Glucose monitors for healthy people: data without a diagnosis

A small sensor on the upper arm, a curve on the phone, a red alert after a bowl of pasta. Continuous glucose monitors (CGMs) were built to keep people with diabetes safe from dangerous highs and lows. Since 2024 they can be bought without a prescription in the United States, and wellness companies now sell them to people with perfectly normal blood sugar as a window into their "metabolic health". The data look precise. The question is whether they tell a healthy person anything useful — and whether acting on them does any good.

What normal glucose actually looks like

The best reference for healthy people comes from a multicentre study by Shah and colleagues, published in 2019. Researchers fitted 153 people without diabetes, aged 7 to 80, with a blinded CGM (they could not see the readings) for up to ten days while they lived normally.

Two points stand out. First, healthy glucose is not a flat line: its coefficient of variation (a measure of how much the curve moves around its average) was about 17 %, and average glucose was slightly higher in people over 60 (104 mg/dL). Second, brief readings above 140 mg/dL and short dips below 70 mg/dL are part of normal physiology. A healthy person who starts wearing a sensor will almost certainly see "spikes" and "lows" that are entirely ordinary.

How accurate are the sensors?

CGMs do not measure blood. They measure glucose in the fluid between cells under the skin, which lags behind blood glucose, especially when levels change quickly after a meal or during exercise. Accuracy is usually reported as MARD — the mean absolute relative difference, i.e. how far, on average, sensor readings are from a reference measurement. The Dexcom G6 used in the Shah study has a reported MARD of about 9 %. That is good enough to guide insulin dosing, but in the narrow healthy range a 10 % error is large relative to the differences people try to interpret.

Recent studies in healthy people make this concrete. In a randomised crossover trial of 15 healthy adults published in 2025, a CGM systematically overestimated glucose compared with finger-prick measurements after standard meals, and the size of the error varied between meals and between individuals. In a study of 394 healthy adults who wore two sensors at the same time, two devices of the same brand agreed closely, but devices from two different brands disagreed much more: in 19 % of comparisons, a meal placed in a person's "best" fifth by one sensor was placed in the "worst" fifth by the other. (That study was funded by a personalised-nutrition company.) A 2024 review in Diabetic Medicine illustrated the problem: with a 20 % accuracy margin, a truly normal reading could display anywhere from 2.6 to 9.4 mmol/L.

The "spike" problem: no outcome evidence

Glucose rises after a carbohydrate-containing meal and falls again; in people without diabetes, insulin takes care of this. Much CGM marketing implies that flattening these post-meal rises improves energy, weight or long-term health. There are no trials showing that reducing post-meal glucose peaks in healthy people prevents diabetes, heart disease or anything else. The 2024 review, which examined 25 studies, found a lack of consistent, high-quality evidence for CGM use in people not living with diabetes, both on accuracy and on health benefit.

There is also a psychological cost to consider. The same review warned that constant feedback can create anxiety about what counts as "normal", lead people to cut out foods unnecessarily, and could contribute to disordered eating such as orthorexia — an unhealthy fixation on eating "correctly". A curve that reacts to a banana can turn ordinary food into a source of worry.

Regulation: what has actually been cleared

In the US, the FDA cleared the first over-the-counter CGM, Dexcom's Stelo, on 5 March 2024, for adults who do not use insulin — including people without diabetes who want to see how diet and exercise affect their glucose. The FDA noted that it is not for people with problematic hypoglycaemia and that users should not make medical decisions based on it without talking to a healthcare provider. Abbott's Lingo (a wellness product, not intended to diagnose diabetes) and Libre Rio (for adults with type 2 diabetes not using insulin) followed on 10 June 2024. These are US clearances; clearance means the device is acceptably safe and performs like comparable devices, not that it improves health in healthy users.

In Germany, statutory health insurance has covered real-time CGM since a 2016 decision of the Federal Joint Committee (G-BA) — for people with insulin-dependent diabetes who need intensified insulin therapy. Wellness use is a private purchase.

Where a CGM might help

The most plausible use is short-term education in people with prediabetes or overweight: seeing how meals, walks and sleep affect glucose may motivate change. The evidence is thin. A 2024 meta-analysis of 25 randomised trials found modest benefits of CGM feedback overall, but only three small trials (130 participants in total) were in people without diabetes, there was no separate estimate for their HbA1c, and the pooled effect on weight (−0.7 kg) was not statistically significant.

This article is general information and not individual medical advice. Decisions about tests and treatment belong in a conversation with your own physician.