Continuous Glucose Monitor (CGM): When to Order This Test

What Is a Continuous Glucose Monitor?

A CGM is a wearable sensor that measures interstitial glucose every 1 to 5 minutes and transmits readings to a receiver or smartphone. Unlike fingerstick blood glucose testing, which captures a single point in time, CGM generates 288 or more data points per day, revealing patterns of hyperglycemia, hypoglycemia, and glycemic variability that isolated readings miss entirely.

Real-Time CGM vs. Intermittently Scanned CGM

Two categories exist. Real-time CGM (rtCGM) devices like the Dexcom G7 push readings automatically and trigger alerts for high or low glucose. Intermittently scanned CGM (isCGM) devices like the FreeStyle Libre 3 can now also send continuous readings but originated as "flash" monitors requiring a manual scan. The ADA 2024 Standards of Care treat both as acceptable tools, though rtCGM with alerts carries stronger evidence for hypoglycemia reduction in type 1 diabetes [1].

How the Sensor Works

A thin filament sits in the subcutaneous interstitial fluid, typically on the back of the upper arm or the abdomen. An enzymatic reaction between glucose oxidase and interstitial glucose generates an electrical signal proportional to glucose concentration. There is a physiologic lag of approximately 5 to 15 minutes between blood glucose and interstitial glucose, which matters most during rapid glucose changes such as post-meal spikes or exercise-induced drops [2].

Key Metrics a CGM Report Provides

The international consensus Ambulatory Glucose Profile (AGP) report includes time in range (TIR, 70 to 180 mg/dL), time below range (TBR, <70 mg/dL and <54 mg/dL), time above range (TAR, >180 mg/dL and >250 mg/dL), glucose management indicator (GMI, an estimated HbA1c), and coefficient of variation (CV). A 2019 consensus from Battelino et al. (N=42 international experts) established that TIR >70% correlates with an HbA1c of approximately 7.0%, and each 5% increase in TIR is clinically meaningful [3].

Clinical Indications: Who Should Get a CGM Order?

The strongest evidence supports CGM for anyone using insulin, but guideline bodies have steadily expanded recommended populations. Ordering a CGM is appropriate when standard monitoring fails to explain glucose patterns, when hypoglycemia is recurrent or unrecognized, or when HbA1c alone does not capture the full glycemic picture.

Type 1 Diabetes (All Ages)

The ADA gives a Grade A recommendation for CGM use in adults with type 1 diabetes [1]. The DIAMOND trial (N=158) demonstrated that adults with type 1 diabetes using CGM spent 11% more time in range and reduced HbA1c by 0.6% compared with fingerstick-only monitoring over 24 weeks [4]. Pediatric and adolescent patients also benefit. Order CGM at diagnosis or as soon as the patient can adhere to sensor wear.

Type 2 Diabetes on Multiple Daily Injections or Insulin Pump

For patients using intensive insulin regimens, the MOBILE trial (N=175) showed that CGM use increased TIR by 3.8 hours per day (a 15% absolute improvement) and reduced HbA1c by 0.4% at 8 months compared with blood glucose monitoring alone [5]. The ADA and the Endocrine Society both endorse CGM for this group.

Type 2 Diabetes on Basal Insulin Only

Evidence is building. The FLASH-UK trial (N=156) found that isCGM use in type 2 patients on basal-only insulin improved TIR by 9% at 24 weeks [6]. The ADA 2024 update extended a Grade B recommendation to this population. Order a CGM trial when HbA1c remains above target despite basal insulin titration, or when a patient reports symptoms suggesting unrecognized hypoglycemia.

Type 2 Diabetes on Non-Insulin Therapies

The ADA notes that CGM "can be used" in non-insulin-treated type 2 diabetes for behavior modification, though insurance coverage remains inconsistent. A 2023 retrospective cohort analysis published in the Journal of Diabetes Science and Technology (N=4,072) found that non-insulin-using type 2 patients with CGM access had a 0.3% greater HbA1c reduction over 12 months than matched controls [7]. Consider ordering a 14-day professional (clinic-owned, blinded) CGM study for patients struggling with diet and medication adherence.

Gestational Diabetes and Pregnancy with Pre-existing Diabetes

The CONCEPTT trial (N=325) demonstrated that pregnant women with type 1 diabetes using rtCGM had lower rates of large-for-gestational-age infants (53% vs. 69%) and neonatal hypoglycemia (15% vs. 28%) compared with standard care [8]. For gestational diabetes, evidence is limited, but the Endocrine Society 2024 guidelines suggest considering CGM when postprandial targets are difficult to achieve with fingerstick monitoring alone.

