice training being inadequate and the delegation log not being signed prior to the PI going on leave. The CTU was contacted and it was agreed that the patient could extend her run-in phase to 34 days and be seen for her T0 visit on 7 April 2010. The patient was therefore reviewed by a member of child health's medical team on 1 April 2010 and prescribed another fluticasone propionate Accuhaler 100 µg twice daily (GSK) to ensure that she had an adequate supply until her rescheduled appointment on 7 April 2010^a	Minor
60008	3 June 2010	Patient started the run-in period on 22 April 2010 but it was arranged at site that the child would not begin taking the medications until 3 May 2010. The T0 visit was scheduled for 28 days following 3 May 2010. This is a deviation because the run-in period for this child was 6 weeks (although only 4 weeks were spent on run-in trial medications). T0 visit occurred on 3 June 2010^a	Minor
70008	7 February 2011	Patient's final T48 visit fell outside of the visit window allocated for the closure of the trial because patient failed to attend. Patient attended on 7 February 2011, which is a protocol deviation because of trial closure and shortened follow-up dates; however, 7 February 2011 is exactly 48 weeks from randomisation for this patient	Minor
70011	15 June 2010	Patient's T0 visit fell 3 days outside of the visit window because parents were unable to attend initial appointment^a	Minor
12008	14 July 2010	Patient's T0 visit fell 9 days outside of the visit window; however, the trial had already been discontinued and so this patient could not be randomised^a	Minor
12002	8 February 2011	Patient's final T48 visit fell outside of the visit window allocated for the closure of the trial because of illness. Parent then attended the final visit without the child. This meant that spirometry, physical assessments and child's quality of life assessment could not be completed (research nurse mailed them to parents but did not receive back)	Minor
13007	2 February 2010	Different caregiver completed the quality of life questionnaire at the T24 visit from the T8 visit (mother at T24, father at T8). Father was not present at T24	Minor
10010	15 January 2010	Pre-reversibility test not saved; therefore, initial spirometry data are not available for T0. Also, at T48, long-acting and short-acting bronchodilators had not been withheld for the appropriate time	Minor
13006	15 June 2010	Long-acting bronchodilators had not been withheld for the appropriate time at the T48 spirometry assessment – patient had taken study medication that morning	Minor
13010	21 December 2009	At T0 short-acting bronchodilator taken approximately 2.5 hours prior to spirometry testing	Minor
13014	25 May 2010	At T0 short-acting bronchodilator not withheld for 4 hours prior to spirometry testing	Minor
13015	25 May 2010	At T0 short-acting bronchodilator not withheld for 4 hours prior to spirometry testing	Minor
60001	19 May 2009	Required salbutamol 2 hours before spirometry testing at T0	Minor

a: MASCOT protocol specifes that ‘The next study visit (T0) will be organised with the participant and their carer/s to be in no less than 24 days time and no longer than 30 days’.

TABLE 6

Exacerbations requiring course of oral steroids over 48 weeks (ITT analysis)

	Fluticasone (n = 11)	Fluticasone plus salmeterol (n = 15)	Fluticasone plus montelukast (n = 12)	Total (n = 38)
Children with at least one exacerbation, n (%)	1 (9.1)	5 (33.3)	1 (8.3)	7 (18.4)
Total no. of exacerbations over 48 weeks,^a n (%)
0	10 (90.9)	10 (66.7)	11 (91.7)	31 (81.6)
1	0 (0.0)	4 (26.7)	0 (0.0)	4 (10.5)
2	0 (0.0)	1 (6.7)	0 (0.0)	1 (2.6)
3	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
4	1 (9.1)	0 (0.0)	1 (8.3)	2 (5.3)
Exposure (weeks)
Total	528.4	721.3	580.1	1829.8
Mean (SD)	48.03 (1.06)	48.09 (1.96)	48.35 (2.35)	48.15 (1.85)
Range	46.14 to 50.00	43.00 to 50.00	43.00 to 53.00	43.00 to 53.00

a: Patients included if exposure time is within the range of 43–53 weeks.

TABLE 7

Exacerbations requiring course of oral steroids over 24 weeks (ITT analysis)

	Fluticasone (n = 18)	Fluticasone plus salmeterol (n = 7)	Fluticasone plus montelukast (n = 19)	Total (n = 54)
Children with at least one exacerbation, n (%)	4 (22.2)	3 (17.6)	3 (15.8)	10 (18.5)
Total no. of exacerbations over 24 weeks,^a n (%)
0	14 (77.8)	14 (82.4)	16 (84.2)	44 (81.5)
1	1 (5.6)	2 (11.8)	3 (15.8)	6 (11.1)
2	2 (11.1)	1 (5.9)	0 (0.0)	3 (5.6)
3	1 (5.6)	0 (0.0)	0 (0.0)	1 (1.9)
Exposure (weeks)
Total	436.3	421.3	464.6	1322.2
Mean (SD)	24.24 (1.03)	24.78 (1.43)	24.45 (1.28)	24.48 (1.25)
Range	22.00 to 26.00	22.29 to 29.00	22.00 to 28.14	22.00 to 29.00

a: Patients included if exposure time is within the range of 19–29 weeks.

TABLE 8

Relative treatment effects from Poisson regression model adjusted for overdispersion: number of exacerbations requiring oral corticosteroids

Analysis set	Number of patients	Time point	Scale parameter^a	Fluticasone vs fluticasone plus salmeterol			Fluticasone vs fluticasone plus montelukast			Fluticasone plus salmeterol vs futicasone plus montelukast
Analysis set	Number of patients	Time point	Scale parameter^a	RR	CI	p-value^b	RR	CI	p-value^b	RR	CI	p-value^b
ITT	n_A = 11, n_B = 15, n_C = 12	48 weeks	1.68	0.91	0.07 to 12.05^c	0.93	1.10	0.06 to 18.6^c	0.94	1.21	0.09 to 15.97^c	0.86
	n_A = 11, n_B = 15, n_C = 12	48 weeks	1.68	0.91	0.11 to 7.59^d	0.93	1.10	0.11 to 11.21^d	0.94	1.21	0.14 to 10.07^d	0.86
	n_A = 18, n_B = 17, n_C = 19	24 weeks	1.17	1.93	0.35 to 10.67^c	0.36	2.84	0.43 to 18.79^c	0.19	1.47	0.17 to 12.39^c	0.67
	n_A = 18, n_B = 17, n_C = 19	24 weeks	1.17	1.93	0.47 to 7.86^d	0.36	2.84	0.60 to 13.40^d	0.19	1.47	0.26 to 8.47^d	0.67
Per protocol	n_A = 9, >n_B = 11, n_C = 8	48 weeks	1.27	1.64	0.16 to 16.54^c	0.61	–^e	–^e	–^e	–^e	–^e	–^e
	n_A = 9, >n_B = 11, n_C = 8	48 weeks	1.27	1.64	0.25 to 10.94^d	0.61	–^e	–^e	–^e	–^e	–^e	–^e
	n_A = 15, N_B = 13, >n_C = 16	24 weeks	1.18	2.36	0.35 to 15.93^c	0.28	2.87	0.42 to 19.35^c	0.19	1.22	0.12 to 12.15^c	0.84
	n_A = 15, N_B = 13, >n_C = 16	24 weeks	1.18	2.36	0.49 to 11.32^d	0.28	2.87	0.60 to 13.76^d	0.19	1.22	0.18 to 8.05^d	0.84

: A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

a: The scale parameter was estimated as the square root of Pearson's chi-squared statistic/df.

b: p-value = Wald chi-square p-value.

c: 98.3% CI.

d: 95% CI.

e: Unable to calculate RRs for futicasone vs futicasone plus montelukast and futicasone plus salmeterol vs futicasone plus montelukast in the 48-week per-protocol analysis as none of the patients in the futicasone plus montelukast group had a course of oral corticosteroids prescribed throughout the duration of the trial.

