The Rivaroxaban FDA Panel and Large Pragmatic Trials

In case you don’t want to read my entire blow-by-blow account of yesterday’s FDA advisory panel on rivaroxaban (which is only slightly shorter than Gibbon’s Decline & Fall of the Roman Empire) here’s the brief summary I wrote for CardioExchange:

The FDA Cardiovascular and Renal Drugs Advisory Committee voted 9-2 (with 1 abstention) in support of rivaroxaban (Xarelto, Johnson & Johnson) for stroke prevention in patients with atrial fibrillation, but the vote did not represent a ringing endorsement of the drug. The committee wrestled throughout the day with the numerous concerns raised by FDA reviewers, who had recommended that the drug be rejected. The committee left three key issues for the FDA to resolve if it approves the drug:

  • Although rivaroxaban was noninferior to warfarin in the pivotal ROCKET trial, warfarin was not used as skillfully in ROCKET as in RE-LY and other recent trials, thereby making it difficult to assess the true efficacy of rivaroxaban.
  • Because of a high number of events observed in patients when they discontinued use of rivaroxaban, there is no firm guidance for how to transition patients to warfarin or other drugs when rivaroxaban is discontinued.
  • The optimal dose of rivaroxaban is still unclear. Many on the committee thought rivaroxaban might perform better as a twice-a-day drug.

Just a few thoughts about the panel:

In the end, after an impressive display of intellectual firepower, the committee relied on a few, fairly basic principles. For all the issues raised by the FDA reviewers, the committee agreed with panel member Darren McGuire, who said that ROCKET was “a large pragmatic trial.” Though far from “perfect,” he said ROCKET was well-designed and well-executed.

McGuire’s practical perspective was evident throughout the day. For instance, there was a great deal of discussion and anguish about the lack of knowledge about how to transition patients off rivaroxaban. McGuire helped deflate the issue by sensibly pointing out that we don’t know much about how to handle transitions off any of the other commonly used anticoagulants. (But McGuire didn’t respond to Nissen, who agreed with McGuire’s point but also noted that this topic may be more problematic with rivaroxaban due to the particular danger signal seen in people discontinuing rivaroxaban after the end of  ROCKET.)

McGuire also offered a pragmatic perspective about the TTR (time in the therapeutic range) issue, which was probably the biggest red flag raised by the FDA reviewers. (In short, TTR is a measure of how well warfarin is used. Because TTR times in ROCKET were shorter than TTR times in other trials, the FDA reviewers claimed that warfarin was not used skillfully in the ROCKET trial, and that therefore it is impossible to accurately assess the true efficacy of rivaroxaban based on a comparison with warfarin as it was used in ROCKET.)

McGuire pointed out that the TTR in ROCKET is probably a fair reflection of TTR in most real-world settings. He also noted that TTR is not completely under the control of health care providers and that cultural and other factors can affect TTR. More importantly, McGuire cast doubt on the importance of TTR itself, saying that although he wished the TTR had been higher in ROCKET he was “completely unconvinced” that TTR actually measures the quality of warfarin therapy. Moreover, early on in the proceedings McGuire predicted that TTR will “join the graveyard of failed surrogate endpoints.”

McGuire’s perspective shouldn’t be a surprise. He did his cardiology fellowship at Duke  and spent a lot of time at the Duke Clinical Research Institute and with Rob Califf, its director, who was the point man for rivaroxaban yesterday. To reject ROCKET would have required him to reject the work of his teachers and much of his education and philosophy.

On the other hand, yesterday’s panel shouldn’t be construed as a wholesale endorsement of the large, pragmatic trial philosophy. Although the committee vote favored approval of rivaroxaban for the indication of stroke prevention in AF patients, the label and other FDA requirements are likely to limit the widespread adoption of rivaroxaban. Big pragmatic clinical trials will still be around, but it seems clear they will be undergoing increasing scrutiny and critical attention.

Full disclosure: Darren and I collaborated on several joint projects when he was the chief cardiology fellow at the DCRI and I was the editor of TheHeart.Org.


  1. Large simple randomized trials.

    I loved your coverage, Larry, and agree with your points and those of Dr McGuire. I hope he reads your column.

    I read the FDA statistical reports and do not really see what all the fuss is about. ARISTOTLE was a well-run trial with a TTR far above of what we normally see in practice. Very few patients are blessed to be followed in an oral anticoagulation clinic, and consequently their TTRs are much lower. Ample evidence has documented that patients with new stroke and known atrial fibrillation will often present with INR’s well below 2.0 – papers by Gladstone and Hylek have shown this repeatedly.

    As an aside, if rivaroxaban is approved, I may prescribe it as a half tablet BID (unless of course it comes as a formulated capsule, in which case I would stick with the once daily preparation, or go with BID dabigatran). I agree with the FDA reviewers and Nissen that the drug should really have been dosed twice daily, given its short half-life and pharmacodynamics.


  1. […] the recent FDA advisory committee panel about the ROCKET AF trial many panel members expressed concern about the once daily dosage of rivaroxaban used in ROCKET AF. […]

Speak Your Mind