De Novo Pathway: Classifying a Device With No Predicate
A De Novo request asks FDA to classify a novel device into Class I or Class II when there is no legally marketed predicate to compare it to.
Without it, the default is unforgiving. A device that is not substantially equivalent to anything already on the market is automatically Class III by statute, which means PMA, which for a low-risk novel device is wildly disproportionate. De Novo exists to correct that default.
When it applies
Two conditions, both required:
- No predicate exists. Nothing legally marketed is substantially equivalent, so 510(k) is closed to you.
- The risk is low to moderate, and general controls, or general plus special controls, are sufficient to provide reasonable assurance of safety and effectiveness.
If the risk is genuinely high, De Novo is not the answer and PMA is.
The two routes in
Direct De Novo. Submit the request without filing a 510(k) first. This has been available since FDASIA in 2012 and is now the normal route, because filing a 510(k) you expect to fail wastes a cycle.
After an NSE determination. You filed a 510(k), FDA found the device not substantially equivalent, and you request De Novo classification within 30 days of that determination.
Most sponsors who know they lack a predicate go direct.
What the request contains
- A device description and its proposed intended use.
- Classification summary and recommendation: why Class I or II, with the reasoning.
- A benefit-risk discussion, including probable benefits and probable risks to health.
- Proposed special controls, if general controls alone are not sufficient. This is the part sponsors underestimate: you are proposing the regulatory framework that will govern your device and every future competitor.
- Supporting evidence: bench, animal and clinical data as the risks require.
- Labelling, and a summary of any prior submissions on the device.
FDA's MDUFA decision goal is 150 days, though the real elapsed time depends heavily on how many information requests the file attracts.
The consequence people miss
A granted De Novo creates a new classification regulation and product code. Your device becomes the predicate.
That cuts both ways, and the trade is worth thinking about before you file.
You get first-mover position and you shape the special controls, which means the testing and labelling every follower has to meet is the framework you proposed. Competitors then arrive by 510(k), which is cheaper and faster than what you just went through. Your De Novo effectively lowers the barrier for everyone behind you.
Proposing special controls that are genuinely appropriate rather than strategically onerous matters here, because FDA will not adopt controls that are not justified by the risks, and trying reads badly.
De Novo, 510(k) and PMA
| 510(k) | De Novo | PMA | |
|---|---|---|---|
| Requires a predicate | Yes | No | No |
| Resulting class | I or II | I or II | III |
| Basis | Substantial equivalence | Risk-based classification | Reasonable assurance of safety and effectiveness |
| Creates a predicate | No | Yes | No |
| FDA decision goal | 90 days | 150 days | 180 days |
Where De Novo requests go wrong
- Filing a 510(k) first, expecting it to fail. A wasted cycle when the direct route exists.
- Intended use written too broadly. It defines the classification and the special controls, so a broad claim invites a broader evidence burden.
- Special controls proposed without thought. They become the regulation. Vague ones get rewritten by FDA; onerous ones do not survive review.
- Assuming De Novo means light evidence. It means proportionate evidence. A novel device with real risks still needs data supporting the classification you are asking for.
- Treating it as a fallback. De Novo is a deliberate strategy decided early, not a place to land after a failed 510(k).
Frequently asked questions
What is the De Novo pathway?
A request under section 513(f)(2) asking FDA to classify a novel device into Class I or Class II when no legally marketed predicate exists, instead of the automatic Class III designation that would otherwise apply.
When should I use De Novo instead of a 510(k)?
When no predicate is substantially equivalent to your device. If a predicate exists, the 510(k) route is faster and cheaper.
Do I have to file a 510(k) first?
No. Direct De Novo has been available since FDASIA in 2012 and is the usual route. You may also request De Novo within 30 days of a not-substantially- equivalent determination on a 510(k).
How long does a De Novo take?
FDA's MDUFA decision goal is 150 days, though elapsed time depends on how many information requests the submission attracts.
What are special controls?
Requirements beyond general controls, such as performance standards, labelling requirements or specific testing, that provide reasonable assurance of safety and effectiveness. You propose them, and once granted they govern your device and every follower.
Does a De Novo create a predicate?
Yes. A granted De Novo establishes a new classification regulation and product code, so competitors can reach the market afterwards through a 510(k) citing your device.