How It Works
Continuous glucose monitors (CGMs) operate by measuring interstitial fluid glucose through a subcutaneous sensor, which then transmits real-time readings to a receiver or mobile device. The sensor uses an enzyme, typically glucose oxidase, to catalyze a reaction that generates an electrical signal proportional to glucose concentration. This signal is converted into a glucose value displayed to the user. The core mechanism is identical across FDA-regulated CGMs, but the rigor of validation, calibration, and performance thresholds differs by device classification.
FDA Class II devices are cleared via the 510(k) premarket notification pathway, requiring substantial equivalence to an already legally marketed predicate device. Class III devices, by contrast, require a Premarket Approval (PMA) application, which demands comprehensive clinical data demonstrating safety and effectiveness. For CGMs, Class II clearance is the standard for most consumer-facing systems, while Class III designation applies to devices intended to support critical treatment decisions—such as insulin dosing—without physician oversight. The classification determines the depth of clinical evidence and the allowable margin of error.
Key terms defined: MARD (Mean Absolute Relative Difference) is the average deviation between CGM readings and reference blood glucose values. Accuracy is assessed across the full range of glucose levels, including hypoglycemia and hyperglycemia. Calibration refers to the process of adjusting sensor output using fingerstick blood glucose values, though many modern CGMs are factory-calated. Time in Range (TIR) is the percentage of readings within target glucose bounds. These metrics are reported in regulatory submissions but are not always directly comparable across devices due to differing study populations and reference methods.
The mechanism of action is consistent, but the regulatory pathway dictates how accuracy is demonstrated. Class II devices may rely on smaller, retrospective studies against a predicate, while Class III devices require prospective, multi-site trials with stricter endpoints. This distinction matters because a Class II device may meet minimum clearance thresholds without demonstrating performance in extreme glucose fluctuations or during rapid changes. A Class III device, by contrast, must prove reliability under dynamic conditions to justify automated insulin delivery integration.
For treatment decisions, the classification implies a different risk-benefit calculus. A Class II CGM is suitable for trend monitoring and lifestyle adjustments, but its accuracy may not support insulin dosing without confirmatory fingersticks. A Class III device is engineered to meet higher performance standards, enabling therapeutic decisions with reduced redundancy. The user must verify whether their device’s classification aligns with the intended use—especially when integrating with automated insulin pumps or closed-loop systems.
| Term | Definition | Regulatory Relevance |
|---|---|---|
| MARD | Average deviation from reference glucose | Primary accuracy metric in 510(k) and PMA submissions |
| Class II | 510(k) clearance pathway | Substantial equivalence to predicate; lower evidence bar |
| Class III | PMA approval pathway | Full clinical trial data required; supports autonomous treatment |
| Time in Range | % of readings within target glucose bounds | Clinical outcome metric; increasingly used in regulatory reviews |
Key Factors to Consider
When choosing a continuous glucose monitor (CGM) system, the first decision criterion is the regulatory classification of the device. In the United States, the FDA assigns either Class II or Class III status, and this distinction directly impacts the level of clinical evidence required for clearance. Class II devices typically receive clearance through the 510(k) pathway, which requires demonstrating substantial equivalence to an already marketed device. Class III devices, on the other hand, must undergo the more rigorous premarket approval (PMA) process, which demands comprehensive clinical data proving safety and effectiveness. This regulatory tiering is not merely bureaucratic; it reflects the device's intended use and risk profile, which in turn influences how clinicians interpret the data for treatment decisions.
The second critical factor involves the sensor's stated accuracy metrics, specifically the Mean Absolute Relative Difference (MARD). MARD quantifies the average deviation between CGM readings and reference blood glucose measurements, and lower values indicate higher reliability. For treatment decisions such as insulin dosing, a lower MARD reduces the risk of hypo- or hyperglycemic events. The acceptable MARD threshold varies between device classes: per the FDA criteria cited in this guide, Class II devices must not exceed 9.0% MARD across at least 95% of sensor glucose values, while Class III devices must not exceed 8.0% MARD across at least 99% of readings. Clinicians should verify the published MARD for each device and consider how this accuracy aligns with the patient's glycemic variability and treatment regimen.
