In the past decade, the introduction of newer potent P2Y₁₂ inhibitors (e.g., ticagrelor and prasugrel) has helped to further reduce the occurrence of ischemic events in coronary artery disease (CAD) patients [5]. Meanwhile, the development of new generations of drug-eluting stents, such as biodegradable polymer stents, also appears to have lowered the thrombotic risks following a percutaneous coronary intervention (PCI), when compared with the older bare metal stents [6]. In view of the improved APT potency and stent safety, the balance between ischemic and bleeding risks must also be managed in further detail.

Physicians working in the area may often wonder, why do results from the large number of clinical studies appear to be conflicting? For example, while some trials (e.g., PEGASUS-TIMI 54 [7]) suggest better outcomes with an extended DAPT duration, others support shortened DAPT (e.g., DAPT-STEMI [8]). Other trials suggest switching from DAPT to P2Y₁₂i monotherapy by dropping aspirin (e.g., TWILIGHT [9]), or to a different P2Y₁₂i dose or agent (e.g., HOST-REDUCE POLYTECH-ACS [10] and TOPIC [11]). To answer this question, it is essential to realize that these trials target different patient populations and are concerned with different research questions and objectives.

Depending on the specific aims of APT, different strategies may be adopted (Table 1). Table 1 defines short-, medium-, and long-term APT as approximately <1 month, 1–12 months, and >12 months, which are arbitrary divisions that coincide with common designs of randomized controlled trials (RCTs) of APT. In practice, APT duration is often a moving target [12] that is contingent on patient factors and treatment tolerance. While, at hospital discharge, it may not be possible to determine a patient’s risk over time, risk assessment should be re-evaluated regularly [13].

In addition to the differences in medication strategy, the trials were conducted in different patient groups (e.g., those with acute coronary syndrome [ACS] or stable CAD [sCAD]) and regions (e.g., U.S., Europe, or the Asia-Pacific), using various measurement criteria (e.g., Thrombolysis in Myocardial Infarction [TIMI] or Bleeding Academic Research Consortium [BARC] bleeding criteria). This review aims to categorize the recent results, and layout an important conceptual framework that underlies these studies, namely, ischemic and bleeding risks may vary for different patients at different time points.

Patients who recently had an ACS or are indicated for PCI have an elevated risk of experiencing an ischemic event (including recurrent MI), particularly in the first 30 days [14,15]. Although there are some suggestions of a decreasing trend in recurrent coronary hospitalization in recent years [16], the risk remains high, especially for patients with additional risk factors [17]. The aim of APT in these patients, by and large, is to aggressively reduce their ischemic risk, while avoiding any excessive increase in bleeding risk (Table 1).

Landmark RCTs that have established a standard DAPT duration of 12 months include CURE [18], PLATO [19], and TRITON [20], in which the ischemic benefits appeared to outweigh the bleeding risks. For example, in PLATO [19], where ACS patients were randomized to receive ticagrelor 90 mg twice daily (BID) versus clopidogrel 75 mg once daily (QD), the occurrences of vascular death, MI, or stroke at 12 months were 9.8% versus 11.7%, respectively (p < 0.001), and the rates of major bleeding were 11.6% versus 11.2% (non-significant [N.S.]). All-cause deaths occurred in 4.5% versus 5.9% (p < 0.001) of patients in the two arms, respectively.

Recent Asian studies of 1-year DAPT in ACS patients, such as PHILO [21], TICAKOREA [22], and PRASFIT-Practice-II [23,24], reported somewhat lower rates of ischemic events. TICAKOREA [22] also reported significantly reduced bleeding rates for patients treated with clopidogrel versus ticagrelor. While these results might reflect the more recent and Asian clinical scenarios, these studies also had smaller sample sizes compared with the trials above. APT for Asian patients will be further discussed in Section 5.

Hypothetically, P2Y₁₂i monotherapy may provide two potential benefits over traditional DAPT: first, it may reduce bleeding while providing similar ischemic protection in the medium term; second, it reduces the medication burden in the longer term (e.g., when administered beyond 1 year).

