The field of percutaneous coronary intervention (PCI) remains extremely active, with major advances involving revascularisation strategies and techniques, adjuvant antithrombotic therapy (with a particular emphasis this year on the optimal duration of antiplatelet therapy after drug-eluting stent (DES) stenting), concerns regarding the importance and risks of bleeding and methods to decrease those risks and a strong focus on the study of newer generation DES and bioresorbable devices with lingering concerns regarding the safety of the latter.
Given the improved clinical outcomes of newer generation DES, comparisons between these and coronary artery bypass surgery (CABG) are important. The BEST trial1 randomly assigned 880 patients with multivessel disease to either PCI with everolimus-eluting stents (EES) or to CABG. After a median follow-up of 4.6 years, major adverse cardiac events were higher in the PCI group than in the CABG group (15.3 vs. 10.6%, P = 0.04), driven by a reduction of repeat revascularization but also spontaneous myocardial infarction (MI) in the surgical group, reaffirming the role of CABG surgery in the management of patients with multivessel disease in the era of newer DESs.
The management of in-stent restenosis can be difficult. The PROSPERO network meta-analysis of 27 trials2 attempted to synthesize the evidence and suggested that PCI with EES yielded the best angiographic and clinical outcomes and that drug-coated balloons provided favourable angiographic results without adding an additional metal layer. The RIBS-IV trial3 compared prospectively these two strategies in 309 patients with in-stent restenosis after DESs and found that EES provided superior 1-year clinical (lower composite of cardiac death, MI, and target vessel revascularization, 10 vs. 18%, P = 0.04) and angiographic outcomes compared with drug-eluting balloons. Thus, while drug-eluting balloons provide effective treatment for restenosis in bare metal stents, they are less effective than EES for restenosis in DES.
The optimal management of bifurcation lesions remains uncertain.4 The TRYTON trial5 randomly assigned 704 patients with bifurcation lesions to either a bifurcation stent (the TRYTON stent) or stenting of the main vessel with provisional stenting of the side branch. At 9 months, target vessel failure was numerically more frequent with the bifurcation stent, largely due to a higher periprocedural MI rate, and the bifurcation stent failed to achieve non-inferiority with the strategy of provisional stenting, illustrating once more that, in interventional cardiology, ‘less is more’… .
Likewise, the role of PCI in the management of chronic total occlusions is still debated, in the absence of a properly powered randomized outcomes trial. An observational analysis from the Italian registry of chronic total occlusions6 which enrolled 1777 patients with CTOs found that after propensity score matching, rates of cardiac death, acute MI, and rehospitalization were lower in patients managed with PCI than managed with medical therapy or CABG, suggesting that PCI may have a role in improving outcomes in these patients.
As interventionalists treat ever older patient populations with more advanced, complex, and calcified disease, the potential role of rotational atherectomy has increased. However, in many sites, the use of this technique has been hampered by the lack of a standardized protocol. A group of experienced operators has provided a consensus document,7 recapitulating recommendations for the performance of the technique, which will be useful to sites and operators with less experience.
