Live Blog– FDA Advisory Panel on Rivaroxaban (Xarelto) 7

Click here for a brief summary and commentary on the FDA panel hearing.

Panel Adjourns! Good night everyone.

4:58– Lincoff asks if anyone thinks the transitioning study should be done pre-approval? Everyone seems ok with a post-marketing study.

4:55– Nissen wants to see the drug tested in a BID regimen. Papdemetriou wants a post-marketing study on transitioning.

4:54: Last question:

9. If rivaroxaban were to be approved for stroke prevention in patients with atrial fibrillation, …

9.1. … are there any constraints you would place on the population in whom it would be indicated?

9.2. … are there any issues you would want to resolve post- marketing?

4:53– I missed this earlier: Fleming was the abstention.

4:50– Kaul recommends R be given a third-line indication. Nissen says it can be used in rare situations. I don’t think anyone has supported a superiority claim. McGuire makes the point that it can’t be claimed to be as effective as warfarin but that it is effective. Most panelists seem to go with this idea.

4:49– 10 minutes to answer the last 2 questions:

8. If you voted to approve rivaroxaban to prevent strokes in patients with

atrial fibrillation, does it merit …

8.1. … a superiority claim to warfarin?

8.2. … a claim as an effective alternative to warfarin?

8.3. … a claim as effective?

8.4. … a claim for patients failing other anticoagulant therapy? If so, what constitutes failure?

Folks– despite the lopsided vote this is NOT a strong endorsement of rivaroxaban. This committee has LOTS of concerns about R, but they want to see alternatives to warfarin and dabigatran.

4:46– Papademetrious (I’m glad he didn’t speak more often or my fingers would have fallen off!) voted yes. Kindzelski also voted yes.

4:45– Emerson is the other no vote. Cocnerned about documentation, dosing, creeping noninferiority, etc.

[Readers: would love to hear your feedback, either via email, comments, or twitter. Thanks!]

4:42– Coukell (consumer rep) voted yes, but nothing he’s heard  suggests that it should be a first line therapy. Debra McCall (patient rep) voted yes.

4:38– Krantz was a cautious yes. Lincoff was a yes, agreeing with McGuire. Lincoff thinks additional data needed. Sanjay Kaul voted yes (!) but not enthusiastic. Kaul is still concerned about what to do when you stop the treatment.

4:37– McGuire voted yes. Well-designed and executed study, a large pragmatic trial, though not perfect. Large, representative population and the results convincingly show the benefits of the drug. The concerns raised are legitimate academic issues but shouldn’t affect approval.

4:33– Nissen voted against approval. Sager voted for approval. Nissen concerned about the dose issue and we won’t find out what the right dose is if we approve it now. Also concerned about creeping noninferiority and all the other issues mentioned during the meeting.

4:30– 9 in favor of approval, 2 against, 1 abstention!

4:26: Voting Questions:

7. VOTE: Should rivaroxaban be approved for the reduction of stroke and non-CNS systemic embolism in patients with non-valvular atrial fibrillation?

4:23– McGuire points out we don’t know how to handle transitions from other drugs as well, including warfarin, heparin, low molecular weight heparin, etc. Nissen agrees but says that what makes this poignant is that there were a lot of bad events after people discontinued R at the end of the trial.

4:21– Lincoff said he would be amenable to the FDA’s suggestion about a smaller trial about to find out how to handle the transition issue.

4:20– Nissen: I don’t know how to transition the drug. We need to know what to do and we don’t have the information. In the absence of information physicians will use what their gut tells them to use. They may be right and they may be wrong. We just don’t know the answer.

4:16– One more question before the voting starts:

6. Are there adequate instructions for use with regard to …

6.1. … what regimen to use in most patients? If not, does this matter?

6.2. … what dose adjustments are needed in patients at extremes of exposure or risk? If not, does this matter?

6.3. … transitioning between rivaroxaban and other anticoagulant therapy? If not, does this matter?

6.4. … actions to take in the event of serious bleeding? If not, does this matter?

4:14– McGuire: “we don’t have to chase the tail of efficacy”–from a practical level you can’t change the comparator every time a new agent comes along. It’s good for patients to have different therapies available.

4:10– Fleming says that if dabigatran is superior to warfarin then that should be the comparator.

4:05– So far no new ground covered with this question.