Notable trials include TWILIGHT [9,26], SMART-CHOICE [28], and STOPDAPT-2 [29,30]. TWILIGHT demonstrated significantly reduced bleeding at 15 months in patients treated with ticagrelor monotherapy after 3 months of DAPT, compared with those who continued DAPT, both in the overall population (4.0% [ticagrelor alone] vs. 7.1% [ticagrelor + aspirin], p < 0.001) [26] and the ACS subgroup (3.6% [ticagrelor alone] vs. 7.6% [ticagrelor + aspirin], p < 0.001), but not in the sCAD subgroup (4.8% [ticagrelor alone] vs. 6.2% [ticagrelor + aspirin]; N.S.) [9].

In the Asian studies SMART-CHOICE [28] and STOP-DAPT2 [29], with PCI patients, switching to P2Y₁₂i monotherapy also reduced bleeding without compromising ischemic event prevention. However, STOP-DAPT2-ACS [30], where ACS patients were switched from DAPT to clopidogrel monotherapy, did not achieve noninferiority, and there was a marginal increase in the major composite ischemic endpoint (2.8% vs. 1.9%, hazard ratio [HR] = 1.50, 95% confidence interval [CI]: 0.99–2.26), including a HR of 1.91 (95% CI: 1.06–3.44) for MI. One explanation could be that 1 month of DAPT was too short for ACS patients, whose conditions are more severe and unstable, and clopidogrel resistance might also have affected ischemic outcomes.

Another strategy is de-escalation, where DAPT continues at a reduced dose or duration, or with a less potent P2Y₁₂i. Both “unguided” (by randomized allocation only) and “guided” (e.g., by platelet function test [PFT] or genotyping) de-escalation approaches have produced favorable results. A recent network meta-analysis [46] compared APT trials that shortened DAPT with those that reduced P2Y₁₂i dosage or potency (total 29 trials; 50,602 patients), and found no difference in all-cause death between the two. Reducing P2Y₁₂i dosage or potency was favored in terms of trial-defined net adverse CV events (NACE; risk ratio [RR] = 0.87, 95% CI: 0.70–0.94), but not with respect to bleeding (RR = 1.54, 95% CI: 1.07–2.21). However, because some of the sample sizes in the escalation and de-escalation studies were relatively small, and most were open-label, adjudicator-blinded studies, there could potentially be some effects of patient selection, as well as bias in the reporting of both physician- and patient-reported clinical outcomes. More large-scale studies are required for further comparison.

It is worth noting that the time of de-escalation chosen in these trials vary in aggressiveness, from 1, 3 to 6 months after starting DAPT, i.e., when ischemic and bleeding risks remain high to becoming more stable. While these trials generally demonstrated a reduction in bleeding events without increasing ischemic events significantly, in real-life, the time chosen for de-escalation will depend on the patient’s characteristics and evolving risks.

Currently, two kinds of test are available for helping to select patients for the different APT strategies: PFT and genotyping. PFT measures platelet activation levels and may be performed at baseline and during APT [47]. Different laboratory techniques may be used, including light transmission, electrical impedance, and flow cytometry [47]. The RPFA-VerifyNow^® P2Y12 test is a point-of-care whole blood test for monitoring clopidogrel resistance; results are expressed as P2Y₁₂ reaction units (PRU) [47]. Genotyping identifies cytochrome P450 loss-of-function (LOF) mutations, which are associated with clopidogrel resistance because they reduce the liver’s ability to metabolize clopidogrel into its active form [48].

In ANTARTIC [38], depending on PFT results, patients receiving DAPT could be escalated from prasugrel 5 mg QD to 10 mg QD (for those with high platelet reactivity [HPR]) or de-escalated to clopidogrel 75 mg QD (for those with low platelet reactivity). However, the trial failed to achieve superiority over DAPT with prasugrel 5 mg QD. In TROPICAL-ACS [39] and POPular Genetics [40], noninferiority was demonstrated for guided de-escalation from a potent P2Y₁₂i to clopidogrel based on PFT results. PATH-PCI [42] escalated patients with high platelet maximum aggregation rate (>55%) from clopidogrel to ticagrelor, and produced a significant net clinical benefit.

