Critical appraisal of the AUGUSTUS trial Free

Previous trials comparing dual antithrombotic therapy (DAT) with rivaroxaban or dabigatran plus a P₂Y₁₂ inhibitor to triple antithrombotic therapy (TAT) with warfarin, aspirin, and a P₂Y₁₂ blocker revealed a significant reduction of bleeding events with DAT. However, the lower bleeding risk with DAT in these studies may be attributable to the use of a direct oral anticoagulant (DOAC) instead of a vitamin K antagonist (VKA), or to the omission of aspirin, or to both interventions. AUGUSTUS is the first trial with a two-by-two factorial design in patients with atrial fibrillation (AF) on combined antithrombotic therapy after an acute coronary syndrome (ACS) and/or percutaneous coronary intervention (PCI), respectively.¹ Thereby, it allows the independent evaluation of the effects of prescribing apixaban instead of warfarin and skipping aspirin in the same patient population.

Risk reduction and net clinical benefit

Apixaban was associated with a significantly lower bleeding risk than warfarin, with a number needed-to-treat over 6 months of 24 to prevent one event. The lowest incidence of the primary safety endpoint was observed in patients on DAT with apixaban (16.8/100 patient-years) with a relative risk reduction (RRR) of 65.8% compared with TAT including warfarin (49.1). Of note, DAT with warfarin carried a lower bleeding risk (26.7) than TAT with apixaban (33.6). Together with the results of PIONEER AF-PCI and RE-DUAL PCI, the major clinical implications of these findings are that aspirin should be omitted whenever the ischaemic risk does not exceed the risk of bleeding. Further, the DOAC apixaban should be preferred to a VKA in the case of DAT and TAT. These results may not be directly translated to other DOACs as it remains unclear how much the DOAC contributed to bleeding risk reduction in the other studies. Further, differences between the studies regarding the duration of aspirin treatment and the proportion of follow-up on aspirin influence the magnitude of the effect of skipping aspirin: while RE-DUAL PCI showed a smaller RRR with DAT than AUGUSTUS, particularly with dabigatran 150 mg with –28%, the study design may have also resulted in a smaller contribution of skipping aspirin than in AUGUSTUS. In PIONEER AF-PCI, the RRR of –37% with DAT using rivaroxaban 15 mg plus a P₂Y₁₂ inhibitor and no aspirin was also smaller than in AUGUSTUS, with a study design likely to induce a comparable or smaller effect of skipping aspirin compared with AUGUSTUS. With regard to published subgroup analyses of AUGUSTUS, the safety and efficacy results were consistent, including high-risk patients, with one exception: in patients with diabetes (36%), a significant effect modification was observed, with a weaker effect of apixaban for the combined endpoint of death and ischaemic events as well as a borderline effect modification for the combined endpoint of death and hospitalization. The net benefit (including bleeding, death, and ischaemic events) is likely to be preserved in patients with diabetes due to a consistent reduction of bleeding. Further analyses of effect modification indicated a trend (P = 0.052) towards a particular safety benefit for apixaban in ACS patients without PCI, which also requires further investigations.

The right dosage of DOACs

In contrast to previous DOAC studies in this indication, AUGUSTUS predominantly tested the approved standard dose for AF (5 mg b.i.d.) and achieved a reduction of bleeding including ISTH major bleeding. Based on the reduction of major bleeding in ARISTOTLE, there was a well-grounded rationale for using the higher dose. Indeed, the standard dose with only 10% of patients with dose reduction based on specific criteria showed a consistent bleeding reduction even in patients on TAT. These results may not be translated to other DOACs as dabigatran and rivaroxaban were never investigated in a TAT regimen. Furthermore, the other tested DOACs struggle with the dose to be used in patients with AF and PCI or ACS. Rivaroxaban was primarily evaluated in a lower dose of 15 mg without having the power to assess thrombo-embolic events. However, from ROCKET AF, we know that even the higher dose of 20 mg did not significantly reduce thrombo-embolic events compared with warfarin in the intention-to-treat analysis. In contrast to rivaroxaban, dabigatran was tested in both dosages approved for AF: while dabigatran 110 mg b.i.d. was associated with a numerically higher number of myocardial infarctions and stent thrombosis, dabigatran 150 mg b.i.d. showed a more moderate reduction of bleeding but without signs of increased ischaemic endpoints.