Prediabetes and Metabolic Health Optimization

No major guideline body formally recommends CGM for prediabetes. Direct-to-consumer CGM programs have grown rapidly, yet the clinical utility in this population remains unproven in randomized trials. A professional CGM study may reveal postprandial glucose excursions exceeding 180 mg/dL that fingerstick testing would miss, but insurers rarely cover CGM for prediabetes. Order with caution and counsel patients on out-of-pocket costs.

What Do Normal CGM Readings Look Like?

"Normal" on a CGM is defined by consensus targets, not a single number. The international consensus targets from Battelino et al. Apply to most adults with diabetes, while tighter targets exist for pregnancy [3].

Standard Targets for Most Adults with Diabetes

The primary goal is TIR >70% (roughly 16 hours and 48 minutes per day between 70 and 180 mg/dL). TBR should be <4% below 70 mg/dL (<1% below 54 mg/dL). TAR should be <25% above 180 mg/dL (<5% above 250 mg/dL). CV should remain below 36%, indicating stable glucose without large swings. Dr. Tadej Battelino, lead author of the consensus, has stated: "Time in range gives patients a daily, actionable number that HbA1c never could. It changes behavior in real time" [3].

Pregnancy Targets

Tighter targets apply: TIR >70% between 63 and 140 mg/dL. TAR (<25% above 140 mg/dL) and TBR (<4% below 63 mg/dL) remain the same percentage thresholds but within narrower glucose bounds [8].

Older Adults and High-Risk Populations

For older adults with significant comorbidities or hypoglycemia risk, the ADA suggests a relaxed target: TIR >50% (70 to 180 mg/dL) with TBR <1% below 70 mg/dL. Preventing hypoglycemia takes priority over reducing hyperglycemia in this group [1].

What Does a High CGM Reading Mean?

Persistent hyperglycemia on CGM (TAR consistently >25%) signals inadequate glucose control. The pattern matters as much as the numbers.

Fasting Hyperglycemia

Elevated overnight and fasting glucose (above 130 mg/dL at wake) often reflects insufficient basal insulin dosing, the dawn phenomenon (a cortisol-driven glucose rise between 4 AM and 8 AM), or late-night eating. CGM data clearly differentiates the dawn phenomenon from Somogyi rebound (nocturnal hypoglycemia followed by counter-regulatory hyperglycemia), which requires opposite treatment strategies.

Postprandial Spikes

Glucose peaks exceeding 180 mg/dL within 1 to 2 hours of eating point toward inadequate mealtime insulin coverage, high-glycemic carbohydrate loads, or delayed gastric emptying. In type 2 diabetes, adding a GLP-1 receptor agonist (e.g., semaglutide) or an SGLT2 inhibitor (e.g., empagliflozin) to the regimen can blunt postprandial excursions by 30 to 50 mg/dL [9].

Sustained Elevation

A GMI above 8.0% or TAR consistently >50% warrants therapy intensification. Reassess medication adherence, carbohydrate intake, and physical activity before adding new agents.

What Does a Low CGM Reading Mean?

CGM is the single most effective tool for detecting hypoglycemia, including nocturnal episodes that patients sleep through.

Level 1 Hypoglycemia (54 to 69 mg/dL)

This range requires attention but not emergency action. Patients should consume 15 g of fast-acting carbohydrate and recheck in 15 minutes. Recurrent level 1 events (more than two per week) warrant dose reduction of the offending agent, typically a sulfonylurea or insulin.

Level 2 Hypoglycemia (<54 mg/dL)

This is clinically significant. The ADA defines <54 mg/dL as the threshold for "serious, clinically important hypoglycemia" [1]. The I-HART CGM trial (N=153) found that rtCGM use in older adults with type 1 diabetes reduced time spent below 54 mg/dL by 60% over 6 months compared with standard monitoring [10].

Hypoglycemia Unawareness

Patients who no longer sense low blood sugar represent the highest-risk group. CGM with predictive low-glucose alerts (available on Dexcom G7 and Medtronic Guardian 4) can warn patients 20 minutes before projected hypoglycemia. Dr. Irl Hirsch at the University of Washington has noted: "For patients with hypoglycemia unawareness, CGM is not optional. It is the standard of care" [11].

How to Order a CGM: Practical Steps

Ordering a CGM involves a prescription, a diagnosis code, and (for insurance coverage) documentation that specific criteria are met.

Prescription Requirements

Write the prescription for the specific device (e.g., "Dexcom G7 real-time CGM system, 3 sensors per 30-day supply, transmitter as needed"). Include the ICD-10 code: E11.65 (type 2 diabetes with hyperglycemia), E10.65 (type 1), or O24.4 (gestational diabetes). Specify "medical necessity for glucose pattern analysis" in the clinical notes.

