October 8th, 2012
Triple Antithrombotic Therapy: What — and When — Is the Bleeding Risk?
In a national registry study, researchers in Denmark studied the effect of multiple antithrombotic drugs — including triple therapy with warfarin, aspirin, and clopidogrel — in patients with atrial fibrillation and coronary artery disease (indicated by hospitalization for PCI or MI).
Among more than11,000 patients, 30-day bleeding rates were:
- Warfarin + ASA + clopidogrel (triple therapy): 22.6 per 100 patient-years
- Warfarin + single antiplatelet: 20.3 per 100 patient-years
- ASA + clopidogrel: 14.3 per 100 patient-years.
The increased bleeding risk observed with triple therapy persisted for up to 1 year after the initiation of therapy.
Ryan: Dr Lamberts, congratulations on this wonderful paper. Can you give our readers an idea of why you believed it was important to do this study?
Lamberts: Thank you very much. In everyday practice, we face the problem of bleeding complications when balancing the addition of antiplatelet(s) on top of oral anticoagulation therapy in AF patients who suffer MI or undergo PCI. We know that antiplatelet therapy is indicated for CAD and oral anticoagulant therapy for AF, and current guidelines support an additive approach, although no firm evidence of the safety of triple therapy (ASA, clopidogrel, oral anticoagulant) exists. Our perception that short-term triple treatment is safe is based on expert opinion, which, if unfounded, could potentially put patients at even greater hazard.
Ryan: The mean CHADS2 score in this study was 1.5 (SD, 1.5). I think it is important for people to recognize that these results were obtained in a well-treated cohort that is not necessarily at high risk for bleeding. Similarly, from what I understand, you did not observe any significant differences in risk between various subgroups. Can you expand on that for our readers?
Lamberts: We looked at several clinically relevant subgroups for any differences in our main finding of early serious bleeding with triple therapy use. Regardless of sex, age, CHADS2 score (≥2 or ≤1), or MI or PCI status at inclusion, the relative risk for bleeding events was increased, particularly within the first 90 days after the index MI or PCI.
Ryan: At inclusion, 13% of patients were on triple therapy, whereas by day 180, only 8% were. Do you have a sense as to why treatment was altered in almost 40% of patients? And also, which agent was typically stopped when physicians took people off triple therapy?
Lamberts: We defined antithrombotic treatment from ongoing medication, which allowed us to continually update exposure status based on actual redeemed prescriptions; thus, patients could change exposure status during the 1-year follow-up period. Current guidelines arbitrarily recommend triple therapy for 1 to 12 months, based on the severity of CAD presentation, type of stent implanted, and presumed bleeding risk. We also know from several studies that antithrombotic treatment in real-life patients after MI or PCI is fairly consistent and adheres to recommendations, but antithrombotic treatment in AF patients is highly variable. In this light I am not surprised that the proportion of triple therapy users declined. We did not specifically look at which agent was stopped, but it seems to be the case that if an oral anticoagulant is prescribed, this treatment is continued.
Ryan: It is noteworthy that the combined endpoint of cardiovascular death, MI, and ischemic stroke was similar in patients on triple therapy and in those on warfarin plus a single antiplatelet agent. Did this surprise you, and what recommendations can you make based on that data?
Lamberts: Although the study was designed to evaluate safety issues, it somewhat surprised us that we could not find a benefit for the combined thrombosis outcome with triple therapy. With regard to recommendations, I think we need further evaluation of specific treatments and more details on outcomes (i.e., ischemic stroke, death, stent thrombosis, and recurrent MI), before the data can support advice against use of triple therapy. Nevertheless, I believe these data raise important questions for patients with dual indications for antithrombotic therapy. When the patient is already anticoagulated, do we really need two drugs to inhibit platelet aggregation? And, are there differences between the available antiplatelet agents?
Ryan: So, let me pose a case: If I have a 71 year old man with AF and diabetes mellitus who needs PCI for critical LAD disease, should I be thinking about placing a bare-metal stent so I can discontinue his triple therapy as soon as possible?
Lamberts: This is a very difficult case; interventionalists especially know this problem well. I believe that in this setting, a BMS should be preferred, which also accords with current American and European guidelines. The reasoning, as you say, is that the risk for stent thrombosis is lower with a BMS compared with a drug-eluting stent; consequently, an earlier removal of an antiplatelet agent is possible. Although our study suggests that even 30 days of triple therapy is associated with increased bleeding risk, at this point, we cannot firmly say that an oral anticoagulant and a single antiplatelet are sufficient to protect from recurrent coronary events in the initial vulnerable period. Until solid evidence is available, treating physicians must assess bleeding and thrombosis risks individually. They should also be aware that our data suggests that triple therapy does not have a safe therapeutic window.