Duration of triple antithrombotic therapy: a personalized approach

In contrast to current international guidelines, which recommend a TAT of 6 months only in AF patients at very high risk of ischaemic events after PCI, all patients in the TAT group of AUGUSTUS were treated with aspirin for 6 months.² Consequently, a substantial number of low- to moderate-risk patients received long-term TAT in AUGUSTUS. This may have resulted in an (avoidable) increase in bleeding events in the TAT arm. On the other hand, AUGUSTUS reported a numeric increase in stent thrombosis when aspirin was omitted. In order to separate patients in need of intensified antithrombotic treatment from those at risk of net harm due to overtreatment, it would be of particular interest to analyse safety and efficacy in subgroups with low, moderate, and high risk of adverse ischaemic outcomes in AUGUSTUS. It may be speculated that well-defined high-risk patients may require a prolonged time of TAT or a more potent P₂Y₁₂ antagonist than clopidogrel, the inhibitor of choice in most patients with DAT. Unfortunately, information on newer P₂Y₁₂ antagonists as part of DAT or TAT remains scarce, although AUGUSTUS is already the third DOAC trial on this topic. Regarding the ideal timing of omitting aspirin, individual patient characteristics favouring prolonged TAT, e.g. the acuteness of the coronary event, but also the complexity of coronary lesions and PCI, need to be taken into account. Of note, there is currently no evidence-based approach to identify patients in need of TAT. However, since individuals were randomized several days after the index event (up to 2 weeks in AUGUSTUS), a minimal duration of TAT of 1 week seems justified in all patients based on current evidence.

Methodological considerations

With regard to the study design, AUGUSTUS was an open-label trial for OAC which might introduce information bias. On the other hand, double blinding of OAC using sham international normalized ratio (INR) measurements may have led to a very complex study protocol. While the other studies on this topic were not blinded for the treatment regimen at all, AUGUSTUS introduced partial blinding for aspirin. Furthermore, a comparably low time in therapeutic range (TTR) with a median of only 59% within the VKA arm may raise concerns about potential conclusions from AUGUSTUS. However, although the majority of INR values outside the target range were <2 (median 23%) and only rarely >3 (median 3%), suggesting undertreatment in the VKA arm, even fewer bleeding events were observed within the apixaban arm, reflecting a major safety benefit for the DOAC. With regard to the power of AUGUSTUS, the higher than expected number of bleeding endpoints indicates sufficient power for safety analyses. Additionally, AUGUSTUS represents the largest study in this field also increasing the power for less frequent ischaemic endpoints. However, it needs to be considered that none of the mentioned trials was adequately powered to draw a definitive conclusion on ischaemic endpoints. Future meta-analyses may help to fill this gap of knowledge.

Conclusion of current evidence

Like the Roman Emperor Augustus, the study of the same name has resolved many conflicts, claiming dominance with a superior study design for the tested drug regime in patients with AF undergoing PCI or suffering an ACS. The study demonstrated superiority of apixaban with the dose approved for AF compared with a VKA regardless of skipping aspirin. However, TAT for 1 week after stent implantation in the DAT arm needs to be taken into consideration, based on the late randomization in many individuals. It remains to be proven if the reduction of bleeding directly related to apixaban indicates a class effect of DOACs in this clinical situation. While skipping aspirin resulted in an even greater reduction of bleeding in AUGUSTUS, uncertainty regarding the efficacy of DAT in the prevention of ischaemic outcomes remains. Consequently, aspirin may be safely omitted in most patients at hospital discharge, but a minority of patients with high risk of ischaemic events may still benefit from prolonged TAT or a stronger P₂Y₁₂ inhibitor than clopidogrel. This group of patients needs to be more clearly defined in future.

Conflicts of interest: T.G. reports personal fees from Bayer, Daiichi Sankyo, Boehringer-Ingelheim, Bristol Myers Squibb, and Pfizer, outside the submitted work. P.S. reports grants from Daiichi Sankyo and Boehringer-Ingelheim, outside the submitted work. A.N. reports personal fees from Bayer, BMS, and Pfizer; and grants and personal fees from Boehringer Ingelheim and Daiichi Sankyo, outside the submitted work.

References