FDA 510(k) Predicate Best Practices: Recall History, Creep, and the 2026 QMSR Gate

Executive Summary

What does FDA's 'Best Practices for Selecting a Predicate Device' guidance require, how does the recall-history criterion limit predicate creep, and how does the 2026 QMSR intersect with 510(k) predicate selection?

In September 2023, the FDA released the draft guidance "Best Practices for Selecting a Predicate Device to Support a Premarket Notification [510(k)] Submission" (Docket FDA-2023-D-3134). It recommends four best practices: the predicate should be (1) cleared using well-established, scientifically accepted methods, (2) meet or exceed the expected safety and performance of predicates, (3) be without unmitigated use-related or design-related safety issues, and (4) be without an associated design-related recall. FDA recommends manufacturers describe in the 510(k) Summary how these best practices were used to select the predicate. The recall-history criterion directly targets "predicate creep" documented in the JAMA 2023 Kadakia study: of 127 devices that received Class I recalls in 2017-2021 with predicate data, 44.1% were cleared using a predicate that itself had a Class I recall, and devices built on such predicates had a 6.4-fold higher risk of their own future Class I recall. The guidance remains in draft, but it signals the criteria reviewers use when scrutinizing predicate choice. Separately, the Quality Management System Regulation (QMSR) took effect February 2, 2026, aligning 21 CFR Part 820 with ISO 13485:2016 and raising the quality-system baseline that inspectors now enforce under Compliance Program 7382.850 — making QMS maturity and design-control rigor a practical companion consideration to predicate selection, even though 510(k) substantial-equivalence review and QMS inspection remain distinct enforcement tracks.

---

The Evolution of the 510(k) Pathway and Predicate Modernization

The Premarket Notification [510(k)] pathway is the most common route for commercializing medical devices in the United States. Established under the Medical Device Amendments of 1976, the pathway requires a manufacturer to demonstrate that their new device is "substantially equivalent" to a legally marketed predicate device. Substantial equivalence means the new device has the same intended use and the same technological characteristics, or different technological characteristics that do not raise new questions of safety and effectiveness.

Over the past five decades, the complexity of medical devices has grown exponentially. Software algorithms, complex materials, and wireless connectivity are now standard features. However, the regulatory framework governing 510(k) clearances has remained structurally unchanged. This has led to the phenomenon of "predicate chaining"—where a new device is cleared by comparing it to a predicate cleared in 2015, which in turn was compared to a predicate cleared in 2005, tracing back to a pre-amendment device from 1975.

1975 Device (Pre-amendment)
       |
       v (Substantial Equivalence Comparison)
1995 Device (Incremental change)
       |
       v (Substantial Equivalence Comparison)
2015 Device (Software & wireless added)
       |
       v (Substantial Equivalence Comparison)
2026 Device (Subject of new 510(k) filing)

In 2018, the FDA's Center for Devices and Radiological Health (CDRH) proposed a controversial modernization initiative that included "sunsetting" or banning the use of predicates older than 10 years. This proposal faced fierce industry pushback. Industry groups argued that age does not correlate with safety, and that many older predicates represent well-established, highly reliable technologies.

Recognizing that a rigid age-based limit was politically and operationally unfeasible, the FDA shifted its focus toward qualitative selection criteria. This culminated in the September 2023 release of three interconnected draft guidances designed to strengthen the 510(k) program. Among these, the guidance "Best Practices for Selecting a Predicate Device to Support a Premarket Notification [510(k)] Submission" (Docket FDA-2023-D-3134) stands as the primary framework for choosing safe and effective predicates.

---

Verbatim: The Four Best Practices for Predicate Selection

Rather than focusing on the chronological age of a device, the FDA's draft guidance outlines four qualitative best practices. The agency recommends that manufacturers select a predicate device that meets as many of these criteria as possible.

Best Practice 1: Select a predicate device cleared using well-established, scientifically accepted methods.

