in pictures and a few words
What it is and why it matters
A glucose sensor is a small disc on the skin. A thin filament under it measures sugar on its own every 1–5 minutes, day and night, and sends the numbers to your phone.
- Instead of one number from a finger prick — the whole day on a graph
- Insertion feels like a light click; after that the filament is usually not felt
- One sensor is worn for 10 to 15 days
Time in range
The main number is how much of the day sugar stays within the right limits. In diabetes outside pregnancy, the goal is more than 16 h 48 min a day.
- The right limits are 3.9–10.0 mmol/L (70–180 mg/dL)
- Below 3.9 mmol/L (70 mg/dL) — less than 58 minutes a day
- More time in range usually means a lower long-term average sugar test
Type 1 diabetes
Together with an insulin pump, the sensor works as a “smart” system: it adjusts background insulin by itself based on the sensor readings.
- In adults with a sensor, average sugar over months was lower and less time was spent with low sugar
- Some systems pause insulin by themselves if sugar is about to drop
Type 2 diabetes and prediabetes
The sensor shows how specific food, movement and stress change your sugar. A clear graph helps you change habits.
- Without insulin, even a short course of 2–4 weeks is useful
- In prediabetes, two weeks with a sensor can help you spot hidden sugar rises
When low sugar isn’t felt
Some people with long-standing type 1 diabetes stop feeling low sugar. Some sensors have an alert about 20 minutes before sugar is forecast to drop below the threshold.
- With a sensor, episodes of very low sugar became much less frequent
Pregnancy
In type 1 diabetes, the sensor is put on from the start of pregnancy and worn until delivery. The goal here is stricter: 3.5–7.8 mmol/L (63–140 mg/dL).
- Very large babies are born less often
- Low sugar in newborns happens less often
- After delivery, the sensor helps the mother avoid low sugar, especially while breastfeeding
Children and teenagers
Children have the same goal as adults. In a large observation of children and adults with type 1 diabetes, those who wear a sensor have lower average sugar.
- For teenagers, alerts are set more gently for the first weeks so they don’t wear them out
- The model is chosen by the child’s age: one of them is not yet indicated for children
How to start
Decide why you need the sensor and wear it for at least 14 days — that’s how you see the full picture.
- Write down what and when you eat, and compare it with the graph
- Show the 14-day report to an endocrinologist
- Protective spray before insertion and changing the site every 8–14 days protect the skin
Who should see a doctor first
Sugar repeatedly below 3.0 mmol/L (54 mg/dL) or above 8.9 mmol/L (160 mg/dL) — see a doctor. Below 3.9 mmol/L (70 mg/dL) and not matching how you feel — be sure to recheck with a glucometer.
- On insulin: for the first 12 hours don’t change the dose by sensor numbers — watch the arrow
- In some sensors, paracetamol over 1 g and vitamin C over 500 mg distort the numbers
- Rash, itching or blisters beyond the patch edge after 1–3 days — see a dermatologist
Important. Check with a glucometer also with swelling late in pregnancy. Dehydration distorts the numbers too. Dialysis, severe kidney disease — use the sensor with caution.
In one sentence
The sensor shows sugar all day, not one point: this makes it easier to keep it in range, miss low sugar less often, and choose treatment together with a doctor.
Key points
Continuous glucose monitoring (CGM) measures glucose every 1–5 minutes — up to 288 data points a day instead of a single fingerstick reading. Here is where that helps.
| Where it helps | How it helps | What human studies showed |
|---|---|---|
| Type 1 diabetes | a continuous picture of glucose; with an insulin pump — a hybrid closed-loop system: the algorithm adjusts basal insulin automatically from sensor data | FLASH-UK (NEJM 2022): in adults with type 1 diabetes, HbA1c was 0.5 percentage points lower and time in hypoglycemia 43 minutes a day shorter |
| Type 2 diabetes on insulin | shows how specific foods, physical activity and stress affect glucose | — |
| Pregnancy with type 1 diabetes | CGM from confirmation of pregnancy until delivery; a narrow target range of 3.5–7.8 mmol/L (63–140 mg/dL) | CONCEPTT RCT (n=325), CGM vs self-monitoring: large-for-gestational-age infants 53% vs 69%, neonatal hypoglycemia 15% vs 28%, the infant’s hospital stay one day shorter |
| Children and adolescents with type 1 diabetes | the same TIR target as adults; alarms for nocturnal hypoglycemia | T1D Exchange registry (n>22,000, children and adults): HbA1c lower in CGM users than in non-users |
| Impaired awareness of hypoglycemia in type 1 diabetes | real-time CGM (Dexcom G5 in HypoDE) | HypoDE RCT (n=149, 26 weeks): episodes with glucose at or below 3.0 mmol/L (54 mg/dL) fell from a mean of 10.8 to 3.5 per 28 days, 72% fewer |
| Prediabetes and metabolic health | a two-week course shows your own response to foods, hidden glucose spikes, and the effect of sleep, stress and exercise | — |
How it helps
How the sensor works
A glucometer gives a single value at the moment of the fingerstick — like assessing the weather from one snapshot a day. A CGM sensor is a miniature electrochemical biosensor inserted under the skin (typically on the back of the upper arm or the abdomen). A thin filament electrode (5–7 mm long, <0.4 mm in diameter) is coated with the enzyme glucose oxidase, which catalyzes the oxidation of glucose in interstitial fluid; the resulting current is proportional to glucose concentration.
