In the United States, there are currently four direct oral anticoagulants (DOACs). All four DOACs are approved for the treatment of venous thromboembolism (VTE) and nonvalvular atrial fibrillation (NVAF), among other indications.1-4 Despite differences in pharmacology, pharmacokinetics, and clinical trial efficacy and safety data, current guidelines do not prefer a specific DOAC. Given the lack of guideline-based recommendations for a particular DOAC, clinicians are frequently left without clear guidance of the most appropriate DOAC for a particular patient beyond the preferences of an insurance company or the availability of manufacturers’ coupons. After a careful analysis of the existing data, a very strong case can be made to make apixaban (Eliquis) the preferred DOAC for both VTE and NVAF.
What is the role of DOACs versus warfarin in VTE and NVAF?
For the treatment of venous thromboembolism (VTE), the CHEST 2016 guidelines recommend any of the four DOACs over warfarin therapy for long-term anticoagulation therapy in patients without cancer (grade 2B).5 The guidelines do not specifically endorse any DOAC, although a table is provided that outlines factors that may influence the selection of a “preferred” anticoagulant. For example, patients wanting to avoid parenteral therapy may prefer rivaroxaban or apixaban because dabigatran and edoxaban require five to ten days of parenteral therapy prior to initiation.
For the treatment of nonvalvular atrial fibrillation (NVAF), the 2014 AHA/ACC/HRS guidelines recommend warfarin (class IA) or a DOAC (class IB).6 The guidelines provide the same level of recommendation (IB) for dabigatran, rivaroxaban, and apixaban. At the time of publication, edoxaban (Savaysa) was not approved for use in the United States and was not mentioned within the guidelines.
Comparison of DOAC Efficacy and Safety in VTE Clinical Trials
To date, there are no trials directly comparing one DOAC to another for the treatment of VTE. Instead, all DOACs have been compared to warfarin in phase III trials. Although a direct DOAC comparison would be very desirable, it is unlikely that this type of clinical trial will come to fruition in the near future.
As a substitute for directly comparing DOACs, two recent indirect meta-analyses were conducted to evaluate the performance of each DOAC using warfarin as a consistent comparator.7, 8 Not surprisingly given the small number of clinical trials, both meta-analyses came to the same conclusion. All DOACs appear to be similar in efficacy with regards to the prevention of recurrent VTE or VTE-related mortality. With regards to safety, however, apixaban was shown to have the most desirable adverse effect profile with regards to bleeding. Compared to other DOACs, apixaban demonstrated a reduction in major or clinically relevant bleeding ranging from 53% less (versus rivaroxaban) to 31% less (versus dabigatran).
Given the available clinical trial data, apixaban has been shown to be equally effective in the treatment of VTE and superior in causing fewer bleeding events versus the other DOACs. With a more favorable risk:benefit ratio, one could argue that apixaban should be the preferred DOAC for the treatment of VTE.
Comparison of DOAC Efficacy and Safety in NVAF Clinical Trials
As with VTE, there are currently no trials directly comparing one DOAC to another for the treatment of NVAF. All phase III DOAC NVAF trials have used warfarin as a comparator, and it is similarly unlikely that a trial comparing two DOACs will be published in the near future.
A recent meta-analysis using indirect comparison analysis was conducted to evaluate the performance of each DOAC versus using warfarin as a consistent comparator among patients with NVAF.9 Among the DOAC doses that were eventually FDA approved, there was no difference in efficacy (stroke or systemic embolism) between any DOAC. With regards to safety, however, apixaban demonstrated a lower risk of major bleeding versus dabigatran and rivaroxaban, but was similar in major bleeding to edoxaban. Finally, apixaban was likely more “tolerable” than all other DOACs as reflected by a lower treatment discontinuation rate.
In consideration of the best available evidence, apixaban has been shown to be equally effective in preventing stroke or systemic embolism in patients with NVAF, is the most tolerable as measured by treatment discontinuation rates, and shares the lowest rate of major bleeding with edoxaban.
