CGM
Continuous Glucose Monitor · 24/7 Skin Sensor
Evidence grade 2 of 5, Thin. Read how the Atlas grades evidence.
8 human trials, about 300 studies, >3,000 participants across non-diabetes-focused behaviour-change RCTs; millions of device users overall. The written rationale for this grade is still being prepared.
- FDA
- FDA-cleared as a medical device. Stelo by Dexcom cleared March 5, 2024 as the first OTC glucose sensor in FDA history; Abbott Lingo and Libre Rio cleared June 10, 2024. Prescription CGMs for diabetes long predate these OTC clearances.
- WADA
- Not a prohibited substance or method. WADA has discussed CGM in-competition restrictions for some sports, but CGMs are not on the Prohibited List.
Reference only. Not medical advice.
What can go wrong
In healthy people the main risks are psychological and behavioral rather than physical: anxiety from reading a normal glucose swing as disease, and restrictive or disordered eating patterns. Nicola Guess (Practical Diabetes, 2023) warned that non-diabetic use risks pathologizing normal post-meal excursions and pushing people to avoid healthful foods. Interstitial readings lag blood and can be inaccurate at the extremes, and consumer sensors have been shown to overstate time above range in healthy adults (Gonzalez et al., 2025), so a flagged spike may be noise. Over-the-counter consumer sensors such as Stelo are not designed to detect dangerously low blood sugar and, per FDA labeling, will not alert to it, so they are not for people on insulin or with problematic hypoglycemia. Anyone who suspects diabetes, has an eating-disorder history, or feels increasing anxiety around food should work with a qualified clinician or registered dietitian and, if diabetes is the question, a validated medical CGM rather than a wellness sensor or an app dashboard.
A CGM is a skin-worn enzyme sensor that estimates glucose in the fluid between cells and streams it to a phone. The chemistry goes back to Leland Clark and Ann Lyons (1962) and to Updike and Hicks (1967), and the first wearable sensor-driven system came from Motoaki Shichiri's group at Osaka University (The Lancet, 1982). For diabetes the benefit is well established: the DIAMOND and GOLD randomized trials (JAMA, 2017) both lowered HbA1c in adults with type 1 on insulin injections, and the American Diabetes Association grades CGM as Grade A for type 1 and for anyone on insulin. For healthy, non-diabetic users the picture is thin. Hall, Snyder and colleagues (PLOS Biology, 2018) showed even normoglycemic people have variable glucose patterns they called glucotypes, which fueled wellness interest, though that work describes variation rather than showing that acting on it improves health, and a later analysis (Hulman et al., PLOS Biology, 2021) argued the glucotype label mostly tracks average glucose. Aronson et al. (2025) reviewed non-insulin type 2 trials and again found benefit tied to disease. Nicola Guess (Practical Diabetes, 2023) called non-diabetic use an evidence-free zone, and a University of Bath study (Gonzalez et al., American Journal of Clinical Nutrition, 2025) found consumer sensors overstate time above range in healthy people. Real-time data can change short-term food choices, but durable health gains in metabolically healthy adults are unproven.
A coin-sized arm patch that streams your blood sugar to your phone every few minutes.
What people take it for
These are the reasons people give for taking it. They are not effects it has been shown to have.
What people get wrong
Often saidA glucose spike after a meal means something is wrong with you.
ActuallyGlucose rising after carbohydrate is normal physiology. Hall and Snyder (PLOS Biology, 2018) found even healthy normoglycemic people spend time in ranges that look impaired on a sensor; a spike by itself is not a diagnosis.
Often saidCGMs are proven to make healthy people healthier.
ActuallyNo trial shows that. The benefit in meta-analyses is concentrated in people with diabetes (Aronson 2025 in non-insulin type 2; Richardson 2024 across mixed groups), and Guess (2023) labeled non-diabetic use an evidence-free zone.
Often saidThe number on the app is your blood sugar in real time.
ActuallyIt estimates glucose in interstitial fluid with roughly a 5 to 15 minute lag plus device error, so it approximates rather than equals a blood reading. A University of Bath study (Gonzalez et al., American Journal of Clinical Nutrition, 2025) found consumer sensors overstated time above range in healthy adults.
Often saidYour personal glucotype tells you something your regular labs miss.
ActuallyHulman and colleagues (PLOS Biology, 2021) reproduced the glucotype label from average glucose alone in 92 percent of people and found the original prevalence estimates did not hold in outside groups, so the category may mostly restate mean glucose.
Who profits, and who pushes back
Nicola Guess
For people at no risk of diabetes there is probably no need to track glucose; the growing use of continuous glucose monitors in people without diabetes is an evidence-free zone that risks pathologizing normal post-meal glucose and the unnecessary avoidance of healthful foods (Practical Diabetes, 2023).
Javier Gonzalez
For someone with good glucose control these monitors can be misleading, and relying on them could lead to unnecessary food restrictions or poor dietary choices (University of Bath, 2025).
Levels (Casey Means, Sam Corcos, Josh Clemente)
Sells this. Has a commercial interest in a product sold as this compound, as of July 2026.
PAID. Commercial vendor; markets non-diabetic CGM use directly.
David Sinclair
Sells this. Has a commercial interest in a product sold as this compound, as of July 2026.
PAID. Joined Levels as an advisor (announced 2021).
Michael Snyder
Academic proponent of personal glucose profiling; has had ties to wearable ventures.
Where the research comes from
| Place | Share of published research |
|---|---|
| North America | 46% |
| Europe | 30% |
| Asia | 16% |
| Rest of world | 8% |
Research published from: United States.
