Continuous Glucose Monitoring for Non-Diabetics: What It Shows and What It Misses

Continuous glucose monitors, or CGMs, were developed for people with type 1 and type 2 diabetes who need to track their glucose minute by minute to dose insulin safely. In the last five years, startups have rebranded the same hardware for healthy non-diabetics, promising to reveal your personal glucose response to foods, optimize your energy, and prevent metabolic disease decades before it shows up on a fasting glucose test. The pitch is seductive, and the devices genuinely produce fascinating data. But the interpretation is far more ambiguous than the marketing suggests, because non-diabetic glucose dynamics are tightly regulated, because the devices have meaningful accuracy limits, and because the link between post-meal glucose spikes in healthy people and long-term health outcomes is not well established. The most cited serious research is Kevin Hall's group at the National Institute of Diabetes and Digestive and Kidney Diseases, which has used CGMs in controlled feeding trials in non-diabetics, and the American Diabetes Association has issued position statements that explicitly limit their recommendation of CGMs to people with diabetes and prediabetes. This guide covers what a CGM actually measures, what the data shows in non-diabetics, what it does not show, and whether the hype is justified.

How a CGM works

A continuous glucose monitor is a small sensor, about the size of a coin, applied to the back of the upper arm or the abdomen with a self-adhesive patch. A tiny flexible filament penetrates the skin to sit in the interstitial fluid (the fluid between cells), where it measures glucose electrochemically every 1 to 5 minutes. The sensor transmits readings via Bluetooth to a smartphone app. Most sensors last 10 to 14 days before they need to be replaced. The two dominant consumer CGM systems are the Abbott FreeStyle Libre (which requires a manual scan to retrieve readings in its basic version) and the Dexcom G7 (which streams continuously and can integrate with insulin pumps).

The key technical fact is that a CGM measures interstitial glucose, not blood glucose. Interstitial glucose lags blood glucose by about 5 to 15 minutes, depending on the direction of change. When blood glucose is rising rapidly after a meal, interstitial glucose will read lower than the actual blood value. When blood glucose is falling rapidly during exercise, interstitial glucose will read higher. This lag matters most during rapid changes, which is exactly when people are most alarmed by their CGM data. A spike that reads 180 mg per dL on the CGM may correspond to a blood glucose peak of 200 mg per dL 10 minutes earlier, or to a peak of 165 mg per dL 10 minutes later. The accuracy is generally within 15 percent of blood glucose when glucose is stable, and within 20 to 30 percent during rapid changes.

What normal glucose dynamics look like in non-diabetics

In a healthy non-diabetic adult, fasting blood glucose is 70 to 99 mg per dL. After a meal, blood glucose rises to a peak that depends on the meal's carbohydrate content, glycemic index, fiber, fat, and protein, plus the person's insulin sensitivity, fitness, and recent activity. The American Diabetes Association defines a normal post-meal peak as under 140 mg per dL at 2 hours, but many healthy adults briefly exceed this in the first 30 to 60 minutes after a high-glycemic meal. CGM studies in non-diabetics have shown that the average peak is around 120 to 140 mg per dL, the average time above 140 mg per dL is 30 to 60 minutes per day, and the average time above 180 mg per dL is essentially zero in metabolically healthy adults.

A 2018 study by Hall and colleagues at the NIDDK, published in Cell Metabolism and related publications, used CGMs in non-diabetic adults during controlled feeding trials and showed substantial individual variability in post-meal glucose response. The same standardized meal produced glucose peaks ranging from 100 to 180 mg per dL across participants, with some individuals spiking much higher than expected for a "healthy" meal and others spiking much lower than expected for a "high-glycemic" meal. This variability is consistent with the Personalized Nutrition Project (Zeevi and colleagues, Cell, 2015), which showed that individual glycemic responses to identical foods differ widely and are poorly predicted by the food's glycemic index alone.

The Hall 2018 trial and what it revealed

Kevin Hall's group at the NIDDK has been among the most rigorous users of CGMs in non-diabetics. Their 2018 work, building on earlier ultra-processed food trials, used CGMs to track glucose dynamics in adults fed controlled diets of ultra-processed or minimally processed foods, matched for calories, macronutrients, sugar, sodium, fiber, and energy density. The trial found that participants ate about 500 more calories per day on the ultra-processed diet and gained weight rapidly, but the glucose dynamics were less dramatic than expected: post-meal glucose peaks were modestly higher on the ultra-processed diet, but the differences were small compared to the differences in calorie intake and weight gain.

