Continuous Glucose Monitoring Beyond Insulin: Evidence, OTC Systems, and Data Interpretation for Family Medicine

Understanding Why CGM Use Beyond Insulin Presents a Different Clinical Challenge

When managing intensive insulin therapy, continuous glucose monitoring often serves an immediate safety and dosing function. In noninsulin-treated diabetes, however, its clinical value shifts. The primary goals become identifying the dominant source of glycemic exposure, revealing repeated medication-related lows, making the effects of meals and activity visible, and supporting specific treatment decisions.

This distinction matters immensely because a dense stream of glucose values is not inherently beneficial on its own. The raw information must be actively translated into a medication, nutrition, activity, or monitoring plan to provide real value. [1,3]

The 2026 American Diabetes Association Standards of Care have broadened this clinical frame. Recommendation 7.15 supports monitoring at diagnosis or later for people using insulin, people taking noninsulin therapies that can cause hypoglycemia, and people on any diabetes regimen when the data helps guide management. Recommendation 7.18 supports periodic personal or professional monitoring when continuous use is not feasible. Both recommendations require careful individualization rather than automatic, lifelong sensor use. [1]

Evaluating the Current Evidence for Noninsulin-Treated Type 2 Diabetes

The most focused current synthesis included eight randomized controlled trials and 541 participants who were not using insulin. Compared with standard blood glucose monitoring or usual care, continuous monitoring lowered A1C by a weighted mean of 0.37 percentage points, increased time in range by 8.84 percentage points, and reduced time above range by 8.14 percentage points. Statistical heterogeneity for these pooled glycemic outcomes was low, but the studies were generally short and primarily enrolled people with suboptimal baseline control. [2]

A 2026 critical review reached a similar overall conclusion. Across adults with type 2 diabetes, randomized evidence supports an average A1C improvement of about 0.3 percentage points, with some individuals benefiting substantially more. The review appropriately cautioned against default use when there is no defined clinical problem, particularly in people not taking glucose-lowering medication and in those with prediabetes or obesity. [3]

The CONNECT randomized trial is clinically important but still preliminary. At the 2026 American Diabetes Association Scientific Sessions, investigators reported results from 283 adults with type 2 diabetes who were not using insulin and were treated in 22 U.S. primary care practices. The mean baseline A1C was 8.8 percent. At 26 weeks, the intervention group had a 0.9 percentage point greater A1C reduction and about five more hours per day in the 70 to 180 mg/dL range than the routine care group. Because the findings were presented as an oral abstract and the 12-month extension was still underway, the result should not yet be treated as a final, peer-reviewed durability or outcomes trial. [4]

Summary of Current Evidence

Evidence source	Main finding	What it does not prove
Updated meta-analysis, 8 RCTs, 541 noninsulin-treated adults [2]	A1C decreased by 0.37 percent; time in range increased by 8.84 percentage points; time above range decreased by 8.14 percentage points.	Long-term durability or fewer diabetes complications.
CONNECT RCT, 283 adults, 26 weeks, oral abstract [4]	A1C benefit was 0.9 percent greater; about 5 more hours per day in the 70 to 180 mg/dL range.	Final peer-reviewed effect estimate or 12-month durability.
Matched primary care cohort, 182 adults [5]	Associated with a 0.62 percent greater A1C reduction at 3 months; time in range improved from 39.7 percent to 61.9 percent.	Causality; the cohort included both noninsulin and basal-insulin therapy.
MOBILE RCT, 175 adults on basal insulin [6]	Adjusted A1C difference of -0.4 percent at 8 months versus meter monitoring.	Direct evidence for patients using absolutely no insulin.

Acknowledging the Evidence Gap Regarding Hard Clinical Outcomes

The strongest evidence currently supports intermediate glycemic outcomes, not hard clinical endpoints. The noninsulin trials have not established that continuous monitoring reduces retinopathy, kidney failure, neuropathy, myocardial infarction, stroke, hospitalization, or mortality. Furthermore, cost-effectiveness findings depend heavily on device price, persistence of use, baseline A1C, subsequent treatment changes, and the clinical workflow surrounding the device. [2,3]

Identifying Which Family Medicine Patients Are Most Likely to Benefit

A trial of continuous monitoring is most defensible when the resulting data can directly change clinical management. High-yield situations include the following:

