Addition to inhaled corticosteroids of long-acting beta2-agonists versus anti-leukotrienes for chronic asthma

  • Review
  • Intervention

Authors


Abstract

Background

Asthma patients who continue to experience symptoms despite taking regular inhaled corticosteroids (ICS) represent a management challenge. Long-acting beta2-agonists (LABA) and anti-leukotrienes (LTRA) are two treatment options that could be considered as add-on therapy to ICS.

Objectives

To compare the safety and efficacy of adding LABA versus LTRA to the treatment regimen for children and adults with asthma who remain symptomatic in spite of regular treatment with ICS. We specifically wished to examine the relative impact of the two agents on asthma exacerbations, lung function, symptoms, quality of life, adverse health events and withdrawals.

Search methods

We searched the Cochrane Airways Group Specialised Register until December 2012. We consulted reference lists of all included studies and contacted pharmaceutical manufacturers to ask about other published or unpublished studies.

Selection criteria

We included randomised controlled trials (RCTs) conducted in adults or children with recurrent asthma that was treated with ICS along with a fixed dose of a LABA or an LTRA for a minimum of four weeks.

Data collection and analysis

Two review authors independently assessed the risk of bias of included studies and extracted data. We sought unpublished data and further details of study design when necessary.

Main results

We included 18 RCTs (7208 participants), of which 16 recruited adults and adolescents (6872) and two recruited children six to 17 years of age (336) with asthma and significant reversibility to bronchodilator at baseline. Fourteen (79%) trials were of high methodological quality.

The risk of exacerbations requiring systemic corticosteroids (primary outcome of the review) was significantly lower with the combination of LABA + ICS compared with LTRA + ICS—from 13% to 11% (eight studies, 5923 adults and 334 children; risk ratio (RR) 0.87, 95% confidence interval (CI) 0.76 to 0.99; high-quality evidence). The number needed to treat for an additional beneficial outcome (NNTB) with LABA compared with LTRA to prevent one additional exacerbation over four to 102 weeks was 62 (95% CI 34 to 794). The choice of LTRA, the dose of ICS and the participants' age group did not significantly influence the magnitude of effect. Although results were inconclusive, the effect appeared stronger in trials that used a single device rather than two devices to administer ICS and LABA and in trials of less than 12 weeks' duration.

The addition of LABA to ICS was associated with a statistically greater improvement from baseline in lung function, as well as in symptoms, rescue medication use and quality of life, although the latter effects were modest. LTRA was superior in the prevention of exercise-induced bronchospasm. More participants were satisfied with the combination of LABA + ICS than LTRA + ICS (three studies, 1625 adults; RR 1.12, 95% CI 1.04 to 1.20; moderate-quality evidence). The overall risk of withdrawal was significantly lower with LABA + ICS than with LTRA + ICS (13 studies, 6652 adults and 308 children; RR 0.84, 95% CI 0.74 to 0.96; moderate-quality evidence). Although the risk of overall adverse events was equivalent between the two groups, the risk of serious adverse events (SAE) approached statistical significance in disfavour of LABA compared with LTRA (nine studies, 5658 adults and 630 children; RR 1.33, 95% CI 0.99 to 1.79; P value 0.06; moderate-quality evidence), with no apparent impact of participants' age group.

The following adverse events were reported, but no significant differences were demonstrated between groups: headache (11 studies, N = 6538); cardiovascular events (five studies, N = 5163), osteopenia and osteoporosis (two studies, N = 2963), adverse events (10 studies, N = 5977 adults and 300 children). A significant difference in the risk of oral moniliasis was noted, but this represents a low occurrence rate.

Authors' conclusions

In adults with asthma that is inadequately controlled by predominantly low-dose ICS with significant bronchodilator reversibility, the addition of LABA to ICS is modestly superior to the addition of LTRA in reducing oral corticosteroid–treated exacerbations, with an absolute reduction of two percentage points. Differences favouring LABA over LTRA as adjunct therapy were observed in lung function and, to a lesser extend, in rescue medication use, symptoms and quality of life. The lower overall withdrawal rate and the higher proportion of participants satisfied with their therapy indirectly favour the combination of LABA + ICS over LTRA + ICS. Evidence showed a slightly increased risk of SAE with LABA compared with LTRA, with an absolute increase of one percentage point. Our findings modestly support the use of a single inhaler for the delivery of both LABA and low- or medium-dose ICS. Because of the paucity of paediatric trials, we are unable to draw firm conclusions about the best adjunct therapy in children.

Résumé scientifique

Ajout de bêta2-agonistes à action prolongée versus ajout d'antileucotriènes aux corticoïdes inhalés dans l'asthme chronique

Contexte

La persistance des symptômes chez certains patients asthmatiques malgré la prise régulière de corticoïdes inhalés (CSI) constitue un défi pour la prise en charge thérapeutique. Les bêta2-agonistes à action prolongée (BAAP) et les antileucotriènes (ARL) sont deux options qui pourraient être envisagées comme traitement d'appoint en complément des CSI.

Objectifs

Comparer l'efficacité et l'innocuité de l'ajout d'un BAAP par rapport à un ARL au schéma thérapeutique pour les enfants et les adultes asthmatiques demeurant symptomatiques malgré un traitement régulier aux CSI. Nous avons spécifiquement voulu examiner l'impact relatif des deux agents sur les exacerbations de l'asthme, la fonction pulmonaire, les symptômes, la qualité de vie, les événements indésirables et les arrêts prématurés.

Stratégie de recherche documentaire

Nous avons effectué des recherches dans le registre spécialisé du groupe Cochrane sur les voies respiratoires jusqu'en décembre 2012. Nous avons consulté les références bibliographiques de toutes les études incluses et nous avons contacté des fabricants de produits pharmaceutiques pour les interroger sur d'autres études publiées ou non publiées.

Critères de sélection

Nous avons inclus des essais contrôlés randomisés (ECR) réalisés chez des adultes ou des enfants atteints d'asthme récurrent traités aux CSI associés à une dose fixe d'un BAAP ou à d'un ARL pendant un minimum de quatre semaines.

Recueil et analyse des données

Deux auteurs de la revue ont évalué le risque de biais des études incluses et extrait les données de manière indépendante. Lorsque cela s'avérait nécessaire nous avons cherché à obtenir des données non publiées et des informations supplémentaires concernant les plans d'étude.

Résultats principaux

Nous avons inclus 18 ECR (7208 participants), dont 16 portaient sur des adultes et des adolescents (6872) et deux portaient sur des enfants de six à 17 ans (336) atteints d'asthme et présentant une réversibilité significative aux bronchodilatateurs à l'inclusion dans l'étude. Quatorze essais (79 %) étaient de qualité méthodologique élevée.

Le risque d'exacerbations nécessitant des corticostéroïdes systémiques (critère de jugement principal de la revue) était significativement inférieur avec la combinaison BAAP + CSI par rapport à ARL + CSI passant de 13 % à 11 % (huit études, 5923 adultes et 334 enfants ; risque relatif (RR) 0,87, intervalle de confiance à 95 % (IC) de 0,76 à 0,99 ; preuves de qualité élevée). Le nombre de sujets à traiter pour obtenir un résultat bénéfique supplémentaire (NSTB) avec les BAAP par rapport aux ARL pour prévenir une nouvelle exacerbation pendant quatre à 102 semaines était de 62 (IC à 95 % 34 à 794). Le choix de l'ARL, la dose de CSI et le groupe d'âge des participants n'avaient pas d'influence significative sur l'ampleur de l'effet. Bien que les résultats n'étaient pas concluants, l'effet semblait plus important dans les essais qui utilisaient un seul dispositif plutôt que deux pour administrer les CSI + BAAP et dans les essais de moins de 12 semaines de durée.

L'ajout d'un BAAP aux ICS était associé à une amélioration statistiquement supérieure de la fonction pulmonaire par rapport à l'inclusion dans l'étude, ainsi que dans les symptômes, l'utilisation de médicaments de secours et la qualité de vie, bien que ces derniers effets étaient modestes. Un ARL était supérieur dans la prévention du bronchospasme induit par l'exercice. Davantage de participants étaient satisfaits avec la combinaison BAAP + CSI par rapport à ARL + CSI (trois études, 1625 adultes ; RR 1,12, IC à 95 % 1,04 à 1,20 ; preuves de qualité modérée). Le risque global d'arrêt prématuré était significativement plus faible avec BAAP + CSI par rapport à ARL + CSI (13 études, 6652 adultes et 308 enfants ; RR 0,84, IC à 95 % 0,74 à 0,96 ; preuves de qualité modérée). Bien que le risque des événements indésirables dans leur ensemble était équivalent entre les deux groupes, le risque d'événements indésirables graves (EIG) était proche de la significativité statistique en défaveur des BAAP par rapport aux ARL (neuf études, 5658 adultes et 630 enfants ; RR 1,33, IC à 95 % 0,99 à 1,79 ; P = 0,06 ; preuves de qualité moyenne), avec apparemment aucun impact du groupe d'âge des participants.

Les événements indésirables suivants ont été signalés, mais aucune différence significative n'était démontrée entre les groupes: maux de tête (11 études, N = 6538); événements cardio-vasculaires (cinq études, N = 5163), ostéopénie et ostéoporose (deux études, N = 2963), événements indésirables (10 études, N = 5977 adultes et 300 enfants). Une différence significative dans le risque de candidose de la cavité buccale a été notée, mais cela représente un faible taux d'occurrence.

Conclusions des auteurs

Chez les adultes asthmatiques insuffisamment contrôlés par, principalement, des CSI à dose faible avec une réversibilité significative aux bronchodilatateurs, l'ajout d'un BAAP aux ICS est légèrement supérieur à l'ajout d'un ARL pour réduire les exacerbations traitées par la prise de corticoïdes oraux, avec une réduction absolue de deux points de pourcentage . Les différences en faveur des BAAP par rapport aux ARL comme traitement d'appoint ont été observées dans la fonction pulmonaire et, dans une moindre mesure, dans l'utilisation de médicaments de secours, les symptômes et la qualité de vie. Le plus faible taux global d'arrêt prématuré et la plus grande proportion de participants satisfaits de leur traitement sont indirectement en faveur de l'association BAAP + CSI par rapport à ARL + CSI. Des données probantes ont montré un risque légèrement accru d'EIG avec les BAAP par rapport aux ARL, avec une augmentation absolue d'un point de pourcentage. Nos résultats sont légèrement en faveur de l'utilisation d'un seul inhalateur pour l'administration conjointe BAAP + CSI à dose faible ou moyenne. En raison de la rareté des essais pédiatriques, nous ne sommes pas en mesure de tirer des conclusions définitives concernant le meilleur traitement d'appoint chez les enfants.

Plain language summary

Effects of long-acting beta2-agonists compared with anti-leukotrienes when added to inhaled corticosteroids?

Background

People who continue to experience asthma symptoms despite regular use of inhaled corticosteroids (ICS) represent an asthma management challenge. The addition of a long-acting beta2-agonist (LABA) and the addition of an oral anti-leukotriene (LTRA) are two therapeutic options.

Review question

What is the best add-on treatment in adults and children with asthma whose condition is inadequately controlled with ICS alone: LABA or LTRA?

What evidence did we find?

From available evidence until December 2012, we found 16 trials involving 6872 adults and two trials involving 336 children contributing to the review. The risk of asthma exacerbations requiring the use of corticosteroids was lower with the combination of LABA + ICS compared with LTRA + ICS—from 13% to 11%. The choice of LTRA (montelukast or zafirlukast), the dose of ICS and the age of patients did not significantly affect the results. The effect appeared stronger in trials of short duration and in those using a single device to administer both ICS and LABA. Serious adverse events were more common with LABA than with LTRA, particularly in adults. The combination of LABA + ICS was superior to LTRA + ICS in terms of lung function and was modestly superior in other indicators of the control of asthma and quality of life. LTRA was found superior in preventing deterioration during exercise. The risk of withdrawal from a trial for any reason was significantly lower with LABA than with LTRA. More patients were satisfied with the combination of LABA + ICS, and fewer changed therapy if they started with LABA instead of LTRA.

Conclusion

In adults whose asthma is inadequately controlled with ICS, the addition of LABA to ICS was found to be modestly superior to LTRA + ICS. Although both options appeared safe, evidence suggests that slightly more serious adverse events (SAE) may be seen with LABA than with LTRA, particularly when separate devices are used to administer LABA + ICS. Because only two paediatric trials contributed data to the review, the best adjunct strategy to ICS remains uncertain for children.

Quality of the evidence

Our confidence in the quality of evidence is high in the primary efficacy outcome (i.e. patients with exacerbation requiring systemic corticosteroids and high to moderate efficacy and safety measures). One open-label study of adult patients with asthma contributed to certain outcomes that led to downgrading of the quality of evidence from high to moderate. Of note, only two paediatric trials contributed to this review.

Résumé simplifié

Effets des bêta2-agonistes à action prolongée par rapport aux antileucotriènes en complément des corticoïdes inhalés ?

Contexte

Les personnes qui continuent de présenter des symptômes d'asthme malgré l'utilisation régulière de corticoïdes inhalés (CSI) constituent un défi de prise en charge de l'asthme. L'ajout d'un bêta2-agoniste à action prolongée (BAAP) et l'ajout d'un antileucotriène oral (ARL) sont deux options thérapeutiques.

Question de la revue

Quel est le meilleur traitement d'appoint chez les adultes et enfants asthmatiques dont la pathologie est mal contrôlée sous CSI seul: BAAP ou ARL ?

Quelles preuves avons-nous trouvées ?

Des données disponibles jusqu'en décembre 2012, nous avons trouvé 16 essais portant sur 6872 adultes et deux essais portant sur 336 enfants qui ont contribué à la revue. Le risque d'exacerbations de l'asthme nécessitant l'utilisation de corticostéroïdes était inférieur avec la combinaison BAAP + CSI par rapport à ARL + CSI — passant de 13 % à 11 %. Le choix de l'ARL (montelukast ou zafirlukast), la dose de CSI et l'âge des patients n'ont pas d'effet significatif sur les résultats. L'effet semblait plus important dans les essais de courte durée et chez ceux utilisant un dispositif unique pour administrer les CSI + BAAP. Les événements indésirables graves étaient plus fréquents avec les BAAP qu'avec les ARL, en particulier chez l'adulte. La combinaison BAAP + CSI était supérieure à ARL + CSI en terme de fonction pulmonaire et était légèrement supérieure dans d'autres indicateurs du contrôle de l'asthme et de la qualité de vie. Les ARL ont été trouvés supérieurs pour prévenir la détérioration pendant l'exercice. Le risque d'arrêt prématuré d'un essai pour toutes raisons confondues était significativement plus faible avec les BAAP qu'avec les ARL. Davantage de patients étaient satisfaits avec la combinaison BAAP + CSI, et moins de patients ont changé de traitement, s'ils l'avaient commencé avec un BAAP au lieu d'un ARL.

Conclusion

Chez les adultes dont l'asthme est mal contrôlé sous CSI, l'ajout d'un BAAP aux CSI a été trouvé légèrement supérieur à un ARL + CSI. Bien que les deux options semblaient sûres, les preuves suggèrent que légèrement plus d'événements indésirables graves (EIG) puissent être observés avec les BAAP qu'avec les ARL, en particulier lorsque des dispositifs séparés sont utilisés pour administrer les BAAP + CSI. La meilleure stratégie de traitement complémentaire aux CSI reste incertaine pour les enfants car seulement deux essais pédiatriques ont fourni des données pour la revue,.

Qualité des preuves

Notre confiance dans la qualité des preuves est élevé dans le principal critère de jugement de l'efficacité (c'est-à-dire les patients atteints d'exacerbation nécessitant des corticostéroïdes systémiques et les mesures de l'efficacité grande à modérée et de l'innocuité). Une étude ouverte portant sur des patients adultes souffrant d'asthme a contribué à certains résultats ayant conduit à rétrograder la qualité des preuves de haute à modérée. Il est à noter que seuls deux essais pédiatriques ont contribué à cette revue.

Notes de traduction

Traduit par: French Cochrane Centre 6th July, 2014
Traduction financée par: Financeurs pour le Canada : Instituts de Recherche en Santé du Canada, Ministère de la Santé et des Services Sociaux du Québec, Fonds de recherche du Québec-Santé et Institut National d'Excellence en Santé et en Services Sociaux; pour la France : Ministère en charge de la Santé

Summary of findings(Explanation)

Summary of findings for the main comparison. LABA + ICS compared with LTRA + ICS for chronic asthma
  1. 1Grosclaude 2003 was an open-label study.

LABA + ICS compared with LTRA + ICS for chronic asthma
Patient or population: patients with chronic asthma
Settings: outpatients
Intervention: LABA + ICS
Comparison: LTRA + ICS
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
LTRA + ICS LABA + ICS
Participants with one or more exacerbations requiring systemic corticosteroids
Follow-up: 4 to 102 weeks
126 per 1000 109 per 1000
(96 to 125)
RR 0.87
(0.76 to 0.99)
6257
(8 studies)
⊕⊕⊕⊕
high
Adults are overrepresented, with 5923 adults and 334 children
Morning PEFR: L/min change from baseline
Follow-up: 4 to 48 weeks
Data were presented as generic inverse varianceMean morning PEFR: L/min change from baseline in anti-leukotrienes versus long-acting beta2 agonists in the intervention groups was
15.36 higher
(11.35 to 19.37 higher)
 6161
(11 studies)
⊕⊕⊕⊝
moderate 1
 
FEV1: L change from baseline
Follow-up: 4 to 48 weeks
Data were presented as generic inverse varianceMean FEV1: change from baseline with LABA + ICS was
0.08 L higher
(0.05 to 0.1 higher)
 4999
(10 studies)
⊕⊕⊕⊕
high
 
Rescue-free days: % change from baseline
Follow-up: 4 to 12 weeks
Mean rescue-free days: % change from baseline with LRTA + ICS was
22.58% change from baseline
Mean rescue-free days: % change from baseline with LABA + ICS was
9.18% higher
(5.39 to 12.98 higher)
 2612
(5 studies)
⊕⊕⊕⊝
moderate 1
 
Rescue medication use: puffs/d change from baseline
Follow-up: 4 to 48 weeks
1.9 lower use of puffs per dayMean rescue medication use: puffs/d change from baseline in the intervention groups was
0.49 lower
(0.75 to 0.24 lower)
 4055
(7 studies)
⊕⊕⊕⊕
high
 
Withdrawals for any reason
Follow-up: 4 to 48 weeks
133 per 1000 111 per 1000
(98 to 127)
RR 0.84
(0.74 to 0.96)
6960
(13 studies)
⊕⊕⊕⊝
moderate 1
 
Serious adverse events
Follow-up: 4 to 48 weeks
23 per 1000 31 per 1000
(23 to 41)
RR 1.33
(0.99 to 1.79)
6288
(9 studies)
⊕⊕⊕⊝
moderate 1
 
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio.
GRADE Working Group grades of evidence.
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Background

Leukotrienes are inflammatory molecules released along with several other substances by mast cells during the immediate response to an inhaled allergen. They are derived from arachidonic acid, the precursor of prostaglandins (Wasserman 1988; Wenzel 1997). Two families of leukotrienes have been identified. Leukotriene B4 acts primarily in conditions in which inflammation is dependent on neutrophils, such as cystic fibrosis, inflammatory bowel disease and psoriasis. The second group (C4, D4, E4), called cysteinyl-leukotrienes, bind to highly selective receptors to induce eosinophil- and mast cell-induced bronchoconstriction and inflammation associated with asthma (Davis 1997). Drugs that can interfere with the production (leukotriene synthesis inhibitors) and activity (leukotriene receptor antagonists) of leukotrienes have been designed. Leukotriene synthesis inhibitors (e.g. zileuton) inhibit the enzyme 5-lipoxygenase, thus blocking the production of many leukotrienes (e.g. B4, C4, D4, E4) (Georgitis 1999). Leukotriene (cysteinyl) receptor antagonists (e.g. montelukast, zafirlukast, pranlukast) block leukotriene D4 (LTD4) receptors (Georgitis 1999). Both types of leukotriene modifiers are administered orally as tablets.

Two Cochrane reviews have concluded that leukotriene receptor antagonists (LTRA) are mild anti-inflammatory agents when used as monotherapy (Chauhan 2012), and they bring modest benefit as add-on therapy to inhaled corticosteroids (ICS) (Ducharme 2004). Long-acting beta2 (beta2)-agonists (LABA) have a similar mode of action to that of short-acting beta2-agonists (SABA). Some LABA may have a slightly slower onset of action than beta2SABA (Lotvall 1996) but display prolonged activation of beta2-receptors (Johnson 1995) in bronchial smooth muscle, resulting in prolonged duration of action for up to 12 hours (Rees 1995). LABA are recommended solely as add-on therapy to ICS in patients with moderate to severe asthma who remain symptomatic despite anti-inflammatory therapy (BTS 2012; GINA 2012; Lougheed 2012; NAEPP 2011). Several concerns have been raised about the safety of LABA, predominantly when used without concomitant ICS (Cates 2008a; Cates 2008b; Salpeter 2006; Walters 2007). Because evidence suggests increased risk of severe exacerbations and death, a recent call was made for the withdrawal of inhalation devices containing only LABA. The combination of LABA with ICS has been carefully examined and is superior to placebo when introduced as second-line therapy in adults treated with ICS (Ducharme 2010; Ernst 2006). In contrast to adults, the addition of LABA to maintenance dose ICS in children with persistent asthma failed to reach a significant reduction in, but rather showed a trend towards, increased risk of oral corticosteroid-treated exacerbations and hospital admissions. These trends raised concern about the safety of combination therapy in view of modest improvement in children younger than 12 years of age (Ducharme 2010a). Despite non-statistically significant results for outcomes related to serious adverse events, the data do not prove conclusively that the risk of serious adverse events is abolished by the presence of ICS (Cates 2009a; Cates 2009b; Cates 2012).

