Overview of the non-vitamin K oral anticoagulants for atrial fibrillation

For over half a century physicians had only one choice for oral anticoagulation, namely, vitamin K antagonists (VKAs), such as warfarin. Despite their use for a number of medical and post-surgical conditions, it was not until the completion of six randomized placebo-controlled trials of warfarin in 2900 patients with atrial fibrillation (AF) in the early 1990s that VKAs gained widespread acceptance for the prevention of thromboembolic stroke. ¹ The 67% reduction in ischemic stroke with warfarin compared with placebo outweighed the risk of hemorrhagic stroke or other serious bleeding events, such that even among patients with low risk of stroke but high risk of bleeding, the risk–benefit balance favored VKAs over placebo. ² Analyses demonstrated superiority of VKAs over aspirin ³ and thus guidelines in the first decade of this century firmly endorsed the use of VKAs for almost all patients with AF, except those at the very lowest risk of stroke (i.e. ‘lone’ AF).^4,5

Nevertheless, the well-known limitations of VKAs, including inter- and intra-patient variability in pharmacodynamic effect, multiple drug–drug and food–drug interaction, delayed onset and offset of action, genetic variability in metabolism, increased risk of intracranial hemorrhage (ICH) and fatal bleeding, and need for frequent monitoring and dose adjustment, led to the development and rapid introduction of non-VKA oral anticoagulants (NOACs) between 2009–2013. This themed issue of Therapeutic Advances in Cardiovascular Disease is dedicated to the NOACs in AF, summarizing the key findings from the large phase III clinical trials of the four currently available NOACs (dabigatran, rivaroxaban, apixaban, edoxaban) and the many important secondary papers that have followed.

A meta-analysis ⁶ of the four NOACs versus warfarin trials in over 70,000 patients demonstrated three clinical advantages of NOACs: (1) a 19% reduction in stroke or systemic embolism, (2) 10% reduction in mortality, and (3) >50% reduction in ICH (that was largely responsible for the decreased mortality). Furthermore, two of the NOACs (apixaban and edoxaban) significantly reduced major bleeding compared with warfarin, while dabigatran, 150 mg twice daily, was the only regimen to significantly reduced ischemic stroke.

These findings led to the class I recommendation in the 2016 European Society of Cardiology AF Guidelines for NOACs preferentially over VKAs in patients who are eligible for anticoagulation to prevent stroke. ⁷ I agree with this recommendation and in my practice the list of patient scenarios in which a NOAC is clearly the better option continues to lengthen, including:

Patients in whom a VKA is not in range at least 65% of the time

Patients who have a major event (stroke, major bleeding) while therapeutically on a VKA

Patients with complex or multiple drug–drug interactions with a VKA

Patient preference to avoid frequent monitoring and dose adjustment

Patient age ⩾75 years

Patients at a high risk of bleeding

Patients with mild-to-moderate renal insufficiency

(Note: for scenarios 5–7 above, apixaban and edoxaban are associated with significantly less bleeding than warfarin, while they retain their efficacy).

However, our patients are complex and it behooves practitioners to be familiar with the important differences between the NOACs and appreciation their limitations. I am often asked “which NOAC is best” as some clinicians would prefer to stick with only one agent for the majority of their patients with AF. That seemingly simple question is deceptive. While the NOACs have several features in common (rapid onset and offset of activity, predictable pharmacodynamics with oral dosing), there are critical distinctions that may help guide selection of the optimal NOAC for an individual patient.

