OUP user menu

Combined antiplatelet and novel oral anticoagulant therapy after acute coronary syndrome: is three a crowd?

Freek W.A. Verheugt
DOI: http://dx.doi.org/10.1093/eurheartj/eht075 1618-1620 First published online: 12 March 2013

This editorial refers to ‘New oral anticoagulants in addition to single or dual antiplatelet therapy after an acute coronary syndrome: a systematic review and meta-analysis’, by J. Oldgren et al., on page 1670

The past 20 years have seen substantial advances in the short- and long-term outcomes for patients after acute coronary syndromes (ACS). Improvements in therapy and adherence to guidelines have resulted in a reduction in hospital mortality from 10.4% in 1994 to 6.3% in 2006.1 Despite this, ∼17% of patients surviving an ACS event would experience recurrent events without secondary prophylaxis, and even with the newer antithrombotics there remains an ∼10% risk of recurrence over 12 months.2,3 The high rates of recurrence highlight the need for more effective secondary prevention strategies. Besides novel antiplatelet therapies.35 novel anticoagulants have also been developed. They have first been tested in atrial fibrillation and shown to be effective in stroke prevention.6 Given the positive results obtained with warfarin on top of aspirin after myocardial infarction published >10 years ago,7,8 they have also been tested on a large scale after ACS.9,10

A meta-analysis about the use of the new-generation oral anticoagulant agents on top of single or dual antiplatelet therapy after ACS has now been published.11 Properly randomized controlled trials are lumped together for the evaluation of efficacy against the usual endpoints (cardiovascular) death, (re)infarction, or stroke. Also, the safety with these new agents on top of the standard of dual antiplatelet therapy has been evaluated. As expected, bleeding is increased by adding a second antithrombotic strategy on top of dual antiplatelet therapy. However, efficacy is increased as well; but here the trials clearly are heterogeneous in outcome. The strength of the paper is the large number of trials and patients included. The weakness is that it is just a meta-analysis, and that all types of trial designs have been included: phase II dose-finding studies as well as phase III registration trials. Especially in phase II trials the safety of the agent tested is more carefully scrutinized than its efficacy, whereas in phase III trials efficacy is the primary outcome for ischaemic endpoints, which is subsequently mirrored against the expected increase in bleeding. Doses of the novel anticoagulants with excessive bleeding, that never made it to phase III evaluation, have been included, which has overestimated the bleeding risk in this meta-analysis. Therefore, only the phase III trials APPRAISE-29 and ATLAS-210 should have been included. The former trial has been stopped prematurely because of excess bleeding. Apparently, the dose of apixaban was too high and in fact similar to the dose successfully applied in stroke prevention in atrial fibrillation.12 In ATLAS-2, only a quarter or half of the dose used in stroke prevention in atrial fibrillation13 was tested, and found to be successful, albeit with more bleeding.

Secondly, the comparator in this meta-analysis is either dual or single antiplatelet therapy, the former being the standard of care in patients who have survived an ACS whether it has been treated invasively or conservatively, or whether a stent has been implanted or not. Therefore, only the components of the phase III trials with dual antiplatelet therapy exclusively should have been included.

The 15% relative risk reduction in cardiovascular death, myocardial infarction, and stroke in the two phase III trials is also accompanied by a reduction in stent thrombosis seen with both rivaroxaban and apixaban (Table 1), which is comparable with the reduction seen with, for example, the novel antiplatelet agent ticagrelor when compared with clopidogrel.3 This remarkable finding has not been seen before with anticoagulants and suggests a role for the coagulation cascade in the pathogenesis of stent thrombosis formerly thought to be a platelet activation-derived phenomenon. Interestingly, triple antiplatelet therapy with aspirin, clopidogrel, and vorapaxar for ACS did not reduce stent thrombosis when compared with clopidogrel and aspirin alone.14

View this table:
Table 1

Stent thrombosis in the APPRAISE-29 and ATLAS-210 studies

Trialf/u (m)nReported stent thrombosis (%/year)HR (95% CI)NNT
APPRAISE-2873920.91.30.73 (0.47–1.12)*250
ATLAS-2a1315 5261.11.50.69 (0.51–0.93)**250
  • CI, confidence interval; f/u (m), follow-up (months); HR, hazard ratio; n, number of patients; NNT, number needed to treat; NOAC, novel oral anticoagulant.

