Medical device classification: how Class I, II, and III are assigned, and how a device type gets reclassified

In the United States, a medical device is not approved or cleared in a vacuum. It first has to be placed into one of three regulatory classes, and that class determines almost everything downstream: what controls apply, what premarket submission (if any) is required, and how much clinical and manufacturing evidence FDA expects. Class is a function of risk and intended use, not of the marketing name a company chooses. This article walks through how the three classes work in practice, illustrates each with real product codes, and shows how FDA moves an entire device type from one class to another, in both directions.

The three-class, risk-based framework

Every device that meets the statutory definition is assigned to a generic device type, each of which carries an FDA product code and a classification regulation (a citation in Title 21 of the CFR). The product code is the practical anchor: it fixes the class, the governing regulation, the medical specialty panel, and the expected submission type. The three classes escalate with risk:

• Class I (low risk). Subject to "general controls" only (registration and listing, labeling, adverse-event reporting, and quality system requirements). The large majority are exempt from premarket notification, so no 510(k) is filed before marketing.
• Class II (moderate risk). General controls are not sufficient to provide reasonable assurance of safety and effectiveness, so FDA adds special controls (for example, performance standards, specific labeling, or mandatory testing). Most Class II devices reach the market through a 510(k) premarket notification demonstrating substantial equivalence to a legally marketed predicate.
• Class III (high risk). Devices that support or sustain life, are of substantial importance in preventing impairment of human health, or present a potential unreasonable risk of illness or injury. General and special controls are deemed insufficient, so these devices generally require premarket approval (PMA), the most demanding pathway, supported by valid scientific evidence of safety and effectiveness.

The product code database makes this hierarchy concrete through its "submission type" field, which records the pathway FDA associates with each generic type (1 = 510(k), 2 = PMA, 4 = 510(k)-exempt).

Class I in practice

Low-risk, general-controls-only devices dominated by 510(k)-exempt product codes. For example, the non-surgical tongue depressor (product code FMA, 21 CFR 880.6230) is Class I and 510(k)-exempt (submission type 4), and is also flagged GMP-exempt ⁵. Common wound-care and support items follow the same pattern: the elastic bandage (FQM, 21 CFR 880.5075) is Class I and 510(k)-exempt ¹⁷⁰. Devices like these carry no implant or life-support flags, consistent with their low-risk status ⁵¹⁷⁰.

Class II in practice

Moderate-risk devices that clear through 510(k) under special controls. The general-purpose infusion pump (FRN, 21 CFR 880.5725) is Class II with a 510(k) submission type (1), as are its close relatives in the same regulation: infusion pump accessories (MRZ), ophthalmic (MRH), elastomeric (MEB), patient-controlled analgesia (MEA), and enteral (LZH) pumps ¹²⁶. The recurrence of one regulation number (880.5725) across a family of related product codes is typical: FDA groups technologically similar devices under a shared classification regulation while distinguishing them by product code.

Class III in practice

High-risk devices, typically implants or life-sustaining technologies, that require PMA. Coronary stents are the textbook case: the coronary drug-eluting stent (NIQ) and the bare coronary stent (MAF) are both Class III with a PMA submission type (2), and are flagged as implants; the drug-eluting stent additionally carries the life-sustain/support flag ¹⁷⁹¹⁸². The PMA designation reflects the judgment that neither general nor special controls alone can assure safety and effectiveness for these devices.

How an entire device type gets reclassified

Classification is not permanent. FDA can move a whole generic type up or down the risk ladder by administrative action, and it does so in both directions. Reclassification is distinct from the case-by-case novelty pathway (De Novo), although the two interact. Below are three concrete, database-verifiable examples that illustrate the mechanics.

Up-classification: transvaginal surgical mesh for pelvic organ prolapse

The clearest up-classification example is surgical mesh used to repair pelvic organ prolapse (POP) transvaginally. Today the mesh itself is Class III and requires PMA:

• Synthetic urogynecologic mesh for transvaginal POP repair (OTP, 21 CFR 884.5980) is Class III with a PMA submission type (2) ⁸⁹¹⁸³.
• The non-synthetic counterpart (PAI, 21 CFR 884.5980) is likewise Class III ⁹².

What makes this a reclassification story rather than a simple high-risk designation is the contrast with the tooling used to place the same mesh. The instrumentation for transvaginal POP mesh placement (PWI, 21 CFR 884.4910) remains Class II with a 510(k) submission type (1) ⁸⁸¹⁸⁴. In other words, FDA separated the surgical mesh (moved to the most stringent PMA pathway) from the delivery instruments (left in the moderate-risk 510(k) pathway), a targeted up-classification of the higher-risk component of a formerly lower-risk product category. The product code database reflects the end state of that action but does not itself narrate the effective dates of the underlying reclassification order, which a reviewer should confirm against the corresponding Federal Register final order ⁸⁹¹⁸³.