The FDA prefers predicates where the substantial equivalence decision was supported by:

• FDA-Recognized Consensus Standards: The predicate was shown to comply with international standards (e.g., ISO 10993-1 for biocompatibility, IEC 60601-1 for electrical safety, or ISO 14971 for risk management) that are still recognized by the FDA.
• FDA Guidance Documents: The testing performed on the predicate aligned with specific, current FDA product-class guidance documents.
• Qualified Third-Party Review: The predicate underwent a thorough review by an accredited third-party review organization.
• Qualified Medical Device Development Tools (MDDTs): The testing utilized FDA-qualified models, biomarker assays, or clinical outcome assessments.

Selecting a predicate that utilized obsolete testing methodologies or non-standard protocols increases the risk that the reviewer will reject the comparison, forcing the manufacturer to conduct new, costly testing.

Best Practice 2: Select a predicate device that meets or exceeds the expected safety and performance of predicates.

A predicate should represent the current consensus of safety and performance in its product category. The FDA advises against selecting a predicate if:

• The predicate utilizes obsolete materials (e.g., certain plastics or coatings since shown to cause chronic inflammation or tissue reactions).
• The predicate lacks basic safety features that have since become standard in the industry (e.g., needle-stick protection on syringes or cybersecurity locks on connected infusion pumps).
• The predicate's clinical performance data represents a baseline that has been consistently outperformed by newer, legally marketed devices.

Best Practice 3: Select a predicate device without unmitigated use-related or design-related safety issues.

Even if a device has not been formally recalled, post-market surveillance data may reveal persistent safety issues. Manufacturers must search public databases (including the FDA's MAUDE and MedSun databases) to verify that the candidate predicate does not have a history of:

• High rates of adverse events associated with user interface flaws or labeling deficiencies.
• Repetitive design-related malfunctions (e.g., battery degradation, sensor drift, or connector failures).
• Unmitigated hazards identified in safety communications or public warning letters.

Best Practice 4: Select a predicate device without an associated design-related recall.

This is the most critical and objectively verifiable of the four best practices. The FDA explicitly recommends that manufacturers avoid using a predicate device that has been subject to a Class I recall or a design-related Class II recall. If a predicate has had its clearance suspended or if the product has been withdrawn from the market due to a design defect, using it to demonstrate the safety of a new device is a major regulatory risk.

---

What is Predicate Creep? The Kadakia JAMA 2023 Evidence

The recall-history criterion (Best Practice 4) was designed to combat a phenomenon known as "predicate creep." Predicate creep occurs when incremental, unreviewed technological changes accumulate across multiple generations of 510(k) clearances, eventually resulting in a cleared device that is fundamentally different—and potentially less safe—than the original predicate.

Baseline Device (Cleared 1990: Simple mechanical design)
       |
       v (+ Plastic component change - Cleared 1998)
Incremental Step 1 (Slightly higher friction)
       |
       v (+ Digital sensor added - Cleared 2007)
Incremental Step 2 (Battery heat generation)
       |
       v (+ Bluetooth transmitter - Cleared 2018)
Modern Device (Accumulated changes cause thermal degradation of plastic)

The Kadakia JAMA Study

The clinical and regulatory consequences of predicate creep were quantified in a landmark cross-sectional study published in JAMA in January 2023 by Kushal T. Kadakia, Dr. Harlan M. Krumholz, and colleagues ("Safety and Premarket Scientific Evidence of Predicate Devices for 510(k) Cleared Medical Devices Subject to Class I Recalls", doi:10.1001/jama.2022.23279).

