| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Vanderbilt University Medical Center, Nashville, TN
Correspondence to: John J. Murray, MD, PhD, Associate Professor of Medicine and Pharmacology, Vanderbilt University Medical Center, 21st Ave S and Garland Ave, Nashville, TN 37203; e-mail: john.murray{at}vanderbilt.edu
To the Editor:
The recent metaanalysis that was published in CHEST (June 2004)1 examining the cardiovascular risks associated with ß-agonist therapy for obstructive lung diseases has significant implications for the millions of individuals who are currently using this therapy. The potential therapeutic impact and concerns that were raised for the management of these patients from the conclusions of this metaanalysis is demonstrated by the widespread reporting of this extremely crucial information on numerous news services.234 However, I disagree strongly that "physicians who recommend regular use of ß-agonists [such as myself] may actually be putting their patients at risk."4 The regular use of long-acting ß-agonists as maintenance therapy, as well as the use of short-acting ß-agonists as rescue medication, is certainly in agreement with both national and international evidence-based medicine guidelines for the treatment of asthma and COPD.5678 In this regard, I also strongly disagree with the conclusion drawn by the authors of this metaanalysis that short-term or long-term ß-agonist use increases the risk for untoward cardiovascular events to the point of suggesting that the recommendations of these widely accepted guidelines be revised. The fact that the information from this metaanalysis was rapidly incorporated into personal injury legal web sites9 even further speaks to the magnitude of this recommendation and the impact that this may have on the patient use of, and even the physician prescribing of, these agents.
Despite the apparent concerns raised by the authors from the interpretation of their results, I was reassured that there was no evidence of a statistically significant increase in major cardiovascular effects from ß-agonist use in their analysis. Only when the authors’ so called "minor" cardiovascular risk factor of sinus tachycardia was included did the analysis show a statistically significant increase in "risk" associated with ß-agonist use. The rationale for the inclusion of this arrhythmia as a risk factor was that an increase in heart rate portends a possible increased risk of major cardiovascular events. The reference10 cited to support this premise, however, refers to sinus tachycardia found at baseline in patients with varying cardiac disease conditions such as congestive failure or other conditions reflecting increased adrenergic stimulation. There are no data to support the concept that inducing sinus tachycardia pharmacologically is in any way a comparable marker for major cardiovascular events. Additionally, nearly two thirds of the episodes of tachycardia were contributed by only one study,11 which, importantly, was conducted in young children whose heart rates were already significantly higher at baseline than older individuals and, hence, had less of a net increase in the heart rate to produce sinus tachycardia. I would not presume that drug-induced sinus tachycardia, especially in this group of young patients, would have any of the detrimental cardiovascular associations that are asserted in this study. To this point, the purported increased cardiovascular risk (reported by the authors), which is derived solely from an "arrhythmia that can herald a poor prognosis" (as contended by the authors), appears to be unfounded based on the authors’ own data. Excluding this risk factor from the analysis reduces the total number of events by > 86%, resulting in no significant difference between the ß-agonist-treated group and the placebo-treated group.
Furthermore, I am concerned that a significant study selection bias could have played a role in the results presented. The literature search ultimately provided 20 studies for the inclusion of the long-term cardiovascular events, since only these 20 studies reported at least one extractable adverse event. An additional 115 studies were excluded because no "extractable" data were available. If indeed the latter studies were rejected for inclusion because there were in fact no events to report, the results would obviously be substantially biased.
Another important issue is the vast heterogeneity of studies dating back > 40 years, which included a variety of ß-agonists (short-acting and long-acting, and even some drugs or delivery devices that are not or never have been available in the United States) delivered by various routes of administration (eg, by mouth, IM, and nebulized). Although the weighting of each study appears to be based on the inverse of the variance, the significant test for heterogeneity for the heart rate and potassium analyses indicate that the weighting scheme may not be the most appropriate method. In addition, for the heart rate analysis, the studies by Braun and Levy12 and Burgess et al13 account for half of all patients studied but account for only one fourth of the weight, while both studies were not significant and had the smallest difference. Furthermore, for the tachycardia analysis, the study by Milgrom et al11 accounted for 62% of the events, with < 10% of the total number of patients and a weight of 23.6%, and, more importantly, it was the only significant study. Although the test for heterogeneity in this case was not significant, this one study clearly appears to be driving the results and may not be representative across all selected studies. Finally, as indicated above, all analyses may be overly biased by the omission of the 115 studies with no events that were not included. However, the analysis of major cardiovascular events indicated no significant difference, even without a correction for the studies with no events.
The most serious factor impacting the current metaanalysis was that many of the studies specifically chosen were phase II dose-ranging trials or cumulative dosing studies that were designed to evaluate the maximum tolerated doses, from the perspective of safety, for a variety of ß-agonist agents. Surprisingly, the authors included all data from these studies for adverse events, with many of the adverse events observed at higher than currently approved doses (some nearly 10-fold higher). For example, one study included in the metaanalysis11 provided nearly two thirds of the most common minor adverse event, such as sinus tachycardia, and nearly half of all the cardiovascular events. However, this study in children aged 4 to 11 years, was conducted using a nebulizer delivery system at doses up to double the recommended dose of the ß-agonist agents. When examining only those studies that did not exceed the recommended doses or in which a determination of adverse events at the clinically relevant doses could be made, one is struck by the lack of adverse effects, and a very different conclusion would be drawn. In fact, a metaanalysis14 also published in CHEST of studies in which a long-acting ß-agonist agent was administered continuously at recommended therapeutic doses over a mean period of 6 months showed no increased risk of adverse cardiovascular events compared with placebo.
Certainly, a statement advocating a concept that is at odds with current therapy and evidenced-based guidelines warrants a very careful and thoughtful analysis, which I found wanting in this article. For instance, in the list of studies in Table 1 of the article, various studies were misrepresented. For instance, the study by Burgess et al13 was listed twice, rather than the study by Burggraaf et al,15 and in the latter study salbutamol, not formoterol, was examined. In the study by D’Urzo et al16 it was indicated that nebulized salmeterol was studied despite the fact that there is no such commercially available product, and in the study by Spector and Garza Gomez17 inhaled, not nebulized, albuterol was studied. Likewise, the extraction of the data from the studies was not, in several instances, accurate (eg, reviewing Fig 3 from the metaanalysis). The study by Pearlman et al18 included three groups and hence 67 patients rather than 46 patients in the ß-agonist treatment. The studies by Spector and Garza Gomez17 and Anderson et al19 included two ß-agonist treatments rather than one, and, hence, the study groups should have been twice the number reported. The study by Nathan et al20 included only the data from the salmeterol group (4 of 369 patients), not the data for the albuterol group, and, most importantly, it improperly stated that the placebo rate was 0 of 187 patients, although this was not the information presented in the article.
Given the substantial issues that, in my view, exist with this analysis, I believe that the authors need to rethink their current conclusions and carefully consider analyzing only studies that utilize approved doses or preparations of ß-agonists and should reexamine each study to be certain that the correct number of patients are included in the analyses. If this were done, I feel certain that the authors would come to a distinctly different conclusion that would be in line with the current treatment recommendations found in the carefully substantiated guidelines for asthma and COPD.5678 Until then, I would advocate that the current guidelines for asthma and COPD, supported by a multitude of studies showing the benefit of therapy with ß-agonists, both short-term and long-term, be continued. I believe that this is in the best interest of patient care.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
