August 28th, 2011
A “Straight A” Trial: Answers About Apixaban from ARISTOTLE
Samuel Goldhaber, MD
ESC 2011 is being held in a gigantic venue called the Exposition Center, close to Charles de Gaulle Airport. This enormous campus is the perfect place to showcase ESC, which can now boast that it’s the biggest cardiology meeting in the world. Its attendance and registration numbers surpass those of AHA and ACC. And the scientific quality of the presentations is excellent.
That is the backdrop for the presentation of ARISTOTLE, where Chris Granger unveiled results of an 18,000-plus patient megatrial of apixaban versus warfarin for stroke prevention in patients with nonvalvular atrial fibrillation. Imagine a trial with about 1000 institutions participating from 39 countries. Try to think of the logistics involved in a randomized double-dummy protocol: Each patient underwent finger-stick INR testing; and half of the test results were sham INRs, leading to dose adjustment of the placebo for warfarin.
The presentation auditorium was aptly named “Paris,” and all seats (I’m guessing at least 3000) were filled. We knew from a manufacturer press release 2 months ago that apixaban was superior to warfarin for both efficacy and safety. But we didn’t know the magnitude of the difference between the two groups. Prior to the presentation, there were more questions than answers.
Questions Answered by Today’s Presentation
- Was the time within the therapeutic range (TTR) for warfarin acceptable (defined by most clot experts as 60% or greater)? Yes. It was 62%.
- Were there any subgroups that didn’t benefit from apixaban’s improved efficacy over warfarin’s? No.
- Were there any types of major bleeding that occurred more often in the apixaban group? No.
- Did the reduction in stroke and reduction in major bleeding ultimately translate into a lower all-cause mortality rate in the apixaban group? Yes.
The ARISTOTLE investigators can be proud of the rigor with which they conducted this trial. If anyone doubted that we’ve entered a new era for stroke prevention in nonvalvular atrial fibrillation, fear no more.
9 Responses to “A “Straight A” Trial: Answers About Apixaban from ARISTOTLE”
NEJM Journal Watch — Recent Cardiology Articles- Strategies and Criteria for the Diagnosis and Management of Myocarditis
- Sex-Based Disparity in CABG Outcomes: Are We Moving Toward Parity?
- Can Genetic and Social Risk Scores Improve Cardiac Risk Prediction?
- Extensive Ablation vs. Pulmonary Vein Isolation in Persistent Atrial Fibrillation
- Can a Novel Stent Design Improve Long-Term Outcomes of Percutaneous Coronary Intervention?
Curious-was ‘nonvalvular’ defined strictly for the trial?
Competing interests pertaining specifically to this post, comment, or both:
None
The ARISTOTLE apixaban data: is it flawed or good enough?
August 28, 2011
The full ARISTOTLE trial data are now published. As we reported last week, the biggest controversy for the data arise from the clinical trial design that uses a combination endpoint of two very different types of stroke: the kind caused by bleeding into the brain (hemorrhagic), and the kind caused by clots forming on the heart as a result of atrial fibrillation and flicking off to the brain starving portions of blood flow (ischemic). The trial would have failed to show that apixaban was as good as warfarin at preventing blood clot formations and ischemic stroke if the endpoint just measured ischemic stroke. Instead, it measured also hemorrhagic stroke, and the lower bleeding from apixaban drove the overall endpoint to success.
Supporters of this trial design, common to all of the A-fib trials (RE-LY for Pradaxa, ROCKET-AF for rivaroxaban, etc), say that patient does not care what type of stroke they get, just as long as the drug causes fewer of them.
Critics of the design say that by double counting bleeding strokes as a safety endpoint as well as the efficacy endpoint, it makes the risk/benefit analysis meaningless. The conclusion in ARISTOTLE published in the NEJM states that apixaban is superior to warfarin in safety as it causes fewer adverse bleeding events. Is that fair to say since the same bleeding events were double counted in the stroke-prevention efficacy endpoint?
Other critics will say that a drug should be prescribed for a specific reason, such as reducing clot formation in this case. If the drug fails at that endpoint in a trial, as apixaban did in this trial (was not statistically significant with ischemic stroke), but accidentally benefits the patient in another way (lowers risk of bleeding stroke) then it muddies the waters too much to be a clinically relevant hypothesis-answering study.
Moreover, trials studying very rare events like stroke are well known to be unrepresentative of the real world experiences on millions of patients. If the FDA were to approve apixaban because it was less effective than warfarin at reducing ischemic stroke, but was safer at reducing bleeding accidents, and millions of patients started using the drug, is the ARISTOTLE data strong enough to assure that more ischemic strokes will not flare up?
This raises the second controversy of ARISTOTLE: it is yet another large trial conducted around the globe collecting data from all of the continents, in unreliable medical centers where corruption and bribery are common. The FDA is concerned about this trend of outsourcing trials. The geographical subgroup analysis (see bottom of Figure 2, page 8) shows that the drug actually failed to meet endpoints in all of the regions with the exception of Asia Pacific, etc., yet the overall global pool endpoint was met. The P-value for interaction was not significant at 0.44, which suggests that there are no true geographical differences. However, the primary analysis performed by the actual FDA staff almost always reveals surprises.
In summary, Bristol-Myers Squibb’s apixaban, as well as the other Factor Xa inhibitor class JNJ/Bayer drug Xarelto (rivaroxaban) which is heading to an FDA panel in two weeks, will likely be approved by the FDA. However, the FDA panels will focus on these two controversies.
