Across the Vyvata catalog, one product sits alone at the top. Dexcom Stelo scored 91 out of 100 — the only Elite-tier product we have graded so far. That score is not a marketing claim. It is the output of a six-dimension rubric that examines evidence, design, build quality, safety, transparency, and sustainability, with sources cited and a published methodology. When a product earns Elite, we are saying the documentation holds up to scrutiny.
This article is for someone who is curious whether a CGM is worth wearing and wants a straight answer rather than a sales pitch.
The only Elite product in our catalog
Vyvata grades products on a 0 to 100 scale across four tiers: Rejected (0 to 59), Standard (60 to 74), Verified (75 to 89), and Elite (90 to 100). Elite is rare on purpose. To reach 90, a product has to demonstrate independent clinical evidence, manufacturer transparency, regulatory clearance from a credible body, and a documented safety record — not just one of those things. Most of the supplement and device market does not clear the Verified bar, much less Elite. Stelo did.
What a CGM actually does
A continuous glucose monitor is a small disposable sensor worn on the back of the upper arm. A spring-loaded applicator inserts a tiny flexible filament — thinner than a strand of hair — just under the skin. That filament sits in the interstitial fluid between your cells and measures glucose concentration. The sensor sends a reading to your phone every one to five minutes by Bluetooth.
It is not a blood draw. Interstitial glucose lags blood glucose by roughly five to ten minutes, which matters in some edge cases but is fine for the patterns most users care about. Stelo sensors last fifteen days; a two-pack covers a month. There is no fingerstick calibration — you tap the sensor with your phone to pair it, and the app handles the rest.
Why non-diabetics started caring
For decades CGMs were a prescription tool for insulin-dependent diabetics. That changed in 2024, when the FDA cleared Stelo for adults who do not take insulin. Dexcom built the OTC category around a different audience: people who want to understand their metabolic patterns before there is a problem to fix.
The motivating data is uncomfortable. The CDC estimates that roughly 88 million American adults — about one in three — have prediabetes, and around 80 percent of them do not know it. A standard annual physical typically captures a single fasting glucose number and, if you are lucky, an A1C. Neither tells you what happens to your glucose after you eat oatmeal, or after you sleep poorly, or after a stressful meeting.
Two patterns matter most for non-diabetics:
- Glucose variability — large swings up and down across the day. Independent of average glucose, variability is associated with oxidative stress markers and cardiovascular risk in observational studies.
- Post-meal spikes — how high glucose goes after a specific meal, and how long it stays elevated. People respond very differently to the same food, and you cannot guess your own response from a nutrition label.
One honest note: long-term outcome data on CGM use in non-diabetic populations is still thin. Randomized trials showing that healthy adults live longer because they wore a CGM do not yet exist. What the device gives you is information about your own physiology. What you do with it is up to you.
What you will actually see in your first month
Here is what most first-time wearers notice.
The oatmeal surprise
Plain oats with banana and honey — the prototypical "healthy breakfast" — spikes a lot of people into the 160s or 180s mg/dL within forty-five minutes. Add a scoop of protein powder, two eggs on the side, or a tablespoon of nut butter, and the same meal might peak in the 130s. This is not a moral judgment about oats. It is a useful pattern: when you eat fast carbohydrate alone, you get a steeper spike than when you eat it with fat, fiber, and protein.
The workout effect
A twenty-minute walk after dinner often cuts the post-meal peak in half. Resistance training tends to flatten the post-workout glucose curve for hours. High-intensity intervals can briefly raise glucose because of adrenaline, then drop it lower than baseline later. None of this is theoretical when you can see it on the graph.
The post-dinner crash
A big spike is usually followed by a reactive dip. After a late high-carb dinner, you may see a glucose value below your morning fasting number around bedtime — registering as cravings, restlessness, or a poor first hour of sleep.
The cortisol bump
Glucose often rises in the early morning, peaking around six or seven a.m. even if you have not eaten. That is the dawn phenomenon — cortisol-driven hepatic glucose release. It is normal, and seeing it once explains why fasting glucose varies day to day for reasons that have nothing to do with food.
Combined, these four observations are usually enough to change how someone eats and trains, without anyone telling them to.
Why Stelo earned Elite (the dimension breakdown)
Vyvata grades on six dimensions. Here is how Stelo performed on each.
Evidence
Stelo is based on the same sensor platform as the Dexcom G7, which has published clinical accuracy data with a Mean Absolute Relative Difference (MARD) around 8 to 9 percent against laboratory reference. MARD is the standard metric for CGM accuracy: lower is better, and under 10 percent is considered clinically reliable. The accuracy validation studies are peer-reviewed and listed in the FDA 510(k) summary. This is not vendor marketing — it is filed evidence.
Design
One-step applicator. No calibration required. Fifteen-day wear time, the longest in the OTC category. iPhone and Android apps. Waterproof rated for showering and swimming. The friction cost is low enough that most users actually finish a sensor cycle, which is the bar that matters for a behavior-change tool.
Build quality
Manufactured by Dexcom under ISO 13485, the international medical-device quality standard. Stelo has an FDA 510(k) clearance as an integrated continuous glucose monitor for adults not on insulin. The regulatory paper trail is verifiable in the FDA's public database.