• Patients with an A1C above their individualized goal when it is unclear whether fasting, overnight, or postprandial glucose is the main contributor. [1,3]
• Patients starting a new medication, undergoing a dose change, initiating a nutrition plan, starting an exercise program, or experiencing glucocorticoid exposure where short-term pattern feedback would guide the next clinical step. [1]
• Patients using sulfonylureas, meglitinides, or another noninsulin therapy capable of causing hypoglycemia, especially if they have older age, irregular meals, kidney disease, weight loss, or symptoms that are difficult to characterize. [1,14]
• Cases featuring discordance between laboratory A1C, home glucose measurements, reported symptoms, and the apparent clinical picture. [1,13]
• Patients who are highly motivated to test one or two specific behavioral hypotheses and are willing to actively review the findings rather than simply collect data. [1,3]

Lower-value use includes stable diabetes that is already at goal when no clinical decision is pending, repeated sensor use without a structured review plan, and routine metabolic tracking in healthy adults solely to eliminate every postmeal rise. A 2026 systematic review of 23 studies in people without diabetes found possible short-term effects on mean glucose and behavior, but no significant body mass index benefit and no appreciable glycemic benefit in healthy normoglycemic participants. Most interventions combined the device with structured education, making the independent effect of the sensor uncertain. [8]

Selecting the Most Appropriate Continuous Glucose Monitoring Model

Model	Best use in family medicine	Main limitations
Professional CGM	A focused 10 to 14 day diagnostic-pattern question; can be blinded or unblinded; highly useful when long-term access is uncertain.	Requires clinical placement, data download, interpretation, and a scheduled action visit.
Prescription personal CGM	Ongoing self-management, repeated medication adjustment, data sharing, and alarm-enabled safety when the selected device provides alerts.	Subject to prescription and payer requirements; requires training and supply continuity.
OTC glucose biosensor	Low-barrier pattern feedback for eligible people not using insulin, usually as a self-pay option.	Lacks alarms; carries label exclusions; may encourage overinterpretation without clinical framing.

Professional and intermittent monitoring should always be paired with interpretation, education, and a concrete action plan. [1]

Professional monitoring is often underused in primary care. A short, structured exposure can answer a narrow clinical question without committing the patient to continuous, long-term use. It is particularly useful when the clinician wants blinded data to reduce behavior changes during the monitoring period, when the patient wants a supervised trial, or when insurance coverage for a personal device is uncertain. The clinical value is entirely lost if the report is downloaded but never reviewed with the patient. [1]

Navigating Over-the-Counter CGM Systems and FDA Labeling in 2026

Over-the-counter clearance improves patient access, but clearance is not a blanket recommendation for every potential user and does not establish long-term clinical benefit in every cleared subgroup. Crucially, current over-the-counter devices lack glucose alarms and should never be substituted for a prescription device when real-time hypoglycemia protection or insulin dosing support is needed. [1,10,11]

System	FDA-cleared population	Important boundaries	Wear session
Stelo Glucose Biosensor System	People age 2 years and older who are not using insulin.	Do not use with insulin, problematic hypoglycemia, or dialysis; requires caregiver supervision under age 18; lacks alarms.	Up to 15 days plus a 12-hour grace period.
Lingo Glucose System	Adults age 18 years and older who are not using insulin.	Intended for lifestyle and behavior feedback; no medical action should be taken without a qualified clinician; not for dialysis or critical illness; lacks alarms.	Up to 14 days.

Note: FDA labeling and software can change. Always verify the current instructions for use before recommending or interpreting a specific device. [10,11]

The June 12, 2026 FDA decision for Stelo expanded the cleared age range to 2 years and older for people not using insulin. The label explicitly states that users with problematic hypoglycemia or those on dialysis should not use the system, that users younger than 18 require caregiver supervision, and that acetaminophen above 1,000 mg every 6 hours and hydroxyurea can produce falsely high readings. The Lingo label is limited to adults not using insulin and warns that vitamin C intake above 1,000 mg per day may falsely elevate readings. Product-specific interfering substances must be checked individually rather than generalized across the entire device class. [10,11]

A Practical Step-by-Step Sequence for Interpreting the Ambulatory Glucose Profile

1. Start With the Clinical Question

Before opening the report, define exactly what the sensor was intended to clarify. Examples include distinguishing fasting versus postmeal hyperglycemia, investigating suspected sulfonylurea-related lows, evaluating the effect of a medication change, or determining whether evening eating is driving overnight exposure. Without a predefined question, the review is highly vulnerable to data fishing and overreaction. [1,3]

2. Confirm That the Data Are Interpretable

For reliable pattern interpretation, the international consensus recommends approximately 14 days of data with at least 70 percent active sensor time. Shorter data sets can still identify a repeated safety problem, but confidence in estimates such as time in range and glucose management indicator falls as wear time becomes sparse or behavior during monitoring becomes atypical. [12]

