Author + information
- Donald E. Cutlip, MD⁎ (, )
- Mitchell W. Krucoff, MD,
- Roxana Mehran, MD and
- Renu Virmani, MD
- ↵⁎Beth Israel Deaconess Medical Center, Interventional Cardiology, 330 Brookline Avenue, Boston, Massachusetts 02215-5400
We thank Dr. Pfisterer and colleagues for their interest in our study and for sharing their data from the BASKET (Basel Stent Cost-Effectiveness) trial. We agree our results are limited by relatively short-term follow-up after stenting (25% of patients followed beyond 1 year) and a population defined by autopsy. On the basis of our objective to assess validity of clinical adjudication according to the sensitivity and specificity of Academic Research Consortium (ARC) criteria, the autopsy data were the necessary gold standard.
The data provided by these investigators on the basis of detailed clinical assessment from a real-world clinical trial are also helpful. Given that autopsy data will not be available in all cases of suspected stent thrombosis within a clinical trial, we agree with the importance of detailed assessment of clinical data to minimize uncertainty in the cause of death as stated in our paper (1). As they suggest, such a clinical assessment that is based on the criteria proposed for modification of the ARC possible criteria could be useful to determine which late deaths are more likely to be associated with late or very late stent thrombosis.
Thus, we concur that the most accurate assessment of stent thrombosis within a clinical trial requires detailed and complete collection of data regarding the cause of death. Whenever possible, we would encourage that this include autopsy specimens of the stented segment, especially in clinical trials for which stent thrombosis is a primary endpoint of interest. We are encouraged that the clinical trial data are consistent with our findings from pathological data. Although both pathological and clinical data suggest that modification of the ARC possible criteria might improve the estimate of the true event rate, specifically related to the underreporting of events with only the more sensitive but less specific definite or probable criteria, more evidence will be needed to assess whether such a modification improves discernment of safety differences between coronary stents or pharmacological therapies designed to reduce stent thrombosis. Finally, it is important to recognize that all clinical definitions of a pathological event will necessarily be limited. Despite these imperfections, consistency inherent with the use of consensus definitions remains the key objective and enormously more informative across the published data involving the reporting of infrequent events, wherein comparability between clinical trials or “poolability” of clinical trial data might be required for optimal assessment.
- American College of Cardiology Foundation