4:02– Question 5:

5. If you conclude that the policy does apply and that rivaroxaban needs to be “as effective as” something, …

5.1. … what does “as effective” mean operationally?

5.2. … is it sufficient to be “as effective” as warfarin? If so, is it?

5.3. … is it necessary to be “as effective” as something else? If so, …

5.3.1. … do you need a direct comparison to the “something else”?

5.3.2. … is it?

4:02– McGuire says there’s no reason to consider dabigatran in their decision.

3:58– Kaul and Temple agree that this is a complicated question. It’s made far more complicated because dabigatran came on the scene after ROCKET was initiated. [Shocked that no one so far has used the “moving targets” phrase.]

3:55– Nissen: compliments them for studying this sicker population but there’s clearly a lot of overlap. Most panelists agree with the position that the ROCKET population was sicker but that they’re not fundamentally different and there’s a lot of overlap.

3:52– 6 more questions remaining. Here’s question 4:

As part of the Clinton administration “Reinventing Government” initiative, FDA published (Federal Register 1995 60(147):39180-1) a policy that said, in part,

“The agency does not require new human drug products or medical devices to be more effective than existing therapies nor does it necessarily require the product to be compared to other products. However, for products intended to treat life- threatening diseases, diseases with irreversible morbidity, and contagious diseases that pose serious health risks to others, it is essential for public health protection that a new therapy be as effective as existing, approved therapies.”

4. The “as effective” policy explicitly does not apply if the new therapy is studied in a new population. In considering how this exclusion might apply to rivaroxaban, here are some points for comparison of the warfarin arms in RE-LY and ROCKET-AF.

Is the population in ROCKET-AF sufficiently distinct from the population in RE-LY that the “as effective” policy does not apply? If so, how?


3:41– McGuire points out that TTR is not only a reflection of physician care. Many other factors related to the patients themselves, for instance.

3:40– Kaul says that you have to accept that TTR is a marker of quality of care.

3:39– Nissen: it is very common to interrupt therapy with warfarin. Some patients willdo this on their own. It’s slow-on, slow-off effect is very forgiving. But in the first few days after you stop R “pretty bad things happen.” But we don’t have a clear way to stop taking R and “we’ve not studied that.”

3:35– Krantz says R as good as warfarin as used in the US in the real world.

3:34– McGuire says the fact that  R at least comes close to warfarin is “impressive.”

3:33– Nissen says we have to judge the trial on the way it was conducted and not on the way it should have been conducted. Concerned about what happens when people stop taking the drug.

3:28– Fleming punts. Says he doesn’t know what the TTR data means.

3:25– Kaul says that R has not been shown to be as effective as warfarin when warfarin is well managed.

3:22– Nissen: “when warfarin is well-managed I’m not sure I know the answer… I worry about the slippery slope of noninferiority.”

3:20– Now they’re on question 3:

3. Please comment on effectiveness. How does rivaroxaban compare with warfarin …

3.1. … as used in ROCKET-AF?

3.2. … as used in the US?

3.3. … when it is well managed?

3:11– Superiority really seems to be off the table. Whether R is noninferior to warfarin will be the crux, I think.

3:09– Fleming says it’s difficult to understand importance of TTR. It doesn’t invalidate but it does compromise the results of ROCKET.

3:05– McGuire says we all wish the TTR would have been higher but “I’m completely unconvinced” that TTR reflects what it is supposed to measure. In other words, the TTR of 55% is a “worst case scenario” but probably represents the real world.

3:03– Nissen says you can’t guess what would have happened if TTRs had been as high in other trials and it’s unfair to try to torture the data to figure it out.

3:01– Kaul: clearly R is NOT superior to warfarin.

3:00– Emerson states that most panelists believe that the TTR issue is a reflection of poor use of warfarin, not a reflection of higher risk patients.

2:57– Now we’re on question 2.

2. The interpretation of a non-inferiority study depends upon certain understanding of the effect of the active control. If the active control is used to achieve less than its expected effect, a finding of non- inferiority may not be informative regarding the effectiveness of the study drug. Similarly, a finding of superiority to a suboptimally administered active control cannot be used to support superiority of the study drug. Please comment on the adequacy of the conduct of ROCKET-AF.

2.1. One measure of the quality of warfarin management, time in therapeutic range (TTR), was not as good in ROCKET-AF as in many recent randomized, controlled studies.

