Introduction The optimum treatment strategy for patients who undergo PCI but also require anticoagulation remains undetermined. There have been a handful of studies to date that suggest altered doses of rivaroxaban may lead to reduced bleeding, with similar ischaemic endpoints. Rivaroxaban 2.5 mg BD (Riv2.5) has attracted a lot of attention, namely due to 1) COMPASS, which suggested benefits in addition to aspirin in patients with coronary artery disease; 2) The approval of Riv2.5 by the National Institute for Health and Clinical Excellence for acute coronary syndrome (ACS); and 3) PIONEER-AF, suggesting reduced bleeding with Riv2.5 vs vitamin K antagonist (VKA) in those with atrial fibrillation. It could therefore be postulated that Riv2.5 may be used in all patients having PCI whether stable, acute, or requiring anticoagulation. We present real world’ data of Riv2.5 vs VKA in patients on dual antiplatelet therapy (DAPT).

Methods All consecutive patients undergoing PCI in a single centre from 2013–17 were included in the analysis and DAPT with aspirin and clopidogrel was used as the comparator. Patients on DAPT with ticagrelor or prasugrel were not considered here. Triple therapy was defined as the use of DAPT with either warfarin (VKA), Riv2.5, or full dose rivaroxaban (Riv20). The primary outcome was bleeding as defined by Bleeding Academic Research Consortium (BARC) type 3 to 5 bleeding at 12 months. Secondary outcomes included all-cause bleeding, and a composite of myocardial infarction, stroke, ischaemia driven revascularisation and cardiac death (MACCE).

Results 562 patients were included in the analysis with 20% treated with triple therapy (VKA 42%, Riv2.5 33%, Riv20 25%). There was no significant difference in age and presentation (male 70%; ACS 54%). Triple therapy patients tended to be older (74±14 years) than those treated with DAPT (70±11 years, p<0.002).

The primary outcome of BARC type 3–5 bleeding occurred in 3% of DAPT, 4% of VKA, 16% of Riv2.5, and 10% of Riv20 patients (Figure 1). BARC 3–5 bleeds were significantly higher in all patients on triple therapy with NOAC vs DAPT (p<0.01). Using a logistic regression model and adjusting for age, sex, diabetes and ACS, no predictor except anticoagulant use was associated with this increased bleeding. There was no significant difference in all-cause bleeding events between the triple therapy groups. MACCE was similar, however the DAPT group had a CVA event rate of 0.9%, with no events in the triple therapy groups.

Conclusion Triple therapy with NOAC increases major bleeding risk by up to four-fold compared with DAPT, this figure is slightly higher than found in the literature. Reassuringly, this increased bleeding risk does not translate into MACCE, however the perception of a reduced bleeding risk with a lower NOAC dose is not supported in this cohort.