The Case against Dabigatran (Pradaxa) as the DOAC of Choice
Dabigatran has a concerning adverse effect profile unique to other DOACs that may impact its tolerability.1 In clinical trials, the rates of dyspepsia and GERD-like symptoms was more common than in patients taking warfarin. In addition, and more concerning, there may be a signal of a small but elevated risk of myocardial infarction events compared to warfarin in the VTE clinical trials (0.66 vs. 0.17 events per 100 patient-years).
Another complicating factor of dabigatran is its precautions regarding administration and storage.1 Dabigatran capsules must be swallowed whole (not broken, crushed, chewed, or emptied); doing so will increase overall drug exposure and potentially increase the risk for bleeding. Furthermore, dabigatran should be kept in its original bottle and it expires within four months of opening.
While dabigatran can be reversed with idarucizumab (Praxbind), there is a promising new reversal agent for the other DOACs on the horizon in the form of andexanet alfa (AndexXa).
The Case against Rivaroxaban (Xarelto) as the DOAC of Choice
Rivaroxaban has many qualities that make it a desirable DOAC agent.2 Unlike dabigatran, it does not have any unique or concerning adverse effects. Rivaroxaban’s unique quality, however, is that most doses (15 or 20 mg) must be taken with food. Taking rivaroxaban with food increases drug exposure by 39% and the maximum concentration by 76%, meaning that failure to take rivaroxaban appropriately can result in low serum concentrations of the drug and a risk of therapeutic failure.
The Case against Edoxaban (Savaysa) as the DOAC of Choice
At first glance, edoxaban has many advantageous traits.4 It is taken once daily without regards to meals and does not have any of the unique adverse effects seen with dabigatran. Further, it was the only DOAC shown to have similarly low major bleeding rates to apixaban in patients with NVAF.9
Unlike all other DOACs, however, edoxaban warrants serious concern regarding its boxed warning in patients with excellent renal function (creatinine clearance above 95 mL/min). In the trial for NVAF, edoxaban was less effective (associated with a higher incidence of stroke or systemic embolism) versus warfarin among the cohort of patients with excellent renal function. There is no mention or analysis of this cohort within the VTE clinical trials, but it stands to reason that similar precaution among this patient group would be prudent.
The Final Case for Apixaban (Eliquis) as the DOAC of Choice
Apixaban embodies nearly all of the qualities desirable in a DOAC for the treatment of either VTE or NVAF. As previously mentioned:
- Apixaban is just as effective as other DOACs for the treatment of VTE and NVAF
- In the treatment of VTE, apixaban likely has the lowest bleeding risk compared to all other DOACs
- In the treatment of NVAF, apixaban and edoxaban have the lowest bleeding risk compared to the other DOACs
- In the treatment of NVAF, apixaban demonstrated the lowest discontinuation rate (a surrogate for tolerability)
Furthermore, apixaban does not require five to ten days of parenteral anticoagulation for the treatment of VTE (unlike dabigatran and edoxaban).3 It can be crushed and does not have any unique administration or storage concerns. It does not have any unique adverse effects (unlike dabigatran) and does not have any specific concerns in patients with excellent renal function (unlike edoxaban).
The most significant downside to apixaban is that it requires twice daily dosing, whereas rivaroxaban and edoxaban tout the quality of once daily administration. While this may be less convenient, it seems like a small price to pay for so many other redeeming qualities.
Given the similar efficacy, favorable safety, greater tolerability, and a lack of unique administration or storage warnings, a compelling case can be made for selecting apixaban as the DOAC of choice for the treatment of VTE or NVAF.
- Pradaxa [package insert]. Ridgefield, CT: Boehringer Ingelheim Pharmaceuticals, Inc; 2015.
- Xarelto [package insert]. Titusville, NJ: Janssen Pharmaceuticals, Inc; 2015.
