Statistics from Altmetric.com
Since their introduction, bioresorbable vascular scaffolds (BVS) has been labelled as the forth revolution in interventional cardiology. Its development was mainly driven by the willing to overcome late metallic stent failure, such as neoatherosclerosis and thrombosis, and to avoid a permanent prosthesis inside coronary artery aiming to the so-called vascular restoration therapy.1 Efficacy and safety results in early observational studies and randomised clinical trials were comparable with those obtained with new-generation Drug eluting stent (DES) and accepted widely by the interventional cardiology community.2
Though this initial enthusiasm, recent findings from long-term studies have cooled the interest on these devices. Concerns have been raised in particular on an increased risk for both stent thrombosis (ST) and myocardial infarction (MI) after BVS implantation.3 4 On the top of this, the recent 3-year results of ABSORB II, which randomly assigned 501 patients to ABSORB or everolimus-eluting stent, showed not only significantly higher target vessel-related MI (6% vs 1%) and rates of ST and very late ST (VLST) (3% vs 0%) in the BVS arm but also lack of vasomotor reactivity difference in the two arms, with a larger late lumen loss in the ABSORB group.4
In the attempt to understand the real weight of these very late events, Nairooz et al 5 herein report the results of a meta-analysis of those studies comparing BVS with DES across the whole spectrum of coronary artery disease with at least 2-year of follow-up. They selected five studies with a total of 2360 patients, finding that at 2-year follow-up, ABSORB was associated with higher rates of device-oriented endpoint (OR 1.53, 95% CI 1.06 to 2.23), target vessel myocardial infarction (OR 1.94, 95% CI 1.02 to 3.67) and definite ST (OR 3.39, 95% CI 1.46 to 7.88) as compared with metallic drug-eluting stents, without any difference in all-cause mortality and target lesion. Interestingly, no differences were found in terms of very late definite ST (0.7% vs 0%, p=0.25) or definite/probable ST (1.8% vs 0.7%, p=0.10).
Although this analysis may contribute to expand our knowledge on very late ABSORB events, some points have to be taken into account to understand not only the value of this meta-analysis but also the issue of scaffold thrombosis.
First of all, it is important to note that the value of a meta-analysis depends on many factors and in particular on the quality of the studies included. In case of ABSORB, no many trials with long follow-up are so far available, so that some registries or propensity-score comparisons are included either in this specific meta-analysis or in others.3 The risk is that quality of data of controlled randomised trials could be diluted by presence of registries of propensity-score comparison, which do not have adjudication event committee, standardised event definition and may suffer from patient selection. This is reflected by the fact that the findings reported may differ according to the studies selected (table 1). Comparing the meta-analysis by Nairooz et al with another one recently published on the same topic, ORs for the events reported as well as statistical significance and time occurrence of ST varies.3 Specifically, in the present meta-analysis, rate of very late scaffold thrombosis did not differ between the two arms and rate of ST is surprisingly lower considering definite/probable ST together than considering definite ST alone. Moreover, both meta-analysis missed the opportunity to include the recent reports of a 3-year outcome of the ABSORB II trial, limiting at 2-year follow-up.4
Considering its study-level pooled methodology, it is also difficult to understand the time occurrence of thrombosis and to construct landmark analysis. By contrast, a patient-level pooled meta-analysis offers the advantages of having time-to-event curves (and landmark analysis for estimation of very late events rate, for example), multivariable analyses to ascertain the individual predictors of outcomes and analysis of treatment effects in clinically relevant subgroups.
For all these reasons, the present meta-analysis represents a first attempt to understand the real incidence of scaffold thrombosis problem beyond 1 year after implantation, but many questions remain unanswered. As doctors having the safety of the patient as objective and as scientists with a close critical scrutiny on the clinical performance of these novel devices, we need to understand when ST may occur more frequently, which are the causes, how and if we can solve it.
Various hypotheses have been so far formulated. Imaging studies have provided, for example, important new insights on the pathophysiological mechanisms involved in late failures of ABSORB: disintegration of the polymeric scaffold struts, endoluminal dislocation or dismantling with subsequent flow disturbances and thrombus formation.6 Determining whether meticulous imaging-guided ABSORB implantation can overcome device-related late complications would require prospective evaluation in well-designed controlled clinical trials. The Optical Coherence Tomography to Improve Outcome for Coronary Revascularisation Using Bioresorbable Vascular Scaffolds (OPTICO-BVS)—number clinical trial (NCT) 02683356—which is currently ongoing, randomising patients to receive ABSORB implantation guided by optical coherence tomography or angiography will help us in this point.
The importance of optimal BVS implantation, evaluated by the PSP (accurate Predilatation, Sizing and post-Dilatation) ensuring full strut apposition has also been recently emphasised, as these stents appear less forgiving than new-generation metallic DES, if too optimistically implanted.7 Nevertheless, blaming suboptimal implantation techniques for some of the late adverse ABSORB results may result excessive and would need to be further investigated.
Eventually, this old problem of very late thrombosis for a new-generation stent remembers the story of first-generation DES, accepted with enthusiasm among interventional cardiologist until it was realised that although DES represented a marked improvement over bare-metal stents, they were still not perfect with a 2%–3% frequency per year of both restenosis and thrombosis.8 As it happened in that case, technological evolution is currently ongoing for filling our unmet need of having safe and efficacious biodegradable devices for coronary artery disease treatment. New-generation polymeric scaffolds, with thinner struts, enhanced expansive capability and improved resorption kinetics will be soon available. These iterations are expected to overcome most of the limitations encountered by first-generation ABSORB. We also need to understand if this more than expected high rate of very late events would affect only the ABSORB scaffold or all the bioresorbable scaffolds currently on the market. Data from these patients, once they would reach at least 2-year follow-up would be very welcome.
For all these reasons, we should learn our lessons and keep a close critical attention on the clinical performance of these novel devices. Patients’ care should be maintained on clinical data with robust methodology more than on expectations.
Competing interests None declared.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.