: Note: Both the 98.3% and 95% CIs have been presented for completeness.

TABLE 9

Quality of life of children measured using the PAQLQ(S) over 48 weeks

Domain	Fluticasone			Fluticasone plus salmeterol			Fluticasone plus montelukast			Fluticasone vs futicasone plus salmeterol	Fluticasone vs futicasone plus montelukast	Fluticasone plus salmeterol vs futicasone plus montelukast
Domain	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value
Activity limitations (n_A = 10, n_B = 14,^an_C = 12)	4.76 (0.91), 3.60 to 6.00	6.20 (1.07), 3.40 to 7.00	1.44 (1.00), −0.20 to 3.40	4.39 (1.30), 2.60 to 6.60	5.60 (1.47), 3.00 to 7.00	1.21 (2.12), −2.80 to 4.40	4.25 (1.47), 1.80 to 6.80	6.30 (0.85), 4.40 to 7.00	2.05 (1.61), −0.60 to 4.20	0.59 (−0.43 to 1.61), p = 0.25	−0.12 (−1.18 to 0.94), p = 0.82	−0.70 (−1.67 to 0.26), p = 0.16
Symptoms (n_A = 10, n_B = 15, n_C = 12)	4.90 (0.92), 3.70 to 6.30	6.28 (0.84), 4.40 to 7.00	1.38 (0.93), 0.00 to 3.10	4.35 (1.51), 1.30 to 6.80	5.30 (1.77), 1.50 to 7.00	0.95 (2.17), −2.20 to 5.70	4.32 (1.63), 1.50 to 6.70	6.23 (0.92), 4.70 to 7.00	1.91 (1.79), −0.80 to 4.40	0.90 (−0.22 to 2.02), p = 0.11	−0.03 (−1.20 to 1.15), p = 0.96	−0.93 (−1.98 to 0.12), p = 0.08
Emotional function (n_A = 10, n_B = 15, n_C = 12)	5.31 (0.82), 3.50 to 6.25	6.40 (0.86), 4.25 to 7.00	1.09 (0.62), 0.13 to 2.13	4.33 (1.68), 1.75 to 7.00	5.58 (1.62), 2.13 to 7.00	1.25 (1.72), −1.50 to 4.38	4.79 (1.54), 2.00 to 6.75	6.42 (0.83), 4.88 to 7.00	1.63 (1.66), −0.63 to 4.13	0.50 (−0.48 to 1.49), p = 0.31	−0.19 (−1.19 to 0.82), p = 0.71	−0.69 (−1.60 to 0.22), p = 0.13
Total (n_A = 10, n_B = 15, n_C = 12)	5.01 (0.64), 3.91 to 5.78	6.30 (0.88), 4.13 to 7.00	1.29 (0.61), 0.22 to 2.52	4.34 (1.39), 2.04 to 6.83	5.44 (1.56), 2.26 to 7.00	1.10 (1.86), −2.09 to 4.96	4.47 (1.50), 1.74 to 6.65	6.31 (0.85), 4.78 to 7.00	1.84 (1.66), −0.70 to 4.26	0.73 (−0.29 to 1.74), p = 0.15	−0.12 (−1.17 to 0.94), p = 0.82	−0.84 (−1.78 to 0.10), p = 0.08

: A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

a: One patient with missing question on the activity limitations domain was excluded from the domain score calculation but included in the total score as the missing data made up < 10% of the total.

: Scores for PAQLQ(S) range from 1 to 7; lower scores are worse.

TABLE 10

Quality of life of children measured using the PAQLQ(S) over 24 weeks

Domain	Fluticasone			Fluticasone plus salmeterol			Fluticasone plus montelukast			Fluticasone vs futicasone plus salmeterol	Fluticasone vs futicasone plus montelukast	Fluticasone plus salmeterol vs futicasone plus montelukast
Domain	Baseline mean (SD), range	T24 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T24 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T24 mean (SD), range	Change mean (SD), range	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value
Activity limitations (n_A = 17, n_B = 15,^an_C = 17)	4.52 (1.18), 2.20 to 6.00	5.60 (1.50), 2.20 to 7.00	1.08 (1.02), −0.40 to 3.60	4.40 (1.26), 2.60 to 6.60	6.00 (0.91), 4.40 to 7.00	1.60 (1.39), −0.80 to 4.20	4.21 (1.32), 1.80 to 6.80	5.75 (1.48), 2.40 to 7.00	1.54 (1.33), −0.60 to 3.80	−0.47 (−1.28 to 0.35), p = 0.26	−0.33 (−1.12 to 0.47), p = 0.41	0.14 (−0.68 to 0.96), p = 0.73
Symptoms (n_A = 17, n_B = 16, n_C = 17)	4.64 (1.15), 2.60 to 6.60	5.49 (1.31), 2.70 to 7.00	0.85 (0.97), −1.00 to 2.90	4.45 (1.51), 1.30 to 6.80	5.91 (1.18), 2.70 to 7.00	1.46 (1.42), −1.10 to 3.60	4.24 (1.42), 1.50 to 6.70	5.46 (1.67), 1.50 to 7.00	1.22 (1.42), −1.40 to 3.20	−0.52 (−1.34 to 0.29), p = 0.20	−0.20 (−1.01 to 0.61), p = 0.63	0.33 (−0.49 to 1.14), p = 0.43
Emotional function (n_A = 17, n_B = 16, n_C = 17)	5.21 (1.13), 3.38 to 6.88	5.86 (1.38), 1.50 to 7.00	0.65 (1.11), −1.88 to 2.63	4.45 (1.70), 1.75 to 7.00	6.22 (1.04), 3.75 to 7.00	1.77 (1.65), 0 to 4.88	4.79 (1.32), 2.00 to 6.75	5.93 (1.66), 1.38 to 7.00	1.14 (1.41), −1.50 to 3.63	−0.73 (−1.61 to 0.15), p = 0.10	−0.28 (−1.13 to 0.57), p = 0.51	0.45 (−0.41 to 1.31), p = 0.30
Total (n_A = 17, n_B = 16, n_C = 17)	4.81 (1.01), 2.96 to 6.52	5.64 (1.29), 2.17 to 7.00	0.83 (0.90), −1.09 to 2.96	4.43 (1.39), 2.04 to 6.83	6.05 (1.01), 3.43 to 7.00	1.62 (1.38), −0.52 to 3.74	4.43 (1.29), 1.74 to 6.65	5.69 (1.59), 1.65 to 7.00	1.26 (1.30), −0.83 to 3.13	−0.63 (−1.42 to 0.15), p = 0.11	−0.27 (−1.05 to 0.50), p = 0.48	0.36 (−0.42 to 1.14), p = 0.36

: A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

a: One patient with missing question on the activity limitations domain was excluded from the domain score calculation but included in the total score as the missing data made up < 10% of the total.