Down-classification: Medical Device Data Systems (MDDS)

Reclassification also runs downhill. A Medical Device Data System, defined as a device intended only for the electronic transfer, storage, conversion, or display of medical device data without controlling or altering the function of any connected device, is now classified as Class I (OUG, 21 CFR 880.6310), with a 510(k)-exempt submission type (4) ⁶²¹⁸⁶. The definition explicitly excludes systems intended for active patient monitoring, which keeps genuinely higher-risk functions out of this low-risk bucket ⁶². A device performing only passive data handling now sits in the lowest-risk class with no premarket submission, the down-classified end state for this technology type. As with the mesh example, the database shows the current Class I status but not the prior classification or the order that effected the change ⁶²¹⁸⁶.

High-risk designation with PMA: automated external defibrillators

Automated external defibrillators (AEDs) illustrate a life-critical type held to the PMA standard. The over-the-counter AED (NSA, 21 CFR 870.5310) is Class III with a PMA submission type (2); its definition describes a device that senses, detects, classifies, and treats ventricular fibrillation with an electrical shock in suspected sudden-cardiac-arrest victims ¹¹⁸⁵. A wearable AED variant (MVK) is likewise Class III ². AEDs are a useful reminder that even widely distributed, over-the-counter devices can sit in the highest class when the failure mode is death.

Where De Novo fits

The three-class system has a gap: a genuinely novel device with no legally marketed predicate is, by default, automatically placed in Class III, even when its actual risk is low or moderate. The De Novo pathway exists to correct that mismatch by granting a new, lower classification (Class I or Class II, with special controls) for such devices, which then becomes available as a predicate for future 510(k)s. This is a classification decision made for an individual novel device, as distinct from the administrative reclassification of an existing generic type described above. (The product code database used here catalogs existing generic types and their current class and pathway; it does not narrate the De Novo grant mechanism itself, so the process detail should be drawn from FDA's De Novo classification guidance and the governing statute.)

A note on terminology: device class vs. recall class

Two different "Class I/II/III" scales coexist in device regulation, and they are easy to conflate. This article addresses device classification (risk-based regulatory class governing market entry). A separate scale, recall classification, describes the severity of a specific recall event (Class I recalls involve a reasonable probability of serious adverse health consequences or death; Class II and III recalls are progressively less serious). A Class II device can be the subject of a Class I recall; the two labels answer different questions and should never be equated.

How other jurisdictions compare

The risk-based logic is broadly shared, but the number of tiers and the mechanics differ. Under the EU Medical Device Regulation (MDR), devices are divided into Class I, IIa, IIb, and III based on the device's intended purpose and inherent risks, applying the classification rules in Annex VIII ¹⁰⁹. The manufacturer must consider every applicable rule and sub-rule, and where more than one applies, the strictest rule producing the highest class governs the whole device ⁹⁷¹¹⁰. Intended purpose as specified by the manufacturer controls the outcome, not accidental use or how similar products are classified ¹¹⁰. The escalation mirrors the US logic: Class I is the lowest-risk default when no higher rule applies ^95107108; Class IIa and IIb capture moderate and higher-risk monitoring, diagnostic, and therapeutic functions ^106107116; and Class III applies where incorrect performance could cause death or irreversible deterioration of health ¹⁰⁶¹⁰⁷. The practical difference for a US-focused team is that the EU adds a fourth tier (splitting the US "Class II" band into IIa and IIb) and routes conformity assessment through notified bodies rather than an FDA premarket submission.

Practical takeaways for regulatory teams

• Start from the product code, not the marketing claim. The code fixes class, regulation, panel, and expected submission type in one lookup.
• The submission type on a product code is your fastest read on the pathway: 510(k)-exempt for most Class I, 510(k) for most Class II, PMA for Class III ^5126179.
• Classification is mutable. If your device type sits near a risk boundary, monitor FDA reclassification activity; a device type you rely on as a 510(k) predicate can be up-classified (as with transvaginal POP mesh) or down-classified (as with MDDS) ⁸⁹⁶².
• The product code database tells you the current state, not the history. When reclassification timing or the specific order matters (for legacy predicates, grandfathering, or transition provisions), confirm the effective date against the underlying Federal Register final order rather than inferring it from the current record ¹⁸³¹⁸⁶.