Main findings
The MASCOT study faced a number of challenges from the outset. Issues relating to study set-up, the pharmaceutical companies and the lack of a NHS facility for packaging and distribution are dealt with in the section on strengths and weaknesses. The most significant issue was the difficulty in recruitment of children with asthma into the study, particularly from primary care. A number of novel strategies were developed to improve recruitment; however, although recruitment did increase towards the end of the study, the percentage of children eligible for randomisation after the run-in fell and an application for additional funding for the study was rejected.
The strategies are clearly outlined in Appendix 3 and should be considered in any study in the future that attempts to recruit children from both primary and secondary care. Had these strategies been considered from the outset of the study the recruitment issue may have been less of a problem. Unfortunately, although all centres agreed that they would institute the strategies appropriate to them from January 2010 onwards, it often takes time before the rewards of such strategies take hold. Had the study been able to progress into and through the 2010 autumn peak seen in children with asthma there was a feeling among the MASCOT staff that we would have shown an upturn in both registration and randomisation of suitable patients. We still have no doubt that such patients exist in the UK.
Because of the early study closure and resulting small sample size there was insufficient power to detect any statistically significant differences between the three groups for the efficacy analyses. At 48 weeks the RR of exacerbations requiring treatment with oral corticosteroids was 0.91 (98.3% CI 0.07 to 12.05, p = 0.93) for fluticasone compared with fluticasone plus salmeterol; 1.10 (98.3% CI 0.06 to 18.6, p = 0.94) for fluticasone compared with fluticasone plus montelukast; and 1.21 (98.3% CI 0.09 to 15.97, p = 0.86) for fluticasone plus salmeterol compared with fluticasone plus montelukast. The CIs are extremely wide and include clinically important RRs that could favour any of the treatments. The results for the 24-week analysis were similarly inconclusive as were the results for the time to first exacerbation comparing fluticasone with fluticasone plus salmeterol (HR 0.63, 95% CI 0.19 to 2.08); time to first exacerbation comparing fluticasone with fluticasone plus montelukast (HR 1.52, 95% CI 0.34 to 6.7); and time to first exacerbation comparing fluticasone plus salmeterol with fluticasone plus montelukast (HR 2.37, 95% CI 0.68 to 8.2).
Although there were no statistically significant differences in mean quality of life scores adjusted for baseline values for any of the pair-wise treatment comparisons, the mean quality of life score had improved at 48 weeks and at 24 weeks for all treatment groups across all domains, both for the child and for the caregiver. 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%), whereas more children on fluticasone plus salmeterol (71.4%) required at least one beta-2 agonist than children on fluticasone (54.5%) or fluticasone plus montelukast (58.3%) over 48 weeks. These patterns were not supported by the 24-week data and wide CIs for pair-wise comparisons of relative treatment effects make conclusions difficult to draw. Only a few children required a hospital admission during the study, with relative treatment effects difficult to estimate.
The majority of patients (84%) experienced at least one mild or moderate adverse event during follow-up. The most common events reported were respiratory disorders, nervous system disorders, infections and infestations, general disorders and gastrointestinal disorders. The percentage of patients reporting each adverse event type was similar across treatment groups except for nervous system disorders, which were reported by fewer patients on fluticasone plus salmeterol [one patient (4.3%)] than patients on fluticasone plus montelukast [eight patients (38.1%)] and fluticasone [five patients (26.3%)], with a much greater number of events on fluticasone plus montelukast because of one patient experiencing 20 events.
Health economic analyses were extremely limited by the small sample size and very small non-significant differences in quality of life outcomes between the three treatment groups, which mean that very little confidence can be placed in the ICER and CEAC results. Nevertheless, interpreting the results with great caution indicates that fluticasone plus montelukast could be the economic treatment of choice. A key component of this treatment is montelukast, which lost patent status in August 2012. Historically, prices of drugs that lose patent status have fallen between 30% and 70%, particularly within 4–5 years.23,24 This will further improve the cost-effectiveness profile of fluticasone plus montelukast relative to fluticasone plus salmeterol; however, this would be based solely on intervention cost.