- Eliquis [package insert]. Princeton, NJ: Bristol-Myers Squibb Company; 2016.
- Savaysa [package insert]. Parsippany, NJ: Daiichi Sankyo, Inc; 2016.
- Kearon C, et al. Antithrombotic therapy for VTE disease: CHEST guideline and expert panel report. Chest. 2016 Feb;149(2):315-352. doi: 10.1016/j.chest.2015.11.026.
- January CT, Wann LS, Alpert JS, et al. 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol. 2014 Dec 2;64(21):e1-76. doi: 10.1016/j.jacc.2014.03.022.
- Cohen AT, Hamilton M, Mitchell SA, et al. Comparison of the Novel Oral Anticoagulants Apixaban, Dabigatran, Edoxaban, and Rivaroxaban in the Initial and Long-Term Treatment and Prevention of Venous Thromboembolism: Systematic Review and Network Meta-Analysis. PLoS One. 2015 Dec 30;10(12):e0144856. doi: 10.1371/journal.pone.0144856.
- Mantha S, Ansell J. Indirect comparison of dabigatran, rivaroxaban, apixaban and edoxaban for the treatment of acute venous thromboembolism. J Thromb Thrombolysis. 2015 Feb;39(2):155-65. doi: 10.1007/s11239-014-1102-5.
- Lip GY, Mitchell SA, Liu X, et al. Relative efficacy and safety of non-Vitamin K oral anticoagulants for non-valvular atrial fibrillation: Network meta-analysis comparing apixaban, dabigatran, rivaroxaban and edoxaban in three patient subgroups. Int J Cardiol. 2016 Feb 1;204:88-94. doi: 10.1016/j.ijcard.2015.11.084.
Thanks for the article; it’s very interesting. But who authored it? What is their affiliation? Where are their disclosures? It’s important to observe the same guidelines that print or scholarly journals observe to reveal conflicts of interest or bias
That’s fair! I wrote this post (Sean Kane, the author of ClinCalc.com). I have no affiliations or conflicts of interest with any entity relevant to the content of this post.
Great! May I suggest placing that disclosure statement in your “About” page and at the bottom of each article. (since the way that people will interface with / cite this article is the individual page or article in itself, rather than clicking around the website. Thanks for providing this valuable service!
Thanks fr the Article but you do not mention the article by David J Graham in which Rivaroxaban and Dabigatran are directly compared. What is your opinion about this study? Graham DJ, Reichman ME, Wernecke M, et al. Stroke, bleeding, and mortality risks in elderly Medicare beneficiaries treated with dabigatran or rivaroxaban for nonvalvular atrial fibrillation [published online October 3, 2016]. JAMA Intern Med. doi:10.1001/jamainternmed.2016.5954
This is a retrospective study — it’s going to be at risk for several biases that a prospective, randomized controlled trial is not at risk for. It’s absolutely interesting data to review, but I would find it hard to apply this directly to clinical practice given the issues with the trial design. Obviously this particular study does not address apixaban, but I could understand why one might use this data to favor dabigatran over rivaroxaban in the absence of other data (e.g., an RCT comparing the two).