Data is transmitted via Bluetooth or NFC to a smartphone or reader. CGM measures glucose in interstitial fluid, not in blood: there is a physiological lag of 5–15 minutes between the two (lag time). The technology has expanded well beyond type 1 diabetes: CGM is used in type 2 diabetes and prediabetes, and even healthy people use it to optimize metabolic health. The CGM market reached $8.2 billion in 2024 and is projected to reach $15 billion by 2028.
Time in range
An international consensus (Battelino et al., Diabetes Care, 2019) set targets for people with diabetes outside pregnancy (pregnancy targets are below). Optimal time in the 3.9–10.0 mmol/L (70–180 mg/dL) range is above 70%, i.e. more than 16 h 48 min a day.
| Metric | Target |
|---|---|
| Time in range 3.9–10.0 mmol/L (70–180 mg/dL) | >70% |
| Hypoglycemia: below 3.9 mmol/L (70 mg/dL) | <4% (<58 minutes) |
| Severe hypoglycemia: below 3.0 mmol/L (54 mg/dL) | <1% |
| Hyperglycemia: above 10.0 mmol/L (180 mg/dL) | <25% |
| Severe hyperglycemia: above 13.9 mmol/L (250 mg/dL) | <5% |
Each 5% increase in TIR is associated with an approximately 0.5% reduction in HbA1c (Beck et al., Annals of Internal Medicine, 2019). TIR >70% correlates with HbA1c <7%.
Glucose stability and the AGP report
Coefficient of variation (CV) is the key stability indicator: CV <36% means stable glycemia, above 36% unstable, with an increased risk of hypoglycemia. Monnier et al. (Diabetes Care, 2017) showed that glycemic variability is an independent risk factor for cardiovascular complications of diabetes.
Ambulatory Glucose Profile (AGP) is a standardized report that combines 14 days of data into one picture: median, 25th and 75th, and 5th and 95th percentiles.
Type 1 and type 2 diabetes
CGM + insulin pump = hybrid closed-loop system: the algorithm automatically adjusts basal insulin based on CGM data. AID systems (automated insulin delivery — Tandem Control-IQ, Medtronic 780G, Omnipod 5) suspend basal delivery when hypoglycemia is predicted and reduce its frequency by 30–50%. FLASH-UK results in type 1 diabetes are in “Key points” (PMID: 36198143).
In type 2 diabetes, CGM helps you see how specific foods, physical activity and stress affect glucose.
Prediabetes and metabolic health
A two-week CGM course lets you identify individual responses to foods, detect hidden glucose spikes (reactive hypoglycemia, nocturnal rises), and assess the effect of sleep, stress and physical activity, while visual feedback motivates changes in eating behavior.
Pregnancy
Pregnancy with type 1 diabetes carries a twofold risk of congenital anomalies, preeclampsia and fetal macrosomia. The multicenter CONCEPTT RCT compared CGM with standard self-monitoring (outcomes are in “Key points”); HbA1c in the CGM group was 0.19% lower by week 34 (PMID: 28923465).
Targets in pregnancy with T1D (Battelino consensus, 2019): a range of 3.5–7.8 mmol/L (63–140 mg/dL), time in it >70% (at least 16 h 48 min a day), above range <25%, below range <4%, including <1% below 3.0 mmol/L (54 mg/dL) (PMID: 31177185).
CGM is started once pregnancy is confirmed (4–6 weeks) and continued until delivery. After delivery, insulin requirements fall by 30–50% in the first 24–48 hours — CGM helps avoid postpartum hypoglycemia, especially during breastfeeding. In gestational diabetes, observational data show a benefit of CGM with insulin therapy. FreeStyle Libre 2/3 and Dexcom G6/G7 are approved for use in pregnancy (the FDA lifted the off-label status for Dexcom G7 in 2024).