Individuals with asthma who continue to experience symptoms and have ongoing airway obstruction despite taking regular ICS represent a management challenge. Both LTRA and LABA agents may be considered as add-on therapy to ICS (Adams 2007). Several reasons are known to support the synergistic effect of either combination at the cellular and pathophysiological level. LABA reduce airway hyperresponsiveness by means of functional antagonism (Lipworth 2002), and corticosteroids increase the expression of beta2-adrenergic receptors (Baraniuk 1997); this represents a good combination for synergy. LTRA inhibit the production of cysteinyl leukotrienes—important pro-inflammatory mediators in asthma that are unaffected by corticosteroid treatment. LTRA are particularly effective in allergen-, exercise- and aspirin-induced asthma (Krawiec 2002). Thus, the addition of LTRA or LABA could potentiate the anti-inflammatory effect of ICS and may lead to better asthma control. The current review compares the relative benefits and the safety profile of adding an LTRA or a LABA to the treatment of patients with asthma whose condition is inadequately controlled by ICS and updates a previous Cochrane review on the same topic (Ducharme 2011).

Objectives

To compare the safety and efficacy of adding LABA versus LTRA in children and adults with asthma who remain symptomatic in spite of regular treatment with ICS. We specifically wished to examine the relative impact of the two agents on asthma exacerbations, lung function, symptoms, quality of life, adverse health events and withdrawals.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs) conducted in adults or children for whom a LABA or an LTRA was added, as a fixed-dose combination, to ICS treatment.

Types of participants

Children or adults with recurrent or persistent asthma treated with ICS as the only asthma control medication before study entry.

Types of interventions

Interventions included LABA (e.g. salmeterol, formoterol) or LTRA (e.g. montelukast, pranlukast, zafirlukast, zileuton). Participants were required to be taking a stable dose of ICS throughout the treatment period. The intervention must have been administered for a minimum of four weeks. Inhaled short-acting beta2-agonists and short courses of oral corticosteroids were permitted as rescue interventions.

Types of outcome measures

Primary outcomes
  • Number of participants with asthma exacerbations requiring a rescue short course of systemic corticosteroids.

Secondary outcomes
  • Other measures of severity of exacerbations, such as hospital admissions.

  • Measures reflecting long-term asthma control, such as pulmonary function tests, symptom scores, days or nights or both without symptoms, quality of life, use of rescue fast-acting beta2-agonists and participant satisfaction.

  • Measures of inflammation, such as eosinophilia, serum eosinophil cationic protein and sputum eosinophils.

  • Adverse events including rates of clinical and biochemical adverse events.

  • Withdrawals.

Search methods for identification of studies

Electronic searches

The literature search was updated until December 2012 using the Cochrane Airways Group Specialised Register of trials, which is derived from systematic searches of bibliographic databases, including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE and CINAHL; and from handsearching of respiratory journals and meeting abstracts (see Appendix 1 for further details). All records in the Specialised Register coded as 'asthma' were searched using the following terms.

(((beta* AND agonist*) AND (long-acting OR "long acting")) OR ((beta* AND adrenergic*) AND (long-acting OR "long acting")) OR (bronchodilat* AND (long-acting OR "long acting")) OR (salmeterol OR formoterol OR eformoterol OR advair OR symbicort)) AND (((steroid* OR glucocorticoid* OR corticosteroid*) AND inhal*) OR (budesonide OR beclomethasone OR fluticasone OR triamcinolone OR flunisolide)) AND (leucotrien* OR leukotrien* OR anti-leukotrien* OR anti-leucotrien* OR *lukast).

An additional search of CENTRAL was completed using the above search strategy.

Searching other resources

We reviewed reference lists of all included studies and reviews to identify potentially relevant citations. We asked authors of included studies and pharmaceutical companies to identify other published or unpublished studies until March 2010. We searched by hand abstract books of the American Thoracic Society (ATS) and the European Respiratory Society (ERS) (1998 to 2005) until March 2010. For the 2006, 2010 and 2013 updates, we accessed a register of study results posted by pharmaceutical manufacturers (www.clinicaltrials.gov).

Data collection and analysis

Selection of studies

One review author (BFC) screened the titles, abstracts or descriptors and excluded all studies that clearly were not RCTs or that clearly did not fit the inclusion criteria. Two review authors (BFC and FMD) reviewed the full-text documents of remaining trials, assessing for inclusion on the basis of population, intervention, study design and outcomes. We searched the bibliographies of articles that we retrieved in full to identify additional studies.

Data extraction and management

Data were extracted from the trials by two review authors (BFC and FMD) and were entered into Review Manager 5.2. When necessary, expansions of graphic reproductions and estimations from other data presented in the papers were performed.

We contacted primary study authors to confirm methodology and data extraction, as well as to obtain additional information and clarification, if needed.

Assessment of risk of bias in included studies

We assessed the risk of bias for each study in terms of randomisation, allocation concealment, blinding, handling of withdrawals, selective reporting bias and other sources of bias (see Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions) (Higgins 2008). This replaced the methodology used previously to assess study quality (see Differences between protocol and review).

We assessed the risk of bias of each study for the following items.

  • Randomisation

  • Allocation concealment

  • Blinding

  • Incomplete data

  • Selective reporting

  • Other potential sources of bias

Our judgements of high, low and unclear risk of bias were corroborated by quotations from trial reports, correspondence or summarised information from the relevant sections of individual study reports. The study was deemed to have high methodological quality if reported randomisation procedures and blinding were adequate and a low and balanced group attrition was noted, supporting a low risk of bias.

Assessment of heterogeneity

We measured heterogeneity of effect sizes between studies using the I2 statistic (Higgins 2003). If heterogeneity was suggested by I2 > 40%, a random-effects model was applied to the summary estimates and was reported in the results.

Subgroup analyses were planned to explore possible effect modifications associated with a priori identified variables or to identify the cause of heterogeneity of study results, if any, for the main outcome (i.e. participants with one or more exacerbations requiring systemic corticosteroids). Differences in the magnitude of effect attributable to these subgroups were examined by using the residual Chi2 test from the odds ratios (Deeks 2001).

Data synthesis

All included trials were combined using Review Manager 5.2. For dichotomous variables, we combined data as a pooled fixed-effect model risk ratio (RR) with 95% confidence interval (95% CI). For continuous outcomes, we combined data as a pooled fixed-effect model mean difference (MD) or standard mean difference (SMD) with 95% CI. We calculated the number needed to treat for an additional beneficial outcome (NNTB) for the primary outcome using Visual Rx, a web-based programme available via www.nntonline.net (Cates 2002).

Odds ratios were used for NNTBs, as the results were not affected by selection of the reference treatment (LABA or LTRA). In view of the different duration of trials, the pooled odds ratio was applied to the average exacerbation rate in trials to yield NNTBs for 12 and 48 weeks of treatment.

Subgroup analysis and investigation of heterogeneity

  • Number of inhaler devices used to deliver LABA + ICS therapy (added after publication of the protocol, see Differences between protocol and review).

  • Dose and type of long-acting beta2-agonist (salmeterol, formoterol).

  • Dose and type of anti-leukotriene (montelukast, pranlukast, zafirlukast, zileuton).

  • Dose and type of ICS (in beclomethasone equivalents).

  • Children versus adults.

  • Baseline severity of airway obstruction based on the percentage of predicted forced expiratory volume in one second (FEV1) or the peak expiratory flow (PEF): severe < 60%, moderate 61% to < 80%, mild ≥ 80% (GINA 2012).

Sensitivity analysis

For the primary outcome, we planned to perform the following sensitivity analyses to investigate the potential effects of study duration (≤ 12 weeks, > 12 weeks), publication bias, risk of bias and funding source (trials funded by producers of LABA, studies funded by producers of LTRA, independently funded studies) on the study results. Funnel plots were used to test for the presence of publication bias and other biases in trials contributing data to the main outcomes, if more than 10 studies contributed to the outcome (Egger 1997).

Results

Description of studies

Results of the search

See Table 1 for details of the search history, which formed the basis of the previous version of the review (all years to December 2012).

Table 1. Search history
YearsDetail
All years to January 2004

Citations: 184 (181 from the literature search and three unpublished trials provided by pharmaceutical companies for a total of 184 citations)

Citations excluded: 172: (1) duplicate citations (N = 29), (2) abstracts of considered full-text publications or secondary analyses of the same study (N = 18), (3) not a randomised controlled trial (N = 72), (4) protocol of ongoing trial (N = 1), (5) no consistent co-treatment with ICS (N = 21), (6) one of the tested interventions was not daily LTRA (N = 18), (7) one of the tested interventions was not daily LABA (N = 5), (8) interventions were administered for less than four weeks (N = 6), and (9) use of prohibited co-interventions such as LABA in both groups (N = 2)

Studies meeting the entry criteria of the review: 12 (six full-text publications (Bjermer 2003; Fish 2001; Ilowite 2004; Nelson 2000; Nelson 2001; Ringdal 2003), two unpublished full-text reports (Hultquist 2000; McCarthy 2003) and four abstracts (Gold 2001 (abs); Green (abs) 2002; Leibman (abs) 2002; Nsouli 2001). The abstracts did not provide data in sufficient detail to contribute to the meta-analyses

January 2004 to January 2006

Citations: 60

Citations excluded: 55: The study was a duplicate (i.e. identical citation of one trial report or a subsequent report of a trial) (N = 22); the study was not randomised (N = 2); the study was ongoing (N = 5); the administration of LTRA or LABA was not standardised across treatment groups (N = 3); no consistent co-treatment with ICS was provided (N = 8); one of the tested interventions was not daily LTRA as add-on to ICS (N = 9); one of the tested interventions was not daily LABA as add-on to ICS (N = 2); the tested interventions were administered for less than four weeks (N = 1); the study used prohibited co-intervention (i.e. maintenance oral steroids, theophylline, non-steroidal anti-inflammatory drugs, anticholinergics) (N = 3)

One paediatric study (Fogel 2010), obtained from searches conducted between March 2010 and December 2012, met the eligibility criteria of the review. We replaced BADGER and ELEVATE studies with full-text published papers—Lemanske 2010 and Price 2011—and added information to the Characteristics of included studies tables and data analyses. The addition of one eligible data contributing study resulted in a total of 18 included studies (see Figure 1).

Figure 1.

Study flow diagram.

We conducted an update literature search in January 2014. We identified 13 references for possible inclusion in the next update of the review (Studies awaiting classification).

Included studies

The studies were reported as 14 full-text journal publications (Bjermer 2003; Ceylan 2004; Fogel 2010; Price 2011; Fish 2001; Green 2006; Grosclaude 2003; Ilowite 2004; Lemanske 2010; Nelson 2000; Nelson 2001; Pavord 2007; Ringdal 2003; Storms 2004), two unpublished full-text company reports (SAM40030; SD-004-0216) and two conference abstracts (Hendeles 2004; Nsouli 2001). The conference abstracts did not provide data in sufficient detail to contribute to the meta-analyses, and we have not been able to obtain data from the investigators through correspondence (Hendeles 2004; Nsouli 2001). We describe below the characteristics of all 18 included studies that contributed data to the review.

Design

All but two paediatric trials (Lemanske 2010 and Fogel 2010) employed a parallel-group design.

Participants

Fifteen studies focused on adults, with mean age ranging from 35 to 44 years, a similar sex representation and mean asthma duration ranging from 10 to 26 years. Two studies recruited children six to 17 years of age (Lemanske 2010 and Fogel 2010), and one study (Nsouli 2001) failed to report the age of participants. Most trials (Bjermer 2003; Ceylan 2004; Fish 2001; Grosclaude 2003; Ilowite 2004; Nelson 2000; Ringdal 2003; Storms 2004) allowed the inclusion of adolescents aged ≥ 15 years (or ≥ 12 years for SD-004-0216 and Nelson 2001), although the number of teenagers randomly assigned, if any, was not reported. All trials enrolled only participants who were symptomatic at enrolment despite ICS at doses of 200 to 1000 µg/d of chlorofluorocarbon (CFC)-propelled beclomethasone or equivalent (CFC-BDP) (when ICS doses were reported). Severity of asthma, measured by degree of airway obstruction, was available for all but one study and is reported on the basis of categorisations outlined by GINA 2012 of mild obstruction, FEV1 80% of predicted or higher; moderate obstruction, FEV1 60% to 80% of predicted; severe, less than 60% of predicted. Based on the mean reported FEV1, participants enrolled in two trials had mild airway obstruction (Lemanske 2010; Storms 2004), and moderate airway obstruction was found in 11 trials (Bjermer 2003; Ceylan 2004; Fish 2001; Green 2006; Ilowite 2004; Nelson 2000; Nelson 2001; Pavord 2007; Ringdal 2003; SAM40030; SD-004-0216). Children with a baseline FEV1 70% of predicted or higher with exercise-induced bronchoconstriction were recruited in one study (Fogel 2010). We were not able to ascertain baseline FEV1 in two studies (Price 2011; Grosclaude 2003).

The baseline allergy status of enrolled participants was reported in two studies (Bjermer 2003; Lemanske 2010), with 65% and 77%, respectively, of participants with atopy. Three studies (Bjermer 2003; Ceylan 2004; Grosclaude 2003) reported that 60%, 65% and 51%, respectively, of participants suffered from allergic rhinitis.

Withdrawal rates ranged from 3% to 17% in the LTRA group and from 3% to 27% in the LABA group.

Intervention

During the intervention period, all participants remained on a stable dose of ICS. For the purposes of this review, we considered low ICS doses to be 400 µg/d or less (CFC-BDP equivalent), moderate doses to be 400 to 800 µg/d (CFC-BDP equivalent) and high doses to be 800 µg/d (CFC-BDP equivalent) or higher. Based on this categorisation, most studies assessed the addition of LABA or LTRA to low and moderate doses of ICS (see Table 2). Two trials failed to report the dose of ICS (Price 2011; Nelson 2001).

Table 2. ICS at CFC-BDP equivalent dose (µg/d)
StudyActual dose of ICS (total per day)CFC- BDP equivalent/dPopulationLow/medium/high
Basu 2011Fluticasone; dose not specifiedNot specifiedChildrenUnclear
Bjermer 2003200 μg fluticasone400 μgAdultsLow
Ceylan 2004400 μg budesonide400 μgAdultsLow
Fish 2001Range of doses between 175 and 1700 μg560 µg (range 175 to 1700)AdultsMixed
Fogel 2010200 μg fluticasone400 μgChildrenLow
Gold 2001100 μg fluticasone200 μgAdultsLow
Green 2006200 μg budesonide200 μgAdultsLow
Grosclaude 20031000 μg CFC-BDP and fluticasone 500 μg1000 μgAdultsHigh
Hendeles 2004250 μg fluticasone500 μgAdultsMedium
Ilowite 2004250 μg fluticasone500 μgAdultsMedium
Lemanske 2010200 μg fluticasone400 μgChildrenLow
Nelson 2000200 μg fluticasone400 μgAdultsLow
Nelson 2001Not specifiedRequested from author 08/03AdultsUnclear
Nsouli 2001Unclear500 μgAdultsUnclear
Pavord 2007200 μg fluticasone400 μgAdultsLow
Ringdal 2003200 μg fluticasone400 μgAdultsLow
SAM40030200 μg fluticasone400 μgAdultsLow
SD-004-0216400 μg budesonide400 μgAdultsLow
Price 2011Beclomethasone, budesonide, fluticasoneNot specifiedAdultsUnclear
Storms 2004200 μg fluticasone400 μgAdultsLow

LTRA administered were zafirlukast 20 mg twice daily (Nelson 2001; SD-004-0216) and montelukast 5 mg or 10 mg once daily (Bjermer 2003; Ceylan 2004; Fish 2001; Fogel 2010; Green 2006; Grosclaude 2003; Ilowite 2004; Lemanske 2010; Nelson 2000; Pavord 2007; Ringdal 2003; SAM40030; Storms 2004), depending on the age of participants. Price 2011 was a pragmatic study in which participants were allocated to montelukast 10 mg once daily or zafirlukast 20 mg twice daily.

LABA used included salmeterol 50 µg twice daily in 11 trials (Bjermer 2003; Fish 2001; Grosclaude 2003; Ilowite 2004; Lemanske 2010; Nelson 2000; Nelson 2001; Pavord 2007; Ringdal 2003; SAM40030; Storms 2004), salmeterol 50 µg once daily in one paediatric trial (Fogel 2010) and formoterol 12 µg twice daily in three trials (Ceylan 2004; Green 2006; SD-004-0216). In one study (Price 2011), participants were allocated to salmeterol (as Serevent®, GlaxoSmithKline, Uxbridge, UK) or formoterol (as Foradil®, Novartis Pharmaceuticals UK Ltd, Camberley, UK, or Oxis®, AstraZeneca Ltd, Kings Langley, UK), along with ICS, in separate devices or in fixed-dose combinations with ICS (as Seretide™, GlaxoSmithKline, Uxbridge, UK, and Symbicort®, AstraZeneca Ltd, Kings Langley, UK). In seven studies, the combination therapy (Seretide®, Advair® or Symbicort®) was administered in a single device (Green 2006; Grosclaude 2003; Lemanske 2010; Nelson 2000; Pavord 2007; Ringdal 2003; SAM40030), and in eight studies by separate inhaler devices (Bjermer 2003; Ceylan 2004; Fish 2001; Fogel 2010; Ilowite 2004; Nelson 2001; SD-004-0216; Storms 2004). In the remaining trial (Price 2011), we were unable to determine the number of participants who received combination therapy in a single device versus separate devices.

The intervention period varied from four weeks (Fogel 2010; Nelson 2001; Storms 2004) to six weeks (Green 2006), eight weeks (Ceylan 2004; SD-004-0216), 12 weeks (Fish 2001; Grosclaude 2003; Pavord 2007; Nelson 2000; Ringdal 2003; SAM40030), 16 weeks (Lemanske 2010), 48 weeks (Bjermer 2003; Ilowite 2004) and two years (Price 2011).

Outcomes

The primary outcome (number of participants with exacerbations requiring rescue systemic corticosteroids) was available in eight (seven adult: one paediatric) trials (Bjermer 2003; Fish 2001; Ilowite 2004; Lemanske 2010; Nelson 2000; Nelson 2001; Price 2011; Ringdal 2003), representing 76.5% of the total number of participants randomly assigned to trials included in this review. In two trials, we could not satisfactorily identify the requirements for oral corticosteroids as binary data (Green 2006; Grosclaude 2003). None of the studies identified since the first version of the review contributed additional data to our primary outcome.

Other measures of asthma control (e.g. pulmonary function tests, % fall in FEV1 postexercise, symptoms, use of rescue beta2-agonist, quality of life), withdrawals and adverse events were reported by several included studies.

Excluded studies

In the 2010 update and in the present update, 66 of 69 and 31 of 34 identified citations, respectively, were excluded for the following reasons and were reported as total exclusions (exclusions from 2010 update and present update).

  • The study was a duplicate (i.e. identical citation of a trial report or a subsequent report of a trial), N = 39 (29 + 10).

  • The study was not randomised, N = 6 (2 + 4).

  • The study was ongoing, N = 5 (5 + 0).

  • Administration of LTRA or LABA was not standardised across treatment groups, N = 3 (3 + 0).

  • No consistent co-treatment with inhaled glucocorticoids was provided, N = 8 (8 + 0).

  • One of the tested interventions was not daily LTRA as add-on to ICS, N = 22 (9 + 13).

  • One of the tested interventions was not daily LABA as add-on to ICS, N = 4 (2 + 2).

  • Tested interventions were administered for less than four weeks, N = 2 (1 + 1).

  • The study used prohibited co-interventions, N = 6 (3 + 3).

  • The study did not recruit participants at the step twoa level (i.e. the study recruited corticosteroid-naive participants or participants given combination therapy, N = 3 (3 + 0)).

aSteps one, two and three refer to levels of asthma treatment (BTS 2012).

Risk of bias in included studies

An overview of our judgements for the risk of bias of each study is provided in Figure 2.

Figure 2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Randomisation was clearly described and appropriate in all trials, with the exception of three trials that contributed data to the review (Ceylan 2004; Grosclaude 2003; SD-004-0216) and two abstracts that did not contribute data in sufficient detail to be meta-analysed (Hendeles 2004; Nsouli 2001). All data included in the primary outcome were drawn from eight studies with robust random sequence generation and allocation concealment.

Blinding

Fourteen trials reported double-blinding with identical 'dummy' treatments, and three trials (Ceylan 2004; Grosclaude 2003; Nsouli 2001) had an open-label design. One trial failed to clearly report the means of blinding (Hendeles 2004).

The methodology was confirmed by the authors of all trials contributing data, with the exception of four (Ceylan 2004; Fogel 2010; Grosclaude 2003; Storms 2004). No confirmation was obtained for two studies reported as abstracts (Hendeles 2004; Nsouli 2001).

The eight trials contributing data to the primary outcome were double-blind, double-dummy studies.

