Letter by Rosenstein and Parra Regarding Article, “Greater Clinical Benefit of More Intensive Oral Antiplatelet Therapy With Prasugrel in Patients With Diabetes Mellitus in the Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel: Thrombolysis in Myocardial Infarction 38”
To the Editor:
In the article by Wiviott et al,1 we read about the use of prasugrel compared with clopidogrel in a subgroup of the original Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel: Thrombolysis in Myocardial Infarction (TRITON–TIMI) 38. We have several concerns about this article.
First, when interpreting a subgroup analysis from within a larger trial, regardless of whether it is post hoc or prespecified, it is essential for the reader to know how many comparisons (or subgroups) are being made. The risk of a false-positive result increases with an increasing number of subgroups, and the traditional probability value of <0.05 should not be used to determine statistical significance. For example, in this trial, if 15 subgroups were prespecified, then it might be reasonable to consider a P value of 0.05/15 or <0.0033 as significant. We believe that information about the number of different comparisons needs to be transparent, which would allow complete interpretation of the data.2
Second, the reduction in the primary end point appears to be driven by the reduction in the lower incidence of myocardial infarction, which occurred in the prasugrel group, and this difference appeared early (within the first 3 days when the efficacy of the thienopyridine clearly favored prasugrel because of the particulars of the dosages and loading conditions used in this study). Beyond this time frame, prasugrel did not appear to yield an advantage, yet bleeding risk may be cumulative.3
We are concerned that large subgroups of patients have failed to show significant efficacy with prasugrel (women aged >65 years and creatinine clearance <60 mL/min). In these patients, a higher risk of bleeding is found with prasugrel, which would suggest a net clinical harm in these groups.4
We concur with the authors that prasugrel may offer potential benefit over clopidogrel in selected groups of patients. This study highlights the possibility that diabetic patients may derive a particular benefit from more intense antiplatelet therapy in the acute coronary syndrome setting undergoing percutaneous coronary intervention. Certainly, this subgroup analysis has been thought provoking and would hopefully generate further prospective randomized trials of various antiplatelet strategies in this setting, which could test the hypothesis put forward by these results.
Wiviott SD, Braunwald E, Angiolillo DJ, Meisel S, Dalby AJ, Verheugt FWA, Goodman SG, Corbalan R, Purdy DA, Murphy S, McCabe CH, Antman EM. Greater clinical benefit of more intensive oral antiplatelet therapy with prasugrel in patients with diabetes mellitus in the Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition with Prasugrel: Thrombolysis in Myocardial Infarction 38. Circulation. 2008; 118: 1626–1636.
Wiviott SD, Braunwald E, McCabe CH, Montalescot G, Ruzyllo W, Gottlieb S, Neumann FJ, Ardissino D, De Servi S, Murphy SA, Riesmeyer J, Weerakkody G, Gibson CM, Antman EM. Prasugrel versus clopidogrel in patients with acute coronary syndromes. N Engl J Med. 2007; 357: 20: 2001–2015.