TWENTE: Abbott’s Xience V Stent and Medtronic’s Resolute Go Head-to-Head

What is the history of comparisons between the Xience V and Resolute stents?

The TWENTE trial is actually the second, prospective, randomized trial between Resolute and Xience V in patients with complex coronary artery disease. The first, the RESOLUTE All Comers trial, was published in 2010 in the New England Journal of Medicine by Prof. Patrick Serruys.¹ Our trial started a few months later; it is a single-center trial, so we needed more time for the inclusion and enrollment.

First-generation drug-eluting stents did not reduce mortality, but reduced the need for re-intervention. One thought was that the biocompatibility of the stent coating might have been one of the important reasons for its lack of effect on mortality. Therefore, the second-generation DES were developed, with Resolute and Xience V being two of the most prominent devices. Xience V has been available in the U.S., and Resolute has been available for several years in the European Union and other parts of the world outside America. Both stents have a thin, novel coating and were developed to improve the biocompatibility of drug-eluting stents.

Actually, it is quite amazing that while two million drug-eluting stents were implanted annually worldwide, there has been only one comparative study between these big players. It makes sense to perform a second one to confirm the findings. There may even be room for another dedicated trial, because we are dealing with devices in patients’ hearts. It may have important consequences for patients if these devices fail.

What are the main differences between the two stents?

While both are “limus” analog drugs, they have a slight difference in complete drug elution, with Resolute taking about 60 days longer for complete drug elution. Xience uses a fluoropolymer and Resolute uses BioLynx^™, a combination of hydrophilic and hydrophobic polymers which is believed to increase biocompatibility. Coating thickness is particularly thin for the Resolute stent at 5.6 µm (microns), but also very thin for the Xience V stent, at 7.8 µm. Resolute uses the Driver platform and Xience V uses the Multi-Link Vision. We published bench side work on the Xience V and Resolute stents (among others) with micro CT in EuroIntervention in 2010,² looking at the effect of extreme post dilatation. The largest cell opening was achieved with the Resolute — slightly more even than with the Xience V. This has to do with the connections between the “stent rings.” Stents can be regarded as a stack of metallic rings, with each ring being connected to the adjacent ring at certain connection spots. Xience typically has three connections, and Resolute has two for most of its length, but this differs for Resolute depending on the stent diameter. Both stents are made from cobalt chromium alloys. Each has a slightly different alloy, but that makes little difference. The strut thickness is low for both, but a little lower for Xience, with 81 µm. Resolute is 91 µm.

How was the TWENTE trial designed?

We randomized a total of 1,391 patients 1:1 to treatment with the zotarolimus-eluting Resolute stent or the everolimus-eluting Xience V stent. There were no exclusion criteria with regards to the type or number of lesions or vessels to be treated, but the trial did not include ST-elevation myocardial infarction (MI) patients. The analysis was based on intention to treat, and the primary endpoint was the composite of target vessel failure (cardiac death, target vessel-related myocardial infarction or clinically indicated target vessel revascularization) at one year. 

We were able to include more than 80% of all eligible patients, a very high inclusion rate. Two patients within each arm withdrew their consent to participate in the study. We have 12-month follow-up at 100% in both arms, thanks to the great work and effort of the study team, which was able to trace patients who immigrated to Africa and Asia. We were very persistent. For example, we had one patient who immigrated to South Africa, but we were able to contact the buyer of his house. The buyer gave us an email address that he had, and our patient was able to phone back and give the necessary information. For all parameters, there was no difference between the two study arms, because the randomization went well. Diabetes was present in 22% of the patients. Fifty-two percent presented with acute coronary syndromes, so there were many patients with acute disease. Complex lesions (lesion type B2 or C) were present in 70% of lesions. Bifurcation lesions were treated in a quarter of the population. Multi-vessel treatment was performed in 24% of the patients, and at least one off-label indication for drug-eluting stent use was present in 77% (in RESOLUTE All Comer, off-label use was 66%). Direct stenting was performed in 39%, similar to the LEADERS trial.³ If absolutely required, operators could switch to their preferred “workhorse” stent, but this ended up being a very low percentage. Device success (meaning the device the patient was assigned to was ultimately used with success) was 98% for both devices, and lesion success (lesion was successfully treated independent of which device was ultimately used) was 99.8% in both.

