What Does a Normal CGM Graph Look Like?

A normal continuous glucose monitor (CGM) graph in someone without diabetes looks like a series of gentle hills and valleys, with readings staying roughly between 70 and 140 mg/dL for the vast majority of the day. In a large multicenter study of healthy adults, the average 24-hour glucose was about 99 mg/dL, and people spent around 96% of their time in the 70-to-140 range.¹ But that smooth-sounding average hides real-world variation from meals, exercise, sleep, and even sensor quirks that can make a perfectly healthy graph look alarming if you don’t know what to expect.

The Baseline Numbers

Studies consistently put the 24-hour mean glucose for healthy, non-diabetic adults somewhere between roughly 89 and 102 mg/dL, depending on the population studied and the sensor used. One early study of 21 healthy volunteers eating their usual diets found a 24-hour mean of about 89 mg/dL.² A later study of 32 people with confirmed normal glucose tolerance reported a mean CGM reading of about 102 mg/dL.³ These numbers are not contradictory. Differences in sensor calibration, study design, and participant demographics explain the spread. What matters is the range they share: for most healthy people, average glucose sits comfortably under 110 mg/dL.

How much the readings bounce around also matters. In healthy adults, the coefficient of variation, a measure of glucose swings relative to the average, tends to land around 17%.¹ A large Chinese reference study of 434 healthy subjects set the upper limit of normal glucose swings (using a metric called MAGE) at about 70 mg/dL, with a typical median swing well below that.⁴ If your CGM graph looks like a gentle rolling landscape rather than a jagged mountain range, your variability is probably fine.

The Overnight Dip

One of the most recognizable features of a normal CGM graph is a smooth, low stretch during sleep. Glucose drifts down at night because you are not eating, your body’s metabolic demands drop, and insulin sensitivity follows a circadian rhythm. The same study that found a daytime average of about 93 mg/dL recorded an overnight average of roughly 82 mg/dL.² Other work puts the overnight average somewhat higher, around 97 mg/dL, but the pattern is the same: nighttime glucose runs noticeably lower than daytime glucose.³

A healthy overnight trace is typically flat to gently declining, sometimes with a small rise in the early morning hours as cortisol kicks in before waking. If your overnight graph shows a sudden sharp drop into the 50s or 60s, that spike of alarm might actually be a sensor artifact rather than real hypoglycemia, a phenomenon covered in the sensor quirks section below.

What Happens After You Eat

Meals create the peaks in your CGM graph. In a healthy person, glucose rises after eating, reaches a peak roughly 30 to 60 minutes after the meal, and returns close to baseline within about two hours. The size of that peak depends on what you ate. A carbohydrate-heavy meal produces a taller, faster spike. Adding protein or fat to the same meal tends to blunt and delay the peak, though the relationship is not as simple as “fat and protein always help.”

A pilot trial in women with gestational diabetes found that a high-fat meal actually produced a higher peak glucose (about 146 mg/dL) and a larger overall glycemic excursion than a high-protein meal (peak around 137 mg/dL), even though both meals contained the same amount of carbohydrate.⁵ Modeling work in healthy adults confirms that the interactions between macronutrients meaningfully change both the shape and height of the glucose curve, so you cannot predict post-meal glucose from carbohydrate content alone.⁶

Even the order in which you eat your food can alter the graph. A systematic review found that eating vegetables, protein, or fiber-rich foods before carbohydrates generally lowered the post-meal glucose spike compared with eating carbohydrates first or eating everything mixed together.⁷ The effect varied across studies, but the direction was consistent. If you want flatter post-meal hills on your CGM graph, eating your salad or protein before your rice or bread is a reasonable strategy.

How Exercise Changes the Curve

If meals create the peaks, exercise creates the valleys. During physical activity, your muscles pull glucose out of the bloodstream at an accelerated rate, and CGM readings drop accordingly. In healthy people without diabetes, aerobic exercise lowered sensor glucose by an average of about 15 mg/dL from the pre-exercise baseline, while resistance exercise produced a smaller drop of about 9 mg/dL.⁸ Neither drop is dangerous in a healthy person, but if you are already at the lower end of normal when you start exercising, the dip can briefly touch the high 60s or low 70s and trigger a low-glucose alert on your device.

Exercise also affects what happens hours later. The same study found that overnight glucose nadirs were slightly lower on nights following resistance exercise compared to aerobic exercise (about 76 vs. 83 mg/dL), suggesting that the type of workout you do can shape your graph well into the next sleep cycle.⁸ A meta-analysis of exercise interventions in non-diabetic populations confirmed that exercise type and participant age both influence how the 24-hour glucose profile shifts.⁹

Sensor Quirks That Can Fool You

Not every alarming dip on a CGM graph reflects what is actually happening in your blood. CGMs measure glucose in the fluid between your cells (interstitial fluid), not directly in the bloodstream. There is a built-in physiological delay of roughly 5 to 6 minutes for glucose to move from blood to interstitial space in healthy adults.¹⁰ In people with type 1 diabetes, that delay stretches to under 10 minutes.¹¹ This lag means that during a rapidly rising glucose spike after a meal, your CGM reading slightly trails reality, and during a rapid drop, it reads a bit higher than your actual blood glucose for a few minutes. The lag is small enough that it rarely matters for day-to-day decisions, but it explains why a fingerstick and a CGM reading taken at the same moment may not perfectly match.

