Author + information
- Pierfrancesco Agostoni, MD⁎ ( and )
- Giuseppe G.L. Biondi-Zoccai, MD
- ↵⁎Antwerp Cardiovascular Institute Middelheim, AZ Middelheim, Lindendreef 1, 2020 Antwerp, Belgium
We read with great attention the paper by Pocock et al. (1). We welcome their thorough analysis; however, we have a few concerns related to some methodologic aspects of the manuscript.
The first relates to the assessment of treatment differences among angiographic or clinical end point selected with the z-score. We believe the authors should also report the rates of target lesion revascularization (TLR) in the patients who were not supposed to receive angiographic follow-up in the trials where a control angiogram was mandated only in a subgroup of patients. This would allow comparing the difference among stents (“delta”) in TLR rate between the angiographic and the clinical cohorts and indirectly comparing the “delta” in late loss (LL) and percentage diameter stenosis (%DS) with the “delta” in TLR in the clinical cohort. Indeed, despite the fact that the authors already acknowledged in their discussion that systematic angiographic follow-up can increase the rate of TLR, it would be useful to systematically quantify this phenomenon and to observe if it can impact the treatment differences between devices.
The second concern refers to the relationship between the size of treatment effect on TLR and on LL/%DS, using the Hughes criterion of surrogacy by visually plotting mean LL or %DS and TLR rate. The authors should limit this analysis only to the drug-eluting stent (DES) arms of the trials, avoiding inclusion of the bare-metal stent (BMS) arms. Indeed, much of the visual assessment of the relationships shown seems dependent on the BMS arms of the trials included. We believe that the relationship between mean LL and TLR rate is a crucial point in clinical practice to decide which DES performs better and in the development of new DES to understand whether a new device can compete with the ones already on the market. Thus a more accurate and definitive analysis of this correlation should be clearly made and should only include patients treated with DES. In light of this, the authors correctly recognized that, in the REALITY (Prospective Randomized Multi-Center Head-to-Head Comparison of the Sirolimus-Eluting Stent [Cypher] and the Paclitaxel-Eluting Stent [Taxus]) trial (2), which directly compared 2 DES, the TLR rates were similar despite a highly significant difference in LL and %DS. We suggest a possible explanation for this apparent paradox. Bare metal stents have been shown to have a bimodal distribution of angiographic measures of restenosis (3). We have suggested that this distribution can be possible also with DES (4,5). Sirolimus-eluting stents showed a sort of “all-or-none” phenomenon, having a very low average LL driven by the lack of LL in nonrestenotic lesions. Paclitaxel-eluting stents, in contrast, accommodated some LL also in nonrestenotic lesions and this led to a higher average LL value.
We urge the authors to test again this hypothesis and to perform a detailed analysis of the distribution of LL and %DS, given the large database of prospectively enrolled patients with independently performed quantitative coronary angiographic analysis that they can access. If the bimodal distribution of angiographic parameters of restenosis were confirmed also for DES, this could have a major impact in the design, analysis, and conduct of future comparative DES trials.
- American College of Cardiology Foundation
- Pocock S.J.,
- Lansky A.J.,
- Mehran R.,
- et al.
- Schomig A.,
- Kastrati A.,
- Elezi S.,
- et al.
- Cosgrave J.,
- Melzi G.,
- Corbett S.,
- et al.