Intended for healthcare professionals


Randomised, placebo controlled trial of effect of a leukotriene receptor antagonist, montelukast, on tapering inhaled corticosteroids in asthmatic patients

BMJ 1999; 319 doi: (Published 10 July 1999) Cite this as: BMJ 1999;319:87
  1. Claes-Göran Löfdahl, professor of respiratory medicinea,
  2. Theodore F Reiss, senior directorb,
  3. Jonathan A Leff, associate directorb,
  4. Elliot Israel, director, clinical researchd,
  5. Michael J Noonan, clinical investigatore,
  6. Albert F Finn, clinical associate professor of medicine, microbiology, and immunologyf,
  7. Beth C Seidenberg, vice president, clinical researchb,
  8. Thomas Capizzi, senior director, clinical biostatisticsc,
  9. Sudeep Kundu, senior biometricianc,
  10. Philippe Godard, professor of medicineg
  1. a University Hospital, Lund, Sweden
  2. b Department of Pulmonary and Immunology, Merck Research Laboratories, Rahway, NJ 07065, USA
  3. c Department of Biostatistic and Research Data System, Merck Research Laboratories
  4. d Brigham and Women's Hospital, Boston, MA, USA
  5. e Allergy Associate PC Research Center, Portland, OR, USA
  6. f Allergy and Asthma Centers of Charleston, North Charleston, SC, USA
  7. g Arnaud de Vilenueve Hospital, Montpellier, France
  1. Correspondence to: T F Reiss
  • Accepted 19 April 1999


Objective: To determine the ability of montelukast, a leukotriene receptor antagonist, to allow tapering of inhaled corticosteroids in clinically stable asthmatic patients.

Design: Double blind, randomised, placebo controlled, parallel group study. After a single blind placebo run in period, during which (at most) two inhaled corticosteroids dose decreases occurred, qualifying, clinically stable patients were allocated randomly to receive montelukast (10 mg tablet) or matching placebo once daily at bedtime for up to 12 weeks.

Setting: 23 academic asthma centres in United States, Canada, and Europe.

Participants: 226 clinically stable patients with chronic asthma receiving high doses of inhaled corticosteroids (113 randomised to montelukast and 113 to placebo).

Interventions: Every 2 weeks, the inhaled corticosteroids dose was tapered, maintained, or increased (rescue) based on a standardised clinical score.

Main outcome measures: Last tolerated dose of inhaled corticosteroids.

Results: Compared with placebo, montelukast allowed significant (P=0.046) reduction in the inhaled corticosteroid dose (montelukast 47% v placebo 30%; least square mean difference 17.6%, 95% confidence interval 0.3 to 34.8). Fewer patients on montelukast (18 (16%) v 34 (30%) placebo, P=0.01) required discontinuation because of failed rescue.

Conclusions: Montelukast reduces the need for inhaled corticosteroids among patients requiring moderate to high doses of corticosteroid to maintain asthma control.

Key messages

  • Leukotriene receptor antagonists have complementary action to inhaled corticosteroids in asthma

  • Many patients receive higher doses of inhaled corticosteroids than clinically required

  • In this placebo controlled trial, montelukast allowed significant reduction of inhaled corticosteroid doses

  • Fewer patients receiving montelukast had failed rescue than patients receiving placebo


People with persistent asthma often require daily treatment to control symptoms, usually with inhaled corticosteroids.1 Recent publications have shown that in patients whose asthma is not completely controlled by inhaled corticosteroids, adding a second drug rather than increasing the dose of corticosteroid achieves better control.2 3 Leukotriene receptor antagonists have complementary effects to corticosteroids on inflammatory events in asthma,4 5 providing a basis for these additive clinical effects. Montelukast is a potent, specific, cysteinyl leukotriene receptor antagonist which, given daily, improves asthma control in adults 6 7 and children,8 protects against exercise induced bronchoconstriction,9 and decreases sputum eosinophil counts.10 Subgroup analyses of previously reported clinical trials have suggested that montelukast has an additive effect with inhaled corticosteroids.6 11 12

We performed a placebo controlled, double blind, randomised trial to investigate whether daily treatment with montelukast (10 mg) allowed reduction of inhaled corticosteroids while maintaining asthma control.

Participants and methods


We recruited 226 non-smoking adults with a clinical history of asthma for at least 1 year. Their forced expiratory volume in 1 second had to be ≥70% of the predicted value (after withholding β agonist for ≥6 hours) at the prestudy visit and to improve by ≥15% (absolute value) after inhaled β agonist. Patients were excluded if they had had emergency treatment for an asthma exacerbation within 1 month, been in hospital for asthma within 3 months, or had an unresolved upper respiratory tract infection within 3 weeks of the prestudy visit. Patients must have been treated with a stable, twice daily dose of inhaled corticosteroids for at least 3 weeks before the prestudy visit.

