Antithrombotic drugs in coronary artery disease: risk benefit ratio and bleeding


Sam Schulman, Thrombosis Service, HHS-General Hospital, 237 Barton Street East, Hamilton, ON L8L 2X2, Canada.
Tel.: +1 905 527 0271/44479; fax: +1 905 521 1551.


Summary.  The antithrombotic treatment of coronary artery disease is becoming increasingly complex. Aspirin is often combined with more potent antiplatelet agents such as thienopyridines and glycoprotein IIb/IIIa inhibitors. The classic anticoagulant unfractionated heparin is giving way to low-molecular-weight heparin, the pentasaccharide fondaparinux and the direct thrombin inhibitor bivalirudin. Warfarin (or another vitamin K antagonist) and antiplatelet agents are often required in combination for several months. Patients and physicians who have experienced major bleeding complications sometimes question the benefit of these treatment strategies. It is therefore crucial to try and weigh the impact on efficacy against safety. In this review the net benefit is discussed both numerically, comparing absolute reductions vs. increases in risks, and also by addressing the qualitative importance of each component in reaching the net benefit. Except for primary prophylaxis in patients at low-moderate risk for coronary events, there is a net benefit of antithrombotic therapy. With increasing severity of the coronary condition the net benefit generally prevails even with an increasing number of antithrombotic drugs combined. However, as the patient slowly stabilizes after appropriate interventions, it is necessary to de-escalate the treatment in accordance with decreasing net benefit of prolonged combination therapy.

Antithrombotic therapy in coronary artery disease (CAD) is becoming increasingly complex with the use of intracoronary stents, requiring highly efficient prevention of stent thrombosis. Combined antiplatelet therapy with aspirin and clopidogrel becomes a challenge when the patient also has a strong indication for anticoagulant therapy. In the aftermath of acute coronary syndromes (ACS), even without primary coronary intervention (PCI), the discussion about the optimal antithrombotic therapy (i.e. aspirin alone or in combination with clopidogrel or a vitamin K antagonist) is still open [1–3]. The problems clinicians have to deal with include (i) weighing the benefits of improved protection against recurrent ACS against the risk of bleeding, (ii) assessing when an apparent short-term net benefit of combination therapy disappears during long-term management and (iii) identifying the subgroups of patients that benefit the most and/or have the smallest risk of harm.

Moreover, the antithrombotic drugs currently emerging on the clinical stage will pose new challenges. These drugs include the P2Y12 receptor blockers prasugrel [4] and ticagrelor [5], specific coagulation inhibitors directed against factor Xa for acute, parenteral therapy, otamixaban [6], or rivaroxaban [7] or apixaban [8]for oral, secondary prophylaxis, or against thrombin [9,10]. Data from clinical phase II or III trials have been published but little is known about the risk of these therapies in different combinations with other antithrombotic agents.

Thrombolytic therapy is not part of the topic reviewed here but adds another practical dimension to the era of combination therapies. Instead this review will focus on the different platelet inhibitors and anticoagulants although, due to space limitations, without going into details regarding different bolus or maintenance doses or exact timing.

The rationale for combinations

Plaque rupture involves platelet activation with adhesion and aggregation as well as ensuing thrombin generation and formation of a fibrin clot [11]. It is therefore logical to suggest a therapeutic strategy reducing the activity of many components in this process. Platelets, with their large numbers of agonists and receptors, are only moderately inhibited when a single target is aimed for, such as aspirin-mediated cyclooxygenase inhibition with reduced formation of thromboxane A2. Additional blockade of the fibrinogen receptor, gpIIb/IIIa, or the ADP receptor, P2Y12, seems crucial to minimize platelet aggregation.

This review of antithrombotic therapy will start from the simplest setting of primary prevention, where only one antithrombotic drug may be indicated. Continuing with established CAD the review ends with the three conditions requiring highly effective antithrombotic treatment: ST-segment elevation myocardial infarction (STEMI), non- ST-segment elevation myocardial infarction (NSTEMI) and PCI. The effectiveness of these agents (singly and in combination) will be balanced against the harm in terms of bleeding.

The implication of hemorrhage

Bleeding is often merely reported as ‘major bleeding’, for which not only various definitions exist [12] but the components have very different clinical importance. A gastro-intestinal bleeding event, requiring hospitalization for a couple of days and a few blood transfusions but thereafter full recovery, and an intracerebral hemorrhage with residual hemiparesis and complete dependence will count the same in the definition. Whenever possible, the internal balance between the types of hemorrhage will also be accounted for.

