Meta-analysis: effects of adding salmeterol to inhaled corticosteroids on serious asthma-related events.

BACKGROUND Recent analyses have suggested an increased risk for serious asthma-related adverse events in patients receiving long-acting beta-agonists. PURPOSE To examine whether the incidences of severe asthma-related events (hospitalizations, intubations, deaths, and severe exacerbations) differ in persons receiving salmeterol plus inhaled corticosteroids compared with inhaled corticosteroids alone. DATA SOURCES The GlaxoSmithKline (Research Triangle Park, North Carolina) database, MEDLINE, EMBASE, CINAHL, and the Cochrane Database of Systemic Reviews (1982 to September 2007) were searched without language restriction. STUDY SELECTION Randomized, controlled trials reported in any language that compared inhaled corticosteroids plus salmeterol (administered as fluticasone propionate/salmeterol by means of a single device or concomitant administration of inhaled corticosteroids and salmeterol) versus inhaled corticosteroids alone in participants with asthma. DATA EXTRACTION Three physicians independently reviewed and adjudicated blinded case narratives on serious adverse events that were reported in the GlaxoSmithKline trials. DATA SYNTHESIS Data from 66 GlaxoSmithKline trials involving a total of 20 966 participants with persistent asthma were summarized quantitatively. The summary risk difference for asthma-related hospitalizations from these trials was 0.0002 (95% CI, -0.0019 to 0.00231; P = 0.84) for participants receiving inhaled corticosteroids plus salmeterol (n = 35 events) versus those receiving inhaled corticosteroids alone (n = 34 events). One asthma-related intubation and 1 asthma-related death occurred among participants receiving inhaled corticosteroids with salmeterol; no such events occurred among participants receiving inhaled corticosteroids alone. A subset of 24 trials showed a decreased risk for severe asthma-related exacerbations for inhaled corticosteroids plus salmeterol versus inhaled corticosteroids alone (risk difference, -0.025 [CI, -0.036 to -0.014]; P <0.001). LIMITATIONS The included trials involved selected patients who received careful follow-up. Only 26 trials were longer than 12 weeks. Few deaths and intubations limited the ability to measure risk for these outcomes. CONCLUSION Salmeterol combined with inhaled corticosteroids decreases the risk for severe exacerbations, does not seem to alter the risk for asthma-related hospitalizations, and may not alter the risk for asthma-related deaths or intubations compared with inhaled corticosteroids alone.