Key design elements of the Xarelto pivotal trial in atrial fibrillation have come back to haunt sponsors Bayer and Johnson & Johnson during the course of the rivaroxaban NDA review.

Decisions to study a geographically diverse group of high-risk patients, and to allow clinical sites to manage warfarin dosing according to local practice rather than following a standard protocol, contributed to what FDA has determined was suboptimal use of the vitamin K antagonist in the trial’s control arm, according to briefing documents for – and discussion at – the Cardiovascular and Renal Drugs Advisory Committee’s Sept. 8 review of an atrial fibrillation claim.

The agency also challenged the sponsors’ decision to study only the 20 mg once-daily dosing regimen in the ROCKET AF study. A single daily dose would give the Factor Xa inhibitor a leg-up over current and future competitors in the novel oral anticoagulant market, but FDA believes a twice-a-day regimen could offer efficacy and safety advantages.

If FDA approves rivaroxaban for atrial fibrillation, look for a post-marketing study requirement or commitment on twice-daily dosing.

The advisory committee voted 9-2, with one abstention, in favor of approving rivaroxaban for preventing stroke and systemic embolism in atrial fibrillation patients. However, FDA’s lingering concerns about how well warfarin was used in the pivotal trial still could pose an approval hurdle, and the agency’s desire for data on twice-daily dosing could make its way into a post-marketing study requirement or commitment.

Targeting A High Risk Population …

At the committee meeting, rivaroxaban’s sponsors were forced to defend the study population, design and dosing regimen employed in ROCKET AF.

The 14,000-patient blinded trial enrolled patients with two or more risk factors. This represented a higher risk population than was studied in some other recent trials of novel anticoagulants. These include the RE-LY study of Boehringer Ingelheim GmbH’s direct thrombin inhibitor Pradaxa (dabigatran), which was approved in October 2010. Warfarin use in RE-LY was open-label, and the study enrolled atrial fibrillation patients with at least one additional risk factor for stroke. Similarly, ARISTOTLE, the blinded pivotal trial for Bristol-Myers Squibb/Pfizer’s investigational Factor Xa inhibitor apixaban, also enrolled patients with at least one additional risk factor (Also see "Apixaban Shows Mortality Advantage, But Will It Survive FDA Scrutiny?" - Pink Sheet, 5 September, 2011.).

When compared to RE-LY, a greater percentage of patients in ROCKET AF had higher CHADS scores, meaning that they were at a greater risk for stroke. FDA’s analysis showed that 87% of patients in ROCKET AF had a CHADS score above 2, compared to only 32% in RE-LY. More than half of patients in ROCKET AF had a history of prior stroke, transient ischemic attack or systemic embolism, compared with 21% in RE-LY.

However, briefing documents revealed that FDA did not agree with the sponsors’ plans to study this higher-risk population because the patients were much sicker than had been studied in historical trials of stroke prevention in atrial fibrillation, which compared warfarin’s efficacy against placebo. “There was sparse evidence for the safety and efficacy of warfarin in preventing strokes and systemic embolism in this much sicker population,” the agency’s clinical review states.

The higher-risk nature of the study population, combined with its greater geographic diversity, proved problematic for the sponsors.

Rivaroxaban demonstrated non-inferiority to warfarin on the primary analysis, but failed to show superiority under FDA’s preferred intention-to-treat analysis. Agency reviewers asserted the suboptimal use of warfarin in the study’s control arm made it difficult to discern rivaroxaban’s true efficacy compared to warfarin when the latter is well managed (Also see "Xarelto Faces Approval, Labeling Obstacles Even After Cmte. Endorsement" - Pink Sheet, 12 September, 2011.).

The mean time in therapeutic range (TTR) for International Normalized Ratio in the warfarin arm was 55%, compared with 64% in RE-LY.

Speaking on behalf of the sponsors, Robert Califf, Duke University, asserted the TTR in ROCKET AF was consistent with other recent studies considering that more complex patients have a lower TTR and there are well-recognized regional variations in INR control.

“Our results, taking into account the severity of illness factors, … we think, are directly comparable to the results that were seen in RE-LY,” Califf said. “If you take into account the severity of illness, we believe we did as well or perhaps even better than the other trials in terms of management of anticoagulation.”

… Rather Than “Namby-Pamby” Patients

FDA reviewers’ criticisms notwithstanding, Califf defended the decision to enroll a high-risk population, even though the sponsors are pursuing a broad claim across atrial fibrillation patients.

“We routinely do trials in namby-pamby populations of low-risk patients for a lot of reasons – it’s easier to do the studies – and then we extrapolate to the 80-year old person on 15 other medications,” Califf said. “We think it’s time to begin to think about doing trials in the more sick people and having enough overlap that you can talk about the general population. We’re somewhat responding to what practitioners are saying, which is you’re doing these trials, they’re not relevant to the patients that are the biggest concern to us.”

