FDA Review of Data From the RE-LY Trial on September 20th, 2010
Editor-In-Chief: C. Michael Gibson, M.S., M.D. [1]
Design of RE-LY
RE-LY (Randomized Evaluation of Long-Term Anticoagulant Therapy) was a 18,113 patient randomized, non-inferiority trial comparing unblinded warfarin to blinded dosing of dabigatran (either a dose of 110 mg or 150 mg twice a day) [1]. RE-LY patients included those patients with non-valvular atrial fibrillation. Both patients who were warfarin naive and non-naive were enrolled. The primary endpoint was a composite of stroke and systemic embolism. The median duration of follow-up was 2 years.
FDA Commentary on the Design of RE-LY
An expert external panel was convened by the FDA to review data supporting the clinical use of Pradaxa in the management of patients with non-valvular atrial fibrillation. What follows are areas or questions of active interest on the part of the FDA.
Blinding
The trial was unblinded. The FDA has indicated that it could be difficult if not impossible to garner a superiority claim in the face of an unblinded trial design. Events that are submitted to an independent clinical event committee should be blinded to safeguard the adjudication process. The FDA, however, found that in 20% of patients, there was data in the source documents that could have unblinded the CEC.
Follow-up
Patients who dropped out of the study could be followed up by either telephone contact or in person in clinic. The FDA expressed concern that if a patient is evaluated for a follow-up visit in clinic this may yield a more complete ascertainment of endpoints than if the patient follow-up is conducted by phone. Similar rates of phone follow-up were observed for the three strategies (8%).
Drop outs and Drug Discontinuation
The FDA expressed concern if there are differential rates of drop out in the comparator arms. 96% of subjects completed the trial. 19% of dabigatran patients and 15% of warfarin patients discontinued therapy.
Event Rates for Warfarin
One question the FDA had was whether Warfarin was administered in the optimal fashion (a sufficient time in therapeutic range (TTR) is maintained) and whether the event rates were similar to those of patients enrolled in prior RCTs. The event rates for the warfarin control arm were similar to those of other RCTs.
Time in Therapeutic Range
Time in therapeutic range (TTR) of 64% is not too dissimilar to other RCTs. TTR in RELY is similar to that in SPORTIF III and V. The TTR rate may depend upon frequency of TTR monitoring. A high TTR may reflect a low frequency of INR monitoring, and the TTR must placed in the context of the frequency of INR monitoring.
Quality of Data and Adjudication
The rate of concordance between investigator and CEC reported events was 50% to 85%. In 20% of cases, the CEC could have been unblinded by the source documents.
Efficacy Results
Efficacy Results Published in the New England Journal of Medicine
Among patients treated with warfarin, the primary endpoint event rate was 1.69% per year. Among patients treated with 110 mg of dabigatran, the rate was 1.53% per year(relative risk with dabigatran, 0.91; 95% confidence interval [CI], 0.74 to 1.11; P<0.001 for noninferiority), and among patients treated with 150 mg of dabigatran the rate was 1.11% per year (relative risk, 0.66; 95% CI, 0.53 to 0.82; P<0.001 for superiority).
FDA Interpretation of Efficacy Results
The FDA viewed the data as showing that both doses of dabigatran were non-inferior to warfarin and were within the pre-specified hazard ratio margin of 1.38. The 150 mg dose of dabigatran was actually superior to warfarin in the prevention of the primary endpoint. One FDA reviewer was not certain that dabigatran achieved superiority given the 1) open label nature of the study; 2) given the adequate but moderate time in a therapeutic range for warfarin use; and 3) given that this is a single study without replication. A p-value of 0.00125 is often required to establish superiority in a single trial.
There was an increased risk of clinically manifest MI associated with Dabigatran dosing. Although this was of concern, it was inconsistent with the observation that there was no increased risk of revascularization associated with dabigatran. There was no data available regarding study drug discontinuation in the day or days prior to the occurrence of an MI that might link rebound thrombin generation associated with drug discontinuation to an increased risk of MI. There was numerically a higher risk of major bleeds the week before an MI in the dabigatran group, although this number did not reach statistical significance. The excess number of MIs in the dabigatran group was observed months after drug discontinuation, raising questions as to the biologic plausability of the relationship of MI to study drug.
Bleeding Results
Bleeding Results from RE-LY Reported in the New England Journal of Medicine
Among patients treated with Warfarin, the rate of major bleeding was 3.36% per year versus 2.71% per year in the group receiving dabigatran at a dose of 110 mg (P = 0.003) versus 3.11% per year among patients treated with 150 mg of dabigatran (P = 0.31). Warfarin was associated with a 0.38% per year risk of hemorrhagic stroke versus 0.12% per year among patients treated with 110 mg of dabigatran (P<0.001) versus 0.10% per year among patients treated with 150 mg of dabigatran (P<0.001).
