Are Regulators the Root of the Problem of Serious Device Recalls?

Among cardiovascular devices that were deemed unsafe enough to warrant urgent regulatory action, many had been allowed to skip rigorous premarket clinical testing in the first place, a report showed.

Researchers identified 137 Class I recalls affecting 157 unique cardiovascular devices logged by the FDA from 2013 to 2022. Of these recalled devices, just 19.1% had undergone premarket clinical testing as part of their 510(k), premarket approval (PMA), or PMA supplement applications, reported Claudia See, MD, MBA, of the University of California San Francisco, and colleagues.

And when premarket testing was done, it usually involved studies using surrogate (79.4%) and composite measures (70.6%) for their primary endpoints.

“Even when premarket clinical testing was conducted, evidence was usually generated from a single nonrandomized, unblinded clinical trial using surrogate measures with no active control. In other words, medical devices later recalled due to safety issues often had little clinical evidence supporting their original authorization,” See’s group wrote in the Annals of Internal Medicine.

One example of the trouble with relying on surrogate endpoints is the case of the Vici venous stent system. Approved in 2019 based on the surrogate measure of vessel patency at 12 months, the Vici was later recalled in 2021 due to numerous reports of stent migration.

Short follow-up is another problem that often appears in the backstory of recalled cardiovascular devices, such as the Valiant Navion thoracic stent graft system, which was approved in 2018 based on 30-day data, and was subsequently linked to type IIIb endoleaks, stent fractures, and stent ring enlargement when used commercially.

“The public health implications of recalled cardiovascular devices were significant, with patients and clinicians usually advised to stop further device use. Given the centrality of medical devices in cardiovascular medicine, our study highlights important opportunities to advance patient safety,” the study authors wrote.

In an accompanying editorial, Ezekiel Emanuel, MD, PhD, a health policy expert and oncologist at the University of Pennsylvania in Philadelphia, took issue specifically with how different approval processes and assumptions of equivalency allow new medical devices — unlike medications — to slip into the market without strict proof of meeting standards.

“If the FDA applied the same logic it uses to regulate devices to drugs, it could be argued that one statin is substantially equivalent to another statin or one PD-1 inhibitor is substantially equivalent to another. But the FDA does not accept this chain of reasoning for drugs. Being a ‘me-too’ drug does not permit the skipping of phase I, II, and III clinical testing, while being a ‘me-too’ device usually does,” Emanuel wrote.

“For all its limitations, drug regulation is more standardized, has greater reliance on the accumulation of substantial, controlled clinical research, with equivalent evidentiary requirements for ‘me-too’ drugs. Devices need to be treated more like drugs and postmarket surveillance of both drugs and devices should become routine,” he stressed.

For their cross-sectional study, See and colleagues probed the FDA’s log of cardiovascular device recalls over a decade.

Of the devices with Class I recalls included in the study, 71.3% were moderate-risk 510(k) devices — considered “substantially equivalent to a device that is already on the market” — and 28.7% were high-risk PMA devices.

The recalls were attributed to device design (31.4%), process control (16.1%), or component design or selection (7.3%).

Each device recall affected a median 7,649 units, and 26.8% of these devices had multiple Class I recalls.

As of September 2023, 48.9% of PMA devices had their required post-approval studies done. No 510(k) devices were subject to postmarket surveillance.

See and colleagues acknowledged that the FDA database they used may have lacked important information, including details about clinical testing.

“If there are limitations of See and colleagues’ analysis it is that it does not go far enough. They focus on cardiovascular devices and Class 1 recalls while many other devices also warrant thorough study and Class 2 recalls are also problematic,” Emanuel wrote.

As for specific solutions, he suggested three fixes for the FDA: a shift in its attitude of how to balance innovation and safety, more emphasis on postmarket surveillance of devices, and requirements for better-quality clinical studies before devices go to market.

“Finally, Congress needs to act, too,” he added. “Like the FDA, Congress must emphasize safety along with innovation. In addition, legislation could require UDI [unique device identifier] integration into interoperable EMRs [electronic medical records] to ensure that device data are available wherever and whenever the patient seeks care.”