What it's actually good for
A continuous glucose monitor is a coin-sized sensor worn on the upper arm that measures glucose in the fluid just under your skin (interstitial fluid, not blood directly) and reports an estimate every few minutes for one to two weeks per sensor. For people with diabetes — especially those on insulin — this is genuinely transformative: it replaces multiple daily fingersticks with continuous data and is backed by decades of validation against blood glucose in that population.
This page grades a different use case: the wave of over-the-counter CGMs (Dexcom Stelo, Abbott Lingo, Abbott Libre Rio, and apps like Levels that layer coaching on the same hardware) cleared by the FDA in 2024 for people not on insulin, marketed heavily at people without diabetes who want to "optimize" metabolic health. That's a different evidence question than diabetes management — the readings only matter if they mean something in a body that already regulates glucose normally, and if looking at them changes anything.
What the research says
Accuracy against the gold-standard test (Grade C). HbA1c — a blood test reflecting average blood sugar over roughly three months — is the clinical benchmark for glucose control. A 2025 Mass General Brigham analysis of 972 adults found CGM-derived average glucose closely tracked HbA1c in people with type 2 diabetes, but that relationship weakened substantially in prediabetes and was essentially absent in people with normal glucose regulation. In plain terms: the device works as intended in the population it was originally built for, and its numbers stop meaning what they appear to mean once you no longer have a glucose problem.
Does watching it change your weight or metabolic health (Grade C). This is the marketing claim on nearly every OTC CGM box, and it's the thinnest part of the evidence. A 2024 systematic review and meta-analysis pooling 25 RCTs and nearly 3,000 participants found CGM-based feedback produced non-significant effects on weight and BMI overall, with a modest HbA1c benefit driven almost entirely by trials in people who already had diabetes. Only 3 of the 25 trials involved non-diabetic overweight or obese adults — not enough to conclude the device itself moves the needle for that group. One small 8-week trial did find a slightly larger weight loss in a CGM-plus-coaching arm versus coaching alone, but both groups got the same dietary education, so it doesn't cleanly separate "the sensor helped" from "eight weeks of any structured feedback helps a little."
Precision of individual readings (Grade C). A crossover trial in 18 healthy adults found a modern CGM (Dexcom G6) performed well at rest (under 6% error) but degraded to roughly 11% error immediately after a glucose drink — not because the sensor malfunctioned, but because interstitial fluid lags blood glucose by 5-10 minutes as it diffuses across tissue. That lag barely matters for spotting a multi-hour trend. It matters more if you're staring at the number 20 minutes after a meal and treating it as a precise verdict on that meal.
Who this is actually for
If you have diabetes or prediabetes, a CGM under clinical guidance is well-supported, and this grade doesn't apply to that use case. Without a diagnosed glucose problem, the honest framing is: a curiosity tool that can show real, individual glucose responses to specific foods or stress — motivating for some — but no trial has shown that owning one, by itself, improves metabolic health when glucose regulation is already normal.
Safety & interactions
Physical risk is low — adhesive skin irritation is the main complaint. The FDA-cleared OTC devices are explicitly indicated for adults not on insulin and are not validated for insulin-dosing decisions. Because of the interstitial lag described above, don't treat a single reading taken during a rapid glucose swing (right after eating, mid-workout) as precise. This is informational, not medical advice — talk to a clinician before using CGM data to make health decisions, especially if you have or suspect a metabolic condition.
How we picked the brand
A CGM earns a spot here on FDA clearance for the specific non-insulin, no-prescription use case (not a diabetes-only device repurposed off-label), a multi-week wear time, and a manufacturer with a long independent validation track record. Stelo is Dexcom's purpose-built OTC device for exactly this population, cleared in March 2024.