The long-term safety of DESs vis à vis that of bare metal stents has long been a contentious issue. With the advent of newer generation DESs, both efficacy and safety have been improved. A comprehensive network meta-analysis reviewed these issues8 in 51 trials including a total of 52 158 randomized patients with follow-up duration ≥3 years. After a median follow-up of 3.8 years, all DES demonstrated superior efficacy compared with BMS. Among DES, second-generation devices have substantially improved long-term safety and efficacy outcomes compared with first-generation devices. Since there is continuous improvement and innovation in the field, an ESC-EAPCI task force has released a report reviewing the existing evidence and confirming the improved efficacy and safety of new DES (Figure 1A and B). This report also outlines the desirable pathway for preclinical and clinical evaluation and regulatory approval of future devices.9
Finally, the effect of PCI on long-term survival in patients with stable coronary artery disease was studied in the extended follow-up of the COURAGE trial10 In 2287 patients randomized to optimal medical therapy alone or combined with PCI, the initial median follow-up of 4.6 years had found no difference in the composite primary outcome of death or MI, or in survival. An updated report with an extended median follow-up of 11.9 years found no difference in mortality (adjusted HR 1.03; 95% CI 0.83–1.21; P = 0.76). These results are consistent with the BARI-2D results. Controversy regarding the lack of mortality benefit with PCI vs. either best of medical care or CABG was revisited by Windecker et al.11 in a 93 553 patient large network meta-analysis showing survival benefit with newer generation DES (HR everolimus DES 0.75; 95% CI 0.59–0.96; HR zotarolimus Resolute DES 0.65; 95% CI 0.42–1.00) vs. medical therapy. These results emphasize the importance of the ongoing International Study of Comparative Health Effectiveness with Medical and Invasive Appproaches (ISCHEMIA) trial (Clinical trials.gov number NCT01471522) to assess the role of cardiac catheterization and contemporary revascularization techniques in improving the outcomes of patients with stable coronary artery disease and at least moderate ischaemia.
Adjunctive pharmacologic therapies
Duration of antiplatelet therapy
The issue of the benefits and risks of prolonged dual antiplatelet therapy (DAPT) following DES placement has continued to attract interest and spark controversy, particularly given the heterogeneity of results. Many of the randomized trials comparing various durations of DAPT were underpowered.12,13 In addition, in the large DAPT trial, despite clear benefit of prolonged DAPT beyond 1 year in reducing the risks of MACCE, stent thrombosis, and MI, there was a signal of increased non-cardiovascular mortality. Both of these issues have sparked interest in meta-analyses of the data. With respect to mortality, a pairwise and Bayesian network meta-analysis of 10 randomized trials comparing different DAPT durations after DES stenting in 31 666 patients14 found that although treatment with DAPT beyond 1 year after DES implantation reduces MI and stent thrombosis, it is associated with increased mortality because of an increased risk of non-cardiovascular mortality not offset by a reduction in cardiac mortality. However, this signal had not been seen in a prior broader meta-analysis comparing various durations of DAPT across a variety of cardiovascular conditions.15 With respect to the risk benefit risk of shorter vs. longer durations of DAPT, several meta-analyses16–18 have concurred in suggesting that prolonged DAPT should be considered on a case-by-case basis, after weighing both ischaemic and bleeding risks.
One patient profile that is most likely a good candidate for consideration of prolonged DAPT after stenting is patients with prior MI. Both the analysis of the MI subset from the large DAPT trial19 and a comprehensive meta-analysis of randomized trials in patients with prior MI20 have found that patients with prior MI derive a clear benefit from prolonging DAPT beyond 1 year (compared with aspirin alone). In the meta-analysis, prolonged DAPT reduced rates of MACE (rate ratio, RR 0.78, 95% CI 0.67–0.90; P = 0.001), MI (RR 0.70, 95% CI 0.55–0.88; P = 0.003), stent thrombosis, but also cardiovascular mortality (RR 0.85, 95% CI 0.74–0.98; P = 0.03), at the cost of an increase in major bleeding (RR 1.73, 95% CI 1.19–2.50; P = 0.004) but without increase in fatal bleeding or in non-cardiovascular death.
The DAPT investigators developed a decision tool to identify whether an individual patient is more likely to derive benefit or harm from continuation of DAPT beyond 1 year.21 A nine-item score, including patient characteristics (age, diabetes, smoking, prior PCI or MI, congestive heart failure, or left ventricular ejection fraction <30%) and index procedure characteristics (MI at presentation, vein graft PCI, and stent diameter <3 mm) was able to discriminate patients with a score of <2 (half of the population) who derived no net clinical benefit from prolonged DAPT and had increased mortality (NNT 153 and NNH 64) from patients with a score of ≥2 in whom there was a marked reduction in ischaemic outcomes (−3.02% for the composite of MI or stent thrombosis) and marked net clinical benefit (−2.7%) at the expense of a minimal increase in GUSTO severe or moderate bleeding (0.37%) and no increase in mortality. The population with a DAPT score of ≥2 or more derived greater benefit and less harm from prolonged DAPT (NNT 34, NNH 272). This score, which was validated externally in the PROTECT trial database,22 will help clinicians make decisions regarding selection of candidates for prolonged DAPT.
