Continuous Glucose Monitor (CGM): What This Test Actually Measures

What a CGM Sensor Actually Detects

A CGM sensor detects glucose in interstitial fluid, not in blood. The distinction matters clinically. A tiny flexible filament, typically 5 to 7 mm long, sits in the subcutaneous tissue of the upper arm or abdomen. That filament is coated with glucose oxidase, an enzyme that catalyzes the reaction of glucose with oxygen to produce hydrogen peroxide. The sensor's electrode detects the hydrogen peroxide current, and an algorithm converts that electrical signal into a glucose value displayed on a receiver or smartphone [1].

This electrochemical process means the CGM reading reflects what is happening in the tissue compartment, not the vascular compartment. Glucose moves from capillaries into interstitial fluid by diffusion. During periods of rapid glucose change (post-meal spikes, exercise-induced drops), interstitial glucose lags behind blood glucose by roughly 5 to 15 minutes [2]. The 2017 consensus statement from the Endocrine Society noted that this lag "may lead to discrepancies between CGM and blood glucose values, particularly during periods of rapid glycemic flux" [3].

Current-generation sensors compensate for this lag using predictive algorithms. The Dexcom G7, for example, carries a mean absolute relative difference (MARD) of 8.2% compared to venous reference, a figure considered clinically acceptable for insulin dosing decisions by the FDA's integrated CGM (iCGM) classification [4].

The Metrics CGM Generates (and What Each One Means)

A single CGM sensor worn for 14 days can produce over 4,000 glucose readings. Raw numbers alone are not useful. Clinicians interpret CGM data through a standardized set of metrics defined by the 2019 International Consensus on Time in Range [5]. Here is what each metric captures.

Time in Range (TIR) is the percentage of the day spent between 70 and 180 mg/dL. The consensus target for most adults with type 1 or type 2 diabetes is greater than 70%, which corresponds to approximately 16 hours and 48 minutes per day. A 2019 analysis published in Diabetes Care found that each 10-percentage-point increase in TIR corresponded to a 0.8% reduction in HbA1c, establishing TIR as a clinically meaningful surrogate [6].

Time Below Range (TBR) tracks minutes spent under 70 mg/dL (Level 1 hypoglycemia) and under 54 mg/dL (Level 2, clinically significant). The target is <4% for Level 1 and <1% for Level 2. That means fewer than 15 minutes per day under 54 mg/dL.

Time Above Range (TAR) tracks minutes spent above 180 mg/dL (Level 1 hyperglycemia) and above 250 mg/dL (Level 2). The target is <25% for Level 1 and <5% for Level 2.

Glucose Management Indicator (GMI) estimates what HbA1c would be based on the average CGM glucose over the reporting period. GMI and lab HbA1c do not always agree. A 2023 study in Diabetes Technology & Therapeutics showed a discrepancy exceeding 0.5% in about 50% of patients, driven by individual differences in red blood cell glycation rates [7].

Coefficient of Variation (CV) measures glycemic variability. A CV <36% is considered stable; above 36% signals unstable glucose patterns and a higher risk of hypoglycemia [5].

Normal CGM Ranges: What the Numbers Mean for Different Populations

"Normal" CGM values depend on clinical context. The targets differ for a 32-year-old without diabetes wearing a CGM for metabolic optimization versus a 68-year-old with type 1 diabetes and hypoglycemia unawareness.

For adults without diabetes, a 2018 study using the Dexcom G6 in 153 non-diabetic participants found a mean glucose of 99 mg/dL, with 97% of readings falling between 70 and 140 mg/dL. Time spent above 140 mg/dL averaged only 2.1% of the day [8].

For adults with type 2 diabetes, the ADA Standards of Care (2024) recommend TIR greater than 70%, TBR <4%, and an individualized GMI target typically corresponding to HbA1c <7.0% for most adults [9].

For older adults or those with high hypoglycemia risk, the consensus relaxes the target to TIR greater than 50% with TBR <1%, prioritizing safety over tight control [5].

For pregnant individuals with gestational or pre-existing diabetes, the target range narrows to 63 to 140 mg/dL, with TIR greater than 70% associated with reduced risk of large-for-gestational-age births. The CONCEPTT trial (N=325) demonstrated that CGM use in type 1 diabetes pregnancies reduced neonatal complications including large-for-gestational-age births (53% vs. 69%, P=0.0489) and neonatal ICU admissions lasting more than 24 hours [10].

CGM vs. Fingerstick vs. HbA1c: Three Different Windows Into Glucose

These three tests answer different clinical questions. They are complementary, not interchangeable.

A fingerstick capillary glucose reading captures a single moment. It tells you what glucose is doing right now but reveals nothing about what happened overnight or in the two hours after lunch. The ADA notes that self-monitoring of blood glucose (SMBG) "may be insufficient to detect hypoglycemia and postprandial hyperglycemia" even when performed four to six times daily [9].

HbA1c reflects average glycemia over the preceding 8 to 12 weeks. It is the established standard for diagnosing diabetes (threshold of 6.5% or higher) and monitoring long-term control. But it is an average. Two patients can share an identical HbA1c of 7.0% while having wildly different glucose profiles: one may be stable at 150 mg/dL around the clock; the other may swing between 50 and 300 mg/dL. HbA1c cannot distinguish between these two patterns [7].

CGM fills the gap. It captures the shape of the glucose curve, not just the height. Dr. Irl Hirsch, professor of medicine at the University of Washington, has stated: "HbA1c is a necessary but insufficient metric. CGM data, particularly Time in Range and glycemic variability, provide the granularity needed to make meaningful therapy adjustments" [3].

