Abstract
Background: High doses of maintenance inhaled corticosteroids (ICS) in asthma may achieve only modest additional clinical benefit beyond low-to-medium doses and are associated with an increased risk of adverse systemic effects. The ICS dose-response relationship when administered as maintenance combination ICS/long-acting beta2-agonist (LABA) therapy is uncertain.
Research question: What is the ICS dose-response of maintenance ICS/LABA therapy?
Methods: A systematic review of randomized controlled trials (RCTs) allocating participants to >1 ICS dose category, per Global Initiative for Asthma categorization, administered in combination ICS/LABA inhalers was conducted. Meta-analysis compared outcomes of high-dose (HD) and medium-dose (MD), HD and low-dose (LD), and MD and LD ICS/LABA. The primary outcome was the proportion of participants with ≥1 severe asthma exacerbation; secondary outcomes were patient reported outcome measures of asthma control, spirometry and serious adverse events. Certainty of evidence was assessed using the Grading of Recommendations, Assessment, Development and Evaluations domains.
Results: We identified 12 RCTs (6373 participants), seven comparing HD versus MD ICS/LABA, one HD versus LD ICS/LABA, and four MD versus LD ICS/LABA. HD versus MD ICS/LABA reduced the odds of a severe asthma exacerbation: Peto's Odds Ratio 0.81 (95% CI 0.67 to 0.98) with high certainty. There were no other clinically important differences in efficacy or safety outcomes of HD versus MD ICS/LABA. There was no difference in all outcomes comparing HD with LD or MD with LD ICS/LABA.
Interpretation: Maintenance HD ICS/LABA reduces the odds of a severe exacerbation by about 20% compared to MD ICS/LABA. The absolute reduction in severe exacerbation risk with HD ICS/LABA is determined by patients' exacerbation risk, and this effect size may be clinically relevant for patients if this risk is high. Comparisons of other doses of ICS/LABA were limited by the number of identified studies although no large difference in effect sizes were observed.