Children and adolescents
In children and adolescents with T1D the target is >70% of time in the 3.9–10.0 mmol/L (70–180 mg/dL) range, as in adults; it is harder to reach because of pubertal hormonal fluctuations, physical activity and eating behavior. Data from the T1D Exchange registry are in “Key points” (PMID: 30657336).
Dexcom G7 is approved from age 2, FreeStyle Libre 3 from age 4 (officially from 14 in some countries; off-label use from age 4 is common). In children under 6 the sensor is placed on the buttocks, away from abdominal fat: in lean children subcutaneous tissue is thinner than 5 mm, which raises the risk of the sensor detaching. For adolescents, alarm thresholds are individualized for the first 2–4 weeks: low 3.9 mmol/L (70 mg/dL), high 13.9 mmol/L (250 mg/dL) instead of the standard 4.4 and 10.0 mmol/L (80 and 180 mg/dL).
When hypoglycemia goes unnoticed
Impaired awareness of hypoglycemia (IAH) develops in 20–25% of people with T1D with a diabetes duration over 10 years because of an impaired counterregulatory response of the sympathetic nervous system, and it raises the risk of severe hypoglycemia 6-fold. It is assessed with the Gold and Clarke scales (≥4 points — IAH present). In the multicenter HypoDE RCT in people with T1D and IAH, CGM (Dexcom G5) reduced episodes with glucose at or below 3.0 mmol/L (54 mg/dL) by 72% (p<0.0001) (PMID: 29459019).
Predictive low glucose alerts on Dexcom G7 and FreeStyle Libre 3 fire 20 minutes before a forecast drop below the threshold; the algorithm relies on the rate of change and the trend. Sensitivity is about 85%, specificity 75%. ADA 2024 Standards of Care place people with IAH in the category for whom CGM is mandatory (category A) and AID integration desirable (category B).
How to use it
1. Define your goal: diabetes management, metabolic optimization, ketosis monitoring 2. Choose a system: FreeStyle Libre 3 (best value), Dexcom G7 (best pump integration) 3. Set the period: at least 14 days for a complete picture 4. Keep a food diary: match meals with glycemic responses 5. Interpret with your doctor: share the AGP report with your endocrinologist
Available systems (2025–2026)
- FreeStyle Libre 3 (Abbott) — 21.2 mm, the smallest CGM in the world; 14 days; a reading every minute, continuous Bluetooth Low Energy transmission; MARD 7.9% (one of the best figures); factory calibration; customizable hypo- and hyperglycemia alerts; IP27 water resistance (submersion to 1 m for 30 minutes).
- Dexcom G7 — 28 mm (60% smaller than G6); 10 days + a 12-hour grace period; every 5 minutes, Bluetooth; MARD 8.2% (adults); no calibration required; integration with Omnipod 5 and Tandem Mobi pumps; trend arrows (↑↑, ↑, →, ↓, ↓↓) show the rate of change.
- Dexcom Stelo (2025) — the first CGM cleared by the FDA for people not on insulin (type 2 diabetes without insulin, prediabetes); 15 days; over the counter (OTC) in the US; a simplified interface without hypoglycemia alerts.
- SiBio KS1 — the world's first continuous ketone monitor: 14 days, measures beta-hydroxybutyrate (BHB) in interstitial fluid; ideal for monitoring ketosis (ketogenic diet, intermittent fasting), and combined with CGM gives a complete metabolic picture.
Accuracy and trend
Dexcom G7 with factory calibration delivers MARD 8.2% from activation (30-minute warm-up). The ISO 15197:2013 standard requires glucometers to have 95% of values within ±0.8 mmol/L (±15 mg/dL) at glucose ≤5.6 mmol/L (100 mg/dL) or within ±15% above 5.6 mmol/L (100 mg/dL); modern CGMs meet this standard, but not in all ranges (see “Who should see a doctor first”).
The lag between capillary and interstitial glucose is 5–10 minutes when glycemia is stable and 15–20 minutes during rapid changes (after meals, during exercise, during hypoglycemia). Pressure on the sensor during sleep can cause compression false lows.
Skin under the sensor
Contact dermatitis at the insertion site occurs in 5–18% of users: irritant in 65% of cases, allergic in 35%. The main allergen is isobornyl acrylate (IBOA), a component of the FreeStyle Libre adhesive (PMID: 28804907). Additional allergens are N,N-dimethylacrylamide and colophony.
