TCT 2016: Three-year data fail to show Absorb to be superior or even non-inferior to Xience

Patrick Serruys
Patrick Serruys

According to the three-year results of the ABSORB II trial, a bioresorbable vascular scaffold (Absorb, Abbott Vascular) does not provide superior vasomotor reactivity than an everolimus-eluting, permanent polymer stent (Xience, Abbott Vascular). These results also indicate that, at three years, late lumen loss is significantly greater with Absorb than it is with Xience. Additionally, the scaffold was associated with significantly higher rates of device thrombosis and target-vessel myocardial infarction.

Presenting the results at the 2016 Transcatheter Cardiovascular Therapeutics (TCT) meeting (29 October–2 November, Washington DC, USA), Patrick Serruys (The National Heart and Lung Institute, Imperial College London, UK) noted that the aim of ABSORB II was to provide data for how the scaffold compares with existing devices (ie. Xience). He added that prior to the study, clinical evidence for the device had been attained without the use of a comparator (the CE mark for the device, awarded in December 2010, was based on first-in-man data). Overall, in ABSORB II, 501 patients were randomised in a 2:1 fashion to undergo percutaneous coronary intervention (PCI) with Absorb (335 patients; 364 lesions) or PCI with Xience (166 patients; 182 lesions).

Serruys reported that the study was “powered to demonstrate two mechanistic co primary endpoints: superiority of the bioresorbable scaffold Absorb on the metallic stent Xience in vasomotion post intracoronary nitrate and non-inferiority in late lumen loss post intracoronary nitrate”. He added that true changes in mean lumen diameter (to indicate vasomotion) would be 0.07mm for Absorb and 0.0mm for Xience, explaining that lumen diameter was assessed with quantitative coronary analysis (QCA) pre and post intracoronary nitrate.

However at three years, the vasomotor reactivity was not significantly different between groups (in a paired lesion analysis): increase in mean lumen diameter after intracoronary injection of nitrate was 0.047mm for Absorb vs 0.056mm for Xience (p=0.49)—meaning that the criterion for superiority was not met. Writing in The Lancet (ABSORB II was published in the journal to coincide with the TCT presentation), Serruys and colleagues comment that they assumed that Xience would not exhibit vasomotion because previous studies of metallic stents have indicated that this would be case. They add that the overall vasodilation of the scaffold observed in ABSORB was similar to that seen in the ABSORB first-in-man three-year and the ABSORB Japan two-year data, stating: “Future trials might consider alternative imaging, different vasodilator responses, or late follow-up to confirm whether there is a true vasomotor advantage for the bioresorbable scaffold or whether this proposed benefit is not realised in practice when compared with contemporary metallic drug-eluting stents.”

The study’s other co-primary endpoint was also not met—late luminal loss at three years was 0.37mm for Absorb compared with 0.25mm for Absorb (p=0.78 for non-inferiority). According to Serruys et al, the implant technique used in the study may have affected the outcomes. They report: “Very late lumen size at three years for the Absorb scaffold was compromised by the relative under-expansion at baseline, which is very different from what is being observed in clinical practice with the proper lesion preparation, adequate sizing, and systematic post dilation. Future studies will be needed to show the effect of these new implant techniques on late outcomes.” In an interview with Cardiovascular News (issue 42; page 4) about optimising outcomes with the scaffold, Nick West (Papworth Hospital, Cambridge, UK) said: “The key tips for success are to prepare the lesion optimally, to size and implant the device per instructions for use, and to routinely optimise the result with post dilation.” In ABSORB II, by contrast, post dilation was only performed with 61% of Absorb devices (59% with Xience).

Some of the secondary endpoints of ABSORB II also indicated that, in certain respects, Absorb may be inferior to Xience. The device-oriented composite endpoint was significantly higher in the Absorb arm—10% vs. 5% for Xience (p=0.045)—which Serruys et al report was driven by a significantly higher rate of target vessel myocardial infarction with Absorb (6% vs. 1%, respectively; p=0.0108). Furthermore, the rate of definite or probable scaffold thrombosis (including late and very late) was higher with Absorb. The authors state: “Over three years, there were eight definite scaffold thromboses (one acute, one subacute, six very late) and one late probable scaffold thrombosis in the Absorb group (3%) and no definite or probable stent thrombosis in the Xience group (p=0.0031). They suggest that that there may be connection between the incidences of late or very late scaffold thrombosis that occurred in the Absorb arm and the lack of dual antiplatelet therapy, commenting: “No late and very late thrombosis occurred in 63 patients who never interrupted dual antiplatelet therapy up to three years. The benefit and need for prolonged dual antiplatelet therapy bioresorbable scaffold implantation is at the present time, entirely speculative.” Speaking at TCT, Serruys commented: “The incidence of very late scaffold thrombosis warrants further careful monitoring of the patient having a clinical follow-up of longer than two years.”

Concluding, Serruys et al write: “The co-primary endpoint of superior vasomotion was not met due to the fact that the metallic stents exhibited an unexpected change in diameter so far not reported in the literature. The significant difference in clinical events such as the device-oriented composite endpoint could be a type 1 error.”

AT TCT, commenting on the results, Serruys said “as a pioneer, I am disappointed but that is the reality”. However, he stressed that Absorb was a first-generation device “and there was room for improvement by reducing the thickness of the struts, and having a faster absorption time and stronger struts”.

In an accompanying editorial in The Lancet, Robert Byrne and Adnan Kastrait (both Deutsches Herzzentrum München, Klinik an der Technischen Universität München, Munich, Germany) note: “The data reported show that the hypothesised advantages of bioresorbable stents were not readily detectable; neither vasomotion in the stented segment not angina relief as reported by patients were superior to conventional stents. An increase in vessel dimensions due to positive remodelling as the scaffold degrades could also not be demonstrated. This finding suggests that in testing these scaffolds the well-established endpoints used in the evaluation of conventional stents remain the benchmark of choice.” However, they add: “Although the expectations regarding late performance have not been realised to date, the obvious advantage of a stent that disappears remains a goal worth pursuing.”