  • *P = 0.15; **P = 0.02

  • aBoth doses of rivaroxaban.

Unfortunately, the meta-analysis is not based on individual patient data. So, which ACS patients are eligible for a novel oral anticoagulant on top of dual antiplatelet therapy after ACS? Basically, all patients randomized in ATLAS-2 are eligible in that their acute event occurred at least 4 days earlier, that they should not have experienced a previous stroke while on dual antiplatelet agents, and that they are on low-dose rather than high-dose aspirin. Clearly, the novel antiplatelet agents ticagrelor and prasugrel have not been tested in combination with rivaroxaban. Though more effective than clopidogrel, both antiplatelet agents increase short- and long-term bleeding, and as of now should not be used in conjunction with a novel oral anticoagulant, at least not in the acute phase of ACS.

Should dual antiplatelet therapy with ticagrelor or prasugrel be a preferred antithrombotic strategy over rivaroxaban plus dual antiplatelet therapy with aspirin and clopidogrel? For prasugrel, the answer is clear. In the TRILOGY trial,5 prasugrel was started 4 days after ACS, which is at exactly the same time after ACS when patients were started on therapy in ATLAS-2, but there was no benefit with prasugrel over clopidogrel. So, rivaroxaban can be an excellent and effective alternative, albeit that it takes three medicines rather than two in such a secondary prevention strategy. For ticagrelor, the answer is unclear. Although one should be cautious with intertrial comparisons, ticagrelor reduced cardiovascular events including total mortality and stent thrombosis to the same extent as rivaroxaban 2.5 mg daily did in ATLAS-2, but on the other hand has the relative charm of two medicines rather than three. Another problem is that ticagrelor induced excess fatal intracranial bleedings in the PLATO study3 (11 for ticagrelor and 1 for clopidogrel), whereas low-dose rivaroxaban in ATLAS-2 did not. Anyway, only direct comparison trials in secondary prevention of ACS will give a definitive answer on whether a single pathway (dual antiplatelet therapy) or a dual pathway strategy (dual antiplatelet plus novel anticoagulant therapy) will be the most effective and safe. However, it is unlikely that such trials will be carried out.

Bleeding is the Achilles heel of all combined antithrombotic strategies. To prevent bleeding, omitting one component may be helpful. The first randomized trial in this field was done in stented patients on warfarin plus dual vs. single antiplatelet therapy, where omitting aspirin was much safer and did not seem to increases ischaemic events such as stroke, myocardial infarction, or stent thrombosis.15 More and much larger antithrombotic trials without aspirin are underway, such as the large GLOBAL LEADERS and COMPASS (NCT01776424) trials. In the former, ticagrelor plus aspirin will be compared with ticagrelor without aspirin in ACS patients undergoing percutaneous coronary intervention and followed for 24 months. In COMPASS, rivaroxaban 2.5 mg daily with aspirin will be tested vs. rivaroxaban 5 mg b.i.d. without aspirin in patients with remote ACS without the need for dual antiplatelet therapy.

In conclusion, in secondary prevention of ACS, novel oral anticoagulant therapy on top of dual antiplatelet therapy reduces recurrent ischaemic events by ∼15%, but triples the risk of clinically significant bleeding. Furthermore, a significant mortality benefit is seen with a low dose of rivaroxaban in the largest trial. These findings underscore the need to find the optimal dosing with these agents or, in other words, to find the ‘sweet spot’.

Conflict of interest: F.W.A.V. has received educational and research grants from Bayer Healthcare, Roche, Eli Lilly, and Boehringer Ingelheim, as well as honoraria for consultancies/presentations from Daiichi-Sankyo, Eli Lilly, Merck, AstraZeneca, The Medicines Company, and Bayer Healthcare


  • The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.

  • doi:10.1093/eurheartj/eht049.