Sean, I’m happy your site is all about evidence based medicine (EBM). Given that, your opinion is no more than that, an opinion. First of all, the AMPLIFY study excluded provoked DVTs and PEs as well as those associated with trauma/surgery. I guess you consider that meaningless – they must not occur in real life, therefore we don’t need any data to treat them. Also, since it’s been well studied and proven that compliance is better with once a day dosing rather than multiple doses, then possibly Rivaroxaban or Edoxaban should be chosen over apixaban regardless of concomitant food administration. And then you muddy the waters further by comparing bleeding rates in NVAF studies. You are correct that apixaban has low bleeding rates – they better, because they also had the lowest HAS-BLED scores. Let’s face it, in ARISTOTLE 1/3 of the patients had a CHADS 1 score – meaning they had very low risk of a stroke or a bleed. They could just as effectively been treated with aspirin instead of a OAC (per AHA/ACC/ESC/HRS). There is no way to compare bleeding rates between the ARISTOTLE, ROCKET AND HOKUSAI STUDIES. They had completely different pt populations with very different CHADS2 scores. EBM tells us that the higher the CHADS score, the higher the HAS-BLED score, and the higher the risk of bleeding. In fact, I see the comparable bleeding rates of Edoxaban and Apixaban in the HOKUSAI and ARISTOTLE studies respectively as being a win for Edoxaban over Apixaban given the higher CHADS score in the former study. Additionally, there’s Dabigatran’s reveral agent, idarucizumab, which has been clinically available for a year or more, yet you still believe Apixaban is superior because “there is a promising new reversal agent for the other DOACs on the horizon in the form of andexanet alfa (AndexXa)”. Until these reversal agents are FDA approved, they are non-existent. Interestingly, your blog is 9 months old and we’re still waiting for approval and availability of these agents.
Meta-analyses have a role in medicine, but only when you compare apples to apples and when you leave your opinions at the door. Let’s stick to the EBM.
I am disclosing that I am a speaker for Jansen Pharmaceuticals, yet I am not waving a Rivaroxaban flag here. I will be the first to state that the “best” DOAC for a patient is as individualized as is that patient. I see unique benefits in all of these agents. Until there is a head to head randomized prospective study, we will never be able to choose what the “best” DOAC is.
Thank you for the comments — I love the discussion! You’re absolutely correct that this is an opinion-based commentary. As mentioned in the article, we simply don’t have head-to-head trials between the DOACs for NVAF or VTE to definitively show whether one agent has a better efficacy or safety profile versus another.
AMPLIFY (with apixaban) did exclude provoked VTEs among patients who did not have a persistent risk factor for VTE recurrence. Given that this criterion excluded a patient group with a low risk of VTE recurrence (versus unprovoked VTE), I don’t view this as a major limitation. My guess is that this criterion was partially done to improve statistical power by selecting out patients who were more likely to have VTE recurrence.
You are correct that compliance is going to be better for a once daily medication versus a twice daily medication; however, the two major trials involving apixaban (AMPLIFY and ARISTOTLE) presumably capture the compliance issue as part of the primary endpoints. In other words, apixaban performed the way it did in these two trials with a potential reduced compliance versus a once-daily medication. I don’t think it’s fair to say that a BID medication is inferior to a daily medication simply because of its dosing regimen; this would be akin to always preferring metoprolol succinate over carvedilol in systolic heart failure simply due to the dosing regimen difference. As mentioned in this article, the dosing frequency is certainly a consideration and a great point to bring up with a patient during the shared decision making process when discussing anticoagulation options.
With regards to study comparison, you’re correct that you cannot compare stroke, VTE recurrence, or bleeding rate percentages between different studies. However, in the meta-analyses cited in the article, one must remember that all studies used warfarin as a consistent comparator. Without having head-to-head DOAC trials, this is the best alternative to have an idea of comparative efficacy and safety. By using the indirect comparator method in these meta-analyses, one observes that drug A is superior to warfarin, drug B is non-inferior to warfarin, therefore, without other data being available, drug A may be a preferred agent over drug B due to its comparative efficacy versus warfarin. By using a consistent comparator, differences in baseline characteristics (CHA2DS2VASC, HAS-BLED) become less meaningful.
I do think that the availability of idarucizumab is a win for dabigatran, but this benefit may be exaggerated to some extent. For most bleeding and urgent surgical procedures, I do believe that idarucizumab can impact patient outcomes. For intracranial hemorrhage, though, I am much more skeptical as to whether dabigatran reversal with idarucizumab actually changes patient outcomes versus no reversal (e.g., a change in modified Rankin scale). You are correct that andexanet alfa is not on the market. The manufacturer (Portola Pharmaceuticals) resubmitted its FDA application on August 3, 2017. I have no knowledge of a timeline for the FDA’s decision for this new application.