A major hemorrhage is associated with worse prognosis, expressed as a 5-fold higher risk of death within 30 days in patients with ACS [13]. These fatal events may or may not be included in the efficacy outcome, depending on the definition including any death or only vascular death.

Prevention of ischemic heart disease (IHD) in asymptomatic patients

In a recent meta-analysis of individual data from six studies with 95 000 individuals with or without risk factors for IHD, the benefits and risks of aspirin (at doses between 75 and 500 mg) were assessed [14]. The absolute reduction of serious vascular events (myocardial infarction, stroke or vascular death) was modest compared with control (0.06% per year, = 0.0001), with a similar effect on major coronary events (0.06%) and on non-fatal myocardial infarction (0.05%). There was no reduction of vascular death or overall mortality. At the same time there was a borderline significant absolute increase in the risk of hemorrhagic stroke (0.01% per year, = 0.05) and a significant increase of major extracranial hemorrhage (0.03%, < 0.0001). Fatal hemorrhagic strokes were more frequent than fatal ischemic strokes. Taking into account that some of the extracranial hemorrhages were gastrointestinal, there may be a small net-benefit clinically by using aspirin, but this is uncertain. Interestingly, the small proportional reduction in serious vascular events did not vary significantly based on age, sex, or other co-morbidities (e.g. hypertension, hyperlipidemia, diabetes). The authors also questioned whether any observed benefit of aspirin for the prevention of vascular events might be even less in the current era given increased use of statins and antihypertensive medications in this population Table 1.

Table 1.   Effect and harm of antithrombotic prophylaxis in primary prevention of coronary artery disease
Agent [reference]PopulationEffectARR*, %HarmARI*, %
  1. ARR, absolute risk reduction in % per year; ARI, absolute risk increase in % per year; MI, myocardial infarction; INR, International Normalized Ratio.

Aspirin [14]No risk factorsMI/stroke/vascular death0.06Hemorrhagic stroke
Extracranial major hemorrhage
Warfarin, INR 1.5 [18]Men, high riskCoronary death, fatal & non-fatal MI0.26Hemorrhagic stroke
Extracranial major hemorrhage

In the largest single trial, the Women’s Health Study, primary prevention with aspirin, 100 mg, did not reduce the incidence of fatal or non-fatal myocardial infarction or death from cardiovascular causes, whereas there was an absolute reduction of ischemic stroke (0.02% per year, = 0.009) [15]. Subgroup analysis indicated that there might be more benefit in women who are at least 65 years old. There was a significant increase in gastrointestinal bleeding and a non-significant increase in intracerebral hemorrhage with aspirin.

A recent meta-analysis including 10 117 subjects enrolled in six randomized trials has also examined the role of aspirin in the prevention of cardiovascular events in subjects with diabetes [16]. Although a trend was seen toward the reduction of major cardiovascular events [cardiovascular death, non-fatal myocardial infarction, non-fatal stroke, and all-cause mortality; relative risk (RR) 0.90, 95% confidence interval (CI) 0.81–1.00; = 0.06], there were no significant decreases in the risk of myocardial infarction (0.86, 0.61–1.21), stroke (0.83, 0.60–1.14), vascular death (0.94, 0.72–1.23), or all-cause mortality (0.93, 0.82–1.05).

The combination of clopidogrel and aspirin compared with aspirin alone was studied in the CHARISMA trial, but 78% of the patients had already manifested cardiovascular disease [17]. Among the remaining asymptomatic patients with multiple risk factors, the rate of major vascular events did not differ significantly between patients on combination therapy vs. aspirin alone (6.6% vs. 5.2%, = 0.2). In addition, the rate of cardiovascular death was higher with clopidogrel (3.9% vs. 2.2%, = 0.01).

Warfarin at a low intensity, targeted at an International Normalized Ratio (INR) of 1.5 was compared with aspirin, 75 mg per day, in a factorial design in the Thrombosis Prevention Trial with 5085 men at high risk of ischemic heart disease [18]. Both aspirin and warfarin reduced the risk of a first episode of IHD by 20%, corresponding to an absolute risk reduction of 0.26% and 0.23%, respectively, per year. There was an absolute increase of hemorrhagic stroke and of extracranial bleeding of 0.03% and 0.04% per year, respectively, with warfarin compared with no warfarin and of 0.05% and 0.04% per year, respectively, with aspirin compared with no aspirin, thus still retaining a net benefit with either alternative. This applies to men in the top 20–25% risk score stratum.

In summary, primary prevention provides a clear net benefit only in patients at high risk and may then be given with either low-intensity warfarin or with aspirin. The combination of aspirin and clopidogrel for primary prevention is not indicated and may be associated with harm.