Was anticoagulation on warfarin in ROCKET-AF good enough so that the warfarin group is an appropriate comparator to show …

2.1.1. … effectiveness of rivaroxaban?

2.1.2. … superiority of rivaroxaban to warfarin?

2.2. Disposition of subjects in ROCKET-AF is summarized below:

Please comment on how the disposition data affect your ability to infer …

2.2.1. … effectiveness of rivaroxaban?

2.2.2. … superiority of rivaroxaban to warfarin?

  1. 2.3.  Was follow-up for end point events adequate in both treatment groups?
  2. 2.4.  Are there other aspects of study conduct that importantly affect interpretation of the study?

2:54– Lots of discussion again about ischemic and hemorrhagic strokes and double counting hemorrhagic strokes for both the efficacy and safety endpoints, but unlikely this will have any practical effect on the decision.

2:47– Now they’re talking about question 1.4:

1.4. Are there other aspects of study design that importantly affect interpretation of the study?

2:45– Panelists don’t seem really worked up about these late events.

2:38– Webcast back up. Panel discussing how long you need to follow patients after they discontinue study drug.

1.3. The primary analysis included events that occurred within 2 days of discontinuing study drug. For how many days should end point events that occurred after discontinuation of study drug—during the study or at its end—be counted?

2:32– Sorry, experiencing some technical problems connecting to the webcast. 

2:27– Nissen says they should have tested more than one regimen.

2:23– Next question:

Was it reasonable to test a single regimen of rivaroxaban in ROCKET-AF? Was the specific choice of regimen reasonable, given the short half-life and nonlinear kinetics of rivaroxaban?

2:22– Lincoff says the management in the trial was “reasonable,” and a “defensible” strategy.

2:20– Kaul complimentary on trial but says a centralized dosing strategy would have overcome the trial’s limitations.

2:18– Now starting discussion questions. First: Was the planned warfarin management strategy reasonable? Nissen is first: NO!

2:15– Lots of technical/statistical questions. Panel discussion starting. Now the fun should begin. 

1:44– Temple says that most of the effect of INR is on embolic stroke but most of the benefit of R is on hemorrhagic stroke, so the INR shouldn’t actually have a big impact on the efficacy of R.

1:42– Rose: I personally believe this drug would beat placebo in a placebo-controlled trial. Temple agrees that the standard is higher now that dabigatran is available.

1:33– Everyone (Kaul, Temple, Rose) agrees that noninferiority is in the mind of the beholder. Interesting discussion about this, but not sure if it means anything for rivaroxaban decision.

1:26– Kaul points out that in countries with low TTRs the event rate was actually pretty low, and he says that it’s event rates we should be concerned about.

1:22– Rose agrees that when compared to sites with high TTR rivaroxaban was noninferior to warfarin.

1:15– And we’re off again. Sanjay Kaul has 3 questions.

1:00 PM– Earlier today FDA reviewer Martin Rose said (and I reported) that J&J was no longer pursuing a superiority claim. I just received the following from J&J stating that they were not withdrawing the superiority claim:

“We are seeking a label that reflects the data showing superiority to warfarin on treatment based on the 21% risk reduction shown in the Rocket AF trial, which was our pre specified primary endpoint. That has not changed.

I don’t know if J&J is being delusional about this or if they know something the rest of us don’t. Right now it’s hard to imagine how it could get superiority, since they’re fighting for their life just to prove noninferiority.

LUNCH BREAK– Questions for sponsor will resume at 1:15

12:23– Rose: A lot of patients leave the study and have events– why is it worse in the R group? Rose speculates that R-treated patients may have been withdrawn after they ran into trouble. Very complex stuff here.

12:18– Rose: Current landscape of anticoagulant therapy is important. Now there is an alternative to warfarin. RE-LY demonstrated that dabigatran was robustly noninferior and that it is reasonable to ask the same of rivaroxaban. Rivaroxaban does not meet this test.

12:10– Now Rose takes issue with J&J’s plan to transition patients on R to warfarin. He finds lots of holes in their proposal. He leaves open that the transition problem could be solved with a PD study and not a clinical trial.

12:04– Rose agrees with J&J that the lower TTRs in ROCKET compared to other trials can be explained by geography. “We’re not quibbling with that.” “ROCKET enrolled a majority of its patients in countries where warfarin is not used well. That’s our major concern about this study.”

