The More Tweaks Made to PMA-Approved Devices, the More Likely a Recall

Medical devices cleared under the US Food and Drug Administration’s premarket approval (PMA) pathway are more likely to undergo a recall if their manufacturers have filed additional documentation with the agency over the years to report changes in their design, labeling, or production, a new analysis shows.

Cardiovascular products contributed the greatest number of these supplements, as compared with devices in other specialties, and they accounted for nearly half of all recalled devices.

The practice of submitting supplements to the original PMA application is meant “to facilitate incremental improvements in safety and effectiveness,” Jonathan R. Dubin, MD (University of Missouri–Kansas City), and colleagues explain in JAMA Network Open. Yet, they note, “the accrual of supplements over time has raised concerns that the modern device may bear little resemblance to its approved predecessor.”

The investigators say their results pose questions as to whether the agency’s process for approving these supplements needs more oversight, or whether better postmarketing surveillance systems are warranted.

“Clinicians should be aware that the devices that we are using are often modified over time. There is a regulatory process, but we should still be mindful that sometimes these devices that we’ve implanted have changed since they were approved,” Dubin told TCTMD, adding that these updates are often made with little clinical data to inform them and lack a “track record of proof.”

This is not meant to be the definitive study or nail in the coffin. This is meant to open up ideas for other researchers now to go out and explore and diver deeper into. Jonathan R. Dubin

Dubin acknowledged that the analysis can’t show a causal relationship—ie, that the modifications were responsible for the problems at the root of the recalls—and noted that there could be other explanations, such as devices being under extra scrutiny after manufacturers submitted supplements.

“This is not meant to be the definitive study or nail in the coffin. This is meant to open up ideas for other researchers now to go out and explore and diver deeper into,” he said.

Commenting for TCTMD, Kushal T. Kadakia, MSc (Harvard Medical School, Boston, MA), described the new paper as a “really valuable study that continues to extend the literature about medical device safety and recalls.”

He pointed out that the more rigorous PMA process is very different from the 510(k) pathway, in which devices are approved on the basis of a previously authorized products. The latter pathway has received more attention from researchers, understandably so since the proportion of devices approved through 510(k) far exceeds those that go through PMA, said Kadakia.

Even though they’re in the minority, the devices that fall under the PMA umbrella are quite relevant to cardiologists, because many are meant to address cardiovascular conditions and many are high risk, Kadakia continued. Also relevant is that the number of supplement submissions per device has been rising over time, he noted. “So it’s interesting in this study that’s now coming out to see how having more supplements can be associated with a higher risk of a recall.”

And while the study doesn’t prove causation, Kadakia said it does bring to mind situations where devices are approved, only be to be recalled decades later after dozens of supplements have accumulated. “It stands to reason that the latest version of the device is substantially different from the underlying one, and maybe some more clinical testing along the way could have helped identify [the issue behind the recall] and catch it,” he observed, stressing that postmarket surveillance is critical to gather clues on device performance in the real world.

Risk of Recall Around 30% Higher

Dubin and colleagues identified 373 devices that originally received PMA between 2008 and 2019. There were a total of 10,776 supplements for these products by the end of 2021, working out to a median of 2.5 modifications per device each year.

There are several types of PMA approval supplements. Among them are panel track supplements (PTS), which “pose the greatest risk and typically handle expanding indications and some major design change. Used least frequently, they are the only supplements to require clinical data for support,” Dubin et al note.

The other types include: 180 day (major design changes, labeling changes), real time (minor design changes), special (labeling designed to improve safety), and 30-day notice or 135-day supplement related to manufacturing changes (eg, sterilization or welding of parts).

The 30-day type stood out as most frequent (67.3%) in their analysis and PTS as least common (1.2%). Most of the supplements (67.7%) were due to changes in manufacturing procedures. On the other end of the spectrum, device and labelling changes were the least common reason for submitting supplements to the FDA (15.2% and 7.8%, respectively).

Supplements were most often seen for cardiovascular (37.0%) and microbiology devices (12.1%), while no other medical specialty contributed more than 10% of the total.

During the study period, which lasted through 2021, 97 of the devices were recalled. Thanks to some devices being affected more than once, there were a total of 192 recalls. Fully 66 devices had multiple recalls—notably, the HeartWare ventricular assist system (Medtronic) was recalled 26 times.

Of the recalled devices, 45.4% were cardiovascular. Of all the recalls, 57.5% were for cardiovascular devices. There were 20 devices with a Class I recall, the most serious type, with 75% of those being cardiovascular.

Multivariable analysis showed that each additional supplement annually was linked to a greater risk of recall overall (HR 1.28; CI 1.15-1.44) and to a Class I recall (HR 1.32; 95% CI 1.06-1.64). Cardiovascular devices were more likely than others to be the subject of a Class I recall (HR 3.51; 1.15-10.72).

“Interestingly, we did not find cardiovascular devices to be associated with greater risk of recall when combing all classes of recall, likely reflecting the elevated risks that these devices possess in addressing cardiovascular conditions in typically high-risk patients compared with other, lower-acuity fields such as ophthalmology or orthopedics,” the authors note.

Clarify Language, Strengthen Surveillance

Dubin pointed out that one reason it’s hard to draw a direct, causal link between the supplements and the recalls is because the FDA’s terminology can be vague. For instance, the agency describes the reasons for recall “often with overlapping terms such as device design vs component design or labeling design vs labeling error,” the researchers explain in their paper. “While the FDA publishes a brief manufacturer’s reason for recall, [we found] it usually added little clinical insight with which to analyze the recall from a clinician’s perspective.”

Kadakia noted another reason why it’s difficult to establish cause and effect.

“The thing that is really hard for us as researchers and clinicians is our recall process is driven by the manufacturers, not by the FDA, so it’s beholden to them to report it,” he said. “It’s hard for us to piece together a time line when there’s a delay in which manufacturers identify problems and then submit recalls, a delay between when the FDA received the recall and then publishes the recall, and a delay between when the FDA [announces] the recall, manufacturers may address it, and the FDA finally rules that it has been resolved.”

Hope Caughron, MD, and Sanket S. Dhruva, MD (both from University of California San Francisco and San Francisco Veterans Affairs Medical Center), writing in an accompanying editorial, agree that the study sheds useful light. “Although the PMA supplement pathway allows for expeditious device innovation that could improve patient care, the findings of Dubin et al suggest that when device alterations multiply over time without supporting clinical data, patient safety may be compromised,” they write.

Yet “these findings should not disincentivize device iteration through the PMA supplement pathway; indeed, some PMA supplements may make devices safer, in addition to more effective,” they note. “Instead, to ensure that device modifications approved through PMA supplements do not imperil patient safety, three important steps are needed.”

As a starting point, the agency could make sure companies submit any changes to their devices through the right type of PMA supplement, to ensure the appropriate data are collected. Moreover, the agency could set a threshold for how many supplements without clinical data can accumulate before the manufacturer must submit new safety information.

And finally, “while targeted requirements for evidence would help to reduce the chance that PMA supplements unintentionally introduce risk to patients, the third and more comprehensive solution is to strengthen the pathways for safety surveillance when medical devices are used in clinical practice,” they stress.