I agree that the meta-analyses have their flaws, but the indirect comparison method is better than examining the raw percentage rates between two trials (without using indirect comparison). Of course a head-to-head DOAC trial will provide the best quality of evidence, but I sincerely doubt that a drug company will take the risk to invest millions of dollars to potentially lose a fight against another DOAC. I do think that shared decision making with patients is important in general, but in particular with the decision regarding anticoagulation selection. Insurance coverage, cost, compliance with the regimen, reversibility, renal function, and a host of other factors should play a role in the process of selecting a DOAC. I too see value in all of the DOAC agents; however, I still contend that the best available evidence is most supportive of apixaban primarily from the perspective of its safety profile regarding bleeding events.
You mentioned that you prefer rivaroxaban despite your role as a speaker for Jansen — what factors have you used to make this decision? Aside from the BID versus daily dosing, I would be curious to know more about your perspective. Thank you again for your comments!
Not so fast re: once daily dosing and compliance. This data suggests greater cumulative pills taken with QD dosing (with specific meds, NOACs in this case) but it necessarily greater compliance nor outcomes. Patient behavior is a finicky thing.
Thanks for sharing! It’s always interesting when something you would intuitively believe to be correct (such as the superiority of once daily dosing) may not actually be valid in a real-world environment, hence the need for actually studying some of these clinical questions.
Sean, did you mention anywhere in your article about the statistically significant decrease in ischemic strokes seen with dabigatran vs. warfarin in RELY? And that no significantly decreased stroke rate was observed for either ARISTOTLE (for apixaban) or ROCKET AF (for rivaraxaban)? As a vascular (stroke) neurologist, this is an important decision-point for myself and my colleagues. Round with me in the Neuro-ICU and you’ll see just how devestating strokes can be — if just a few more can be prevented by a superior drug, in exchange for a bit more dyspepsia, then I’m an advocate. Also, it is now the end of 2017 and the FDA has yet to approve any reversal agent for rivaroxaban, edoxaban, or apixaban. What I tell my stroke/TIA patients is that Praxbind is like a spare tire in your car: you don’t want to have to use it, but if you need it, it’s there!
You’re absolutely correct that ischemic stroke was reduced with dabigatran 150 versus warfarin, which was not observed in other DOAC trials. Using the same logic, though, apixaban demonstrated a reduction in all-cause mortality whereas dabigatran did not. Of course ischemic strokes can be devastating, but so can major or clinically relevant bleeding events, which the network meta-analyses suggested is worse with dabigatran versus apixaban. Furthermore, the risk of MI associated with dabigatran seen in RELY isn’t something to be ignored.
I think it’s fantastic that Praxbind is on the market, and I do believe it offers an advantage toward dabigatran that the other DOACs do not have (for now). My concern is whether the presence of Praxbind is enough to favor dabigatran over other agents. For something like a GI bleed or emergent surgery, I think it’s a fantastic option and likely does change clinical outcomes. For patients with intracranial hemorrhage, however, I am skeptical whether it actually changes long-term outcomes or not. As was seen in the FAST trial (Mayer 2008) with rFVIIa in patients with a large ICH, reversal may prevent the bleeding from expanding, but the damage may have already been done and the administration of a reversal agent may not actually change long-term clinical outcomes.
Would cost per dose comparison be useful? Or would that vary too much from country to country?
Of course! At this point, insurance coverage, cash price, and even patient assistance programs through the manufacturers is a consideration, but is usually considered on a patient-by-patient basis. For example, one DOAC may be covered and have the lowest copay for patient A, whereas a different DOAC is the least expensive for patient B. Like any medication, if a patient can’t afford a DOAC they aren’t likely to take it!