Check for prolonged gaps, a sensor that detached early, unusual illness or travel, major diet changes undertaken only because the patient was wearing the sensor, and values that conflict with symptoms or capillary glucose. Remember that the device measures interstitial glucose and can lag behind blood glucose during rapid changes. Confirm suspicious readings with a standard blood glucose meter, particularly when symptoms and sensor values disagree. [1]

3. Review Time Below Range Before Chasing Hyperglycemia

Hypoglycemia always holds the highest immediate safety priority. For many nonpregnant adults with diabetes, consensus targets are less than 4 percent of time below 70 mg/dL and less than 1 percent below 54 mg/dL. A repeated low pattern should be verified and connected to medication timing, meals, kidney function, alcohol, exercise, and sleep before any glucose-lowering therapy is intensified. [12,14]

4. Review the Core Metrics

Metric	General target for many nonpregnant adults with diabetes	Office question to consider
Time below 54 mg/dL	Less than 1%	Is this confirmed, repeated, and medication related?
Time 54 to 69 mg/dL	Less than 4%	When does the low pattern typically occur?
Time 70 to 180 mg/dL	Greater than 70%	Which recurring pattern keeps the patient out of range?
Time 181 to 250 mg/dL	Less than 25%	Is exposure mainly fasting, postmeal, or all day?
Time above 250 mg/dL	Less than 5%	Is treatment failure, illness, steroid use, or insulin deficiency possible?
Coefficient of variation	36% or less	Is variability itself creating risk or treatment complexity?

These are consensus targets, not rigid diagnostic thresholds. Always individualize goals for frailty, comorbidity, hypoglycemia risk, pregnancy, age, and overall treatment burden. [12,14]

Time in range should be interpreted alongside the baseline A1C, current medications, and the specific clinical goal. Moving from 45 percent to 60 percent time in range may represent a highly meaningful response even though the consensus target has not been reached. Conversely, a high time in range can easily coexist with clinically dangerous lows if the report is reviewed only as a single headline number. [12]

5. Treat Glucose Management Indicator as an Estimate

The glucose management indicator converts mean sensor glucose into an A1C-like estimate. It is not a laboratory measurement, should never be used to diagnose diabetes, and can differ from measured A1C due to recent changes in glucose, sensor sampling, red-cell lifespan, anemia, chronic kidney disease, hemoglobin variants, transfusion, or other factors. Persistent discordance is a valuable clinical clue, not a reason to simply choose whichever number is more convenient. [7,13]

6. Read the Shape of the Ambulatory Glucose Profile

The ambulatory glucose profile compresses multiple days onto a single 24-hour curve. Begin with the median line, then examine the central and outer variability bands. A consistently elevated median suggests a reproducible target for treatment. A wide band suggests day-to-day variation stemming from meal timing, adherence, activity, stress, illness, sleep, or medication exposure. A narrow but high overnight trace represents a completely different problem than a normal overnight trace followed by a sharp breakfast rise. [12]

• Overnight and fasting elevation: Review evening intake, medication duration, sleep schedule, glucocorticoid use, and whether progressive insulin deficiency is plausible. [1,12]
• Repeated postbreakfast or postdinner rise: Review meal composition, portion size, timing, beverage carbohydrate content, activity levels, and whether the current medication plan adequately addresses postprandial exposure. [1,12]
• Lows after activity or during the night: Verify the event and review sulfonylurea or meglitinide dosing, meal omission, alcohol intake, renal function, and exercise timing. [1,14]
• Large day-to-day variability: Look for inconsistent medication use, variable meals, weekend patterns, shift work, acute illness, and sensor artifacts before adding new medications. [12]

7. Make One or Two Changes and Define the Reassessment

A comprehensive review should end with a small number of explicit actions. Examples include reducing a hypoglycemia-prone medication, changing the timing or composition of one recurring meal, adding a short postmeal walk when feasible, addressing missed doses, or escalating pharmacotherapy when hyperglycemia is persistent. Define whether success will be assessed by symptoms, laboratory A1C, a repeat professional monitoring period, or continued personal use. Avoid changing several variables at once when the objective is to learn which specific intervention actually worked. [1,3]

Avoiding Common Data Interpretation Errors in Clinical Practice

• Using the device to diagnose diabetes or prediabetes. The American Diabetes Association states that evidence is insufficient for screening or diagnosis with these devices; always use standard laboratory criteria. [7]
• Treating every postmeal excursion as pathology. In a large cohort without diabetes, normoglycemic adults still spent an average of about 12 percent of their time above 140 mg/dL and more than 15 minutes per day above 180 mg/dL. Context, duration, baseline glycemic status, and reproducibility all matter. [9]
• Equating a flat trace with optimal health. Aggressively minimizing normal glucose variation can foster restrictive eating, anxiety, or unnecessary self-experimentation. [3,8]
• Ignoring sensor limitations. Interstitial lag, compression lows, poor adhesion, dehydration, and product-specific interfering substances can all create misleading values. [1,10,11]
• Letting the device replace the clinical plan. Raw data without education, interpretation, and follow-through has very limited clinical value. [1]

