We read with interest the study by Palmer and colleagues.1 Their findings suggest that patients who are Arg/Arg homozygous at position 16 of the β₂ adrenoceptor may be at increased risk of asthma exacerbations, particularly when treated with inhaled salmeterol. These results add to the body of evidence questioning the safety of long-acting β₂ agonists in this subgroup of patients with asthma.2 ,3 The authors imply that the Arg16 genotype may identify individuals who respond poorly when salmeterol is added to the use of regular inhaled corticosteroids and advocate further prospective randomised controlled trials in this area.

We recently published the results of a randomised double-blind placebo-controlled crossover study of treatment given in addition to inhaled corticosteroids in patients with symptomatic asthma.3 Compared with heterozygotes and Gly-Gly homozygotes, patients homozygous for Arg16 responded less well when formoterol was added to budesonide 100 μg twice daily for 1 month. Treatment was associated with a deterioration in methacholine hyperresponsiveness and forced expiratory volume in 1 s (FEV₁) that was both statistically and clinically significant (doubling dose change in concentration of methacholine provoking a fall in FEV₁ of 20% or more (PC₂₀) −1.08 vs 1.05, mean difference −2.13 (95% CI −0.04 to −4.2), p = 0.046; change in FEV₁ −0.38 l vs +0.02 l, mean difference 0.40 (95% CI 0.09 to 0.71), p = 0.014). While the number of patients we studied was small (n = 37, of which 6 were Arg16 homozygotes), our study was fully blinded with respect to both treatment allocation and to β₂ adrenoceptor genotype, and we believe that the large deterioration in objective markers of airway physiology seen following treatment with formoterol supports the suggestion that long-acting β₂ agonists as a class may be harmful in this minority of patients. We too would encourage further work in this important area.