The protocol was approved by the ethics review committees or institutional review boards of all participating centres. Written informed consent was obtained from all participants.

Study design

The study was conducted at 23 academic medical centres or large group practices in the United States, Canada, and Europe. We used a randomised, double blind, placebo controlled, two period, parallel group design to investigate the effect of montelukast (10 mg tablet once daily at bedtime, irrespective of food) on the ability to taper the dose of inhaled corticosteroids over 12 weeks. Patients were recruited over 3 months. A prestudy visit preceded a single blind, placebo run in period, which lasted at most 7 weeks. The purpose of the run in was to achieve or approach the minimum inhaled corticosteroid dose necessary to maintain clinical stability. After, at most, two dose reductions, clinically stable patients were eligible for randomisation if they demonstrated during the 7-10 days before or at the allocation visit a forced expiratory volume in 1 second ≥90% of the run in baseline value, less than pre-established levels of asthma symptoms and β agonist use, peak flow at least 65% of the maximum, and required a prespecified minimum dose of inhaled corticosteroids. Allocation was according to a computer generated allocation schedule with a blocking factor of 4. Three blocks were initially sent to each centre. Patients remained on their randomised treatment (10 mg montelukast or placebo) throughout the study. Follow up was stopped at the end of 12 weeks' treatment.


Participants returned to the clinic every 2 weeks, and their inhaled corticosteroids dose was adjusted (decreased, increased, or maintained) based on a previously validated composite clinical score (box).13 The dose of inhaled corticosteroids was tapered or increased by about 25% (according to pre-established criteria). Unstable patients who failed to regain clinical stability after an increase in dose of corticosteroids were withdrawn from the trial and classified as “failed rescue.” Patients were also withdrawn from the study if they interrupted their study drug for more than 5 consecutive days or had worsening asthma requiring treatment with oral, intravenous, or intramuscular corticosteroids.

Criteria for tapering dose of inhaled corticosteroids: composite clinical score

View this table:

Diary cards —As part of the taper criteria, patients recorded daytime symptoms,14 “as needed” use of β agonists, and morning and afternoon peak expiratory flow on a daily diary card.

Spirometry —Standard equipment and quality control procedures were used across all study sites. Participants performed at least three forced vital capacity manoeuvres achieving American Thoracic Society standards.15 The best forced expiratory volume was recorded from each set of measurements.

Statistical analysis

The primary analysis was by intention to treat. All patients with postallocation data were included in the analysis. The last tolerated dose of inhaled corticosteroids was the primary end point and was defined as the last dose during the treatment period at which the subject had a composite clinical score of 2 or 3. An endpoint committee, blind to subject allocation, determined all last tolerated doses for all allocated patients based on pre-established criteria. This end point was compared between treatment groups as the percentage change from preallocation baseline (to adequately summarise changes in corticosteroids of different potencies) with an analysis of variance model. This model was used to construct 95% confidence intervals for the least square means and differences of means. All statistical tests were two tailed, and a P value ≥0.05 was considered significant.

The number (percentage) of patients with failed rescue and tapering were prespecified as a secondary end point and were compared between treatment groups by the Cochran-Mantel-Haenszel test. The time to failed rescue was compared with a log rank test.

The study was designed with a sample size of 100 patients in each group to have 90% power to detect (α=0.050, two tailed test) a mean difference between the two treatment groups of about 30 percentage points in the percentage change from baseline in dose of inhaled corticosteroid.13


Randomisation and withdrawals

Of the 226 patients (113 randomised to placebo, 113 to montelukast), one could not be included in the analysis (montelukast) because of insufficient postallocation data. A further 48 patients did not complete 12 weeks' treatment: 31 in the placebo group (16 for “failed rescue,” nine for clinical adverse experiences, three for protocol deviations, and three withdrew consent) and 17 in the montelukast group (seven for “failed rescue,” four for clinical adverse experiences, four for protocol deviations, one withdrew consent, and one lost to follow up). The two treatment groups did not differ in baseline values (table 1).

Table 1.

Characteristics of patients before randomisation. Values are mean (SD) unless stated otherwise

View this table:

The mean inhaled corticosteroid dose for the 226 allocated patients was reduced 37.3% before randomisation (tables 2 and 3). The preallocation doses of inhaled corticosteroids, daily symptoms, β agonist use, and forced expiratory volumes were comparable between treatment groups (tables 1 and 3). Compared with placebo, montelukast significantly (P=0.046) reduced the last tolerated dose of inhaled corticosteroids. Mean percentage changes from preallocation baseline were 47% and 30% for the montelukast and placebo groups, respectively (least square mean difference 17.6%, 95% confidence interval 0.3 to 34.8; table 3). Tapering was consistent across all inhaled corticosteroids.