A more dramatic artifact is pressure-induced sensor attenuation, sometimes called a “compression low.” When you sleep on the arm or body site where the sensor sits, tissue compression reduces local blood flow, and the sensor reports a sudden, steep glucose drop that can look like serious hypoglycemia. One study that had subjects wear four sensors simultaneously found that individual sensors intermittently showed readings more than 25 mg/dL away from the median of the other sensors, and these aberrant readings were strongly correlated with lying on top of the sensor.¹² These false lows can trigger hypoglycemia alarms and, in people using automated insulin delivery systems, may even cause an insulin pump to shut off inappropriately.¹³ If you see a sharp V-shaped dip during the night that corrects itself within an hour without you eating anything, it was almost certainly a compression artifact. The telltale sign is the speed of the drop and recovery: real hypoglycemia follows a more gradual curve.

How Age Shifts the Picture

A “normal” CGM graph looks slightly different at 25 than it does at 65. In the multicenter study of healthy adults, average glucose was 98 to 99 mg/dL across most age groups, but people over 60 averaged about 104 mg/dL and spent less time in the 70-to-140 range (about 93% compared with 96% or higher in younger groups).¹ A large-cohort analysis that mapped CGM-derived measures across ages found that the maximum glucose value rose by about 0.5 mg/dL per year of age, and glucose variability (measured by MAGE) also crept upward at roughly 0.13 mg/dL per year.¹⁴ These are small annual shifts, but they accumulate. An older adult’s graph will naturally show slightly higher peaks and a bit more bounce than a younger adult’s, even if both are metabolically healthy.

The same analysis revealed subtle sex differences. The minimum glucose value rose about twice as fast per year of age in women (0.18 mg/dL per year) compared to men (0.11 mg/dL per year), meaning that the overnight floor of the graph gradually lifts more steeply for women as they age.¹⁴

Menstrual Cycle Patterns

For people who menstruate, the CGM graph can shift across the month. A study using CGM data alongside hormone tracking found a biphasic glucose pattern over the menstrual cycle: daily median glucose peaked during the luteal phase (roughly the two weeks before a period) and dropped during the late-follicular phase (the days leading up to ovulation).¹⁵ Higher estrogen levels were associated with lower glucose, while the luteal rise in progesterone appeared to push glucose upward. The shift was statistically real but modest, not enough to push a healthy person out of the normal range but enough to be visible if you compare your CGM data across weeks. If your graph consistently looks “higher” in the week before your period, hormones are the likely explanation.

Stress and Sleep Deprivation

Acute stress raises glucose. An exploratory trial in healthy young adults found that undergoing a standardized psychological stress test caused glucose to rise compared with a control condition, with glucose 60 minutes before the stressor running significantly higher than 60 minutes afterward, as the body mobilized glucose in preparation and then recovered.¹⁶ On a CGM graph, a stressful meeting or a near-miss in traffic can show up as a small but visible glucose bump even without eating anything.

Sleep deprivation has a more sustained effect. An experimental study restricting healthy young adults to limited sleep for three nights found a roughly 5% increase in the total amount of glucose circulating over 24 hours compared with baseline, with both overnight and daytime glucose rising.¹⁷ Interestingly, normal night-to-night variation in sleep duration among well-rested university students did not meaningfully affect average glucose. While longer sleep was linked to slightly lower nocturnal variability, the predicted change was less than 1 mg/dL per additional hour of sleep, likely too small to notice on a graph.¹⁸ In other words, one bad night probably will not wreck your CGM trace, but several short nights in a row will visibly raise the entire curve.

How Prediabetes Looks Different on CGM

One of the most practical reasons to understand a normal CGM graph is so you can spot when things start drifting. A matched-pair study comparing people with prediabetes to normoglycemic controls found that both groups spent the vast majority of time in range, but the differences were measurable. People with prediabetes had a median time in the 70-to-180 range of about 98.5% versus 99.9% for normal, and their mean 24-hour glucose averaged about 114 mg/dL compared with 109 mg/dL in the normal group.¹⁹ Glucose variability was also higher in the prediabetes group, with wider swings after meals.

A larger community-based study painted a similar picture but added more detail: normoglycemic participants spent about 98% of time between 70 and 180 mg/dL and about 87% in the tighter 70-to-140 range, with roughly 3 hours per day above 140 and about 15 minutes per day above 180. People with prediabetes spent about 77% in the 70-to-140 range, while those with diabetes spent about 46%.²⁰ The takeaway is that even healthy people do not stay between 70 and 140 all day. Spending some time above 140 after meals is completely normal. Spending sustained periods there, or seeing frequent readings above 180, is where the graph starts to diverge from healthy.