In a meta-analysis [49] of guided-DAPT, encompassing 11 RCTs (six PFT-guided and five genotype-guided trials) and three observational studies (all genotype-guided studies) with 20,743 patients, guided APT was associated with reduced trial-defined major adverse CV events (MACEs; RR = 0.78, p = 0.015), CV death (RR = 0.77, p = 0.049), MI (RR = 0.76, p = 0.021), stent thrombosis (RR = 0.64, p = 0.011), stroke (RR = 0.66, p = 0.010), and minor bleeding (RR = 0.78, p = 0.003), but not all-cause death and major bleeding. The authors noted that, generally, guided escalation was associated with a reduction in ischemic risks without safety tradeoffs, whereas guided de-escalation was associated with bleeding reductions without efficacy tradeoffs [49].

TAILOR-PCI [41] enrolled 5,302 patients to receive genotype-guided or conventional DAPT. CYP2C19 carriers in the genotype-guided arm received ticagrelor, and all other patients received clopidogrel. In a primary analysis of 1,849 CYP2C19 LOF carriers, composite CV death, MI, stroke, stent thrombosis, and severe recurrent ischemia occurred in 4.0% (35/903) and 5.9% (54/946) of patients in the genotype-guided and conventional arms, respectively, but the difference did not reach statistical significance (p = 0.06). None of the 11 prespecified secondary endpoints, including major or minor bleeding, demonstrated statistical significance, except marginally for stent thrombosis (p = 0.05).

Nevertheless, an updated meta-analysis [50] of 11 RCTs (11,740 patients) on genotype-guided APT vs. standard treatment demonstrated significant reductions across all reported efficacy outcomes, including trial-reported MACEs (RR = 0.60, p = 0.001), all-cause death (RR = 0.70, p = 0.02), CV death (RR = 0.71, p = 0.02), MI (RR = 0.53, p < 0.0001), stroke (RR = 0.64, p = 0.04), stent thrombosis (RR = 0.63, p = 0.01), and target vessel revascularization (RR = 0.79, p = 0.003). Differences in all bleeding outcomes were non-significant: BARC types 2,3,5: RR = 0.87, p = 0.13; BARC types 3,5: RR = 1.14, p = 0.44; TIMI major: RR = 1.05, p = 0.81; TIMI minor: RR = 1.04, p = 0.88. Of note, the subgroup analysis suggested that genotype-guided APT was more likely to reduce MACEs in populations that consist of more ACS or Chinese patients [50].

Because point-of-care PFT is common, and genotyping results can be produced within a few days (in POPular Genetics, the median time between blood collection and genotyping result was 4 h only [51]), guided escalation and de-escalation may be performed quite readily, even within the first 2 weeks after PCI, as in the trials. However, Angiolillo et al. [4] cautioned that patients who are de-escalated to clopidogrel could in fact have HPR, and because 7–14 days of maintenance clopidogrel is required after de-escalation to assess platelet function, they can be subject to an increased risk of thrombosis.

Studies on MI recurrence generally suggest that, in 30-day survivors of acute MI, mortality rates plateau at about 3 years after the first index MI [52]. To prevent long-term ischemic events, several large-scale studies have investigated the efficacy and safety of extending DAPT from 1 year to about 3 years, most notably the DAPT [43] and PEGASUS TIMI-54 [7] trials. The DAPT trial [43] reported a 1.6% absolute reduction in all-cause death, MI, or stroke after 30 versus 12 months of DAPT with prasugrel or clopidogrel, which was coupled with a 0.9% absolute increase in moderate or severe bleeding according to the GUSTO (Global Use of Streptokinase and Tissue plasminogen activator to Open occluded coronary arteries) criteria.

PEGASUS [7] recruited patients who had a prior MI 1–3 years previously. Extended DAPT with ticagrelor plus aspirin achieved a 1.1% (ticagrelor 60 mg BID vs. aspirin alone, p = 0.004) or 1.2% (ticagrelor 90 mg BID vs. aspirin alone, p = 0.008) absolute reduction in CV death, MI, or stroke at 36 months, which was accompanied by a 1.2% or 1.5% absolute increase in TIMI major bleeding, for the two ticagrelor doses respectively (both p < 0.001). A post-hoc subgroup analysis of PEGASUS [53] illustrated that in patients with no bleeding risk indicators and ≥2 ischemic risk indicators (59% of 13,938 patients), ticagrelor significantly reduced the primary composite efficacy endpoint of CV death, MI, or stroke by 1.9% (p = 0.0024), and TIMI major bleeding (primary safety endpoint) only by 1.0% (p < 0.001). Given a moderate increase in bleeding, extended DAPT would likely benefit those who have elevated ischemic risks (e.g., impaired renal function, large atherosclerotic burden, multiple stents) and relatively low bleeding risks (e.g., young age; See Section 5).

