Fourteen RCTs (5,768 children) were included in the review. All trials were randomised with the use of double-blinding. Methods of randomisation were described in five trials. Allocation concealment was reported in one trial. There were nine placebo-controlled trials. Six trials used active comparators. Three trials had a duration of 52 weeks, 10 trials of 12 weeks duration and one trial 54 weeks.
There were statistically significant benefits observed with moderate doses of inhaled corticosteroids in FEV1 (SMD 0.11, 95% CI 0.01 to 0.21; six trials; 1,601 children) compared with low doses. Moderate, non significant benefits were observed in peak expiratory flow in the morning and evening, asthma symptom scores, and the requirement for beta-2-agonist use. No statistically significant heterogeneity observed across the trials except for beta-2-agonist use (I2=65%). There were no differences between moderate and low dose regimens in withdrawals due to lack of efficacy.
There were no significant differences observed between low and moderate doses of inhaled corticosteroids on linear growth (two trials) or on local adverse events and withdrawal due to adverse events.
The most common adverse event was pharyngitis/sore throat (overall rate 7.7%). Some heterogeneity was observed by dysphonia/hoarseness (I2=39%).
Planned subgroup analyses were not performed because of the small numbers of included trials. The sensitivity analysis using three different statistical models found similar results.