RESOLUTE All Comers did include ST-elevation MI patients. Why were these patients not included in TWENTE?

When we designed the trial and started inclusion of patients, the standard of care in our center was bare-metal stent implantation for treatment of STEMI patients. Therefore, we did not include STEMI patients in TWENTE.

In TWENTE, the rates of the primary endpoint “target vessel failure” and the secondary endpoint “target lesion failure” were slightly lower than in the RESOLUTE All Comers trial, the first study to compare both stents. Intuitively, you may think that the reason for this may be because TWENTE did not include STEMI patients, which classically are patients with higher risk. If we look at the data of the STEMI sub-group in RESOLUTE All Comers (12% of study patients), presented by Prof. Stephen Windecker in 2010 at TCT,⁴ you will see that that STEMI sub-group actually had lower event rates than the overall study population. This is partly because of the huge cardiac marker release that takes place with ST-elevation, meaning you are virtually unable to trace peri-procedural MI, which is, in these days, the vast majority of myocardial infarctions following stent implantation. As a result, the event rate of patients with STEMI is actually slightly lower than for patients without STEMI. In other words, if TWENTE would have had a similar population to RESOLUTE All Comers with the inclusion of a comparable sub-group of STEMI patients, the event rates of TWENTE may have been better, rather than worse. This is an important point, because intuitively one may think otherwise.

What did the data show for the primary endpoint of target vessel failure?

Xience V had a target vessel failure rate of 8.1% and Resolute of 8.2%. The p value for non-inferiority was highly significant, so this is quite convincing data for the non-inferiority of these devices. For the secondary endpoint of the components of the primary endpoint — cardiac death, target vessel-related myocardial infarction and clinically indicated target vessel revascularization, there were no significant differences. For the secondary endpoint of stent thrombosis, the ARC definition (Academic Research Consortium) was used. The analysis was externally performed, and it showed that the composite of definite plus probable stent thrombosis, which is clinically the most relevant, was similar for both stents. For the entire population, it was 0.86% for the Resolute versus 1.16% for the Xience V group. Keep in mind that these were complex patients and complex lesions, with a great deal of multi-vessel treatment. These are relatively low values.

We looked at whether stent thromboses were definite or probable. Definite stent thrombosis means stent thrombosis is certain; thrombus can be seen on angiography. Probable means it is very likely, depending on the timeframe and other definitions that are clearly outlined in the ARC definition. There were a total of eight probable stent thromboses in the Xience V arm. In the Resolute arm, there were six stent thromboses, with four being definite. There were four deaths in the Xience V arm related to probable stent thrombosis, and there was one death related to a definite stent thrombosis in the Resolute arm. This allowed us to calculate 0.58% definite stent thrombosis for Resolute and 0% for Xience V. We also considered dual antiplatelet therapy adherence in relation to stent thrombosis. First, we determined from the pharmacist whether patients picked up their drugs. We found that the two Xience V late stent thromboses, beyond day 8, were in patients not on dual antiplatelet therapy. This allows us to conclude, based on the data, that there were no (probable or definite) stent thromboses in the Xience V arm beyond day 8 for those patients on dual antiplatelet therapy. In the Resolute arm, we found that one of the definite stent thromboses was in a patient not on dual antiplatelet therapy. The only lethal definite stent thrombosis in the Resolute arm was on day 5 in a patient who had, seven days prior, bare metal stenting of the circumflex for a relatively large lateral myocardial infarction. After a week, he was then enrolled in the study, and received study stents in the left anterior descending coronary artery and the right coronary artery. Five days later, the patient died. Autopsy showed coagulated blood in all three vessels. According to the definition, this event must be counted as a stent thrombosis, but considering the recent medical history of that patient, one cannot rule out an arrhythmogenic death after the myocardial infarction. We cannot say. But according to the ARC definition, there is no other way to define it than counting it as a stent thrombosis.

The conclusion is that zotarolimus-eluting Resolute stents were non-inferior to everolimus-eluting Xience V stents in terms of safety and efficacy for treating real-world patients, with a vast majority having complex lesions and off-label indications for DES, which were implanted with a liberal use of post dilatation.

What is the importance of post dilatation?

