Continuous glucose monitors went from prescription-only diabetes devices to over-the-counter wellness products in the span of about eighteen months. Dexcom Stelo was cleared by the FDA in 2024 for adults not on insulin. Abbott's Lingo followed. And a whole tier of subscription apps — Levels, Nutrisense, January AI, Signos — is now built on top of pharmacy-dispensed sensors, promising to turn glucose data into personalized nutrition coaching.
The marketing is confident. The evidence for non-diabetic use is more complicated. This piece is a straight look at who actually benefits from wearing a CGM without diabetes, who probably does not, and what the honest state of the evidence looks like as of now.
What a CGM does and does not measure
A continuous glucose monitor is a small filament sensor inserted just under the skin on the back of the upper arm. It reads glucose concentration in interstitial fluid — the fluid between your cells — every one to five minutes and sends the data to your phone by Bluetooth. Stelo sensors last fifteen days; Abbott's Lingo lasts fourteen. Neither requires fingerstick calibration.
A CGM measures one thing: glucose. It does not measure insulin, which is what actually rises first in insulin resistance. It does not measure fructose meaningfully — fructose is metabolized primarily in the liver and does not spike blood glucose the way glucose does. It does not measure inflammation, cortisol, or any of the other pieces of the metabolic-syndrome puzzle. When people talk about a CGM as a "metabolic health tool," they mean one piece of metabolic health rendered continuously in high resolution.
That distinction matters because the strongest use case for CGM in a non-diabetic population is behavioral, not diagnostic. You are not going to catch a disease with a CGM that a good annual physical would miss. You are going to see, in real time, how your own body responds to things you already do.
The evidence gap for healthy adults
Let us be direct about what the peer-reviewed literature does and does not show.
What is well established: in diabetic and prediabetic populations, CGM use is associated with improved glycemic control (lower HbA1c) and reduced hypoglycemia. This is the basis for insurance reimbursement and clinical guidelines. The DIAMOND, GOLD, and Mobile studies are the anchor trials here.
What is not established: in metabolically healthy adults with normal BMI, normal HbA1c, and no family history of diabetes, no randomized controlled trial has shown that wearing a CGM leads to better cardiovascular outcomes, longer lifespan, better body composition, or any hard clinical endpoint. That is not a gotcha — it is the current state of the evidence.
Observational work suggests glycemic variability may be associated with cardiovascular risk in some populations, and post-meal spikes above 140 mg/dL have been linked to oxidative stress markers in mechanistic studies. But "associated with" and "causes when we lower it" are different claims. The intervention studies that would answer the second question have not been done in healthy adults.
This is a fair note to make when someone claims a CGM will "optimize your longevity." The device gives you real-time information about your physiology. What that information is worth for a healthy person depends on what they do with it. That is not the same as saying it extends lifespan.
Who probably benefits
Here are the profiles where the case for wearing a CGM for a month or two is strongest.
Anyone with prediabetes or metabolic-syndrome markers
The CDC estimates roughly 88 million American adults — about one in three — have prediabetes, and around 80 percent do not know it. If your last HbA1c was between 5.7 and 6.4 percent, or your fasting glucose has been in the 100 to 125 mg/dL range, or you have elevated triglycerides and low HDL and central adiposity, a CGM is more than a curiosity. It is a behavior-change tool for exactly the population where CGM data has the strongest mechanistic rationale.
For this profile, the highest-leverage learning is usually about post-meal spikes — which foods drive them, which combinations flatten them, and how much movement after eating cuts them down. A month of Stelo plus a fasting insulin lab (which measures the piece a CGM cannot see) is a reasonable metabolic self-assessment for anyone in this bucket.
If a CGM feels like an overcommitment, an HbA1c test kit gives you a three-month backward-looking average glucose from a fingerstick and is a good first data point.
Endurance athletes titrating fuel
For a marathoner, cyclist, or triathlete, glucose during and around training is not a wellness metric — it is fuel logistics. Was the gel taken at the right time? Did the pre-workout meal actually stabilize glucose, or did it produce a reactive dip mid-effort? For an athlete testing a race-day nutrition strategy, a CGM turns qualitative guesses into a curve you can look at.
This use case is time-limited by design. You wear one for a training block, learn the pattern for the specific fueling protocol you will use, and take the sensor off. The value is real. The utility is transactional.
Anyone with a family history of Type 2 diabetes
Type 2 diabetes has a strong hereditary component — first-degree relative risk is roughly two to four times higher. If you know Type 2 runs in your family and you are in the age bracket where prediabetes typically starts (35 and up for many populations, earlier for South Asian and Hispanic populations), a month of CGM data is a legitimate screening exercise. You will see whether your postprandial responses look normal or whether variability is creeping into a range that warrants clinical follow-up.