Incomplete outcome data

Withdrawal rates were described in all but one study (Nsouli 2001). Withdrawals were not reported by treatment group in two trials (Ceylan 2004; Lemanske 2010).

Analyses within the studies were frequently described as performed by intention-to-treat. However, further information as to how missing data were handled in trials was limited.

Selective reporting

Although representing eight of 18 included trials, the quantity of data pertaining to the primary outcome was important, as data were derived from 76.5% of all participants randomly assigned in the 18 included trials. We were unable to use data from most other studies because of their broader definition of exacerbations or because of inadequate reporting of outcome data (Green 2006; Grosclaude 2003). In one study, exacerbations were not measured (Pavord 2007); in another study, no exacerbations occurred (Storms 2004). In the remaining studies, we could not ascertain whether exacerbations were measured.

Effects of interventions

See: Summary of findings for the main comparison LABA + ICS compared with LTRA + ICS for chronic asthma

Primary outcome: exacerbations requiring oral (systemic) corticosteroids

Eight trials (5923 adults and 334 children) contributed data to the primary outcome (Bjermer 2003; Fish 2001; Ilowite 2004; Nelson 2000; Ringdal 2003; Lemanske 2010; Nelson 2001; Price 2011). A statistically significant reduction in the risk of having an exacerbation requiring systemic corticosteroids was seen with the use of LABA + ICS compared with LTRA + ICS (RR 0.87, 95% CI 0.76 to 0.99) (Figure 3). The addition of LABA lowered the risk of an exacerbation from 13% to 11%—a 2% (95% CI 0 to 3%) risk difference in exacerbations requiring systemic corticosteroids over the use of LTRA. The NNTB from the combination of LABA + ICS compared with LTRA + ICS was 62 (95% CI 34 to 794) to prevent one exacerbation requiring rescue oral corticosteroids over four to 102 weeks (Figure 4).

Figure 3.

Forest plot of comparison: 1 Leukotriene receptor antagonists + ICS versus Long-acting β2-agonists + ICS, outcome: 1.1 Participants with one or more exacerbations requiring systemic corticosteroids.

Figure 4.

In the LRTA + ICS group, 13 out of 100 people had one or more exacerbations requiring OCS over four to 102 weeks, compared with 11 (95% CI 10 to 12) of 100 for the LABA + ICS group.

The results were homogeneous despite different treatment and participant characteristics (I2 = 0%). Although the funnel plot intercept suggested no evidence of publication bias (-0.01, 95% CI -4.23 to 2.10), visual inspection of the funnel plot precludes firm reassurance (Figure 5).

Figure 5.

Funnel plot of comparison: 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, outcome: 1.1 Participants with one or more exacerbations requiring systemic corticosteroids.

Although no heterogeneity was apparent between trials, we had planned a priori to perform subgroup analyses on the following variables to explore their possible influence on the magnitude of effect (effect modification). The subgroup comparison between single and combined inhalers was added after the initial protocol was published as a result of work subsequently published on this topic (Nelson 2003).

Children versus adults

Of the eight trials contributing data to the main outcome, only one trial enrolled children six to 17 years of age with uncontrolled asthma given a low IICS dose; all remaining trials enrolled adults with mild or moderate asthma (Analysis 1.1). The confidence intervals for the paediatric study were much wider than in the pooled results from the studies in adults, but no statistically significant subgroup difference was apparent between results in adults and results in children (Chi² = 0.02, df = 1 (P value 0.90), I² = 0%).

Dose and type of LABA

All included studies contributing data towards the meta-analysis used only salmeterol as the LABA, preventing assessment of the within-class effect of LABA.

Number of inhaler devices used to deliver LABA + ICS

Of the eight studies contributing data to the primary efficacy outcome, three studies (Nelson 2000; Ringdal 2003; Lemanske 2010) provided LABA + ICS delivered in a single inhaler device, and four studies (Bjermer 2003; Fish 2001; Ilowite 2004; Nelson 2001) used separate inhalers; in one study (Price 2011), both separate and combination inhalers were allowed. Use of a single inhaler to deliver LABA + ICS appeared to provide significantly greater protection against exacerbations compared with LTRA + ICS (RR 0.64, 95% CI 0.45 to 0.91) versus use of two separate inhalers (RR 1.06, 95% CI 0.80 to 1.41), yet because of the wide confidence interval around each estimate, no statistically significant difference in the magnitude of effect was noted between subgroups of trials using a single inhaler versus two separate inhalers (Chi² = 2.69, df = 1 (P value 0.10), I² = 62.9%) (Analysis 2.1). Of note, this indirect comparison may be confounded by other differences between trials, including trial duration and ICS dose.

Dose and type of ICS (in beclomethasone-equivalent doses)

No effect modification was apparent in trials that used low, moderate, mixed or unclear ICS doses (Analysis 2.2). Fluticasone was used in five trials (Bjermer 2003; Ilowite 2004; Nelson 2000; Ringdal 2003; Lemanske 2010). Two trials used a variety of ICS, as investigators kept participants on their usual ICS (Fish 2001; Price 2011); we were unable to obtain details on the ICS used in Nelson 2001.

Duration of trials

The four trials of ≥ 48 weeks' duration appeared to display a smaller difference in treatment effect (RR 0.92, 95% CI 0.80 to 1.05) when compared with the other trials of 12 weeks' duration or less (RR 0.66, 95% CI 0.43 to 1.02), although confidence intervals were overlapping and the difference between longer and shorter trials was not statistically significant (Chi2 = 1.95, df = 1 (P value 0.16), random-effects model) (Analysis 2.3). However, differences in trial duration could be confounded by other differences between trials, including compliance, the number of devices and so forth.

All studies contributing to the primary efficacy outcome were at a low risk of bias, were funded by manufacturers of the study drugs and were published as full text. The pooled result was not sensitive to bias from any of the sources we assessed, and we could not ascertain whether funding source or publication status affected the estimated effect.

Dose and type of LTRA

The type of LTRA used in the included trials did not appear to affect the magnitude of effect in the primary efficacy outcome. The RRs of exacerbations for montelukast + ICS and zafirlukast + ICS were 0.84 (95% CI 0.72 to 0.99) and 0.86 (95% CI 0.28 to 2.59), respectively, with no statistically significant difference noted between the two subgroups (Chi² = 2.24, df = 2 (P value 0.33), I² = 10.6%) (Analysis 2.4).

Baseline severity of airway obstruction

As studies that contributed data to the main outcome pertained to participants with moderate airway obstruction who were relatively homogeneous in the average baseline FEV1 (all within 66% to 76% of predicted value), subgroup analyses on baseline severity could not be performed.

Secondary outcomes

Morning peak expiratory flow (PEF) (L/min change from baseline)

Eleven studies in 5723 adults contributed data to the morning PEF results. Greater improvement was seen in morning PEF with LABA compared with LTRA (15.36 L/min, 95% CI 11.35 to 19.37) (Analysis 1.2).

Evening PEF (L/min change from baseline)

Ten studies in 4012 adults contributed data to the meta-analysis of evening PEF. Significantly greater improvement was observed in evening PEF with LABA compared with LTRA (12.64 L/min, 95% CI 10.11 to 15.17) (Analysis 1.3).

Forced expiratory volume in one second (FEV1) (L/s change from baseline)

Ten studies in 4538 adults contributed data to changes in FEV1. Greater improvement in FEV1 was seen with LABA compared with LTRA (0.08 L/s, 95% CI 0.05 to 0.10; random-effects model) (Analysis 1.4).

FEV1 (% fall postexercise)

Two studies in 358 children and adults with asthma contributed data to the outcome. A significant reduction in the % fall in FEV1 due to exercise was reported with LTRA compared with LABA, favouring the combination of LTRA + ICS (Analysis 1.5).

FEV1 (L/s, % change from baseline) and FEV1 (% of predicted)

As a single adult trial contributed date to these outcomes (Storms 2004), we were unable to pool data (Analysis 1.6 and Analysis 1.7).

Rescue-free days (% change from baseline)

Five studies involving 2612 adults contributed data to the change in percentage of rescue-free days. Participants treated with LABA + ICS displayed statistically significant greater improvement in the percentage of days with no rescue medication use compared with LTRA + ICS (MD 9.18, 95% CI 5.39 to 12.98; random-effects model) (Analysis 1.8).

Rescue medication use (puffs/d)

Seven studies in 4055 adults contributed data on rescue medication. The overall estimate showed a statistically significant lower use of rescue medication with LABA + ICS versus LTRA + ICS by an average of one-half puff per day (MD -0.49 puffs/d, 95% CI -0.75 to -0.24; random-effects model) (Analysis 1.9).

Change in asthma quality of life score (higher score is better; change from baseline)

Four studies in 3243 adults reported asthma-specific quality of life: Bjermer 2003 and Ilowite 2004 using montelukast; Nelson 2001 using zafirlukast and Price 2011 using both montelukast and zafirlukast. The overall estimate showed a statistically significant greater improvement in quality of life with LABA + ICS than with LTRA + ICS (SMD 0.12, 95% CI 0.05 to 0.19) (Analysis 1.10).

Symptom-free days (% change from baseline)

Six studies in 2692 adults reported symptom-free days (Fish 2001; Grosclaude 2003; Nelson 2000; Nelson 2001; Pavord 2007; Ringdal 2003). The pooled effect revealed that the addition of LABA (salmeterol was used in all of these trials) increased the proportion of symptom-free days by 7.09% (95% CI 3.66 to 10.52; random-effects model) compared with LTRA (Analysis 1.11).

Daytime symptom scores (high score is worse; change from baseline)

Five studies in 3823 adults reported daytime symptom scores (Fish 2001; Ilowite 2004; Nelson 2000; Nelson 2001; Ringdal 2003). A statistically significant greater improvement from baseline in daytime symptom scores favoured LABA + ICS over montelukast + ICS (SMD -0.18, 95% CI -0.25 to -0.12) (Analysis 1.12).

Change in morning symptom scores

As only one adult trial contributed data to the outcome (Ceylan 2004), we were unable to pool data (Analysis 1.13).

Night-time symptom score (five-point scale, higher score is worse; change from baseline)

As only one trial contributed data to the outcome (Nelson 2001), we were unable to pool data (Analysis 1.14).

Change in number of night-time awakenings per week (change from baseline)

Four studies in 4214 adults reported night-time awakenings (Bjermer 2003; Fish 2001; Ilowite 2004; Nelson 2000). The overall estimate showed that LABA + ICS led to a significantly greater decrease from baseline in night-time awakenings compared with LTRA + ICS (MD -0.12, 95% CI -0.19 to -0.06) (Analysis 1.15).

Change in % of nights with no awakenings per week (change from baseline)

Two studies reported this outcome in 673 adults (Grosclaude 2003; Nelson 2001) and showed a significantly greater increase in the percentage of awakening-free nights per week with LABA + ICS than with LTRA + ICS (MD 6.89%, 95% CI 2.87 to 10.91) (Analysis 1.16).

Change in % of rescue-free nights

As only one trial contributed data to the outcome (Grosclaude 2003), we were unable to pool data (Analysis 1.17).

Withdrawals for any reason

Thirteen studies involving 6652 adults and 308 children reported withdrawals for any reason. Overall, a statistically significant reduction in the risk of withdrawals was seen with LABA (11%) compared with LTRA (13%) as adjunct to ICS (RR 0.84, 95% CI 0.74 to 0.96) (Analysis 1.18).

Withdrawals due to adverse events

Twelve studies in 6290 adults and 300 children reported withdrawals due to adverse events. The overall estimate comparing LABA + ICS versus LTRA + ICS did not show a statistically significant group difference (RR 1.01, 95% CI 0.79 to 1.29) (Analysis 1.19).

Withdrawals due to poor asthma control (exacerbations)

Eight studies (5354 adults) reported withdrawals due to exacerbations. No statistically significant differences were observed in the overall estimate (RR 0.87, 95% CI 0.49 to 1.56; random-effects model) (Analysis 1.20).

Participants with one or more exacerbations requiring hospital admission

Six studies in 4345 adults and 330 children contributed data on the outcome (Bjermer 2003; Grosclaude 2003; Ilowite 2004; Lemanske 2010; Ringdal 2003; Price 2011). No statistically significant difference was noted between the two adjunct therapies (RR 1.34, 95% CI 0.68 to 2.67) (Analysis 1.21).

Serious adverse events (SAE)

Nine studies in 5658 adults and 630 children reported this outcome. No statistically significant group difference was noted in the risk of serious adverse events (SAE) (RR 1.33, 95% CI 0.99 to 1.79, P value 0.06), although a trend approaching significance in favour of LTRA over LABA was observed (2.3% vs 3.1%) (Analysis 1.22 and Summary of findings for the main comparison). This corresponds to a number needed to treat for an additional harmful outcome for one participant to suffer an additional adverse event or NNTH of 132 (95% CI NNTH of 56 to NNTB of 4348) (Figure 6).

Figure 6.

In the LRTA + ICS group, two of 100 people had serious adverse events over 4 to 102 weeks, compared with three (95% CI 2 to 4) of 100 for the LABA + ICS group.

In view of the proposed reasons behind the increased risk of SAE with LABA (Cates 2008a; Cates 2008b; Cates 2009a; Cates 2009b; Cates 2012), we undertook a post hoc subgroup analysis of the pooled estimate by exploring the relationship between the age of participants, the number of inhaler devices and the risk of SAE. The RR of SAE was 0.77 (95% CI 0.31 to 1.90) in the four studies using a single inhaler to deliver LABA + ICS compared with LTRA + ICS, and the RR was 1.43 (95% CI 1.04 to 1.96) in the five studies in which LABA + ICS were delivered by separate inhalers (Analysis 2.5). The difference between the two subgroup estimates did not reach statistical significance (subgroup differences: Chi² = 1.59, df = 1 (P value 0.21), I² = 37.2%). The RR of SAE was 1.35 in the seven adult studies (N = 5658 adults) (95% CI 1.00 to 1.82) compared with 1.00 (95% CI 0.20 to 4.91) in the two paediatric studies (630 children) (Analysis 2.6). The difference between the two subgroup estimates did not reach statistical significance (test for subgroup differences: Chi² = 0.13, df = 1 (P value 0.72); I² = 0%).

Deaths

One study reported deaths (Bjermer 2003) with no statistically significant difference between the two study groups (one death occurred in the LABA group) (Analysis 1.23).

Headache

Eleven studies in 6538 adults reported headache as an adverse event. No statistically significant group difference was reported (RR 1.06, 95% CI 0.90 to 1.25) (Analysis 1.24).

Cardiovascular events

Five studies with 5163 adults reported cardiovascular events (Bjermer 2003; Fish 2001; Ilowite 2004; Nelson 2000; Ringdal 2003). No statistically significant group difference was reported (RR 1.09, 95% CI 0.77 to 1.53) (Analysis 1.25).

Oral moniliasis

Six studies in 5203 adults reported the number of participants with oral moniliasis and described a statistically significant increased risk of oral moniliasis with the addition of LABA compared with montelukast to ICS (RR 1.86, 95% CI 1.00 to 3.44) (Analysis 1.26). Yet, the occurrence rates were low, and this represents an average risk of oral moniliasis of 1% for LABA and 0.5% for LTRA. The risk difference for this outcome was 0.01 (95% CI 0 to 0.01).

Osteopenia and osteoporosis

Two studies in 2963 adults reported the outcome (Bjermer 2003; Ilowite 2004) with no statistically significant difference between study groups (RR 0.56, 95% CI 0.12 to 2.63) (Analysis 1.27).

Elevated liver enzymes

As only one trial contributed data to this outcome (Bjermer 2003), we were unable to pool data (Analysis 1.28).

Overall adverse events

Ten studies in 5977 adults and 300 children reported adverse events with no statistically significant group difference. In fact, the absence of group differences (RR 1.02, 95% CI 0.99 to 1.06) (Analysis 1.29) met our a priori definition of equivalence.

Participants' treatment satisfaction and participants switched to alternative treatment

Three studies in 2020 adults (Fish 2001; Nelson 2001; Ringdal 2003) reported significantly higher participant satisfaction with the combination of LABA + ICS than LTRA + ICS (RR 1.12, 95% CI 1.04 to 1.20) (Analysis 1.30). A random-effects model was used because of the high level of heterogeneity (I2 = 62%). Only one trial reported data on participants switched to alternative treatment during the trial period, again clearly favouring the combination of LABA + ICS (RR 0.01, 95% CI 0.00 to 0.16) (Analysis 1.31).

Change from baseline in serum eosinophils (× 109/L)

Two adult studies reported this outcome (Bjermer 2003; Ilowite 2004), which showed a statistically significant greater decrease from baseline in serum eosinophils with LTRA + ICS than with LABA + ICS (MD 0.04, 95% CI 0.02 to 0.05) (Analysis 1.32).

Discussion

Primary outcome

This review demonstrates that, in adult study participants who remain symptomatic on low or moderate doses of ICS, the addition of a LABA significantly reduces by 13% the risk of participants experiencing one or more exacerbations requiring oral corticosteroids compared with the addition of a LTRA. The proportion of exacerbations requiring oral corticosteroids was 13% in those treated with a combination of LTRA + ICS versus 11% with the use of LABA, for an absolute risk reduction of 2%. Sixty-two participants had to be treated over four to 102 weeks with LABA + ICS instead of LTRA + ICS to prevent one participant from experiencing an exacerbation requiring rescue oral corticosteroids.

Results were homogeneous between trials. The choice of LTRA did not appear to affect the magnitude of the benefit related to LABA. When compared with LTRA + ICS, a single inhaler containing LABA + ICS was associated with a 30% reduction in the risk of exacerbations requiring systemic corticosteroids, and a 10% reduction was observed when the two drugs were delivered separately. This difference was not statistically significant and should be interpreted with caution because it is not based on a head-to-head comparison and may be confounded by other differences between trials that used single or separate inhalers (such as trial duration, ICS dose used and age of participants). However, the direction of effect is congruent with findings of head-to-head comparisons of single versus separate inhalers (Nelson 2003). No effect of age, ICS dose to which to add adjunct therapy and choice of LTRA was apparent in terms of the magnitude of effect. Although subgroup analyses showed no apparent effect of trial duration on the magnitude of effect, study findings suggest that the superiority of LABA over LTRA was less apparent in children and in trials lasting 48 weeks or longer compared with those lasting 12 weeks or less. Baseline severity of airway obstruction, choice of LABA, type of ICS and methodological quality were too similar between the studies to allow exploration of any effect modification. Additional trials are needed to confirm these trends to enable meta-regression and untangle the effect of other trial and participant characteristics. Consequently, we were unable to identify participant characteristics that would allow us to better guide the selection of treatment for specific patients and, because of the paucity of paediatric trials, we could not firmly conclude which adjunct therapy is best for use in children.

Secondary outcomes

Statistically significant improvements were seen with LABA + ICS versus LTRA + ICS for most secondary outcomes. The average group difference in improvement from baseline in FEV1 between LABA and LTRA was 81 mL (95% CI 60 to 100 mL), specifically, 215 mL with the addition of LABA to ICS compared with 134 mL with the addition of LTRA to ICS. A change of 200 mL or more in FEV1 is considered a clinically important difference, as it exceeds normal intrasubject variation; this would suggest that in general the change in LTRA did not exceed normal variation (ATS 1991). In the absence of a placebo group, the magnitude of benefit from ICS alone is unclear. Yet, these changes from baseline exceed the previously reported benefit for the additive effect of each drug in comparison with the use of ICS alone, suggesting that other participant or study characteristics may be at play. Indeed, previous Cochrane reviews have quantified the magnitude of improvement in FEV1 attributable to each drug over that of ICS and placebo. The addition of LABA to ICS was associated with an increase of 110 to 120 mL (Ducharme 2010), and an increase of 60 mL was observed for the combination of LTRA + ICS over ICS alone (Ducharme 2004). The clinical importance of the observed differences in favour of LABA over LTRA (81 mL in FEV1, 15 mL in morning PEF and 12 mL in evening PEF) is debatable, particularly as LABA are specifically meant to achieve bronchodilation, and participants were generally selected if they demonstrated significant bronchodilator reversibility. In contrast, compared with the combination of LABA + ICS, the combination of montelukast + ICS was found superior for preventing exercise-induced % fall in FEV1 in children and adults with asthma.

Other outcomes such as rescue-free days, rescue medication use, asthma quality of life, symptom-free days, daytime symptom score, number and percentage of night-time awakenings, participant satisfaction with treatment and number of participants who switched their allocated medication also reported statistically significant better values with LABA. Moreover, significantly fewer participants allocated to the combination of LABA + ICS withdrew from the study for any reason. Yet, for most secondary outcomes (other than lung function, participant treatment satisfaction and participants switching to other treatments), the magnitude of the observed difference appeared modest. No group difference was observed in the risk of withdrawals due to poor asthma control and hospitalisation.

Only two trials examined the impact of both strategies on inflammatory markers, namely, serum eosinophils. The addition of a LTRA to ICS was associated with greater (4%) reduction in serum eosinophils from baseline when compared with LABA + ICS.