We performed post dilatation in 82% of stented lesions. This number is very high. Why did we do that? Well, the idea is that post dilatation improves the apposition of the stent to the vessel wall, which may facilitate the transition of the drug into the vessel wall, and this could have a beneficial effect on stent thrombosis and other related adverse events. We are treating more and more long lesions and diffuse disease, and in these scenarios, the vessel is tapering. Each side branch means that the mother vessel’s diameter is reduced in size, because the myocardium supplied is reduced. So if there are various side branches along a stented segment, the vessel diameter will taper. You do not want to overexpand the segment at the distal stent edge, which generally has the smallest vessel dimensions of the treated vessel segment, to avoid dissection (we actually saw very little dissection with both stents). To mimic the biology of the vessel, you can post dilate with higher pressure in the proximal part. Then, in the mid portion, you do so with somewhat lower pressure, and in the distal, lower still. All of this is done with the goal of matching the anatomy of the native vessel. This philosophy is something all of the interventional cardiologists in our center have been following for several years. Our center has 5 operators, each having completed 4,000 or more PCI procedures. Our post dilatation strategy could be part of the explanation for the beneficial outcome data in TWENTE. 

Diabetics were 22% of the patient population. How did they fare?

We did a subgroup analysis. It is important to mention that this is exploratory, because we did not predefine the subgroups. But to handle this limitation, we stuck to the same subgroups used for the RESOLUTE All Comer trial. The primary endpoint, target vessel failure, was consistent throughout all the subgroups. In diabetic patients, the p value for interaction between diabetes mellitus and stent type was 0.045, slightly significant. This meant that there could be a relative difference between stents, which required further statistical analysis. But when further analyses were performed between diabetic patients treated with Resolute and Xience V, no statistically significant difference was found. To conclude, also for diabetics, no difference has been proven for the two stents with regard to the primary endpoint of target vessel failure.

What were the particular strengths of the TWENTE study?

We assessed a real-world patient population with advanced disease and lesions. We had more than 80% of all eligible patients in an investigator-initiated study. We had a systematic post-procedural measurement of cardiac biomarkers which was available in 99% of patients, not standard for all investigator-initiated comparative stent studies. The TWENTE trial had entirely clinical endpoints. No routine angiographic follow-up was performed, meaning bias from angiography — the “occulo-stenotic reflex” — was avoided. We had 100% clinical follow-up. We verified all patient-reported clinical event triggers from the source, which was possible because we are located in the eastern part of the Netherlands and the next PCI center is at least a one-hour drive from us. We have a network with the referring centers and referring cardiologists that were involved in the study, and some of them are also co-authors on the paper. They opened their databases and we had a very close collaboration, allowing for very reliable clinical follow-up data. If a patient said, I had a problem, I was with my cardiologist, but I am not sure it was a myocardial infarction, we visited their office or asked for copies of all reports to get the information. External clinical event adjucation was performed by the independent clinical research organization Cardialysis in Rotterdam, which has a long history of performing and analyzing clinical studies. This is the same core lab that did the clinical event adjucation for the RESOLUTE All Comers trial, the only other head-to-head comparative study of these two stents, which may facilitate meaningful comparisons between both trials.

Dr. Clemens von Birgelen can be contacted at c.vonbirgelen@mst.nl.

In the U.S., the Resolute drug-eluting stent is currently limited to investigational use.

References

1. Serruys PW, Silber S, Garg S, et al. Comparison of zotarolimus-eluting and everolimus-eluting coronary stents. N Engl J Med 2010;363:136-346; doi:10.1056/NEJMoa1004130.
2. Basalus MW, van Houwelingen KG, Ankone MJ, Feijen J, von Birgelen C. Micro-computed tomographic assessment following extremely oversized partial postdilatation of drug-eluting stents. EuroIntervention 2010 May;6(1):141-148. doi: 10.4244/.
3. Windecker S, Serruys PW, Wandel S, et al. Biolimus-eluting stent with biodegradable polymer versus Sirolimus-eluting stent with durable polymer for coronary revascularization (LEADERS): a randomised non inferiority trial. The Lancet 2008; 372(9644): 1163-1173.
4. Windecker S. Clinical and angiographic outcomes in patients with acute myocardial infarction from the RESOLUTE All Comers open label randomized trial: RESOLUTE-AMI. Presented at TCT 2010; September 22, 2010; Washington, D.C.