: Scores for PAQLQ(S) range from 1 to 7; lower scores are worse.

TABLE 11

Quality of life of caregivers measured using the PACQLQ over 48 weeks

Domain	Fluticasone			Fluticasone plus salmeterol			Fluticasone plus montelukast			Fluticasone vs futicasone plus salmeterol	Fluticasone vs futicasone plus montelukast	Fluticasone plus salmeterol vs futicasone plus montelukast
Domain	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value
Activity limitations (n_A = 11, n_B = 15, n_C = 12)	4.75 (1.34), 2.50 to 7.00	6.07 (1.16), 3.50 to 7.00	1.32 (1.41), −1.25 to 3.50	5.08 (1.43), 1.25 to 6.75	6.30 (0.99), 4.25 to 7.00	1.22 (1.30), −1.00 to 3.75	3.85 (1.53), 1.75 to 6.50	5.81 (1.41), 3.50 to 7.00	1.96 (1.21), 0.50 to 4.75	−0.09 (−0.92 to 0.75), p = 0.83	−0.13 (−1.03 to 0.77), p = 0.77	−0.04 (−0.90 to 0.83), p = 0.93
Emotional function (n_A = 11, n_B = 15, n_C = 12)	4.44 (1.09), 2.22 to 6.22	5.80 (1.56), 1.67 to 7.00	1.36 (1.05), −0.56 to 3.00	4.35 (1.35), 1.00 to 6.00	5.73 (1.27), 2.33 to 7.00	1.38 (1.45), −0.56 to 4.44	4.05 (1.49), 2.00 to 6.00	5.67 (1.27), 3.78 to 7.00	1.62 (1.48), −1.11 to 4.11	0.02 (−0.95 to 0.99), p = 0.97	−0.07 (−1.10 to 0.96), p = 0.89	−0.08 (−1.04 to 0.87), p = 0.86
Total (n_A = 11, n_B = 15, n_C = 12)	4.54 (1.08), 2.46 to 6.46	5.88 (1.42), 2.23 to 7.00	1.34 (1.07), −0.23 to 3.15	4.57 (1.28), 1.08 to 6.15	5.91 (1.13), 3.08 to 7.00	1.34 (1.25), −0.38 to 4.00	3.99 (1.43), 2.00 to 6.15	5.71 (1.25), 3.77 to 7.00	1.72 (1.27), −0.38 to 4.08	−0.01 (−0.87 to 0.85), p = 0.99	−0.12 (−1.04 to 0.80), p = 0.79	−0.12 (−0.97 to 0.74), p = 0.78

: A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

: Scores for PAQLQ(S) range from 1 to 7; lower scores are worse.

TABLE 12

Quality of life of caregivers measured using the PACQLQ over 24 weeks

Domain	Fluticasone			Fluticasone plus salmeterol			Fluticasone plus montelukast			Fluticasone vs fluticasone plus salmeterol	Fluticasone vs fluticasone plus montelukast	Fluticasone plus salmeterol vs fluticasone plus montelukast
Domain	Baseline mean (SD), range	T24 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T24 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T24 mean (SD), range	Change mean (SD), range	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value
Activity limitations (n_A = 17, n_B = 14, n_C = 16)	4.49 (1.49), 2.00 to 7.00	6.00 (1.31), 3.25 to 7.00	1.51 (1.29), −1.25 to 3.50	5.36 (0.99), 3.25 to 6.75	6.04 (1.24), 3.25 to 7.00	0.68 (1.33), −2.00 to 3.25	4.09 (1.64), 1.75 to 7.00	5.92 (1.07), 4.00 to 7.00	1.83 (1.55), −2.25 to 4.25	0.30 (−0.53 to 1.12), p = 0.47	−0.07 (−0.85 to 0.70), p = 0.85	−0.37 (−1.23 to 0.49), p = 0.39
Emotional function (n_A = 17, n_B = 16, n_C = 16)	4.22 (1.36), 1.67 to 6.78	5.66 (1.42), 2.33 to 7.00	1.44 (1.27), −0.56 to 4.44	4.35 (1.31), 1.00 to 6.00	5.42 (1.82), 1.78 to 7.00	1.07 (1.22), −1.78 to 2.89	4.32 (1.43), 2.00 to 6.33	5.63 (1.18), 3.00 to 7.00	1.31 (1.57), −3.33 to 4.11	0.32 (−0.57 to 1.20), p = 0.47	0.09 (−0.79 to 0.98), p = 0.83	−0.23 (−1.12 to 0.67), p = 0.62
Total (n_A = 17, n_B = 16, n_C = 16)	4.30 (1.33), 2.00 to 6.85	5.76 (1.36), 2.62 to 7.00	1.46 (1.18), −0.77 to 4.15	4.54 (1.24), 1.08 to 6.15	5.57 (1.66), 2.28 to 7.00	1.03 (1.21), −1.85 to 2.85	4.25 (1.41), 2.00 to 6.54	5.72 (1.08), 3.54 to 7.00	1.47 (1.51), −3.00 to 4.08	0.32 (−0.51 to 1.15), p = 0.44	0.02 (−0.81 to 0.85), p = 0.96	−0.30 (−1.15 to 0.54), p = 0.47

: A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

: Scores for PAQLQ(S) range from 1 to 7; lower scores are worse.

TABLE 13

Time to first exacerbation

Treatment	No. of patients	No. of events (%)	Median exacerbation-free time	24-week exacerbation-free probability (95% CI)
Fluticasone	19	4 (21.1)	Not estimable	0.79 (0.52 to 0.91)
Fluticasone plus salmeterol	23	7 (30.4)		0.80 (0.55 to 0.92)
Fluticasone plus montelukast	21	3 (14.3)		0.83 (0.57 to 0.94)
Total	63	14 (22.2)

FIGURE 2

Kaplan–Meier survival curves for time to first exacerbation by treatment.