Oral anticoagulation and percutaneous coronary intervention
The optimal strategy for managing antiplatelet therapy in patients who are treated with OAC (e.g. for stroke prevention in atrial fibrillation) is a common and difficult clinical problem, for which there is still little clinical evidence. One of the key randomized trials, ISAR TRIPLE, was released this year:23 it evaluated whether shortening the duration of clopidogrel therapy from 6 months to 6 weeks after DES implantation was associated with a superior net clinical outcome in 614 patients receiving concomitant aspirin and OAC. The trial failed to establish the superiority of shorter therapy, as the primary endpoint occurred in 30 patients (9.8%) in the 6-week group compared with 27 patients (8.8%) in the 6-month group [hazard ratio (HR) 1.14; 95% CI 0.68–1.91; P = 0.63]. There were no significant differences for the secondary combined ischaemic endpoint of cardiac death, MI, definite stent thrombosis, and ischaemic stroke or in terms of TIMI major bleeding. More trial evidence is needed to guide clinical decisions. Pending newer trials, an important and useful consensus document24 provides precious guidance. A key aspect is the sequential assessment of stroke risk, bleeding risk, and the clinical setting before making a decision on the type and duration of antiplatelet therapy (Figure 2). Generally, oral anticoagulation is the ‘foundation’ therapy to which antiplatelet therapy should be added for the shortest possible duration, given the increased bleeding risk associated with combination therapies.
The RIVER-PCI randomized placebo-controlled trial tested the value of adding the anti-anginal agent ranolazine to medical therapy in 2651 patients with a history of chronic angina and incomplete revascularization after PCI.25 The primary endpoint was time to first occurrence of ischaemia-driven revascularization or ischaemia-driven hospitalization without revascularization. After a median follow-up of 643 days (IQR 575–758), ranolazine did not affect the composite primary endpoint (hazard ratio 0.95, 95% CI 0.82–1.10; P = 0·48) or the incidence of ischaemia-driven revascularization and ischaemia-driven hospitalization. Thus, routine adjunctive use of ranolazine in this setting does not improve clinical outcomes. A subsequent analysis also found no incremental improvement in angina or quality of life.26
Bleeding as a complication of percutaneous coronary intervention
Bleeding remains an important concern as a complication of PCI, given the need for invasive arterial access, the use of potent anticoagulant, and antiplatelet agents. Two strategies that have some benefit in addressing bleeding are vascular closure devices and radial access. ISAR CLOSURE27 was a randomized non-inferiority trial comparing vascular closure devices to manual haemostasis in 4524 patients undergoing angiography via femoral access. The trial found that vascular closure devices were non-inferior to manual haemostasis in avoiding access site-related vascular complications (6.9 vs. 7.9%) with a shorter time to haemostasis. The MATRIX trial28 was a large multicentre trial comparing transradial against transfemoral access in 8404 patients with acute coronary syndrome with or without ST-segment elevation who were about to undergo coronary angiography and PCI. Radial compared with femoral access did not significantly reduce the co-primary outcome of MACE [8.8 vs. 10.3%, RR 0.85, 95% CI 0.74–0.99; P = 0.0307, non-significant at α of 0.025] but did reduce the other co-primary outcome of net adverse clinical events, through a reduction in BARC major bleeding unrelated to CABG (1.6 vs. 2.3%, RR 0.67, 95% CI 0.49–0.92; P = 0.013) but also importantly all-cause mortality (1.6 vs. 2.2%, RR 0.72, 95% CI 0.53–0.99; P = 0·045). These data suggest that radial access has a major role in modern coronary intervention. However, bleeding during or after PCI is not limited to the access site and to the periprocedural period. Data from the ADAPT-DES29 have confirmed prior findings showing that post-discharge bleeding after PCI have a strong detrimental prognostic impact, which may even exceed that of post-discharge MI. These findings have bearing on the interpretation of the balance of benefit to risk of prolonged antiplatelet therapy after stenting.
