Idarucizumab: Plenty of optimism, not enough science

Optimism is essential. Few things are as powerful as hope for the future. When facing an onslaught of critically ill patients, optimism allows emergency physicians to persist. Optimism, however, can cloud also cloud our judgement; allow us to focus only on the good, not the bad. As physicians, we cannot afford optimism blindness. We need to be objective. We need to be scientists.

This month I was distressed to hear overly-optimistic, unscientific statements about idarucizumab on two of my favourite emergency medicine education programs: EM:RAP and EM Cases. I have incredible respect for these sources. EM:RAP has been irreplaceable in my emergency medicine education. On EM Cases, the statements were made by Dr. Walter Himmel, who is one of the smartest individuals I have ever had the opportunity to meet. But in this instance, I think that they were both wrong.

What has got me so riled up?

In the January 2018 EM:RAP, the paper chase segment gave us a fairly uncritical look at the one big study “supporting” the use of idarucizumab. They conclude: “Put a fork in it. This one is done. Idarucizumab is here, it works, it’s safe. If you don’t have it, get it, and if you do have it start using it.” That’s a pretty glowing review, but unfortunately it’s dead wrong.

Around the same time, Emergency Medicine Cases released an (overall excellent) episode on the management of intracerebral hemorrhage with Dr. Scott Weingart and Dr. Walter Himmel. There is some great stuff in there, but then they talked about idarucizumab. Their conclusion is a little less glowing, but the end result is the same: “The best evidence to date – and I’m not saying it’s great – but the best evidence is to use idarucizumab.”

Here are two very widely listened to, highly respected programs suggesting the use of idarucizumab. I understand their optimism, but their science is way off.

So let’s start with the science:

Pollack CV, Reilly PA, van Ryn J. Idarucizumab for Dabigatran Reversal – Full Cohort Analysis. The New England journal of medicine. 2017; 377(5):431-441. PMID: 28693366

What kind of study is it? It is a multicenter, prospective, observational trial. (So no blinding, and not even a control group.)

Who ran the study? Boehringer Ingelheim was involved in all aspects of the design, implementation, and write up of the trial.

What did they do? They looked at 503 adult patients on dabigatran. Group A consisted of 301 patients with “uncontrollable or life threatening bleeding” that the treating physician judged to require rapid anticoagulant reversal (45% had GI bleeds, 33% intracranial hemorrhage, and 25% trauma). Group B consistent of 202 patients who were about to undergo surgical or invasive procedures that could not be delayed 8 hours. They gave all of these patients 5 grams of idarucizumab IV, and a second 5 gram dose could be given if needed. (It really has to be emphasized that there is no control group here.)

What were they looking for? Their primary outcome was biochemical: the maximum percentage reversal of the anticoagulant effect of dabigatran. (So their primary outcome was not a patient oriented outcome, but they did look at some clinical outcomes as secondary outcomes.)

What did they find? On entry to the trial, 92% of patients had prolonged clotting times. They were able to completely reverse these lab numbers in 100% of patients by 4 hours. (23% did have their numbers come back up in the first 24 hours.)

So let’s just pause there for a moment. Think back to the most basic critical appraisal class you have ever attended. This is an industry run trial, with no control group, no blinding, and that looked at a surrogate outcome. It doesn’t check even a single science checkbox.

They did look at some clinical outcomes, but without a control group, it is impossible to make any valid conclusions. They looked at “time to confirmed bleeding cessation” in group A. They couldn’t measure anything in the intracranial hemorrhage group, so they ignore them. In the remaining patients, they state that the median time to hemostasis was 2.5 hours, although I am not sure how they can be certain when most of the patients had GI bleeds. More importantly, without a control group, we don’t know what that 2.5 hour number means. It is possible these patients would have all stopped bleeding by 2 hours without the drug. There is no way to know. Finally, in an unblinded study controlled by the drug company, there is certainly room for some bias in determining exactly how long it took to get “confirmed bleeding cessation”.

They tell us that the 30 day mortality was 13.5% in group A and 12.6% in group B. Again, this information is meaningless without a control group. The EM:RAP commentators tried to compare these mortality numbers to previous cohorts, claiming these are great numbers, but you just can’t do that. They aren’t the same patients. If I was the manufacturer running this trial, you can be sure I would be enrolling the healthiest possible patients in order to show the best possible outcomes. (Speaking of which, this trial does not provide the standard flow diagram, so we have no idea how many patients they screened to arrive at the final study population. We do know that the 503 patients came from 173 different hospitals over 2 years, and that there were another 196 hospitals ‘initiated’ into the study. We don’t know how many patients were excluded, but we know very few made it in: about 1.5 per year from the hospitals that actually contributed data.)

