OUP user menu

Benefits and risks with antiplatelet therapy: how great a problem is bleeding?

Steen Husted
DOI: http://dx.doi.org/10.1093/eurheartj/sun036 I19-I24 First published online: 31 October 2008


Significant advances in treatment strategies for patients with acute coronary syndromes (ACS) have been made over the past 10 years, including the introduction of new antiplatelet therapies. Additionally, modifications have been made in the dosing of these agents, the duration of treatment has been increased, and they are used in dual- and triple-therapeutic regimens. On the basis of a number of large-scale clinical trials, the current guidelines recommend the use of aspirin in combination with clopidogrel as a first-line therapy in patients with ACS. While data show significantly reduced ischaemic events with combination antithrombotic therapy, this is often associated with an increase in major bleeding episodes. Because of these risks, it is important that treatment with antithrombotics follow appropriate guidelines with a risk/benefit ratio calculated for each patient. Additionally, agents with the lowest bleeding risk should be used first. This article reviews the benefits and risks of antiplatelet therapy with emphasis on the potential for serious bleeding episodes.

  • Acute coronary syndrome
  • Antithrombotic agents
  • Antiplatelet agents
  • Clopidogrel
  • Bleeding


Antithrombotic agents, antiplatelet and anticoagulant agents, are supposed to reduce the risk of death from myocardial infarction (MI). Over the past 20 years, antithrombotic agents have been shown to lower the risk of death and MI from ∼20% to between 4 and 8% in patients with non–ST-segment elevation acute coronary syndromes (non-STE ACS).1 The original treatment for these patients was aspirin and heparin, then heparin in combination with glycoprotein (GP) IIb/IIIa-receptor blockers, low–molecular-weight heparin (LMWH) and, most recently, clopidogrel. As a result, in-hospital mortality from non-STE ACS is low, but with this comes an increased risk of bleeding.

This paper focuses on three main topics: the relationship between bleeding and mortality in patients receiving antiplatelet therapy, the extent of the bleeding risk associated with each antiplatelet drug administered alone or in combination, and the ways in which this risk can be managed.

Bleeding and mortality

A database of major randomized trials showed that bleeding in the acute setting leads to an increased risk of death even when the bleeding is not very severe (Figure 1).2 Rao et al. showed that it is not only the performance of coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI) that can lead to an increased risk of bleeding and death, but also the potential of spontaneous bleeding in these patients.

Figure 1

Bleed is associated with an increased 30-day mortality in NSTEMI patients. NSTEMI, non-ST elevation myocardial infarction; GUSTO IIb, Global Use of Strategies To Open occluded arteries in acute coronary syndromes; PURSUIT, Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy; PARAGON A&B, Platelet IIb/IIIa Antagonism for the Reduction of Acute Coronary Syndrome Events in a Global Organization Network A and B trials. n = 26 452 acute coronary syndrome patients from GUSTO IIb, PURSUIT and PARAGON A&B. From Rao et al.2

Data from Eikelboom et al.3 showed a very high risk of bleeding in the acute setting when patients were treated with antithrombotics. Patients who bled had a noticeably increased risk of death over the 30 day observation period. Again, these deaths were not attributable to a specific bleeding episode but to something else, which may be the undertreatment of ACS. The question that needs to be asked is why there is a persistent increased risk of death after 30 days? Data from the Clopidogrel in Unstable angina to prevent Recurrent ischaemic Events (CURE) trial shows a relationship between the severity of bleeding and the subsequent risk of death.4

An analysis by Alexander et al.5 from the Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA guidelines (CRUSADE) database shows that patients are not always treated with the correct dose of antithrombotic agents. Even though dosing guidelines have been established, physicians do not always choose the appropriate dose, considering age and kidney function among other factors. It may well be that physicians prescribe doses that are higher than necessary, especially with GP-receptor antagonists. Patients who are severely overdosed have a high risk of bleeding. This underlines the importance of prescribing these medications appropriately.