Stents and scaffolds
In the BASKET-PROVE II trial,30 the second-generation biolimus-A9-eluting biodegradable-polymer (BP) stainless steel DES (Nobori, Terumo) was compared with the best-in-class durable-polymer (DP) DES (Xience Prime, Abbott Vascular) for efficacy and safety and to a last-generation thin-strut BMS coated with a biocompatible silicone-carbide layer (ProKinetik, Biotronik) for late safety in large-vessel stenting (≥3 mm). BP-DES (7.6%) was non-inferior compared with DP-DES (6.8%) and superior to BMS (12.7%) in terms of the composite primary endpoint of cardiac death, MI, or clinically indicated target-vessel revascularization (TVR) within 2 years. The composite secondary safety endpoint of very late (>1 year) stent thrombosis (VLST), MI, and cardiac death did not differ between the three stent groups. Those findings challenge the concept that durable polymers are the key initiator of VLST.
The open-label, randomized SORT OUT-VI trial (n = 2999) showed non-inferiority of the biocompatible DP zotarolimus-eluting stent, Resolute Integrity (Medtronic CardioVascular, Santa Rosa, CA, USA) compared with the BP-polymer biolimus-eluting stent, BioMatrix Flex (Biosensors Interventional Technologies, Singapore City, Singapore) in unselected patients.31 At 12 months, 79 (5.3%) and 75 (5.0%) patients, respectively, (Pnon-inferiority = 0.004) met the primary endpoint [a composite of safety (cardiac death and non-target lesion MI) and efficacy (target-lesion revascularization)]. The individual components of the primary endpoint also did not differ between stent types.
In the randomized, double-blind LEADERS FREE trial (n = 2466 patients), BioFreedom (a polymer-free and carrier-free, umirolimus-coated stent) was compared with a similar bare metal stent (the Gazelle stent, Biosensors Interventional Technologies, Singapore) in patients with a high bleeding risk, all received 1 month of DAPT.32 At 390 days, the rate of the primary safety endpoint (a composite of cardiac death, MI, or stent thrombosis) was 9.4% in the drug-coated stent group and 12.9% in the bare metal stent group (Psuperiority = 0.005). Clinically driven target-lesion revascularization was needed in 5.1% of patients in the drug-coated stent group and in 9.8% of patients in the bare metal stent group (P < 0.001). In spite of the short course of DAPT, the rate of BARC types 3–5 bleeding was high (7.2%) and similar in both groups.
In the ZEUS trial,33 compared with BMS, a biocompatible-polymer DES with a fast drug-eluting characteristics, combined with a short, tailored DAPT regimen, resulted in a lower risk of 1-year MACE in uncertain candidates for DES implantation (a unique patient population that was largely excluded from pivotal DES trials). Patients (n = 1606) with stable or unstable symptoms who, on the basis of thrombotic, bleeding, or restenosis risk criteria, were qualified as uncertain candidates for DES were randomized to receive zotarolimus-eluting stent [Endeavor stent (Medtronic Vascular, Minneapolis, Minnesota)] or thin-strut (<100 mm) BMS under similar duration (median 32 days) of DAPT. The primary endpoint (1-year MACE, including death, MI, or TVR) was lower in the DES group (17.5 vs. 22.1%, P = 0.011) as a result of lower rates of MI (2.9 vs. 8.1%; P < 0.001) and TVR (5.9 vs. 10.7%; P = 0.001). Definite or probable stent thrombosis was also reduced in DES-treated patients (2.0 vs. 4.1%; P = 0.019). Overall, these trials suggest that the indications for bare metal stents may be rapidly disappearing, as newer DES provides superior efficacy without safety concerns and no longer require protracted antiplatelet therapy.
