Over-the-Counter CGM Clearance: FDA 510(k) Pathway and Differences From Prescription iCGMs

Over-the-counter CGMs in the US: how FDA cleared them and how they differ from prescription iCGMs

Key points

• FDA did not create a new device classification for over-the-counter (OTC) continuous glucose monitors. OTC CGMs have been cleared through 510(k) under the existing integrated continuous glucose monitoring system (iCGM) regulation, 21 CFR 862.1355, a Class II device type with special controls ^155218180292.
• That regulation traces to the 2018 De Novo for the Dexcom G6 (DEN170088), which was authorized for prescription use only ⁷⁴²⁴².
• The first OTC CGM was the Dexcom Stelo Glucose Biosensor System (K234070), cleared March 5, 2024, with the prescription Dexcom G7 (K231081) as predicate ^4156255. Abbott's Lingo (K233655, May 29, 2024) and Libre Rio (K233861, June 7, 2024) followed, both using the prescription FreeStyle Libre 2 (K222447) as predicate ^32191181.
• FDA separates OTC iCGMs from prescription iCGMs mainly through dedicated OTC product codes (SAF and SBH). Both codes are limited to people who are not on insulin and who do not have a significant risk of problematic hypoglycemia, and both exclude connection to insulin delivery controllers ⁹²⁹⁵. The prescription code QBJ explicitly includes automated insulin dosing systems ⁹³.
• The codified special controls in 21 CFR 862.1355 apply to OTC and prescription iCGMs alike. The retrieved regulation text has no separate OTC-specific special controls ²⁹⁰. OTC clearances met the same clinical performance special controls ^194223348. OTC-specific risks were handled through narrower indications, restricted display ranges, removal of glucose alarms, and lay-user self-selection testing ^{157220182195277}.

Regulatory foundation: one regulation, multiple product codes

21 CFR 862.1355 and the G6 De Novo

FDA classified the iCGM device type through the De Novo for the Dexcom G6 Continuous Glucose Monitoring System (DEN170088), decided March 27, 2018 ⁷⁴. The regulation defines an iCGM as a device that automatically measures glucose in bodily fluids continuously or frequently over a specified period and reliably and securely transmits the data to digitally connected devices, "including automated insulin dosing systems," for use alone or with those devices to manage a condition related to glycemic control ²⁹¹. It is Class II (special controls) ²⁹².

The G6 De Novo indication was diabetes management in people age 2 years and older. It covered replacing fingerstick testing for treatment decisions and autonomous communication with automated insulin dosing systems, and it carried the condition "This device is for prescription use only" ²⁴¹²⁴².

OTC product codes under the same regulation

FDA later created OTC product codes that sit under 21 CFR 862.1355 and are classified as Class II ⁹²⁹⁵:

The difference between SAF ("monitoring") and SBH ("management") matters for labeling. SAF's definition carries the statement that output should not drive treatment decisions without an HCP. SBH's definition does not, and it refers to the full clinically significant hypo/hyperglycemic range ⁹²⁹⁵.

The OTC clearances

Two features of this pathway stand out:

1. Prescription predicates for OTC devices. Stelo and Lingo/Libre Rio were found substantially equivalent to prescription iCGMs (G7 and FreeStyle Libre 2), and the change in prescription status was listed as a difference in the predicate comparison tables ^157220182. Stelo uses the same hardware design as G7. Its firmware and app were changed to support a longer wear period, restrict connections, and tailor the interface to Stelo users ¹⁵⁸.
2. Software-only OTC entrants. Signos is an OTC mobile application that receives data from an iCGM sensor. It was cleared under SAF with Stelo as predicate, and its indication adds that the information "may be useful in helping users to maintain a healthy weight" ^106107109.

