The periprocedural management of OAC in patients treated with vitamin K antagonists (VKAs) who require elective CORO and/or PCI remains controversial as both OAC continuation and interruption are widely used in clinical practice. However, in 2014 the European Society of Cardiology (ESC) Working Group on Thrombosis recommended to continue OAC with VKA in a therapeutic window (i.e., INR ≥2.0) in patients undergoing PCI [1] (Table 1.2). This strategy has also been endorsed in the 2014 ESC Guidelines on myocardial revascularization [2]. Although there is no randomized evidence on this issue, available registries on VKAs reported no difference in the rate of bleeding or thrombotic events in patients undergoing PCI with uninterrupted OAC compared to OAC discontinuation and bridging with unfractionated (UFH) or low-molecular-weight (LMWH) heparin [3, 4]. These findings were recently corroborated by a subgroup analysis of the WOEST study reporting the absence of significant differences in bleeding and major adverse cardiac and cerebrovascular events (MACCEs), including death, myocardial infarction, stroke, target vessel revascularization, and stent thrombosis, in patients undergoing PCI with uninterrupted OAC or bridging therapy [5]. A recent analysis from the Atrial Fibrillation undergoing Coronary Artery Stenting (AFCAS) registry reported a higher rate of major bleeding at three months in AF patients who underwent PCI and were discharged on triple therapy (TT) of VKA, aspirin, and clopidogrel in combination with bridging LMWH compared to those discharged on TT alone [6]. Thus, it is possible that the short time period of exposure to a quadruple therapy, including VKA, aspirin, clopidogrel, and LMWH, may increase the risk of bleeding without preventing MACCEs. An additional theoretical advantage of uninterrupted OAC is the elimination of the prothrombotic state occurring early after VKA initiation in relation to the transient suppression of the anticoagulant protein C and S, whose half-life is shorter than that of most of the coagulation factors targeted by VKAs (Table 1.3). Therefore, uninterrupted, effective (i.e., INR ≥2.0), periprocedural OAC seems as safe as, and potentially more effective than, OAC interruption plus bridging strategy, with the additional advantage of reduced length of hospitalization. The concern of bleeding events occurring during ongoing, effective (i.e., INR ≥2.0) OAC should be considered in the context of rapid reversal of the effect of VKAs with the administration of several therapies, including prothrombin complex concentrates, fresh frozen plasma, and recombinant factor VII, in combination or not with oral or intravenous vitamin K [7].

As regards the intra-procedural anticoagulation, additional heparin in patients undergoing PCI (mostly by the femoral approach) during effective OAC has been reported to increase the risk of vascular access site complications with no benefit on the risk of MACCEs [6]. In contrast, a significantly higher rate of radial occlusion in patients undergoing CORO/PCI by the radial route and receiving OAC without additional UFH compared to patients receiving UFH has been recently reported [8]. The intra-procedural administration of additional UFH, at the reduced dose of 50 IU/kg currently recommended for radial procedures [9], is therefore an option to be routinely considered (Table 1.2). Although the use of bivalirudin is considered to be associated with a lower risk of bleeding than UFH [10], a recent randomized trial comparing bivalirudin vs. UFH in patients at high risk of bleeding undergoing elective PCI reported similar rates of major in-hospital bleeding [11]. Therefore, while having been reported more effective and safer than the combination of UFH and glycoprotein IIb/IIIa inhibitors in small cohorts [12], the use of bivalirudin in patients with uninterrupted OAC undergoing CORO/PCI may not afford a greater prevention of bleeding compared with UFH and is therefore currently not recommended (Table 1.2). Moreover, the anticoagulant effect of UFH may be rapidly reverted by the use of protamine. In a meta-analysis of 6,762 patients undergoing PCI, the use of protamine after coronary stenting was safe, by allowing early sheath removal and reducing the risk of bleeding complications [13].

As regards the choice of the vascular access site in patients on uninterrupted OAC with VKA undergoing CORO/PCI, there is broad agreement to prefer the radial approach over the standard femoral route [1] (Table 1.2). Several randomized trials and meta-analyses demonstrated that the radial approach reduces the risk of vascular complications and major bleeding compared to the femoral approach [14, 15]. Similar results have also been reported in a small, single-center experience with patients on OAC with VKA [16]. However, an important caveat is that the radial approach may have several shortcomings in patients with previous coronary artery bypass graft surgery (CABG). Indeed, a randomized trial, comparing radial vs. femoral approach in patients who had previously undergone CABG, showed that radial diagnostic CORO is associated with greater contrast use, longer procedure time, and greater access crossover and operator radiation exposure compared with femoral angiography [17]. Thus, in these patients, as in other patients where the radial approach is not feasible or has failed, the puncture of the femoral artery, provided that it is carried out with the proper technique (i.e., puncture of the anterior wall only) and possibly with ultrasound guidance, may be considered. When the femoral approach has been used, the routine use of vascular closure devices (VCD) is advised, even though data on VKA patients undergoing PCI are currently not available. In patients at increased risk of periprocedural bleeding and/or submitted to aggressive antithrombotic therapy, VCD have indeed been shown to reduce the time to ambulation, and hence hospital discharge, while appearing comparably effective and safe [18].