Prevention of ischemic heart disease (IHD) in symptomatic patients

The meta-analysis by the Antithrombotic Therapy Trialists’ Collaboration also addressed studies on secondary prophylaxis with aspirin [14]. The absolute risk reduction of serious vascular events was more impressive than in primary prevention, 1.5% per year compared with placebo (< 0.0001) and specifically for coronary events 1.0%. Hemorrhagic strokes and other major bleeding were incompletely reported in the trials on secondary prophylaxis but the authors identified an incidence of hemorrhagic stroke of 0.17% per year with aspirin vs. 0.09% per year on placebo Table 2.

Table 2.   Effect and harm of antithrombotic prophylaxis in secondary prevention of coronary artery disease
Agent [reference]PopulationEffectARR*, %HarmARI*, %
  1. ARR, absolute risk reduction in % per year; ARI, absolute risk increase in % per year; MI, myocardial infarction; NE, non-evaluable; ACS, acute coronary syndrome; CABG, coronary artery bypass grafting.

Aspirin [14]Previous vascular eventMI/stroke/vascular death
Major coronary events
Hemorrhagic stroke
Extracranial major hemorrhage
Clopidogrel + aspirin (vs. aspirin alone) [21]After ACSMajor coronary events1.2Major hemorrhage,
long treatment,
short treatment

Ticagrelor (+ASA) vs. Clopidogrel (+ASA) [5]After ACSMI/stroke/vascular death
All-cause mortality
Major hemorrhage (non-CABG)0.7
Warfarin + aspirin (vs. aspirin alone) [22]After ACSAll-cause mortality, non-fatal MI or ischemic stroke3.0Hemorrhagic stroke
Major hemorrhage

The efficacy and safety of combination of clopidogrel and aspirin compared with antiplatelet monotherapy was assessed in a Cochrane review [19] including two randomized clinical trials (the CHARISMA trial [17] and the Clopidogrel in Unstable Angina to Prevent Recurrent Events, CURE, study [20]). The use of combination therapy vs. aspirin alone was associated with a lower risk of cardiovascular events (OR 0.87, 95% CI 0.81–0.94) but a higher risk of major bleeding [odds ratio (OR) 1.34, 95% CI 1.14–1.57]. Interestingly, treatment effect differed by study. In the CURE study, which treated patients following NSTEMI for 3–12 months, 23 vascular events would be avoided and 10 major bleeds would be caused by treating 1000 patients with combination therapy for a mean period of 9 months. In CHARISMA, which enrolled subjects with existing vascular disease or at high risk for vascular disease, five vascular events would be avoided and three major bleeds caused for every 1000 subjects treated for 28 months [17]. Furthermore, in a meta-analysis of eight trials with 91 744 patients [21] three of the studies were in 61 905 patients with unstable angina or myocardial infarction. The follow-up in these studies ranged from 28 days to 12 months. The absolute risk reduction for major coronary events was 1.2% with dual therapy but again without any reduction in all-cause mortality. The risk of major bleeding was not increased significantly by dual therapy in these three trials (OR 1.24, 95% CI 0.97–1.59). However, when studies in patients with various indications and with long treatment (8–28 months) were analyzed there was an absolute increase of major bleeding of 1.9% vs. a non-significant increase of 0.05% in patients from studies with 1 month of treatment. Combined treatment specifically in ACS patients during the first 30 days is discussed below.

Given the success of dual antiplatelet therapy, subsequent studies have evaluated agents that may prove even more effective than clopidogrel in combination with aspirin. Clopidogrel is a prodrug and an irreversible inhibitor of the adenosine diphosphate receptor P2Y12. It was recently compared with the direct and reversible inhibitor ticagrelor in 18 624 patients with ACS, out of whom about 60% had unstable angina or NSTEMI [5]. Aspirin was also given to 97.5% of the patients. After 12 months ticagrelor provided an absolute risk reduction for the composite outcome (vascular death, myocardial infarction or stroke) of 1.9%, driven by reductions in myocardial infarction and vascular death. There was also an absolute risk reduction of all-cause mortality of 1.4%. Against this there was an absolute increase of non-bypass surgery-related major bleeding of 0.6% and life-threatening bleeding of 0.5%.