FDA's March 5, 2024 announcement described Stelo as the first OTC CGM. It was intended for adults 18 and older who do not use insulin, including people with diabetes on oral medication and people without diabetes who want to understand how diet and exercise affect glucose. FDA cited expanded access without HCP involvement as the rationale ^255260261. The announcement stated that Stelo is not for people with problematic hypoglycemia because it does not alert users to that condition, and that users should consult a health care provider before making medical decisions based on its output ²⁶⁰²⁶².

Indications and intended populations: OTC vs prescription

Stelo's labeling also states that it should not be used by people with problematic hypoglycemia, because it was not designed for that population. It also should not be used on dialysis, because performance was not evaluated there ²⁷⁹. For Lingo and Libre Rio, the retrieved review text did not state a problematic-hypoglycemia warning. The not-on-insulin limit and the absence of alarms were the clearest stated limitations ²⁸⁶³¹⁰. The SAF/SBH code definitions, however, both exclude persons with a significant risk of problematic hypoglycemia ⁹²⁹⁵.

Special controls: identical text, different implementation

What 21 CFR 862.1355 requires

The special controls apply to every device under the regulation, OTC or prescription ²⁹²:

1. Clinical accuracy. Data must come from the intended-use population and cover the full measuring range and wear period, compared against a traceable, FDA-accepted laboratory method ²⁹³. Adult performance must meet lower one-sided 95% confidence bounds ²⁹⁴:

   • Below 70 mg/dL: more than 85% within ±15 mg/dL and more than 98% within ±40 mg/dL ²⁹⁵²⁹⁶
   • 70 to 180 mg/dL: more than 70% within ±15% and more than 99% within ±40% ²⁹⁷²⁹⁸
   • Above 180 mg/dL: more than 80% within ±15% and more than 99% within ±40% ²⁹⁹³⁰⁰
   • Across the full measuring range: more than 87% within ±20% ³⁰¹

   Readings must never cross between below 70 and above 180 mg/dL ³⁰². Rate-of-change limits apply ³⁰³³⁰⁴. Pediatric performance must be similar to adult performance unless FDA accepts a justification ³⁰⁵.

2. Secure, reliable data transmission at clinically meaningful intervals, with no clinically significant data gaps that would prevent connected devices from achieving their intended use ²⁷⁰²⁶⁹.

3. Manufacturing and release controls that ensure released devices meet the performance specifications ³⁰⁶.

4. Interference testing ³⁰⁷.

5. Wear-limit enforcement that prevents sensors from being used beyond their claimed wear period ³⁰⁸.

6. Usability study showing intended users can use the device safely and obtain the expected accuracy ³⁰⁹.

7. Labeling of clinical performance for each intended-use population and insertion site. This covers accuracy by glucose band (<54, 54 to <70, 70 to 180, >180 to 250, >250 mg/dL), rate-of-change accuracy, data gaps, alert performance (if applicable), and observed sensor life ²⁶⁸.

The retrieved text of 21 CFR 862.1355 does not contain separate special controls for OTC devices or non-insulin users, and it does not itself prescribe a prescription-use statement ²⁹⁰.

How OTC sponsors met the same controls

• Clinical performance. Stelo's decision summary states that the device met the iCGM clinical-performance special controls ¹⁹⁴. Against YSI venous plasma over about 15 days of wear, it reported MARD of 7.3% to 9.1% and 92.1% to 96.0% of readings within 20 mg/dL or 20% ¹⁹⁰¹⁹¹. Readings outside the 70 to 250 mg/dL display window were evaluated separately, and the full 40 to 400 mg/dL sensor range was available retrospectively ¹⁹²²⁷³. Lingo used the predicate's clinical studies with a new analysis for the changed age range, measuring range and wear period, and met the special-control requirements ^222223224. Libre Rio referenced the 862.1355 special controls and relied on analytical performance established in K222447 ¹⁸³¹⁸⁴.
• Usability (special control 6) adapted to OTC. Stelo's human factors program tested whether lay users could correctly identify themselves as appropriate users from pre-purchase labeling and then use the system safely ¹⁹⁵¹⁹⁶. Sixty participants evaluated four e-commerce self-selection platforms, and the 15 who self-identified correctly went on to simulated-use testing ²⁷⁷. Lingo's app includes onboarding with a self-selection questionnaire that users must consent to before use ²⁸⁷. The Libre Rio review text cited human factors support for substantial equivalence but did not describe OTC-specific self-selection methods ³¹².
• Alerts and labeling (special control 7). The labeling control requires alert performance data only "if applicable" ²⁶⁸. All three initial OTC devices removed glucose alerts and alarms. For Stelo, this included urgent-low, low, high and rapid-change alerts ^281221182.
• Interoperability (special control 2). Prescription iCGMs are cleared to feed AID systems ¹⁶⁸. The OTC product codes exclude insulin delivery controllers ⁹²⁹⁵. Libre Rio's comparison table lists it as not interoperable with AID ¹⁸², and Stelo's firmware restricts connections ¹⁵⁸.

