Background: A case-control study was undertaken to investigate the hypothesis that the use of the long acting beta agonist salmeterol increases the risk of a near-fatal attack of asthma.
Methods: The cases comprised admissions to the intensive care unit (ICU) for asthma in 14 major hospitals within the Wessex region in 1992. For each of the cases four age-matched controls were selected from asthma admissions to the same hospital during the same period. Information on prescribed drug therapy for the 48 cases and 185 controls was collected from the hospital admission records.
Results: The patients admitted to the ICU had greater chronic asthma severity and had generally been prescribed more asthma drugs than the control admissions to hospital. The relative risk of a near-fatal attack of asthma in patients prescribed inhaled salmeterol was 2.32 (95% CI 1.05 to 5.16), p = 0.04. However, the salmeterol relative risk decreased to 1.42 (95% CI 0.49 to 4.10), p = 0.52 when the analysis was restricted to the more chronically severe patients (those in the subgroup of patients with a hospital admission for asthma in the previous 12 months). These findings suggest that the increased unadjusted relative risk with salmeterol is predominantly due to confounding by severity--that is, the increased relative risk is due to patients with more severe asthma (at greatest risk of a near-fatal asthma attack) being preferentially prescribed salmeterol. This interpretation is supported by the finding in this study that, within the control group (selected from the population of asthmatics requiring hospital admission), salmeterol was preferentially prescribed to the most severe patients (a threefold greater prescription of salmeterol to control patients if they had been admitted to hospital in the 12 months prior to the index admission). There was no increased risk of a near-fatal attack of asthma in patients prescribed a beta agonist by metered dose inhaler (OR 0.75 (95% CI 0.31 to 1.78), p = 0.51). In contrast, the relative risks for beta agonists delivered by nebulisation (OR 3.86 (95% CI 1.99 to 7.50), p < 0.001) and oral theophylline (OR 2.45 (95% CI 1.26 to 4.78), p < 0.01) were increased and did not markedly decrease when the analysis was restricted to the more severe asthmatic subjects.
Conclusions: Although these findings are not conclusive, particularly because of the small numbers involved in some subgroup analyses, they suggest that the use of salmeterol by patients with chronic severe asthma is not associated with a significantly increased risk of a near-fatal attack of asthma. If a near-fatal asthma attack is considered to be an intermediate step in a process by which a severe attack of asthma may become fatal, these results would suggest that salmeterol is unlikely to be associated with an increased risk of death, at least by this mechanism.