The researchers analyzed 156 medical devices that received the FDA's highest-risk Class I recalls between 2017 and 2021. Of these, 127 devices had traceable predicate data available in public FDA databases. The findings were stark:

• High Recall Chaining (44.1%): Out of the 127 recalled devices, 56 (44.1%) were cleared using at least one predicate device that had itself been subject to a Class I recall prior to the new device's clearance.
• The 6.4x Risk Multiplier: Devices that were cleared based on predicates with a prior Class I recall had a 6.4-fold higher risk of subsequently being recalled themselves compared to devices cleared on recall-free predicates.
• Lack of Premarket Clinical Data: Across the 229 unique predicate devices supporting these recalled products, only 10 (4.4%) had undergone any form of premarket clinical testing. The remaining 95.6% were cleared entirely on laboratory and benchtop engineering data.
• High Patient Volume Impact: The recalled devices in the study represented more than 10,000 recalled units per index recall, affecting millions of patients undergoing cardiovascular, orthopedic, and general surgical procedures.

The Kadakia study proved that "safety deficiencies are transmissible through the 510(k) clearance process." If a manufacturer uses a flawed predicate, they are highly likely to inherit the same underlying design vulnerability, leading to a repeat of the same post-market clinical failures. The FDA's Best Practices guidance is a direct effort to break this chain of inherited defects.

---

How to Implement the Recall-History Criterion in Practice

To operationalize the recall-history criterion under Best Practice 4, regulatory affairs teams must establish a rigorous search protocol before selecting a predicate. This search must go beyond the standard FDA 510(k) database.

Step 1: Query the Medical Device Recalls Database

For every candidate predicate, the RA team must search the FDA's Medical Device Recalls Database using the predicate's 510(k) number (K-number), product code, or trade name.

• Analyze the Classification: Determine if any recalls were classified as Class I (reasonable probability of serious adverse health consequences or death) or Class II (temporary or medically reversible adverse health consequences).
• Identify the Root Cause: The FDA categorizes recalls by root cause (e.g., design, process control, component change, software error). Under the guidance, only design-related or use-related recalls act as a barrier to predicate selection. Recalls caused strictly by a one-off manufacturing deviation or a packaging error do not necessarily disqualify the predicate, provided the manufacturer can show the design itself is sound.

Step 2: Establish a Recall-History Decision Matrix

When selecting among multiple candidate predicates, use a structured scoring system to evaluate risk:

[Candidate Predicates]
         |
         v
Are there any Class I recalls?
  +-- YES: High Risk (Avoid / Pivot to alternative predicate)
  +-- NO:
       |
       v
Are there any Design-Related Class II recalls?
  +-- YES: Moderate Risk (Must document mitigations in 510(k) Summary)
  +-- NO: Low Risk (Preferred Predicate)

Predicate Status	Regulatory Risk Level	Action Required in 510(k) Submission
No Recalls / Clean Post-Market History	Low	Standard substantial equivalence comparison. Reference the clean record as evidence of safety.
Manufacturing-Related Class II Recall	Low-Moderate	Document the recall. Show that the new manufacturing process utilizes controls (e.g., under the new QMSR) that prevent the specific defect.
Design-Related Class II Recall	Moderate-High	Conduct a gap analysis. Document in the 510(k) Summary how the new design resolves or bypasses the defect that caused the predicate's recall.
Class I Recall (Any Cause)	High	Avoid. Selecting this device is likely to trigger an Additional Information (AI) request or a Not Substantially Equivalent (NSE) decision unless no other predicate exists.

---

Comparison: US Substantial Equivalence vs. EU Equivalence Pathway

To understand the unique positioning of the US FDA's 510(k) predicate-selection system, it is highly valuable to contrast it with the European Union's regulatory framework under the EU Medical Device Regulation (MDR 2017/745) and In Vitro Diagnostic Regulation (IVDR 2017/746).

While the US relies on "substantial equivalence" to a cleared predicate, Europe implements a much stricter "equivalence pathway" governed by MDCG 2020-5 (Clinical evaluation - Equivalence).

1. The Data Accessibility Barrier

In the United States, a 510(k) applicant does not need the consent of the predicate manufacturer, nor do they need access to the predicate's proprietary technical file. They rely entirely on public summaries, device labels, and comparative bench testing.