The advent of bare-metal stents (BMSs) resolved the threat of abrupt vessel closure following balloon angioplasty, thus eliminating the need for standby surgical backup. Subsequently, DESs with release of antiproliferative drugs during the first months after implantation successfully addressed the problem of restenosis inherent to BMSs due to potent suppression of neointimal hyperplasia [19] (Table 1.1). A relevant shortcoming of early-generation DESs was a delayed healing response of the stented coronary vessel that was associated with a small but notable increase in late thrombotic events [20] (Table 1.4). New-generation DESs were developed featuring thinner stent struts, novel-durable or biodegradable polymer coatings or even no coating, and various antiproliferative agents at lower dosages [21]. These refinements resulted not only in a 50 % reduction in the risk of definite or probable stent thrombosis compared to early-generation DESs during long-term follow-up but also in improved efficacy (10–20 % lower risk of repeat revascularization) and safety (lower risk of death and myocardial infarction). A network meta-analysis of 76 trials with 117,762 patient-years of follow-up comparing DESs with BMSs revealed a lower risk (18–37 %) of myocardial infarction among patients treated with DESs (except early-generation, paclitaxel-eluting stents) compared to those treated with BMSs [22]. A pooled analysis of 26 randomized trials in 11,557 women showed a significantly lower rate of death or myocardial infarction in patients treated with new-generation DESs compared with early-generation DESs and BMSs at 3 years [23]. Therefore, new-generation DESs represent the standard of care in patients undergoing PCI and are currently indicated in nearly all patient and lesion subsets [2]. No data are available regarding the use of bioresorbable vascular scaffolds (BVSs) in patients on VKA undergoing PCI. Although undergoing progressive reabsorption to complete disappearance, the time required for such process (i.e., approximately 2 years) and the recommended duration of dual antiplatelet therapy (DAPT) (i.e., at least 6 months) [24] currently make these devices of limited applicability for this patient subset. Similarly, no data are available with drug-eluting balloons (DEBs), whose indication for a short duration of DAPT (i.e., 1–3 months) [25] makes these devices attractive for patients on OAC. Whereas they may be considered when accepted indications (i.e., in-stent restenosis and possibly small vessel disease) in non-OAC patients are present, DEBs are currently not routinely recommended in OAC patients undergoing PCI.

In AF patients undergoing PCI, current guidelines recommend the use of new-generation DES over BMSs in patients requiring OAC who are at low bleeding risk (i.e., HAS-BLED score ≤2) (Table 1.5) [1, 2] (Tables 1.5 and 1.6). In contrast, among patients undergoing PCI who require OAC and have a high bleeding risk (i.e., HAS-BLED score ≥3) (Tables 1.5 and 1.6), the choice between BMSs and new-generation DESs needs to be decided on an individual basis [1, 2]. However, in patients with uncertain indication for DESs, the Zotarolimus-eluting Endeavor Sprint Stent in Uncertain DES Candidates (ZEUS) trial reported a significantly lower rate of major adverse events in patients randomly allocated to the new-generation zotarolimus-eluting stent compared to those treated with BMSs, driven by reductions in target vessel revascularization, myocardial infarction, and definite/probable stent thrombosis [26]. Of note, among 1,606 patients included in the ZEUS trial, 52 % were deemed at high risk of bleeding, while 45 % and 63 % of patients discontinued DAPT at 1 and 2 months, respectively [26]. At variance with the clinical context of acute coronary syndrome (ACS), where there is indication to prolong DAPT up to 12 months irrespective of the type (i.e., BMS or DES) of stent implanted, and even regardless of whether a stent has been implanted or not, the duration of DAPT in the elective setting is driven by the type of stent, with 1 and 6 (or even 1–3 only) months being required in the case of BMSs or new-generation DESs implantation, respectively [2].

In elective PCI patients on OAC with VKA, antiplatelet therapy should include aspirin and clopidogrel [1, 27, 28] (Table 1.6). Although there is no guideline-based recommendation, the choice of clopidogrel loading dose (300 vs. 600 mg) should be carefully evaluated in patients who are going to be treated with triple therapy (TT). ESC Guidelines on myocardial revascularization strongly recommend the higher 600 mg clopidogrel loading [2], given the more rapid and intense antiplatelet effect. Indeed, the Intracoronary Stenting and Antithrombotic Regimen: Choose between 3 high Oral doses for immediate Clopidogrel Effect (ISAR-CH) trial did not find an additional suppression of platelet function with single doses of clopidogrel higher than 600 mg [29], while meta-analysis data showed a greater protection from ischemic events with a loading dose >300 mg [30]. Therefore, it seems reasonable to use a high clopidogrel loading dose (600 mg) in the majority of cases and reserve the standard loading dose (300 mg) to patients with very high risk of bleeding. At the same time, the risk of periprocedural myocardial infarction and stent thrombosis should be also weighted, again favoring the 600 mg loading dose in lesion and patient subsets at increased risk of periprocedural thrombotic events. Because of the time required for clopidogrel to be effective (i.e., approximately 2–4 and 6–8 h with 600 and 300 mg, respectively), timely pretreatment should generally be considered in elective patients referred for CORO/PCI. Newer P2Y₁₂ ADP-receptor inhibitors such as prasugrel and ticagrelor are currently not approved for elective patients, and their use as part of TT should be avoided due to the excessive risk of bleeding [1] (Table 1.6). In a series of 377 patients who underwent DES implantation (for both stable and unstable coronary artery disease) and had indication for OAC, the use of prasugrel compared with clopidogrel significantly increased the risk of bleeding at 6 months, without differences in terms of ischemic events [31]. At variance, in a small retrospective study, the use of ticagrelor in combination with VKA in patients with ACS was not associated with an increased risk of bleeding compared to TT including VKA, aspirin, and clopidogrel [32], possibly making this combination an option needing however still to be proven.