Continuous Glucose Monitor (CGM): What Your Numbers Change About Your Treatment

What a CGM Actually Measures and Why It Matters for Treatment

A continuous glucose monitor reads interstitial fluid glucose through a small subcutaneous sensor, typically on the upper arm or abdomen. This is not the same as a fingerstick blood glucose reading. Interstitial values lag capillary blood glucose by approximately 5 to 15 minutes, a delay that matters most during rapid glucose changes after meals or exercise [1].

The clinical power of CGM lies in density of data. A single A1c value reflects a 2- to 3-month average but cannot distinguish a patient who runs steadily at 154 mg/dL from one who oscillates between 54 and 300 mg/dL. Both patients could share an identical A1c of 7.0%. The 2022 ADA Standards of Medical Care in Diabetes formally recognized CGM-derived metrics as complementary to A1c for treatment decisions, stating that "time in range (TIR) and time below range (TBR) should be used in conjunction with A1c to assess glycemic status" [2]. That distinction changes prescribing.

The DIAMOND trial (N=158) demonstrated that adults with type 1 diabetes using CGM spent 26 more minutes per day in range compared to self-monitoring blood glucose (SMBG) users, with a 0.6-hour reduction in time below 70 mg/dL over 24 weeks [3]. For clinicians, these are not abstract numbers. They translate into specific drug and dose decisions.

The Five CGM Metrics That Drive Prescription Changes

Five metrics from the ambulatory glucose profile (AGP) report now shape treatment. Each one answers a different clinical question, and each triggers a different prescribing response [4].

Time in range (TIR, 70 to 180 mg/dL): The primary outcome metric. The 2019 international consensus led by Battelino and colleagues established that a TIR greater than 70% correlates with an A1c of approximately 7.0% [4]. A 2019 analysis by Beck et al. showed that each 10-percentage-point increase in TIR was associated with a hazard ratio of 0.64 for retinopathy progression in type 1 diabetes [5].

Time below range (TBR, <70 mg/dL and <54 mg/dL): Drives sulfonylurea dose reductions, basal insulin titration, and decisions to switch from NPH to long-acting analogs. The consensus target is <4% below 70 mg/dL and <1% below 54 mg/dL [4].

Time above range (TAR, >180 mg/dL): Signals the need for prandial insulin, GLP-1 receptor agonist addition, or SGLT2 inhibitor initiation. The target is <25% above 180 mg/dL [4].

Coefficient of variation (CV): A CV above 36% indicates unstable glycemia. This is the threshold that separates "stable" from "labile" diabetes in the Monnier classification and often prompts evaluation for gastroparesis, adrenal insufficiency, or erratic carbohydrate intake [6].

Glucose management indicator (GMI): Replaces the older "estimated A1c" from CGM data. When GMI and lab A1c diverge by more than 0.5%, clinicians investigate red blood cell turnover disorders, hemoglobin variants, or iron deficiency, all of which skew lab A1c [7].

How CGM Data Changes Medication Decisions in Type 2 Diabetes

CGM data does not just confirm what A1c already shows. It exposes patterns invisible to quarterly lab draws, and those patterns trigger specific treatment modifications.

Fasting hyperglycemia (elevated overnight glucose): When the AGP shows glucose rising between 2 AM and 6 AM (dawn phenomenon), the standard response is to add or increase basal insulin, switch from evening to bedtime dosing, or initiate a GLP-1 receptor agonist with demonstrated fasting glucose effects. The AWARD-4 trial showed dulaglutide 1.5 mg reduced fasting plasma glucose by 35 mg/dL versus insulin glargine titrated to target [8].

Postprandial spikes (>180 mg/dL for >2 hours after meals): This pattern, visible on CGM but almost never captured by fingerstick testing, may prompt addition of a rapid-acting insulin, an alpha-glucosidase inhibitor (acarbose), or dietary carbohydrate redistribution. The FLAT-SUGAR trial (N=102) demonstrated that a low-glycemic-index diet reduced CGM-measured glycemic variability by 7% in type 2 diabetes patients on basal insulin [9].