Table 2.

Mean (range) doses* of inhaled corticosteroids before study and before allocation

View this table:
Table 3.

Mean (SD) changes in inhaled corticosteroid dose (μg/day) during study

View this table:

Forty five (40%) patients on montelukast and 33 (29%) on placebo tapered completely off inhaled corticosteroids. Seventy (62%) patients on montelukast compared with 57 (50%) on placebo had their dose tapered by 50% or more of the prerandomisation baseline; 31 (28%) on montelukast and 41 (36%) on placebo could not taper at all (P=0.055).

Thirty four (30%) patients on placebo had failed rescues compared with 18 (16%) on montelukast (P=0.01). In addition, time to failed rescue was better with montelukast compared with placebo (P=0.042, log rank test).

No significant differences occurred between treatment groups in the change in spirometry measurements (% change in forced expiratory volume 2.7% montelukast v 5.4% placebo), asthma symptom score (0.07 v 0.12), and β agonist use (0.29 v 0.36 puffs/day) at the last tolerated dose.

There were no significant differences in the frequency of clinical or laboratory adverse experiences between placebo and montelukast patients. Nine patients on placebo and four on montelukast stopped the trial because of adverse effects (worsening asthma for all montelukast patients).


We found that, compared with placebo, montelukast (10 mg, administered once daily at bedtime) allowed reduction of moderate to high doses of inhaled corticosteroids while maintaining clinical stability in patients with chronic asthma. These data demonstrate that montelukast has an additive benefit to inhaled corticosteroids in this setting. Another leukotriene receptor antagonist, pranlukast, has been shown to decrease requirements for inhaled corticosteroids, although the study used a different design.16 Given the potential limitations of prolonged exposure to high doses of inhaled corticosteroids,1720 therapeutic strategies providing clinical control of asthma with the lowest corticosteroid dose would be valuable and are consistent with present treatment guidelines.1

Patients often receive higher doses of inhaled corticosteroids than required.17 For example, during the placebo run in period of this trial, patients reduced their inhaled corticosteroids dose by about 37%, consistent with previous publications.21 Additionally, a pilot study for our trial that carefully evaluated the taper and rescue protocols also found appreciable reductions of inhaled corticosteroids. Because we expected a large placebo response we included a 5-7 week, single blind, placebo run in period.

At allocation, patients appeared well matched. Although the inhaled corticosteroids dose was slightly lower in montelukast patients, the similarity in forced expiratory volume, symptoms, and use of β agonists supports similar disease severity in these patients. Furthermore, a lower inhaled corticosteroid dose at allocation might provide less opportunity for tapering in montelukast patients.

A second, important finding of this study was that the number of failed rescues was lower with montelukast than placebo. Patients taking montelukast were more likely to be successfully rescued after an episode of worsening asthma and therefore more likely to remain clinically stable.

Since the study was time based (12 weeks after randomisation) not event based, it had certain limitations. In some patients, tapering (based on clinical stability) could have continued beyond 12 weeks; therefore, the true number of patients who could have stopped inhaled corticosteroids is unknown. Additionally, since patients were not followed beyond the 12 weeks we do not know how many required the reintroduction of inhaled corticosteroids to maintain clinical stability. A study with longer follow up would allow more complete understanding of the clinical importance of the effect of montelukast observed in this study.

Montelukast was generally well tolerated in this asthmatic population taking high doses of inhaled corticosteroids. Double blind, placebo controlled, clinical trials have confirmed the short term (≤3 month) safety profile in adults and children (ages 6-14).69 Additional long term clinical trials and real world clinical experience are necessary to understand the long term safety profile of leukotriene receptor antagonists.

In conclusion, our results suggest that leukotriene receptor antagonists such as montelukast may be useful for long term treatment of asthmatic patients requiring high doses of inhaled corticosteroids.


The members of the Montelukast Study Group are listed on the BMJ ‘s website. We thank Ji Zhang for expert statistical advice and Lars-Olof Eriksson for study monitoring.

Contributors: The manuscript was written collaboratively by TFR, JAL, C-GL, EI, PG, and SK. TFR, JAL, BCS, SK, C-GL, and TC participated in the design of the clinical trial. C-GL, EI, MJN, AFF, and PG participated in data collection. SK and TC participated in the statistical planning and analysis. TFR is the study guarantor.


  • Funding This study was supported by a grant from Merck Research Laboratories, Rahway, NJ.

  • Competing interests C-GL is a member of a Merck advisory board, AFF owns shares of Merck stock and is a member of the Merck speakers bureau, MJN is a member of Merck speakers bureau and a member of a scientific council sponsored by Merck, and EI is a member of a scientific council sponsored by Merck.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
View Abstract