The risk of SAE just failed to achieve a statistically significant group difference between LABA and LTRA (3.1% with LABA vs 2.4% with LTRA—a risk difference of 0.7%). SAE associated with LABA have raised concerns, so we performed a post hoc analysis to explore whether the use of single or separate devices for delivering LABA + ICS and the age of participants influenced the risk of SAE. The increased risk may be limited to the studies using two separate inhalers to deliver LABA + ICS, but again, the subgroup differences were not statistically significant. Yet, the trend is concordant with accumulating evidence of an increased risk of SAE in participants using LABA without ICS and is possibly mediated by non-compliance with concurrent ICS (Perera 2003). Similarly, adults, but not children, showed a significant group difference with higher numbers of SAE with LABA over LTRA; however, in view of the large confidence interval around the paediatric estimate, no firm conclusion on age can be drawn.

No statistically significant group difference was observed in the risk of cardiovascular events, headaches and osteopenia or osteoporosis. Only the risk of oral moniliasis was significantly higher in the LABA group than in the LTRA group, although the risk difference was clinically small (1% for LABA in comparison with 0.5% for LTRA). The risk of overall adverse events was similar in the two groups, meeting our a priori definition of equivalence and suggesting a similar overall safety profile of the two treatment options. No difference was noted between LABA and LTRA in withdrawals due to adverse effects.

One of the entry criteria common to all included trials was the need to demonstrate significant reversibility in FEV1 (≥ 12% improvement postbronchodilation). It is possible that the requirement to demonstrate significant reversibility with beta2 SABA resulted in the selection of participants who were more likely to show a response on lung function outcomes. This may explain the greater differences in favour of LABA that were observed with measures of lung function compared with other outcomes. Although reversibility to a bronchodilator is one of the standard diagnostic criteria of asthma (BTS 2012; GINA 2012; NAEPP 2011; Lougheed 2012), only a minority of asthmatic participants display significant reversibility at a given point in time (Storms 2003). It is quite possible that the selection of participants with significant reversibility has favoured the combination of LABA + ICS over LTRA + ICS.

The major limitation of relevant studies in this area is the strikingly low number of studies examining the best step three in children. Only one of two paediatric trials reported the number of exacerbations requiring the use of rescue oral corticosteroids (Lemanske 2010). Several other paediatric trials were excluded, as they tested add-on therapy in children still on step one or in those already on step three (Karaman 2007; Miraglia del Giudice 2007; Stelmach 2007; Stelmach 2008). As wide variations in the definition of exacerbations have been identified as an important difficulty in comparing data across studies, reporting of data restricted to the use of rescue oral corticosteroids appeared important. Although this review could not provide conclusions regarding the best therapy, many asthma guidelines still recommend LABA as preferred add-on therapy in children. In view of the potential harms associated with LABA, it is particularly critical that future paediatric studies carefully examine the best option as add-on therapy for those who remain inadequately controlled when taking ICS alone using direct comparison of adjunct (LABA vs LTRA) therapy.

The relative homogeneity of adult trials may limit the applicability of the results in individuals older than 65 years of age, smokers and those with asthma with no significant reversibility to SABA. Moreover, how well these add-on therapies perform when added to doses of ICS outside the range of those we have reviewed remains uncertain. Inadequate documentation and reporting also limit the generalisability of results to adolescents. We recommend that trialists including adolescents specify the number of teenagers included and perform subgroup analyses on this age group to begin to address this considerable gap in knowledge. An individual data meta-analysis might provide critical information to reveal whether the presence of allergic rhinitis modifies the observed superiority of LABA over LTRA as add-on therapy to ICS.

With well-documented decreases in adherence over time (Storms 2003), one wonders whether an undocumented lack of compliance may have affected the results. Was there poor adherence to a twice-daily regimen for LABA? With a flat dose-response curve to ICS (Powell 2003), one may even wonder whether similar improvement observed with LTRA and LABA is derived from enhanced compliance with ICS per se as a result of study participation, rather than the selected add-on therapies. Is the greater improvement associated with concomitant rather than separate delivery of LABA and ICS mostly attributable to better lung deposition of, and interaction between, the two drugs (Buhl 2003; Rosenhall 2003) or better adherence with ICS? In the absence of adherence measures, these questions remain unanswered. The perception of more rapid and greater benefit for participants with LABA is often regarded by clinicians as justification for selecting LABA over LTRA as add-on therapy. In the three trials reporting satisfaction, significantly more participants were satisfied in the group receiving LABA + ICS (85%) than in the group given LTRA + ICS (76%), and 27% of participants treated with the combination of LTRA + ICS switched to alternative treatment options over the period of two years, which indicates better satisfaction or clinical benefit with LABA + ICS.

Although derived from close to 6000 adult participants in eight trials, the direction of results may be influenced by participant selection. It is possible that a differential effect of add-on options may be influenced by age, airway reversibility, smoking status, severity of baseline airway obstruction, type of asthma (eosinophilic vs non-eosinophilic), triggers (such as allergic rhinitis), adherence, etc. Future studies should focus on comparing these add-on strategies in selected groups of participants, so that characteristics of responders to either option may be better delineated. Measures of adherence (before and after randomisation) should also be incorporated into the design of future studies.

Study results apply predominantly to adult asthmatic patients who remain symptomatic despite predominantly low doses of ICS and who present with a moderate (baseline FEV1 65% to 75% of predicted value) reversible airway obstruction. Data are insufficient to allow firm conclusions regarding children and adults over 65 years of age.

The extensive search strategy and assessment of methodological quality and data extraction provided by two review authors yielded the high-quality outputs. The high methodological quality of all trials contributing data and the confirmation of methodology and extracted data by review authors or of study sponsors for studies contributing to the primary outcome strengthen our findings.

Authors' conclusions

Implications for practice

In asthmatic adults with persistent mild or moderate airway obstruction despite low or medium doses of ICS who demonstrate significant reversibility to a short-acting bronchodilator, the absolute risk of an exacerbation requiring oral corticosteroids in patients is 2% lower if patients are treated with LABA rather than LTRA. This finding was compatible with an NNTB of 62 over four to 102 weeks. Compared with LTRA, the addition of LABA to ICS is associated with statistically significant greater improvement in lung function, symptom-free days, use of rescue beta2-agonists, symptoms, symptom-free days, night-time awakenings and quality of life, although group differences in several markers are generally modest. The findings are indirectly supported by higher participant satisfaction and lower numbers of participants switching to the alternative treatment with LABA over LTRA as adjunct to ICS. Findings suggest that LABA increase the risk of SAE from 2.4% to 3.1% in adults when compared with LTRA, but the data are much more uncertain in children. These findings would support the use of a single inhaler as the preferred device for delivery of LABA + ICS.

Implications for research

Future trials should address the main gaps in knowledge, namely, the generalisability of results to the following.

  • Children, adolescents and elderly individuals.

  • Patients with severe (or milder) airway obstruction.

  • Asthmatic patients with minimal or no (< 12%) airway reversibility to bronchodilators at the time of enrolment but with positive provocation challenge or other convincing criteria of the diagnosis of asthma.

  • Patients with co-morbidities, such as allergic rhinitis or aspirin-induced asthma, smokers and those having environmental exposure to cigarette smoke, etc.

  • Add-on therapy to higher doses of ICS than 200 to 280 µg/d of hydrofluoroalkane-beclomethasone dipropionate (HFA-BDP), or equivalent.

  • Monitoring of adherence to both combination therapies.

  • Use of single inhalers for delivery of LABA + ICS compared with LTRA + ICS.-

  • Measurement and reporting of the impact of each adjunct therapy on inflammatory markers (preferably using induced sputum) and airway hyperresponsiveness over time.

  • Careful monitoring and reporting of outcomes that are important to the patient, particularly exacerbations requiring systemic corticosteroids or hospital admission, symptoms, symptom-free days, night-time awakenings, quality of life, satisfaction, number of participants switching to alternative treatment and life-threatening asthma, as defined by admission to the intensive care unit (ICU) or requirement for intubation or ventilation.

Acknowledgements

We are indebted to Emma Welsh, Elizabeth Stovold, Susan Hansen and Veronica Stewart from the Cochrane Airways Review Group editorial base for providing extensive ongoing support with identifying, retrieving and translating the literature.

We acknowledge the contribution of Felix Ram to the first iteration of this review. We are indebted to the following individuals, who replied to our request for confirmation of methodology and data extraction and graciously provided additional data when possible: Karen Richardson and Inge Vestbo from GlaxoSmithKline, UK; Ian Naya from AstraZeneca, Sweden; Nitesh Shah and Graham Debney from AstraZeneca, UK; and Peter Polos and Steven Bird from Merck Frosst, USA.

CRG funding acknowledgement: The National Institute for Health Research (NIHR) is the largest single sponsor of the Cochrane Airways Review Group.

Disclaimer: The views and opinions expressed herein are those of the review authors and do not necessarily reflect those of the NIHR, the NHS or the Department of Health.

Data and analyses

Download statistical data

Comparison 1. Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Participants with one or more exacerbations requiring systemic corticosteroids86257Risk Ratio (M-H, Fixed, 95% CI)0.87 [0.76, 0.99]
1.1 Adults75923Risk Ratio (M-H, Fixed, 95% CI)0.87 [0.76, 1.00]
1.2 Children1334Risk Ratio (M-H, Fixed, 95% CI)0.84 [0.53, 1.34]
2 Morning PEF: L/min change from baseline11 Mean Difference (Random, 95% CI)15.36 [11.35, 19.37]
2.1 Adults11 Mean Difference (Random, 95% CI)15.36 [11.35, 19.37]
3 Evening PEF: L/min change from baseline10 Mean Difference (Random, 95% CI)12.64 [10.11, 15.17]
3.1 Adults10 Mean Difference (Random, 95% CI)12.64 [10.11, 15.17]
4 FEV1: L, % change from baseline10 Mean Difference (Random, 95% CI)0.08 [0.05, 0.10]
4.1 Adults10 Mean Difference (Random, 95% CI)0.08 [0.05, 0.10]
5 % fall in FEV1 postexercise2358Mean Difference (IV, Fixed, 95% CI)3.07 [0.68, 5.47]
5.1 Adults170Mean Difference (IV, Fixed, 95% CI)2.8 [-1.48, 7.08]
5.2 Children1288Mean Difference (IV, Fixed, 95% CI)3.20 [0.31, 6.09]
6 FEV1: L, % change from baseline1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
6.1 Adults1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
7 FEV1: % of predicted end of treatment1 Mean Difference (IV, Random, 95% CI)Totals not selected
7.1 Adults1 Mean Difference (IV, Random, 95% CI)0.0 [0.0, 0.0]
8 Rescue-free days: % change from baseline52612Mean Difference (IV, Random, 95% CI)9.18 [5.39, 12.98]
8.1 Adults52612Mean Difference (IV, Random, 95% CI)9.18 [5.39, 12.98]
9 Rescue medication use: puffs/d change from baseline74055Mean Difference (IV, Random, 95% CI)-0.49 [-0.75, -0.24]
9.1 Adults74055Mean Difference (IV, Random, 95% CI)-0.49 [-0.75, -0.24]
10 Global asthma QoL AQLQ score (higher is better): change from baseline43243Std. Mean Difference (IV, Random, 95% CI)0.12 [0.05, 0.19]
10.1 Adults43243Std. Mean Difference (IV, Random, 95% CI)0.12 [0.05, 0.19]
11 Symptom-free days: % change from baseline6 Mean Difference (Random, 95% CI)7.09 [3.66, 10.52]
11.1 Adults6 Mean Difference (Random, 95% CI)7.09 [3.66, 10.52]
12 Daytime symptom scores (high is worse): change from baseline53823Std. Mean Difference (IV, Fixed, 95% CI)-0.18 [-0.25, -0.12]
12.1 Adults53823Std. Mean Difference (IV, Fixed, 95% CI)-0.18 [-0.25, -0.12]
13 Morning symptoms: change from baseline1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
13.1 Adults1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
14 Night-time symptom score (five-point scale, higher score is worse): change from baseline1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
14.1 Adults1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
15 Change in number of night awakenings per week: change from baseline44214Mean Difference (IV, Fixed, 95% CI)-0.12 [-0.19, -0.06]
15.1 Adults44214Mean Difference (IV, Fixed, 95% CI)-0.12 [-0.19, -0.06]
16 Change in % of nights with no awakenings per week: change from baseline2673Mean Difference (IV, Fixed, 95% CI)6.89 [2.87, 10.91]
16.1 Adults2673Mean Difference (IV, Fixed, 95% CI)6.89 [2.87, 10.91]
17 Rescue-free nights (%): change from baseline1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
17.1 Adults1 Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]
18 Withdrawals for any reason136960Risk Ratio (M-H, Fixed, 95% CI)0.84 [0.74, 0.96]
18.1 Adults126652Risk Ratio (M-H, Fixed, 95% CI)0.84 [0.74, 0.95]
18.2 Children1308Risk Ratio (M-H, Fixed, 95% CI)1.25 [0.34, 4.57]
19 Withdrawals due to adverse events126591Risk Ratio (M-H, Fixed, 95% CI)1.01 [0.79, 1.29]
19.1 Adults116291Risk Ratio (M-H, Fixed, 95% CI)1.01 [0.79, 1.29]
19.2 Children1300Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
20 Withdrawals due to poor asthma control/asthma exacerbation85354Risk Ratio (M-H, Random, 95% CI)0.87 [0.49, 1.56]
20.1 Adults85354Risk Ratio (M-H, Random, 95% CI)0.87 [0.49, 1.56]
21 Participants with one or more exacerbations requiring hospital admission64675Risk Ratio (M-H, Fixed, 95% CI)1.34 [0.68, 2.67]
21.1 Adults54345Risk Ratio (M-H, Fixed, 95% CI)1.37 [0.67, 2.79]
21.2 Children1330Risk Ratio (M-H, Fixed, 95% CI)1.0 [0.06, 15.85]
22 Serious adverse events96288Risk Ratio (M-H, Fixed, 95% CI)1.33 [0.99, 1.79]
22.1 Adults75658Risk Ratio (M-H, Fixed, 95% CI)1.35 [1.00, 1.82]
22.2 Children2630Risk Ratio (M-H, Fixed, 95% CI)1.0 [0.20, 4.91]
23 Death1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
23.1 Adults1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
24 Headache106187Risk Ratio (M-H, Fixed, 95% CI)1.07 [0.90, 1.26]
24.1 Adults106187Risk Ratio (M-H, Fixed, 95% CI)1.07 [0.90, 1.26]
25 Cardiovascular events55163Risk Ratio (M-H, Fixed, 95% CI)1.09 [0.77, 1.53]
25.1 Adults55163Risk Ratio (M-H, Fixed, 95% CI)1.09 [0.77, 1.53]
26 Oral moniliasis65203Risk Ratio (M-H, Fixed, 95% CI)1.86 [1.00, 3.44]
26.1 Adults65203Risk Ratio (M-H, Fixed, 95% CI)1.86 [1.00, 3.44]
27 Osteopenia/osteoporosis22963Risk Ratio (M-H, Fixed, 95% CI)0.56 [0.12, 2.63]
27.1 Adults22963Risk Ratio (M-H, Fixed, 95% CI)0.56 [0.12, 2.63]
28 Elevated liver enzymes1 Risk Ratio (M-H, Fixed, 95% CI)Totals not selected
28.1 Adults1 Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
29 Overall adverse events106277Risk Ratio (M-H, Fixed, 95% CI)1.02 [0.99, 1.06]
29.1 Adults95977Risk Ratio (M-H, Fixed, 95% CI)1.03 [0.99, 1.07]
29.2 Children1300Risk Ratio (M-H, Fixed, 95% CI)0.79 [0.47, 1.31]
30 Participants' treatment satisfaction32020Risk Ratio (M-H, Random, 95% CI)1.12 [1.04, 1.20]
30.1 Adults32020Risk Ratio (M-H, Random, 95% CI)1.12 [1.04, 1.20]
31 Participants switched/received additional treatment1342Risk Ratio (M-H, Random, 95% CI)0.01 [0.00, 0.16]
31.1 Adults1342Risk Ratio (M-H, Random, 95% CI)0.01 [0.00, 0.16]
32 Change from baseline in serum eosinophils (× 109/L)22787Mean Difference (IV, Fixed, 95% CI)0.04 [0.02, 0.05]
32.1 Adults22787Mean Difference (IV, Fixed, 95% CI)0.04 [0.02, 0.05]
Analysis 1.1.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 1 Participants with one or more exacerbations requiring systemic corticosteroids.

Analysis 1.2.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 2 Morning PEF: L/min change from baseline.

Analysis 1.3.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 3 Evening PEF: L/min change from baseline.

Analysis 1.4.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 4 FEV1: L, % change from baseline.

Analysis 1.5.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 5 % fall in FEV1 postexercise.

Analysis 1.6.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 6 FEV1: L, % change from baseline.

Analysis 1.7.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 7 FEV1: % of predicted end of treatment.

Analysis 1.8.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 8 Rescue-free days: % change from baseline.

Analysis 1.9.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 9 Rescue medication use: puffs/d change from baseline.

Analysis 1.10.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 10 Global asthma QoL AQLQ score (higher is better): change from baseline.

Analysis 1.11.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 11 Symptom-free days: % change from baseline.

Analysis 1.12.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 12 Daytime symptom scores (high is worse): change from baseline.

Analysis 1.13.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 13 Morning symptoms: change from baseline.

Analysis 1.14.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 14 Night-time symptom score (five-point scale, higher score is worse): change from baseline.

Analysis 1.15.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 15 Change in number of night awakenings per week: change from baseline.

Analysis 1.16.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 16 Change in % of nights with no awakenings per week: change from baseline.

Analysis 1.17.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 17 Rescue-free nights (%): change from baseline.

Analysis 1.18.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 18 Withdrawals for any reason.

Analysis 1.19.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 19 Withdrawals due to adverse events.

Analysis 1.20.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 20 Withdrawals due to poor asthma control/asthma exacerbation.

Analysis 1.21.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 21 Participants with one or more exacerbations requiring hospital admission.

Analysis 1.22.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 22 Serious adverse events.

Analysis 1.23.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 23 Death.

Analysis 1.24.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 24 Headache.

Analysis 1.25.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 25 Cardiovascular events.

Analysis 1.26.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 26 Oral moniliasis.

Analysis 1.27.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 27 Osteopenia/osteoporosis.

Analysis 1.28.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 28 Elevated liver enzymes.

Analysis 1.29.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 29 Overall adverse events.

Analysis 1.30.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 30 Participants' treatment satisfaction.

Analysis 1.31.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 31 Participants switched/received additional treatment.

Analysis 1.32.

Comparison 1 Long-acting β2-agonists + ICS versus leukotriene receptor antagonists + ICS, Outcome 32 Change from baseline in serum eosinophils (× 109/L).

Comparison 2. Subgroup and sensitivity analyses
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Participants with one or more exacerbations requiring systemic corticosteroids: number of inhaler devices86257Risk Ratio (M-H, Fixed, 95% CI)0.87 [0.76, 0.99]
1.1 Single device for LABA + ICS31586Risk Ratio (M-H, Fixed, 95% CI)0.64 [0.45, 0.91]
1.2 Two devices for LABA + ICS44319Risk Ratio (M-H, Fixed, 95% CI)0.89 [0.75, 1.04]
1.3 Single device or two separate devices for LABA + ICS1352Risk Ratio (M-H, Fixed, 95% CI)1.06 [0.80, 1.41]
2 Participants with one or more exacerbations requiring systemic corticosteroids: dose of ICS86257Risk Ratio (M-H, Fixed, 95% CI)0.87 [0.76, 0.99]
2.1 Low dose of ICS43076Risk Ratio (M-H, Fixed, 95% CI)0.81 [0.66, 0.98]
2.2 Medium dose of ICS11452Risk Ratio (M-H, Fixed, 95% CI)0.85 [0.67, 1.08]
2.3 Mixed1948Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.50, 1.96]
2.4 Unclear2781Risk Ratio (M-H, Fixed, 95% CI)1.04 [0.79, 1.37]
3 Participants with one or more exacerbations requiring systemic corticosteroids: study duration86257Risk Ratio (M-H, Random, 95% CI)0.88 [0.77, 1.01]
3.1 12 weeks or less42629Risk Ratio (M-H, Random, 95% CI)0.66 [0.43, 1.02]
3.2 ≥ 48 weeks43628Risk Ratio (M-H, Random, 95% CI)0.92 [0.80, 1.05]
4 Participants with one or more exacerbations requiring systemic corticosteroids: LTRA86257Risk Ratio (M-H, Fixed, 95% CI)0.87 [0.76, 0.99]
4.1 Montelukast 5 mg or 10 mg once daily65476Risk Ratio (M-H, Fixed, 95% CI)0.83 [0.72, 0.97]
4.2 Zafirlukast 20 mg twice daily1429Risk Ratio (M-H, Fixed, 95% CI)0.86 [0.29, 2.52]
4.3 Monterlukast 10 mg once daily or zafirlukast 20 mg twice daily1352Risk Ratio (M-H, Fixed, 95% CI)1.06 [0.80, 1.41]
5 Serious adverse events stratified by number of inhaler devices used for LABA + ICS96288Risk Ratio (M-H, Fixed, 95% CI)1.33 [0.99, 1.79]
5.1 Single device for LABA + ICS41648Risk Ratio (M-H, Fixed, 95% CI)0.82 [0.35, 1.94]
5.2 Two devices for LABA + ICS54640Risk Ratio (M-H, Fixed, 95% CI)1.42 [1.04, 1.95]
6 Serious adverse events stratified by age of participants (adults vs children)96288Risk Ratio (M-H, Fixed, 95% CI)1.33 [0.99, 1.79]
6.1 Adults75658Risk Ratio (M-H, Fixed, 95% CI)1.35 [1.00, 1.82]
6.2 Children2630Risk Ratio (M-H, Fixed, 95% CI)1.0 [0.20, 4.91]
Analysis 2.1.