The Absorb II trial compared Absorb [Abbott Vascular, Santa Clara, CA, USA (n = 335 patients, 364 lesions)] to Xience [Abbott Vascular, Santa Clara, CA, USA (n = 166 patients, 182 lesions)].34 Although acute gain was lower for the bioresorbable scaffold, at 1-year interim analysis, cumulative rates of new or worsening angina were lower in the bioresorbable scaffold group, whereas performance during maximum exercise and rate of composite device-oriented endpoint were similar. Three patients in the bioresorbable scaffold group had scaffold thrombosis (definite or probable), compared with no patients in the metallic stent group. There were 17 (5%) major cardiac adverse events in the bioresorbable scaffold group compared with 5 (3%) events in the metallic stent group, with the most common adverse events being MI [15 cases (4%) vs. two cases (1%), respectively] and clinically indicated target-lesion revascularization [four cases (1%) vs. three cases (2%), respectively]. Data from Absorb III trial35 further support the non-inferiority of Absorb to Xience in terms of 1-year target lesion failure (TLF), although superiority (which was also targeted in the study design) was not achieved (TLF, 7.8% for Absorb and 6.1% for Xience; Pnon-inferiority = 0.007 and Psuperiority = 0.16). There was no statistical difference between groups in rates of cardiac death, target-vessel MI, ischaemia-driven TLR, or device thrombosis (1.5 vs. 0.7%) within 1 year. Consistent non-inferiority results were seen in the Absorb Japan36 and Absorb China37 trials. In the multicentre GHOST-EU registry,38 1731 bioresorbable scaffolds [Absorb bioresorbable vascular scaffold (BVS); Abbott Vascular, Santa Clara, CA, USA] were implanted in 1189 patients. Technical success was achieved in 99.7% of cases, while the cumulative incidence of TLF and scaffold thrombosis (definite/probable) was 2.2 and 1.5% at 30 days and 4.4 and 2.1% at 6 months. Diabetes mellitus was the only independent predictor of TLF (HR 2.41). Therefore, in spite of acceptable rates of TLF, the rate of scaffold thrombosis (early and midterm) was of concern. While trials have fulfilled statistical non-inferiority for BVS compared with DES, the non-inferiority margin was wide and several important clinical outcomes were numerically worse with BVS. This registry, as well as the results of trials, indicates that the issues of stent thrombosis and clinical efficacy (at least equivalent to that of newer DES) have to be addressed for bioresorbable scaffolds to be widely adopted in clinical practice.
In the BIOSOLVE-II trial,39 a second-generation drug-eluting absorbable metal scaffold (DREAMS 2G) was used in patients with stable or unstable angina or documented silent ischaemia. Routine angiographic surveillance at 6 months revealed a mean in-segment late lumen loss of 0.27 ± 0.37 mm and discernable vasomotion in 80% of 25 tested patients. Intravascular ultrasound and optical coherence tomography showed preservation of the scaffold area, a low mean neointimal area, and no intraluminal masses. Overall, TLF occurred in four (3%) patients with no definite or probable scaffold thrombosis observed.
Coronary sinus reduction
The management of patients with refractory angina who are not amenable to revascularization can be difficult. Verheye et al.40 used a new balloon-expandable device to reduce the coronary sinus in 104 such patients. Implantation led to an improvement of at least two angina classes in 35% and of one class in 71% of patients at 6 months. In the control group (received a sham procedure), angina class improvement by two or one class occurred in 15 and 42% of the patients, respectively (P < 0.05 for both). ‘More mechanistic markers of ischaemia reduction (improvement in exercise time or in the mean change of the wall-motion index) did not differ between the two groups’. Six-month event rates were low in both groups. This device may represent a future option for patients with refractory angina.
