Reading a continuous glucose monitor — a practical guide
Stelo, Lingo, Levels, Nutrisense. Three months of CGM data will change what you eat more than any book. Here's how to read it.
A continuous glucose monitor (CGM) is a small sensor that sits on your upper arm for 10-14 days and measures your interstitial glucose roughly every five minutes. Until recently CGMs required a prescription — as of 2024 in the US, multiple brands sell consumer CGMs over the counter (Stelo from Dexcom, Lingo from Abbott, Levels, Nutrisense, Signos).
What a healthy glucose curve looks like
- Fasting (overnight): 70-95 mg/dL
- Post-meal peak: under 130 mg/dL ideal, under 140 acceptable
- Post-meal low: should not crash below fasting baseline
- Time-in-range (70-140): >90% is excellent, >85% is good
- Variability: flatter is better than both high-and-crashing
What CGM data will show you that's hard to believe
After two weeks of data, nearly every wearer discovers things like:
- Oatmeal alone spikes harder than you think. Oatmeal with walnuts and blueberries barely moves the needle.
- White rice with stir-fried vegetables spikes less than white rice alone. Protein and fat flatten the curve.
- Sweet potato with butter produces a flatter response than sweet potato with skin removed. Fat slows absorption.
- An apple eaten whole produces a smaller response than the same apple juiced.
- A 20-minute walk after a meal reduces peak glucose by 20-30 points.
- Your morning coffee probably spikes you (cortisol-mediated).
- Sleep deprivation tanks your glucose tolerance the next day.
- Stress spikes glucose without food.
The three levers for flatter curves
- Pair carbs with protein + fat + fiber at the same meal. Never eat a carb in isolation.
- Eat the vegetables first. Literally. Same meal, different order = measurably different glucose response.
- Walk 10-20 minutes after the larger meal of the day. Muscle uptake is the fastest way to clear glucose from blood.
What a CGM is not
A CGM is not a medical device for non-diabetics — it's an information tool. Glucose is one biomarker among many. Over-optimizing for flat glucose at the expense of sleep, movement, or social connection is a trap. Use the data to calibrate meals, then take the sensor off and go live.
Recommended: a 14-day self-experiment
Wear one sensor for 14 days. Eat your normal foods. Note what spikes you and what doesn't. Throw out your assumptions about which foods are "healthy" for you — almost everyone has a few personalized surprises. Then cook from the Glucose Steady chapter of this site for the next month with the knowledge in hand.