Comparison 2 Subgroup and sensitivity analyses, Outcome 1 Participants with one or more exacerbations requiring systemic corticosteroids: number of inhaler devices.

Analysis 2.2.

Comparison 2 Subgroup and sensitivity analyses, Outcome 2 Participants with one or more exacerbations requiring systemic corticosteroids: dose of ICS.

Analysis 2.3.

Comparison 2 Subgroup and sensitivity analyses, Outcome 3 Participants with one or more exacerbations requiring systemic corticosteroids: study duration.

Analysis 2.4.

Comparison 2 Subgroup and sensitivity analyses, Outcome 4 Participants with one or more exacerbations requiring systemic corticosteroids: LTRA.

Analysis 2.5.

Comparison 2 Subgroup and sensitivity analyses, Outcome 5 Serious adverse events stratified by number of inhaler devices used for LABA + ICS.

Analysis 2.6.

Comparison 2 Subgroup and sensitivity analyses, Outcome 6 Serious adverse events stratified by age of participants (adults vs children).

Comparison 3. MD archive from previous review version
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Morning PEF (L/min): change from baseline10 Mean Difference (IV, Fixed, 95% CI)Totals not selected
2 Evening PEF (L/min): change from baseline9 Mean Difference (IV, Fixed, 95% CI)Totals not selected
3 FEV1 (L): change from baseline8 Mean Difference (IV, Fixed, 95% CI)Totals not selected
4 Symptom-free days (%): change from baseline5 Mean Difference (IV, Random, 95% CI)Totals not selected
Analysis 3.1.

Comparison 3 MD archive from previous review version, Outcome 1 Morning PEF (L/min): change from baseline.

Analysis 3.2.

Comparison 3 MD archive from previous review version, Outcome 2 Evening PEF (L/min): change from baseline.

Analysis 3.3.

Comparison 3 MD archive from previous review version, Outcome 3 FEV1 (L): change from baseline.

Analysis 3.4.

Comparison 3 MD archive from previous review version, Outcome 4 Symptom-free days (%): change from baseline.

Appendices

Appendix 1.  Sources and search methods for the Cochrane Airways Group Specialised Register (CAGR)

Electronic searches: core databases

Database Frequency of search
CENTRALMonthly
MEDLINE (Ovid)Weekly
EMBASE (Ovid)Weekly
PsycINFO (Ovid)Monthly
CINAHL (EBSCO)Monthly
AMED (EBSCO)Monthly

 

Handsearches: core respiratory conference abstracts

Conference Years searched
American Academy of Allergy, Asthma and Immunology (AAAAI)2001 onwards
American Thoracic Society (ATS)2001 onwards
Asia Pacific Society of Respirology (APSR)2004 onwards
British Thoracic Society Winter Meeting (BTS)2000 onwards
Chest Meeting2003 onwards
European Respiratory Society (ERS)1992, 1994, 2000 onwards
International Primary Care Respiratory Group Congress (IPCRG)2002 onwards
Thoracic Society of Australia and New Zealand (TSANZ)1999 onwards

 

MEDLINE search strategy used to identify trials for the CAGR

Asthma search

1. exp Asthma/

2. asthma$.mp.

3. (antiasthma$ or anti-asthma$).mp.

4. Respiratory Sounds/

5. wheez$.mp.

6. Bronchial Spasm/

7. bronchospas$.mp.

8. (bronch$ adj3 spasm$).mp.

9. bronchoconstrict$.mp.

10. exp Bronchoconstriction/

11. (bronch$ adj3 constrict$).mp.

12. Bronchial Hyperreactivity/

13. Respiratory Hypersensitivity/

14. ((bronchial$ or respiratory or airway$ or lung$) adj3 (hypersensitiv$ or hyperreactiv$ or allerg$ or insufficiency)).mp.

15. ((dust or mite$) adj3 (allerg$ or hypersensitiv$)).mp.

16. or/1-15

Filter to identify RCTs

1. exp "clinical trial [publication type]"/

2. (randomised or randomised).ab,ti.

3. placebo.ab,ti.

4. dt.fs.

5. randomly.ab,ti.

6. trial.ab,ti.

7. groups.ab,ti.

8. or/1-7

9. Animals/

10. Humans/

11. 9 not (9 and 10)

12. 8 not 11

The MEDLINE strategy and RCT filter are adapted to identify trials in other electronic databases.

Appendix 2. GSK randomisation procedures

The procedures for randomising GSK-sponsored studies have been detailed in correspondence between Richard Follows and TL, the details of which are given below.

The randomisation software is a computer-generated, centralised programme (RandAll). After verification that the randomisation sequence is suitable for the study design (cross-over, block or stratification), Clinical Supplies then packages the treatments according to the randomisation list generated. Concealment of allocation is maintained by a third party because the sites phone in and are allocated treatments on that basis. Alternatively, a third party may dispense the drug at the sites. Unblinding of data for interim analyses can be done only through RandAll, and this is restricted so that only those reviewing the data are unblinded to treatment group allocation.

What's new

DateEventDescription
13 December 2012New search has been performedNew literature search run.
13 December 2012New citation required but conclusions have not changedOne new study added. Summary of findings table added. Plain language summary updated according to the new guidelines. Change in author team: TL and CJC stepped down and BFC stepped in.

History

Protocol first published: Issue 3, 2003
Review first published: Issue 1, 2005

DateEventDescription
5 July 2011AmendedClarification made to abstract regarding subgroup analysis.
2 February 2011New citation required and conclusions have changed

Full risk of bias assessment has been incorporated into the review.

Data on secondary outcomes were provided by the new studies. Serious adverse events are more common with LABA and ICS than LTRA and ICS. This result is not definitive and could be influenced by separate administration of LABA and ICS.

17 March 2010New search has been performedLiterature search re-run. Three new studies were included (Lemanske 2010; Pavord 2007; Price 2011). One previously included study was excluded (Stelmach 2008); this study had not contributed data to the primary outcome.
4 August 2008AmendedConverted to new review format.
20 June 2006New citation required and conclusions have changedFive new studies met the entry criteria of the review (Ceylan 2004; Grosclaude 2003; Hendeles 2004; Stelmach 2008; Storms 2004). Of these, two studies contributed data to this updated review. The additional data did not alter the conclusions of the review.

Contributions of authors

Bhupendrasinh Chauhan (update 2013) assessed studies for inclusion or exclusion, extracted and entered data, assessed methodological quality, revised results and discussion sections and drafted the response to comments and the final version of the Cochrane review.

Francine M Ducharme reviewed the protocol design; supervised the literature search; reviewed all citations; participated in the selection of trials, methodology assessment, and data extraction; corresponded with study authors and pharmaceutical companies to identify other relevant trials, verify methodology and data extraction and request additional information; analysed and interpreted results of the meta-analysis; and edited and approved the final review.

Toby Lasserson (update 2006, 2010) assessed studies for inclusion or exclusion, extracted and entered data, revised results and discussion sections and solicited additional data from study authors.

Christopher Cates edited the review, checked the methodology and contributed to writing of the final review.

Felix Ram participated in the initial version of the review (2005): reviewed the protocol design, identified and reviewed the full-text publication of all citations of eligible or potentially eligible RCTs, extracted the methodology and data, analysed and interpreted results of the meta-analysis and wrote the first draft of the review.

Declarations of interest

Bhupendrasinh Chauhan holds a postdoctoral scholarship from one of Dr Ducharme's grants from the Canadian Institute of Health Research and has no conflicts of interest.

Francine M Ducharme has received travel support for meeting attendance, research funds, fees for speaking and consulting fees from Merck Frosst Inc (producer of montelukast), GlaxoSmithKline (producer of fluticasone, beclomethasone, salmeterol), Novartis (producer of formoterol) and/or Nycomed (producer of combination of mometasone and formoterol).

Sources of support

Internal sources

  • No sources of support supplied

External sources

  • Nederlands Asthma Fonds, Netherlands.

  • Francine M Ducharme, Canada.

  • NIHR, UK.

    Funder

Differences between protocol and review

  • In the protocol published in April 2001, we had planned to examine the impact of the ICS (subgroups: three) and of baseline severity (subgroups: five) as sensitivity analyses but changed to subgroup analyses because this enhanced the clarity of interpretation. The last two subgroup analyses (six and seven), not initially considered in 2001, were added to the list subsequently, as recent data indicated that the number of inhaler devices used to deliver LABA + ICS might be an important effect modifier (Nelson 2003), and a peer reviewer suggested that differences might not be the same over 12 weeks and 48 weeks.

  • Because of recent concerns over the association between LABA and SAE, we included a subgroup analysis of data by number of inhaler devices.

  • Study assessment was amended to reflect changes in the recommended approach to risk of bias evaluation. In the original protocol and in the first version of the review, we assessed studies by using the Jadad Scale and by grading concealment of allocation. For the 2010 update, we used a tool to assess the degree of protection offered by the study design against systematic error. This is outlined in the section Assessment of risk of bias in included studies.

  • For the 2013 update, we raised the threshold for performing a random-effects analysis from I2 = 25% to I2 = 40%

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Bjermer 2003

MethodsParallel-group, multi-centre trial (148 centres in 37 countries)
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE

LTRA: 638 ± 285 µg of beclomethasone-equivalent/d

LABA: 647 ± 229 µg of beclomethasone-equivalent/d

RANDOMLY ASSIGNED: 1490 (LTRA: 747; LABA: 743)

WITHDRAWALS

LTRA: 125 (17%)

LABA: 110 (15%)

AGE in years: mean ± SD

LTRA: 41.2 ± 13.6

LABA: 41.0 ± 13.7

GENDER (% male)

LTRA: 45.4

LABA: 44.8

SEVERITY
MODERATE asthma

BASELINE % of PREDICTED FEV1 (L)

LTRA: 71.3 ± 13.2

LABA: 72.7 ± 13.9

ALLERGIC RHINITIS (%)

LTRA: 61.7

LABA: 60.4

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years: mean ± SD

LTRA: 16.3 ± 13.0

LABA: 16.2 ± 12.7

ELIGIBILITY CRITERIA: age 15 to 72 years; clear history of chronic asthma for at least one year; regular use of ICS over eight weeks before study entry; FEV1 values between 50% and 90% of predicted; ≥ 12% improvement in FEV1 or PEFR after β-agonists; minimum predetermined level of daytime and night-time inhaled short-acting β-agonist use (≥ one puff/d); minimum asthma symptom score (biweekly score ≥ 56 on a scale of zero to 336); current treatment includes only short-acting beta2-agonists and ICS (200 to 1000 µg/d or equivalent); women with negative urine pregnancy test at screening

EXCLUSION CRITERIA: emergency treatment for asthma within one month of first visit; hospitalisation for asthma within three months; unresolved upper respiratory tract infection within three weeks; active sinus infection; received the following asthma medications: oral corticosteroids within one month; cromolyn, nedocromil, leukotriene-receptor antagonists, long-acting or oral β-agonists, inhaled anticholinergics within two weeks; theophylline, terfenadine, fexofenadine, loratadine or cetirizine within one week

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION

Run-in period: four weeks

Intervention period: 48 weeks

INTERVENTION GROUP ONE

LTRA: montelukast @ 10 mg/d PO + ICS (FP 100 µg bid via discus)

INTERVENTION GROUP TWO

LABA: salmeterol 50 μg bid via MDI + ICS (FP 100 µg bid via discus)

Two inhalers used for combination therapy

CO-TREATMENT: none

Outcomes

INTENTION-TO-TREAT ANALYSES
Outcomes used at endpoint or at 48 weeks

PULMONARY FUNCTION TESTS
Change from baseline FEV; change from baseline in AM PEFR

SYMPTOM SCORES
Change from baseline NIGHT-TIME awakenings

*EXACERBATIONS
Exacerbations requiring systemic corticosteroids; exacerbations requiring hospital admission; exacerbations requiring unscheduled office visit; exacerbations requiring emergency room visit; time to first exacerbation

Definition: An asthma attack was defined by one or all of the following: hospitalisation; unscheduled office visit; ER visit; CS use (oral, IM, IV or rectal use)

FUNCTIONAL STATUS
Change in quality of life; change in night-time awakenings

INFLAMMATORY MARKERS
Change in serum eosinophils

ADVERSE EFFECTS
Elevated liver enzymes, headache, nausea, death, neutropenia, increased lymphocytes

WITHDRAWALS
Due to adverse effects; due to poor control; overall

(* denotes primary outcome)

Notes

Full-text report; additional unpublished data provided by Peter Polos, June 2003

Funder: Merck Frosst

Confirmation of methodology and data extraction: received (Peter Polos, June 2003)

User-defined number: 48 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomisation by computer-generated random numbers
Allocation concealment (selection bias)Low riskAllocation occurred at pharmacy and was not conducted by investigator
Blinding (performance bias and detection bias)
All outcomes
Low risk

Number-coded MDI/tablets supplied by pharmacy

Triple-blind (participant, assessor and treating physician); double-dummy (identical placebo)

Incomplete outcome data (attrition bias)
Exacerbations
Unclear riskIntention-to-treat analysis reported, but method not available
Selective reporting (reporting bias)Low riskData obtained from investigators
Other biasLow riskNo risk of bias was observed

Ceylan 2004

MethodsParallel-group; single-centre study (Turkey)
Participants

INADEQUATELY controlled participants taking ICS and SABA prn

BASELINE ICS DOSAGE
Not reported (400 µg/d BUD given as standard during 4 week run-in period)

RANDOMLY ASSIGNED: 48. NB: baseline data reported only for those who completed the study: LTRA: 20; LABA: 20

WITHDRAWALS
Not stated by treatment group

AGE in years, mean
LTRA: 33.2
LABA: 39.1

GENDER (% male)
LTRA: 55
LABA: 50

SEVERITY
Moderate persistent asthma

BASELINE % OF PREDICTED FEV1 (L)
LTRA: 69.7
LABA: 71.2

ALLERGIC RHINITIS (%)
LTRA: 60
LABA: 70

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years: mean ± SD
LTRA: 8.1 ± 4
LABA: 9 ± 8.8

ELIGIBILITY CRITERIA: age 15 to 60 years; diagnosis of asthma (GINA); persistent asthma symptoms for at least 1 year; use of ICS for at least 6 months; post-run-in period: FEV1 or PEF ≥ 60% and ≤ 80% of predicted
≥ 15% reversibility increase in FEV1; mean AM PEF value ≤ 85% max after SABA; use of SABA ≥ 2 times per day or AM/night symptom score ≥ 2 on 4 or fewer days per week

EXCLUSION CRITERIA: smokers; pregnant or lactating women; life-threatening asthma; patients hospitalised for asthma in last 3 months; upper/lower RTI

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: 4 weeks
Intervention period: 8 weeks
Outcomes at 4, 8, and 12 weeks

INTERVENTION GROUP 1
LTRA = montelukast @ 10 mg/d PO + BUD 200 µg bid, unclear inhaler device

INTERVENTION GROUP 2
LABA = formoterol 12 µg bid + BUD 200 µg bid, unclear inhaler device

CO-TREATMENT
SABA prn

Outcomes

PULMONARY FUNCTION TESTS
*Change in AM PEF

Change in PM PEF

Change in FEV1 %predicted

Change in FEV1 (L)

SYMPTOM SCORES
Morning symptom scores; night symptom scores

EXACERBATIONS
Not reported (participants whose condition was exacerbated were excluded from the study)

FUNCTIONAL STATUS
Rescue medication usage (puffs/d); % of days without rescue medication usage

INFLAMMATORY MARKERS
Not reported

EXACERBATIONS
Need for a drug not included in the protocol

ADVERSE EFFECTS
Candidiasis; sore throat; voice problems; headache

WITHDRAWALS
Not clear

Due to ADVERSE EFFECTS
Not reported

Due to poor control
Not reported

Overall
Stated

(* denotes primary outcome)

Notes

Full-text report

No funding body

User-defined number: 8 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot reported
Allocation concealment (selection bias)Unclear riskNot reported
Blinding (performance bias and detection bias)
All outcomes
High riskOpen-label study
Incomplete outcome data (attrition bias)
Exacerbations
Unclear riskStated ITT for efficacy and safety; however, baseline data are presented for only 40 participants who completed the study
Selective reporting (reporting bias)Unclear riskUnable to verify whether primary outcome measured in the review
Other biasLow riskNo apparent risk of bias was observed

Fish 2001

MethodsParallel-group study, multi-centre trial (71 centres in USA and Puerto Rico)
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE
84 to 4000 µg of beclomethasone-equivalent/d

RANDOMLY ASSIGNED: 948
LTRA: 472
LABA: 476

WITHDRAWALS
LTRA: 70 (15%)
LABA: 61 (13%)

AGE in years: mean ± SD
LTRA: 39.5 ± 14.0
LABA: 39.9 ± 13.5

GENDER (% male)
LTRA: 38
LABA: 39

SEVERITY
Not described

BASELINE %PREDICTED FEV1 (L)
LTRA: 68.6 (0.4) SE
LABA: 68.1 (0.4) SE

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years (%)
Less than 10 years
LTRA: 26
LABA: 24

Longer than 10 years
LTRA: 74
LABA: 76

ELIGIBILITY CRITERIA: aged ≥ 15 years; male or non-pregnant, non-lactating female; asthma for ≥ 6 months; symptomatic despite ICS for at least 6 weeks before screening; 50% to 80% predicted FEV1; ≥ 12% increase in FEV1 postbronchodilator (200 µg albuterol)

In the 7 to 14 days before randomisation, one or more of the following

  • FEV1 50% to 80% of predicted

  • Average of 4 or more puffs per day albuterol

  • Symptom score of 2 or higher for 3 or more days

  • 3 or more nights when participant awoke at night because of asthma symptoms

EXCLUSION CRITERIA
Not described

SETTING: outpatients in private and university clinics

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION:
Run-in period: 1 to 2 weeks
Intervention period: 12 weeks

INTERVENTION GROUP 1
LTRA: montelukast 10 mg qd + ICS: continued current medication (which included fluticasone, triamcinolone, BDP, BUD and flunisolide) (mean 565 µg in CFC-BDP equivalent)

INTERVENTION GROUP 2
LABA: salmeterol 50 µg bid via discus + ICS: continued current medication
Mean 546 µg in CFC-BDP equivalent

2 inhalers used for combination therapy

CO-TREATMENT: none permitted

Outcomes

INTENTION-TO-TREAT ANALYSES
Outcomes used at endpoint

PULMONARY FUNCTION TESTS
*Change from baseline in AM PEFR; change from baseline in PM PEFR

SYMPTOM SCORES
Change from baseline overall symptom scores; change in symptom-free days; participant satisfaction

EXACERBATION
Definition: any worsening of asthma symptoms requiring treatment beyond the use of blinded study drug and/or supplemental albuterol. Participants who experienced an asthma exacerbation were withdrawn from the study

FUNCTIONAL STATUS
Change from baseline in mean overall use of β2-agonists (puffs/d); change from baseline in mean DAYTIME use of β2-agonists (puffs/d); change from baseline in mean NIGHT-TIME use of β2-agonists (puffs/d); change in rescue-free days; change in night-time awakenings

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Drug-related and non-drug-related

WITHDRAWALS
Due to adverse effects reported

(* denotes primary outcome)

Notes

Full-text report

Received additional unpublished data provided by Karen Richardson, GSK, UK, August 2003

Funded by Glaxo Wellcome, studies SMS40003 and SMS40004

Confirmation of methodology and data extraction received

User-defined order: 12 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskMethods of randomisation: by computer-generated random number
Allocation concealment (selection bias)Low riskMeans of assignment by number-coded inhaler/pills supplied by pharmacy
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-blind, double-dummy design
Incomplete outcome data (attrition bias)
Exacerbations
Low riskReceived additional unpublished data provided by Karen Richardson, GSK, UK, August 2003
Selective reporting (reporting bias)Low riskPrimary outcome data available for meta-analysis
Other biasLow riskNo apparent risk of bias was observed

Fogel 2010

MethodsMulti-centre, double-blind, double-dummy, cross-over study (30 centres in Europe, Asia, Mexico and South America)
Participants

ADEQUATELY controlled participants taking ICS at baseline with at least 70% of predicted value of FEV1

BASELINE ICS DOSAGE
200 µg of fluticasone/d

RANDOMLY ASSIGNED: 154
WITHDRAWALS
LTRA: 3 (1.95%)
LABA: 5 (3.25%)

AGE in years: mean ± SD
LTRA: 10.2 ± 2.0
LABA: 9.8 ± 2.0

GENDER (% male)
LTRA: 55.1
LABA: 60.5

SEVERITY
Not described

BASELINE %PREDICTED FEV1 (L) (SD)
LTRA: 99.9 (32.5)
LABA: 100.5 (15.6)

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years (%)
Not reported

ELIGIBILITY CRITERIA: children aged 6 to 14 years with at least a 1 year clinical history of asthma who were using daily ICS for at least 8 weeks before enrolling in the study. Participants were eligible for randomisation after demonstrating, while taking open-label fluticasone, an FEV1 70% or more of the predicted value and exercise-induced bronchoconstriction, defined as a decrease in FEV1 of at least 15% compared with preexercise baseline FEV1

