CORRESPONDENCE: LETTER TO THE EDITOR
Reply
Stuart J. Pocock, PhD,
Martin P. Fahy, MSc,
Roxana Mehran, MD, FACC and
Gregg W. Stone, MD, FACC*
* The Cardiovascular Research Foundation, Columbia University Medical Center, 111 East 59th Street, 11th Floor, New York, New York 10022 (Email: gs2184{at}columbia.edu).
We appreciate the insightful comments of Drs. Agostini and Biondi-Zoccai about our study (1). Regarding their first point, we felt it was relevant to concentrate on the direct comparisons of late loss (LL) and percentage diameter stenosis (%DS) with target lesion revascularization (TLR) in the same patients. We recognize the potential impact of the oculostenotic reflex on risk of TLR, but can confirm that the treatment differences in TLR in angiographic and clinical cohorts were in fact similar in those large trials with over 150 patients in both cohorts (DELIVER, ENDEAVOR II, SIRIUS, and TAXUS IV). We have also previously reported that in the TAXUS IV trial routine angiographic follow-up increased TLR rates by approximately 40%, though to a similar degree with bare-metal stents (BMS) and drug-eluting stents (DES) (2).
Regarding their second point, we distinguish between DES and BMS arms in the visual plots of mean LL or %DS and TLR rates, so readers can readily assess the association for DES arms alone. In the 11 trials, the distributions of LL in both the DES and BMS arms are not bimodal but somewhat skewed to the right.
That is why we propose the use of the predictive formulas in the Appendix to work out what TLR rate one expects for a given DES, given the observed distributions (rather than just the mean) of LL or %DS.
We have an alternative explanation for the apparent paradox whereby highly significant differences in LL and %DS for 2 different DES nevertheless lead to somewhat similar TLR rates. The logistic curvilinear relation between TLR and LL means that the majority of patients will have LL at the rather flat left-hand end of the curve, and therefore there is little difference in risk of TLR in such patients. Thus a highly significant difference in LL between 2 stents but with typically low LL in most patients can be expected to lead to a rather small (but real) difference in true TLR rates. However, most trials are of insufficient size to distinguish such small true differences from the null hypothesis of no difference. That is why we believe insight into vessel-related efficacy is best done first with fairly small trials based on these surrogates, to be followed by much larger simpler trials, without planned follow-up angiograms, to evaluate patient-oriented stent safety.
Finally, we do believe that rare phenomena such as strut fracture or hypersensitivity reactions may occasionally result in restenosis unrelated to the logistic equations we have proposed. These, however, are unusual occurrences, and as such LL and %DS may be used as angiographic surrogates for the clinical event TLR.
 |
References
|
|---|
1. Pocock SJ, Lansky AJ, Mehran R, et al. Angiographic surrogate end points in drug-eluting stent trials: a systematic evaluation based on individual patient data from 11 randomized, controlled trials J Am Coll Cardiol 2008;51:23-32.[Abstract/Free Full Text]2. Ellis SG, Popma JJ, Lasala JM, et al. Relationship between angiographic late loss and target lesion revascularization after coronary stent implantation: analysis from the TAXUS-IV trial J Am Coll Cardiol 2005;45:1193-1200.[Abstract/Free Full Text]
Related Article
-
Are Surrogate End Points in Drug-Eluting Stent Trials Reliable?
- Pierfrancesco Agostoni and Giuseppe G.L. Biondi-Zoccai
J. Am. Coll. Cardiol. 2008 51: 1991-1992.
[Full Text]
[PDF]
|
Top
References