Safety
The FDA's MAUDE database tracks adverse event reports for medical devices. Dexcom's platform has millions of patient-years of use logged there, and the complication profile — mostly mild skin reactions at the adhesive site — is well-characterized with no signal of systemic harm. For a wearable medical device, that history is what "safe" looks like.
Transparency
Dexcom publishes its accuracy studies, lists principal investigators, and discloses funding. The privacy policy is readable and clearly explains what data leaves the device. There is no opaque AI-coaching layer reselling user metabolic data to third parties — unusually clean compared to startup CGM-adjacent apps.
Sustainability
This is where Stelo gets the smallest boost. Dexcom offers a sensor recycling program in select regions, but coverage is uneven and the device itself is single-use electronics with adhesive and a small battery. We credit the program honestly and do not pretend a disposable medical sensor is a green product. This is one reason Stelo lands at 91 rather than higher.
The honest caveats
Earning Elite does not mean Stelo solves metabolic health. Several things deserve a flat statement.
- A CGM measures glucose. Nothing else. It does not tell you your insulin level, which is what actually changes first in insulin resistance. For deeper metabolic assessment, fasting insulin and a HOMA-IR calculation from a blood draw remain useful.
- Fructose does not register the same way. Pure fructose is metabolized primarily by the liver and shows less acute glucose response. The CGM will undercount the metabolic impact of high-fructose foods.
- Cost adds up. Stelo is around $89 for a two-pack — roughly one month of continuous wear, or close to $1,000 per year. Most non-diabetic users do not need to wear one continuously; a month or two of targeted data collection is usually enough to learn your patterns.
- Accuracy is lower at extremes. CGM sensors are most reliable in the typical adult range, roughly 70 to 180 mg/dL. At very low or very high values, the gap between sensor and fingerstick widens.
- Data without context can mislead. Glucose responses vary by sleep, stress, cycle phase, and prior meal. A single high reading does not mean a food is "bad" for you. Patterns over many exposures matter.
A 30-day Stelo protocol for a beginner
The worst use of a CGM is reacting to every reading. The best use is structured curiosity. Here is a four-week protocol.
- Week 1 — Observe only. Eat and live as you normally would. Check the app a few times a day, not every hour. Note your fasting number, your typical daytime range, and where your highest spikes happen. Do not change anything yet. Your baseline is the most valuable data you will collect.
- Week 2 — Test foods. Pick three meals you eat regularly and pair them with simple alternatives. Oatmeal versus eggs and avocado. White rice versus rice plus protein and vegetables. A smoothie versus the same ingredients eaten whole. Record peak value and time-to-return-to-baseline for each. Two trials per pairing minimum.
- Week 3 — Test movement. Take a ten- to twenty-minute walk within thirty minutes after dinner on three days, and skip it on the others. Compare the post-meal curves. Try a resistance workout in the morning versus the evening. The point is not optimization yet — it is seeing the effect size of movement on your own glucose.
- Week 4 — Test sleep and stress. Note your sleep duration each night. Look at the next morning's fasting glucose and the next day's post-meal responses. On a stressful workday, watch what happens to glucose during a difficult meeting. This is where most people learn that sleep and stress affect glucose at least as much as food choices.
At the end of the month you will have a short list of foods that spike you, a sense of how movement and sleep affect your numbers, and enough self-knowledge to make a handful of durable changes. That is the win.
Who Stelo is not for
Stelo is cleared for adults who do not take insulin. If you have Type 1 diabetes or insulin-dependent Type 2, you need the prescription Dexcom G7 or equivalent — it has tighter accuracy specifications at low glucose values, alarms, and integration with insulin pumps. Stelo is not a substitute.
If you have a history of eating disorders, or you know that quantified self-data tends to drive anxiety for you, a CGM may not be the right tool. Continuous data invites continuous monitoring. Some people thrive on that. Others find it exhausting and stop sleeping well because they are checking a number at three a.m. Be honest with yourself before you order one.
Children, pregnant users, and anyone with severe skin sensitivity to medical adhesives should talk to a clinician first.
The CGM market and what is coming
Stelo is currently the only Elite product in our catalog, but the OTC CGM category is moving fast.
Abbott's Lingo is the other major OTC entrant, built on the FreeStyle Libre sensor platform with a non-diabetic-focused app. The hardware lineage is strong and the accuracy data is comparable to Dexcom's. When we score it, Lingo has a credible path to Verified or Elite — the dimensions we evaluate will look broadly similar.
Levels Health is a different category: a subscription software layer that pairs with a prescription Libre or Dexcom sensor and adds coaching and nutrient scoring. The hardware grade flows through from the underlying sensor; the question we evaluate separately is whether the software layer adds evidence-backed value over the native Dexcom or Abbott apps.
The broader trajectory is clear: sensor cost is dropping, wear time is getting longer, and the regulatory door for non-diabetic use is open. Stelo is the product that defines what "good" looks like in that future — which is why it sits, for now, alone at the top of the Vyvata score sheet.
If you wear one for thirty days and learn what your body does with breakfast, a walk, a bad night of sleep, and a hard week of work, that is the highest-leverage thing this kind of device can do. The Elite score reflects that Stelo is built honestly enough to be worth trusting with the answer.