Practice·September 7, 2026 · 4 min read

The K-number doesn't track today's device

21 CFR 807.81(a)(3) lets a manufacturer decide, in its own quality records, that a design or manufacturing change didn't rise to needing a new 510(k) — a judgment FDA never reviews before an inspection asks for it, and the public clearance record never shows either way.

By Connor Griggs — Regulatory & Quality Strategist

A 510(k) clearance record reads like a fixed description of a device: a K-number, an applicant, a product code, a decision date. It is tempting to read it as current — as if the device sold under that number today is the device FDA reviewed on the date the letter was signed. For a device that has been on the market for years, that reading is often wrong, and it is wrong for a reason the regulation states on purpose, not by accident.

The threshold the rule actually sets

21 CFR 807.81(a)(3) requires a new premarket notification only when a device already in commercial distribution is about to change in one of two specific ways: a change or modification that could significantly affect the device’s safety or effectiveness — design, material, chemical composition, energy source, or manufacturing process are the rule’s own examples — or a major change or modification in the device’s intended use. Anything a manufacturer changes that stays under both of those bars triggers nothing. No new K-number, no new decision date, no new record of any kind.

The decision that never leaves the building

Deciding which side of that threshold a specific change falls on is the manufacturer’s own call, made under a structured, risk-based framework FDA laid out in its 2017 final guidance on deciding when to submit a 510(k) for a device change — itself a replacement for guidance FDA had run since 1997. When the firm’s own analysis concludes a new 510(k) isn’t required, standard practice is to write that conclusion down: a letter to file, sometimes called a memo to file, recording what changed, why, and the reasoning that kept it under the regulatory threshold. FDA does not receive that document when it is written. 21 CFR 820.35 governs the records a device manufacturer’s quality system has to control and keep available — readily accessible for FDA review, which is a requirement that a letter to file exist and hold up, not a requirement that it be filed with anyone in advance.

A 510(k) clearance describes the device FDA reviewed on the day the letter was signed. It says nothing about whether that description still matches what shipped last week — only that if it doesn’t, the gap stayed under a threshold the manufacturer measured itself.

What the public record was never built to carry

The 510(k) database entry for a given K-number doesn’t change after clearance. It carries no field for a subsequent letter to file, no change log, no count of how many times a firm has run this analysis since the decision date. Two devices selling under the same K-number today — one whose manufacturer has documented five design or manufacturing changes since clearance, one that hasn’t touched the device at all — look identical from the public side. The database was built to record a clearance decision, not to track a device’s condition afterward.

The practice

This cuts two ways for a regulatory or quality operator. On your own device, the letter to file is the document an inspector asks for by name during a design-control or CAPA review — §820.35 requires that it exist and be available, which means the rationale has to be written and defensible at the time of the change, not reconstructed afterward under questioning. On a competitor’s device, or on a predicate you’re citing in your own submission, reading an old K-number as a live specification is the mistake this rule exists to explain: the clearance is a floor the device met once, not a running record of what it is today. FDA Radar’s own 510(k) corpus mirrors the clearance record exactly as openFDA publishes it — K-number, applicant, product code, decision date — because that is the public record; a device’s later, internal change history was never going to be in it, because no source publishes that.

Whether a specific change to a specific device crosses the 807.81(a)(3) threshold is a risk-based judgment for that device’s own manufacturer to make and document, under FDA’s guidance and their own regulatory professional’s reasoning — regulatory intelligence and method, never a determination this post makes for anyone’s specific device.