Given the importance of both platelet activity and plasma coagulation for the development of ACS and plaque development, the combination of aspirin and warfarin has been heavily studied in patients with CAD. A meta-analysis of 14 studies with 25 307 patients treated after ACS did not reveal a significant reduction of the composite of all-cause mortality and non-fatal myocardial infarction or ischemic stroke with the combination compared with aspirin alone [22]. However, when four studies with the therapeutic range INR below 2.0 were excluded, the absolute risk reduction for these events was 3.0%, driven by reductions of non-fatal myocardial infarction or ischemic stroke but not all-cause mortality. The absolute increase in the risk of major bleeding with combination therapy was 1.5% and for intracranial hemorrhage it was 0.2%, which was not statistically significant.

The largest individual study with the therapeutic range of at least INR 2.0 was the warfarin sub-study of OASIS 2 with 3712 patients [23]. The authors further sub-grouped the study population as belonging to good-complier countries, where at least 70% were still on warfarin at 35 days, and poor-complier countries, where this proportion was lower. The absolute risk reduction in these two sub-groups for the primary outcome (cardiovascular death, myocardial infarction or stroke) was 2.8% and 1.2%, respectively, whereas the absolute increase in the risk of life-threatening bleeding or major bleeding was 0.4% and 1.6% vs. 0.8% and 1.0%, respectively.

In WARIS II, which continued for a mean of 4 years, the combination of warfarin and aspirin provided an absolute risk reduction of death, non-fatal myocardial infarction or thrombotic stroke of 1.2% per year vs. aspirin alone and of 0.4% vs. higher intensity warfarin alone [24]. The risk of major bleeding increased by 0.45% per year with both the combined treatment and with warfarin alone vs. aspirin alone. The vast majority of the major bleedings were gastrointestinal. There was neither any decrease in mortality nor any increase in intracerebral hemorrhage with the combination. It should be borne in mind that 77% of cross-sectional INR results were at least 2.0 and that warfarin management is traditionally well performed in Norway.

In summary, there is a substantial net benefit of aspirin alone vs. nothing after acute coronary syndromes. The additional net benefit of aspirin plus clopidogrel is convincing for up to 3–12 months following acute coronary syndrome but is of no clear benefit thereafter. Ticagrelor may prove to be a better alternative than clopidogrel, offering increased net benefit but with an increased risk of important bleeding. The net benefit identified for aspirin plus warfarin seems to be present mainly when the treatment continues for several months or perhaps years in the case of access to good management, as seen in WARIS II.

Antithrombotic treatment in unstable angina and NSTEMI

It can be assumed that the vast majority of patients with STEMI (except those with aspirin allergy) receive aspirin, which therefore will not be discussed here.

The net-benefit derived from clopidogrel in addition to aspirin has been discussed above. The CURE trial was specifically in patients (n = 12 562) with NSTEMI [20]. During the first 30 days addition of clopidogrel to aspirin provided an absolute risk reduction of cardiovascular death and non-fatal myocardial infarction of 0.9%, whereas the absolute increase of the risk of major bleeding during the first month was 0.5%. Excess bleeding was either gastrointestinal or at the site of arterial punctures, whereas there was no difference in fatal or intracerebral bleeding. At the end of treatment (3–12 months) the absolute risk reduction of the primary outcome had increased to 2.1% and the absolute increase of risk of major hemorrhage amounted to 1.0%. Thus, the net benefit is maintained for up to 1 year Table 3.

Table 3.   Effect and harm of antithrombotic prophylaxis in unstable angina and NSTEMI
Agent [reference]PopulationEffectARR*, %HarmARI*, %
  1. ARR, absolute risk reduction in % per year; ARI, absolute risk increase in % per year; LMWH, low-molecular-weight heparin; UFH, unfractionated heparin; NSTEMI, non-ST-segment elevation myocardial infarction; MI, myocardial infarction.

Clopidogrel + aspirin (vs. aspirin alone) [20]After NSTEMI, first 30 days, 3–12 monthsCardiovascular death or
non-fatal MI
Idem plus stroke
Major hemorrhage Idem0.5
GP IIb/IIIa inhibitor [30]NSTEMI, before PCIDeath or MI before PCI  1.3Major hemorrhage0.7
LMWH vs. UFH [26]Unstable angina, NSTEMIMI
Major hemorrhage0.9
Fondaparinux vs. LMWH [27]IdemDeath at 30 days,
MI at 30 days
Major hemorrhage−1.9

A pooled analysis of three trials on aspirin with or without unfractionated heparin (UFH) demonstrated an absolute risk reduction for the composite of death and myocardial infarction of 2.7%. Bleeding was not addressed although it was slightly more common in each of the trials [25].

