Time to Question Long-Term Safety of Routine Scheduled Inhaled β-2-agonist Treatment for COPD

How many times have I heard in my general medical clinic a patient complain that he has stopped using his anticholinergic inhaler because it doesn't seem to work as well as his β-agonist inhaler? On further questioning such patients often cite the lack of an immediate improvement in ease of breathing as their criterion for efficacy. Explaining how both inhalers may help them, but in different time frames, usually will be enough encouragement to get them to comply. I fear, however, that my patients are not the only ones who are transfixed on short-term symptomatic relief to the exclusion of long-term benefit (or harm). We physicians as well as trialists and policy makers are often complicit in this short sightedness.

The meta-analysis by Salpeter et al.¹ focuses our attention on longer-term consequences of inhaled bronchodilator therapy for COPD. Such drugs are commonly used, and guidelines recommended them with little direction as to whether anticholinergic or β-agonists or both should be used. For example, in the Global Initiative for Chronic Obstructive Lung Disease (GOLD) guideline “the choice between β2-agonist, anticholinergic, theophylline, or combination therapy depends on availability and individual response in terms of symptom relief and side effects” and “bronchodilators are prescribed on an as-needed or on a regular basis to prevent or reduce symptoms.”²

Salpeter et al. have carefully selected and analyzed nearly 2 dozen randomized controlled trials that compare β-agonist with anticholinergics. The noteworthy aspect of their analysis is that they tease out data on respiratory outcomes that are undoubtedly of clinical importance, but for which individual trials are seldom designed to detect as primary outcome measures. This meta-analysis thus provides an opportunity to see data in a new way by aggregating individual controlled trial data on infrequent outcomes. This approach is a double-edged sword. On the one hand, it improves power; on the other, it emphasizes events that may not have been subjected to the same scrutiny in collection as primary outcomes.

Salpeter et al. found that inhaled anticholinergics are associated with fewer COPD exacerbations associated with withdrawal from trials and hospitalization when compared with β-agonists; in addition, there was a trend toward more respiratory deaths among patients treated with inhaled β-agonists than among those treated with inhaled anticholinergics.

Meta-analyses have sometimes met with harsh criticism particularly when the results go against conventional wisdom. Occasionally, meta-analyses may be misleading due to publication bias, incomplete (and biased) collection of primary studies, inaccurate transcription of data, or flawed analysis of data. For example, one recent meta-analysis of the efficacy of inhaled corticosteroids was shown to overestimate the effectiveness of these medications because multiple exacerbations for an individual patient were counted the same as exacerbations in different patients. This type of analysis ignored the variability in the number of exacerbations per patient in a disease where multiple exacerbations are commonplace.³

Salpeter et al. have conducted a rigorous meta-analysis of good quality data; they have steered clear of several of the methodological pot holes described above. They tested for and found no evidence of heterogeneity. In addition, they have evaluated both direct comparisons (trials of β-agonist vs anticholinergics) and indirect comparisons (trials of β-agonists vs placebo or anticholinergics vs placebo). Similar findings from these complementary approaches suggest the conclusions are robust. However, the findings are not identical; in direct comparisons, the increase in respiratory death for inhaled β-agonists compared with anticholinergics was not statistically significant.

So how should clinicians respond to these data? For one, we must recognize that this analysis portrays only a part of the data required for evaluating inhaled therapy in stable COPD. The effectiveness of these drugs hinges not only on the risk of adverse effects, but also on the benefits of treatment and, in particular, on the relative value one places on these risks and benefits.

The relative benefit of inhaled ipratropium vs inhaled short-acting β-agonists in stable COPD was reported by a recent systematic review from the Cochrane Collaboration.⁴ This review showed a small benefit to ipratropium alone or in combination with short-acting β-agonist when compared to inhaled short-acting β-agonist alone. Taken with the current analysis, these data suggest that use of inhaled ipratropium for regular scheduled therapy in stable COPD is both slightly more effective and somewhat less risky than using short-acting inhaled β-agonists in the same way.

These data should not be ignored. I look forward to seeing how the several guidelines on the management of stable COPD will deal with these new insights. In the meantime, I may try withdrawing the regular scheduled use of inhaled β-agonists in my patients with stable COPD, instructing them instead to use it on an as needed basis only for symptomatic relief. I will have to rethink my use of combined β-agonist and ipratropium inhalers, a practice I adopted several years ago to improve my patients' compliance with ipratropium based on the results of one of the trials included by Salpeter et al.⁵ And, I will likely use an inhaled anticholinergic before adding a β-agonist for newly diagnosed patients. It is important to note that such changes are all consistent with the GOLD guidelines as currently written.

I may have difficulty in explaining to my patients that as needed use of inhaled β-agonists for symptomatic relief is permitted while discouraging routine scheduled use. Particularly difficult will be explaining why I would suggest resuming regular inhaled β-agonists in patients who do not tolerate their withdrawal. Indeed, sometimes short-term benefits do trump longer-term risks; the important thing is making those decisions with complete and accurate information and with eyes wide open!