A biosensor is a measuring principle, not a regulatory category. The same electrochemical cell can sit inside a fitness patch, a continuous glucose monitor or a hospital troponin analyser, and each of those products lives under different rules.

What decides which rules apply is rarely the chemistry. It is what the manufacturer says the product is for — on the label, in the instructions, and on the website.

That has always been true. In 2026 it became more visible. The US Food and Drug Administration rewrote its wellness policy in January and, in doing so, opened a route for non-invasive wearables that estimate blood pressure or glucose. In Europe, a proposal to revise the MDR and IVDR is working its way through Parliament and Council and would change how much clinical data some devices need. Neither development changes the first question every biosensor company has to answer: what exactly are you claiming?

This article goes family by family through the main types of biosensor, and for each one sets out the framework, the typical class, and the claim that most often moves a product into a stricter category. Part 2, next week, covers the evidence each family needs to reach the market. For the regulatory pathways themselves, see our biosensor regulatory pathway guide.

1. The First Decision: Wellness, Medical Device or In Vitro Diagnostic

In the EU, the intended purpose is whatever you say it is — everywhere

Under Article 2(12) of the MDR, the intended purpose is “the use for which a device is intended according to the data supplied by the manufacturer on the label, in the instructions for use or in promotional or sales materials or statements and as specified by the manufacturer in the clinical evaluation.”

Promotional materials are named explicitly. Your website, app store listing, sales deck and social media posts are part of your intended purpose.

Article 7 then prohibits text, names, trademarks and images that “may mislead the user or the patient with regard to the device’s intended purpose, safety and performance”, including suggesting uses beyond those covered by the conformity assessment.

There is no wellness exemption anywhere in the articles of the MDR or the IVDR. The closest thing is Recital 19, which is not binding and applies to software only: “software intended for life-style and well-being purposes is not a medical device.” MDCG 2019-11 Rev.1 says the same for wellness and fitness apps (see the MDSW qualification test). A product is outside the regulation because its stated purpose is not medical — not because it belongs to a category that is exempt.

MDR or IVDR: where the measurement happens

The IVDR covers devices “intended by the manufacturer to be used in vitro for the examination of specimens … derived from the human body” (Article 2(2)). The dividing line is whether the sample has left the body before it is analysed.

For some biosensors this line is obvious. For others — sweat, tears, breath — it is not, and we come back to them in section 4.

In the US, the January 2026 wellness guidance

On 6 January 2026 the FDA issued a revised version of General Wellness: Policy for Low Risk Devices, replacing the 2019 version. For the first time it addresses products that “use non-invasive sensing (e.g. optical sensing) to estimate, infer, or output physiologic parameters (e.g. blood pressure, oxygen saturation, blood glucose, heart rate variability).”

Such products may be treated as wellness products when their outputs are intended solely for wellness uses and they:

They may display “values, ranges, trends, baselines, or longitudinal summaries.” They fall back into device regulation if their labelling, advertising or interface uses disease names or diagnostic thresholds, issues alerts that recommend clinical action, gives treatment guidance, or claims “clinical equivalence, clinical accuracy, medical or clinical grade.”

The guidance defines invasive as anything that “penetrates or pierces the skin or mucous membranes.” That definition matters for section 3.

The point to take from both sides of the Atlantic is the same: the framework is decided by the claims, and the claims are read across everything you publish.

2. Non-Invasive Wearables: ECG, PPG, SpO₂, Cuffless Blood Pressure, Temperature

EU. These are active devices for monitoring vital physiological processes, classified under Rule 10 of MDR Annex VIII. MDCG 2021-24 Rev.1 lists electrocardiographs, electronic thermometers and electronic blood pressure equipment as class IIa examples, and explains that devices for self-monitoring are class IIa, whereas those intended for continuous surveillance in anaesthesia, intensive care or emergency care — such as multi-parameter monitors with alarms — are class IIb.

If the interpretation of the signal is done by software, Rule 11 also applies. MDCG 2019-11 gives the example of a smartwatch app that alerts the user to irregular heartbeats “for the purpose of detecting cardiac arrhythmia”: that is medical device software.

US. Over-the-counter ECG software (21 CFR 870.2345) and pulse oximeters (21 CFR 870.2700) are class II. The FDA’s January 2025 draft guidance on pulse oximeter performance, which addresses accuracy across skin pigmentation, was still a draft as of October 2026.

The claim that moves you. Cuffless blood pressure is the clearest example of how fast this can change:

The sensor did not change between July 2025 and June 2026. The policy did, and so did the way the product was presented.