How we got here
This is metabolic-sensing lineage, not the discovery of insulin, so Banting and Best (1921) are not the origin here. The thread begins with Leland Clark Jr.'s oxygen electrode (1956) and the first glucose enzyme electrode he built with Ann Lyons at Cincinnati Children's Hospital (1962), which detected the oxygen consumed as glucose oxidase broke down glucose. Updike and Hicks (1967) bound the enzyme in a polyacrylamide gel, the step that made the electrode practical for biological fluids. The first genuinely wearable, sensor-driven system was not Western: Motoaki Shichiri, Yoshimitsu Yamasaki, Ryuzo Kawamori and colleagues at Osaka University Medical School reported a needle-type sensor wearable artificial endocrine pancreas in The Lancet in November 1982. Medtronic MiniMed's CGMS won FDA approval in 1999 as a blinded clinic tool. Abbott's factory-calibrated FreeStyle Libre received its European CE Mark on September 3, 2014, cutting routine fingersticks and pushing CGM toward the mainstream. The consumer-wellness era opened in 2024 when the FDA cleared Stelo (March) and Abbott's Lingo and Libre Rio (June) for sale without a prescription to people who do not use insulin.
- 1956Leland Clark presents the oxygen electrode
- 1962Clark and Ann Lyons build the first glucose enzyme electrode at Cincinnati Children's Hospital
- 1967Updike and Hicks immobilize glucose oxidase in gel, making the sensor practical
- 1982Shichiri's Osaka group reports a wearable needle-type sensor system in The Lancet
- 1999Medtronic MiniMed CGMS cleared by FDA as a blinded clinic device
- 2014Abbott FreeStyle Libre receives its European CE Mark, factory-calibrated, no fingersticks
- 2017DIAMOND and GOLD RCTs published in JAMA show HbA1c benefit in type 1 diabetes on injections
- 2018Hall and Snyder publish glucotypes in PLOS Biology, sparking wellness interest
- 2021David Sinclair joins the metabolic-health startup Levels as an advisor (October); Hulman et al. publish a critique of glucotypes
- 2023Nicola Guess publishes the evidence-free zone critique in Practical Diabetes
- 2024FDA clears Dexcom Stelo (March 5) and Abbott Lingo and Libre Rio (June 10) for over-the-counter use
- 2025Aronson et al. review non-insulin type 2 trials; Gonzalez et al. (Bath) report consumer sensors overstate glucose in healthy adults
At a glance
- Sensor lifespan
- Stelo up to 15 days; Lingo up to 14 days
- Reading interval
- Glucose every 1 to 15 minutes depending on device
- What it measures
- Interstitial fluid glucose, not direct blood glucose (5 to 15 min lag)
- First OTC clearance
- Dexcom Stelo, March 2024
What this entry does not yet cover
What follows is unfinished work on our side, not a finding about CGM. A short entry is not a safe entry.
- Cost and market. We have not documented what it sells for or who sells it.
- Supply chain. We have not mapped where it is made.
- Who uses it. We have not documented the user groups.
- What could change. No forward-looking summary yet.
- How it compares. We have not set it against the alternatives.
- Children and teenagers. We have not checked what is known about younger people.
Sources
Dexcom Stelo cleared March 5, 2024 as first OTC glucose biosensor; for non-insulin users 18+, 15-day wear, readings every 15 min, not for people with hypoglycemia
Dexcom Stelo first FDA-cleared OTC glucose sensor, pricing and August 2024 availability
Abbott Lingo and Libre Rio cleared by FDA June 10, 2024 as OTC CGM systems
DIAMOND RCT: CGM lowered HbA1c by 0.6 percent vs usual care in 158 adults with type 1 diabetes on injections, 24 weeks, JAMA 2017
GOLD RCT: mean HbA1c difference of 0.43 percent favoring CGM in 161 adults with type 1 diabetes on injections, JAMA 2017
Aronson 2025: meta-analysis of 8 RCTs, 541 participants in non-insulin type 2 diabetes, HbA1c reduction 0.37 percent, benefit tied to disease management
Richardson 2024: meta-analysis of 25 RCTs, 2,996 participants, HbA1c reduction 0.28 percent, mostly diabetic populations, unclear benefit in non-diabetic groups
International Journal of Behavioral Nutrition and Physical Activity (via PMC)
Non-diabetic CGM use is an evidence-free zone that risks pathologizing normal glucose and avoidance of healthful foods
Consumer CGM overstated time above range by nearly 400 percent and overstated a food's glycemic index in healthy adults vs finger-prick
University of Bath (Gonzalez et al., American Journal of Clinical Nutrition, 2025)
Glucotypes: even non-diabetic people show variable glucose patterns; proposed as risk-identification tool
Critique of glucotypes: label reproducible from mean glucose alone in 92 percent of people; original prevalence did not replicate
Shichiri et al. reported a wearable artificial endocrine pancreas with a needle-type glucose sensor, Osaka University, The Lancet, November 1982
Clark and Lyons built the first glucose enzyme electrode (1962); Updike and Hicks immobilized glucose oxidase in gel (1967)
Abbott FreeStyle Libre received its European CE Mark on September 3, 2014, factory-calibrated, no fingerstick calibration
ADA Standards of Care grade CGM as Grade A for type 1 diabetes and for anyone on insulin
Diabetes Care, ADA Standards of Care 2025, Section 7 Diabetes Technology
David Sinclair joined the metabolic-health company Levels as an advisor in October 2021