| 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 |
|||||||||
St. George’s University of London, London, UK
Correspondence to: Christopher Cates, BM, BCh, St. George’s University of London, Community Health Sciences, Cranmer Terrace Tooting, London SW17 0RE, UK; e-mail: ccates{at}sgul.ac.uk
To the Editor:
Currie et al (October 2005)1 in their review article on add-on therapy in patients with persistent asthma concluded that "the studies highlighted in the review article have shown that the addition of an LRTA (leukotriene receptor antagonist) to an inhaled corticosteroid is generally as effective at reducing exacerbations as adding a LABA (long-acting ß2-agonist)." This conclusion is based on the fact that in the two largest studies the difference was statistically and clinically nonsignificant.23
The authors decided not to combine the trials in this review due to differences between the trials, but this decision carries the risk of a type 2 error as the individual studies may not have the power to demonstrate an important difference in exacerbations. The confidence interval around the relative risk (RR) of exacerbations for the two large studies is quite wide (Bjermer study: 95% confidence interval [CI], 0.72 to 1.16; Ilowite study: 95% CI, 0.67 to 1.08). These studies therefore include the possibility of a roughly 30% reduction in RR of exacerbation with LABA and cannot be cited as evidence that an LRTA is generally as effective as a LABA.
Our Cochrane review4 addressing the same question was published in January 2006. We chose to combine the results of the same studies and found that overall the "risk of exacerbations requiring systemic corticosteroids was significantly lower with LABA + ICS (inhaled corticosteroid) when compared to LTRA + ICS (RR = 0.83, 95% CI = 0.71, 0.97): the number needed to treat with LABA compared to LTRA, to prevent one exacerbation over 48 weeks, was 38 (95% CI: 23 to 247)."
Perhaps this gives an alternative picture of the size of the possible difference between LABAs and LRTAs, and allows the reader to draw their own conclusions about the clinical significance of the difference that was found.
References
Aberdeen Royal Infirmary, Aberdeen, Scotland Cambridge Hospital, Cambridge, UK
Correspondence to: Graeme P. Currie, MD, Department of Respiratory Medicine, Aberdeen Royal Infirmary, Foresterhill, Aberdeen, Scotland, UK; e-mail: graeme.currie{at}nhs.net
To the Editor:
We thank Dr. Cates for his interest in and comments regarding our recent systematic review (October 2005).1 He quite correctly points out that the two largest multicenter trials23 comparing the effects of leukotriene receptor antagonists vs long-acting ß2-agonists when used as add-on therapy to inhaled corticosteroids demonstrated no statistically or clinically significance differences between exacerbation rates. It is pertinent to point out that the authors of both these studies23 considered that there was in fact sufficient power to detect potential differences in their primary end point of exacerbation frequency.
Due to the heterogeneity of the trials highlighted in our review, we believed that performing a metaanalysis per se risked arriving at a relatively meaningless conclusion. Indeed, one of the main problems with any metaanalysis is the potential to include a hotchpotch of different studies and patient populations in an attempt to arrive at a uniform answer to suit all. For example, one must consider whether it is appropriate and valid to merge together asthmatics of different severities who are using different drug delivery devices, different inhaled corticosteroid doses, different long-acting ß2-agonists with different pharmacologic properties who take part in studies of different duration and different design, all of which have different primary end points. It must surely stand to reason that an adequately powered, well-conducted, randomized controlled trial with well-defined inclusion criteria should remain the "gold standard" tool by which clinicians extrapolate answers to clinical questions.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