TABLE 14

Hazard ratios for time to first exacerbation

Comparison	HR (95% CI)
Fluticasone vs fluticasone plus salmeterol	0.63 (0.19 to 2.08)
Fluticasone vs fluticasone plus montelukast	1.52 (0.34 to 6.70)
Fluticasone plus salmeterol vs fluticasone plus montelukast	2.37 (0.68 to 8.20)

TABLE 15

Number of school days missed over 48 weeks (ITT analysis)

	Fluticasone (n = 11)	Fluticasone plus salmeterol (n = 15)	Fluticasone plus montelukast (n = 11)	Total (n = 37)
Children with at least 1 school day missed, n (%)	7 (63.6)	9 (60.0)	2 (18.2)	18 (48.6)
Total no. of school days missed over 48 weeks,^a n (%)
0	4 (36.4)	6 (40.0)	9 (81.8)	19 (51.4)
1	2 (18.2)	1 (6.7)	0 (0.0)	3 (8.1)
2	1 (9.1)	3 (20.0)	0 (0.0)	4 (10.8)
3	1 (9.1)	1 (6.7)	0 (0.0)	2 (5.4)
4	1 (9.1)	0 (0.0)	0 (0.0)	1 (2.7)
5	1 (9.1)	0 (0)	1 (9.1)	2 (5.4)
6	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
7	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
8	1 (9.1)	0 (0.0)	0 (0.0)	1 (2.7)
9	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
10	0 (0.0)	3 (20.0)	0 (0.0)	3 (8.1)
11	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
12	0 (0.0)	0 (0.0)	1 (9.1)	1 (2.7)
13	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
14	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
15	0 (0.0)	1 (6.7)	0 (0.0)	1 (2.7)
Median (IQR)	1 (0 to 4)	2 (0 to 10)	0 (0 to 0)	0 (0 to 3)
Range	0 to 8	0 to 15	0 to 12	0 to 15
Total exposure time (weeks)
Mean (SD)	48.04 (1.06)	48.09 (1.96)	48.38 (2.46)	48.16 (1.87)
Range	46.14 to 50.00	43.00 to 50.00	43.00 to 53.00	43.00 to 53.00

a: Patients included if total exposure time is within the range of 43–53 weeks.

TABLE 16

Number of school days missed over 24 weeks (ITT analysis)

	Fluticasone (n = 18)	Fluticasone plus salmeterol (n = 17)	Fluticasone plus montelukast (n = 19)	Total (n = 54)
Children with at least 1 school day missed, n (%)	6 (33.3)	8 (47.1)	6 (31.6)	20 (37.0)
Total no. of school days missed over 24 weeks,^a n (%)
0	12 (66.7)	9 (52.9)	13 (68.4)	34 (63.0)
1	0 (0.0)	3 (17.6)	0 (0.0)	3 (5.6)
2	2 (11.1)	2 (11.8)	3 (15.8)	7 (13.0)
3	2 (11.1)	2 (11.8)	0 (0.0)	4 (7.4)
4	1 (5.6)	0 (0.0)	0 (0.0)	1 (1.9)
5	0 (0.0)	1 (5.9)	1 (5.3)	2 (3.7)
6	1 (5.6)	0 (0.0)	0 (0.0)	1 (1.9)
7	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
8	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
9	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
10	0 (0.0)	0 (0.0)	2 (10.5)	2 (3.7)
Median (IQR)	0 (0 to 2)	0 (0 to 2)	0 (0 to 2)	0 (0 to 2)
Range	0 to 6	0 to 5	0 to 10	0 to 10
Total exposure time (weeks)
Mean (SD)	24.24 (1.03)	24.78 (1.43)	24.45 (1.27)	24.48 (1.25)
Range	22.00 to 26.00	22.29 to 29.00	22.00 to 28.14	22.00 to 29.00

a: Patients included if total exposure time is within the range of 19–29 weeks.

TABLE 17

Relative treatment effects from Poisson regression model adjusted for overdispersion: number of school days missed

Analysis set	No. of patients	Time point (weeks)	Scale parameter^a	Fluticasone vs fluticasone plus salmeterol			Fluticasone vs fluticasone plus montelukast			Fluticasone plus salmeterol vs futicasone plus montelukast
Analysis set	No. of patients	Time point (weeks)	Scale parameter^a	RR	95% CI	p-value^b	RR	95% CI	p-value^b	RR	95% CI	p-value^b
ITT	n_A = 11, n_B = 15, n_C = 11	48	2.58	0.60	0.17 to 2.05	0.41	1.42	0.29 to 7.07	0.67	2.39	0.59 to 9.71	0.22
ITT	n_A = 18, n_B = 17, n_C = 19	24	2.00	1.07	0.30 to 3.84	0.91	0.69	0.22 to 2.12	0.51	0.64	0.20 to 2.05	0.45
Per protocol	n_A = 9, n_B = 11, n_C = 8	48	2.72	0.69	0.14 to 3.33	0.65	0.93	0.15 to 5.66	0.94	1.35	0.27 to 6.68	0.72
Per protocol	n_A = 15, n_B = 13, n_C = 16	24	2.00	1.97	0.41 to 9.46	0.40	0.74	0.24 to 2.31	0.61	0.38	0.08 to 1.68	0.20

: A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

a: The scale parameter was estimated as the square root of Pearson's chi-squared statistic/df.

b: p-value = Wald chi-square p-value.

TABLE 18

Number of hospital admissions over 48 weeks (ITT analysis)

	Fluticasone (n = 11)	Fluticasone plus salmeterol (n = 15)	Fluticasone plus montelukast (n = 11)	Total (n = 37)
Children with at least one hospital admission, n (%)	0 (0.0)	2 (13.3)	0 (0.0)	2 (5.4)
Total no. of hospital admissions over 48 weeks,^a n (%)
0	11 (100.0)	13 (86.7)	11 (100.0)	35 (94.6)
1	0 (0.0)	2 (13.3)	0 (0.0)	2 (5.4)
Total exposure time (weeks)
Mean (SD)	48.04 (1.06)	8.09 (1.96)	48.38 (2.46)	48.16 (1.87)
Range	46.14 to 50.00	43.00 to 50.00	43.00 to 53.00	43.00 to 53.00

a: Patients included if total exposure time is within the range of 43–53 weeks.

TABLE 19

Number of hospital admissions over 24 weeks (ITT analysis)

	Fluticasone (n = 18)	Fluticasone plus salmeterol (n = 17)	Fluticasone plus montelukast (n = 19)	Total (n = 54)
Children with at least one school day missed, n (%)	2 (11.1)	1 (5.9)	1 (5.3)	4 (7.4)
Total no. of hospital admissions over 48 weeks,^a n (%)
0	16 (88.9)	16 (94.1)	18 (94.7)	50 (92.6)
1	1 (5.6)	1 (5.9)	1 (5.3)	3 (5.6)
2	1 (5.6)	0 (0.0)	0 (0.0)	1 (1.9)
Total exposure time (weeks)
Mean (SD)	24.24 (1.03)	24.78 (1.43)	24.45 (1.27)	24.48 (1.25)
Range	22.00 to 26.00	22.29 to 29.00	22.00 to 28.14	22.00 to 29.00

a: Patients included if total exposure time is within the range of 19 and 29 weeks.