We know from the Global Registry of Acute Coronary Events (GRACE) registry that there are independent predictors, including older age, female gender, renal failure, history of bleeding, right-side heart catheterization, and the use of GP-receptor antagonists, for increased risk of bleeding that can be identified in patients.6

Bleeding may lead to the complete cessation of antithrombotic therapy with drugs such as aspirin, clopidogrel, and the anticoagulants. It is also important to watch for hypotension, which can result from excessive bleeding. However, the sudden cessation of antithrombotic therapy may lead to the recurrence of a thrombotic episode or even a kind of rebound so that the patients may experience a higher than normal risk of thrombi.

Data from CRUSADE show the age-related use of packed red blood cells (RBCs) in patients with ACS.7 Elderly patients have an increased risk of bleeding, and women bleed more than men. Although the reason is not known, blood transfusions are associated with a worse outcome. Packed RBCs are depleted of nitrous oxide and function as a nitrous oxide ‘sink'. Further, the depletion of nitrous oxide leads to vasoconstriction, increased risk of platelet aggregation, and insufficient oxygen delivery. A study addressing this question showed an increase in C-reactive protein (CRP) and interleukin-6 levels in patients receiving blood transfusions. It is known that an inflammatory response is a risk factor for death or even new MIs.

In summary, bleeding occurs more frequently than commonly thought, and clinical bleeding and transfusions are associated with a worsening of clinical outcome. The studies cited here and the other, more recent studies suggest that strategies, which maintain an adequate antithrombotic effect to reduce ischaemia while minimizing the risk of bleeding, may improve patient survival. So if, in the acute setting, a safer therapy is selected, it may result in lower mortality even though it may not be more efficacious. This has been confirmed in at least two recent anticoagulant studies, the Organization to Assess Strategies for Ischemic Syndromes (OASIS) V and VI trials.

Bleeding and the activities of the antiplatelet agents

In the study by Patrono et al.8 investigating the use of aspirin, the absolute risk reduction of aspirin was very low in low-risk populations (Figure 2), and the risk/benefit ratio was not favourable. In patients who have already had an acute ischaemic event, the use of aspirin can lead to an acceptable risk/benefit ratio. The risk of having major bleeds is one to two events per 1000 patients treated, which is about twice that of placebo. The risk of intracranial bleeding is also increased, with a rate of one to two per 10 000 patients treated. Risk factors for aspirin-related bleeding include taking more aspirin than 75–100 mg/day, increased age, a history of peptic ulcer disease, and concomitant treatment with non-steroidal anti-inflammatory drugs, clopidogrel, and warfarin.

Figure 2

Benefits and risks of low-dose aspirin in primary-prevent trials. WHS, Women’s Health Study; PHS, Physicians Health Study; PPP, Primary Prevention Project; HOT, Hypertension Optimal Treatment Study; BDT, British Doctors Trial; TPT, Thrombosis Prevention Trial; SAPAT, Swedish Angina Pectoris Aspirin Trial. From Patrono et al.8

The Antithrombotic Trialists’ Collaboration showed a similar risk reduction of ∼25% in major ischaemic events among patients taking either 75–150 or 160–325 mg of aspirin per day.9 However, data from the CURE trial showed that the administration of >200 mg of aspirin per day increases the risk of bleeding from the combination of clopidogrel and aspirin.4 As a result, the use of low-dose aspirin is recommended in combination with clopidogrel.