The September 2025 safety communication is still published. The close-out letter applies to one modified product, not to the category.

3. Minimally Invasive Continuous Monitors: Glucose, Ketones, Lactate, Microneedles

EU. A sensor filament under the skin measures interstitial fluid inside the body, so by the definitions above it falls under the MDR rather than the IVDR. No official EU document states this for continuous glucose monitors specifically; it follows from the wording of IVDR Article 2(2).

Anything that penetrates the skin is surgically invasive. Depending on wear time, Rule 6 (under 60 minutes), Rule 7 (60 minutes to 30 days) or Rule 8 (more than 30 days, or implantable) applies, alongside Rule 10 for the monitoring function. Where several rules apply, the strictest one decides the class (Annex VIII, implementing rule 3.5). A sensor intended to warn of dangerous variations — hypoglycaemia alarms, for instance — invites the Rule 10 class IIb reading.

US. Integrated continuous glucose monitors are class II under 21 CFR 862.1355, with special controls that fix accuracy thresholds against a laboratory reference. Over-the-counter CGMs have their own product codes under the same regulation. Two 2026 authorisations show where the field is going:

The claim that moves you. Microneedles cannot be wellness products in the US, whatever the claim. The January 2026 guidance includes the example of a microneedle glucose wearable, labelled as not for use by people with diabetes or prediabetes, and concludes that because it “penetrates the stratum corneum” it “is therefore not a low risk general wellness product.” The invasiveness alone decides it.

4. Non-Invasive Fluid Sensors: Sweat, Tears, Saliva, Breath

This is where the line between the MDR and the IVDR is hardest to draw, and where official guidance is thinnest.

Breath. The Commission’s Manual on borderline and classification (version 5, April 2026) addresses a device measuring exhaled nitric oxide. Its conclusion: “The exhaled air is no longer part of the human body and therefore the exhaled air is considered to be a gaseous specimen derived from the human body, which is subsequently analysed by a device outside of the body.” The device qualifies as an IVD.

Saliva collected and then tested. MDCG 2020-16 treats saliva collection kits as specimen receptacles and the test as an IVD.

Sweat and tears measured on the body. There is currently no official EU position. Version 5 of the borderline manual contains no entry for them. An older edition, written under the previous Medical Devices Directive, treated a sweat collection card applied to the skin as a medical device rather than an IVD — but that edition has been superseded and its decisions were not carried over.

The breath reasoning points one way (the fluid has left the body, so it is a specimen); the old sweat case points the other (the device sits on the body, so it is a medical device). For an on-skin sweat patch or a tear-sensing contact lens, the answer depends on how the product works and what it is said to do. It is a genuine borderline case, and the borderline manual itself is not legally binding.

US. The FDA’s wellness guidance includes the example of a product estimating lactate, electrolytes and haemoglobin for elite athletes using an optical sensor, treated as wellness provided the values are validated. The same conditions apply: non-invasive, wellness outputs only, no disease claims.

The claim that moves you. For this family, it is not only the medical claim that matters but the technical description. Whether you say the patch “collects” sweat or “measures on the skin” may be the detail on which the borderline decision rests.

5. Diagnostic Biosensors for Blood Biomarkers

Electrochemical immunosensors, lateral flow strips, lab-on-chip cartridges and point-of-care readers are the largest family by far. Troponin, C-reactive protein, procalcitonin, HbA1c, brain injury markers such as S100B or GFAP, hormones, drug levels: the list of analytes is long and keeps growing.

The class depends on the clinical use, not the molecule

MDCG 2020-16 rev.5 (September 2026) is explicit: “The classification of a device is defined by its intended purpose, as specified by the manufacturer,” and, “unless otherwise specified,” the rules “apply equally to all technologies, principles of detection or analytical procedures.”

The guidance shows this with a single marker. C-reactive protein or procalcitonin used to monitor the response to treatment in sepsis is class C (Rule 3(j)). A test measuring C-reactive protein to detect systemic inflammatory processes is class B (Rule 6). Same analyte, same technology, different class.

