Ninety-two RCTs were included in the review (n>74,000 patients). Summary trial quality scores ranged from 3 to 5 out of 5 points; all trials were reported to be double blinded. Details on individual items fulfilled were not reported. Trial duration ranged from four to 56 weeks.
Long-acting beta agonist versus placebo (40 RCTs, n>40 000 patients): Long-acting beta agonists resulted in a significant decrease in asthma exacerbations requiring systemic corticosteroids (RR 0.80, 95% CI 0.73 to 0.88; NNT 24 patients, 95% CI 17 to 39; 24 RCTs) and withdrawals related to asthma exacerbations (RR 0.68, 95% CI 0.61 to 0.76; NNT 68 patients, 95% CI 53 to 93; 24 RCTs) compared with placebo. There was no significant difference between treatment groups for the incidence of acute exacerbations requiring hospitalisation (17 RCTs), or incidence of life threatening exacerbations (nine RCTs). However, there was an increased risk of asthma related deaths among patients receiving long-acting beta agonists (RR 3.83, 95% CI 1.21 to 12.14; NNH 1,226 patients, 95% CI 703 to 10,585; two RCTs). There was no evidence of heterogeneity for any of these analyses. Some outcomes showed evidence of publication bias.
Long-acting beta agonist plus inhaled corticosteroids versus inhaled corticosteroids: (57 RCTs, n=34 747 patients) Compared with inhaled corticosteroids, long-acting beta agonists plus inhaled corticosteroids resulted in a significant decrease in asthma exacerbations requiring systemic corticosteroids (RR 0.73, 95% CI 0.67 to 0.79; NNT 20 patients, 95% CI 16 to 26; 30 RCTs), asthma exacerbations requiring hospitalisation (RR 0.58, 95% CI 0.45 to 0.74; NNT 135 patients, 95% CI 90 to 282; 25 RCTs), and withdrawals related to asthma exacerbations (RR 0.64, 95% CI 0.52 to 0.78; NNT 91 patients, 95% CI 54 to 310; 25 RCTs) . There was no significant difference between treatment groups for incidence of life threatening exacerbations (11 RCTs) or asthma related deaths (three RCTs). There was no evidence of heterogeneity for any of these analyses. There was evidence of publication bias for withdrawals due to acute exacerbations, but not for other outcomes.
Most subgroup analyses showed similar results across groups.