A public debate escalated last week over how device recall data reflects the success or failure of the 510(k) review program in keeping unsafe devices off the market.

Kicking things off: A journal article calling for serious reforms of the pre-market program based on the large proportion of high-risk recalls that are 510(k)-cleared products.

The Archives of Internal Medicine paper, which attracted significant attention in the media and on Capitol Hill, pressed for FDA device reviewers to more broadly define the types of risks that should be considered when determining appropriate pre-market requirements and, ultimately require clinical trial and pre-market manufacturing inspections for more products.

Industry was quick to point to problems with both the data and the conclusions, and received support from House Republicans on the Energy and Commerce Health Subcommittee, who had called a hearing for Feb. 17 in response to complaints by industry that FDA device oversight was inhibiting innovation and job growth.

In the middle of the debate is the agency itself, which on one hand defends the overall success of 510(k) reviews in ensuring safety, but on the other hand acknowledges that some changes are needed, and is firing back at industry claims that FDA is primarily to blame for a U.S. device innovation slow-down. (See in this issue: “ Also see "Shuren On The Hill: Industry Should Take More Blame For Review Delays" - Medtech Insight, 21 February, 2011..”)

Zuckerman, Nissen Press For Tighter Controls

The Archives paper was co-authored by patient advocate Diana Zuckerman, of the National Research Center for Women and Families, and Steven Nissen, a high-profile cardiologist and frequent FDA critic, of the Cleveland Clinic.

They looked at FDA high-risk Class I device recalls from 2005 through 2009 and determined that 71% (80 out of 113) involved 510(k)-cleared products, with the rest involving PMA or 510(k)-exempt devices.

FDA is somewhat critical of the Zuckerman/Nissen study, echoing industry's main points.

So many products that are considered intermediate risk prior to pre-market clearance should not be triggering "high risk" recalls associated with the potential for serious injury or death, they argue. Thus, they conclude, there is something wrong with how FDA determines risk pre-market.

"If a product can kill you if it fails, that makes it a high-risk product," Zuckerman said in an interview. "The fact that we are looking only at the highest risk recalls means they virtually all – granted there could be exceptions – but virtually all should be for PMA products."

The paper recommends that the agency more fully consider the risks of device failure when determining if a device class should require a PMA or 510(k) review (or, at least, clinical data) and expand its authority to include performance of pre-510(k)-clearance facility inspections, among other things.

Don't Forget The Denominator, Industry Says

Industry disputes the findings, pointing out that the proportion of 510(k)-cleared products that are recalled compared to the total number of 510(k)-cleared products is extremely small.

According to a study by University of Minnesota law professor Ralph Hall looking at Class I recalls during the same period, that proportion is, in fact, 0.45% (89 recalls out of 19,873 submissions). (See box, "Maisel's Take.")

Hall, who represents device clients as a part-time attorney with Baker & Daniels and is CEO of an early-stage device start-up, summarized his research as a witness at the Energy and Commerce Subcommittee hearing Feb. 17. He had presented it in full at an Institute of Medicine committee meeting on the 510(k) program last July. (See " Also see "Broad 510(k) Reform Still On The Table; IOM Panel Enters Closed-Door Phase" - Medtech Insight, 2 August, 2010. " – "The Gray Sheet" Aug. 2, 2010.)

Maisel's Take

Another study that addressed links between 510(k) clearances and recalls was presented alongside Hall's analysis at last summer's IOM meeting by then-Harvard electrophysiologist and current CDRH Deputy Director William Maisel. Maisel looked at all recalls, not just high-risk Class I's. He found the annual rate of recalls for 510(k) products to be about 1.3%-1.5% per year for the first four years after clearance, and about 1% for devices that have been cleared for five to six years. He also found that the most common causes of 510(k) product recalls are manufacturing process and device design issues.

"If you just look at the numerator, that does not tell you if you have a big problem or a small problem," he told the panel. "And most importantly it does not tell you how to fix it."