11:59– Rose: “You have to wonder if this drug is as good as warfarin when it [warfarin] is used well.”

Too bad statistics are really important, because they can really drive you crazy.

I’ll bet some of you could come up with some funny alternative expansions for the “TTR” acronym.

11:51– I can’t help wondering if Darren McGuire was right and that although there will be a whole lot of discussion today about TTR, in the end the committee will just conclude that TTR is just another surrogate endpoint that shouldn’t be taken too seriously.

11:49– Rose says you can’t trust the J&J calculation of TTR because they did it in a nonstandard way.

11:42– Rose responds to earlier comments by Califf that TTR was lower in ROCKET because it was double blind, noting that other trials, including ARISTOTLE, have been similarly blinded and had higher TTRs.

11:37– Adequacy of warfarin management: now Rose is showing data from the briefing document. I’m not going to summarize all this as it’s in the briefing documents.

11:35– Rose now making the point that the quality of warfarin management was NOT included in the noninferiority analyses of ROCKET.

11:33– now Martin Rose will discuss the interpretation of the efficacy data. J&J has agreed that the superiority claim is now “off the table” so he won’t spend a lot of time discussing it. (This is news.)

11:31– Dunmon: clinical pharmacology and VTE studies suggest BID dosing of rivaroxaban.

11:19– “Dabigatran has a longer half life than rivaroxaban and it is dosed BID.”

11:15– FDA presentation now beginning from Preston Dunmon, Clinical Reviewer. He will discuss the dose selection of R in ROCKET. This is important because there’s a lot of talk that J&J is ideally a twice-a-day drug, but the commercial appeal of once-daily dosing was too tempting.

11:13– Under questioning from Tom Fleming, Califf shows slides that in the ITT analysis in North America the primary endpoint point estimates go against rivaroxaban when you follow out to 30 days after the trial has ended. Califf notes that this data has never been shown for other trials or drugs.

11:04– Califf said J&J prepared 4,000 slides. I’ll be live-blogging every single one of them later this week. Stay tuned.

10:57– Panelist Darren McGuire predicts that TTR will “join the graveyard of failed surrogate endpoints.” (McGuire was a Duke cardiology fellow who studied under Califf and Mahaffey. It shouldn’t make any difference, but Darren is probably having fun grilling his old teachers.)

10:55– Panel not yet restarted. Just a quick thought. J&J did a great job deflecting the FDA review. Now the FDA will have to convincingly demonstrate that TTR makes a big difference.

BREAK. Panel resumes at 10:55

10:33– Temple goes back to the question about different types of strokes, noting that nearly all of the effect of R is on hemorrhagic stroke, also largely true for dabigatran and apixaban.

10:30– Califf notes that although the ROCKET population is higher risk than RE-LY, there’s still a lot of overlap so that the general indication for stroke prevention is valid. Califf also points out that too often trials are done in “namby pamby” populations, which makes it easier to do the trial, but may not apply to more complex populations.

10:23– Questions will run longer than scheduled. No public presentations scheduled for afternoon so there’s extra time.

10:12–  On twitter, @aliciaault makes a good summary of Califf’s point about INR: “Califf’s main point: that INR levels had to be managed given cultural, economic, social constraints.”

10:10– Nissen gets J&J to show slide that ischemic endpoint point estimate goes the wrong way (but still noninferior) in the complete ITT analysis through the end of the trial.

10:05– Califf giving Nissen grief about Nissen’s use of the term “third world,” noting that those third world countries now own most of the US debt, and says the term is offensive.

10:02– Nissen asks the second question. Asks Mahaffey if the ROCKET investigators educated centers about using warfarin well. Mahaffey says they educated the centers but did not provide them with a cookbook recipe about INR management.

10:00– Long discussion about whether hemorrhagic stroke counts as an efficacy endpoint or a safety endpoint.

9:55– Califf finished. Now the committee asks clarifying questions to the sponsor. Sanjay Kaul is up first (no surprise).

9:53– Califf criticizes FDA for comparing ROCKET to RE-LY and not mentioning that RE-LY was not double blinded. He says he doesn’t want to bash the FDA but clearly he’s upset at the criticism of the trial by the FDA briefing documents.

9:49– I’m really curious to see how the FDA will respond to Califf’s withering dissection of the importance of TTR.  Califf wins this years edition of “So You Think You Can Dance.” He can dance.

