authors’ reply Bertil Lindmark of Astra Draco makes five points about our study on the risk of non-fatal cardiac failure and ischaemic heart disease with long acting β₂agonists. Firstly, he points out that all β₂ agonists should be used with caution in patients with severe cardiovascular disease. The cardiac effects of β₂ agonists are well described,1-1 but there is limited evidence available on whether or not the risks of adverse cardiac effects differ depending on the dose and method of administration of the drug. Clearly, these are important questions for prescribing doctors faced with treating asthmatic patients with concomitant cardiac disease. An observational cohort study formed from health insurance databases from the Province of Saskatchewan, Canada found an increased relative risk of death from cardiovascular disease in users of β₂ agonists taken orally or by nebuliser, but not in users of β₂ agonists administered by metered dose inhaler.1-2 The deaths occurred in patients with significant cardiac disease, suggesting that β₂ agonists taken orally or by nebuliser should be avoided in patients at high risk of cardiovascular events. We found that the oral β₂ agonist bambuterol, but not the inhaled β₂ agonist salmeterol, was associated with an increased risk of non-fatal cardiac failure. The results from both these studies are plausible as oral β agonists provide a greater systemic dose than that achieved with metered dose inhalers1-1 and tachycardia and prolonged Q-T interval have been reported principally with nebulised or oral β agonists.1-2 The advised total daily dose of oral bambuterol (20 mg) is 200 times that of inhaled salmeterol (100 μg).1-3

Secondly, Dr Lindmark reiterates our point that the results must be interpreted with caution because the study was observational, and more definitive evidence would come from a prospective randomised trial. Nevertheless, hypotheses about drug safety concerns are often generated from observational studies.1-4 Such studies drive further research because they provide an “a priori” hypothesis and allow the formulation of a clinically relevant end point. Until results from prospective trials become available, observational research using cohort or case-control techniques remains an important source of evidence about the safety of drugs.

Thirdly, he states that a review by Astra Draco has found no evidence from pre-marketing or post-marketing studies of an association between bambuterol and cardiac failure. In general, pre-marketing studies have their own limitations,1-5 as evidenced by the recent withdrawal on safety grounds of two newly launched drugs.1-6Similarly, different types of post-marketing surveillance studies, including PEM, have different advantages and disadvantages and, in general, one system cannot be relied upon to provide all the evidence needed.1-7 This point also applies in response to the fourth comment. In particular, it should be noted that there is gross under-reporting of suspected adverse drug reactions to the Committee on Safety of Medicines1-8 and other regulatory authorities, and there are many difficulties associated with interpreting data from spontaneous reporting schemes.1-5

Finally, as is stated clearly in our paper, it is possible that the association may be explained by factors such as confounding by concomitant disease and disease severity. Interestingly, the rate of cardiac failure associated with bambuterol in the first month of treatment was higher than for 11 cardiovascular drugs previously studied by PEM (table 1-1). Only two cardiac drugs, including the inotropic sympathomimetic xamoterol (licensed for use in mild heart failure) had higher rates of cardiac failure. Since it is highly unlikely that the rate of cardiovascular disease in the bambuterol cohort was higher than in cohorts of patients taking cardiac drugs, and the bambuterol cohort was the youngest, these data provide further evidence that an association cannot be discounted. Our findings require confirmation, but we remain concerned about the size and biological plausibility of the association.