EXCLUSION CRITERIA: unresolved signs and symptoms of an upper respiratory tract infection within 1 week; evidence of active, clinically significant sinus infection, history of intubation for asthma; short-term asthma therapy in an emergency department, urgent care facility, or office setting within 1 month or hospitalisation for asthma within 3 months; and systemic administration of corticosteroids within 1 month; cromolyn, nedocromil or LTRA within 1 week; theophylline within 1 week or oral beta2-agonists or LABA within 1 week; or astemizole within 3 months before the first study visit

SETTING: outpatient

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION:
Run-in period: 4 weeks
Intervention period: 4 weeks

INTERVENTION GROUP 1
LTRA: montelukast 5 mg qd + fluticasone 50 μg 2 puffs twice daily

INTERVENTION GROUP 2
LABA: salmeterol 50 µg + fluticasone 50 μg 2 puffs twice daily

CO-TREATMENT: beta2-agonists (salbutamol or albuterol) were permitted on an as needed basis. Use of nasal corticosteroids or nasal cromolyn for allergic rhinitis and continuation of immunotherapy (initiated at least 6 months before the first study visit) were allowed for participants taking stable doses at baseline

Outcomes

INTENTION-TO-TREAT ANALYSES: Efficacy analysis included all randomly assigned participants who took at least 1 dose of study drug and had spirometry measurements for analysis in both treatment periods
Outcomes used at endpoint

PULMONARY FUNCTION TESTS
*Maximum % decrease in FEV1 after exercise and before beta2-agonist administration

Area under the curve for the first 20 minutes after exercise

Time to recovery within 5% of per exercise FEV1

Maximum FEV1 %predicted after beta2-agonists

Average % change from preexercise baseline in FEV1 after beta2-agonists

SYMPTOM SCORES
Not reported

EXACERBATIONS

Not reported

FUNCTIONAL STATUS
Need for rescue medication

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Clinical adverse experiments

Serious adverse experiments

Drug-related adverse experiments

Adverse experiments resulting in discontinuation

WITHDRAWALS
Overall and due to adverse effects reported

(* denotes primary outcome)

Notes

Full-text report

Funded by Merck Research Laboratories

Did not confirm methodology and extracted data

User-defined order: 4 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomly assigned according to a computer-generated allocation schedule.
Allocation concealment (selection bias)Low riskNumbered inhalers and bottles were used to implement allocation. All study personnel, including investigators, study site personnel, participants, monitors and central laboratory personnel, remained masked to treatment allocation throughout the study
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble blind, double-dummy study
Incomplete outcome data (attrition bias)
Exacerbations
Low riskWell-balanced number of participants withdrawn in both comparison groups; reported reasons for withdrawal
Selective reporting (reporting bias)Low riskPrimary and secondary outcomes were specified and all data were reported
Other biasLow riskNo risk of bias was observed

Green 2006

MethodsCross-over, single-centre study in UK
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE
≤ 400 µg BDP equivalent

NUMBER RANDOMLY ASSIGNED: 49

NUMBER COMPLETED: 39

M = 25

F = 24

MEAN AGE in years: 42

SEVERITY: not stated

BASELINE %PREDICTED FEV1 (L): 74.8

ATOPIC: 93%

INCLUSION CRITERIA: 18 to 75 years of age, diagnosed with asthma; receiving treatment with ≤ 400 µg/d CFC-BDP per day; one or more of (1) > 15% increase in FEV1 post-SABA; (2) > 20% within-day variability in PEF assessed twice daily over a 2-week period; (3) provocative concentration of methacholine causing a 20% fall in FEV1 (PC20) < 8 mg/mL-1; following run-in on 200 μg/d BUD, participants were eligible if they had recorded day- or night-time asthma symptoms on their diary cards on at least 4 days in the third or fourth baseline week

EXCLUSION: current smokers or smoking history of > 10 pack-years, significant co-morbidity, treated with oral corticosteroids, long-acting β2-agonists, leukotriene antagonists or theophylline; asthma exacerbation or lower respiratory tract infection within 4 weeks before trial entry

Interventions

LTRA + ICS versus LABA + ICS (stable low dose of ICS)

INTERVENTION GROUP 1

LTRA: oral montelukast 10 mg qd + budesonide 100 mg bid

INTERVENTION GROUP 2

LABA: formoterol 12 mg bid + budesonide 100 mg bid

TREATMENT PERIOD: 6 weeks (washout period: 4 weeks)

RUN-IN PERIOD: 4 weeks

CO-TREATMENT: not reported

Outcomes

INTENTION-TO-TREAT ANALYSES: cross-over data analysed for completers

PULMONARY FUNCTION TESTS
FEV1, PEFR but only improvements when groups compared; no individual group results were presented

SYMPTOM SCORES
VAS (individual group values not presented, but rather differences between groups)

EXACERBATIONS
Reported as events

FUNCTIONAL STATUS
Not stated

INFLAMMATORY MARKERS
Not stated

ADVERSE EFFECTS
Not stated

WITHDRAWALS
Not reported

Notes

Funding source not disclosed

Confirmation of methodology and data extraction received

User-defined order: 4 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk'I believe that this was generated using a computer statistical package generating a random sequence. I don't know the package that was used and unfortunately the individual has left our organisation but had extensive clinical trials expertise'
Allocation concealment (selection bias)Low risk'...this was indeed generated by a third party, namely the pharmacist responsible for dispensing the double blind medication (...) None of the study investigators were aware of the randomisation schedule until the last patient had completed the cross-over study'
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-dummy
Incomplete outcome data (attrition bias)
Exacerbations
High riskCompleters used for analysis
Selective reporting (reporting bias)Low riskOCS-treated exacerbations reported. Data could not be extracted, as only data on events without number of participants were made available
Other biasLow riskNo risk of bias was observed

Grosclaude 2003

MethodsParallel-group, open-label study; multi-centre study (115 centres in France)
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE
1000 µg CFC beclomethasone or equivalent daily

RANDOMLY ASSIGNED: 253

LTRA: 130
LABA: 123

WITHDRAWALS
LTRA: 16 (12%)
LABA: 7 (6%)

AGE in years: mean ± SD
LTRA: 44.6 (18.2)
LABA: 43.1 (17.8)

GENDER (% male)
LTRA: 39

LABA: 39

SEVERITY
Not described

BASELINE %PREDICTED FEV1 (L)
Not reported

BASELINE PEF (L/min)
LTRA: 327
LABA: 344

ALLERGIC RHINITIS (%)
LTRA: 51
LABA: 52

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years: mean ± SD
Reported as % with asthma duration
<1 year: 6; between 1 and 5 years: 17; between 5 and 10 years: 15; between 10 and 15 years: 19; longer than 15 years: 43

ELIGIBILITY CRITERIA: 15 years of age or younger, diagnosed with asthma; treatment for at least four weeks with CFC-BDP equivalent of ≥ 1000 µg/d and inhaled SABA prn; able to use Mini Wright PEF metre; able to fill in daily record card

Over last seven days of run-in

  • Mean AM PEF between 60% and 80% of predicted best (as obtained post-BD at visit 2)

  • Asthma symptoms on at least two days

  • Used SABA at least four times

EXCLUSION CRITERIA: use of systemic CS, anti-leukotriene agent, LABA, lower RTI within previous four weeks; hospitalisation within previous 4 weeks; hypersensitivity to one of compound study drugs; serious uncontrolled concurrent disease; allergen-specific immunotherapy in incremental phase; smoker or ex-smoker with 10 pack-year history; participation in clinical study in previous month

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: 1 to 2 weeks
Intervention period: 12 weeks

INTERVENTION GROUP 1
LTRA: montelukast @ 10 mg/d PO + CFC-BDP 250 μg two puffs bid via pMDI

INTERVENTION GROUP 2
LABA: salmeterol 50 μg bid via MDI + FP 250 μg one puff bid via discus (single/combination inhaler)

CO-TREATMENT: SABA prn

Outcomes

INTENTION-TO-TREAT ANALYSES: outcomes used at endpoint or 12 weeks

PULMONARY FUNCTION TESTS
*Change from baseline in AM PEF; change from baseline in PM PEFR

SYMPTOM SCORES
Change from baseline in % of nights with awakenings; change from baseline in % of days with no symptoms; change from baseline in % of nights with no symptoms

EXACERBATIONS
One or more of the following
Mild: reduction in AM PEF > 20% of baseline; increased bronchodilator usage; awakenings due to asthma on one or more consecutive nights
Moderate: reduction in AM PEF > 30% of baseline; change in maintenance therapy or premature termination of trial therapy; oral steroids
Severe: hospitalisation

FUNCTIONAL STATUS
Change from baseline in % of nights without rescue medication usage; change from baseline in % of days without rescue medication usage; % of participants with good asthma control for 10 of 12 weeks, as defined by the presence of two of the following

  • PEF ≥ 80 %predicted

  • No more than four puffs of BD on no more than 2 days

  • Symptom-free for at least two days

  • Presence of all the following criteria on a weekly basis: no nocturnal awakening; no exacerbation; no unscheduled medical contact; no adverse effect of treatment leading to withdrawal

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Headache; gastroenteritis; upper respiratory inflammation; pharyngitis; viral respiratory infection; malaise and fatigue; allergic rhinitis; diarrhoea; digestive discomfort and pain; ENT symptoms; muscle cramps and spasms; regurgitation and reflux; nasal inflammation; vertigo; nausea and vomiting; cough; lower respiratory infection; dyspeptic symptoms

WITHDRAWALS
Due to poor completion of diary cards; due to adverse effects; due to poor control
Overall (all reported)

(* denotes primary outcome)

Notes

Full-text report and unpublished trial report

Received additional unpublished data (SFCF4007) from GSK website

Funded by GSK

User-defined number: 12 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskSee Appendix 2
Allocation concealment (selection bias)Low riskSee Appendix 2
Blinding (performance bias and detection bias)
All outcomes
High riskOpen-label study
Incomplete outcome data (attrition bias)
Exacerbations
Unclear risk'The safety population included all subjects who received at least 1 dose of the study medication. The intent-to-treat population (ITT) included all randomised subjects who received at least one dose of the study medication, and from whom daily record card (DRC) data were available during the run-in period and the treatment period'
Selective reporting (reporting bias)Low riskData for exacerbations reported in pharmaceutical company download. The definition of exacerbation was not explicit, and we could not use the outcome data for this study in the meta-analysis
Other biasLow riskNo risk of bias was observed

Hendeles 2004

MethodsParallel-group; numbers of sites and countries unclear
Participants

INADEQUATELY controlled participants taking ICS + SABA prn with history of EIB at baseline

BASELINE ICS DOSAGE
Not reported

RANDOMLY ASSIGNED: 91 (unclear allocation between groups)

WITHDRAWALS
Not reported

AGE in years (range)
15 to 60

GENDER (% male)
Not reported

SEVERITY
Not described

BASELINE %PREDICTED FEV1 (L)
LTRA: 81.3
LABA: 78.9

ALLERGIC RHINITIS (%)
Not reported

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years: mean ± SD
Not reported

ELIGIBILITY CRITERIA: participants with asthma who remained symptomatic when taking ICS; age 15 to 60 years; history of EIB

EXCLUSION CRITERIA: not reported

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Intervention period: 4 weeks

INTERVENTION GROUP 1
LTRA: montelukast @ 10 mg/d PO + FP 125 µg one puff bid via inhalation

INTERVENTION GROUP 2
LABA: salmeterol 50 µg bid via MDI + montelukast placebo + ICS = FP 125 µg one puff bid via inhalation (separate inhalers)

CO-TREATMENT
Not reported

Outcomes

INTENTION-TO-TREAT ANALYSES: outcomes used at endpoint or 4 weeks

PULMONARY FUNCTION TESTS
Challenge FEV1 %predicted; change in FEV1 %predicted; rescue bronchodilation

SYMPTOM SCORES
Not reported

EXACERBATIONS
Not reported

FUNCTIONAL STATUS
Not reported

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Not reported

WITHDRAWALS

Due to adverse events: not reported
Due to poor control: not reported
Overall: reported

Primary outcome not identified

Notes

Unpublished: conference abstract

Funded by Merck

User-defined number: 4 weeks

No data could be used for aggregation

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskDescribed as randomised; other information not available
Allocation concealment (selection bias)Unclear riskInformation not available
Blinding (performance bias and detection bias)
All outcomes
Unclear riskInformation not available
Incomplete outcome data (attrition bias)
Exacerbations
Unclear riskInformation not available
Selective reporting (reporting bias)Unclear riskUnable to verify whether primary outcome measured in the review
Other biasUnclear riskNot enough details about the study could be ascertained from the available abstract

Ilowite 2004

MethodsParallel-group study, multi-centre trial (132 centres in USA for 48 weeks)
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE: 220 µg of ICS per day

RANDOMISED: 1473
LTRA: 743
LABA: 730

WITHDRAWALS
LTRA: 128/743
LABA: 113/730

AGE in years: mean ± SD
LTRA: 39.0 (range 14 to 73)
LABA: 38.1 (range 15 to 70)

GENDER (% male)
LTRA: 41.2
LABA: 37.5

SEVERITY
Moderate to severe persistent asthma

BASELINE %PREDICTED FEV1 (L)
LTRA: 74.3 ± (SD 11.5)
LABA: 74.3 ± (SD 11.7)

ALLERGEN TRIGGERS
Not reported

ALLERGIC RHINITIS
Not reported

ASTHMA DURATION in years
NS

ELIGIBILITY CRITERIA: age 15 to 65 years; asthma for ≥ 1 year; ICS use daily for at least 8 weeks before first visit; baseline FEV1 50% to 90% of predicted

≥ 12% change in FEV1 after albuterol and, in the 14 days before randomisation, one or more of the following

  • Asthma symptom that required the use of β2-agonist medication on average once per day

  • Minimum biweekly daytime symptom score of 56 for a 14-day period

EXCLUSION CRITERIA: emergency department visit in < 1 month; admission for asthma in < 3 months; upper respiratory infection in < 3 weeks of 1st visit or during run-in; pregnant or lactating women; use of LABA within 1 month before visit 1; use < 1 month of oral, intravenous, intramuscular or intra-articular corticosteroids; use < 2 weeks of leukotriene antagonist, cromolyn or nedocromil; use of theophylline in < 1 week; use in < 2 weeks of oral or inhaled long-acting β2-agonists or inhaled anti-cholinergics

SETTING: not described

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION:
Run-in period: 2 weeks
Intervention period: 48 weeks

INTERVENTION GROUP 1
LTRA: montelukast 10 mg once daily + fluticasone 125 µg bid via MDI

INTERVENTION GROUP 2
LABA: salmeterol 50 µg bid via MDI + fluticasone 125 µg bid via MDI

2 inhalers used for combination therapy

CO-TREATMENT: not specified

Outcomes

INTENTION-TO-TREAT ANALYSES for participants who received at least one dose of medication

Outcomes used at endpoint

PULMONARY FUNCTION TESTS

*Change from baseline in AM PEFR; change from baseline in PM PEFR; change from baseline in FEV1

SYMPTOM SCORES

Change from baseline DAYTIME symptom scores; change from baseline NIGHT-TIME symptom scores

EXACERBATIONS
Exacerbations requiring systemic steroids

FUNCTIONAL STATUS
Change from baseline in mean OVERALL use of beta2-agonists (puffs/d); change from baseline in mean DAYTIME use of beta2-agonists (puffs/d); change from baseline in mean NIGHT-TIME use of beta2-agonists (puffs/d); % of night-time awakenings

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Drug-related and non-drug-related

WITHDRAWALS
Due to adverse effects reported

(* denotes trial's primary outcome)

Notes

Unpublished data

Received full disclosure of unpublished data provided by Peter Polos, March 2004

Funded by Merck & Co

Confirmation with supportive documents received for methodology and data extraction

User-defined number: 48 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomisation was determined by computer-generated allocation schedule (block size 4)
Allocation concealment (selection bias)Low riskCentralised, third-party randomisation
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-dummy
Incomplete outcome data (attrition bias)
Exacerbations
Unclear riskITT
Selective reporting (reporting bias)Low riskData available for meta-analysis of primary outcome
Other biasLow riskNo risk of bias was observed

Lemanske 2010

MethodsCross-over study, multi-centre study in USA
Participants

INADEQUATELY controlled children taking ICS at baseline

BASELINE ICS DOSAGE: 100 µg FP ICS per day during run-in

RANDOMLY ASSIGNED: 182

WITHDRAWALS: 25

AGE in years: mean ± SD

LTRA: 10.5 ± 3.1
LABA: 11.2 ± 2.9

GENDER (% male)

LTRA: 65.0
LABA: 72.1

SEVERITY: Mild to moderate asthma

BASELINE %PREDICTED FEV1 (L)

LTRA: 98.8 ± 13.9
LABA: 96.5 ± 12.1

ALLERGEN TRIGGERS (%)
+ve aeroallergen test: 77

+ve perennial allergen test: 70

ALLERGIC RHINITIS: not reported

Age at MD asthma diagnosis (years)

LTRA: 3.8 ± 2.6
LABA: 4.2 ± 3.6

ELIGIBILITY CRITERIA: age 6 to 17 years; physician-diagnosed mild to moderate asthma, based on NAEPP criteria; FEV1 %predicted > 60%; increase in FEV1 > 12% predicted or PC20 12.5 mg/mL or less

During run-in, children had to exhibit uncontrolled asthma, defined as one or more of the following

  • Diary-reported symptoms (coughing rated as moderate or severe or wheezing rated as mild, moderate or severe)

  • Rescue use of reliever medication (two or more puffs per day)

  • Peak flows < 80 %predicted

EXCLUSION CRITERIA: corticosteroid treatment within 2 weeks (unless ingested nasally, in which case at discretion of investigator; current or prior use of medications known to interact with corticosteroids; more than three hospitalisations for asthma in the past year; lung disease other than asthma; significant medical illness other than asthma; history of cataracts, glaucoma or medical disorder associated with adverse effects related to corticosteroids; uncontrolled gastroesophageal reflux symptoms; significant asthma exacerbation within 2 weeks of visit 1 or more than 5 courses of systemic corticosteroids in the past year; life-threatening asthma exacerbation requiring intubation or mechanical ventilation or resulting in a hypoxic seizure in last 5 years; adverse reactions to ICS, LTRA or LABA preparations; hyposensitisation therapy other than an established maintenance regimen (continuous regimen for >3 months); pregnancy or lactation; failure to practice abstinence or use of an acceptable birth control if of child-bearing potential; inability to perform study procedures; refusal to consent to a genotype evaluation; inability to ingest the study drug; evidence that the family may be unreliable or nonadherent, or that they may move from the clinical centre area before trial completion

SETTING: not described

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: 2 to 8 weeks
Intervention period: 48 weeks (3 × 16 weeks)

INTERVENTION GROUP 1
LTRA: montelukast 5 or 10 mg once daily + fluticasone 100 µg bid via DPI

INTERVENTION GROUP 2
LABA: salmeterol 50 µg bid via MDI + fluticasone 100 µg bid via DPI

1 inhaler used for combination therapy

CO-TREATMENT: not specified

COMPLIANCE: 84% for study tablets (as measured by an electronic monitor in the bottle cap) and 87% for study inhalers (as measured by a disk counter)

Outcomes

INTENTION-TO-TREAT ANALYSES: no completers analysed

PULMONARY FUNCTION TESTS

Collected as part of a composite outcome (differential response*): FEV1

SYMPTOM SCORES

Collected as part of a  composite outcome (differential response*): symptom-free days

EXACERBATIONS
Collected as part of a composite outcome (differential response*): exacerbations requiring systemic steroids

FUNCTIONAL STATUS
Quality of life (AQLQ)

INFLAMMATORY MARKERS
Collected

ADVERSE EFFECTS
Drug-related and non-drug-related

WITHDRAWALS: reported

(* denotes trial's primary outcome)

Notes

Full-text article

Funded by National Heart, Lung, and Blood Institute (HL064307, HL064288, HL064295, HL064287, HL064305 and HL064313), the National Institute of Allergy and Infectious Diseases (T32AI007635) and the Clinical Translational Science Award programme of the National Center for Research Resources (UL1-RR025011 (Wisconsin), UL1-RR025780 (Colorado) and UL1-RR024992 (St Louis))

Confirmation of data: not obtained

User-defined number: 16 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk

Computer-generated randomisation schedule

'The pattern of treatment assignment will utilize the complete set of orthogonal Latin squares...'

