LETTER TO THE EDITOR
Composite confusion: Reply
David A. Cox, MD, FACC
Interventional Cardiologist, Mid Carolina Cardiology, 1718 East 4th Street, Suite 501, Charlotte, NC 28204, USA
Gregg W. Stone, MD, FACC
The Cardiovascular Research Foundation, New York, NY, USA
dcox{at}mccardiology.com
We thank Dr. Massel for his interest in our study (1), and though we agree with some of his points, there is one major flaw with his reasoning: The optimal percutaneous transluminal coronary angioplasty (PTCA) group was defined in the study as having near perfect angiographic results, whereas the routine stent group as defined included all patients, whether or not an optimal (or even successful!) result was obtained. Most pertinently, 100% of patients in the optimal PTCA group achieved Thrombolysis In Myocardial Infarction (TIMI) flow grade 3 (by definition), compared to only 95.7% of patients in the routine stent group (p < 0.0001), clearly explaining the weak trends toward increased mortality Massel notes. In light of this unfair playing field, it is particularly noteworthy that the benefits of stents in reducing restenosis and infarct artery reocclusion are still strongly apparent. As we stated in the Limitations section of our report, our data are hypothesis generating only; an adequately powered randomized trial of stenting versus no stenting in patients achieving optimal PTCA results is required to definitely address this issue. Such a study, the Florence Randomized Elective Stenting in Acute Coronary Occlusions (FRESCO) trial, was performed in a relatively small number of patients (n = 150) undergoing primary angioplasty using now obsolete first-generation stents, demonstrating not only marked reductions in clinical and angiographic restenosis, but also nonsignificant reductions toward reduced rates of mortality and reinfarction (2).
Where we do agree with Massel is in our disdain for composite end points. Although at times a necessary evil to allow realistic sample sizes in randomized trials, they may obscure the forest for the trees. Hierarchical rankings, unfortunately, introduce as many new problems and vagaries as they solve. A balanced perspective can usually be obtained through careful consideration of the patient populations and methods, and by judicious examination of all component end points. Finally, as important as it is to understand beta error (realizing that real differences between groups may not become statistically apparent with small sample sizes), it is equally vital to recognize that small sample sizes can also by chance suggest possible differences (or even large treatment effects) where none exist.
 |
References
|
|---|
1. Cox DA, Stone GW, Grines CL, et al. Outcomes of optimal or "stent-like" balloon angioplasty in acute myocardial infarction: the CADILLAC trial. J Am Coll Cardiol. 2003;42:971–977[Abstract/Free Full Text]
2. Antoniucci D, Giovanni M, Santoro GM, et al. A clinical trial comparing primary stenting of the infarct-related artery with optimal primary angioplasty for acute myocardial infarction. J Am Coll Cardiol. 1998;31:1234–1239[Abstract/Free Full Text]
|
Top
References