First things first: I have no conflicts of interest regarding the content of this blog entry; I'm merely sharing my best interpretation of the published literature as I'm aware of it.
Rivaroxaban (Xarelto) has become the most widely used oral anticoagulant in America. Choosing one NOAC to primarily use is a good idea because one can become familiar with its dosing requirements and drug interactions and prescribe it well. This is especially true for primary care practitioners who are faced with a constant barage of new medications with which to become familiar.
But why choose Rivaroxaban?
There were three pivotal trials comparing warfarin to each of the three non-vitamin-K-dependent oral anticoagulants (NOACs) currently available in Canada: RE-LY for dabigatran, ARISTOTLE for apixaban, and ROCKET-AF for rivaroxaban. Each of those trials established the superiority of its respective NOAC over warfarin in either effectiveness (reducing stroke and systemic embolism) and/or safety (reducing major bleeding) except for one. That's right: In ROCKET-AF, rivaroxaban was merely "non-inferior" to warfarin.
It is difficult to compare the outcomes across these trials because the populations were all different, but consider this: ROCKET-AF included the highest risk patients with the highest rates of the primary outcome (stroke and systemic embolism), placing it in the best possible position to demonstrate statistical superiority in efficacy if it was present, yet it didn't. Not only that, warfarin-treated patients in ROCKET-AF had the lowest mean time in the therapeutic range (55% in ROCKET-AF vs approximately 62-64% in the others). That means that rivaroxaban was tested against warfarin performing at its worst among these three pivotal trials, and it still couldn't come out ahead even though the others did.*
While a head-to-head comparison of the NOACs will probably never be performed (the number of patients needed to do such a trial would be truly enormous), there is mounting observational community based data that permits head-to-head comparisons of just how these drugs are actually performing in the "real world". Three such studies (available here, here, and here) have been published so far, and compared to dabigatran and apixaban, rivaroxaban has repeatedly been associated with higher rates of major bleeding and/or stroke & systemic embolism (SSE). One must be careful in interpreting these studies because the assignment to each drug is not randomized. Investigators use propensity scores to match patients for covariates that influence treatment assignment, but it is unlikely that all relevant covariates are able to be accounted for. Accordingly, such information is to be considered hypothesis generating. Nevertheless, if one is to find one NOAC to prefer among the others, this data is enough to bolster the suspicion, based upon the pivotal trial results I described above, that rivaroxaban may not perform as well as its competitors.
There exists a ready and plausible explanation for the possible under-performance of rivaroxaban: its once-a-day (OD) dosing schedule. The half lives of rivaroxaban, apixaban and dabigatran are all around 12 hours, yet rivaroxaban is the only one available in Canada that has been tested in a once-daily (OD) dosing regimen. This has undoubtedly proven to be a marketing boon for Bayer, but only because most doctors over-estimate the compliance and adherence benefits of OD dosing versus twice-a-day dosing, which are small (please read the paper that Bayer researchers cite supporting their decision to use OD dosing). And besides, adherence among apixaban users is probably similar to adherence among rivaroxaban users.
What most prescribers aren't aware of when a drug with a half life of 12 hours is dosed OD are the consequences of missed doses and extra doses, which are subtherapeutic drug levels for prolonged periods of time (the equivalent of missing three consecutive BID doses of a competitor drug) or very high peak drug concentrations. I urge you to click on the link in the last sentence to view the paper, free-on-line, and study figures 1 & 2. It is well recognized that the safety and effectiveness of anticoagulants are closely related to the time spent in the therapeutic window, and so it may well be that by dosing rivaroxaban once-a-day, resulting in high peak concentrations, low trough concentrations, and markedly exaggerated swings when doses are accidentally missed or doubled, patients are actually being placed at unnecessary risk. And so it may well be that the consequences of the kinds of minor deviations from perfect compliance that most patients are prone to from time to time are greater for patients on rivaroxaban than its BID-dosed competitors.
I share this information with my patients so that they understand that I'm choosing to prescribe apixaban twice-a-day based on good evidence that it's both safer and more effective than warfarin. I explain to them that the twice-a-day dosing keeps them in the therapeutic window more effectively when the odd dose is missed or the odd extra one is taken. Pragmatically speaking, many patients with atrial fibrillation are already on other treatments that require them to take pills at least twice a day, so fitting apixaban into their medication schedule doesn't add to their frequency of pill-taking. Another pragmatic advantage of apixaban is that it is the least renally excreted NOAC currently available, so dose reductions are rarely required, and changes in renal function are less likely to cause bleeding. That also makes peri-procedural management of apixaban straight forward and easy to remember. Of course, if a patient refuses to consider anything other than a once-a day pill, then I prescribe rivaroxaban (or warfarin), and I'm careful to make sure that they know that they must take it with the largest meal of their day for proper absorption (a feature of rivaroxaban that complicates its implementation that is often overlooked by prescribers, in my experience).
Perhaps you disagree? If you do, please chime in in the comments section below. If I should be prescribing rivaroxaban more often, or apixaban less often, I'd certainly like to know.
*As if that wasn't enough, the warfarin assigned patients used a device to monitor their INRs at home, and Bayer and Johnson & Johnson (the makers of rivaroxaban) were aware that there were problems with this device underestimating the INR, potentially leading to more bleeding in patients on warfarin, yet they kept that information from the FDA.