Allocation concealment (selection bias)Low riskCentralised randomisation. Investigators dialled into server requesting assignment and received a packet number that related to a medication
Blinding (performance bias and detection bias)
All outcomes
Low risk

'The drug assignments were masked with the use of placebo tablets and dummy disk devices that discharged powder without the active drug'

'...investigators and the children, along with their caregivers, will not know which treatment is being received during each treatment period'

Incomplete outcome data (attrition bias)
Exacerbations
Low riskThe primary analysis involved two stages: a one-sided exact test for binomial proportions to test the null hypothesis that the percentage of participants with a differential response would be ≤ 25% at the 0.01 significance level, and if there was a significant result in stage one, a rank-ordered logistic regression to test whether four prespecified baseline co-variates would predict distinct patterns of differential responses
Selective reporting (reporting bias)Low riskStudy protocol is available. Primary outcome analysed as events
Other biasLow riskNo risk of bias was observed

Nelson 2000

MethodsParallel-group study, multi-centre trial (39 centres)
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE
For the 3 week run-in period, 100 µg twice daily FP (equivalent to 400 µg daily of beclomethasone)

RANDOMLY ASSIGNED: 447
LTRA: 225
LABA: 222

WITHDRAWALS
LTRA: 30 (13%)
LABA: 24 (11%)

AGE in years: mean ± SD
LTRA: 43 ± 13.7
LABA: 40.2 ± 14.4

GENDER (% male)
LTRA: 40
LABA: 39

SEVERITY
Not reported

BASELINE %PREDICTED FEV1 (L)
LTRA: 70.8 ± 0.05 (SEM)
LABA: 70.0 ± 0.05

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years (%)
Less than 10 years
LTRA: 23
LABA: 24

10 years or longer
LTRA: 77
LABA: 76

ELIGIBILITY CRITERIA: age ≥ 15 years; asthma ≥ 6 months; low-moderate dose of ICS for ≥ 1month CFC-BDP: 252 to 420 µg/d; BUD 400 µg/d; FP 176to 220 µg/d; triamcinolone 600 to 800 µg/d); 50% to 80% of predicted normal
≥ 12% increase in FEV1 post-200 µg albuterol

At randomisation: FEV1 50% to 80% of predicted; 1 additional sign of inadequate asthma control in the preceding 7 days

  • ≥ 4 puffs/d albuterol

  • Symptom score ≥2 on a scale of (0 to 5) for ≥ 3 days

  • ≥3 nights waking for asthma

EXCLUSION CRITERIA: pregnant or lactating female patients; life -threatening asthma; hospitalised for asthma in the last three months; significant concurrent disease; < 30 days of screening: use of theophylline, other bronchodilators, other leukotriene modifiers, cromolyn or nedocromil

SETTING: not specified

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: 3 weeks
Intervention period: 12 weeks

INTERVENTION GROUP 1
LTRA: oral montelukast 10 mg once daily + ICS = FP 100 µg twice daily via discus

INTERVENTION GROUP 2
LABA: salmeterol 50 µg twice daily via discus + ICS = FP 100 µg twice daily via Diskus

1 inhaler used for combination therapy

CO-TREATMENT: none

Outcomes

Modified INTENTION-TO-TREAT ANALYSES
Outcomes used at endpoint for exacerbations and withdrawals only (not available for continuous values)

PULMONARY FUNCTION TESTS
Change from baseline in FEV1; change from baseline in AM PEFR; change from baseline in PM PEFR

SYMPTOM SCORES
Change from baseline in OVERALL symptom scores; change from baseline in night-time awakenings; change in symptom-free days

EXACERBATIONS
Exacerbations requiring hospital admission; exacerbations requiring systemic steroids (data provided)

FUNCTIONAL STATUS
Change from baseline in mean OVERALL use of β2-agonists (puffs/d); change in rescue-free days

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Included oral candidiasis, sore throat, hoarseness, headache

WITHDRAWALS
Due to adverse effects
Due to poor control
Overall
(reported)

Notes

Full-text report

Received additional unpublished data provided by Karen Richardson, GSK (July 2003)

Funded by Glaxo Wellcome, study SAS40018

Confirmation of methodology and data extraction received

User-defined order: 12 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer-generated randomisation schedule
Allocation concealment (selection bias)Low riskNumber-coded inhaler/pills supplied by pharmacy
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-dummy
Incomplete outcome data (attrition bias)
Exacerbations
Unclear risk'The primary population for the analyses of demographic/baseline characteristics, efficacy and safety was the Intent-to-Treat (ITT Population), which consisted of all subjects who were randomised to receive study drug'
Selective reporting (reporting bias)Low riskData available for primary outcome
Other biasLow riskNo risk of bias was observed

Nelson 2001

MethodsParallel-group, multi-centre trial (54 centres)
Participants

INADEQUATELY controlled adolescent and adult participants taking ICS at baseline

BASELINE ICS DOSAGE
Not described

RANDOMLY ASSIGNED: 429
LTRA: 215
LABA: 214

WITHDRAWALS
LTRA: 18 (8%)
LABA: 12 (6%)

AGE in years: mean ± SD
LTRA: 39.3 ± 13.20
LABA: 40.9 ± 13.17

GENDER (% male)
LTRA: 44
LABA: 44

SEVERITY
Not described

BASELINE %PREDICTED FEV1 (L)
LTRA: 65.86 ± 0.58 (SEM)
LABA: 66.62 ± 0.58

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years (%)
Less than 10 years
LTRA: 24
LABA: 24

10 years or longer
LTRA: 76
LABA: 76

ELIGIBILITY CRITERIA: age ≥ 12 years; asthma ≥ 6 months; FEV1 50% to 80% of predicted normal; ≥ 12% increase in FEV1 post-200 µg albuterol

Following 7 to 14-day run-in

In the six days before randomisation, one or more of the following

  • Average of 4 or more puffs/d of albuterol

  • Symptom score of 2 or higher on at least 2 days for any of the asthma symptom categories

  • At least one night when the participant awoke because of asthma

  • Two or more days on which PM to AM PEF variation was 20% or more

* Intake of daily ICS before randomisation is NOT specified as inclusion criteria

Participants also must have been using an oral or inhaled SABA for 6 weeks

EXCLUSION CRITERIA: not described

SETTING: clinical centres

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: 1 to 2 weeks
Intervention period: 4 weeks

INTERVENTION GROUP 1
LTRA: zafirlukast 20 mg twice daily + ICS: constant dose of existing ICS medication

INTERVENTION GROUP 2
LABA: salmeterol 42 µg, 2 puffs twice daily via MDI + ICS = constant dose of existing ICS medication

2 inhalers used for combination therapy

CO-TREATMENT
Theophylline or other medications that could potentially interact with study treatment not allowed; albuterol inhalers provided for use on an as needed basis, but all other bronchodilators not permitted; antihistamines, nasal decongestants and intranasal medications for rhinitis were permitted

Outcomes

INTENTION-TO-TREAT ANALYSES: outcomes used at endpoint

PULMONARY FUNCTION TESTS
Change from baseline FEV1; *change from baseline in AM PEFR; change from baseline in PM PEFR; change in PEF variability

SYMPTOM SCORES
Change from baseline DAYTIME symptom scores; change from baseline NIGHT-TIME symptom scores; change in symptom-free days; participant satisfaction

EXACERBATIONS
Exacerbations requiring systemic steroids; exacerbations defined as any worsening of asthma symptoms requiring a change in participant's asthma therapy other than increased use of supplemental albuterol. Participants who experienced an exacerbation were withdrawn from the study

FUNCTIONAL STATUS
Change from baseline in mean DAYTIME use of beta2-agonists (/d); change from baseline in mean NIGHT-TIME use of β2-agonists (/d); change in rescue-free days; change/absolute in rescue-free nights; change in quality of life; change in night-time awakenings

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Upper respiratory tract infection, headache, nausea

WITHDRAWALS
Due to adverse effects
Due to poor control
Overall
(reported)

(* denotes primary outcome)

Notes

Full-text report

Received additional unpublished data provided by Karen Richardson, GSK

Funded by Glaxo Wellcome, protocols SLGA5024 and SLGA5025

Confirmation of methodology and data extraction received

User-defined order: 4 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskSee Appendix 2
Allocation concealment (selection bias)Low riskSee Appendix 2
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-dummy
Incomplete outcome data (attrition bias)
Exacerbations
Unclear risk'The ITT population included all subjects who were randomised to study drug'
Selective reporting (reporting bias)Low riskData available for primary outcome
Other biasLow riskNo risk of bias was observed

Nsouli 2001

MethodsUnclear whether parallel-group or cross-over
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE
FP 100 to 300 or CFC-BDP 200 to 550 or BUD 200 to 400 or flunisolide 500 to 1000 or triamcinolone 400 to 1000

RANDOMLY ASSIGNED: 30
LTRA: unknown
LABA: unknown

WITHDRAWALS
Not described

AGE in years: mean ± SD
Not described

GENDER (% male)
Not described

SEVERITY
Not described

BASELINE %PREDICTED FEV1 (L)
Not described

ALLERGEN TRIGGERS
Not described

ASTHMA DURATION in years: mean ± SD
Not described

ELIGIBILITY CRITERIA
Not described

EXCLUSION CRITERIA
Not described

SETTING
Not described

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: not described
Intervention period: 8 weeks

INTERVENTION GROUP 1
LTRA: montelukast 10 mg QD PM + ICS (low-dose ICS)

INTERVENTION GROUP 2
LABA: salmeterol 50 µg bid + ICS (low-dose ICS)

2 inhalers used for combination therapy

CO-TREATMENT
Not reported

Outcomes

ANALYSES: not reported

PULMONARY FUNCTION TESTS
FEV1 and FEF25-75

SYMPTOM SCORES
None described

EXACERBATIONS
Not described

FUNCTIONAL STATUS
Quality of life

INFLAMMATORY MARKERS
Not described

ADVERSE EFFECTS
Not described

WITHDRAWALS
Not described

Notes

Abstract

Funding of study unknown

Confirmation of methodology and data extraction not obtained

User-defined order: 8 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskDescribed as randomised; other information not available
Allocation concealment (selection bias)Unclear riskInformation not available
Blinding (performance bias and detection bias)
All outcomes
Unclear riskInformation not available
Incomplete outcome data (attrition bias)
Exacerbations
Unclear riskInformation not available
Selective reporting (reporting bias)Unclear riskCannot establish this reliably
Other biasUnclear riskCannot establish this reliably

Pavord 2007

MethodsParallel-group, multi-centre study in the UK
Participants

BASELINE ICS DOSE: up to 400 µg (with BDP or HFA) BDP equivalent

NUMBER SCREENED: 132

NUMBER RANDOMLY ASSIGNED: 66

NUMBER COMPLETED: 54

M = 34

F = 32

MEAN AGE in years: 35

BASELINE DETAILS: PEF 417 L/min; FEV1 %predicted 76%

INCLUSION CRITERIA: 18 to 50 years, non-smokers, receiving a stable dose of up to 400 µg of beclomethasone dipropionate (presumed CFC-BDP equivalent) a day or equivalent ICS, requiring further therapy; likelihood of compliance with protocol requirements and ability to use an Accuhaler and mini-Wright peak flow meter. Post-run-in: baseline %predicted FEV1 61% to 85%; PC20 < 8 mg/mL with methacholine challenge; at least one of the following: diary card recording of symptoms on > 4 of the last seven days of the run-in period; recorded use of relief medication on >2 different days during the last seven days of the run-in period; period variation in PEF > 10% over last seven days of run-in

EXCLUSION: additional medication other than ICS, SABA or OCS in previous 3 months; acute respiratory infection/exacerbation of asthma within 4 weeks of screening; recent or significant smoking history; pregnancy/lactation; inadequate contraceptive methods in women of child-bearing age

Interventions
  • Combination fluticasone/salmeterol 100/50 µg bid

  • Fluticasone 100 μg bid plus montelukast 10 mg OD

RUN-IN PERIOD: 2 weeks

TREATMENT PERIOD: 12 weeks

Outcomes

INTENTION-TO-TREAT ANALYSES: no

PULMONARY FUNCTION TEST

FEV1; AM PEF; PM PEF

SYMPTOMS

Percentages of symptom-free days and nights

FUNCTIONAL STATUS

Rescue medication use

INFLAMMATORY MARKERS

Neutrophils, eosinophils, macrophages, lymphocytes

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low risk'...consecutively randomised according to a pre-defined randomisation list...'
Allocation concealment (selection bias)Low risk'Treatment allocation was concealed from the subject, pharmacist, and investigator'
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-dummy design employed
Incomplete outcome data (attrition bias)
Exacerbations
Unclear risk

'All summaries and analyses are for the intention-to-treat population (all subjects receiving at least one dose of the study drug). No imputations were performed for missing data. Therefore if data were missing for either baseline or one of the time points, it was not possible to calculate a change from baseline. However, all available data have been used for relevant summaries'

Some imbalance between treatment groups in terms of withdrawal, although denominators for lung function outcomes show numbers randomly assigned

Selective reporting (reporting bias)Low riskOutcomes identified as those of interest were presented in the article. Data on exacerbations were not identified as representing an outcome of interest to the investigators
Other biasLow riskNo risk of bias was observed

Price 2011

MethodsParallel-group, multi-centre (53 primary care practices in Norfolk, Suffolk, Essex, Sussex, Cambridgeshire, Bedfordshire, Hampshire and Dorset, UK), pragmatic trial
Participants

INADEQUATELY controlled participants taking ICS for at least 12 weeks

BASELINE ICS DOSAGE: beclomethasone, budesonide, fluticasone
LTRA: montelukast, zafirlukast
LABA: salmeterol, formoterol

RANDOMLY ASSIGNED: 361

LTRA: 176

LABA: 185

WITHDRAWALS
LTRA: 7 (4%)
LABA: 4 (2.2%)

AGE in years: mean ± SD
LTRA: 51.0 ± 16.0
LABA: 49.7 ± 16.1

GENDER (% female)
LTRA: 64
LABA: 61.8

SEVERITY

Uncontrolled asthma with first-line monotherapies (ICS and LTRA)

BASELINE %PREDICTED FEV1 (L)

Not reported

ALLERGIC RHINITIS
Not reported

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years: mean ± SD
Not reported

ELIGIBILITY CRITERIA: male and female patients, 12.80 years of age, with symptoms of poorly controlled asthma, were eligible for inclusion. Other eligibility criteria included screening peak expiratory flow (PEF) > 50% predicted after inhaled beta2-agonist was withheld for 4 hours and, at the baseline visit, a score of < 6 points (range 1 to 7 with 7 the best) on the validated Mini Asthma Quality of Life Questionnaire (MiniAQLQ) and/or .1 point (range 0 to 6 with 6 the worst) on the validated Asthma Control Questionnaire (ACQ). An additional eligibility criterion was attending physician opinion that initiation of or increase in asthma controller therapy was required

EXCLUSION CRITERIA: Patients were excluded if they were experiencing an acute asthma exacerbation, if they had received treatment with systemic, intramuscular or intra-articular corticosteroids within 2 weeks of screening or if they had a substantial change in anti-asthma medication or another active, acute or chronic pulmonary disorder or unresolved respiratory infection within 12 weeks of screening. Patients were excluded from the first controller trial if they had received treatment with an ICS or LTRA within the prior 12 weeks; patients were excluded from the add-on therapy trial if they had received an LTRA or LABA within 12 weeks of screening. Female patients of child-bearing potential agreed to use adequate contraception throughout the study

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: not reported
Intervention period: 2 years

INTERVENTION GROUP 1
LTRA: montelukast 10 mg once daily or zafirlukast 20 mg twice daily + ICS (beclomethasone, budesonide, fluticasone)

INTERVENTION GROUP 2
LABA: salmeterol or formoterol + ICS (beclomethasone, budesonide, fluticasone)

CO-TREATMENT: Other permitted asthma medications included as needed inhaled short-acting beta2-agonist, theophylline, cromoglycate, nedocromil or ipratropium if clinically indicated

Outcomes

INTENTION-TO-TREAT ANALYSES
Outcomes used at endpoint

PULMONARY FUNCTION TESTS
PEFR

SYMPTOM SCORES

Asthma control questionnaire

EXACERBATIONS
Incidence of asthma exacerbations

Frequency of asthma exacerbations, defined as a course of oral corticosteroid or hospitalisation for asthma

FUNCTIONAL STATUS
*Self-administered asthma-specific QoL questionnaire

Royal College of Physicians 3-item questionnaire

Validated 14-item Mini Rhinoconjunctivitis Quality of Life Questionnaire

Validated asthma diary card

Treatment changes

Adherence to therapy

INFLAMMATORY MARKERS
Not recorded

ADVERSE EFFECTS
Frequency of upper and lower respiratory tract infections and consultations for respiratory tract infection

Adverse events

WITHDRAWALS
Reported

(* denotes primary outcome)

Notes

Full-text paper

Funded by National Coordinating Centre for Health Technology Assessment UK and Merck Sharp & Dohme, AstraZeneca, Boehringer Ingelheim

Confirmation of methodology and data extraction not obtained

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskAn automated, computerised telephone centre randomly assigned eligible participants. Randomisation was stratified by
practice, with a block size of six
Allocation concealment (selection bias)Low riskAn automated, computerised telephone centre randomly assigned eligible participants
Blinding (performance bias and detection bias)
All outcomes
High riskBlinding of participants and healthcare providers was not performed to have pragmatic design of the study
Incomplete outcome data (attrition bias)
Exacerbations
Low riskWell-balanced number of participants withdrew in both comparison groups; reasons for withdrawal were reported. Intention-to-treat analysis was performed
Selective reporting (reporting bias)Low riskProtocol is available. Primary and secondary outcomes were specified and all data reported
Other biasLow riskNo risk of bias was noted

Ringdal 2003

MethodsParallel-group, multi-centre trial (114 centres in 19 countries)
Participants

INADEQUATELY controlled participants taking 'moderate or high doses' of ICS at baseline

BASELINE ICS DOSAGE
800 µg or more of beclomethasone-equivalent/d (moderate or high dose)

RANDOMLY ASSIGNED: 725
LTRA: 369
LABA: 356

WITHDRAWALS
LTRA: 37 (10%)
LABA: 19 (5%)

AGE in years: mean ± SD
LTRA: 43 (14 to 79)
LABA: 43 (15 to 75)

GENDER (% male)
LTRA: 45
LABA: 46

SEVERITY: MODERATE PERSISTENT asthma

BASELINE %PREDICTED FEV1 (L)
LTRA: 74.3 ± 16.1
LABA: 75.8 ± 15.3

ALLERGEN TRIGGERS
Not described

ASTHMA DURATION in years: mean ± SD
Not described

ELIGIBILITY CRITERIA: age ≥ 15 years; moderate persistent asthma as per the ATS and NAEPP Report 2; using ICS at moderate or high dose (400 to 1000 µg/d of CFC-BDP, BUD or flunisolide; or 200 to 500 µg/d of FP) for at least 4 weeks; history of reversible airway obstruction; ≥ 15% change in FEV1 after 800 µg of salbutamol

At end of run-in: mean PEF of 50% to < 85% of value in clinic after 400 µg of salbutamol; cumulative symptom score ≥ 8 in past 7 days or ≥4 of the last 7 days of run-in

EXCLUSION CRITERIA: recent change in asthma medication; respiratory tract infection or admission for asthma in < 4 weeks; intake of oral, depot or parenteral corticosteroids in < 4 weeks or ≥ 2 occasions in past 12 weeks; cigarette smoking ≥ 10 pack-year history; pregnant or lactating women or those likely to become pregnant during study; FEV1 < 50%

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: 4 weeks
Intervention period: 12 weeks plus a 2-week follow-up

INTERVENTION GROUP 1
LTRA: montelukast 10 mg/d + ICS: FP 100 µg twice daily via discus

INTERVENTION GROUP 2
LABA: salmeterol 50 µg bid + ICS: FP 100 µg twice daily via discus

1 inhaler used for combination therapy

CO-TREATMENT
Salbutamol provided for relief of symptoms, no other SABA permitted
Other oral, parenteral or depot CS not allowed, except where documented for treatment of exacerbations. Other existing asthma treatment allowed at constant dose

Outcomes

INTENTION-TO-TREAT ANALYSES: yes, but excluding those incorrectly randomly assigned because they failed major inclusion criteria; outcomes used at endpoint

PULMONARY FUNCTION TESTS
Change from baseline in FEV1; *change from baseline in AM PEFR; change from baseline in PM PEFR

SYMPTOM SCORES
Change in total symptom score; % change in symptom-free days and nights; participant satisfaction; physician assessment of effectiveness; compliance with study treatment

EXACERBATIONS
Exacerbations requiring systemic steroids; exacerbations requiring hospital admission; exacerbations defined as MILD: deterioration in asthma requiring a clinically relevant increase in salbutamol use defined as more than 3 additional inhalations per 24-hour period with respect to baseline for more than 2 consecutive days. MODERATE: requiring oral CS and/or antibiotics. SEVERE: requiring hospitalisation

FUNCTIONAL STATUS
% rescue-free days; % change in use of rescue medication (puffs/d); % symptom-free days

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Serious adverse events, headache, oral thrush

WITHDRAWALS
Due to adverse effects
Overall
(reported)

(* denotes primary outcome)

Notes

Full-text report

Received additional unpublished data provided by Karen Richardson, GSK

Funded by Glaxo SmithKline, study SAS40015

Confirmation of methodology and data extraction received

User-defined order: 12 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer-generated randomisation schedule
Allocation concealment (selection bias)Low riskNumber-coded inhalers/pills supplied by pharmacy
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-dummy
Incomplete outcome data (attrition bias)
Exacerbations
Unclear risk'The ITT population (SFC: 404 subjects; FP + montelukast: 401 subjects), which included all randomised subjects, was used for adverse event and concurrent medication data. The modified ITT population (SFC: 356 subjects; FP + montelukast: 369 subjects), which excluded randomised subjects who did not receive treatment as well as subjects who were incorrectly randomised, was used for efficacy, demography and baseline characteristics data'
Selective reporting (reporting bias)Low riskData available for primary outcome
Other biasLow riskNo risk of bias was observed

SAM40030

MethodsParallel-group, multi-centre trial
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE
≤ 400 µg of BDP/d or equivalent

RANDOMLY ASSIGNED: 66
LTRA: 33
LABA: 33

WITHDRAWALS
LTRA: 4/33 (12.12%)
LABA: 9/33 (27.27%)

