CORRESPONDENCE: LETTERS TO THE EDITOR
Reply
Patricia A. Cowper, PhD*,
Krishna Udayakumar, MD, MBA,
Michael H. Sketch, Jr, MD and
Eric D. Peterson, MD, MPH, FACC
* Duke Clinical Research Institute, P.O. Box 17969, Durham, NC 27715 (Email: cowpe001{at}mc.duke.edu).
Dr. Eriksson is correct in noting that our results hinge on the ability of clopidogrel to reduce the incidence of myocardial infarction (MI) between one month and one year after percutaneous coronary intervention (PCI), a point we emphasized in our report (1). Although Dr. Eriksson is also technically correct in pointing out that, as the confidence interval for relative risk (RR) for MI ranged from 0.3 to 1.0, there is a possibility that Clopidogrel has no effect on MI, we do not consider this to be a likely scenario. The point estimate (0.56) is the best estimate of RR. Because the Clopidogrel for the Reduction of Events During Observation (CREDO) trial was not powered to detect differences in individual components of the combined end point, borderline significance (p = 0.05) for MI, one of the components, is not surprising.
Dr. Eriksson also expressed concern that the per-protocol population may not have been balanced between groups with respect to confounders, causing the analysis to favor clopidogrel therapy. However, we found no differences between the treatment groups in baseline clinical characteristics of per-protocol patients. Furthermore, although the reduction in MI between one month and one year in the intention-to-treat population was not significant (p = 0.14), a consistent beneficial effect was observed for all components of the combined end point (MI, death, stroke), with a significant reduction in the combined end point (2).
Dr. Eriksson also suggested that a potential negative effect of bleeding on life expectancy should have been incorporated into the model. Although bleeding may be associated with worse outcomes, the extent to which bleeding independently increases the risk of death (rather than being associated with other factors predictive of death) has not been established. Given the low baseline rate of bleeding in the CREDO trial, incorporating such an effect would have a negligible impact on our results in any case.
Finally, Dr. Eriksson postulated that the low compliance rate in the CREDO trial obscured the risk of bleeding. It is true that the risk of bleeding in the CREDO trial may have been underestimated because of the low compliance rate ( 62%). By the same token, the benefits of treatment in the CREDO trial may also have been underestimated. Compliance also affects the cost of therapy. In the cost-effectiveness analysis, we adopted the conservative approach of assigning to the treatment group the full cost of clopidogrel (assuming full compliance), rather than the actual cost of clopidogrel based on the observed compliance rate in the CREDO trial.
In conclusion, we believe our analysis incorporates the best available estimates of the clinical effectiveness of long-term clopidogrel therapy. We look forward to the results of the Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) study, which should further inform discussion regarding the value of long-term clopidogrel therapy.
 |
References
|
|---|
1. Cowper PA, Udayakumar K, Sketch Jr MH, Peterson ED. Economic effects of prolonged clopidogrel therapy following percutaneous coronary intervention J Am Coll Cardiol 2005;45:369-376.[Abstract/Free Full Text]
2. Steinhubl SR, Berger PB, Mann III JT, et al. Early and sustained dual oral antiplatelet therapy following percutaneous coronary interventiona randomized controlled trial. JAMA 2002;288:2411-2420.[Abstract/Free Full Text]
|
Top
References