Implementing a 10-Minute Primary Care Review Workflow

• Minute 0 to 2: Restate the clinical question, confirm current medications and recent changes, and check whether the report contains adequate days and active wear time. [1,12]
• Minute 2 to 4: Review time below range, assess symptoms, confirm with a meter if needed, and evaluate medication-related hypoglycemia risk. [1,12,14]
• Minute 4 to 7: Review time in range, time above range, glucose management indicator, variability, and the ambulatory profile median and bands. Identify one dominant, repeatable pattern. [12,13]
• Minute 7 to 9: Agree on one or two medication or lifestyle actions that directly address that specific pattern. [1,3]
• Minute 9 to 10: Document the plan, confirm when the patient should use a standard meter, define urgent symptoms, and set the reassessment method and date. [1,14]

Navigating Insurance Coverage and Clinical Access Barriers

Insurance coverage is not synonymous with clinical appropriateness. Current Medicare durable medical equipment coverage requires a diagnosis of diabetes, documented training and a prescription, use within FDA indications, and a stated purpose of improving glycemic control. The patient must be insulin-treated or have documented problematic hypoglycemia.

For noninsulin-treated beneficiaries, qualifying pathways include recurrent level 2 hypoglycemia below 54 mg/dL despite more than one treatment-plan adjustment, or at least one level 3 event requiring third-party assistance. Continued coverage requires a clinical follow-up at least every six months. Commercial policies and professional monitoring benefits vary widely. [15]

Over-the-counter biosensors purchased for lifestyle feedback do not automatically meet Medicare durable medical equipment rules. A family medicine practice should decide in advance who verifies benefits, who trains the patient, how data enter the chart, who interprets the report, and how follow-up is scheduled. The true clinical intervention is the entire workflow, not just the sensor itself. [1,15]

Recognizing Safety Concerns, Escalation Triggers, and Referral Indications

• Rely on capillary blood glucose when the sensor value does not fit the patient’s symptoms or when glucose is changing rapidly. Never ignore a symptomatic patient simply because the sensor appears reassuring. [1]
• Promptly review the treatment plan for recurrent confirmed values below 54 mg/dL, any event requiring assistance, or a pattern suggesting medication-related hypoglycemia. An over-the-counter device without alarms is not appropriate protection for problematic hypoglycemia. [10,12,14]
• Urgently assess persistent hyperglycemia accompanied by vomiting, dehydration, dyspnea, altered mentation, or ketones. Remember that sodium-glucose cotransporter 2 inhibitors can be associated with ketoacidosis even when glucose is not markedly elevated. [14]
• Consider specialist evaluation when the pattern suggests autoimmune or severe insulin-deficient diabetes, when pregnancy is present, when dialysis or another label-excluded condition complicates device selection, or when recurrent unexplained hypoglycemia persists. [7,10,11,14]
• Ask about skin reactions, adhesive intolerance, medications or supplements that interfere with the selected system, and any history of eating disorders when relevant. [1,10,11]

Addressing CGM Use in Prediabetes and Individuals Without Diabetes

Continuous monitoring may be useful as a short-term biofeedback tool inside a structured lifestyle program, but current evidence does not support treating it as a stand-alone weight-loss intervention, a diagnostic test, or a universal wellness requirement. The non-diabetes literature is heterogeneous, short-term, and often combines the device with coaching, nutrition education, or other behavior-change components. Healthy participants have not shown a clear glycemic benefit, and clinically meaningful long-term outcomes remain unproven. [3,7,8]

When a person without diagnosed diabetes brings consumer data to the clinic, first ask what decision the patient is trying to make. Confirm symptoms and risk factors, avoid diagnosing disease from isolated sensor excursions, and use established laboratory testing when screening is indicated. The ultimate goal is to prevent unnecessary dietary restriction and anxiety while still recognizing when conventional medical evaluation is warranted. [7,9]

Conclusion: Integrating CGM Safely into Family Medicine

Continuous glucose monitoring has successfully moved beyond intensive insulin therapy, but its best use in family medicine remains targeted rather than automatic. In noninsulin-treated type 2 diabetes, it can produce modest average improvements in A1C and time in range, with significantly larger gains in carefully selected patients. Over-the-counter systems expand access but lack alarms and carry specific label limits.