TABLE 20

Relative treatment effects from Poisson regression model adjusted for overdispersion: number of hospital admissions

Analysis set	No. of patients	Time point	Scale parameter^a	Fluticasone vs fluticasone plus salmeterol			Fluticasone vs fluticasone plus montelukast			Fluticasone plus salmeterol vs fluticasone plus montelukast
Analysis set	No. of patients	Time point	Scale parameter^a	RR	95% CI	p-value^b	RR	95% CI	p-value^b	RR	95% CI	p-value^b
ITT	n_A = 11, n_B = 15, n_C = 11	48 weeks^c	–	–	–	–	–	–	–	–	–	–
ITT	n_A = 18, n_B = 17, n_C = 19	24 weeks	1.08	2.90	0.25 to 33.11	0.39	3.19	0.28 to 36.51	0.35	1.10	0.06 to 21.79	0.95
Per protocol	n_A = 9, n_B = 11, n_C = 8	48 weeks^c	–	–	–	–	–	–	–	–	–	–
Per protocol	n_A = 15, n_B = 13, n_C = 16	24 weeks	1.14	1.77	0.12 to 27.21	0.68	2.15	0.14 to 33.06	0.58	1.22	0.05 to 28.50	0.90

: A, fluticasone; B, fluticasone plus salmeterol; C, fluticasone plus montelukast.

a: The scale parameter was estimated as the square root of Pearson's chi-squared statistic/df.

b: p-value = Wald chi-square p-value.

c: Unable to calculate RRs in the 48-week ITT and per-protocol analyses as there were only two patients (from the fluticasone plus salmeterol group) who had any hospital admissions throughout the duration of the trial.

TABLE 21

Number of beta-2 agonists prescribed over 48 weeks

	Fluticasone (n = 11)	Fluticasone plus salmeterol (n = 14^a)	Fluticasone plus montelukast (n = 12)	Total (n = 37)
Children with at least one beta-2 agonist prescribed, n (%)	6 (54.5)	10 (71.4)	7 (58.3)	23 (62.2)
Total no. of beta-2 agonists prescribed over 48 weeks,^a n (%)
0	5 (45.5)	4 (28.6)	5 (41.7)	14 (37.8)
1	0 (0.0)	3 (21.4)	1 (8.3)	4 (10.8)
2	0 (0.0)	1 (7.1)	1 (8.3)	2 (5.4)
3	2 (18.2)	1 (7.1)	3 (25.0)	6 (16.2)
4	2 (18.2)	0 (0.0)	0 (0.0)	2 (5.4)
5	1 (9.1)	1 (7.1)	0 (0.0)	2 (5.4)
6	0 (0.0)	2 (14.3)	1 (8.3)	3 (8.1)
7	1 (9.1)	0 (0.0)	1 (8.3)	2 (5.4)
8	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
9	0 (0.0)	1 (7.1)	0 (0.0)	1 (2.7)
40	0 (0.0)	1 (7.1)	0 (0.0)	1 (2.7)
Median (IQR)	3 (0 to 4)	1.5 (0 to 6)	1.5 (0 to 3)	2 (0 to 4)
Range	0 to 7	0 to 40	0 to 7	0 to 40
Amount of beta-2 agonists prescribed over 48 weeks^a (µg)
Median (IQR)	36,000 (0 to 80,000)	22,000 (0 to 120,000)	22,000 (0 to 60,000)	24,000 (0 to 80,000)
Range	0 to 140,000	0 to 800,000	0 to 140,000	0 to 800,000

a: For one patient included in the ITT analysis set, the number of beta-2 agonists prescribed between the T8 and T24 visits was missing.

: Note that one patient on fluticasone plus salmeterol had 28 beta-2 agonists prescribed between T0 and T8.

TABLE 22

Number of beta-2 agonists prescribed over 24 weeks

	Fluticasone (n = 18)	Fluticasone plus salmeterol (n = 16^a)	Fluticasone plus montelukast (n = 19)	Total (n = 53)
Children with at least one beta-2 agonist prescribed, n (%)	9 (50.0)	10 (62.5)	9 (47.4)	28 (52.8)
Total no. of beta-2 agonists prescribed over 24 weeks,^a n (%)
0	9 (50.0)	6 (37.5)	10 (52.6)	25 (47.2)
1	2 (11.1)	4 (25.0)	4 (21.1)	10 (18.9)
2	4 (22.2)	4 (25.0)	4 (21.1)	12 (22.6)
3	1 (5.6)	0 (0.0)	1 (5.3)	2 (3.8)
4	2 (11.1)	1 (6.3)	0 (0.0)	3 (5.7)
31	0 (0.0)	1 (6.3)	0 (0.0)	1 (1.9)
Median (IQR)	0.5 (0 to 2)	1 (0 to 2)	0 (0 to 2)	1 (0 to 2)
Range	0 to 4	0 to 31	0 to 3	0 to 31
Amount of beta-2 agonists prescribed over 24 weeks^a (µg)
Median (IQR)	6000 (0 to 40,000)	20,000 (0 to 40,000)	0 (0 to 24,000)	20,000 (0 to 40,000)
Range	0 to 80,000	0 to 620,000	0 to 60,000	0 to 620,000

a: One patient who is included in the ITT analysis set had missing number of beta-2 agonists prescribed between the T8 and T24 visits.

: Note that one patient on fluticasone plus salmeterol had 28 beta-2 agonists prescribed between T0 and T8.

TABLE 23

Difference in mean amount of beta-2 agonists prescribed (ANOVA)

Comparison of means	Fluticasone vs fluticasone plus salmeterol, mean difference (µg) (95% CI^a)	Fluticasone vs fluticasone plus montelukast, mean difference (µg) (95% CI^a)	Fluticasone plus salmeterol vs fluticasone plus montelukast, mean difference (µg) (95% CI^a)
48 weeks	−59,481 (−195,859 to 76,897)	6758 (−134,532 to 148,048)	66,238 (−66,920 to 199,396)
24 weeks	−34,861 (−107,549 to 37,827)	7942 (−61,642 to 77,525)	42,803 (−28,980 to 114,585)

a: Used Bonferroni's method to calculate 95% CIs as there were an unequal numbers of cases in each group.

TABLE 24

Lung function over 48 weeks assessed by spirome

	Fluticasone			Fluticasone plus salmeterol			Fluticasone plus montelukast			Fluticasone vs futicasone plus salmeterol	Fluticasone vs futicasone plus montelukast	Fluticasone plus salmeterol vs fluticasone plus montelukast
	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Baseline mean (SD), range	T48 mean (SD), range	Change mean (SD), range	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value	Adjusted mean difference (95% CI), p-value
FEV₁% predicted (n_A = 8, n_B = 13, n_C = 9)	86.13 (22.56), 48.00 to 116.00	86.25 (17.77), 49.00 to 106.00	0.12 (12.74), −16.00 to 19.00	74.46 (19.70), 47.00 to 103.00	90.00 (13.91), 64.00 to 107.00	15.54 (19.77), −17.00 to 56.00	84.56 (16.94), 56.00 to 104.00	89.11 (13.63), 73.00 to 120.00	4.55 (15.83), −14.00 to 30.00	−8.58 (−20.71 to 3.56), p = 0.16	−3.51 (−16.24 to 9.22), p = 0.58	5.07 (−6.57 to 16.71), p = 0.38
FVC% predicted (n_A = 8, n_B = 13, n_C = 9)	87.50 (19.93), 57.00 to 112.00	92.00 (12.22), 74.00 to 111.00	4.50 (10.88), −7.00 to 23.00	84.08 (18.29), 49.00 to 111.00	95.15 (16.35), 67.00 to 128.00	11.07 (24.59), −12.00 to 79.00	91.89 (20.19), 53.00 to 128.00	91.67 (13.52), 68.00 to 113.00	−0.22 (14.59), −15.00 to 28.00	−4.17 (−16.79 to 8.45), p = 0.50	1.64 (−12.03 to 15.30), p = 0.81	5.81 (−6.53 to 18.15), p = 0.34