AGE in years: mean ± SD
35

GENDER (% male)
52

SEVERITY
Mild to moderate

BASELINE %PREDICTED FEV1 (L)
76

ALLERGEN TRIGGERS
Not reported

ALLERGIC RHINITIS
Not reported

ASTHMA DURATION in years
Not reported

ELIGIBILITY CRITERIA: age 18 to 50 years; confirmed diagnosis of asthma
Have received constant daily dose of up to 400 μg of inhaled CFC-BDP or equivalent in the last 4 weeks

During run-in period: FEV1 61% to 85% of predicted; ≥ 20% fall in FEV1 on methacholine challenge; symptom score ≥ 1 on 4/7 days; use of rescue beta2-agonists on ≥ 2/7 days; ≥ 10% period variation in PEFR over the last 7 days of run-in

EXCLUSION CRITERIA: intake of asthma medication other than ICS or short-acting beta2-agonists in the past 4 weeks; oral steroids in the past 3 months; respiratory infection within 4 weeks; hospital admission in past 12 months; evidence of underlying chronic lung disease; smoking history of 10 or more pack-years; pregnant or lactating women; other chronic diseases; use of LABA or LTRA within 1 month before visit 1; known intolerance to study drugs or inhaled lactose

SETTING: not described

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: not reported
Intervention period: 12 weeks

INTERVENTION GROUP 1
LTRA: montelukast 10 mg bid+ fluticasone (Flixotide) 100 bidtwice daily via MDI

INTERVENTION GROUP 2
LABA: salmeterol 50 µg bid via MDI + fluticasone 100 µg bid
(in single MDI: Seretide 50)

1 inhaler used for combination therapy

CO-TREATMENT: not specified

Outcomes

Modified INTENTION-TO-TREAT ANALYSES
Outcomes used at 12 weeks or endpoint

PULMONARY FUNCTION TESTS
Change from baseline in AM PEFR (L/min); change from baseline in PM PEFR (L/min); change from baseline in FEV1 (L)

SYMPTOMS (reported as medians)
Change in symptom-free days; change in symptom-free nights

EXACERBATIONS REQUIRING SYSTEMIC STEROIDS
Not reported

FUNCTIONAL STATUS (reported as medians):
Change from baseline in mean DAYTIME use of β2-agonists; change from baseline in mean NIGHT-TIME use of β2-agonists; change in rescue-free days; change in night-time awakenings

INFLAMMATORY MARKERS (reported as medians)
Sputum *eosinophils, neutrophils, total cell counts, C-LT, histamine, IL-8

ADVERSE EFFECTS
Reported

WITHDRAWALS
Reported

(* denotes primary outcome)

Notes

Unpublished data

Received full disclosure of unpublished data provided by Karen Richardson, GSK (July 2003)

Funded by GSK : study #40030

Confirmation with supportive documents received for methodology and data extraction obtained from Karen Richardson, GSK, UK

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer-generated randomisation schedule
Allocation concealment (selection bias)Low riskOpaque consecutive envelopes containing assessment
Blinding (performance bias and detection bias)
All outcomes
Low riskIdentical placebo
Incomplete outcome data (attrition bias)
Exacerbations
Unclear risk'The intention-to-treat (ITT) sample was used for the efficacy and safety analyses. This consisted of all subjects randomised to and receiving at least one dose of study medication'
Selective reporting (reporting bias)Unclear riskNot clear whether OCS exacerbations collected in the study
Other biasLow riskNo risk of bias was observed

SD-004-0216

MethodsParallel-group study; multi-centre trial (49 centres in 6 countries)
Participants

INADEQUATELY controlled participants taking ICS at baseline

BASELINE ICS DOSAGE
400 to 1000 µg of ICS (not specified)/d

RANDOMLY ASSIGNED: 236
LTRA: 118
LABA: 118

WITHDRAWALS
LTRA: 19/118 (16%)
LABA: 12/118 (10%)

AGE in years: mean ± SD
LTRA: 38.3 ± NS
LABA: 38.1 ± NS

GENDER (% male)
LTRA: 47
LABA: 49

SEVERITY
Not described

BASELINE % OF PREDICTED FEV1, L
LTRA: 72.03 ± SD
LABA: 69.71 ± SD

ALLERGEN TRIGGERS
Not reported

ALLERGIC RHINITIS
Not reported

ASTHMA DURATION in years
LTRA: 10.1 ± SD
LABA: 12.1 ± SD

ELIGIBILITY CRITERIA: male or female outpatient; age 12 to 70 years; treated for at least 3 months with 400 to 1000 µg of ICS (presumed CFC-BDP equivalent); asthma diagnosis; FEV1 50% to 80% predicted; ≥ 12% reversibility in FEV1 and at least 200 mL after inhalation of 1 mg of terbutaline; smoking history ≤ 10 pack-years

In the 7 days before randomisation, one or more of the following

  • Symptom score ≥ 1 on 4 days

  • Awakening on ≥ 1 night because of asthma symptoms

  • Use of β2-agonists ≥ 10 puffs as weekly mean

EXCLUSION CRITERIA: respiratory infection; clinical obstructive pulmonary disease or pulmonary dysfunction other than asthma; pregnant or lactating women; use of LABA within 1 month before visit 1; previous use ever of a leukotriene antagonist; known intolerance to study drugs or inhaled lactose

SETTING: not described

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION
Run-in period: 10 to 14 days
Intervention period: 8 weeks

INTERVENTION GROUP 1
LTRA: zafirlukast 20 mg bid + budesonide 200 µg bid via turbuhaler

INTERVENTION GROUP 2
LABA: formoterol 12 µg bid via turbohaler + budesonide 200 µg bid via turbuhaler

2 inhalers used for combination therapy

CO-TREATMENT: not specified

Outcomes

Modified INTENTION-TO-TREAT ANALYSES for participants who received at least one dose of medication. Outcomes used at endpoint

PULMONARY FUNCTION TESTS
*Change from baseline in AM PEFR; change from baseline in PM PEFR; change from baseline in FEV1

SYMPTOM SCORES
Change from baseline DAYTIME symptom scores; change from baseline NIGHT-TIME symptom scores

EXACERBATIONS
Exacerbations requiring systemic steroids

FUNCTIONAL STATUS
Change from baseline in mean OVERALL use of β2-agonists (puffs/d); change from baseline in mean DAYTIME use of β2-agonists (puffs/d); change from baseline in mean NIGHT-TIME use of β2-agonists (puffs/d); % night-time awakenings

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Drug-related and non-drug-related

WITHDRAWALS
Due to adverse effects reported

(* denotes primary outcome)

Notes

Unpublished data

Received full disclosure of unpublished data provided by Roger Metcalf, AstraZeneca, July 2003

Funded by Astra Zeneca. Report #SD-004CR-0216

Confirmation with supportive documents received for methodology and data extraction

User-defined number: 12 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskDescribed as randomised; no other information presented
Allocation concealment (selection bias)Low riskOpaque consecutive numbered envelopes containing assignment
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-dummy
Incomplete outcome data (attrition bias)
Exacerbations
Unclear riskAnalysis described as modified
Selective reporting (reporting bias)Unclear riskNot clear whether the study collected information on exacerbations treated with OCS
Other biasLow riskNo risk of bias was observed

Storms 2004

  1. a

    BDP: beclomethasone; DPI: dry powder inhaler; FEV1: forced expiratory volume in one second; FP: fluticasone; GSK: GlaxoSmithKline; ICS: inhaled corticosteroids; ITT: intention-to-treat; LABA: long-acting β-agonists; LTRA: leukotriene receptor antagonists; MDI: metered dose inhaler; PEFR: peak expiratory flow; challenge FEV1 % predicted: FEV1 measured post-SABA after six minutes of exercise on a treadmill exacerbating heart rate to 80% to 90% of individual's predicted maximum.

MethodsParallel-group; multi-centre study (16 centres in USA)
Participants

INADEQUATELY controlled participants taking ICS and SABA prn with history of EIB at baseline

BASELINE ICS DOSAGE
Not reported

RANDOMLY ASSIGNED: 78
LTRA: 39
LABA: 39

WITHDRAWALS
LTRA: 0 (0%); LABA: 2 (5%)

AGE in years: mean
LTRA: 33.3
LABA: 30

GENDER (% male)
LTRA: 29.2
LABA: 41

SEVERITY
Not described

BASELINE % OF PREDICTED FEV1 (L)
LTRA: 87.5
LABA: 88.1

ALLERGIC RHINITIS (%)
Not reported

ALLERGEN TRIGGERS
Not reported

ASTHMA DURATION in years: mean ± SD
LTRA: 17.4 ± 11.1
LABA: 19.7 ± 12

ELIGIBILITY CRITERIA: age 15 to 45 years with one-year history of asthma; uncontrolled asthma taking ICS for at least 2 months; treatment at randomisation with only SABA and ICS; history of EIB (15% drop in FEV1 taking ICS, 20% if not taking ICS); resting FEV1 ≥ 70% predicted; ≥ 12% increase in baseline FEV1 post-SABA; requirement for SABA on ≥3 days of last week of run-in period

EXCLUSION CRITERIA: respiratory infection within last 3 weeks and emergency asthma care in previous 3 months; systemic corticosteroids in previous month; patients were required to stop an anti-asthma medication, with the exception of ICS two weeks before first study visit; participants requiring oral steroids during the study were withdrawn

Interventions

LTRA + ICS versus LABA + ICS (stable dose of ICS)

DURATION

Run-in period: one to two weeks

Intervention period: 4 weeks

INTERVENTION GROUP 1

LTRA: montelukast @ 10 mg/d PO + placebo salmeterol inhaler + FP 100 µg bid via discus

INTERVENTION GROUP 2

LABA: salmeterol 50 µg bid via MDI + montelukast placebo + FP 100 µg bid via discus (separate inhalers)

CO-TREATMENT: SABA prn

Outcomes

INTENTION-TO-TREAT ANALYSES: outcomes used at endpoint or 4 weeks

PULMONARY FUNCTION TESTS
*Challenge FEV1 % of predicted; absolute FEV1 % of predicted; fall in FEV1 postexercise (%); rescue bronchodilation

SYMPTOM SCORES
Clinic exercise assessment score

EXACERBATIONS
None occurred during the study (requirement for OCS)

FUNCTIONAL STATUS
Not reported

INFLAMMATORY MARKERS
Not reported

ADVERSE EFFECTS
Not reported

WITHDRAWALS
Reported

Due to adverse events: reported
Due to poor control: not reported
Overall: reported

(* denotes primary outcome)

Notes

Full-text report

Funded by Merck

User-defined number: 4 weeks

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer-generated randomisation schedule
Allocation concealment (selection bias)Unclear riskInformation not available
Blinding (performance bias and detection bias)
All outcomes
Low riskDouble-dummy LTRA and LABA
Incomplete outcome data (attrition bias)
Exacerbations
Low risk'A modified intention-to-treat approach was used for efficacy analyses. For FEV1, all randomised patients who had challenge–rescue evaluations at baseline and during treatment were eligible for analyses. There was no imputation of missing values, and prior values were not carried forward'
Selective reporting (reporting bias)Low riskNo exacerbations occurred during the study
Other biasLow riskNo risk of bias was observed

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Adinoff 1998One of the interventions was not LTRA + ICS
Anon 1999Not an RCT - Montelukast vs. zafirlukast review
Anon 2000Not an RCT (review)
Anon 2001Not an RCT (review)
BADGER 2010Duplication of Lemanske 2010
Barnes 1997Not an RCT (review)
Basu 2010Duplication of Basu 2011
Basu 2011Not a randomised clinical trial; it was a cross-sectional survey.
Becker 2000Not an RCT (Review of montelukast)
Bergmann 2004One of the tested interventions was not daily LTRA as add-on to inhaled glucocorticoid
Bleecker 2006Combined analysis of two combination therapy trials versus anti-leukotriene agent alone
Borker 2005No ICS co-treatment in both groups
Bozek 2012The study used prohibited co-intervention: study compared montelukast+ICS+formoterol versus IC+formoterol
Brabson 2002No co-intervention with ICS
Buchvald 2003Study duration was less than 28 days.
Caffey 2005No ICS co-treatment in both groups
Calhoun 2001Non permitted drugs: study compared montelukast vs. placebo with both group receiving ICS and LABA
Cash 2001Not a RCT - Commentary on a previously published trial.
Chanez 2010One of the interventions was not LTRA + ICS
Chopra 2005Comparison between two different LABA + ICS combinations
Chuchalin 2002One of the interventions was not LTRA + ICS
Cowan 2010One of the interventions was not LTRA + ICS
Currie 2002No systematic co-treatment with ICS
Currie 2003aNon permitted drug : salmeterol in both groups
Currie 2003bLTRA in both groups
Currie 2003cDuration of intervention <30 days
Davis 2001No co-treatment with ICS and LTRA
Dekhuijzen 2002Not an RCT but a review article
Delaronde 2005Intervention is educational (not drug)
Dempsey 2000Single dose intervention (not > 28 days)
Deykin 2007Comparison of MON/SAL with FP/SAL
Dicpinigaitis 2002No systematic co-treatment with ICS
Donohue 2001Review of combination therapies
Dorinsky 2001No ICS used
Dorinsky 2002One of the interventions was not LTRA + ICS
Dorinsky 2002aOne of the tested interventions was not daily LTRA as add-on to inhaled glucocorticoid (no ICS in LTRA group)
Dorinsky 2004One of the tested interventions was not daily LTRA as add-on to inhaled glucocorticoid (no ICS in LTRA group)
Drummond 2011One of the interventions was not LTRA + ICS
Dunn 2001Review of zafirlukast
Edelman 2000No co-intervention with ICS
Edin 2002One of the interventions was not LTRA + ICS
Eliraz 2001No co-treatment with LTRA - Compares two dry powder inhalers
Eliraz 2002One of the interventions was not LTRA + ICS
Everden 2002One of the interventions was not LTRA + ICS
Gabrijelcic 2004One of the tested interventions was not daily LTRA as add-on to inhaled glucocorticoid
Giannini 2002One of the interventions was not LTRA + ICS
Grzelewska 2003One of the tested interventions was not daily LTRA as add-on to inhaled glucocorticoid
Gupta 2007Study assessed LTRAs in addition to LABAs.
Hanania 2011One of the interventions was not LTRA + ICS
Havlucu 2005Not an RCT
Horwitz 1998Not an RCT (Review)
Houghton 2004Comparison of propellants in formoterol - no ICS in both groups
Inouhe 2007Single dose study protocol
Jarvis 1998Not an RCT (Review of zafirlukast)
Jarvis 1999Not an RCT but a review article on Zafirlukast.
Jenkins 2005LTRA and LABA not compared as add on to ICS
Jonsson 2004One of the tested interventions was not daily LTRA as add-on to inhaled glucocorticoid
Kalberg 1999Retrospective data analysis, not an RCT
Kanniess 2002No systematic co-treatment with ICS
Kanniess 2002bOne of the interventions was not LABA + ICS
Karaman 2007No prior treatment with ICS.
Kardos 2001One of the interventions was not LABA + ICS
Katial 2010One of the interventions was not LTRA + ICS
Keith 2009Observational study
Kemp 1998Not an RCT (Review)
Kerstjens 2011One of the interventions was not LTRA + ICS
Knorr 2001No consistent co-tx with ICS in all patients (Montelukast vs placebo)
Koenig 2008Study compared LABA and LTRA without background ICS in either group.
Kohrogi 1999Not an RCT (before and after study)
Laviolette 1999One of interventions is not LABA + ICS
Lazarus 2001One of interventions is not LTRA + ICS
Lee 2004No LTRA and No LABA
Lee 2005RCT testing two types of ICS
Lee 2010No LTRA and No LABA
Leflein 2002No systematic co-treatment with ICS
Lipworth 2000Intervention < 28 days (1 week only)
Liu 1996No consistent co-treatment with ICS
(Zileuton vs. placebo as add-on therapy to ICS)
LOCCSComparison of Combination therapy with LRTA alone.
Marogna 2010the study used prohibited co-intervention (birch sublingual immunotherapy)
Maspero 2008Study compared LABA and ICS with LTRA alone
Matsuse 2012Only 14 days intervention (not >=28 days)
McCarthy 2002One of the interventions was not LTRA + ICS
McKinlay 2011Not a RCT
Meltzer 2002No co-treatment with inhaled corticosteroids
Miraglia del Giudice 2007No prior ICS treatment.
Mitchell 2005Intervention is educational (not drug)
Molitor 2005One of the interventions not LTRA
Naedele-Risha 2001Not a RCT, review of LABA+ICS therapy
Narmadha 2011One of the interventions was not LABA + ICS
Nathan 2000Good review of add-on therapy to ICS. Not an RCT
Nathan 2001bNot a RCT, review of antileukotriene agents
Nathan 2005No direct comparison between LABA and LTRA
Nelson 2004Both treatment groups received FP and Salmeterol (LTRA tested as add-on to LABA)
O'Connor 2006Not an RCT (prospective observational study)
O'Sullivan 2003One of the interventions was not LABA + ICS
Ohbayashi 2009Investigation of addition of anti-leukotriene to combination inhaled steroid and long-acting beta-agonist.
Ollendorf 2000Not an RCT, but an economic evaluation
Oppenheimer 2008Study assessed addition of anti-leukotriene (montelukast) in addition to combination LABA and ICS in asthma.
Ortega-Cisneros 1998No leukotriene antagonists used in intervention
Patel 2010the study used prohibited co-intervention (budesonide + formoterol + montelukast versus high dose budesonide + formoterol)
Paterson 1999No systematic co-treatment with ICS
Pearlman 2002No consistent co-tx with ICS (FP + S vs Montelukast alone)
Pedersen 2011Not an RCT (Post hoc analysis)
Perez 2000Not RCT - no control group, all patients treated with montelukast
Peroni 2002Short duration < 28 days
Peroni 2005Inadequate duration.
Petermann 2004Review article
Plaza 2005Intervention is educational (not drug)
Price 2003One of the interventions was not LABA + ICS
Price 2011aDuplication of price 2011.
Riccioni 2002No systematic co-treatment with ICS
Rickard 1998No systematic co-treatment with ICS
Rosenhall 2003One of the interventions is not LTRA + ICS
SAS40036LTRA administered without an ICS.
SAS40037LTRA administered without an ICS.
SAS40066LTRA administered without an ICS.
Serrier 2003One of the interventions is not LTRA + ICS
Sheth 2002Second report - cost effectiveness analyses
Sims 2003Intervention < 28 days
Smith 1998Not an RCT (Review)
Smugar 2009Duplication of study by Fogel 2010
Smugar 2009aDuplication of study by Fogel 2010
Sorkness 2007LTRA administered without an ICS.
Stanford 2003LTRA administered without an ICS.
Stelmach 2001No consistent co-intervention with ICS (RCT of ICS vs. LABA vs. LTRA )
Stelmach 2002No consistent co-intervention with ICS (RCT of ICS vs. LABA vs. LTRA vs. nedocromil )
Stelmach 2002aNo co-intervention with ICS
Stelmach 2007Participants were all on combination therapy ICS + LABA prior to enrollment and all controller medication was withdrawn for the 4-week run-in period. Neither before or during the run-in were the participants on ICS alone prior to randomisation.
Stelmach 2008Participants wereall on combination therapy ICS with either LABA or LTRA prior to enrollment and were removed from all controller medications for the 4-week run-in period: consequently they were not on ICS alone prior to enrollment
Stempel 1998Not an RCT (Review)
Stempel 2002Not an RCT (Review)
Stevenson 2005No LTRA or LABA given.
Sugihara 2010One of the interventions was not LTRA + ICS
Teper 2009One of the interventions was not LABA + ICS
Terzano 2001One of the interventions is not LTRA + ICS
Thien 2000Not an RCT (Review)
Tolley 2002One of the interventions was not LTRA + ICS
Vaquerizo 2003One of the interventions was not LABA + ICS
Volovitz 1999No consistent co-intervention with ICS in all patients (Montelukast vs. beclomethasone )
Wang 2011One of the interventions is not LTRA + ICS
Warner 2001Not an RCT (Review)
Wilson 1999Only 14 days intervention (not >=28 days)
Wilson 2001Only 14 days intervention (not >= 28 days)
Wilson 2010Duplication of price 2011
Wilson 2010aDuplication of Wilson 2010
Wilson 2010bOne of the tested interventions was not daily LTRA as add-on to inhaled glucocorticoids
Wytrychowski 2001Not an RCT - controlled study
Yasui 2012One of the interventions was not LTRA + ICS
Yurdakul 2002Not truly randomised as eligible patients were allocated to each treatment group according to their application month to hospital (consecutive alocation not random).
Zarkovic 1998One of the interventions was not LTRA + ICS
Zimmerman 2002One of the interventions was not LTRA + ICS

Characteristics of ongoing studies [ordered by study ID]

Fardon 2002

Trial name or title 
Methods 
Participants 
Interventions 
Outcomes 
Starting date 
Contact information 
Notes 

Fardon 2004

Trial name or title 
Methods 
Participants 
Interventions 
Outcomes 
Starting date 
Contact information 
Notes 

Price 2001

Trial name or title 
Methods 
Participants 
Interventions 
Outcomes 
Starting date 
Contact informationProfessor D Price (University of East Anglia, Norwich, NR4 7TJ)
NotesISSN: N0254145816

Ruggins 2003

Trial name or title 
Methods 
Participants 
Interventions 
Outcomes 
Starting date 
Contact information 
Notes 

Ancillary