Why FDA’s Draft 510(k) Guidance is Forcing Sponsors to Rethink Device Submissions

The device team thought they were following the playbook.

They chose a predicate that had been cleared two decades ago, crafted their submission, and waited. Instead, the FDA came back with new questions: Why this predicate? Where is the data?

For years, sponsors leaned on the flexibility of the 510(k) process to bring devices to market quickly. But today, the FDA is raising the bar.

The Scale of the Problem

The 510(k) pathway remains the backbone of device regulation. Each year, roughly 3,000 devices enter the U.S. market this way, making it by far the most common route for clearance. The principle is simple: if your device is “substantially equivalent” to a legally marketed predicate, it can move forward without the burdens of a Premarket Approval (PMA).

But cracks in the system have long been evident. Many devices still trace their lineage back to predicates cleared in the 1980s or 1990s, when standards for biocompatibility, cybersecurity, and clinical testing were far less rigorous. That reliance on outdated benchmarks has fueled concern that devices cleared today may not reflect modern safety expectations.

At the same time, data show that novelty carries its own risks. A JAMA Network Open study examining recalls between 2008 and 2017 found that although PMA devices represented only 6% of device approvals, they accounted for 41% of the most serious Class I recalls — making them more than seven times as likely as 510(k) devices to be pulled from the market for high-risk safety concerns. The explanation is not that PMA review is inadequate, but that PMA devices are Class III, high-risk technologies — such as heart valves, implantable defibrillators, and metal-on-metal hips — where failures are more severe and often harder to anticipate despite rigorous premarket trials.

The takeaway is twofold: 510(k) dominates in volume and can propagate outdated standards, while PMA devices, though fewer, highlight the hazards of high-risk innovation. Both trends underscore FDA’s push to modernize the 510(k) program: patients deserve devices that are safe and effective by today’s benchmarks, not yesterday’s.

FDA’s New Playbook

In September 2023, the FDA’s Center for Devices and Radiological Health (CDRH) released three draft guidances to address these concerns head-on:

1. Best Practices for Selecting a Predicate Device
2. Recommendations for the Use of Clinical Data in 510(k) Submissions
3. Evidentiary Expectations for 510(k) Implant Devices

These are not minor tweaks. Together, they aim to make the program more predictable, consistent, and transparent.

As Jeff Shuren, M.D., J.D., director of CDRH, explained, “We intend to finalize these three guidances and use other opportunities to continue to strengthen and modernize other aspects of the 510(k) Program to help deliver safe, effective and high-quality medical devices to patients.”

From Any Predicate to the Right Predicate

For decades, sponsors assumed that almost any legally marketed device could serve as a predicate. That approach often delivered speed, but it also created blind spots. Surgical mesh is a case in point: for years, manufacturers relied on predicates cleared decades earlier, despite mounting reports of erosion and chronic pain. The reliance on outdated benchmarks allowed risk to slip through the system until the FDA finally reclassified certain meshes into a higher-risk category.

The new draft guidance is intended to prevent a repeat of that history. Rather than simply pointing to an old device, the FDA now expects companies to demonstrate that their chosen predicate was cleared using modern test methods, has not been associated with adverse events or safety signals, and continues to reflect current standards of care. For a cardiovascular stent developer, this might mean selecting a predicate that has undergone the latest biocompatibility and fatigue testing, rather than one cleared before these standards were in place. The message is straightforward: choose wisely, and be ready to defend your choice.

When Clinical Data Is No Longer Optional

Traditionally, most 510(k) submissions advanced on bench data alone. Clinical studies were the exception, reserved for unusual circumstances. But the agency is now clear that there are situations where human evidence must be part of the package.

The COVID-19 pandemic illustrated why. Several diagnostics that looked solid in the lab later revealed troubling performance gaps once deployed at scale in diverse patient populations. Orthopedics provides another sobering lesson: metal-on-metal hip implants sailed through bench wear testing but failed catastrophically in patients, releasing metal ions and requiring thousands of revision surgeries. Those episodes proved that bench data can only go so far.

In its new guidance, the FDA highlights situations where clinical data should be anticipated from the outset: when indications for use differ from those of the predicate, when new materials are introduced, when non-clinical testing can’t capture performance, or when safety signals already exist in related technologies. Orthopedic implants again illustrate the point. A spinal cage may pass compression testing, but if the device introduces new risks of migration or subsidence, the FDA now expects sponsors to provide clinical evidence, not just laboratory data.

Raising the Bar for Implants

No category brings the stakes into sharper relief than implants. These are devices intended to remain in the body for months or years, and when they fail, the consequences can be life-threatening. Past insulation failures in ICD leads, for example, triggered dangerous malfunctions that could not be remedied without invasive surgery. Even resorbable orthopedic screws carry hidden risks: while initial fixation may be strong, degradation over time must be shown to align with bone healing.

The new implant guidance makes these expectations explicit. It calls for worst-case fatigue testing, long-term degradation and biocompatibility studies, and labeling that includes implant cards. Hence, patients are aware of precisely what is inside their bodies and how to manage it. For sponsors, the implication is clear: the days of limited testing paired with outdated predicates are over.

Case in Point: Orthopedics

Spinal devices offer a concrete illustration of this shift. Not long ago, a new cage could be cleared based on a 1990s predicate and some basic compression data. That strategy is unlikely to succeed today. A sponsor now needs to ground their submission in a modern predicate, demonstrate biomechanical performance under worst-case conditions, and, in some cases, provide registry or follow-up data to prove that the device performs safely in patients.

The burden may be greater, but it is also more predictable. Instead of being blindsided by a mid-review request for additional information, sponsors can anticipate the FDA’s expectations from the outset — and avoid the costly delays that can stretch a “90-day review” into a year-long ordeal.

Closing Insight

For sponsors, these guidances are not about adding red tape. They are about trading shortcuts for predictability — front-loading evidence rather than risking crippling delays once the clock is running. For patients, they signal a cultural shift: substantial equivalence no longer means “good enough by yesterday’s standards,” but safe and effective by today’s.

The 510(k) pathway is evolving from a backward-looking comparison to a forward-looking validation. That difference isn’t academic. It’s transformative.