Background: In asthmatic patients inadequately controlled on inhaled corticosteroids and/or those with moderate persistent asthma, two main options are recommended: the combination of a long-acting inhaled beta2 agonist (LABA) with inhaled corticosteroids (ICS) or use of a higher dose of inhaled corticosteroids.

Objectives: To determine, in asthmatic patients, the effect of the combination of long-acting beta2 agonists and inhaled corticosteroids compared to a higher dose of inhaled corticosteroids on the incidence of asthma exacerbations, on pulmonary function and on other measures of asthma control and to look for characteristics associated with greater benefit for either treatment option.

Search strategy: We identified randomized controlled trials (RCTs) through electronic database searches (MEDLINE, EMBASE and CINAHL), bibliographies of RCTs and correspondence with manufacturers until April 2004.

Selection criteria: RCTs were included that compared the combination of inhaled LABA and ICS to a higher dose of inhaled corticosteroids, in children aged 2 years and older, and in adults with asthma.

Data collection and analysis: Studies were assessed independently by two authors for methodological quality and data extraction. Confirmation was obtained from the trialists when possible. The primary endpoint was rate of patients experiencing one or more asthma exacerbations requiring oral corticosteroids. Secondary endpoints included pulmonary function tests (PFTs), symptoms, use of rescue beta2 agonists, adverse events and withdrawal rates. The meta-analysis was done with RevMan Analyses and the meta-regression, with Stata.

Main results: Of 593 citations identified, 30 (three pediatric; 27 adult) trials were analysed recruiting 9509 participants, including one study providing two control-intervention comparisons. Only one trial included corticosteroid-naive patients. Participants were symptomatic, generally (N=20 trials) presenting with moderate (FEV1 60-79% of predicted) rather than mild airway obstruction. Trials tested the combination of salmeterol (N=22) or formoterol (N=8) with a median of 400 mcg of beclomethasone or equivalent (BDP-eq) compared to a median of 800 to 1000 mcg/day of BDP-eq. Trial duration was 24 weeks or less in all but four trials. There was no significant group difference in the rate of patients with exacerbations requiring systemic corticosteroids [N=15, RR=0.88 (95% CI: 0.77, 1.02)]. The combination of LABA and ICS resulted in greater improvement from baseline in FEV1 [N=7, WMD=0.10 L (95% CI: 0.07, 0.12)], in symptom-free days [N=8 , WMD=11.90% (95% CI:7.37, 16.44), random effects model], and in the daytime use of rescue beta2 agonists than a higher dose of ICS [N=4, WMD= -0.99 puffs/day (95% CI: -1.41, -0.58), random effects model]. There was no significant group difference in the rate of overall adverse events [N=15, RR=0.93 (95% CI: 0.84, 1.03), random effects model], or specific side effects, with the exception of a three-fold increase rate of tremor in the LABA group [N= 10, RR=2.96 (95%CI: 1.60, 5.45)]. The rate of withdrawals due to poor asthma control favoured the combination of LABA and ICS [N=20, RR=0.69 (95%CI: 0.52, 0.93)].

Authors' conclusions: In adult asthmatics, there was no significant difference between the combination of LABA and ICS and a higher dose of ICS for the prevention of exacerbations requiring systemic corticosteroids. Overall, the combination therapy led to greater improvement in lung function, symptoms and use of rescue beta2 agonists, (although most of the results are from trials of up to 24 weeks duration). There were less withdrawals due to poor asthma control in this group than when using a higher dose of inhaled corticosteroids. Apart from an increased rate of tremor, the two options appear safe although adverse effects associated with long-term ICS treatment were seldom monitored.