The Hall 2018 work is important for two reasons. First, it shows that CGMs can detect real differences in glucose dynamics between dietary patterns in controlled settings. Second, it shows that the differences are often smaller and more variable than the popular discourse suggests, and that the dominant effect of dietary pattern on body weight and metabolic health operates through mechanisms other than acute glucose response (specifically, through calorie intake and the food matrix's effect on satiety). The CGM data was fascinating but not the main story.

The American Diabetes Association position

The American Diabetes Association's Standards of Care explicitly addresses the use of CGMs in non-diabetics. The 2024 Standards of Care recommends CGMs for adults with type 1 diabetes, adults with type 2 diabetes on intensive insulin therapy, and selected adults with type 2 diabetes on less intensive therapy. The Standards do not recommend CGMs for non-diabetic adults, citing the lack of evidence that post-meal glucose spikes in this population predict long-term health outcomes, and noting that the devices can cause unnecessary anxiety and lead to unnecessary dietary restriction. The ADA acknowledges that CGMs may have research and educational value in non-diabetics, but does not endorse routine clinical use.

The ADA position is consistent with the broader clinical literature. A 2022 review in Diabetes Care noted that the relationship between post-meal glucose spikes and cardiovascular outcomes is well established in people with diabetes and prediabetes, but is much weaker in metabolically healthy adults, in part because their glucose returns to baseline within 1 to 2 hours and the cumulative exposure is low. The review concluded that "the clinical significance of transient postprandial hyperglycemia in otherwise healthy adults remains uncertain."

What a CGM does well in non-diabetics

Despite the ADA's caution, a CGM can be genuinely educational for a non-diabetic adult for a few weeks. The most valuable uses are:

• Identifying personal high-glucose-response foods. Some people spike dramatically on oatmeal, others on bananas, others on white rice. The Zeevi 2015 study showed that these individual responses are poorly predicted by glycemic index tables. A CGM can identify which foods spike you personally, allowing targeted substitutions.
• Seeing the effect of food order and walking. Eating protein and vegetables before carbohydrates reduces post-meal glucose spikes by 30 to 50 percent in some studies. A 10-minute walk after a meal similarly reduces the spike. A CGM makes these effects visible in real time.
• Detecting reactive hypoglycemia. Some non-diabetics experience a post-meal glucose crash (glucose below 70 mg per dL) 2 to 3 hours after a high-glycemic meal, with symptoms of shakiness, hunger, and fatigue. A CGM can confirm the pattern.
• Seeing the effect of sleep and stress. Poor sleep and high stress raise fasting glucose and post-meal spikes. A CGM can show the effect of a bad night of sleep on the next day's glucose.
• Confirming prediabetes or early dysglycemia. Some adults with normal fasting glucose have post-meal spikes that exceed the prediabetic threshold (140 to 199 mg per dL at 2 hours). A CGM can detect this pattern, which would warrant a formal oral glucose tolerance test.

For these uses, a 14-day or 28-day CGM trial is enough to extract the educational value. There is little reason to wear one continuously for months.

What a CGM does poorly in non-diabetics

The limitations are substantial:

• Glucose is one signal among many. A flat glucose line does not mean a meal was healthy. A meal of olive oil and vodka produces a flat glucose line. A meal of lentils and vegetables produces a modest spike but is profoundly healthy. Optimizing for glucose flatness can lead people to avoid nutritious high-carbohydrate foods (whole grains, fruit, legumes) and toward low-carb processed foods that happen to produce flat lines.
• The accuracy is limited during rapid changes. The 5 to 15 minute interstitial-to-blood lag means that peak readings can be off by 20 to 30 percent during the post-meal spike. The readings are most accurate when glucose is stable, which is exactly when the data is least interesting.
• "Spikes" are normal and not necessarily harmful. Healthy non-diabetics routinely spike to 140 to 160 mg per dL after high-glycemic meals. This is a normal physiological response, not a sign of metabolic disease. The marketing around "avoiding all spikes" pathologizes normal physiology.
• Long-term outcomes are not established. No randomized trial has shown that reducing post-meal glucose spikes in metabolically healthy adults improves long-term health outcomes. The link is plausible (cumulative oxidative stress, advanced glycation end products) but unproven.
• The cost is non-trivial. CGM sensors cost $50 to $100 each, last 10 to 14 days, and are typically not covered by insurance for non-diabetics. A year of continuous wear runs $1,500 to $3,500, which is a lot to pay for educational data.

A 2023 article in JAMA Internal Medicine on the rise of direct-to-consumer CGM startups noted that the marketing often exceeds the evidence, and that some users develop unhealthy obsessions with their glucose line, restricting their diet in ways that are not supported by clinical evidence.