A safe clinical workflow defines the clinical question upfront, verifies adequate and believable data, addresses hypoglycemia first, interprets the ambulatory profile rather than isolated spikes, implements one or two targeted actions, and schedules a clear reassessment. Used correctly, this technology should inform and enhance care, never replacing laboratory diagnosis or sound clinical judgment. [1,2,3,7,10,11,12]

References

1. American Diabetes Association Professional Practice Committee for Diabetes. 7. Diabetes Technology: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49(Suppl 1):S150-S165. doi:10.2337/dc26-S007. PMID: 41358889. DOI. PubMed.
2. Aronson R, Abitbol A, Bajaj HS, Cheng AYY, Christopoulos S, Harris SB, Jain AB, Goldenberg RM. Continuous glucose monitoring in noninsulin-treated type 2 diabetes: A critical review of reported trials with an updated systematic review and meta-analysis of randomised controlled trials. Diabetes Obes Metab. 2025;27(11):6220-6242. doi:10.1111/dom.70008. PMID: 40757453. DOI. PubMed.
3. Dower JA, Johansson M, Camp AW, Montori VM, Lipska KJ. Continuous Glucose Monitoring in Type 2 Diabetes and Beyond: A Review. JAMA Intern Med. Published online August 3, 2026. doi:10.1001/jamainternmed.2026.2772. PMID: 42545686. DOI. PubMed.
4. American Diabetes Association. 82% of Adults with Type 2 Diabetes Not on Insulin Improve Blood Glucose Levels with Continuous Glucose Monitoring. Press release and oral-abstract summary. June 6, 2026. Accessed August 20, 2026. Official source.
5. Shields S, Thomas R, Durham J, Moran J, Clary J, Ciemins EL. Continuous glucose monitoring among adults with type 2 diabetes receiving noninsulin or basal insulin therapy in primary care. Sci Rep. 2024;14(1):31990. doi:10.1038/s41598-024-83548-4. PMID: 39738725. DOI. PubMed.
6. Martens T, Beck RW, Bailey R, et al; MOBILE Study Group. Effect of Continuous Glucose Monitoring on Glycemic Control in Patients With Type 2 Diabetes Treated With Basal Insulin: A Randomized Clinical Trial. JAMA. 2021;325(22):2262-2272. doi:10.1001/jama.2021.7444. PMID: 34077499. DOI. PubMed.
7. American Diabetes Association Professional Practice Committee for Diabetes. 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49(Suppl 1):S27-S49. doi:10.2337/dc26-S002. PMID: 41358893. DOI. PubMed.
8. Liao X, Li Y, Tang S, Xiao Y, Yu X, Huang R, Zhong T. Continuous glucose monitoring in non-diabetic populations: a systematic review of observational and interventional studies with meta-analysis. Eur J Med Res. 2026;31:397. doi:10.1186/s40001-026-03920-0. PMID: 41588451. DOI. PubMed.
9. Spartano NL, et al. Defining Continuous Glucose Monitor Time in Range in a Large, Community-Based Cohort Without Diabetes. J Clin Endocrinol Metab. 2025;110(4):1128-1134. doi:10.1210/clinem/dgae626. PMID: 39257191. DOI. PubMed.
10. U.S. Food and Drug Administration. Stelo Glucose Biosensor System, 510(k) K260935. Decision date June 12, 2026. FDA database last updated August 17, 2026. Accessed August 20, 2026. FDA record. FDA decision summary.
11. U.S. Food and Drug Administration. Lingo Glucose System, 510(k) K233655. Decision date May 29, 2024. FDA database last updated August 17, 2026. Accessed August 20, 2026. FDA record. FDA decision summary.
12. Battelino T, Danne T, Bergenstal RM, et al. Clinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range. Diabetes Care. 2019;42(8):1593-1603. doi:10.2337/dci19-0028. PMID: 31177185. DOI. PubMed.
13. Bergenstal RM, Beck RW, Close KL, et al. Glucose Management Indicator (GMI): A New Term for Estimating A1C From Continuous Glucose Monitoring. Diabetes Care. 2018;41(11):2275-2280. doi:10.2337/dc18-1581. PMID: 30224348. DOI. PubMed.
14. American Diabetes Association Professional Practice Committee for Diabetes. 6. Glycemic Goals, Hypoglycemia, and Hyperglycemic Crises: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49(Suppl 1):S132-S149. doi:10.2337/dc26-S006. PMID: 41358894. DOI. PubMed.
15. Centers for Medicare & Medicaid Services. Glucose Monitor – Policy Article A52464. Current policy reviewed August 20, 2026. CMS policy.