Key points are not available for this paper at this time.
Design
Editorial
Answers are grounded in this paper and cite their sources.
The ESC PCI guidelines task force defends their evidence-based recommendations regarding drug-eluting stents, emphasizing the need for adequately powered, randomized, multicenter trials with clinical endpoints.
The members of the ESC PCI guidelines task force appreciate the comments made by Dr Nienaber regarding our analysis and recommendations for drug-eluting stents (DES). (1) Dr Nienaber states that our recommendations for DES should not have been based on the TAXUS-VI trial, because this trial investigated the moderate-release form of the Taxus stent which has not been marketed. First, in the TAXUS-II trial, there were no clinically or angiographically relevant differences between the marketed slow-release and the not marketed moderate-release forms in equivalent lesions. Furthermore, comparing the TAXUS-VI results with TAXUS-V also did not reveal clinically relevant differences between the slow-release and the moderate-release forms. From Dr Nienaber's point of view, we should also not have recommended the Cypher stent—at least for Germany: the currently marketed Cypher Select stent was not the one that was investigated in the SIRIUS trial (Cypher Bx Velocity). As another example, in most European countries, clopidogrel is not labelled for use after coronary stent implantation. Should the ESC PCI guidelines therefore not recommend clopidogrel after stent implantation? (2) Regarding the use of DES for in-stent restenosis, Dr Nienaber criticizes that we demanded randomized, controlled studies before making definitive recommendations, ‘even at a time when ARTS II and TROPICAL could not use bare metal controls, but rather had to compare with historical controls for ethical reason’. The ESC PCI guidelines committee strongly believes that for such an important recommendation, randomized, controlled studies comparing DES with an accepted standard method of treatment are both necessary and ethical. The superiority of DES (Cypher and Taxus) over plain balloon angioplasty in patients with bare metal in-stent restenosis has been clearly demonstrated in the randomized ISAR-DESIRE and RIBS-II trials. However, the only evidence-based treatment of in-stent restenosis, as pointed out in our guidelines, is brachytherapy. Therefore, randomizing patients with in-stent restenosis to either brachytherapy or DES is ethical. ARTS II and TROPICAL were registries, both using historical controls—active controls in these trials were avoided not for ethical reasons, as Dr Nienaber suggests, but rather for funding or logistical concerns. The final publication of TROPICAL did not even publish the historical brachytherapy arm. Evidence-based guidelines cannot be changed on the basis of registries like ARTS II or TROPICAL with no active control group. Two randomized, prospective trials (SISR and TAXUS-V-ISR) compared DES with the gold standard for in-stent restenosis, vascular brachytherapy. At last year's Transcatheter Cardiovascular Therapeutics meeting, SISR was reported to be positive, but is not yet published; TAXUS-V-ISR has been presented at the ACC in March this year. If these trials are positive when published, then an evidence-based recommendation for the use of DES for restenosis in bare metal stents can be made. (3) Dr Nienaber's third series of questions relates to the comparison of Cypher and Taxus stents and questions our suggested equivalency of these two stents regarding patients' outcome. In the ESC guidelines, we have stated that at the time the guidelines were finished (March 2005), there was no evidence that one DES was superior to another. Until March 2005, TAXI was the only trial comparing Cypher and Taxus, which was fully published. Because TAXI did not clearly define its primary endpoint and no power calculation was provided, we did not feel it appropriate to draw evidence-based recommendations from this study. In the meantime, two superiority trials have been presented: SIRTAX and REALITY. Although SIRTAX was well conducted and met its primary clinical endpoint, the task force does not feel it is appropriate to make an important recommendation on the basis of only a dual-centre study. REALITY, the largest and the only true multicentre trial performed to date testing the hypothesis that Cypher is superior to Taxus, did not meet its primary endpoint (which was angiographic, not clinical, though there were no major clinical differences apparent between the two stents). After our guidelines were printed, two more small (200–250 patients) 1–2 centre studies comparing Cypher vs. Taxus have been fully published: ISAR-Diabetes and ISAR-SMART. Both were designed as non-inferiority studies and both did not meet their non-clinical primary endpoints. This raises the important question: if a non-inferiority study did not confirm its hypothesis, is it justified to apply the results into a retrospective superiority analysis? If one does, then, for example, for the ISAR-SMART trial, the differences between the Cypher and the Taxus stents would have a power in the range between 50 and 70%. The history of medicine is full of erroneous conclusions drawn from underpowered studies or analyses. In order to make an evidence-based statement regarding Dr Nienaber's question about the superiority of Cypher over Taxus, one would require a study as follows: superiority design primary clinical endpoint adequate power calculated and reached true multicentre study external and independent CEC/DSMB committee Not a single large-scale, true multicentre trial with an appropriate clinical endpoint has been performed to address the relative efficacy of Cypher and Taxus. Until such a trial is presented, no evidence-based recommendation about the superiority of one DES vs. another regarding patients' outcome can be made. In conclusion, ESC guidelines are evidence-based and follow strict rules. The task force of the ESC PCI guidelines was and still is strongly committed to the scientific evidence derived from fully published studies, emphasizing the patients' clinical outcome (rather than surrogate endpoints) demonstrated in appropriately designed, large-scale, adequately powered multicentre trials. Guidelines must be based purely on scientific evidence and be independent of labelling, current or future marketing strategies, or pressures from the industries.
No takes yet. Share an insight, caveat, or question.
Sigmund Silber (2006) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: