This study is, overall, negative with regard to the primary outcome, indicating no benefit from intermittent montelukast treatment in preschool children with wheeze. This supports the recent findings of Valovirta et al.,¹⁶ who compared intermittent and regular montelukast with placebo and found no benefit. There was an increased time to first USMA requiring hospital admission for wheeze in the montelukast group (but not for other types of USMA) and an increased use of rescue oral corticosteroids; however, the study was not powered to demonstrate these effects and the patchiness of the effect makes its validity questionable. There was no apparent influence of wheeze phenotype, use of inhaled steroids at baseline or alternative genotype stratum on USMA, although wheeze phenotype was based on parental reporting and mean daily dose of inhaled steroids was not assessed. The IRR seen in the montelukast group compared with placebo was 0.88 (p = 0.06) in favour of montelukast, not meeting statistical significance. A larger trial might have power to identify a difference of this magnitude, but the clinical benefit may not justify the exercise; this should be considered in the design of future studies.

A possible effect was seen within the 5/5 genotype stratum, with a suggestion of increased responsiveness in this group (contrary to Lima’s et al. finding,¹¹ but consistent with Telleria et al.¹⁴), but the test for genotype–efficacy interaction was not confirmatory. Furthermore, the small effect seen in urinary LTE₄ levels at baseline was not supportive. Future work will prospectively study montelukast efficacy in the 5/5 genotype stratum and explore the role of putative response modifiers like environmental tobacco smoke exposure¹⁷ and air pollution.

The search for an effective therapy for preschool wheezing illness is hampered by the lack of a clearly defined phenotype with robust biomarkers. This study adopted a pragmatic approach, recruiting a heterogeneous population encompassing numerous likely aetiologies, in the hope that inhibition of LT activity might address a mechanistic pathway common to these probably distinct but overlapping clinical entities. There is evidence to implicate cLTs in a proportion of preschool wheezing disease^5,12 and a greater success in assessing LTE₄ levels during wheeze exacerbation (as opposed to at baseline) might have shed light on the value of this hypothesis and thus the viability of montelukast as a therapeutic target. The lack of a clear genotype–urinary LTE₄ level correlation may reflect a lower than anticipated importance of ALOX5 promoter polymorphism genotype or perhaps that the differences become more evident during exacerbation compared with during convalescence. The LT pathway is complex, and it is possible that several mutations (perhaps in combination, perhaps with an epigenetic influence¹⁷) play a more important role in determining LT activity and montelukast response in this population than ALOX5. Future work will investigate the role of other genes on LTE₄ output and montelukast response, and consideration should be given to stratification of montelukast response trials by LTE₄ levels measured during wheeze exacerbation (or perhaps following a standardised challenge).