Starting with the simple, two NOACs (rivaroxaban and edoxaban) are dosed once daily, while dabigatran and apixaban are dosed twice daily. For some patients having a once daily option is highly desirable and results in better compliance with therapy, ⁸ while for other patients who are taking several medications throughout the day, frequency of dosing is not so critical. The bioavailabilities of the factor Xa inhibitors are high (>50%), while dabigatran is formulated with tartaric acid (which can cause dyspepsia in approximately 1/4 of patients) to achieve 7% bioavailability. Perhaps the most important distinction among the NOACs is the variability in renal elimination, which varies widely from 80% for dabigatran to 27% for apixaban. Each of the NOACs requires dose reduction dependent upon renal function; however, the criteria for dose reduction differ across the four agents (and differs across approved labels around the globe). Each of the studies analyzed data according to renal function from their respective trials, and clinicians should carefully review the results as they consider which drug to select in patients with impaired renal function. An underappreciated difference across the NOACs is the variability in drug–drug interactions. Dabigatran and edoxaban have little (if any) metabolism via the cytochrome P-450 system, unlike apixaban and rivaroxaban. Depending upon the strength of the interaction, the number of concomitant therapies involved, and the patient’s renal function, a dose reduction or even complete avoidance of the latter two NOACs may be necessary. Protein binding is lowest for dabigatran, making it the only NOAC that can be cleared with dialysis. Other distinctions between the NOACs related to the individual study designs (e.g. risk profile of patients studied), outcomes in key subgroups (e.g. the elderly), and availability of reversal agents, are detailed in the four pieces that follow. Thus my typical response to the question above is that it is important to be knowledgeable about all the NOACs available on the formulary (preferably all four) so that you can pick the best drug for each of your patients. If the anticoagulant shelf in our pharmacy was a shoe closet, then for the past 50+ years we have had only one pair of worn, but reliable boots to choose from. Now, we have four additional colorful pairs of footwear we could also pick, and it behooves us to consider both the properties of these newer options, as well as the circumstances for which they are needed, when we are making our selection.

Before digging in more deeply to the data for the individual NOACs in this issue, it is important to consider scenarios when a NOAC should not be used. Absolute contraindications include patients with mechanical valves (the only available data with dabigatran demonstrated more bleeding and more thrombosis compared with warfarin) ⁹ and patients with moderate or severe mitral stenosis (not studied in the phase III trials). There are very few data in patients with severe renal insufficiency (despite the United States Food and Drug Administration’s labeling that permits use of apixaban in patients on stable hemodialysis) and no prospective data in patients with advanced liver disease. Data in pregnant and lactating women are lacking, and ongoing trials are investigating dosing in children. Some have raised concerns regarding dosing in patients at extremes of body weight (e.g. <50 kg or >150 kg) ¹⁰ or physiology (e.g. age > 90 years), and more data would be helpful to establish risk, benefit, and optimal dosing in these patients. Since a readily available and reliable measure of the anticoagulant effect of a VKA exists, a VKA would be preferable to a NOAC in patients with impaired gastrointestinal absorption (e.g. short bowel syndrome) or history of treatment noncompliance.

Looking to the future, ongoing investigations should help further refine the optimal use of NOACs. Among the major outstanding questions, the following remain on my top 10 list:

How should we incorporate measures of drug concentration or anticoagulation levels in clinical practice, if at all?

Which drug and what dose when dual antiplatelets are needed (e.g. in patients with acute coronary syndromes or post-stenting)

How can the risk of gastrointestinal bleeding be minimized?

What is the optimal way to manage life-threatening bleeding and what is the role for specific antidotes?

How do we compare the different NOACs in the absence of head-to-head studies?

Would more flexible dosing improve safety: efficacy (e.g. in patients who have had a major bleeding event, a recent ischemic stroke, or are at extremes of age, weight)?

Are NOACs ‘better’ if patients are well-controlled on a VKA?

How should we manage NOACs around the time of invasive procedures to minimize the risk of periprocedural bleeding and thrombosis?

Is the fast offset a problem if patients are not compliant?

How we do best evaluate ‘cost-effectiveness’?

With the above in mind, I hope that you find the four manuscripts in this issue, each written by experts in the field with indepth knowledge and experience with the specific NOAC reviewed, to be a useful summary of the most relevant clinical findings to date in patients with AF.