Practical use: what to do if you try one

If you decide to try a CGM as a non-diabetic, the most useful approach is:

• Use it for 14 to 28 days, not indefinitely. The educational value plateaus after the first month. After that, you are mostly confirming what you already learned.
• Test specific foods, not just "monitor." Eat your typical breakfast on day 1, then test variants (same meal with a 10-minute walk, same meal with protein first, same meal with a different grain) on subsequent days. The controlled comparison is what produces insight.
• Use fingerstick confirmation for surprising readings. If the CGM shows a 220 mg per dL spike, verify with a fingerstick before drawing conclusions. The CGM may be off by 20 to 30 percent.
• Focus on the 2-hour post-meal value, not the peak. The 2-hour post-meal glucose is the clinically validated measure (under 140 mg per dL is normal, 140 to 199 mg per dL is prediabetic). The peak can be elevated in perfectly healthy people.
• Do not optimize for flatness. A modest post-meal spike in response to a nutritious high-carbohydrate food is not a problem. The bigger questions are: does glucose return to baseline within 2 to 3 hours, and is fasting glucose in the normal range (under 100 mg per dL)?
• Pair the CGM with the food diary. The glucose data is most useful when paired with detailed food, activity, and sleep logs. Use the food diary to log what you ate and did, and look for patterns.

Track your fasting glucose and HbA1c annually through your doctor. Use the A1C Calculator and Diabetes Risk Calculator to assess your metabolic health in a clinically validated way. These are the measures that have decades of outcome data behind them.

Who should not use a CGM as a non-diabetic

Some non-diabetics should be cautious about CGMs:

• People with a history of eating disorders may find that constant monitoring of a metabolic signal triggers restrictive behaviors or obsessive food tracking.
• People with health anxiety may fixate on normal glucose variability and interpret it as disease, leading to unnecessary dietary restriction and stress.
• People with limited budgets would get more health benefit from spending the same money on better food, gym membership, or a doctor visit.
• People who already have normal fasting glucose and HbA1c and a healthy dietary pattern are unlikely to learn anything actionable from a CGM.

The marketing of CGMs to non-diabetics often implies that everyone should wear one. The evidence does not support that.

Common pitfalls

Pitfall 1: Treating any spike as a problem

Healthy non-diabetics routinely spike to 140 to 160 mg per dL after high-glycemic meals. This is normal physiology, not metabolic disease. The clinically meaningful thresholds are 140 mg per dL at 2 hours (prediabetic threshold) and 100 mg per dL fasting. Optimizing for zero spikes is both unrealistic and unnecessary.

Pitfall 2: Trusting the peak reading without fingerstick confirmation

CGMs have 20 to 30 percent error during rapid changes, which is when the peak reading occurs. A reading of 200 mg per dL on a CGM may correspond to a true blood glucose of 150 or 250. Always confirm surprising readings with a fingerstick.

Pitfall 3: Optimizing for flatness at the expense of nutrition

A meal of olive oil and vodka produces a flat glucose line. A meal of lentils, brown rice, and vegetables produces a modest spike. The flat line is not healthier. Restricting nutritious high-carbohydrate foods (whole grains, fruit, legumes) to keep glucose flat is misguided.

Pitfall 4: Wearing it for too long

The educational value plateaus after 14 to 28 days. Continuous wear for months produces minimal additional insight and significant cost. Get the educational value, take the lessons, and stop.

Pitfall 5: Believing the marketing over the evidence

Direct-to-consumer CGM marketing implies that the devices will optimize your energy, prevent metabolic disease, and extend your life. The evidence supports none of these claims in metabolically healthy adults. The devices are useful educational tools for a few weeks; they are not life-transforming medical devices for non-diabetics.

The bottom line

Continuous glucose monitors are valuable medical devices for people with diabetes, useful educational tools for some non-diabetics for 14 to 28 days, and oversold by direct-to-consumer startups for the broader non-diabetic market. The Hall 2018 work at the NIDDK showed that CGMs can detect real differences in glucose dynamics between dietary patterns, but that these differences are often smaller and more variable than the marketing suggests, and that the dominant effect of dietary pattern on health operates through mechanisms other than acute glucose response. The American Diabetes Association does not recommend CGMs for non-diabetic adults, citing the lack of evidence that post-meal glucose spikes in healthy people predict long-term outcomes. If you try a CGM, use it for 14 to 28 days to identify personal high-response foods, test the effects of food order and walking, and detect any reactive hypoglycemia or post-meal spikes above the prediabetic threshold. Do not optimize for flatness, confirm surprising readings with a fingerstick, and rely on fasting glucose and HbA1c, tracked annually, as your primary measures of metabolic health.