Insurance and Medicare Coverage Criteria

Medicare covers CGM under the Durable Medical Equipment (DME) benefit for patients who (1) have diabetes requiring insulin, (2) perform 4+ fingerstick tests per day or use an insulin pump, and (3) have an in-person or telehealth visit with the prescribing provider within 6 months of the order. As of 2024, the fingerstick requirement was relaxed for rtCGM, but documentation of intensive glucose management decisions (3+ per day) is still required [12]. Most commercial insurers follow similar criteria. Prior authorization is common for isCGM and almost universal for rtCGM.

Professional (Clinic-Owned) CGM vs. Personal CGM

A professional CGM is a clinic-owned sensor placed during an office visit and returned 10 to 14 days later for data download. It requires no patient prescription or insurance authorization for the device itself (the provider bills the service code 95251 for interpretation). This approach works well as a diagnostic trial before committing a patient to long-term personal CGM, or when insurance denies a personal device.

Interpreting CGM Data: The Ambulatory Glucose Profile

The AGP is a standardized one-page report endorsed by the ADA for clinical decision-making. Reading it correctly takes practice, but the structure is consistent across all major CGM platforms.

The Five Numbers to Check First

Start with TIR, then check TBR (especially <54 mg/dL), TAR, GMI, and CV. If TBR is elevated, address hypoglycemia before targeting hyperglycemia. This sequence prevents the common error of increasing insulin doses while the patient is already having lows at other times of day.

Pattern Recognition

The AGP displays a "spaghetti plot" (all daily glucose tracings overlaid) and a smoothed median/interquartile range graph. Look for recurring spikes at consistent times (often post-lunch or post-dinner), nocturnal dips, and high variability (wide interquartile bands). The AACE 2023 CGM Consensus Statement recommends reviewing AGP data at every diabetes-related visit, ideally every 2 to 4 weeks during active titration [13].

Adjusting Therapy Based on CGM

If fasting glucose runs high (AGP median above 130 at 6 AM), increase basal insulin by 10 to 20% every 3 days. If postprandial spikes are the dominant pattern, consider adding rapid-acting insulin at that meal or introducing a GLP-1 receptor agonist. If CV exceeds 36%, look for erratic carbohydrate intake, inconsistent medication timing, or alcohol use.

CGM Accuracy and Limitations

No sensor is a perfect surrogate for venous blood glucose. Understanding the sources of error helps set realistic expectations.

MARD: The Accuracy Benchmark

Mean Absolute Relative Difference (MARD) is the standard accuracy metric. The Dexcom G7 reports a MARD of 8.2% in adults, while the FreeStyle Libre 3 reports 7.9% [14]. A MARD below 10% is considered accurate enough for insulin dosing decisions. Older-generation sensors had MARDs of 12 to 15%, which sometimes led to clinically meaningful dosing errors.

Situations That Reduce Accuracy

Acetaminophen (Tylenol) interfered with older Dexcom sensors but does not affect the G6 or G7. Sensor compression (lying on the sensor during sleep) can produce false low readings. Severe dehydration, rapid temperature changes, and the first 24 hours of a new sensor insertion (the "warm-up" period) may also reduce accuracy. Teach patients to confirm CGM readings with a fingerstick before treating any unexpected value below 55 mg/dL or above 400 mg/dL.

CGM Is Not a Replacement for HbA1c

HbA1c reflects a 2-to-3-month average and remains the validated surrogate for long-term complications in landmark trials (DCCT, UKPDS). CGM complements HbA1c by adding pattern data and real-time feedback, but clinicians should continue to order HbA1c every 3 to 6 months alongside CGM [1].

Emerging Uses and Future Directions

CGM technology is expanding beyond traditional diabetes management into perioperative care, inpatient settings, and preventive medicine.

The FDA cleared the Dexcom G7 and FreeStyle Libre 2 for inpatient use in 2024 under specific protocols, marking a shift from ICU-only point-of-care glucose testing toward continuous monitoring on general medical wards [15]. Trials are ongoing for CGM-guided insulin dosing in hospitalized patients with type 2 diabetes (ClinicalTrials.gov NCT05620888).

In the outpatient space, integrated CGM-insulin pump systems (automated insulin delivery, or AID) now represent the standard of care for many type 1 patients. The ADA 2024 Standards recommend offering AID to all type 1 adults and adolescents who are candidates, with TIR improvements of 10 to 15% demonstrated across multiple randomized trials [1].

For patients prescribed a CGM for the first time, schedule a follow-up visit at 14 to 21 days to review the initial AGP report, troubleshoot sensor adhesion or alarm fatigue, and make the first round of therapy adjustments.