Drugs That Distort CGM Readings: Medications That Cause False Highs and Lows

How Continuous Glucose Monitors Work and Why Drugs Interfere

A CGM measures glucose in interstitial fluid, not blood. A tiny electrochemical sensor sits beneath the skin and generates an electrical current proportional to glucose concentration. Most current-generation sensors use a glucose oxidase enzyme layer that catalyzes the oxidation of glucose, producing hydrogen peroxide. The sensor's platinum electrode then oxidizes that hydrogen peroxide, and the resulting electrical signal is converted into a glucose reading 1.

This electrochemical process is the vulnerability point. Any substance that is itself electrochemically active at the sensor's operating potential can donate or accept electrons at the electrode surface, producing a signal the sensor misinterprets as glucose. The technical term is "electroactive interference." Substances that are oxidized at the same voltage as hydrogen peroxide generate additional current, inflating the glucose reading. Substances that consume hydrogen peroxide or inhibit the enzyme reaction suppress the signal, driving readings down.

Sensor manufacturers apply membranes and coatings designed to block large or charged molecules from reaching the electrode. These barriers work well for many potential interferents but are imperfect. Small, uncharged, electrochemically active molecules can still penetrate 2. The result: a reading your phone displays as 220 mg/dL when your actual blood glucose is 140.

Acetaminophen: The Most Studied CGM Interferent

Acetaminophen (paracetamol, Tylenol) has been the most extensively documented source of CGM interference. Its primary metabolite, N-acetyl-p-benzoquinone imine (NAPQI), is electrochemically active at the operating potential of glucose oxidase-based sensors.

The interference is dose-dependent. A 2017 study evaluating the Dexcom G5 found that a single 1 to 000 mg dose of acetaminophen produced mean CGM elevations of approximately 30 to 40 mg/dL above simultaneous venous glucose, peaking at 2 to 3 hours after ingestion 3. At the maximum recommended daily dose (4 to 000 mg/day), the discrepancy was larger and more sustained.

Newer sensors have made progress. The Dexcom G6, introduced in 2018, incorporated a redesigned membrane that substantially reduced acetaminophen interference, and the Dexcom G7 received FDA clearance with labeling that permits use during acetaminophen administration at standard therapeutic doses (up to 1 to 000 mg every 6 hours) 4. The Abbott FreeStyle Libre 2 and Libre 3 also demonstrate reduced sensitivity to acetaminophen compared to earlier platforms, though Abbott's labeling still advises caution with high doses 5.

The clinical takeaway is generation-specific. If a patient uses a Dexcom G7, therapeutic acetaminophen is generally acceptable. If a patient uses an older system, or takes supratherapeutic doses, a confirmatory fingerstick before corrective insulin dosing is appropriate.

Vitamin C (Ascorbic Acid) and Other Antioxidants

Ascorbic acid is a potent reducing agent. It donates electrons readily at the CGM electrode surface, generating current the sensor reads as glucose. This produces falsely elevated readings.

A 2020 in-vitro evaluation published in the Journal of Diabetes Science and Technology demonstrated that ascorbic acid concentrations corresponding to oral doses of 1 to 000 mg or higher caused clinically significant positive bias in multiple CGM platforms 6. The FreeStyle Libre system's labeling specifically warns that ascorbic acid may affect sensor readings 5.

The threshold matters. Dietary vitamin C intake from food (typically 50 to 200 mg/day) is unlikely to produce clinically meaningful interference. Supplemental doses of 500 mg or above can push interstitial ascorbic acid concentrations into the interference range. Mega-dose regimens (2,000+ mg/day), which some patients take for immune support or as part of integrative medicine protocols, carry the highest risk.

For patients who need both high-dose vitamin C and accurate CGM data, timing the supplement away from critical dosing decisions (such as pre-meal bolus calculations) reduces risk. Taking the supplement after the last meal of the day, when less active insulin dosing is occurring, is one practical strategy.

Hydroxyurea: A Persistent Interferent Across Platforms

Hydroxyurea, used in sickle cell disease and myeloproliferative disorders, produces one of the most consistent and clinically significant CGM interferences. The drug is a small, uncharged, electrochemically active molecule that penetrates sensor membranes easily.

The interference causes falsely elevated glucose readings. Case reports and pharmacovigilance data reported to the FDA document patients on hydroxyurea experiencing CGM readings 50 to 100 mg/dL above simultaneous laboratory glucose values 7. The effect persists throughout the dosing interval because hydroxyurea has a half-life of approximately 3 to 4 hours but is typically dosed daily, maintaining steady-state tissue levels.

The American Diabetes Association's 2024 Standards of Care notes that patients taking hydroxyurea should use CGM data with caution and verify with capillary blood glucose measurements when making treatment decisions 8. This is not a minor population overlap: sickle cell disease accelerates development of type 2 diabetes, and patients with myeloproliferative neoplasms receiving hydroxyurea may also have steroid-induced hyperglycemia requiring monitoring.

For these patients, fingerstick blood glucose monitoring remains the more reliable option for insulin dosing decisions, even when a CGM is worn for trend data.

Salicylic Acid, Dopamine, and Mannitol

Several other medications appear in CGM manufacturer labeling as potential interferents, though the clinical evidence base for each is thinner than for acetaminophen or vitamin C.