In contrast, under the EU MDR, to claim equivalence for Class III or implantable devices, the manufacturer must enter into a contractual agreement with the competitor (the equivalent device manufacturer) allowing them full access to the competitor's technical documentation on an ongoing basis. Because competitors almost never grant such access, the equivalence pathway under the EU MDR is effectively blocked for third parties.

2. The Definition of Equivalence

MDCG 2020-5 requires three criteria to be met simultaneously:

• Technical Equivalence: Same design, similar specifications, same deployment methods, and similar operating principles.
• Biological Equivalence: Same materials or substances in contact with the same human tissues or bodily fluids.
• Clinical Equivalence: Same clinical condition, same severity, same medical purpose, same site in the body, and similar user interface.

If there is any difference in materials or design, Notified Bodies in Europe will reject the claim of equivalence, forcing the manufacturer to conduct a full clinical investigation (clinical trial) to support their CE marking application.

3. Divergence in Product Lifecycle Speed

This difference creates a clear regulatory divergence. A Class II device that can be cleared in the US via a 510(k) in 90-180 days using a public predicate will often take 12-24 months in Europe, requiring full clinical trials or extensive, Notified Body-audited clinical evaluation reports (CERs) because the equivalence pathway cannot be utilized.

---

The 2026 QMSR Gate: Companion Compliance to Predicate Selection

On February 2, 2026, the FDA's Quality Management System Regulation (QMSR) final rule (89 FR 7496) took effect. The QMSR amends the legacy Quality System Regulation (21 CFR Part 820) by incorporating ISO 13485:2016 (Medical devices — Quality management systems — Requirements for regulatory purposes) by reference. On the same date, the FDA began inspecting manufacturers under the updated Compliance Program 7382.850, retiring the prior QSIT-based program (7382.845).

This transition represents a major operational shift, but it is important to understand its boundaries. The 510(k) substantial-equivalence determination and quality-system compliance are enforced through distinct tracks. A 510(k) clearance decision is based on substantial equivalence to a predicate; unlike a PMA, a 510(k) is generally not conditioned on a pre-clearance QMS inspection. QMS failures are instead enforced through inspections, Form 483 observations, Warning Letters, consent decrees, and — in serious cases — injunction or seizure.

Why QMSR Still Matters for Predicate Strategy

Even though QMSR does not add a formal "QMS pass/fail gate" inside 510(k) review, it raises the compliance baseline in ways that interact directly with predicate selection:

1. Heightened Design-Control and Risk-Management Scrutiny: ISO 13485 Clauses 7.1 (risk management) and 7.3 (design and development) are now the inspection yardstick. Selecting a predicate with a known design-related recall invites reviewers and inspectors to scrutinize whether your own risk-management file (Clause 7.1) and design inputs (Clause 7.3.3) actually incorporate the lessons of that recall.
2. Inspection Posture Feeds Enforcement: A history of predicate-chaining onto recalled devices can surface as an inspection finding under CP 7382.850, where poor design-control linkage becomes a documented compliance deficiency that can delay or disrupt commercialization independently of the 510(k) decision letter.
3. The Link to Best Practice 4: Best Practice 4 (avoid predicates with design-related recalls) is, in effect, the premarket mirror of the postmarket design-control rigor QMSR now demands. Manufacturers who ignore the recall-history criterion in selection often discover the same defect in their own CAPA and design-history files during inspection.

In short, QMSR does not withhold your 510(k) letter, but it makes a weak, recall-tolerant predicate-selection strategy far more expensive to defend under inspection. Predicate selection and quality-system maturity now reinforce each other across the device lifecycle.

---

How to Document Best-Practice Selection in the 510(k) Summary

The FDA's draft guidance recommends that manufacturers include a dedicated section in their 510(k) Summary describing how they applied the four best practices when selecting their predicate. This documentation serves as a proactive defense of the submission's integrity.