FDA Interpretation of Bleeding Results from RE-LY
There was a dose response curve for bleeding associated with dabigatran such that the 150 mg dose was associated with a higher rate of bleeding compared with 110 mg. 150 mg of dabigatran was associated with similar to increased risks of bleeding compared to warfarin while the 110 mg dose was associated with less bleeding than warfarin. In general, the more severe the type of bleeding examined, the lower the risk with dabigatran. While dyspepsia was increased with dabigatran, this was not associated with an increased risk of bleeding. A greater time in TTR was associated with no excess risk of bleeding for warfarin. One FDA reviewer stated that "if a patient is well-controlled on warfarin, then there is no reason to switch to dabigatran."
Net Clinical Benefit (NCB)
The 150 mg dose of dabigatran was associated with a greater reduction in stroke events than an increase in major bleeding events. While the clinical impact of efficacy and safety events must be carefully weighed, there was a net benefit observed for the 150 mg dose of dabigatran. This was true in those patients over the age of 75 as well. The patients over the age of 75 had a higher risk of bleeding, but an even higher risk of efficacy events as well. There was not a benefit of the 110 mg dose in those over the age of 75 and the net benefit of the 110 mg dose was less clear.
Drug Induced Liver Injury (DILI)
Dabigatran was not associated with an excess risks of DILI including Hy's law.
Questions posed by the FDA
Question 1: Was RELY adequately designed?
Unblindinded design was adequate but not optimal. Some argued that an unblinded trial actually parallels the real world practice better. Mortality was still favorable.
Question 2: Were the drug doses appropriate?
Yes.
Question 3: Were events after drug discontinuation handled appropriately?
An ITT analysis for efficacy is the appropriate method with mITT analyses providing supportive information. For safety analyses, mITT may be more informative.
Question 4: Was the trial adequately conducted?
The time in therapeutic range of 65% parallels clinical practice and is adequate for comparison. This rate is similar to other trials. Would not want to compare to unrealistic medical care.
Question 5: Was the follow-up of endpoints adequate?
Yes. Implausible that patients who were missing would have changed the results. Hepatotoxicity data good.
Question 6:
Was dabigatran effective in reducing the primary endpoint?
For stroke yes, non CNS embolism was not sufficiently powered.
Was dabigatran effective at both doses?
Yes was the majority view of the panel, but not by some of the FDA reviewers.
Is the 150 mg dose of dabigatran superior to warfarin?
Yes was the majority view of the panel (with a dissenting view by Dr. Steven Nissen). FDA reviewers were not convinced.
Is 150 mg superior to 110 mg?
Less compelling than the data showing superiority to warfarin. Majority view was yes.
Is there any evidence of hepatic injury
No, although idiosyncratic reactions may take years to appear.
Is there any evidence of less bleeding for the 110 mg dose vs warfarin?
Yes. Mixed reviews as to whether the data split by TTR is of relevance (criticized as a post-randomization event).
Is the risk of bleeding with 150 mg of dabigatran similar to warfarin?
150 mg dabigatran bleeding risk was slightly better but did not reach statistical significance.
Voting Question: Should dabigatran be approved for the reduction of stroke and non CNS embolization for non-valvular atrial fibrillation?
The panel voted 9 to 0 yes.
Should both 110 mg and 150 mg doses be approved?
Mixed views were expressed. 150 mg appears to be the preferred dose. 110 mg may also be appropriate in patients at high risk of bleeding, where the risk of bleeding may outweigh the potential benefit, but it should be realized that this 110 dose is not superior to warfarin with respect to efficacy.
What dose of dabigatran should be approved
There was mixed views on this. 4 Voted in favor of a single dose, 6 voted in favor of approving both doses. Steve Nissen and others recommended approval of both doses. While the primary benefit over warfarin is in the 150 mg dose, people at risk of bleeding may be getting nothing at all and could benefit from the 110 mg dose. Sanjay Kaul, Michael Lincoff and others would approve only the 150 mg dose.
Has the dabigatran dose relationship been adequately defined? Are further studies needed?
Although the lack of monitoring is convenient, some people advocated for PK/PD guided dosing to optimize the risk benefit ratio say in the elderly patients.
- ↑ Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, Pogue J, Reilly PA, Themeles E, Varrone J, Wang S, Alings M, Xavier D, Zhu J, Diaz R, Lewis BS, Darius H, Diener HC, Joyner CD, Wallentin L (2009). "Dabigatran versus warfarin in patients with atrial fibrillation". The New England Journal of Medicine. 361 (12): 1139–51. doi:10.1056/NEJMoa0905561. PMID 19717844. Unknown parameter
|month=ignored (help);|access-date=requires|url=(help)