Most clinical trials for asthma drugs exclude ~95% of potential subjects and test under highly controlled conditions, limiting their results' generalizability. Price et al publish results of 2 "pragmatic" open-label trials set in the real world. In #1, they randomized 300 symptomatic asthmatics in 53 U.K. primary care clinics to get either a leukotriene receptor antagonist (LTRA) or inhaled steroid as 1st-line therapy for asthma. In the other, 350 were randomized to LTRA or long-acting beta-agonist (LABA) as step-up/add-on therapy for asthma uncontrolled by an inhaled steroid.
In both trials, after 2 months the therapies were proved equivalent on the primary endpoint (asthma quality-of life scores) as determined by predefined criteria. At 2 years, equivalence could no longer be proved. As an editorial and "Perspective" point out, high rates of nonadherence (~70% adherence for LTRA, ~43% for inhalers) and crossover likely biased the study toward the null over the longer term. The U.K. government funded the study. (n=650) NEJM 2011;364:1695-1707.