- A barrier spray (Cavilon No-Sting, Marathon, SkinTac) before insertion reduces the rate of reactions by 40–60%.
- A hydrocolloid underlay (DuoDerm Extra Thin, Hydrocolloid) with a slit for the sensor isolates the skin from the adhesive.
- Class III topical corticosteroids (mometasone 0.1%, methylprednisolone aceponate 0.1%) — a short 5–7-day course; do not apply directly under the sensor (it impairs adhesion).
- Rotating insertion sites (forearm ↔ abdomen ↔ buttocks) every 8–14 days reduces local sensitization. Stopping CGM altogether is rarely required.
Who should see a doctor first
Choosing a CGM system and interpreting the data are done together with your doctor. In particular:
- When to see a doctor. If glucose is repeatedly below 3.0 mmol/L (54 mg/dL) or above 8.9 mmol/L (160 mg/dL), see a doctor (the author’s rule).
- Low readings and a mismatch with how you feel. During hypoglycemia CGM may show a value higher than the real one — confirmation with a glucometer is mandatory at values <3.9 mmol/L (70 mg/dL) and when readings contradict symptoms. Accuracy is lower at glucose <3.9 mmol/L (70 mg/dL) and >16.7 mmol/L (300 mg/dL).
- The first hours after insertion. Readings are less accurate in the first 24 hours (stabilization period); do not adjust insulin based on absolute values in the first 12 hours, follow the trend. When calculating an insulin bolus, the trend (up/down arrow) is taken into account, and during a rapid rise the CGM value + 20–30% correction is used.
- What distorts readings. Paracetamol >1 g (Dexcom G6 and older — falsely high values; G7 has an improved electrode with reduced sensitivity), ascorbic acid >500 mg (FreeStyle Libre — falsely high), dehydration (reduced interstitial perfusion — falsely low). Hemodialysis and marked swings in hematocrit also affect accuracy; in stage 5 CKD, CGM is used with caution.
- Pregnancy. With third-trimester edema accuracy may decrease — MARD rises from 8% to 10–12%; if readings seem to diverge from symptoms, confirmation with a glucometer is recommended.
- Children. Dexcom Stelo is not yet indicated for children. In children with obesity and prediabetes there is no routine indication for CGM: prediabetes and T2D in children are diagnosed by HbA1c, OGTT and fasting glucose.
- Skin. Erythema, itching and vesicles over an area larger than the patch outline appearing after 24–72 hours are signs of allergy (delayed type IV hypersensitivity): patch testing by a dermatologist and switching systems (Dexcom uses a different adhesive formula than Abbott). Mechanical irritation follows the outline exactly, without itching; tenderness and a pustule point to infection (folliculitis, S. aureus); cholinergic urticaria is also ruled out.
Bottom line
CGM turns a single reading into a continuous picture: you see time in range, glucose stability and your response to food, sleep and exercise. In human studies, with CGM HbA1c fell and hypoglycemia became less frequent, and in pregnant women with type 1 diabetes large-for-gestational-age infants and neonatal hypoglycemia were less common. A complete picture needs at least 14 days, and the AGP report is worth reviewing with an endocrinologist.
References
- PMID: 36198143. PMID 36198143
- PMID: 28923465. PMID 28923465
- PMID: 31177185. PMID 31177185
- PMID: 30657336. PMID 30657336
- PMID: 29459019. PMID 29459019
- PMID: 28804907. PMID 28804907
Key facts
- CGM measures interstitial fluid glucose every 1–5 minutes — up to 288 data points a day instead of a single fingerstick reading.
- The international consensus target (Battelino, 2019) for people with diabetes outside pregnancy is more than 70% of the day (over 16 h 48 min) in the 3.9–10.0 mmol/L (70–180 mg/dL) range; each +5% of this time is associated with an approximately 0.5% reduction in HbA1c.
- In the HypoDE RCT (n=149) in people with type 1 diabetes and impaired awareness of hypoglycemia, CGM reduced episodes with glucose at or below 3.0 mmol/L (54 mg/dL) by 72% over 26 weeks — from a mean of 10.8 to 3.5 per 28 days.
- In the CONCEPTT RCT (n=325) in pregnant women with type 1 diabetes, CGM compared with self-monitoring reduced large-for-gestational-age infants from 69% to 53% and neonatal hypoglycemia from 28% to 15%.
- FreeStyle Libre 3 is the smallest CGM (21.2 mm), MARD 7.9% accuracy, 14-day wear; Dexcom G7 has MARD 8.2% and integrates with insulin pumps.



