Reading Between the K-Numbers: How to Evaluate FDA 510(k) Clearances During Procurement

A field guide for biomedical engineering teams and procurement officers who need to separate genuine regulatory rigor from substantial-equivalence theater.

Why this matters (specific scenarios)

A 510(k) clearance letter on a vendor's spec sheet is not what most clinicians and many procurement officers think it is. Between 95 and 98 percent of medical devices on sale in the U.S. were cleared by the FDA through the 510(k) process , and a 510(K) is a premarket submission made to FDA to demonstrate that the device to be marketed is as safe and effective, that is, substantially equivalent, to a legally marketed device — not a finding that the device itself is safe and effective in any absolute sense.

That distinction has real procurement consequences:

An infusion pump RFP: Two Class II pumps both carry 510(k) clearance, but one cites a predicate that was subject to a Class I recall while still on the market. In a matched analysis, medical devices approved based on a predicate device that was subject to a class 1 recall were more than six times more likely to also undergo a class 1 recall when compared with devices approved based on recall-free predicates (RR 6.40; 95% CI 3.59 to 11.40).
A surgical robot evaluation: The vendor's K-number traces back through a chain of incremental clearances ("predicate creep"), where each step was a small modification — but the cumulative drift between the device you're buying and the original predicate is substantial.
An AI-enabled diagnostic: The 510(k) summary shows no clinical data. Just 4.4% of predicate devices had undergone any premarket clinical testing.

If your evaluation stops at "FDA-cleared, K######, Class II," you are missing the signal that matters.

The decisions that shape the outcome

1. Pull the actual 510(k) record, not the marketing summary

Every cleared device has a K-number searchable in the FDA's Premarket Notification database at accessdata.fda.gov. Search 174,000+ FDA 510(k) medical device clearances from 1976 to today. Start by entering a K-number, company, or device type to explore approvals and market activity. For each shortlisted device, retrieve:

The 510(k) Summary or Statement (the manufacturer's narrative justifying SE)
The Indications for Use statement (this is the legally-binding scope)
The product code and regulation number (e.g., 21 CFR 880.5725 for infusion pumps)
The decision letter identifying the primary predicate

Note a critical caveat: The 510(k) database contains original application information only. That is, the 510(k) database maintains the name of the original applicant and the original trade name provided in the 510(k). It is not updated to reflect the current owner or distributor of the product nor any changes in trade name. If you're buying from a distributor, verify the manufacturer-of-record matches.

2. Verify intended use against your clinical workflow

Whether your device and predicate have identical intended use. This includes indications for use, patient population, and general device purpose. Different intended use = automatic NSE (not substantially equivalent). The reverse is also true for buyers: if your clinicians plan to use the device outside its cleared Indications for Use, you are entering off-label territory, with attendant liability and reimbursement consequences. Compare the cleared IFU against your protocol, patient population, anatomical site, and duration of use line by line.

3. Trace the predicate lineage

This is the step most procurement teams skip. Pull the predicate's K-number, then the predicate's predicate, going back at least 2–3 generations. You're looking for:

Recalled predicates: By law, the FDA does not allow previously approved predicate devices to form the basis of a new 510(k) clearance if that device has been subjected to a mandatory FDA-initiated recall. The problem is that most devices are not recalled by the FDA itself. "We can count on less than one hand the number of times the FDA has issued a voluntary recall of a medical device," said Kadakia. "The dominant paradigm is that manufacturers usually issue recalls. Within that 510(k) pathway, there is no prohibition on manufacturer-initiated recalls of devices and those devices being used as predicates for future devices." Cross-reference each predicate against the FDA's Medical Device Recalls database.
Predicate creep: The risk of predicate creep, a cycle of technology change through repeated clearance of devices based on predicates with slightly different technological characteristics, such as materials and power sources, or have indications for different anatomical sites. A device whose technology has drifted significantly from its original predicate may carry untested risks.
Anatomically inconsistent predicates: Sometimes the predicate device is not even used in the same part of the body as the new cleared device. An example is the SiBone (iFuse) system which is used at the sacro-ileal region of the lower back and has been attributed to several deaths in the FDA's adverse event database (called MAUDE). The predicate device was used in the foot.