His study also assessed the reasons for each Class I 510(k) recall during the 2005 to 2009 period, and found that 55% were tied to post-market issues like manufacturing or labeling errors.

"The pre-market system and clinical trials have nothing to do with that," he said in an interview prior to the hearing.

Overall, he said, his data show that the 510(k) program is "doing a very good job" of ensuring safety, and that the most improvement would come from focusing on post-market quality systems inspections rather than a general increase in pre-market requirements.

"If you have problems that are not related to the pre-market process, changing the pre-market process would be a useless exercise," he said.

Hall also argued that Zuckerman and Nissen were inappropriately equating pre-market risk determination, which addresses risk of the intended use, with risk assessment for a recall.

During the hearing, he used a theoretical example of a tongue depressor to illustrate: "No one would consider a tongue depressor to be high risk or need a clinical trial. But if a tongue depressor gets contaminated with deadly bacteria, the recall for that tongue depressor should be a high risk recall. You cannot link the approval classification with the recall."

Nissen Sees "Serious Problem"

Steve Nissen, who was an invited witness of the Democrats on the subcommittee, was unconvinced.

While he acknowledged that the percentage of 510(k) devices involved in recalls is very small, he pointed out that the total number of individual device units tied to the 510(k) recalls tabulated in his study was about 112 million, with some of them being implantable products.

"There is a serious problem here," he told lawmakers. “It needs to be looked at very carefully."

In addition to recommending a broader definition of a "high risk" device, in his prepared testimony Nissen recommended establishing "an intermediate regulatory category more rigorous than 510(k), but short of a full PMA process, for moderate risk devices."

Rita Redberg, director of women's cardiovascular services at the University of California-San Francisco, and editor of Archives of Internal Medicine, also testified in favor of a greater reliance by FDA on pre-market clinical studies.

"The only way to know a device that I am going to put in your body is safe and effective is to study it in clinical trials, in humans, with meaningful endpoints, real outcomes and reasonable follow-up," she said.

FDA: Defending From Both Sides

For its part, FDA was somewhat critical of the Zuckerman/Nissen study, echoing industry's main points.

The highlighted recalls "represent a small number of the devices cleared via this program and don't reflect the thousands of people who have benefited from these devices," agency spokeswoman Karen Riley said in a prepared statement.

But FDA also acknowledged some weaknesses in the 510(k) program, which it says are being addressed as part of its near-term reform actions, announced last month, and in its ongoing process to belatedly risk-classify certain pre-amendment devices, such as automated external defibrillators and electroconvulsive therapy devices, that are tagged as high-risk class III products but access the 510(k) review process. (See " Also see "GAO Report On 510(k) Program Calls For Regulations On 20 Device Types" - Medtech Insight, 19 January, 2009. " – "The Gray Sheet" Jan. 19, 2009.)

During the House subcommittee hearing, CDRH Director Jeff Shuren emphasized his center's success and relevance as a gatekeeper for product safety. But most of his comments were not made in response to calls by Nissen and others for more oversight. He was more overtly critical of claims by industry and some Republican lawmakers that the agency is holding up device innovation with unnecessary requirements.

In particular, he countered an argument that has been forwarded by a couple of industry-funded reports that the European CE mark process is quicker and more predictable than the 510(k) process but does not result in any additional serious product recalls.

He pointed out that a recent AdvaMed-commissioned Boston Consulting Group study suggesting parity in the rate of serious recalls in Europe and the U.S. relies on data "in which 85% of safety reports come from only five of the 24 European countries examined, suggesting there is a sign of underreporting of safety problems by some European countries," he said.

Shuren also alluded to devices approved in the European Union and not in the U.S. that "were pulled from the EU market due to safety problems, in some cases, identified in studies requested by FDA to support approval for U.S. patients.”

"If those products were approved here, I have no doubt we would be having a different hearing today," he said.

By David Filmore

510(k) Safety Debate Heats Up With Critical Paper, House Hearing

More from Archive

More from Medtech Insight