In a meta-analysis of the low-molecular-weight heparin (LMWH) enoxaparin vs. UFH across the spectrum of ACS, six studies on 21 945 patients with unstable angina and NSTEMI were included [26]. The treatment with LMWH was always with enoxaparin, 1 mg kg−1 every 12 h, whereas UFH was adjusted according to the activated partial thromboplastin time (aPTT). The absolute reduction of risk for myocardial infarction with LMWH was 1.1%, whereas there was an absolute increase, although non-significant, in the risk of major bleeding of 0.9%. There was no difference in mortality. The net clinical benefit was therefore neutral for the two anticoagulants in this population. In view of the short half-life of intravenous UFH and complete reversibility, this alternative is preferable to LMWH in patients for whom early catheterization and revascularization are likely.

The pentasaccharide fondaparinux at a dose of 2.5 mg daily was compared with enoxaparin 1 mg kg−1 twice daily in 20 078 patients with ACS in the OASIS-5 trial [27]. There was no reduction of the primary outcome (death, myocardial infarction or refractory ischemia) after 9 days but at 30 days there was a statistically significant reduction of deaths (absolute risk reduction 0.6%), which was at least partly explained by the absolute risk reduction of major bleeding of 1.9% both at 9 and 30 days. No difference was observed regarding myocardial infarction or refractory angina. The caveats are a long half-life of fondaparinux and an increased risk of catheter-associated thrombosis [27]. In the case of PCI a strategy with bolus doses of UFH, 50–60 units kg−1, during the procedure is recommended [28].

Bivalirudin has been given to patients with NSTEMI [29] but mainly ‘upstream’ to PCI with very short observation times before the revascularization or during the PCI and will therefore be discussed below in the PCI section.

Glycoprotein (GP) IIb/IIIa inhibitors have also been given ‘upstream’ before PCI and in a meta-analysis of three studies with 12 210 patients and three different GP IIb/IIIa inhibitors (eptifibatide, tirofiban and abciximab) the period before possible PCI could be evaluated [30]. Addition of the GP IIb/IIIa inhibitors reduced the absolute risk of death or myocardial infarction by 1.3% at the expense of an absolute increase of major bleeding of 0.7%. The bleeding was, however, mainly from puncture sites.

In summary, the net benefit of aspirin is not disputed. There is an additional net benefit from adding clopidogrel to aspirin and this seems to be maintained with up to 1 year of therapy. Additional net benefit can be obtained with early treatment with a GP IIb/IIIa inhibitor, even if acute PCI is not performed, but this treatment should be reserved for patients at moderate to high risk of myocardial infarction. There is no difference in net benefit between LMWH and UFH, whereas fondaparinux provides a net benefit in comparison with LMWH.

Antithrombotic treatment in STEMI

As with NSTEMI, aspirin has a high level of evidence for its use in STEMI. In the ISIS-2 trial with 17 187 patients and a multifactorial design with or without streptokinase and aspirin, patients were followed for 5 weeks [31]. The absolute risk reduction with aspirin, 162.5 mg day−1, was 2.4% without any increase in the incidence of intracranial hemorrhage or bleeding requiring transfusion. The benefit prevailed even after an extended follow-up of 10 years [32].

Addition of clopidogrel to aspirin was investigated in the Chinese COMMIT trial with 45 852 patients with myocardial infarction treated until discharge or up to 4 weeks [33]. The combination gave an absolute risk reduction of the composite endpoint (death, re-infarction or stroke) of 0.9% and a reduction of all-cause mortality of 0.6% without any increase of fatal or intracerebral hemorrhage and a non-significant increase of bleeding requiring transfusion of 0.04%. Similar results were obtained in the CLARITY-TIMI 28 trial with 3491 patients, performed in other parts of the world [34].

In a meta-analysis of seven trials with 68 000 patients Collins et al. [35] evaluated the addition of UFH to treatment with aspirin in patients with suspected acute myocardial infarction. There was an absolute risk reduction of deaths or re-infarction of 0.5% and 0.3%, respectively, and an absolute increase in the risk of major bleeding of 0.3%. In a meta-analysis by Eikelboom et al. [36] LMWH was compared with placebo in four trials with 16 943 patients. The absolute risk reduction of death or re-infarction was 0.9% and 0.6%, respectively, whereas the absolute increase of major bleeding or intracranial bleeding was 0.7% and 0.2%, respectively. In the largest trial, CREATE, published after the meta-analysis, the LMWH reviparin was given subcutaneously twice daily [37]. The absolute risk reduction for the composite of death, myocardial infarction or stroke at 7 days was 1.4% and remained at 30 days, driven by reductions of deaths and re-infarctions but not stroke. The absolute increase in intracranial bleeding or major bleeding was 0.2% and 0.5%.