The third decision criterion concerns the integration of CGM data with insulin delivery systems. Some Class III devices are designed as part of closed-loop systems, where CGM readings automatically adjust insulin infusion rates. These systems require not only high accuracy but also low latency and robust fail-safe mechanisms, which are mandated by the PMA process. In contrast, Class II devices may be used standalone or with pump systems that do not automate insulin delivery, offering flexibility but placing greater responsibility on the user to interpret trends and make manual adjustments. The choice between these configurations depends on the patient's ability to manage complex data and the desired level of automation in glycemic control.
Regarding numbers that matter, this guide's FDA criteria state that Class III CGMs must demonstrate a MARD of no more than 8.0% across at least 99% of sensor glucose values, while Class II devices must not exceed 9.0% MARD across at least 95% of readings. These thresholds are stricter for Class III devices than for Class II devices. Additionally, the time-in-range (TIR) metric, defined as the percentage of readings within 70-180 mg/dL, has become a key performance indicator. Before relying on any TIR claim, verify the study population and comparator used, since results may not be directly comparable across devices.
Clinicians should also examine the sensor's wear time and calibration requirements, as these factors affect real-world usability and data continuity. Class III devices frequently feature longer wear durations—up to 14 days—and factory calibration, reducing user burden. Class II devices may require more frequent calibration and have shorter wear times, potentially leading to data gaps that complicate trend analysis. When comparing options, verify that the device's specifications match the patient's lifestyle and the clinical team's data interpretation protocols, ensuring that the chosen system supports consistent, actionable insights for treatment adjustments.
Common Mistakes
One of the most common mistakes is assuming that all CGM accuracy claims are created equal. A device might advertise a MARD figure that appears lower than another's, but without knowing the testing conditions, sample size, and whether those numbers reflect real-world use or controlled lab settings, the lower number can be misleading. Always ask for the full MARD breakdown, including the standard deviation and the population studied, before treating one device as definitively more accurate than another.
Another frequent error is conflating regulatory classification with clinical performance. A Class III device is not automatically superior in accuracy just because it underwent more stringent premarket review. The FDA classification reflects the level of risk and the regulatory pathway, not the day-to-day precision of glucose readings. A Class II CGM may perform just as well in practice, especially if it has been on the market longer and has accumulated more real-world calibration data. Verify the device’s post-market surveillance reports and peer-reviewed studies rather than relying solely on the FDA class label.
A third pitfall involves ignoring the impact of sensor insertion timing and site rotation on accuracy. Both Class II and Class III devices can produce skewed readings if the sensor is inserted into scarred tissue, a site with poor circulation, or if it is not given adequate time to stabilize after insertion. Some users assume that a higher-class device will compensate for poor insertion technique, but no regulatory classification can override the physiological variables of interstitial fluid dynamics. Always follow the manufacturer’s recommended insertion protocol and allow the full warmup period before making treatment decisions based on the readings.
Finally, many users fail to account for the difference between point-in-time accuracy and trend accuracy. A CGM may report a glucose value within an acceptable error range at a single moment but still fail to capture rapid directional changes, which are critical for insulin dosing decisions. Before committing to a device, review its ability to detect and communicate trending directions, not just absolute values, and cross-check with confirmatory fingerstick measurements when trends appear inconsistent with symptoms.
Insider Tactics
Insider Tactics
This section gives non-obvious strategies and timing tips. A non-obvious strategy is to treat the FDA classification as a dynamic filter rather than a static label. Instead of assuming a Class III device is automatically superior, verify the specific clearance letter for the firmware version currently in use, because manufacturers may release iterative updates that shift a device’s regulatory standing without changing the model number on the box.
The timing tip is to schedule your verification check during the manufacturer’s quarterly transparency window, when they typically publish updated accuracy tables and adverse-event summaries. By aligning your review with these releases, you avoid relying on marketing materials that may be up to six months out of date, ensuring you compare like-for-like totals and terms across devices.
For example, a clinician might notice that a Class II sensor’s MARD (Mean Absolute Relative Difference) improved in the latest firmware patch, narrowing the gap with Class III competitors. However, without checking the specific clearance letter, one might miss that the updated algorithm is only cleared for use in adults, not pediatric patients.
Another tactic is to cross-reference the FDA’s 510(k) database with the manufacturer’s own clinical trial registry. This dual-check method reveals whether the accuracy claims are based on real-world data or idealized study conditions, helping you assess how the device will perform in your specific treatment environment.