: A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

TABLE 25

All adverse events^a^,^b

Adverse event	Fluticasone			Fluticasone plus salmeterol			Fluticasone plus montelukast
Adverse event	Patients (n = 19)	%	Events	Patients (n = 23)	%	Events	Patients (n = 21)	%	Events
Ear and labyrinth disorders				1	4.3	1
Eye disorders							1	4.8	1
Gastrointestinal disorders	5	26.3	6	5	21.7	7	5	23.8	9
General disorders/administration site conditions	4	21.1	5	6	26.1	10	5	23.8	6
Immune system disorders	2	10.5	2	3	13.0	4	1	4.8	2
Infections and infestations	7	36.8	11	9	39.1	13	7	33.3	13
Injury, poisoning and procedural complications	3	15.8	3	1	4.3	1	3	14.3	3
Musculoskeletal and connective tissue disorders	1	5.3	1				2	9.5	3
Nervous system disorders	5	26.3	13	1	4.3	2	7	33.3	37
Psychiatric disorders				2	8.7	2
Reproductive system and breast disorders				1	4.3	2
Respiratory, thoracic and mediastinal disorders	13	68.4	28	15	65.2	31	13	61.9	28
Skin and subcutaneous tissue disorders	2	10.5	2	3	13.0	3	2	9.5	5
Total	42		71	47		76	46		107

a: Classified according to MedRA System Organ Class.

b: Five events with a missing onset date have been included in the table.

TABLE 26

All adverse events according to severity^a^b

Adverse event	Severity	No. of events				No. of patients
Adverse event	Severity	Fluticasone	Fluticasone plus salmeterol	Fluticasone plus montelukast	Total	Fluticasone (%)	Fluticasone plus salmeterol (%)	Fluticasone plus montelukast (%)	Total (%)
Ear and labyrinth disorders	Mild		1		1		1 (4.3)		1 (1.6)
Ear and labyrinth disorders	Total		1		1		1 (4.3)		1 (1.6)
Eye disorders	Mild			1	1			1 (4.8)	1 (1.6)
Eye disorders	Total			1	1			1 (4.8)	1 (1.6)
Gastrointestinal disorders	Mild	6	7	9	22	5 (26.3)	5 (21.7)	5 (23.8)	15 (23.8)
Gastrointestinal disorders	Total	6	7	9^c	22	5 (26.3)	5 (21.7)	5^c (23.8)	15 (23.8)
General disorders/administration site conditions	Mild	5	10	6	21	4 (21.1)	6 (26.1)	5 (23.8)	15 (23.8)
General disorders/administration site conditions	Total	5	10	6	21	4 (21.1)	6 (26.1)	5 (23.8)	15 (23.8)
Immune system disorders	Mild		4	2	6		3 (13.0)	1 (4.8)	4 (6.3)
	Moderate	2			2	2 (10.5)			2 (3.2)
	Total	2	4	2	8	2 (10.5)	3 (13.0)	1 (4.8)	6 (9.5)
Infections and infestations	Mild	9	6	10	25	6 (31.6)	5 (21.7)	5 (23.8)	16 (25.4)
	Moderate	2	7	3	12	1 (5.3)	4 (17.4)	2 (9.5)	7 (11.1)
	Total	11	13	13	37	7 (36.8)	9 (39.1)	7 (33.3)	23 (36.5)
Injury, poisoning and procedural complications	Mild	3		3	6	3 (15.8)		3 (14.3)	6 (9.5)
	Moderate		1		1		1 (4.3)		1 (1.6)
	Total	3	1	3	7	3 (15.8)	1 (4.3)	3 (14.3)	7 (11.1)
Musculoskeletal and connective tissue	Mild	1		2	3	1 (5.3)		1 (4.8)	2 (3.2)
	Moderate			1	1			1 (4.8)	1 (1.6)
	Total	1		3	4	1 (5.3)		2 (9.5)	3 (4.8)
Nervous system disorders	Mild	13	2	16	31	5 (26.3)	1 (4.3)	5 (23.8)	11 (11.1)
	Moderate			21	21			2 (9.5)	2 (3.2)
	Total	13	2	37	52	5 (26.3)	1 (4.3)	7 (33.3)	13 (20.6)
Psychiatric disorders	Mild		1		1		1 (4.3)		1 (1.6)
	Moderate		1		1		1 (4.3)		1 (1.6)
	Total		2		2		2 (8.7)		2 (3.2)
Reproductive system and breast disorders	Mild		2		2		1 (4.3)		1 (1.6)
Reproductive system and breast disorders	Total		2		2		1 (4.3)		1 (1.6)
Respiratory, thoracic and mediastinal disorders	Mild	24	23	21	68	11 (57.9)	9 (39.1)	10 (47.6)	30 (47.6)
	Moderate	4	8	7	19	2 (10.5)	6 (26.1)	3 (14.3)	11 (11.1)
	Total	28	31^c	28^c	87	13 (68.4)	15^c (65.2)	13^c (61.9)	41 (65.1)
Skin and subcutaneous tissue disorders	Mild	2	2	3	7	2 (10.5)	2 (8.7)	1 (4.8)	5 (7.9)
	Moderate		1	2	3		1 (4.3)	1 (4.8)	2 (3.2)
	Total	2	3	5	10	2 (10.5)	3 (13.0)	2 (9.5)	7 (11.1)
Total		71	76	107	254	42	47	46	135

a: Classifed according to MedRA System Organ Class.

b: Five events with a missing onset date have been included in the table.

c: Events were ‘possibly related to treatment’: gastrointestinal disorders: one patient with one event (fluticasone plus montelukast); respiratory disorders: one patient with one event (fluticasone plus salmeterol) and one patient with four events (fluticasone plus montelukast).