Comparison of the LMWH enoxaparin 1 mg kg−1 subcutaneously b.i.d. (except for one trial with 40 mg t.i.d.) and aPTT-adjusted intravenous UFH was performed in a meta-analysis of six trials with 27 131 patients with STEMI [38]. The absolute risk reduction of death or non-fatal myocardial infarction was 1.9% with LMWH. The components of this composite endpoint and major bleeding are shown in Table 4. In the largest individual trial from this meta-analysis, EXTRACT-TIMI 25 with 20 749 patients, additional analysis was performed of the efficacy and safety of enoxaparin vs. UFH in patients with or without clopidogrel [39]. The absolute risk reduction for death, recurrent myocardial infarction or ischemia or stroke with or without clopidogrel was 2.9% and 2.0%, respectively, in favor of LMWH. The absolute increase of risk of major hemorrhage on LMWH with or without clopidogrel was 1.7% and 0.9%, respectively, thus preserving a net clinical benefit for enoxaparin vs. UFH in either situation.

Table 4.   Effect and harm of antithrombotic prophylaxis in STEMI
Agent [reference]PopulationEffectARR*, %HarmARI*, %
  1. ARR, absolute risk reduction in % per year; ARI, absolute risk increase in % per year; LMWH, low-molecular-weight heparin; UFH, unfractionated heparin; STEMI, ST-segment elevation myocardial infarction; MI, myocardial infarction.

Clopidogrel + aspirin (vs. aspirin alone) [33]After STEMI, up to 4 weeksDeath, re-infarction or stroke0.9Fatal or intracerebral hemorrhage
Bleeding with transfusion
UFH + aspirin (vs. aspirin alone) [35]Suspected MIDeath
Major hemorrhage0.3
LMWH + aspirin (vs. aspirin alone) [36]STEMI at 7 daysDeath
Intracranial hemorrhage
Major hemorrhage
Fondaparinux (vs. placebo) [40]STEMI at 9 daysDeath or re-infarction1.6Major hemorrhage−0.6
Major hemorrhage0.8
Fondaparinux vs. UFH [40]STEMI at 9 daysDeath or re-infarction0.5Major hemorrhage−0.2
Abciximab [41]STEMI, treated with thrombolysisDeath MI0.3
Intracranial hemorrhage
Severe hemorrhage

Fondaparinux was compared against placebo (stratum 1) or against UFH (stratum 2) in the OASIS 6 study with 12 092 patients [40]. The absolute risk reduction for death or re-infarction after 9 days was 1.6% against placebo and non-significant 0.5% against UFH and these results were essentially preserved through the 30-day assessment until the end of study at 3 months. There was actually an absolute decrease of risk for major bleeding, although not statistically significant, of 0.6% against placebo and 0.2% against UFH and no difference in intracranial bleeding.

In order to investigate whether a GP IIb/IIIa inhibitor improves the benefit of thrombolytic therapy in myocardial infarction, 16 588 patients, treated with aspirin and UFH, were randomized in the GUSTO V trial to reteplase alone or reteplase at half dose in combination with the GP IIb/IIIa inhibitor abciximab [41]. The combination did not result in any significant reduction of death but there was an absolute risk reduction of myocardial infarction and of recurrent ischemia of 1.2% and 1.5%, respectively. On the other hand, there was an absolute increase of the risk of severe bleeding of 0.6% but no difference in intracranial bleeding. Direct thrombin inhibitors will be discussed in the following section on PCI.

In summary, aspirin provides a substantial net benefit and this is augmented by clopidogrel, at least during 4 weeks of combined therapy. The net benefit with UFH is relatively small; it is larger with LMWH and head-to-head comparison between the two confirms the advantage of LMWH. Fondaparinux also provides a net benefit vs. UFH. There may be a small net benefit of GP IIb/IIIa inhibitors added to thrombolytic therapy.

Antithrombotic treatment in PCI

Early trials established the superiority of combined antiplatelet therapy (aspirin plus ticlopidine) initiated prior to PCI and continued for 30 days over aspirin alone or aspirin plus warfarin to prevent early and late stent thrombosis following PCI [42–45]. By virtue of a superior safety profile, similar efficacy and once per day dosing clopidogrel became the thienopyridine of choice for this indication [46–48].

Subsequent trials have shown that extended treatment with aspirin and clopidogrel after PCI/stenting for ACS or after an elective procedure reduces cardiovascular events. The addition of clopidogrel to aspirin was evaluated in a meta-analysis of one trial and two sub-studies of larger trials [21]. The absolute risk reduction of major coronary events was 3.9% and there was a non-significant 0.5% reduction for death, as opposed to an absolute increase of major bleeding of 0.9%Table 5.

