CoAxia, Inc. is out of business because the FDA lured the company into the de novo process.
That’s the warning from industry lawyer Mark DuVal of DuVal & Associates, P.A, in a May 2013 “Client Alert.”
De Novo Process
The de novo process was intended to be applied when a predicate device does not exist and the device was not so risky that it required the PMA (Premarket Approval) pathway.
Duval warns companies considering entering the de novo process that while the program looks like a benign substitute for the 510(k) program, in reality, it disguises PMA-like data requirements and “may devour your company.”
And CoAxia NeuroFlow is exhibit A.
Serving CoAxia Up
CoAxia was the first company to receive an FDA offer of a de novo panel. And they were the first to accept the offer. It occurred after CoAxia failed a de novo approval with divisional staff alone. Says Duval: “The prospect of an independent expert advisory panel was appealing to the company. That was before we discovered how manipulated it could be by divisional staff. The device did not get approval for the labeling the company sought—first as an extension of the 510(k) labeling, and second, as part of a de novo reclassification and approval.”
CoAxia NeuroFlow is a dual balloon catheter which the FDA twice cleared for use in the descending aorta to divert blood flow from the lower extremities to the upper extremities, such as in the cerebral, cardiac and pulmonary vasculature.
In addition to two 510(k) clearances, the device has a Humanitarian Device Exemption (HDE) for use in patients with cerebral vasospasm.
The device would be used in patients who need more blood in the head, such as those with cerebral ischemia or, arguably, ischemic stroke. After its two clearances and HDE approval, CoAxia conducted a 500+ patient randomized trial which demonstrated that the device was safe when used in ischemic stroke patients.
Furthermore, CoAxia’s study documented a 2 to 1 reduction in mortality over standard of care.
So, what happened?
For many companies with FDA experience, this will sound depressingly and alarmingly familiar.
First, the FDA’s staff recommended a particular endpoint for the study which was different than the endpoint the company wanted. Staff wanted “return to normal.” Says Duval: “This is a very difficult endpoint to demonstrate and one to which the company, in hindsight of course, should not have acquiesced.” Sure enough, even though twice as many patients lived versus standard of care, the number who “returned to normal” did not meet the pre-study target. In the FDA’s view, the study did not meet the endpoint—their endpoint.
From PMA to 510(k) to De Novo to Oblivion
CoAxia missed the FDA’s efficacy endpoint but, DuVal points out, the body of clinical data still had substantial worth. Based upon this clinical data, the company abandoned its PMA submission and simply asked to extend the current 510(k) labeling for use in ischemic stroke patients.
The FDA said “no” and fought CoAxia’s request over the course of several years.
The question considered by the FDA was whether the company should be able to clarify the labeling to state the device is a “tool” and not a “treatment.”
The company argued the tool claim is a specific indication falling under the general intended use. However, the FDA’s review division ruled that the device was NSE (not substantially equivalent) because the proposed use constituted a new intended use and therefore was a “treatment.”
DuVal: “FDA’s decision shows how subjective this determination/interpretation is because this device certainly can be used by physicians today in ischemic stroke patients.”
“If FDA wanted to embrace the 510(k) program and Least Burdensome requirements, it could just as easily have justified a decision to find that the proposed use fell comfortably within the general use and made a substantial equivalence determination.”
“Instead, FDA forced the device onto the PMA path (with a de novo stop in between) and to support a request for yet another large clinical trial, thus effectively killing the company and the use of its technology in stroke patients, ” charges DuVal.
ReGen Deja Vu
This should remind readers of the Regen Biologics, Inc.’s 510(k) clearance case where the FDA staff ignored two orthopedic panel recommendations that the device was safe, and probably effective, and in violation of its own rules revoked the company’s existing clearance status. That company also went out of business after the FDA action.
DuVal warns companies that the CoAxia case demonstrates how the FDA can use its power to support a request for more and more data and make a decision that is more Class III, PMA-like than a Class II, moderate risk de novo decision; and how the panel process can be manipulated by divisional staff to achieve the end they desire.
Off-Ramp to De Novo
The agency is “subtlety and indirectly redirecting many submissions that normally would have been effectively handled by the 510(k) program on to the de novo path, ” said DuVal.
Why would the FDA do that?