Finally, always verify the sensor’s exclusion criteria before committing. Some Class III devices, despite their higher regulatory scrutiny, may still exclude patients with certain skin conditions or medications, which could render them unsuitable for your needs despite their superior accuracy metrics.
Comparison
When comparing FDA Class II and Class III CGM options side by side, the first check is not the headline claim, but the proof path. As noted by Riskonnect in "Business Continuity Standards: Are You Aligned or Just Accountable?", a program can meet requirements on paper through certification, while alignment focuses more directly on performance under pressure. Applying that same split here, one regulatory path leans toward meeting a defined set of requirements, while the other leans toward a different level of review. For treatment decisions, the option that makes a live proof check unavoidable before relying on it is the one that must be verified first.
To keep the comparison like-for-like, do not compare marketing phrases. Instead, compare the burden of verification for each option by asking for the exact test conditions used to support its accuracy statement, who ran those tests, and whether those conditions match your intended use. This is a totals-and-terms check, not a guess. If either option cannot produce that matching set of terms on request, treat that as a reason to not commit until a complete answer is provided.
When each option wins depends entirely on that proof. If one option can deliver a complete, matching proof set with traceable conditions, it can win for the specific decision point you need covered. If the other can only point to a general claim without matching conditions, it can only win in the narrow case where that decision point is not needed at all. For most treatment decisions, that narrow case is not the one being made.
With that rule applied, the winner for a verify-before-you-commit approach is the option whose review path makes a matching proof set easier to obtain and harder to substitute with a general claim. This is not about the class name itself, but about which path yields a verifiable comparison. Before committing to either, run that single check first and do not accept a partial answer in its place.
Frequently Asked Questions
What clinical data requirement distinguishes FDA Class III CGMs from Class II devices?
Class III devices require a Premarket Approval (PMA) application with comprehensive clinical data demonstrating safety and effectiveness, while Class II devices are cleared via the 510(k) pathway requiring substantial equivalence to a predicate device.
Which FDA classification applies to CGMs intended to support insulin dosing decisions without physician oversight?
Class III designation applies to CGMs intended to support critical treatment decisions—such as insulin dosing—without physician oversight.
How does the sensor in an FDA-regulated CGM generate a glucose reading?
The sensor uses an enzyme, typically glucose oxidase, to catalyze a reaction that generates an electrical signal proportional to glucose concentration, which is then converted into a glucose value.
What metric is used to assess CGM accuracy against reference blood glucose values?
MARD (Mean Absolute Relative Difference) is the average deviation between CGM readings and reference blood glucose values.
What is the primary difference in validation rigor between Class II and Class III CGM classifications?
The classification determines the depth of clinical evidence and the allowable margin of error, with Class III requiring more rigorous validation than Class II.
What is the standard FDA classification for most consumer-facing CGM systems?
Class II clearance is the standard for most consumer-facing CGM systems.
Quick answers
| What is the primary regulatory pathway for FDA Class II CGMs? | FDA Class II devices are cleared via the 510(k) premarket notification pathway, requiring substantial equivalence to an already legally marketed predicate device. |
| What regulatory pathway is required for FDA Class III CGMs? | Class III devices require a Premarket Approval (PMA) application, which demands comprehensive clinical data demonstrating safety and effectiveness. |
| What is the typical accuracy threshold for FDA Class II CGMs? | The article references a 95% accuracy standard for Class II devices, which corresponds to the 5% margin of error implied by the 95% figure. |
| What is the typical accuracy threshold for FDA Class III CGMs? | The article references a 99% accuracy standard for Class III devices, which corresponds to the 1% margin of error implied by the 99% figure. |
| What type of treatment decisions are FDA Class III CGMs intended to support? | Class III designation applies to devices intended to support critical treatment decisions—such as insulin dosing—without physician oversight. |
Also worth reading: Lower Blood Sugar Levels: Continuous Glucose Monitor (CGM) 0.6% vs Fingerstick: Lower Blood Sugar Levels: Continuous · Continuous Glucose Monitor Results: 1 Result—Open an Evidence Review, Not a Final Call: Continuous Glucose Monitor Results: 1 · Why Nordic Blonde Hair Dye Formulas Differ from Mediterranean Versions A 2024 Chemical Analysis: Why Nordic Blonde Hair Dye