Andreotti et al.10 performed a meta-analysis of studies in patients with coronary artery disease treated with a combination of aspirin and warfarin or aspirin alone. The combination is more effective than aspirin alone in reducing ischaemic events but doubles the risk of major bleeding. One would have to treat 33 patients to avoid an ischaemic event and 100 patients to have one major bleed. Data from the CURE trial showed a 20% relative risk reduction in the incidence of major ischaemic events in patients with non-STE ACS treated with clopidogrel plus aspirin compared with aspirin alone, with an absolute risk reduction of 2.1%, from 11.4 to 9.3%.4 This means that 48 patients would have to be treated to obtain a beneficial effect. But the combination of clopidogrel and aspirin also leads to a 1% absolute risk increase in major or life-threatening bleeds. So that the number needed to treat is 48 and the number needed to harm is 100, which is a very narrow ratio. Therefore, unless patients with an increased bleeding risk are treated with stents, they should avoid the combination of aspirin and clopidogrel.

In the Clopidogrel for the Reduction of Events During Observation (CREDO) study, patients were treated with clopidogrel either before or after stenting.11 One group of patients had prolonged treatment with clopidogrel combined with aspirin and the other group received aspirin monotherapy. The absolute risk reduction in combined endpoint (death, MI, or stroke) was 3% in patients who received long-term (1 year) clopidogrel following stenting. However, clopidogrel also caused a 2% increased risk of major bleeding, which equates to treating 33 patients to achieve a beneficial effect and 50 patients to have a major bleed, a very narrow window.

Results from the Clopidogrel and Metoprolol in Myocardial Infarction Trial (COMMIT)12 and the CLopidogrel as Adjunctive ReperfusIon TherapY (CLARITY) trial13 showed that clopidogrel in combination with aspirin administered to patients who were revascularized using thrombolysis did not result in a greater incidence of bleeding than aspirin alone. This lack of an increased bleeding risk may be related to the fact that the observation period for therapy with clopidogrel and aspirin was relatively short (a maximum of 30 days).

The Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial assessed patients older than 45 years of age who were at high risk for atherothrombotic events.14 Patients were randomized and treated for up to 30 months with a combination of clopidogrel and aspirin or aspirin alone, with no significant beneficial effect seen with the dual-antiplatelet therapy. In a later publication from this study, Bhatt et al.15 evaluated the time-related risk of bleeding (Figure 3) and showed that after 1 year of therapy dual-platelet inhibition did not lead to more bleeding than aspirin alone. The risk of spontaneous bleeding is augmented during the first year, but not beyond.

Figure 3

Timing of moderate or severe bleeding. ASA, aspirin. From Bhatt et al.15

A Danish population-based, case–control study using a prescription database compared 1443 cases of serious upper gastrointestinal (GI) bleeding with 57 720 age- and gender-matched controls.16 The number of patients actually admitted to the hospital with GI bleeding was determined (Table 1), and adjusted odds ratios associating drug use with upper GI bleeding were calculated. Combination therapy was associated with a higher risk for bleeding than monotherapy, and aspirin and clopidogrel had the highest risk, followed by aspirin and vitamin K antagonists. Low-dose aspirin had the highest risk among monotherapies, followed by clopidogrel.

View this table:
Table 1

Single and combined antithrombotic therapy and upper gastrointestinal bleeding

Drug treatmentPopulation-based case-controlled study
Cases (exposed/unexposed)Controls (exposed/unexposed)Crude odds ratio (95% CI)Adjusted odds ratio (95% CI)a
Aspirin alone196/10634123/50 4982.4 (2.0–2.8)1.8 (1.5–2.1)
VKA alone12/1063203/50 4983.1 (1.7–5.6)1.1 (0.6–2.1)
Aspirin and clopidogrel13/106349/50 49812.6 (6.6–24)7.4 (3.5–15)
Aspirin and VKA16/1063114/50 4986.4 (3.7–11)5.3 (2.9–9.5)
Aspirin and dipyridamol44/1063737/50 4983.1 (2.2–4.2)2.3 (1.7–3.3)
  • Adapted from Hallas et al.16

  • VKA, vitamin K antagonist.

  • aAdjusted for previous discharge diagnosis.