Method
The lab is accredited. Not the device.
Practice
Breakthrough is law. STeP is a guidance.
Practice
MDSAP replaces routine. Not for-cause.
Method
No field on a 510(k) says what it treats
Method
Discretion is a policy. Exclusion is a statute.
Practice
Designated by one office. Approved by another.
Method
MedSun sits beside MAUDE, not inside it
Practice
The five-unit device never gets a K-number
Practice
Meeting the criteria isn't a tracking order
Method
The PMA clock doesn't stop at approval
Method
A MAUDE report proves it was filed
Practice
Not every MDR starts with the manufacturer
Method
Completed. Not terminated.
Practice
The device-ban list has three entries
Method
The UDI exception has its own deadline
Method
MDUFA VI is being negotiated in public
Practice
The inspection manual changed, not the rule
Method
The count Congress had to legislate
Practice
The small-business rate isn't retroactive
Method
A detention order has no docket to read
Practice
PMA approvals moved off the Federal Register
Method
A citizen petition names names
Practice
A market withdrawal isn't a recall
Method
Not every device letter comes from CDRH
Practice
Closed on one device. Open on the rest.
Method
No IDE is confirmed until the PMA is
Practice
The remedial-action call sets the clock
Method
Jurisdiction is a letter, not a field
Practice
An address is not a footprint
Method
The final rule that isn't final yet
Practice
The special controls aren't in the record
Method
Exempt from GMP. Not from complaints.
Practice
The notify list isn't the shortage list
Method
A constraint can't be skipped. A script can.
Practice
No cron sends this. A person does.
Method
Corrected, not rewritten
Practice
Quiet is a finding, not a default
Method
The summary is public. The statement isn't.
Practice
A competitor's 513(g) leaves no trace
Practice
An empty response field proves nothing
Method
A petition is public. Its outcome isn't.
Practice
The panel is a category, not a meeting
Method
An accessory carries its own classification
Method
A TPLC report is a category, not a device
Practice
A 522 order is a question, not a verdict
Method
A PCCP change leaves no new record
Practice
Not every signal becomes a recall
Method
eSTAR did not remove the acceptance check
Practice
An early alert has no recall number yet
Practice
A consent decree isn't a database row
Method
The record has no cybersecurity field
Practice
Classification Change is not one event
Method
Publication date is not effective date
Method
One event. Many recall numbers.
Practice
Quantity is a sentence, not a number
Practice
How a firm told you is a field too
Method
A RIN outlives the document number
Method
The docket ID is not the comments
Practice
The 510(k) record doesn't name the predicate
Practice
The 510(k) FDA didn't review first
Method
The letter FDA doesn't have to post
Practice
Cleared doesn't say which track it took
Method
The clock stops. The calendar doesn't.
Method
Why a guessed match never reaches you
Practice
The standards list never names a device
Method
The guidance isn't in the Federal Register.
Practice
Filtered. Not discarded.
Method
There are two recall databases, not one
Practice
Breakthrough status is confidential by law
Practice
One MAUDE report is not one malfunction
Method
A supplement number is not a track
Method
Cleared is not approved
Practice
Nationwide doesn't mean your building
Practice
Not every correction reaches FDA
Method
Registered. Not reviewed.
Method
Contact ODE is not a pathway
Practice
Almost no recall is FDA-mandated
Method
FDA's device names read like a card catalog
Practice
openFDA is free. It is not unlimited.
Method
GUDID stores the model. Not the box.
Practice
Detention doesn't need a warning letter
Method
A Class I device, a Class I recall
Practice
21 CFR 820 didn't move. Its contents did.
Method
The product code that doesn't exist yet
Practice
The classification posts. The 483 behind it doesn't.
Method
A recall has three dates, and the pipeline had to pick one
Practice
The count is real. The rate is not.
Method
The firm on the record is not the firm on the box
Practice
Low confidence is an instruction, not a hedge
Method
The same company, spelled three ways
Practice
A device that was never a medical device
Method
FDA's warning letters, addressed by column number
Practice
Your regulation has a decimal. FDA's watch doesn't.
Method
Three letters is too short to search for
Practice
Most warning letters never close
Method
The guidance that skipped the draft
Practice
Ongoing, as of when?
Method
The least interesting fact in a 510(k)
Practice
No recall arrives with a product code attached
Practice
The deadline that doesn't email you
Method
The warning letter has two dates
Method
How to monitor FDA without drowning
Practice
Your predicate was recalled. Now what?
Method
Why no item reaches you without a human