Long-acting beta-agonists and asthma death: how useful are different study designs to evaluate the potential association?

BACKGROUND AND OBJECTIVE: There is uncertainty whether long-acting beta-agonist (LABA) drugs may increase the risk of asthma mortality. This uncertainty is partly due to the difficulty in estimating the risk of rare adverse outcomes. The aim of this paper is to consider the utility of three approaches to determine the magnitude and statistical significance of this potential association. METHODS: Using a death rate of 9 per 10,000 subjects with asthma, derived from a meta-analysis of randomized controlled trials (RCTs) of formoterol, power calculations for a single RCT, a case-control study, and a meta-analysis of RCTs were determined. RESULTS: For each study design, the number of subjects and events required to have adequate statistical power to detect a 1.5- and 2.0-fold increased risk of death were calculated. For a single RCT, or meta-analyses of RCTs, very large sample sizes are required. In contrast, case-control methodology represents a realistic method of estimating the risk of rare serious adverse events. CONCLUSIONS: Major practical limitations exist in the use of RCTs to determine the potential risk of death with LABAs in the treatment of asthma. Case-control methodology may be more effective in establishing causation; however, if selection bias occurs, estimates of risk may be inaccurate.