It is accepted that stabilizing therapy with GP-receptor antagonists in patients not undergoing CABG or a PCI will not lead to a significant risk reduction. However, upfront treatment before these procedures will reduce the risk of important ischaemic events, with an absolute risk reduction of 3%.17 It is noteworthy that within the entire patient population, two major ischaemic events will be prevented for every 100 patients treated, but at the risk of two major extracranial bleeds. Actually, GP-receptor antagonists do not seem to increase the risk of intracranial bleeding.

Factors that may lead to bleeding are renal insufficiency, female gender, the bail-out use of GP-receptor antagonists, and combination therapy with anticoagulants.18,19

A very important question is how safe triple-antithrombotic therapy is. Can an anticoagulant, long-term LMWH, or long-term warfarin, be safely given to patients who are already receiving dual-platelet inhibition? There are only small studies to address this question. One study from 2006 looked at patients receiving anticoagulant therapy and dual-platelet inhibition because of a PCI.20 Administering dual-platelet inhibition to patients already receiving warfarin will lead to a significant increase in major and minor bleeding episodes.

Another small study was designed to address the risk of bleeding with triple-antithrombotic treatment regimens.21 Data showed a 9.2% incidence of major bleeding over a 4-week period in 66 patients receiving triple therapy with warfarin, aspirin, and clopidogrel after PCI with stent placement.

A larger study was conducted in 21 443 patients over 65 years of age who had survived an MI, and were treated with combination therapy or aspirin or warfarin alone.22 Aspirin plus clopidogrel, aspirin plus warfarin, and the three drugs in combination led to a doubling of bleeding risk over aspirin alone, although the overall risk was relatively small.

Finally, a study by Kajalainen et al.23 evaluated antithrombotic treatment over 12 months following coronary stenting in 239 patients requiring long-term anticoagulation. Treatment regimens included warfarin plus aspirin, clopidogrel plus aspirin, and triple therapy. The data showed that aspirin plus warfarin is inadequate to prevent stent thrombosis. For most patients, dual-platelet inhibition combined with an anticoagulant is the most beneficial treatment and perhaps even the safest if it is not administered for long term. The use of clopidogrel in combination with warfarin may be a reasonable treatment option in patients with a suspected increased risk of bleeding.

Managing the increased risk of bleeding

Treatment with dual- or triple-antithrombotic therapy must follow guidelines, and excessive dosing must be avoided. Antithrombotic agents with a low bleeding risk should be used. Data from the OASIS V trial showed that fondaparinux can reduce the risk of ischaemic events similar to enoxaparin with less major bleeding.24

Patients who are likely to require invasive procedures or surgery should have bare-metal stents inserted when possible and these procedures should be postponed until a safer antithrombotic treatment is possible for these specific patients. Thienopyridine therapy is recommended for 1 month in patients receiving bare-metal stents, 3 months in patients receiving a Cypher® stent (Cordis Corporation, Miami Lakes, FL, USA), and 6 months in patients receiving a Taxus® stent (Boston Scientific Corporation, Natick, MA, USA). Longer term treatment may be indicated, but should not be initiated in these patients if the risk of bleeding is increased. Thus, it is important to consider the kind of stent to be used based on the risk stratification of patients.


Antithrombotic treatment must follow guidelines, and the risk/benefit ratio should be calculated for each patient. Invasive treatment strategies should follow guidelines based on the risk stratification of patients, who should be tested carefully for the possible risk of a future ischaemic event before any intervention is initiated.

Lastly, all healthcare providers treating patients with ACS should have guidelines for handling bleeding complications during antithrombotic therapy and know how to resume therapy once the bleeding has been controlled. Guidelines on how to organize antithrombotic therapy during procedures and operations should be available to physicians handling patients with ACS.


Financial support provided by Daiichi Sankyo, Inc. and Eli Lilly and Company.

Conflict of interest: S.H. has received research support from AstraZeneca Pharmaceuticals LP, Bayer AG, Bristol-Myers Squibb Company, and sanofi-aventis LLC.


View Abstract