TABLE 27

All SAEs from randomisation

Patient registration number	Treatment allocation	Time occurred	Description	MedRA SOC	Seriousness	Severity	Expectedness	Relationship	Cause	Outcome	Patient status	Unblinded
134005	Fluticasone	T0–T8	History of cough requiring regular salbutamol use (4 days out of the previous 7 days). Temperature up to 39°C. Low oral intake. Contact with H1N1. Infective exacerbation with O₂ requirement	Respiratory, thoracic and mediastinal disorders	Required hospitalisation	Moderate	Unexpected	Unrelated	Disease under study	Not resolved/ongoing	Continuing in trial	No
134005	Fluticasone	T8–T24	Exacerbation of asthma? Viral-induced wheeze	Infections and infestations	Required hospitalisation	Mild	Expected	Unrelated	Disease under study	Resolved	Continuing in trial	No
009010	Fluticasone	T0–T8	Patient admitted to children's emergency department following anaphylaxis? as a result of eating ice cream	Immune system disorders	Required hospitalisation	Moderate	Unexpected	Unrelated	Other illness	Resolved	Continuing in trial	No
009003	Fluticasone plus salmeterol	T36–T48	Patient complained of episodes of chest pain after swimming. Diagnosis of muscular sprain from swimming	General disorders/ administration site conditions	Required hospitalisation	Mild	Unexpected	Unrelated	Other illness	Resolved	Continuing in trial	No
036001	Fluticasone plus salmeterol	T24–T36	Been having ‘hayfever’ symptoms past few days. Started multidosing. Came to children's assessment unit 1230. Peak expiratory flow rate 100 l/minute	Respiratory, thoracic and mediastinal disorders	Required hospitalisation	Mild	Expected	Unrelated	Disease under study	Resolved	Continuing in trial	No
522010	Fluticasone plus salmeterol	T8–T24	Admitted as a result of asthma symptoms. Patient reported shortness of breath. Received one nebuliser (salbutamol) and one dose of prednisolone	Respiratory, thoracic and mediastinal disorders	Medically significant/important	Mild	Unexpected	Unrelated	Disease under study	Resolved	Continuing in trial	No
522011	Fluticasone plus montelukast	T8–T24	Diagnosis: exacerbation of asthma. Treatment: salbutamol inhalers and oral prednisolone	Respiratory, thoracic and mediastinal disorders	Required hospitalisation	Moderate	Expected	Unrelated	Disease under study	Resolved	Continuing in trial	No

: SOC, System Organ Class.

TABLE 28

Completion of the health economics booklet at each time period

Group	Baseline	T8	T24	T36	T48
Fluticasone	19	19	17	12	11
Fluticasone plus salmeterol	23	22	17	14	15
Fluticasone plus montelukast	21	19	19	13	12

TABLE 29

Total resource use (events of care)

Resource	Resource use
Resource	Fluticasone	Fluticasone plus salmeterol	Fluticasone plus montelukast
Intervention	68	73	72
GP visit	5	9	6
GP nurse visit		3
Walk-in doctor visit	3	2
GP other visit	2	2	3
Home doctor visit		1
Out-of-hours GP visit		1
A&E visit	0	5	2
Prescribed inhalers	6	12	8
Prescribed medicines	5	13	7
OTC medicines	0	2	2
‘Rescue’ medication (salbutamol)	11	28	6
Total	100	151	106

TABLE 30

Total resource costs

Resource	Total resource cost (£)			Total cost (£)
Resource	Fluticasone	Fluticasone plus salmeterol	Fluticasone plus montelukast	Total cost (£)
Intervention	1536	6300	6439	14,275
GP visit	140	252	168	560
GP nurse visit	0	30	0	30
Walk-in doctor visit	120	80	0	200
GP other visit	80	113.70	290.10	483.80
Home doctor visit	0	94	0	94
Out-of-hours GP visit	0	120	0	120
A&E visit	0	483.50	193.40	676.90
Prescribed inhalers	89	280.08	100.90	469.98
Prescribed medicines	18.20	159.54	22.46	200.20
OTC medicines	0	3.04	0.38	3.42
‘Rescue’ medication (salbutamol)	7.94	85.76	6.26	99.96
Total	1991.14	8001.62	7220.50	17,213.26
Average cost per person	104.80	347.90	343.83	273.23

TABLE 31

Quality-adjusted life-year gains

	Mean	SD
Incremental QALYs
Fluticasone	0.09	0.04
Fluticasone plus salmeterol	0.12	0.14
Fluticasone plus montelukast	0.13	0.12
Intervention + treatment costs
Fluticasone	144.75	82.13
Fluticasone plus salmeterol	458.80	95.24
Fluticasone plus montelukast	447.99	128.46

FIGURE 3

Quality-adjusted life-years over time in the trial. A, fluticasone; B, fluticasone plus salmeterol; C, fluticasone plus montelukast.

TABLE 32

Mean differences in costs and QALYs by group using bootstrapped data

Model	Difference in costs (£)			Difference in QALYs
Model	Mean	SD	p-value^a	Mean	SD	p-value^a
Fluticasone plus salmeterol vs futicasone	314.05	34.90	0.48	0.03	0.04	0.46
Fluticasone plus montelukast vs futicasone	303.24	44.59	0.35	0.04	0.04	0.43
Fluticasone plus montelukast vs futicasone plus salmeterol	−10.81	44.50	0.19	0.02	0.05	0.89

a: t-tests of the mean difference in QALYs between the two groups.

TABLE 33

Incremental cost-effectiveness ratios

	ICER (£/QALY)
Fluticasone plus salmeterol vs fluticasone	12,054
Fluticasone plus montelukast vs fluticasone	6827
Fluticasone plus montelukast vs fluticasone plus salmeterol	–588

FIGURE 4

Cost-effectiveness acceptability curves. A, futicasone; B, futicasone plus salmeterol; C, futicasone plus montelukast.

FIGURE 5

The cost-effectiveness plane.

TABLE 34

Withdrawals from the study and treatment

Time point of withdrawal	Reason	Fluticasone (n = 19), n (%)	Fluticasone plus salmeterol (n = 23), n (%)	Fluticasone plus montelukast (n = 21), n (%)	Total (n = 63), n (%)
T0–T8	Total	0 (0)	3 (13.0)	2 (9.5)	5 (7.9)
	Spiritual reasons	0	1	1	2
	Intercurrent illness	0	0	1	1
	Patient decision	0	2	0	2
T8–24	Total	2 (10.5)	3 (13.0)	2 (9.5)	7 (11.1)
	Lost to follow-up	1	2	0	3
	Study schedule too intrusive	0	1	0	1
	Non-compliance	0	0	1	1
	Unknown	0	0	1	1
	Poor asthma control and intercurrent illness	1	0	0	1
T24–T36	Total	1 (5.3)	0 (0)	0 (0)	1 (1.6)
T24–T36	Lost to follow-up	1	0	0	1
Total		3 (15.8)	6 (26.1)	4 (19.0)	13 (20.6)

TABLE 35

Summary of adherence to therapy

Visit	Fluticasone (n)	Fluticasone plus salmeterol (n)	Fluticasone plus montelukast (n)	Total (n)	Cumulative withdrawals (n)	Proportion of inhaler doses missed (patients with missing information)				Proportion of tablet doses missed (patients with missing information)
Visit	Fluticasone (n)	Fluticasone plus salmeterol (n)	Fluticasone plus montelukast (n)	Total (n)	Cumulative withdrawals (n)	Fluticasone	Fluticasone plus salmeterol	Fluticasone plus montelukast	Total	Fluticasone	Fluticasone plus salmeterol	Fluticasone plus montelukast	Total
T0	19	23	21	63	0	–	–	–	–	–	–	–	–
T8	19	22	20	61	2	0.07 (2/19)	0.04 (3/22)	0.04 (2/20)	0.05 (5/61)	0.05 (2/19)	0.07 (3/22)	0.09 (2/20)	0.07 (5/61)
T24	18	17	19	54	9	0.04 (3/18)	0.05 (4/17)	0.05 (3/19)	0.04 (8/54)	0.04 (2/18)	0.04 (4/17)	0.06 (3/19)	0.04 (7/54)
T36^a	12	14	13	39	13	0.02 (1/12)	0.04 (1/14)	0.04 (3/13)	0.03 (5/39)	0.02 (1/12)	0.03 (1/14)	0.03 (3/13)	0.03 (5/39)
No T36^a	4	3	4	11	–	–	–	–	–	–	–	–	–
T48	16	17	17	50	13	0.07 (1/16)	0.04 (2/17)	0.28 (1/17)	0.13 (4/50)	0.08 (1/16)	0.04 (2/17)	0.35 (1/17)	0.16 (4/50)

a: Of the 50 patients who completed the trial up to T48, 39 had a T36 visit and 11 did not have a T36 visit because of early closure of the trial.