Other examples from the guidance:

Use Class (IVDR) Rule
Troponin I/T, CK-MB in acutely presenting patients, when intended for monitoring cardiac muscle injury C 3(j)
HbA1c and blood glucose for managing diabetes C 3(k)
Therapeutic drug monitoring (e.g. lithium, tacrolimus, gentamicin) C 3(j)
Cancer screening, diagnosis or staging (e.g. CA 125) C 3(h)
Most clinical chemistry: hormones, enzymes, electrolytes, metabolic markers B (likely) 6

Brain injury markers. MDCG 2020-16 gives no example for S100B, GFAP or UCH-L1. In practice, the GFAP/UCH-L1 tests on the EU market intended to help decide whether a patient with a mild head injury needs a CT scan are registered in EUDAMED as class C — for example VIDAS TBI and the i-STAT TBI cartridge. The rule applied is not published; a likely reason is that a false negative can mean a missed intracranial bleed. A triage claim should not be assumed to be class B, and MDCG 2020-16 warns that “ambiguous claims may lead to higher classification.”

Who uses it changes the class

US

Troponin tests are class II, regulated under 21 CFR 862.1215 (product code MMI), and a high-sensitivity troponin cartridge for a handheld analyser was cleared in January 2025. For brain injury, the FDA created a dedicated category in 2018, the “brain trauma assessment test” (21 CFR 866.5830, class II), for devices that detect brain injury biomarkers “to assist in determining the need for head imaging.” Several combined GFAP/UCH-L1 tests have been cleared under it since, including a whole-blood cartridge for a handheld analyser. We found no blood S100B test for brain injury in the FDA databases.

Outside central laboratories, the CLIA complexity category matters. Only waived tests — “simple laboratory examinations and procedures that have an insignificant risk of an erroneous result” — can be run at sites holding only a Certificate of Waiver. Both handheld cartridges mentioned above are categorised as moderate complexity, so the sites using them need a laboratory certificate for that level.

The claim that moves you. For blood biomarkers, the move is almost always clinical. “Aids in assessing” is not “rules out”. “Detects inflammation” is not “monitors sepsis”. Each word in the intended purpose maps to a rule.

6. Implantable Biosensors

Implantable sensors fall under Rule 8 of MDR Annex VIII: class IIb by default, class III if, among other cases, they are active implantable devices or their accessories, have a biological effect or are wholly or mainly absorbed, undergo chemical change in the body, or are intended to be used in direct contact with the heart, the central circulatory system or the central nervous system.

For implantable and class III devices, the MDR requires clinical investigations (Article 61(4)). The exceptions are narrow: modifications of a device the manufacturer already markets, or equivalence to another manufacturer’s device under a contract giving full access to its technical documentation, in both cases endorsed by the notified body (Article 61(4)–(5)); certain devices already on the market under the former directives (Article 61(6)(a)); and a list of well-established device types (Article 61(6)(b)).

Commission Delegated Regulation (EU) 2026/1451 of 20 March 2026, published on 29 June 2026, replaced that list with a much longer one. It covers items such as sutures, needles, catheters, guidewires and leads — not implantable sensors. And, as its recital 5 states, exempted manufacturers “are nonetheless required to plan, conduct and document a clinical evaluation in accordance with Article 61.”

7. What May Change in Europe — and What Has Not Changed Yet

In December 2025 the Commission proposed a targeted revision of the MDR and IVDR (COM(2025) 1023). Among other things, it would:

None of this is law. As of October 2026 the procedure (2025/0404(COD)) is awaiting a committee decision in Parliament, and the Council has not reached a general approach. The current MDR and IVDR apply in full, and products placed on the market today are assessed against them. Part 2 looks at what these proposals would mean for the evidence each biosensor family needs.

8. Before You Write a Single Line of Marketing Copy

  1. List everything you publish about the product. Website, app store, sales deck, investor material, posts, conference posters. All of it is part of your intended purpose in the EU, and all of it is read by the FDA.
  2. Decide the framework per market. A product can be a wellness product in the US and a medical device in the EU, or the other way round. Wording that works in one may not work in the other.
  3. Map each claim to a rule. Disease names, thresholds, alerts, “clinical grade”, “monitors”, “rules out”: each one has a regulatory consequence.
  4. Check the technical description, not only the medical claim. For fluid sensors, how you describe where the measurement happens may decide between the MDR and the IVDR.
  5. Make the evidence match the claim. The stronger the claim, the stronger the evidence it needs — which is the subject of Part 2.

Read your claims the way a regulator will

Reviewing claims across a website, a sales deck and an app store listing is the first thing a notified body or the FDA will effectively do when it reads your file. RegAffairs Hub offers a fixed-scope claims and intended purpose audit: we read everything you say about your product in public, the way a regulator would, and tell you which framework and class those claims point to.

Get in touch

Sources