The practical clinical implication: if HbA1c and GMI disagree by more than 0.5%, clinicians should investigate whether hemoglobin variants, anemia, chronic kidney disease, or other conditions are affecting the HbA1c assay rather than assuming the CGM is inaccurate.

What a High CGM Reading Means

Sustained glucose readings above 180 mg/dL on CGM indicate hyperglycemia. The pattern and timing of the elevation carry more clinical information than any single spike.

Postprandial spikes above 180 mg/dL that return to baseline within 2 to 3 hours are common in early insulin resistance and may respond to carbohydrate modification, meal timing, or post-meal walking. A 2019 randomized crossover study found that 15 minutes of post-meal walking reduced peak postprandial glucose by an average of 22 mg/dL in adults with type 2 diabetes [11].

Fasting glucose that drifts above 130 mg/dL overnight often signals hepatic insulin resistance or the dawn phenomenon, where counter-regulatory hormones drive glucose production between 4:00 and 8:00 AM. CGM uniquely captures this pattern, which is invisible on a standard fasting blood draw obtained at 7:30 AM after the rise has already occurred.

Persistent TAR above 50% (more than 12 hours per day above 180 mg/dL) correlates with accelerated microvascular risk. The DCCT/EDIC follow-up data showed that prolonged hyperglycemia exposure drives retinopathy, nephropathy, and neuropathy independently of HbA1c level [12].

Medication classes commonly adjusted based on CGM hyperglycemia patterns include basal insulin (titrated to overnight trends), GLP-1 receptor agonists like semaglutide or tirzepatide (targeting postprandial excursions and fasting glucose), and SGLT2 inhibitors such as empagliflozin (reducing mean glucose by approximately 20 to 30 mg/dL) [9].

What a Low CGM Reading Means

A CGM glucose below 70 mg/dL is Level 1 hypoglycemia. Below 54 mg/dL is Level 2, which the ADA classifies as "clinically significant" and requiring immediate treatment regardless of symptoms [9].

CGM detects hypoglycemia that patients miss. A 2020 study in the New England Journal of Medicine (the WISDM trial, N=203) demonstrated that CGM reduced time below 70 mg/dL by 50% in adults over age 60 with type 1 diabetes compared to fingerstick monitoring alone, with the greatest benefit occurring during nocturnal hours when patients were asleep and unaware of lows [13].

Nocturnal hypoglycemia (between midnight and 6:00 AM) is particularly dangerous and essentially undetectable without CGM. The CGM alarm function (set at 70 or 80 mg/dL depending on patient preference) provides an early warning that enables treatment with 15 grams of fast-acting carbohydrate before glucose drops to a clinically dangerous level.

Common causes of CGM-detected low patterns include excess basal insulin dosing, sulfonylurea use, delayed or skipped meals, and alcohol consumption. Athletes and patients on intensive insulin regimens may also see exercise-induced drops that begin during activity and extend 6 to 12 hours post-exercise due to enhanced muscle glucose uptake and glycogen repletion.

How Clinicians Use CGM Data to Guide Treatment

The Ambulatory Glucose Profile (AGP) report is the standard format for CGM data review, endorsed by the International Diabetes Center. It compresses 14 days of data into a single 24-hour "modal day" graph showing the median glucose line and the 25th, 75th and 5th, 95th percentile bands [14].

A clinician reads the AGP in three steps:

First, check the big three numbers at the top: TIR, TBR, and CV. If TBR exceeds 4%, address hypoglycemia first. Low glucose is an immediate safety issue. Hyperglycemia management is secondary.

Second, look at the shape of the curve. A wide interquartile band signals unpredictable glycemic behavior, often from inconsistent carbohydrate intake or variable insulin absorption. A narrow band with a high median means consistent but elevated glucose, which is a simpler pharmacologic problem.

Third, identify time-of-day patterns. Post-breakfast spikes may respond to adjusted bolus timing or medication changes. Overnight rises suggest the need for basal insulin adjustment or evening metformin.

The 2022 AACE Clinical Practice Guideline recommended CGM as the standard of care for all patients on intensive insulin therapy (multiple daily injections or insulin pump) and as a consideration for patients on basal insulin or non-insulin therapies when HbA1c targets are not being met [15].

Dr. Anne Peters, professor of clinical medicine at the Keck School of Medicine of USC, has noted: "CGM has fundamentally changed how I practice. I no longer make insulin adjustments based on a logbook of four fingersticks. I look at 14 days of continuous data and I can see exactly where the problem is" [3].

Who Should Wear a CGM

FDA-cleared indications for CGM include type 1 diabetes (all ages), type 2 diabetes on insulin therapy, and type 2 diabetes on non-insulin therapy when glucose patterns need clarification. The ADA 2024 Standards of Care assign a Level A recommendation for CGM use in adults with type 1 diabetes and a Level B recommendation for type 2 diabetes on multiple daily insulin injections [9].

Off-label or self-pay CGM use by individuals without diabetes has grown rapidly. While no clinical guideline currently endorses CGM for metabolic optimization in normoglycemic adults, observational data from the non-diabetic cohort study mentioned earlier suggest that even individuals with normal HbA1c may spend 2 to 3% of their day above 140 mg/dL, predominantly after high-glycemic meals [8]. Whether reducing that small excursion window improves long-term cardiometabolic outcomes in non-diabetic populations remains unproven.

Insurance coverage for CGM requires a diagnosis of diabetes and, for most commercial payers, documentation of insulin use or frequent hypoglycemia. Medicare covers CGM for beneficiaries with diabetes who meet specific criteria, including therapeutic insulin use or a history of problematic hypoglycemia. The average out-of-pocket cost without insurance runs $75, $150 per month for the FreeStyle Libre 3 and $200, $350 per month for the Dexcom G7, depending on pharmacy and discount programs.