Furthermore, geographic and cultural differences in warfarin management and INR control help explain much of the TTR variation seen in ROCKET AF compared with other studies, Califf said. Country was the strongest predictor of TTR in the study, and when the TTR results are analyzed by region, the ROCKET AF and RE-LY data are similar, he said.

FDA Cardiovascular and Renal Products Division Clinical Reviewer Martin Rose said he did not disagree with Califf’s assertion that the greater geographic diversity of patients in ROCKET AF was a key reason behind the lower TTR, but this does not alleviate the agency’s concerns about the efficacy data.

While 19% of ROCKET AF patients came from North America, where mean TTR was the highest at 64.1%, almost 39% of patients were enrolled in Eastern Europe, the region of the lowest TTR, 49.7%. In contrast Eastern Europe accounted for only 12% of the subjects in RE-LY, Rose said.

“When the sponsor says the results of this study are related to geography and cultural patient factors, they’re absolutely right. We’re not quibbling with that at all,” Rose said. “When you look at it in an operational sense, what these numbers mean is that the ROCKET study enrolled the majority of its patients from regions where physicians don’t use warfarin very well. They did a warfarin-controlled study, compared themselves to warfarin in a study where most of the patients were enrolled in sites where warfarin was not used well. That’s our major concern with this study.”

Lack Of Dosing Algorithm: Strength Or Flaw?

Committee members generally did not take issue with the decision to enroll a high-risk population, and some panelists commended the companies for studying sicker patients. The committee concluded the study population was not sufficiently distinct as to exempt rivaroxaban from an FDA policy requirement that the drug be proven to be “as effective” as existing treatments (see related story, Also see "Xarelto Review Puts FDA’s Comparative Effectiveness Stance To The Test" - Pink Sheet, 19 September, 2011.).

However, panelists had concerns with another aspect of the trial design – the absence of a warfarin dosing algorithm – particularly given the low TTRs seen in some geographic regions. Clinical investigators were not given a standard protocol for managing INRs that were out of the 2.0-3.0 target range. Instead, each site managed INR according to local standards.

Committee members disagreed on whether the lack of a centralized warfarin dosing algorithm was a good strategy.

While some panelists said they appreciated that the absence of a dosing algorithm better reflected how warfarin is used in the real world, others viewed this as a flaw that contributed to the sometimes poor warfarin management seen in the study.

“If there was one imperfection in this trial, it was the lack of a standardized algorithm for maintenance warfarin dose adjustment,” committee member Sanjay Kaul, Cedars-Sinai Medical Center, said. A centralized dosing strategy would have overcome the mediocre TTR achieved in the study, he asserted.

Disagreeing Over The Dose

In addition to concerns about study population, FDA reviewers had qualms about the dose studied in ROCKET AF.

Given that rivaroxaban has not demonstrated superiority to warfarin under an ITT analysis and has comparable bleeding safety, the drug’s once-daily dosing has been seen as perhaps the sole marketing advantage over dabigatran and apixaban, both of which are dosed twice daily.

FDA briefing documents reveal that at the time the study was being designed in 2006, agency staff did not believe the sponsors had provided sufficient justification for pursuing only the 20 mg once-a-day dose in the Phase III trial.

The sponsors asserted that two Phase II, dose-ranging studies in deep vein thrombosis supported the dosing regimen used in ROCKET. These data supported either once- or twice-daily dosing, but the one-a-day regimen was clinically preferred and would encourage better adherence, the sponsors said.

However, FDA Clinical Reviewer Preston Dunnmon said rivaroxaban’s clinical pharmacology suggests twice-daily dosing might have efficacy and safety advantages over a once-daily regimen at the same total daily dose. Rivaroxaban has a 5-9 hour half-life in healthy individuals, and there is lower prothombin time fluctuation with the twice-daily dose than with the once-daily regimen, he said.

The 20 mg used in ROCKET AF was twice the dose approved for preventing DVT following hip or knee replacement surgery, and the sponsors’ ongoing Phase III clinical trial in acute coronary syndromes has incorporated lower and split dosing regimens of 2.5 mg bid and 5 mg bid, Dunnmon pointed out.

While most panel members agreed it was reasonable to test the 20 mg once-daily dose based upon the information available at the time ROCKET AF was designed, they said they would have liked to see data on a twice-daily regimen.

Temporary voting member Steven Nissen, Cleveland Clinic, disagreed on the choice of dose. “My concern was that the dose was selected more for marketing advantage rather than for the scientific data that was available, and that was a mistake,” Nissen said. “I think it was not the best strategy and I think it may have contributed to the less than robust results.”

By Sue Sutter

ROCKET AF’s Design Causes Headaches For Xarelto Sponsors

More from Archive

More from Pink Sheet