What CGM Looks Like in Pregnancy

Pregnancy creates its own version of normal. In uncomplicated pregnancies, average glucose was about 103 mg/dL in the first trimester, dipping to about 98 mg/dL by the second and third trimesters as the body adjusts to increased metabolic demands.²¹ Post-meal peaks averaged about 126 mg/dL with a typical excursion of about 36 mg/dL above the pre-meal baseline. The international consensus for CGM targets during pregnancy uses a tighter window of 63 to 140 mg/dL, reflecting the fact that glucose management standards are stricter when a developing fetus is involved.²²

Pregnancies that develop gestational diabetes show the same pattern seen in the prediabetes-versus-normal comparison: higher average glucose, more variability, less time in range, and more time spent above the upper threshold. These differences appear across all three trimesters, often before a formal diagnosis is made on an oral glucose tolerance test.²³ This is one of the reasons CGM use in pregnancy has attracted growing interest: the continuous data can pick up patterns that a single fasting blood draw might miss.

Why “Time in Range” Is the Number That Matters Most

If you take away one thing from all of this, it is that the most useful metric on a CGM report is time in range (TIR), not a single glucose reading or even the average. For people with type 1 or type 2 diabetes, the international consensus target range is 70 to 180 mg/dL, with a goal of spending at least 70% of the day there (or over 50% for older adults or those at higher risk of hypoglycemia).²² For healthy people, that number sits closer to 96 to 98% in the 70-to-140 range, a level that many diabetes patients understandably cannot reach.

Some researchers have proposed even tighter targets for optimal metabolic health. One large study of over 4,800 non-diabetic individuals evaluated time spent in a 70-to-100 mg/dL window and found that more time in this narrower band correlated with favorable dietary patterns and health markers.²⁴ Whether that stricter target is worth pursuing if you are already healthy is still an open question. For most people wearing a CGM out of curiosity rather than medical necessity, staying in the 70-to-140 window the great majority of the time, with post-meal peaks that resolve within a couple of hours and a smooth overnight stretch, is what a healthy graph looks like.

1
PubMed Central. Continuous Glucose Monitoring Profiles in Healthy Nondiabetic Participants: A Multicenter Prospective Study
2
PubMed Central. Continuous glucose profiles in healthy subjects under everyday life conditions and after different meals
3
PubMed. Characterizing glucose exposure for individuals with normal glucose tolerance using continuous glucose monitoring and ambulatory glucose profile analysis
4
PubMed Central. Establishment of normal reference ranges for glycemic variability in Chinese subjects using continuous glucose monitoring
5
PubMed Central. Impact of fat and protein on postprandial glycemia in gestational diabetes: A randomized pilot crossover trial
6
Letters in Biomathematics. Non-Linear Modelling of Glycemic Response Based on Macronutrient Composition in Healthy Adults
7
PubMed Central. Effects of meal sequence intervention on blood glucose response in healthy adults: a systematic review
8
PubMed Central. Effect of Exercise and Meals on Continuous Glucose Monitor Data in Healthy Individuals Without Diabetes
9
PubMed. Effects of exercise interventions on 24-h continuous glucose profiles in non-diabetic populations: A systematic review and meta-analysis
10
PubMed Central. Time lag of glucose from intravascular to interstitial compartment in humans
11
PubMed Central. Time lag of glucose from intravascular to interstitial compartment in type 1 diabetes
12
PubMed Central. Susceptibility of interstitial continuous glucose monitor performance to sleeping position
13
PubMed Central. Prospective Detection of Hypoglycemia and Near-Hypoglycemia Inducing Pressure-Induced Sensor Attenuation Onset in Continuous Glucose Monitoring Time Series
14
Cell Metabolism. CGMap: A characterization of continuous glucose monitoring data in a large healthy cohort
15
PubMed Central. Blood glucose variance measured by continuous glucose monitors across the menstrual cycle
16
PLOS Digital Health. Continuous Glucose Monitoring under standardised conditions regarding diet, exercise and stress in Healthy Young People (CGM-HYPE study): An exploratory clinical trial
17
SLEEP. Three Days of Sleep Restriction Increases 24-Hour Glucose Levels in Healthy Young Adults as Measured by Continuous Glucose Monitoring
18
Sleep. Effects of night-to-night variations in objectively measured sleep on blood glucose in healthy university students
19
PubMed Central. Difference on Glucose Profile From Continuous Glucose Monitoring in People With Prediabetes vs. Normoglycemic Individuals: A Matched-Pair Analysis
20
The Journal of Clinical Endocrinology & Metabolism. Defining Continuous Glucose Monitor Time in Range in a Large, Community-Based Cohort Without Diabetes
21
BMJ Open Diabetes Research & Care. Glucose levels measured with continuous glucose monitoring in uncomplicated pregnancies
22
Diabetes Care. Clinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range
23
Diabetes Care. Continuous Glucose Monitoring Profiles in Pregnancies With and Without Gestational Diabetes Mellitus
24
Current Developments in Nutrition. Optimised Glucose “Time in Range” Using Continuous Glucose Monitors in 4,805 Non-Diabetic Individuals Is Associated With Favourable Diet and Health: The ZOE PREDICT Studies