Competing interests pertaining specifically to this post, comment, or both:
none
If necessary, if a patient is involved in an accident, does anyone know how to reverse apixaban? Also, from what I have read apixaban is a BID medication, I do not see how you could truly do a double blind study when the usual course of warfarin is QD or less. I would like to also mention that the patients in our own practice reach 90% in the range of 2-3 INR for atrial fibrillation treatment for stroke prevention. Finally, I would like to know if there are storage issues with apixaban. I do know that dabigatran breaks down very quickly and becomes useless if exposed to air.(E.G. a frail elder does not close the cap properly or stashes the missed doses for future use). As usual, I am skeptical about this study, and would like to see how this new drug does in the real world. I know stroke prevention in atrial fibrillation works in the real world with warfarin, I can reverse it with vitamin K or fresh frozen plasma, and with proper monitoring, I can keep the patient in the therapeutic range. I have been using it for over 20 years this way, and it is going to take quite a bit of post marketing surveillence to see if apixaban really is any better than warfarin. I’m not going to be convinced by an 18000 patient study named after a famous Greek philosopher or a meeting in Europe named Paris larger than anything the AHA or ACC can put together. It really sounds like a grand publicity stunt to me. I do not see this as the golden anything until it is proven by the true test of time. I will be convinced by seeing a few years of safety and efficacy bebore making the move from getting my knees wet to diving in on this one. One should always harbor fear when diving into dark water, and it is best not to do it unless your life is at stake. “Fear no more” sounds like an advertising phrase for a surfboard manufacturer I am not going to mention.
Competing interests pertaining specifically to this post, comment, or both:
My only conflict of interest is the fact that I know that Warfarin is a very inexpensive drug which has saved many quality adjusted years of life in sufferers of atrial fibrillation, as well as other thromboembolic syndromes. I have no industry conflicts of interest. I do not speak for industry, I collect no monies from industry, and do not own any stock certificates. I do not fear speaking my mind. I did note the sponsors of the study called Aristotle and the involvement of the various authors with the pharmaceutical industry and would consider this a potential cause of skepticism on my part. I am sorry to see the long list of payors for Dr Goldhaber, a cardiologist of esteem whom I respect. Unfortunately the payors we in our practice must deal with are the American “healthcare” insurers who will be as willing to pay for this medication as the average person would be willing to have a tooth extracted without anesthesia.
DETAILS? In ARISTOTLE, about 5% of patients got half dose apixaban based on age, weight, and creatinine ( 2 out of 3 criteria). This may have contributed to lower bleeding rates. No such adjustments were made in RE-LY.
The reduction in the primary endpoint was greater in RE-LY (I know…cant really compare). The primary endpoint reduction with apixaban was as mentioned mostly due to hemorrhagic stroke reduction (accounting for two thirds of. the reduction). Should dabigatran be preferred for higher CHADS/lower ICH risk patients?
A point of care anti-Xa assay may be a big bonus, particularly if pharmacokinetic information delineating a therapeutic range ( like for LMWH ) are available. This might allow for dose
verification in borderline scenarios. I think these assays are available now.
Competing interests pertaining specifically to this post, comment, or both:
BI speaker
Pradaxa we hardly knew ya
Competing interests pertaining specifically to this post, comment, or both:
none
Pradaxa is no better than warfarin in pateints who are at goal INR>2/3 of the time. Though this is pointed out in the product insert the pradaxa ads leave this out- fortunately, the FDA insisted on this analysis. Would love to see this analysis for apixiban.
I’ve read the trial data and what impressed me the most is the fact that even 18,000 patients weren’t enough to prove reduction in embolic events. Only inclusion of hemorrhagic stroke made primary endpoints statistically significant. One can interpret the data that apixaban is a safer drug. On the other hand maybe warfarin wasn’t monitored properly especially in centers outside North America. Point-of-care INR device probably requires some experience to work with. They even used some sort of a program to improve the quality of INR control. Maybe they needed better quality program!?
In addition, therapeutic INR was achieved 62% of the time. What about other 38%? I suspect mostly over 3 which can explain some hemorrhagic complications.I would be interested to know what was average INR in hemorrhagic stroke patients. Why didn’t they report it?
I am really not a big fan of warfarin but it’s funny that after so many years there are no clean trials to prove superiority of new drugs.
Competing interests pertaining specifically to this post, comment, or both:
none
Guys, I don’t know what you are smoking. You cannot be a clinical trials investigator of new agents unless you have some connection to the pharmaceutical manufacturer — and with regard to the comment about Sam Goldhaber above, it is far better to be involved in clinical trials with agents from many manufacturers than just one: it generates expertise and consistency in clinic trial conduct, provides familiarity with the entire field of possibilities, prevents developing allegiance to one drug over another based upon anything other than trial outcome, and allows more objectivity than that acquired from being under a single manufacturer’s influence. Most academic institutions and their hospitals have strict rules concerning trial-related conduct with regard to commercial entities; small, non-academic practices not so much. Dr Goldhaber’s institution has some of the most stringent constraints upon pharmaceutical company relationships with clinical investigators.
And with regard to in-range warfarin data, it would help if folks who manage their own patients on warfarin took a very careful look at their own practice’s data. 62% is hardly shabby — and I would venture to say that US practice is likely not superior to that of many centers outside the US.
Competing interests pertaining specifically to this post, comment, or both:
None.
Disagree with above. First, you said all the right things, however, it is naive to think that there is no bias. Second, as far as managing warfarin is concerned, private practice is a completely different ballgame. You have patients, generally sicker than trial patients, going to numerous physicians with different medical problems and end up on numerous interacting medications.
I don’t know what you are smoking but you must be kidding me to compare private practice to tightly handled trial patients with a bunch of doctors, nurses and research coordinators on top of them.
And one more thing, what do you really know about practice patterns overseas?
Competing interests pertaining specifically to this post, comment, or both:
none