The limited data available for children who were registered at T–4 but who were not randomised at T0 suggest that their regimens over the subsequent 12 months were not entirely successful in spite of the control achieved over the 4-week run-in, with 14.5% having an exacerbation, 80.6% requiring at least one short-acting beta-2 agonist prescription and 29% requiring a prescription for at least one further asthma treatment. These results also require careful interpretation as the data represent only 60.2% of the registered non-randomised patients.
Strengths and weaknesses
MASCOT was planned as the largest long-term paediatric asthma study ever. Although UK guidelines on asthma management2 have advocated that, when control is poor on low-dose ICSs alone, combination therapy is the first-line step-up treatment, the evidence base for this remains poor. Before the development of the MCRN in 2005, clinical studies in children were often inadequate in both number and design. MASCOT, it was hoped, would address this by being designed by a paediatric CTU in Liverpool, supported by key respiratory paediatric doctors, nurses and patients. As in all studies the key to success is appropriate study design, good working relations between all concerned and successful recruitment.
MASCOT was in the first wave of MCRN studies funded by the HTA programme. Funding for MASCOT was approved in January 2006 pending the agreement of GSK and MSD to supply the medications, including the necessary placebos. Reaching this agreement took a considerable length of time. It was then discovered that there was no NHS facility large enough to package and supply the medicines for all of the participating research sites. An application was made to the HTA for additional funding for a commercial company to undertake this procedure. The HTA supported the application.
The protocol indicated that metered dose inhalers (MDIs) would be used for the study. Unfortunately, we were then informed that GSK was closing down its factory which was the only worldwide manufacturer for the supply of their MDIs for research purposes. With the approval of all of the principal investigators the study protocol was amended to allow the GSK medications to be in the form of dry powder through Accuhalers. When the montelukast tablets from MSD arrived from America it was noted that their expiration date was only 5 months later. These challenges delayed the study opening and have been well documented in a recent publication in Thorax.25
The study finally opened fully in all centres in May 2009, some 3 years and 4 months after the initial funding had been approved. When the study was being developed in 2005 there was clinical belief that it would not be difficult to recruit up to 90 children over a 12-month period from each of the major recruitment centres. The reality was very different. Prescribing habits in both primary and secondary care changed between these dates, with increasing numbers of children with asthma being commenced on long-acting beta-2 agonists or montelukast, rendering them ineligible for the MASCOT study. The chief investigator and the trial co-ordinator visited all of the research centres from October 2009 to January 2010 to try to discover ways to improve recruitment. A number of novel ideas were developed and these were put in place by January/February 2010.
The TMG requested a meeting with the HTA to discuss recruitment challenges, the development of novel additional recruitment strategies and the possibility of opening new recruitment sites. The TMG considered closing sites that were failing to recruit successfully and the chief investigator and trial co-ordinator had communicated with other sites where there were good working relationships between the research networks and where the prospective principal investigator believed that it would be practically possible to recruit ≥ 30 patients within the allocated time. The HTA met with the TMG but wanted to see a doubling of patients recruited and randomised over the subsequent 3 months before considering the possibility of opening any new potential sites. Over those 3 months patient recruitment into the study did double but there was little increase in patients with sufficient symptoms to be randomised into the double-blind arm of the study. The HTA therefore closed the study to recruitment in June 2010. Patients who had been randomised up to that date were allowed to continue in the study either for a full 48 weeks or until the end of January 2011, whichever occurred first.
As indicated in the results section, the small number of children recruited and randomised was insufficient to show clinically or statistically significant differences in asthma exacerbations between the three treatment groups. Interestingly, however, quality of life measurements improved from baseline in all three groups for the duration of the study. This phenomenon is well reported in clinical trials and is thought to be the result of participating in the study itself rather than a specific effect of the individual medications. Differences were seen between the results obtained at 24 weeks and those at 48 weeks but these differences were not consistent and were probably once again related to the small number of patients randomised.
Because of previous safety concerns regarding long-acting beta-2 agonists, particularly in adult male US asthma sufferers, it was encouraging that we saw no safety issues related to any of the treatment groups in this study, but our patient numbers were too small to draw any conclusions. Other recently published papers26,27 have, to a certain degree, also allayed previous concerns.
Our health economic data suggest a possible benefit of fluticasone plus montelukast over fluticasone therapy alone but caution is needed because of the small number of children participating in the MASCOT study.