Differentiation of Direct Oral Anticoagulants for Nonvalvular Atrial Fibrillation

Rajat Deo, MD, assistant professor of medicine at the University of Pennsylvania

There are currently 4 non-vitamin K antagonist anticoagulants available. These NOAC (novel anticoagulant) therapies include dabigatran, rivaroxaban, apixaban, and edoxaban. First, it’s important to recognize treatment with any of these agents in an appropriate patient when nonvalvular atrial fibrillation is effective at stroke prevention.

There certainly are small caveats that influence our prescription of one NOAC versus another. Some of those factors include an individual’s past history with side effects, including gastrointestinal (GI) side effects, bleeding side effects, bleeding complications from the past, an individual’s kidney function, and an individual’s likelihood to be compliant on a medication in the wake of side effects as well as timing of therapies. So, these are various components that certainly go into our decision making.

Another important variable in our decision making relates to cost. We recognize that patients have different insurance plans that can influence the choice of anticoagulant to be used. We take that into consideration when we prescribe any one of the NOAC therapies.

The choice of NOAC therapy is also influenced by several patient characteristics. We always take into account an individual’s past medical history, especially anything related to bleeding. If an individual has had a serious bleeding event in the past, we obviously have to weigh the risks and benefits of prescribing anticoagulation therapy. And certainly, depending on the type of bleed in the past, that might influence which particular NOAC therapy we prescribe. In addition, individuals (or atrial fibrillation patients) oftentimes have other medical comorbidities as well. Depending on, for example, whether an individual has GI effects, has had gastrointestinal bleeding, has chronic kidney disease, or is taking other medications, those factors may influence which particular NOAC therapy we prescribe.

In the current era, with 4 available NOAC therapies, clinicians are often wondering which one of the 4 agents is the best one to use. We have 4 large, randomized clinical trials that have enrolled nearly 60,000 to 80,000 patients in total, that have evaluated the safety and efficacy of each of these agents compared with warfarin. Overall, we believe that these are effective agents for stroke prevention and thromboembolic prevention.

However, we don’t have a prospectively-based randomized controlled trial that compares each of the 4 agents. As a result, clever clinical studies need to be designed where we attempt to identify signals that may suggest one agent is slightly better, in a certain respect, than another agent. I think we have to be very careful, however, when we interpret the findings of these studies. Real-world studies are clearly very important, but those real-world phase IV studies are primarily important to understanding the efficacy and safety of any particular NOAC therapy compared with a control, such as warfarin.

It’s very difficult to compare one NOAC to another NOAC using these phase IV study designs. The Graham Study evaluated nearly 120,000 Medicare beneficiaries using a propensity-based analysis. In other words, these individuals (these investigators) compared rivaroxaban to dabigatran in Medicare beneficiaries using a propensity-based analysis. A propensity-based matched analysis will attempt to control for the various comorbidities in the various groups. In other words, they are attempting to compare rivaroxaban to dabigatran in 2 separate Medicare populations (2 separate groups of Medicare patients), and they’re trying to control for the various comorbid factors.

These investigators, essentially, found that rivaroxaban use may be more harmful compared to dabigatran use. However, the real question that arises from this study is, how well was the propensity-based design able to control for the various comorbidities influencing each of these 2 Medicare groups.

I think this type of analysis is very challenging. And my own conclusions from this study are first, we should certainly be aware of these findings. We should not ignore these findings. But second, we need larger studies. We need other observational studies (other phase IV designs) where we can evaluate further whether rivaroxaban is safer or more harmful than dabigatran.