THEMSIS-PCI [44] recruited patients with sCAD and diabetes mellitus, and found that, among those who underwent PCI, 3.3 years of ticagrelor (mostly at the lower 60-mg BID dose) led to a 1.3% absolute decrease in CV death, MI, or stroke, and a 0.9% increase in TIMI major bleeding. The significant ischemic benefit was not observed in patients without PCI.

Trials have also considered long-term P2Y₁₂i monotherapy. GLOBAL LEADERS [31] demonstrated no significant differences between 1-month DAPT plus 23-month ticagrelor monotherapy versus 24-month DAPT, both in terms of ischemic and bleeding events, but these results were not sufficient for establishing superiority. The pre-specified subgroup analysis [54] revealed that BARC type 3 or 5 bleeding occurred in 1.95% versus 2.68% of ACS patients (p = 0.037), compared with 2.13% versus 1.62% in sCAD patients (p = 0.081), while differences in the primary endpoint of all-cause death or new Q-wave MI remained non-significant. In the ACS subgroup, there was a significant reduction in all-cause death, new Q-wave MI, and BARC type 3 or 5 bleeding when taken together (rate ratio = 0.81, p = 0.029). Although the superiority hypothesis was not sustained overall, the subgroup analysis suggests that ACS patients may still benefit from ticagrelor monotherapy following abbreviated DAPT. In the post-hoc landmark analysis of GLOBAL-LEADERS [55], which included patients who were event-free at 12 months, the second year of ticagrelor monotherapy demonstrated lower composite all-cause death, MI, or stroke compared with aspirin monotherapy (1.9% vs. 2.6%, log-rank p = 0.014, adjusted p = 0.022) that was driven by reduced MI (0.7% vs. 1.2%, p = 0.003). The authors also noted that the difference in BARC type 3 or 5 bleeding (0.5% vs. 0.3%, log-rank p = 0.051, adjusted p = 0.005) was significant only after adjustment for characteristics of patients excluded from the second-year analysis due to clinical events or nonadherence.

HOST-EXAM [33] enrolled PCI patients who were event-free after 6–18 months of prior DAPT. After another 24 months, compared with aspirin monotherapy, patients who received clopidogrel monotherapy had a reduced composite outcome of all-cause death, non-fatal MI, stroke, ACS re-admission, and BARC type ≥ 3 bleeding (5.7% vs. 7.7%, p = 0.003). One caution is that while both ischemic and bleeding endpoints decreased, all-cause deaths remained comparable (1.9% vs. 1.3%, p = 0.101).

COMPASS [45] investigated whether low-dose rivaroxaban, alone or in combination with aspirin, would be more effective for secondary CV prevention than aspirin alone. The trial recruited 27,395 patients with sCAD and/or peripheral arterial disease, of whom 62% had previous MI and 21% had heart failure. Patients who were already using anticoagulants were excluded, including those with atrial fibrillation (AF) receiving rivaroxaban at the standard dosage.

Participants were randomized to rivaroxaban plus aspirin, rivaroxaban alone, or aspirin alone. The trial was stopped at a mean follow-up of 23 months for superiority of the rivaroxaban plus aspirin combination. Compared with aspirin alone, there was a 1.3% absolute reduction in CV death, MI, or stroke, together with a 1.2% increase in modified ISTH (International Society on Thrombosis and Haemostasis) bleeding, which included hospitalized bleeding. Detailed analysis [56] also showed a significant reduction in stroke occurrences in the rivaroxaban plus aspirin group over the aspirin alone group (0.9% vs. 1.6% per year, p < 0.0001). There were significantly fewer cardioembolic strokes (p = 0.006) and embolic strokes of undetermined source (p = 0.006) in the rivaroxaban plus aspirin arm, compared with aspirin alone (secondary analysis) [57]. Niessner et al. [58] commented that subclinical AF might have underlain such results, as AF can be quite prevalent among peripheral arterial disease patients. During the 23-month follow-up, 49 patients (0.2% of 27,395) were diagnosed with AF [57].