4. Check MAUDE and the recall database for both subject and predicate

FDA recommends searching the Manufacturer and User Facility Device Experience ("MAUDE") database, which contains medical device reports ("MDRs"), and the MedSun Reports database, which contains safety reports provided by certain clinical sites, for any events involving the predicate device. Procurement teams should do the same. Filter by product code, not just brand name, to capture similar devices. Note MAUDE's caveats: "Although MDRs are a valuable source of information, this passive surveillance system has limitations, including the potential submission of incomplete, inaccurate, untimely, unverified, or biased data."

5. Identify the submission type and look for clinical data

Different types of 510(k) • Traditional: Submission to FDA according to 21 CFR Part 807, Manufacturer established substantial equivalence with an available predicate device • Abbreviated: manufacturer may opt when their submission is based on FDA guidance documents and demonstration of compliance with special controls for the device or voluntary FDA consensus standards • Special: Manufacturer whose device is already in the U.S. market require a change or intent to modify Labelling, Design, manufacturing process while ensuring that safety and performance are acceptable

Be aware that Traditional 510(k) were most common for both 510(k)s affected and unaffected by recalls. Special 510(k)s represented a higher percentage of 510(k) type among recalled 510(k)s than among 510(k)s unaffected by recall (34.2% vs 22.3%). Also check whether the submission included clinical data. When the new device's indications for use differ from the predicate device. When the technological characteristics of the new device vary from the predicate device. When substantial equivalence can't be established through non-clinical testing. When newly discovered or escalated risks associated with the predicate device indicate a need for clinical data for the new device. Absence of clinical data isn't disqualifying, but for implants, software-driven diagnostics, and life-sustaining equipment, ask the vendor directly why bench testing was deemed sufficient.

6. Verify consensus standards conformance

For Class II devices, Class II or moderate-risk devices such as magnetic resonance imaging (MRI) scanners or intravenous medication infusion pumps necessitate pre-market notification and require special controls such as meeting FDA-recognized performance standards or certain post-market surveillance requirements. Demand evidence of conformance to applicable FDA-recognized consensus standards — IEC 60601-1 (general electrical safety), IEC 60601-1-2 (EMC), ISO 14971 (risk management), IEC 62304 (medical device software), ISO 10993 (biocompatibility). When a sponsor is selecting between two similar valid predicate devices, where one identified performing testing using FDA guidance and recognized voluntary consensus standards while the other did not, FDA recommends the sponsor select the valid predicate device in which testing relied on FDA standards. The same logic should drive your vendor selection.

7. Confirm QMS posture for 2026 and beyond

Additionally, the new Quality Management System Regulation (QMSR), effective February 2, 2026, replaces the decades-old QSR (21 CFR 820) and aligns U.S. requirements with ISO 13485:2016. Vendors with mature ISO 13485 certification are better positioned. Also note the manufacturer should be prepared for an FDA quality system (21 CFR 820) inspection at any time after 510(k) clearance. Ask for the most recent FDA Form 483 history.

Common mistakes

Treating "510(k) cleared" as equivalent to "FDA approved." Technically, no. 510(k) devices are "Cleared." PMAs are "Approved." It's a semantic difference that marketing teams often get wrong. Reject vendor materials that conflate the two; it suggests sloppy regulatory communication.
Assuming clinical evidence exists because the device is cleared. More than 10% of devices were designated life-sustaining or support and 7.1% were implanted devices. As part of the 510(k) application, just nine devices included any clinical data, with 119 products not undergoing any clinical testing. If your use case demands evidence beyond bench data, request published peer-reviewed studies separately.
Ignoring Class III devices cleared via 510(k). Notably, until FDA issues a PMA requirement regulation, even some Class III device types can be cleared by 510(k) if predicates exist. GAO found that from 2003–2007, FDA cleared 228 out of 342 (67%) Class III 510(k) submissions, even though class III was intended to be PMA-only.
Not differentiating "indications for use" from "intended use." Marketing materials often blur these. The cleared IFU is the legally binding scope; promotional intended-use claims may exceed it.
Skipping the recall database for the predicate. Buyers focus on the device itself. The supply-chain risk often lives upstream.
Using the database as a one-time check. Updated monthly from FDA records. Browse clearance reports by specialty, review time and manufacturer — re-check before contract execution and at major contract milestones, since post-market actions accumulate.