Outline of the Predicate-Selection Section

When drafting the 510(k) Summary, include the following structured elements:

1. Predicate Identification: List the trade name, manufacturer, and K-number of the primary predicate and any reference devices.
2. Methodology Alignment (Best Practice 1): Identify the FDA-recognized consensus standards or guidance documents that were utilized to clear the predicate, and confirm that the new device has been tested using these same (or updated) methods.
3. Safety and Performance Profile (Best Practice 2 & 3): Summarize the predicate's post-market safety history. Cite MAUDE database queries demonstrating that the predicate has not been subject to high rates of adverse events.
4. Recall Clearance Statement (Best Practice 4): State explicitly:

   "A search of the FDA Medical Device Recalls Database was conducted on [Date] for [Predicate Name/K-Number]. The search confirmed that the selected predicate device has not been subject to any Class I recalls or design-related Class II recalls. The substantial equivalence comparison presented in this submission is therefore supported by a predicate with a verified safety profile, minimizing the risk of inherited design defects."

5. Mitigation Gap Analysis (If a recall exists): If a manufacturer must use a predicate with a recall history (e.g., because it is the only predicate available in a highly niche product code), they must provide a detailed table showing:
   • The recall number and date.
   • The root cause of the recall.
   • The specific design or manufacturing changes implemented in the new device to prevent the recall's recurrence.

---

Frequently Asked Questions (FAQ)

Is the FDA best-practices guidance for selecting a predicate device binding or just a recommendation?

As a draft guidance, it represents the FDA's current thinking but is not legally binding. However, in practice, FDA reviewers are actively using these criteria as a framework for their reviews. Failing to address these best practices in your 510(k) Summary increases the risk of receiving an Additional Information (AI) request.

What does the JAMA 2023 Kadakia study say about recalled predicates and 510(k) devices?

The study found that 44.1% of Class I-recalled 510(k) devices were cleared using a predicate that itself had a Class I recall. Furthermore, devices cleared using a recalled predicate had a 6.4-fold higher risk of being recalled in the future.

Can a predicate that had a Class I recall still be used in a new 510(k)?

Technically, yes, because the guidance is not binding. However, doing so is highly discouraged. If you must use a recalled predicate, you must conduct a detailed design gap analysis and document in your 510(k) Summary exactly how your new design mitigates the vulnerability that caused the predicate's recall.

How does the 2026 QMSR affect 510(k) clearance?

The QMSR (effective Feb 2, 2026) aligns 21 CFR Part 820 with ISO 13485:2016 and is enforced through inspections under Compliance Program 7382.850, not through a new pass/fail step inside the 510(k) substantial-equivalence review. It does not give reviewers a routine authority to withhold a 510(k) clearance based on QMS findings. What it does is raise the design-control and risk-management baseline inspectors now expect — so a predicate-selection strategy that ignores recall history becomes much harder to defend at inspection and across the device lifecycle.

How does the FDA define "design-related" versus "manufacturing-related" recalls?

A design-related recall stems from a fundamental error in the device's specifications, software code, or material composition (affecting all units produced). A manufacturing-related recall is caused by a deviation from those specifications during assembly, packaging, or sterilization (affecting only specific lots).

---

Sources

1. US FDA Draft Guidance Documents:
   • "Best Practices for Selecting a Predicate Device to Support a Premarket Notification [510(k)] Submission": Docket FDA-2023-D-3134 (September 7, 2023).
2. Peer-Reviewed Literature:
   • Kadakia KT, Krumholz HM, et al. "Safety and Premarket Scientific Evidence of Predicate Devices for 510(k) Cleared Medical Devices Subject to Class I Recalls". JAMA. 2023;329(2):136-143. doi:10.1001/jama.2022.23279.
3. FDA Quality Management System Regulation (QMSR) Final Rule:
   • Federal Register: Medical Devices; Quality System Regulation Amendments (89 FR 7496, Effective February 2, 2026).
4. Legal & Industry Commentary:
   • Covington & Burling LLP: FDA Proposes Significant Shift to 510(k) Process in Draft Guidance on Best Practices for 510(k) Predicate Selection (September 18, 2023).
   • RegDesk: Understanding Predicate Devices in 510(k) Submissions (2026).