The Study of the Week Is A Study That Did Not Happen
The RECOVER IV trial of an LV assist device vs standard care was planned, began recruiting and then stopped. The reasons say a lot.
This is the first time the study-of-the-week discusses a study that did not happen.
The main lessons of the cancelled RECOVER IV trial are hubris and dualities of interest. These lessons are key to understanding medical evidence.
Background
RECOVER IV was to study a medical device called a micro-axial flow pump. It is a cardiac assist device, which I previously discussed here and here. Briefly, it is placed across the aortic valve into the left ventricle. The brand name Impella device pulls blood from the LV and pumps it into the aorta.
One of the most common uses of LV assist devices is cardiogenic shock due to an acute MI. When an MI occurs in a large coronary artery, the downstream heart muscle stops contracting. This lack of flow can cause organ failure. A cardiologist can open the blocked artery, and the heart muscle will likely improve. But during the time of low flow, the patient can die due to organ death.
The grand idea of the Impella device is to provide support for hours to days while the injured heart muscle regains function. It makes great sense; it is plausible.
That last sentence should trigger regular readers of Sensible Medicine—because we have covered many plausible ideas that fail to pass muster in the randomized trial.
Complex Regulatory History
Impella was first approved in 2008 through a lax pathway wherein it was deemed sufficiently similar to a cardiac bypass pump. Soon after, FDA came to their senses and reclassified many high-risk devices via the 515 pathway.
In 2015, FDA approved Impella again using data primarily from a trial called PROTECT II, which compared Impella to another LV support device called the intra-aortic balloon pump.
What I type next is not a typo: PROTECT II was a) stopped for futility, b) failed to show any superiority of Impella over IABP, and c) previous studies had shown the IABP to be no better than standard care. FDA still approved the device.
Between 2009-2023, many thousands, perhaps hundreds of thousands of patients, had Impella devices implanted.
Doctors and hospitals are paid extra to use the expensive device. As it is with many such devices, high usage of device brings reputational gravitas.
Finally a Trial
Can you imagine? Hundreds of thousands of devices without a positive trial. Then the Danes came to the rescue.
As I covered in April, the DANGER-Shock trial produced positive results. The Impella device reduced death by a statistically significant 12% vs standard of care. DANGER-Shock, like many trials, had caveats. One of the most important of these were that it took 10 years to recruit 360 patients. These were highly selected Danish and German patients.
Cancellation of RECOVER IV
The clinicaltrials.gov site says that RECOVER IV planned to study Impella vs standard care in cardiogenic shock in 560 patients, mostly in US centers. (More than DANGER-Shock).
Last week, however, it was stopped. The data and safety monitoring board recommended stopping the trial because of the positive results of DANGER-Shock.
Key opinion leader, Dr Gregg Stone captured the idea in this Tweet.
The DSMB, Dr Stone, and many US leaders felt that there was no longer equipoise to conduct a trial. If there is not uncertainty of benefit, it is not ethical to randomize patients.
Hubris and Duality of Interest
I disagree strongly with this sentiment. In last week’s edition of the This Week in Cardiology podcast, I cited 8 reasons that this decision represents over-confidence from US cardiologists.
I took a picture of my thoughts about DANGER-Shock.
In sum, DANGER-Shock was a strong trial but it was not enough to settle uncertainty. US cardiologists are over-confident.
And. At the risk of sounding cynical, I must consider duality of interests.
RECOVER IV was supported by the makers of Impella. Termination of the trial due to perceived benefit of their device was clearly positive for them. US cardiologists too benefited.
I am not against this device. Interventional cardiologists have said there are patients who don’t survive without the device. DANGER-Shock supports this opinion—somewhat. But not enough.
The medical literature is replete with examples of single trials failing to replicate. Impella is a costly and invasive device. If ever we should have a confirmatory trial, it should be for this device.
That we won’t have such a trial speaks to over-confidence and dualities of interest.
Indeed, what gets studied is an important lesson in medical evidence.
Looks like the pump has problems cardiologists should be aware of, nevertheless.
NYT, 3/29/24: "Heart Pump Is Linked to 49 Deaths, the F.D.A. Warns"
https://www.nytimes.com/2024/03/29/health/fda-heart-pump-fatalities.html
But don't ask the company that makes them. They say they're "safe and effective."
https://www.heartrecovery.com/products-and-services/impella
Gone (or so it seems) are the days when a study result needed to be replicated before being accepted as “real”.
Long gone are the days where the FDA barely resembled a regulatory agency worthy of the name.
Again, I think Dr. Mandrola is far too kind and circumspect with his language. These are not merely “dualities” of interest. These are outright naked conflicts of interest where the people who generate the data are also considered the experts in the field who are leveraged to “sell” the data…ultimately for the benefit of the device maker and sometimes by happenstance for the benefit of the patient.
I’m not an interventionalist so I have no direct involvement in any of this stuff but while I think FDA regulation of medical therapy has hit rock bottom, it’s laughable regulation of devices has easily smashed through that barrier. This is yet another example of the latter.