A practical workflow / checklist

For each finalist device on a procurement shortlist:

Locate the K-number in the FDA 510(k) database. Pull the decision letter, 510(k) Summary, and Indications for Use.
Confirm the device classification and product code against 21 CFR Parts 862–892. Identify any applicable special controls.
Map the predicate lineage at least three generations deep. Document each K-number.
Run each K-number through the FDA Medical Device Recalls database. Flag any Class I or Class II recalls — open or closed.
Search MAUDE by product code (not just brand) for the past 36 months. Quantify event types.
Compare cleared IFU to your clinical use case verbatim. Document any gaps.
Request from the vendor: list of recognized consensus standards conformed to; biocompatibility test reports (if applicable); software level of concern documentation per IEC 62304; cybersecurity SBOM (per FDA's premarket cybersecurity guidance for connected devices).
Verify ISO 13485 certification and most recent FDA inspection outcome (Form 483, EIR, or warning letters via FDA's enforcement database).
Identify whether submission was Traditional, Special, or Abbreviated. Special submissions modifying a vendor's own device should be probed for what was modified.
Document realistic timeline expectations: An analysis of ~25,000 510(k)s from 2010–2021 reported an average of ~164 FDA days between submission and decision (counting all calendar time including pauses, which exceeds the MDUFA clock). Another analysis found similar averages (~170 days). FDA's own performance reports strive to meet the 90-day target, but variability exists by product type and complexity. For Class II devices with robust acceptance documentation, many 510(k)s do clear near the 90-day mark, whereas more complex devices with multiple deficiency cycles can stretch longer. Useful for evaluating vendor pipeline claims about next-gen products.

Edge cases worth flagging

Pre-amendment "grandfathered" devices. Pre-amendment devices, those approved pre-1976 before the classification of 510(k) and PMA was established, also play a role in the safety gaps of the 510(k) pathway, as many of those devices were grandfathered, and therefore do not have to undergo the approval process. Unfortunately, some pre-amendment devices have been put into the most severe Class I recalls. If a predicate chain terminates in a pre-1976 device, the original safety baseline may never have been formally established.
De Novo-derived predicates. Following a successful De Novo classification, competitors can then use this device as a predicate and obtain market clearance for their "substantially equivalent" device through the simpler, faster, and cheaper 510(k) application process, without (necessarily) having to provide the same evidentiary depth. The original De Novo holder typically did the heavy clinical lifting; subsequent 510(k) followers may not have.
AI/ML-enabled devices. Since 2018, FDA has issued draft guidances and initiated programs (e.g. the Safety and Performance Based Pathway, expanded Breakthrough Device designations, enhanced cybersecurity requirements, and Predetermined Change Control Plans for AI/ML devices) to update 510(k) review for advanced/complex technologies. Ask whether the vendor has a Predetermined Change Control Plan filed; this affects how algorithm updates will be handled post-purchase.
Devices cleared while a predicate's recall was active. Predicates had low rates of premarket clinical testing (4.4%) and were still used as predicates even after safety concerns arose in the postmarket setting (28.8%), including 28 cases for which devices were authorized while the predicate's Class I recall was still ongoing. This is a specific red flag: cross-check predicate K-number recall dates against your subject device clearance date.
Third-party reviewed submissions. Compared to 510(k)s unaffected by recall, recalled 510(k)s are more likely to have been reviewed by a third party or submitted as a Special application (rather than Traditional or Abbreviated). Not disqualifying, but worth a closer look at supporting documentation.
Distributor relabeling. You distribute another firm's domestically manufactured