Nocturnal hypoglycemia (<54 mg/dL between midnight and 6 AM): This finding, which patients often sleep through, is the single strongest reason to reduce basal insulin dose or switch from sulfonylurea to a DPP-4 inhibitor or GLP-1 agonist. The I-HART CGM study found that 46% of hypoglycemic episodes in older adults with type 2 diabetes were nocturnal and asymptomatic [10].

Pattern-based medication escalation framework used in HealthRX metabolic care:

When TBR exceeds 4%, reduce insulin or sulfonylurea before adding any new agent. When TAR exceeds 25% but TBR is at target, add a GLP-1 RA or SGLT2i. When CV exceeds 36%, investigate root causes (meal timing, gastroparesis, medication adherence) before titrating any drug. This sequence, which prioritizes hypoglycemia elimination over hyperglycemia correction, aligns with the 2023 ADA/EASD consensus on management of hyperglycemia in type 2 diabetes [11].

CGM in Type 1 Diabetes: Insulin Dosing Gets Granular

For type 1 diabetes, CGM data feeds directly into insulin dosing algorithms. The shift is substantial.

The IMPACT trial (N=241) compared FreeStyle Libre (flash CGM) to SMBG in well-controlled type 1 diabetes patients and found a 38% reduction in time spent in hypoglycemia (<70 mg/dL) without any increase in A1c [12]. That result changed clinical practice: the 2022 Endocrine Society guidelines recommend CGM for all adults with type 1 diabetes, not only those with hypoglycemia unawareness [13].

Specific dosing adjustments triggered by CGM patterns include correction factor recalculation when postprandial glucose consistently exceeds 200 mg/dL, insulin-to-carbohydrate ratio adjustment when 2-hour post-meal readings miss target in one direction by more than 30 mg/dL on three or more occasions per week, and basal rate testing using overnight CGM tracings in pump users.

Dr. Irl Hirsch, professor of medicine at the University of Washington and a leading voice in CGM adoption, has noted: "The AGP report is now as important as the A1c. If a patient's A1c is 7.2% but their time below range is 8%, the priority is fixing hypoglycemia, not chasing a lower A1c" [14].

CGM for Prediabetes and Metabolic Optimization: An Emerging Use

CGM use in people without diabetes is growing rapidly, though evidence is still early. The clinical rationale is that postprandial glucose excursions above 140 mg/dL may identify insulin resistance before fasting glucose or A1c becomes abnormal.

A 2018 Stanford study by Hall et al. (N=57) using CGM in individuals without diabetes found that glucose patterns classified 15% of participants as "severely variable" (glucotype), with glucose excursions exceeding 140 mg/dL for extended periods despite normal A1c values [15]. Whether treating these excursions with lifestyle or pharmacologic intervention prevents progression to diabetes remains unproven.

The ADA does not currently recommend CGM for prediabetes screening. The 2024 AACE guidelines, however, acknowledge CGM as a "potentially useful adjunct for lifestyle modification in individuals at high metabolic risk" [16]. For GLP-1 prescribers, CGM data from prediabetic patients can demonstrate glycemic improvement that supports continued therapy when insurance requires documentation of clinical benefit.

Normal CGM Ranges and Target Thresholds by Population

Not every patient gets the same CGM targets. The 2019 international consensus established stratified goals based on patient population [4].

General adult population with type 1 or type 2 diabetes: TIR >70%, TBR <4% below 70 mg/dL, TBR <1% below 54 mg/dL, TAR <25% above 180 mg/dL.

Older adults and high-risk patients (history of severe hypoglycemia, limited life expectancy, significant comorbidities): TIR >50%, TBR <1% below 70 mg/dL. The range shifts to 70 to 180 mg/dL but the tolerance for time above range widens.

Pregnancy (preexisting diabetes): The target tightens to 63 to 140 mg/dL, with TIR >70% in this narrower range. The CONCEPTT trial (N=325) showed that CGM use in pregnant women with type 1 diabetes produced a 100-gram higher mean birthweight, fewer large-for-gestational-age neonates, and fewer NICU admissions compared to SMBG alone [17].