How the OTC category has evolved

The OTC category has moved toward claims closer to prescription iCGMs while keeping the insulin exclusion:

• SBH "management" code. Libre Rio was cleared under SBH in June 2024 ¹¹⁸⁰. In June 2026, Stelo K260935 moved from SAF to SBH ^5132133.
• Pediatric expansion. Stelo K260935 lowered the age threshold from 18 to 2 years and older, still limited to people not on insulin ¹³⁰¹³⁴. Pediatric support came from reanalyzing a prior G7 pediatric study (164 children) with the Stelo algorithm, not from a new Stelo pediatric study. Accuracy and sensor survival beyond about day 10.5 were not directly studied in children ^352353354. FDA concluded the device met the iCGM clinical-performance special controls, with that limitation noted ³⁴⁸.
• Broader display range and removal of the HCP statement. The display range widened from 70 to 250 mg/dL to 55 to 350 mg/dL, and a Direct to Watch feature was added ¹³⁵¹³⁶. The statement that users should not take medical action without consulting an HCP was removed from the indication ¹³⁴. This tracks the SBH code definition, which omits that sentence ⁹⁵.

Related category: glucose range monitoring (21 CFR 862.1359)

A separate, non-iCGM classification was created by the De Novo for the Biolinq Shine Autonomous Time-in-Range Microsensor (DEN240080), decided September 23, 2025, under product code SFU ²⁶³. The 862.1359 "glucose range monitoring system" provides qualitative or semi-quantitative glucose information, not quantitative concentrations, and is not intended for acute decisions such as insulin dosing ⁹¹³⁵¹. Biolinq Shine is indicated for people aged 22 or older who are not on insulin. Its users are not intended to take medical action without HCP consultation, and FDA designated it a prescription device ³⁴⁹³⁵⁰. Its special controls mirror the iCGM structure: clinical validation, data availability, interference, wear limits, usability and population-specific labeling ³¹⁵. Not-on-insulin populations can therefore fall under either regulation, and OTC status depends on the product code and clearance, not only on the user population.

Practical implications for regulatory strategy

• Pathway. An OTC iCGM is a 510(k) under 21 CFR 862.1355, with SAF or SBH as product code. Prescription iCGMs (QBJ) have served as predicates. The predicate comparison must address the shift to OTC, the restricted population, and the loss of AID interoperability and alarms ^157220182.
• Evidence. The quantitative accuracy special controls do not relax for OTC. Sponsors have reused the predicate's clinical datasets with new analyses for the modified algorithm, range and wear period ²²²³⁵². Expect FDA to review out-of-display-range performance separately ¹⁹²³⁵⁶.
• OTC-specific human factors. Self-selection from pre-purchase labeling (including e-commerce platforms) and safe lay use are evaluated under the usability control ¹⁹⁵²⁷⁷.
• Labeling boundaries. The not-on-insulin limit, exclusion of problematic hypoglycemia, and any HCP-consultation language should match the target product code's definition ^9295279.