Plaque imaging and new pathophysiologic insights into coronary artery disease
Risk of adverse events in non-obstructive coronary lesions
Maddox et al.41 compared the rates of all-cause mortality and MI in patients with non-obstructive CAD (n = 20 899), obstructive CAD (n = 8384), and no apparent CAD (n = 8391). Non-obstructive CAD (defined as ≥1 stenosis ≥20% but no stenosis ≥70%), compared with no apparent CAD (no stenosis >20%;), was associated with a significantly greater 1-year risk of MI and all-cause mortality. One-year mortality rates increased with increasing CAD extent, ranging from 1.38% among patients without apparent CAD to 4.3% in patients with three-vessel or LM obstructive CAD. After risk adjustment, there was a stepwise increase in the risk of death or myocardial infarction with increasing extent of non-obstructive and obstructive CAD (Figure 3). These data confirm the prognostic importance of ‘mild lesions’, which is consistent with the previous observation that many infarctions are related to sites of modest luminal obstruction.
Fractional flow reserve
At the same time, 15-year follow-up data of Defer42 and 5-year follow-up of FAME43 confirm that fractional flow reserve guidance for revascularization does not result in worse long-term outcomes than angiographic guidance. At 15 years, deferral of treatment of non-significant stenoses42 was associated with similar mortality and reduced MI rate (2.2%) compared with PCI performance (10%; RR 0.22, 95% CI 0.05–0.99, P = 0.03). The meta-analysis by Johnson et al.44 established a significant relation between outcomes and FFR (clinical events increase as FFR decreases) in 6961 patients at 14 months. Revascularization shows larger benefit for lower baseline FFR, and final post-PCI FFR was inversely related with prognosis (HR 0.86, 95%CI 0.80–0.93; P < 0.001).
The impact of OCT during percutaneous coronary intervention on physician's decision-making
In 418 patients (467 stenoses) with stable or unstable angina or NSTEMI, the procedural decision-making was affected (selecting different stent length and diameter) by OCT imaging prior to PCI in 57% of stenoses.45 In 27% of stenoses, post-deployment OCT abnormalities (14.5% malapposition, 7.6% under-expansion, and 2.7% edge dissection) prompted further stent optimization using additional in-stent post-dilatation (n = 101) or placement of new stents (n = 20).
The impact of IVUS on outcomes following stenting
Intravascular ultrasound can be useful to guide stent deployment. A randomized trial compared IVUS-guided to angiography-guided everolimus stent implantation in 1400 patients with long coronary lesions46 and found that IVUS-guided stenting reduced major adverse cardiac events at 1 year by ∼50% (2.9 vs. 5.8%, HR0.48, 95% CI 0.28–0.83, P = 0.007) due to a reduction in ischaemia-driven target-lesion revascularization (2.5 vs. 5.0%, P = 0.02). Whether routine post-stenting balloon overexpansion would have achieved the same outcomes is uncertain. These results suggest value for IVUS to guide stent deployment in complex lesion subsets.
Current advances in coronary intervention aim primarily at perfecting long-term results, with the premise that bioerodable scaffolds will eventually replace metallic DESs. However, the short-term incremental benefit of technical advances or device iterations becomes increasingly difficult to demonstrate. Today, improved outcomes largely depend on defining the optimal synergy between best medical care and device-based therapies in the individual patient.
S.A. and S.W.: conceived and designed the research. S.A.: drafted the manuscript. S.W.: made critical revision of the manuscript for key intellectual content.
Conflict of interest: P.G.S. discloses the following relationships: research grant (to INSERM U1148) from Sanofi, and Servier; speaking or consulting fees from Amarin, AstraZeneca, Bayer, Boehringer-Ingelheim, Bristol-Myers-Squibb, CSL-Behring, Daiichi-Sankyo, GlaxoSmithKline, Janssen, Lilly, Novartis, Pfizer, Regado, Regeneron, Roche, Sanofi, Servier, The Medicines Company; and owns stocks from Aterovax; P.W.S. discloses relationships with Abbott laboratories, Astrazeneca, Biotronik, Cardialysis B.V, GLC research, Medtronic, Sino medical, Société Europa, Stentys France, Svelte Medical Systems, Volcano Europe; M.A. has no disclosures; W.W. is a non-executive Board member and shareholder of Argonauts partners and Celyad (formerly Cardio3Biosciences). He reports institutional research grants from Abbott Vascular, Biotronik, Boston Scientific, Medtronic, MiCell, Stentys, St Jude, Terumo.