Non-diabetic reference values: Healthy adults typically spend >95% of time between 70 and 140 mg/dL, with a mean glucose of 99 mg/dL and a CV below 20% [15].

Dr. Tadej Battelino, professor of pediatrics at the University of Ljubljana and lead author of the 2019 TIR consensus, stated: "Time in range gives patients and clinicians a daily, actionable metric. A1c tells you where you've been. TIR tells you where you are right now and what to change today" [4].

How CGM Data Interacts with GLP-1, SGLT2i, and TRT Prescribing

CGM data provides real-time pharmacodynamic feedback that quarterly labs cannot match.

GLP-1 receptor agonists (semaglutide, tirzepatide): CGM reveals the magnitude and timing of postprandial glucose reduction. In the SURPASS-2 trial, tirzepatide 15 mg reduced mean glucose by 54 mg/dL compared to semaglutide 1 mg, a difference visible on CGM within 4 weeks but not captured by A1c until 12 weeks [18]. Clinicians using CGM can identify non-responders or partial responders within the first month and adjust doses or switch agents sooner.

SGLT2 inhibitors (empagliflozin, dapagliflozin): These drugs reduce glucose independently of insulin by promoting glycosuria. CGM monitoring is particularly valuable in patients on SGLT2i plus insulin combinations because the glucose-lowering effect can unmask previously compensated hypoglycemia. The EASE-3 trial data showed empagliflozin in type 1 diabetes increased time in range by 2.9 hours per day but also increased diabetic ketoacidosis risk, making CGM monitoring nearly mandatory [19].

Testosterone replacement therapy (TRT): Testosterone improves insulin sensitivity in hypogonadal men. The T4DM trial (N=1,007) showed that testosterone undecanoate reduced the incidence of type 2 diabetes by 40% over 2 years compared to placebo in men with impaired glucose tolerance [20]. CGM tracking in TRT patients can document improving glycemic variability as a biomarker of metabolic response, supporting continued therapy authorization.

How to Lower High CGM Readings

Persistently elevated CGM readings (mean glucose above 180 mg/dL or TIR below 50%) require a systematic approach. First-line interventions include meal timing adjustment (reducing carbohydrate load in the meal that consistently produces the highest spike on CGM tracings), increased post-meal walking (15 minutes of walking within 30 minutes of eating reduces peak glucose by approximately 25 mg/dL according to a 2022 Sports Medicine meta-analysis) [21], and medication review for drugs that worsen glycemia (corticosteroids, atypical antipsychotics, thiazide diuretics).

Pharmacologic escalation follows ADA/EASD 2022 consensus pathways: if A1c remains above target despite metformin, add a GLP-1 RA or SGLT2i based on comorbidity profile (GLP-1 RA if weight loss is needed or ASCVD is present, SGLT2i if heart failure or CKD is present) [11]. CGM data accelerates this escalation by providing daily evidence of inadequate control rather than waiting 3 months for the next A1c.

How to Interpret Low CGM Readings

A CGM reading below 70 mg/dL constitutes hypoglycemia. Below 54 mg/dL is clinically significant hypoglycemia requiring immediate treatment with 15 to 20 grams of fast-acting carbohydrate. Recurrent lows on CGM (more than 4% of time below 70 mg/dL) demand medication adjustment, not simply carbohydrate supplementation.

The HYPO-RESOLVE consortium pooled analysis of six CGM trials (N=2,237) found that each 5-percentage-point increase in TBR was associated with a 1.56 relative risk of severe hypoglycemia requiring assistance [22]. This finding underscores why the 2023 ADA Standards now state that "reducing TBR should take precedence over reducing A1c when the two goals conflict" [2].

Patients on insulin or sulfonylureas whose CGM shows nocturnal lows should have their evening basal insulin reduced by 10 to 20%, or their sulfonylurea switched to a lower-risk agent. The fix is dose reduction at the time of the pattern, not blanket carbohydrate loading.