And there were adverse events. There were thrombotic events in 4.8% of group A and 6.8% of group B. Other serious adverse reactions were reported in 25% of the patients, including cardiac arrest and shock in 3.5% and 3.0% respectively. Although it is possible that these are expected events in sick patients, without a control group it is impossible to know. What is certain is that, despite the EM:RAP conclusions, this data certainly cannot demonstrate the safety of this new drug.

Bottom line: This is not science. Without a control group, it impossible to know what to make of the results. However, a direct reading of the results would be: using idarucizumab will reverse (potentially irrelevant) laboratory numbers in 100% of patients, but causes serious adverse events in 25% of patients and thrombotic events in another 5%. Not a ringing endorsement.

So why wasn’t there a control group? Why spend so much time and money on an effort that provides practicing clinicians with no useable clinical information? The manufacturers of the drug, when designing the trial, stated that “it would be unethical to withhold a reversal agent from a patient who may benefit from it.” (Pollack 2015) We don’t need to study it, they said, because we know it works. That is their argument. In fact, they are so certain that it works that it would be unethical to study it. How convenient for them. I have some snake oil that is similarly effective, if anyone is looking to buy.

From an ethical standpoint, they have everything exactly backwards. It is actually the running of a trial without a control group that is unethical. Trial participants assume the risk of unproven and potentially unsafe interventions for the reward of scientific advancement. However, without a control group, we are left with more questions than answers; we have a trial that cannot inform clinical practice. If it was so unethical to withhold this drug, then this trial should never have been run. The drug should have simply been provided to those in need, with the cost waived for anyone unable to afford it. It was clearly unethical to expose these patients to the potential risks of this experimental intervention when the trial design ensured that the results of the trial would be scientifically uninformative.

What should we do now? Despite the complete lack of evidence for either efficacy or safety, EM:RAP and EM Cases seem to think we should just start using this drug. I think that is misguided. I understand the impulse that we need to do something when faced with critically ill patients, but we have no idea if this medication helps. It is entirely unproven. And when you consider the history of medical innovation, it is far more likely that idarucizumab will end up being ineffectual, or even harmful, than it is to be helpful.

I have heard people argue that even if a RCT was the ideal trial, we shouldn’t run one now; that we have enough data; that a RCT at this point would just expose patients to needless harm by being included in the control group. This logic is also wrong. Let’s run some simple math.

In the best case scenario, I will assume that idarucizumab results in a 10% absolute reduction in mortality. (Which would be an astronomical benefit, essentially unheard of in medicine). In the worst case scenario, I will assume that idarucizumab actually results in harm to about 1% of patients (a much more realistic scenario, when you consider the history of the medical literature).

If we ran a randomized control trial including 1000 patients, 500 would be assigned to the control group. Those 500 would miss out on the 10% benefit, meaning that approximately 50 patients would be harmed. On the other hand, if we decide not to run any further trials and just start using idarucizumab on patients, we can expect that it will be used more than 15,000 times a year in the United states alone. (There were 500,000 prescriptions for dabigatran dispensed in the fourth quarter of 2015, and the rate of major bleeding in the RELY trial was 3%). That means that over the next decade, 150,000 patients would be exposed to idarucizumab in the United States. Using the conservative estimate of 1% harm, 1,500 patients would be harmed in the US alone. So our options are to perform a trial now that may expose 50 individuals to potential harm, or just start using the unproven drug and potentially cause harm in thousands of patients. To me, the right choice is pretty obvious. Science isn’t unethical, it is necessary.

We need to demand a proper RCT of idarucizumab before we start using it. A RCT is what should have been done in the first place. I guarantee that trial will be done if the medication isn’t selling. However, if we just start using this unproven, experimental medication, Boehringer Ingelheim will have no reason to engage in science. They will already have sold us their snake oil.

Optimism is essential in medicine, but it can’t come at the expense of science. We cannot let pharmaceutical companies pretend that marketing is actually science. We cannot allow companies who stand to make billions convince us that science is not necessary; that it would be unethical. Presently, we have no idea whether idarucizumab will help our patients. It might, but based on what we currently know, I think it is just as likely to hurt them. I will not be using idarucizumab until we have real science. I do, however, remain optimistic that in the future we will all recognize science’s essential role in caring for our patients.

Follow up post: Most medical practices are not parachutes

I talked to (debated) Scott Weingart about this post on EMCrit.

I have very similar concerns about andexanet alfa.

References

Pollack CV, Reilly PA, Bernstein R. Design and rationale for RE-VERSE AD: A phase 3 study of idarucizumab, a specific reversal agent for dabigatran. Thrombosis and haemostasis. 2015; 114(1):198-205. PMID: 26020620

Pollack CV, Reilly PA, van Ryn J. Idarucizumab for Dabigatran Reversal – Full Cohort Analysis. The New England journal of medicine. 2017; 377(5):431-441. PMID: 28693366