TABLE 36

Number of exacerbations requiring a course of oral corticosteroids since T–4 and time to first exacerbation

	Length of follow-up
	Up to 1 year (n = 32)	1–1.5 years (n = 19)	1.5–2 years (n = 10)	2–2.5 years (n = 1)	Total (n = 62)
Children with at least one exacerbation, n (%)	2 (6.3)	6 (31.6)	1 (10.0)	0 (0.0)	9 (14.5)
Total no. of exacerbations since registration (T–4), n (%)
0	30 (93.8)	13 (68.4)	9 (90.0)	1 (100.0)	53 (85.5)
1	2 (6.3)	3 (15.8)	1 (10.0)	0 (0.0)	5 (8.1)
2	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)	2 (3.2)
3	0 (0.0)	2 (10.5)	0 (0.0)	0 (0.0)	2 (3.2)
Median (IQR)	0 (0 to 0)	0 (0 to 1)	0 (0 to 0)	1	0 (0 to 0)
Range	0 to 1	0 to 3	0 to 2	–	0 to 3
Total exposure time (weeks)
Median (IQR)	41.57 (33.57 to 45.86)	65.14 (61.57 to 69.14)	86.29 (82.07 to 89.93)	108.14	51.93 (41.57 to 68.43)
Range (n)	28.29 to 52.00 (30)	55.0 to 78.14 (19^a)	78.43 to 102.14 (8^a)		28.29 to 108.14

a: Four patients (two for 1–1.5 years; two for 1.5–2 years) do not have a CRF completion date; the date that the CRF was received back at the CTU was used to categorise into the follow-up time category.

TABLE 37

Number of beta-2 agonists prescribed since T–4

	Length of follow-up
	Up to 1 year (n = 32)	1–1.5 years (n = 19)	1.5–2 years (n = 10)	2–2.5 years (n = 1)	Total (n = 62)
Children with at least one beta-2 agonist prescribed, n (%)	22 (68.8)	18 (94.7)	9 (90.0)	1 (100.0)	50 (80.6)
Total no. of beta-2 agonists prescribed since T–4,^a n (%)
0	10 (31.3)	1 (5.3)	1 (10.0)	0 (0.0)	12 (19.4)
1	2 (6.3)	1 (5.3)	0 (0.0)	0 (0.0)	3 (4.8)
2	4 (12.5)	2 (10.5)	2 (20.0)	0 (0.0)	8 (12.9)
3	2 (6.3)	2 (10.5)	2 (20.0)	0 (0.0)	6 (9.7)
4	5 (15.6)	3 (15.8)	0 (0.0)	0 (0.0)	8 (12.9)
5	4 (12.5)	1 (5.3)	2 (20.0)	0 (0.0)	7 (11.3)
6	2 (6.3)	2 (10.5)	0 (0.0)	0 (0.0)	4 (6.5)
7	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)	1 (1.6)
8	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)	1 (1.6)
9	0 (0.0)	0 (0.0)	1 (10.0)	0 (0.0)	1 (1.6)
10	1 (3.1)	1 (5.3)	0 (0.0)	0 (0.0)	2 (3.2)
11	0 (0.0)	2 (10.5)	0 (0.0)	0 (0.0)	2 (3.2)
12	0 (0.0)	0 (0.0)	0 (0.0)	1 (100.0)	1 (1.6)
13	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
14	1 (3.1)	0 (0.0)	1 (10.0)	0 (0.0)	2 (3.2)
15	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
16	0 (0.0)	0 (0.0)	1 (10.0)	0 (0.0)	1 (1.6)
17	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
18	1 (3.1)	0 (0.0)	0 (0.0)	0 (0.0)	1 (1.6)
19	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)	1 (1.6)
28	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)	1 (1.6)
Median (IQR)	2.5 (0 to 5)	5 (3 to 10)	4 (2 to 9)	–	4 (2 to 6)
Range	0 to 18	0 to 28	0 to 16	–	0 to 28

a: Total number of beta-2 agonist inhalers prescribed from T–4 (start of entry into run-in) until patient withdrawn or the end of the study for that patient.

TABLE 38

Prescriptions since T–4

Prescription	Length of follow-up
Prescription	Up to 1 year (n = 32)	1–1.5 years (n = 19)	1.5–2 years (n = 10)	2–2.5 years (n = 1)	Total (n = 62)
None prescribed, n (%)	24 (75.0)	11 (57.9)	8 (80.0)	1 (100.0)	44 (71.0)
At least one prescription, n (%)	8 (25.0)	8 (42.1)	2 (20.0)	0 (0.0)	18 (29.0)
Montelukast, n (%)	2 (6.3)	0 (0.0)	1 (10.0)	0 (0.0)	3 (4.8)
Combination therapy (ICS and long-acting beta-2 agonist), n (%)	6 (18.8)	4 (21.1)	1 (10.0)	0 (0.0)	11 (17.7)
Montelukast and combination therapy (ICS and long-acting beta-2 agonist), n (%)	0 (0.0)	2 (10.5)	0 (0.0)	0 (0.0)	2 (3.2)
Montelukast and long-acting beta-2 agonist and combination therapy (ICS and long-acting beta-2 agonist), n (%)	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)	1 (1.6)
Long-acting beta-2 agonist and combination therapy (ICS and long-acting beta-2 agonist), n (%)	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)	1 (1.6)

TABLE 39

Number of A&E admissions due to respiratory problems since T–4

Number of admissions	Length of follow-up
Number of admissions	Up to 1 year (n = 32)	1–1.5 years (n = 19)	1.5–2 years (n = 10)	2–2.5 years (n = 1)	Total (n = 62)
Had at least one admission, n (%)	1 (3.1)	2 (10.5)	1 (10.0)	0 (0.0)	4 (6.5)
0, n (%)	31 (96.9)	16 (84.2)	9 (90.0)	1 (100.0)	57 (91.9)
1, n (%)	1 (3.1)	2 (10.5)	0 (0.0)	0 (0.0)	3 (4.8)
2, n (%)	0 (0.0)	0 (0.0)	1 (10.0)	0 (0.0)	1 (1.6)

TABLE 40

Number of inpatient admissions due to respiratory problems since T–4

Number of admissions	Length of follow-up
Number of admissions	Up to 1 year (n = 32)	1–1.5 years (n = 19)	1.5–2 years (n = 10)	2–2.5 years (n = 1)	Total (n = 62)
0, n (%)	32 (100.0)	18 (94.7)	10 (100.0)	1 (100.0)	61 (98.4)
1, n (%)	0 (0.0)	1 (5.3)	0 (0.0)	0 (0.0)	1 (1.6)