DuVal says it is because de novo path provides the agency more administrative control to dictate the quality and quantity of data than would otherwise be necessary under the 510(k) path. When FDA asks for data under the de novo path it is not tethered to the 510(k) standard of “substantial equivalence, ” which requires an applicant to demonstrate safety and effectiveness in a comparative sense to a predicate device.
The de novo path, claims DuVal, allows FDA to use the same standard as with a PMA, albeit in the context of a moderate risk (Class II) device.
By diverting as many 510(k) applications as possible to the de novo path, DuVal says the FDA can require as much data as it wants for approval. “Although FDA may disagree, the de novo path is PMA-like. Some would say it is PMA-lite. FDA finds it difficult to hold itself to a moderate risk (Class II) standard, ” added DuVal.
Under the de novo path, the FDA can get into many definitional battles with applicants regarding the interpretation over the elements of the 510(k) program, i.e. whether a device has the same intended use, same technological characteristics, or has different characteristics and/or the question of whether the new technological characteristic raises new types of questions of safety and effectiveness.
De Novo Trap
“If you thought FDA’s seemingly benign offer to pursue de novo to keep your device dream alive because your device was not going to survive the 510(k) path (i.e. an NSE letter), think again, ” warns DuVal.
“Let’s face it, no matter what path your device is on, FDA wants to ask for whatever information it desires, even if it exceeds the statutory construct for a moderate risk device; even if it exceeds the demands of good scientific judgment; even if it looks like a science project; even if it far exceeds what was required to obtain a CE Mark in Europe. Once FDA has an applicant in the de novo world the only limitation on its far-reaching administrative judgment is the well-intentioned, but often loosely defined and infrequently applied, Least Burdensome requirements.”
DuVal writes that some would say FDA is doing indirectly what it cannot do directly, i.e. require the “science project-like” data it prefers of many applicants of PMAs.
The FDA Angst
“One also suspects that the real reason for asking for a clinical trial is that FDA seems to be attempting to shore up past regulatory clearances where no clinical data was required.”
He believes today’s FDA often feels the “old” FDA did not do its job and should have required far more data for clearance of the previous generation of 510(k)s. “Never mind that the predicate devices have not demonstrated any safety issues or, in the rare case where there are some safety issues, they are not significant enough to merit a clinical trial as solution.”
Changing the Predicates on the Fly
He says the FDA now seems to be asking for clinical data from new 510(k) applicants to create new predicates in the predicate family for which clinical data are required. This will, in turn, enable FDA to request clinical data of future device applicants, whether such data are truly needed or not. Frequently, that desire to have clinical data as the norm, where none was required in the past, seems to be another reason for pushing devices down the de novo path.
Internal FDA Politics
There are also internal FDA political issues here.
DuVal observes that applicants often disagree with a review team regarding the appropriate pathway for a device and makes its case that the device has a legitimate predicate(s) and deserves to be considered under the 510(k) path. “And it seems there are also internal disagreements within the Agency about whether a device belongs on the 510(k) path.”
It is industry’s impression that when internal disagreements occur, management will tend to side with the most conservative view, in many cases regardless of the merits of the various arguments. The dynamic at play with the de novo path is that it has become something it was not intended to be—a convenient “out” for the Agency.
“What seems to happen is that review staff and management debate internally whether a device belongs on the 510(k) path and as they struggle with the definitions of same intended use, etc., FDA often comes to an internal stalemate. Rather than management having the courage to break the stalemate and leave the device on the 510(k) path, they simply punt and suggest or direct the applicant to pursue the de novo path.”
This makes de novo an escape valve for internal disagreement and potential strife, says DuVal. “The de novo program was never intended to be a default position for making tough decisions or a tie-breaker for moments when there is internal controversy over whether or not a device belongs on the 510(k) path.”
DuVal says this is one of the reasons why Congress recently passed Section 603 of the FDASIA. It requires FDA to produce documents to show the various internal opinions expressed during the review process.
He says it is industry’s hope that this provision will help to reduce bad decisions aimed more to minimizing internal controversy than in making the correct decision. “These bad, politically motivated decisions frequently lead to the use of the de novo process as an escape valve. In doing so, the de novo process becomes a de facto substitute for 510(k) reviews.”
“Grandma, What Big Teeth You Have”
DuVal warns his clients to remember the story of Little Red Riding Hood and the Big Bad Wolf as a metaphor for the de novo process. He says don’t be fooled by the wolf dressed like Grandma.