Table 5.   Effect and harm of antithrombotic prophylaxis in PCI
Agent [reference]PopulationEffectARR*, %HarmARI*, %
  1. ARR, absolute risk reduction in % per year; ARI, absolute risk increase in % per year; PCI, percutaneous coronary intervention; STEMI, ST-segment elevation myocardial infarction; NSTEMI, non-ST-segment elevation myocardial infarction; MI, myocardial infarction; CABG, coronary artery by-pass grafting.

Clopidogrel + aspirin (vs. aspirin alone) [21]After PCI, 30 days–12 monthsMajor coronary events
All-cause mortality
Major hemorrhage0.9
Prasugrel vs. clopidogrel [4]PCI in STEMICardiovascular death, MI or stroke3.0Non-CABG major hemorrhage0.3
GP IIb/IIIa inhibitor [52]PCI in NSTEMIDeath or MI1.0Intracranial hemorrhage Major hemorrhage0.03
Bivalirudin (see text) [54–56]PCIDeath≈0Major hemorrhage−0.7 to −6.0

Extended treatment with dual antiplatelet therapy might be particularly important following insertion of drug-eluting stents, which exhibit delayed endothelialization compared with bare metal stents. Late stent thrombosis was noted to occur in 0.19% of patients in a large drug-eluting stent registry and conferred significant associated morbidity and mortality [49]. Discontinuation of antiplatelet therapy was one of the most potent predictors of late stent thrombosis in patients with drug-eluting stents. Data from the Duke Cardiovascular Database found that among patients with drug-eluting stents, continued use of clopidogrel was associated with significantly lower rates of death (0% vs. 3.5%) or death and MI (0.0% vs. 4.5%, < 0.001) [50] although bleeding was not analyzed. These data led to a recent American Heart Association/American College of Cardiology Science advisory panel stressing the importance of 12 months of dual antiplatelet therapy after drug-eluting stent placement.

Prasugrel is a novel oral thienopyridine that in preliminary studies inhibited ADP-induced platelet aggregation more rapidly and to a greater extent than clopidogrel. It was recently compared with clopidogrel for the prevention of cardiovascular events in 13 608 patients with ACS and scheduled PCI [51]. After 12 months follow-up, prasugrel was associated with 2.2% absolute decrease in the primary endpoint of vascular death, myocardial infarction or stroke (9.9% vs. 12.1%, < 0.001). However, there was also a small excess of major bleeding (2.4% vs. 1.8%, = 0.03) and life-threatening bleeding (1.4% vs. 0.9%; = 0.01).

In the subset of 3534 patients with STEMI [4] there was an absolute risk reduction for the primary endpoint (cardiovascular death, non-fatal myocardial infarction or stroke) of 3.0% without any significant increase in the risk of bleeding at 30 days. There was an increase in major bleeding in the subset of patients who required coronary artery by-pass surgery.

Boersma et al. [52] performed a meta-analysis of the effect and safety of GP IIb/IIIa inhibitors based on individual data from patients with NSTEMI and PCI in studies with at least 1000 patients. Six trials with 31 402 patients were identified. At 30 days the absolute risk reduction for death or myocardial infarction was 1.0% with increased benefit in high-risk patients, particularly those with elevated troponin level. This was balanced by a similar absolute increase in major bleeding of 1.0% but no substantial difference in the risk of intracranial hemorrhage.

The patients who are taken to the catheterization lab for primary PCI are usually already on an anticoagulant (e.g. UFH, LMWH or fondaparinux). In the case of the latter, bolus doses and flushing of the catheter with UFH is recommended, as discussed above. An alternative to these indirect, antithrombin-dependent anticoagulants is the direct and reversible thrombin inhibitor bivalirudin. In a meta-analysis of five trials comparing bivalirudin with UFH or LMWH in 25 457 patients with ACS, including patients who had PCI, all were treated with aspirin and a proportion of patients also received a thienopyridine [53]. A GP IIb/IIIa inhibitor was planned in all patients treated with heparins in two of the studies whereas it was either provisional or only planned for a proportion of the patients in the bivalirudin arms. Follow-up varied from 48 h to 1 year. For the primary efficacy outcome of death, myocardial infarction or revascularization there was no difference (relative risk 1.01) between arms, but there was a significantly lower risk of bleeding in the bivalirudin groups (relative risk 0.55, 95% CI 0.40–0.76). In three studies, each with at least 3000 patients, evaluating bivalirudin vs. heparins there was an absolute risk reduction for major bleeding ranging from 0.7% to 6.0% [54–56]. This may be of limited advantage if the routine is to use the radial artery rather than the femoral, because most of the bleeding is related to the groin puncture.