9:42– TTR in ROCKET in North America as good as other trials, especially when taking into account the severity of illness. Regional differences therefore account for difference in TTR. In the US TTR in ROCKET was actually better than common practice.

9:33– Califf uses a slide about TTR from panel member Fleming, but promises that they chose the slide before they knew he would be on the panel. Califf says TTR can be a “useful measure of quality improvement” but not a “useful surrogate.” I’m confused.

9:32– “I would be lying to you” if I said the executive committee wasn’t concerned when they first saw the TTR results.

9:30– In the US where TTR was high the treatment effect was better for rivaroxaban. Despite lower TTR rate in ROCKET, results are similar to RE-LY and ARISTOTLE.

9:26– Now we get to the heart of the issue: TTR. Califf says he won’t bash the FDA but watch out… We found no relationship between center TTR and efficacy. The major trials don’t show that TTR predicts efficacy.

9:22– Califf agrees there is a problem (but very small numbers) with patients discontinuing R at the end of the trial. He says the recommendation to maintain continuous anticoagulation and to keep patients on R until the INR on warfarin is acceptable, but he admits there is no clinical trial data to support this. Califf also notes that we don’t know what to do when discontinuing dabigatran. “We’re not unique in this regard in basing our discontinuation strategy on PK data.” We shouldn’t be held at a higher standard than other drugs.

9:19– Now Califf is addressing the problem of higher risk after patients discontinue R and transition to warfarin or other VKA (vitamin K antagonist).

9:15– Now Califf is entering the minefield of comparing different trials. He has to do this because the FDA review compared ROCKET unfavorably to other trials (especially RE-LY). First he’ll look at the transition off rivaroxaban at the end of treatment

9:13– It doesn’t strike me that Califf gave an especially strong defense of once a day. Sounds like he might have preferred twice a day, actually.

9:11– Now Califf is dealing with questions raised by FDA review. First, dose and regimen selected. (The FDA said that twice a day might have been better than once-a-day rivaroxaban.) Califf argues that at 24 hours most patients on one-a-day rivaroxaban are still anti coagulated, but “we’re not saying that it would be irrational” to dose it twice a day.

9:09–When patients are asked, death and disabling stroke are by far the most important outcomes. Bleeding complications less important. Rivaroxaban wins on death and disabling, but not on the “less important” endpoints. Califf argues that the overall risk-benefit calculation favors rivaroxaban over warfarin.

9:06– Califf notes that docs hate warfarin because of bleeding complications, patients hate it because it’s hard to take. ROCKET executive committee aware that there would be many problems regarding significant discontinuation of study drugs in the trial, since they were studying a high risk population with many concomitant morbidities.

9:02– Califf will now address the FDA’s criticism of ROCKET in the briefing documents. Now the fun should start. He says this will be one of the most difficult talks he’s had to give– and believe me, he’s given a lot of talks!

8:54– J&J’s Nessel presenting the safety results from ROCKET.

8:41– Rob Califf now talking. He will present the efficacy results of ROCKET, and will probably serve as the chief defender of ROCKET today. He’s a former member of the advisory committee and has worked closely with the FDA for many years. He has occasionally differed with Steve Nissen.

8:25– Duke’s Ken Mahaffey now speaking, talking about ROCKET study design.

8:22– J&J trying to downplay the importance of TTR (time in therapeutic range), which suggests that warfarin was not well used in ROCKET. But this may be a tough sell.

8:19– Sponsor presentation about to start. Gary Peters from J&J speaking.

8:15– Temple talking about noninferiority and the standards for proving that a new drug is as good as previous drugs. With the availability now of dabigatran the bar has been raised for rivaroxaban. Another major concern is what happened when people stopped taking rivaroxaban.

8:11– Unlike the press, Stockbridge doesn’t care how the committee actually votes. It’s the thought process that’s important to their actions. Temple says he cares “a little about the vote” but he’s also more interested in the thought process.

8:09– FDA’s Stockbridge and Temple starting off. Stockbridge: characterizes FDA’s internal discussion as fairly ambivalent, that the committee will have a big impact on the FDA’s decision. No one doubts that rivaroxaban is effective and there are no safety issues. The question is whether rivaroxaban is as good as warfarin and whether there are adequate instructions about how to use it.

8:05– Meeting is underway.

This is a live blog of the FDA panel on rivaroxaban (Xarelto).  Click here for background on this panel.