People trying to reverse insulin resistance
Behavior change is easier when you can see the effect. A CGM shows you within an hour whether adding a walk after dinner blunts the spike, whether cutting the third glass of wine improves the fasting number three days later, whether swapping oatmeal for eggs changes anything. For someone who has been told to "lose weight and eat better" and does not know where to start, that visibility often does more than another lecture.
Who probably does not benefit
Being honest about who a device is not for is at least as important.
Metabolically normal adults with no risk factors
If your BMI is in the normal range, your HbA1c is under 5.7, your fasting glucose is under 100, your fasting insulin is under 10, and you have no family history — a CGM is unlikely to change how you eat or move in a way that improves an already-healthy outcome. Interesting to wear once out of curiosity, sure. A recurring subscription-priced habit, probably not.
Anyone with a history of disordered eating or food anxiety
This deserves a plain statement. Continuous data invites continuous monitoring. For someone with a history of anorexia, bulimia, orthorexia, or food-related OCD tendencies, a CGM can worsen the relationship with food rather than improve it. Watching a number every hour and treating each spike as a moral verdict is a documented pattern in this population. If eating already causes you meaningful anxiety, wearing a CGM will very likely amplify it. Talk to a clinician who knows your history before you order one.
People who will not change behavior in response to the data
A CGM is a behavior-change tool. If you know from experience that you look at fitness data, feel briefly guilty, and change nothing, the CGM will follow the same pattern. That is not a moral failing. It is just information about whether this particular intervention will work for you.
Kids, pregnant users, and anyone with severe adhesive sensitivity
The OTC clearance is for adults not on insulin. Pediatric use should be clinician-directed. Pregnancy has different reference ranges. Adhesive sensitivity is common enough that Stelo publishes a specific advisory about it.
The subscription apps — Levels, Nutrisense, and friends
A whole tier of consumer apps now sells CGM-plus-coaching subscriptions. Levels, Nutrisense, January AI, and Signos are the biggest names. Each pairs a pharmacy-dispensed Dexcom G7 or Abbott Libre sensor with a proprietary app that scores meals, suggests substitutions, and layers coaching on top.
Two honest observations about this category.
First, the hardware they ship is the same Dexcom or Abbott sensor you can buy over-the-counter. The accuracy specifications flow through from the underlying sensor — the app cannot make a sensor more accurate than its native readings. If the value proposition is the app layer, that is a software question, not a sensor question.
Second, most of these apps charge $200 to $400 per month once you factor in both the subscription and the sensor cost. That is real money. For someone who wants coaching, community, and a curated learning curve, it may be worth it. For someone comfortable reading the native Dexcom Stelo or Abbott Lingo app and thinking for themselves, most of what these subscriptions add is packaging.
Neither of these observations is a criticism — some people learn better with a coach. But the underlying accuracy is not different, and the price differential is not small.
The 30-day non-diabetic protocol, briefly
If you decide a CGM makes sense for your profile, here is the minimum-viable structure.
- Week 1: baseline. Live and eat normally. Do not react to individual readings. Note your fasting glucose, your typical daytime range, and your highest spikes.
- Week 2: food pairings. Test three meals you eat regularly against simple alternatives. Same meal, one variable changed (protein added, walk after, later timing). Two trials each minimum.
- Week 3: movement. Post-dinner walk on some nights, not others. Compare peak spike and time-to-baseline.
- Week 4: sleep and stress. Track sleep duration and note the next day's fasting glucose and post-meal responses. Watch what happens during stressful work windows.
Two Stelo sensors cover 30 days for around $89. At the end of the month, you should have a short list of foods that consistently spike you, a sense of how much movement moves the needle, and enough self-knowledge to make three or four durable changes. If you can, pair the CGM month with a fasting insulin lab draw — that gives you the piece of the insulin-resistance picture the CGM cannot see.
The honest bottom line
A CGM is a well-validated sensor giving you real information about one metabolic axis. For prediabetics, people with family history, athletes titrating fuel, and anyone actively trying to reverse insulin resistance, that information is genuinely useful. For metabolically normal adults with no risk factors, it is more curiosity than intervention. For anyone with a history of eating disorders, it can actively hurt.
Wearable technology tends to promise more than it delivers. CGM is one of the rare cases where the sensor itself is honest — Dexcom's MARD is documented, the FDA clearance is real, the safety history is long. What is oversold is the leap from "you now have a glucose graph" to "you now have a longer, better life." That leap has not been established in healthy adults, and pretending otherwise is not doing the reader any favors.
Use the tool if the profile fits. Skip it if it does not. And if you do wear one, wear it with structure, learn the patterns, and take the sensor off when the learning tapers. A month of directed attention beats a year of passive data.