In summary, dual antiplatelet therapy is strongly recommended for at least 1 month following stent procedures (and 3–6 months following drug-eluting stents) to prevent stent thrombosis. Extended duration dual antiplatelet therapy (up to 12 months) should be considered for patients at low risk of bleeding. Compared with clopidogrel, prasugrel may offer an additional benefit with respect to prevention against vascular events but was associated with increased bleeding. There is also a small net benefit from adding a GP IIb/IIIa inhibitor in higher risk ACS patients undergoing PCI, based on a comparatively low rate of the most serious bleeding complications. There is a net benefit from bivalirudin compared with UFH/LMWH ± a GP IIb/IIIa inhibitor based on a decrease in bleeding events.

Triple therapy

Use of triple therapy is most commonly considered following PCI with stent placement in a patient receiving warfarin for atrial fibrillation with many risk factors for stroke. There are substantial data supporting the efficacy of clopidogrel plus aspirin over warfarin (alone or in combination with an antiplatelet agent) to prevent stent thrombosis. However, dual antiplatelet therapy is significantly less effective than warfarin for the prevention of stroke in patients with atrial fibrillation. Clinicians must generally choose between continuing all three agents for a finite duration (minimum 1 month for bare metal stent, 3–6 months for drug-eluting stent) vs. holding warfarin for this same period. However, the utilization of all three therapies is likely to be associated with an increase in bleeding. In one review of 12 studies, rates of major bleeding were 4.6% at 30 days and 10.3% at 6–12 months [57]. This risk has to be weighed against the risk of stroke if warfarin is withheld: in one small retrospective study, 8.8% of subjects on aspirin and clopidogrel alone suffered stroke at 12-month follow-up compared with 2.8% of subjects on triple therapy [58].

A recent consensus document of the European Society of Cardiology provides recommendations regarding therapy in this population based on clinical setting (elective procedure vs. acute coronary syndrome-related), type of stent (bare metal vs. drug-eluting), and bleeding risk [59]. In general, utilization of a bare metal stent over drug-eluting stent is strongly recommended to avoid prolonged periods of triple therapy. Utilization of concomitant GP IIb/IIIa inhibitors is also discouraged. After elective placement of a bare metal stent or emergency placement of a bare metal stent following ACS (and high bleeding risk), triple therapy is recommended for 4 weeks followed by return to warfarin alone. For drug-eluting stents or bare metal stents following ACS (and with non-high bleeding risk), triple therapy is recommended for 3–6 months.

Perioperative anticoagulation with pacemaker/defibrillator implantation

Development of pocket site hematoma and subsequent site infection with peri-procedural anticoagulation in patients requiring warfarin (for high-risk atrial fibrillation, valvular disease, etc.) is an all too common clinical problem. Options with respect to perioperative anticoagulation include performance of procedure on warfarin, discontinuation of warfarin with pre- and postprocedure bridging with UFH or LMWH or discontinuation of warfarin without bridging. In a systematic review of eight studies comparing continuation of warfarin with warfarin cessation with bridging (IV UFH), pocket hematomas occurred at a rate of 1.9–6.6% vs. 12–20%, respectively [60]. The incidence of thrombotic events was < 1% irrespective of strategy employed. In a single center observational study of 148 patients undergoing device implantation, four different bridging protocols were used. The different treatment regimens with LWMH were used sequentially over a 2-year period [61]. The use of preoperative LMWH did not affect pocket hematoma rates (12% vs. 17%) but postoperative LMWH was associated with a significant increase compared with omission of postprocedure LMWH (23% vs. 8%, = 0.01). No embolic events were observed in the study cohort. Finally, the periprocedure incidence of pocket hematoma was compared in 109 patients with mechanical heart valves undergoing pacemaker implantation [62]. Although prospective randomized trials are needed, available data suggest that avoidance of full dose UFH or LMWH after device implantation is prudent.


The expanding armada of antithrombotic drugs available to cardiologists has contributed to a substantial reduction of serious cardiovascular events and in some cases also reduced all-cause mortality. The absolute increase in the risk of bleeding balances sometimes that of absolute risk reduction, eliminating the numerical net benefit. When the components of major bleeding are considered, with the majority of those often being puncture-site related or gastro-intestinal, the clinical impact is usually lower than after a myocardial infarction or stroke.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.