Salicylic acid is the active metabolite of aspirin. At high doses (the anti-inflammatory range of 3,000+ mg/day, not the cardioprotective 81 mg dose), salicylic acid can cause falsely elevated CGM readings. Standard low-dose aspirin for cardiovascular protection does not appear to produce clinically meaningful interference 9.

Dopamine, administered intravenously in critical care settings, is electrochemically active at the CGM electrode potential. ICU patients receiving dopamine infusions may show falsely high CGM values 10. This is relevant for hospitals adopting CGM technology for inpatient glucose management, a practice expanding rapidly. The Endocrine Society's 2022 guidance on inpatient CGM use flags vasopressor interference as a consideration 11.

Mannitol, an osmotic diuretic and a sugar alcohol, can interfere with some CGM sensor chemistries because it is structurally similar to glucose. The effect varies by sensor design. Patients receiving IV mannitol for elevated intracranial pressure represent a niche but high-stakes scenario where CGM accuracy is especially important.

Tetracycline antibiotics, including doxycycline, are listed as potential interferents in some CGM labeling. The mechanism may involve chelation effects or direct electrochemical activity, though published clinical data confirming the magnitude of interference remain limited.

How to Identify a Drug-Related CGM Error

A falsely high or low CGM reading looks identical to a real one on the screen. The glucose trace moves smoothly. No alarm indicates "this value may be wrong." Recognizing interference requires clinical suspicion.

Three patterns suggest drug-related CGM error. First, a sudden rise in CGM glucose that does not match carbohydrate intake, activity level, or insulin timing, particularly when it occurs 1 to 3 hours after taking a new medication. Second, a persistent discrepancy between CGM readings and fingerstick values (greater than 20% or greater than 30 mg/dL) that resolves when the medication is discontinued. Third, recurrent "unexplained" hyperglycemia on CGM in a patient whose A1C or fructosamine does not reflect the degree of hyperglycemia the CGM data suggests.

The ADA recommends confirming CGM readings with a capillary blood glucose measurement whenever values are inconsistent with symptoms or clinical expectations 8. Dr. Grazia Aleppo, co-author of the ADA technology guidelines, has stated: "CGM is a powerful tool, but it is not immune to interference. When a reading doesn't match how the patient feels, a fingerstick is the right next step" 12.

A second verification principle applies specifically to insulin dosing. The AACE 2023 consensus statement on CGM use advises that "bolus insulin corrections should not be based solely on a single CGM value that is discordant with the patient's symptomatic state, particularly in patients taking medications known to interfere with sensor accuracy" 13.

Which CGM Platforms Are Most and Least Susceptible

Not all CGMs are equally vulnerable. Sensor chemistry, membrane design, and calibration algorithms all affect interference profiles.

The Dexcom G7 uses a modified glucose oxidase chemistry with improved interference-blocking membranes. It is FDA-cleared for use with therapeutic acetaminophen and demonstrates less sensitivity to ascorbic acid than its predecessors. The G7's factory calibration algorithm also includes interference-detection logic 4.

The Abbott FreeStyle Libre 3 uses a wired enzyme (glucose oxidase) design with a permselective membrane. Its labeling lists ascorbic acid (vitamin C) and salicylic acid as substances that may affect readings. Abbott has not published specific threshold data for vitamin C interference 5.

The Medtronic Guardian 4 sensor, used with the MiniMed 780G system, uses a glucose oxidase platform. Its labeling lists acetaminophen as a potential interferent, though clinical data suggest reduced impact compared to earlier Medtronic sensors 14.

Eversense (Senseonics), an implantable CGM that uses a fluorescence-based measurement rather than electrochemical detection, is theoretically less susceptible to electroactive interferents because it does not rely on electrode current. However, fewer drug-interaction studies have been published for this platform.

A general principle: each time a patient switches CGM platforms, the drug-interference profile changes. A medication that was safe to use alongside one sensor may not be safe with another.

What Normal CGM Readings Look Like and When to Suspect Distortion

For adults with diabetes, the ADA-endorsed international consensus targets define normal CGM performance through Time-in-Range metrics 15. The primary target: at least 70% of readings between 70 and 180 mg/dL. Time below range (<70 mg/dL) should be less than 4%, and time significantly below range (<54 mg/dL) should be less than 1%.

For adults without diabetes who wear CGMs for metabolic optimization, normative data from a 2019 study of 153 healthy adults showed mean sensor glucose of 99 mg/dL, with 96% of time spent between 70 and 140 mg/dL 16. Postprandial peaks rarely exceeded 160 mg/dL.

Drug-related distortion should be suspected when CGM-derived metrics deviate sharply from A1C-predicted averages. The glucose management indicator (GMI), which estimates A1C from mean CGM glucose, normally tracks within 0.3 to 0.5 percentage points of a laboratory A1C. A GMI that exceeds the laboratory A1C by more than 0.5 points (suggesting CGM is reading systematically high) warrants investigation for interferent medications 8.

Practical Steps for Clinicians and Patients

When a patient starts a new medication, check the CGM manufacturer's labeling for listed interferents. The FDA maintains 510(k) summaries for each cleared CGM that include interference testing results 4 5.

For patients taking a known interferent who cannot stop it, implement parallel monitoring: perform 2 to 4 fingerstick glucose checks daily for the first 3 to 5 days of the new medication, comparing each to the simultaneous CGM value. A consistent bias of greater than 20% signals clinically significant interference.

Patients taking high-dose vitamin C supplements should consider reducing the dose to 250 mg or below, or timing the supplement to a period when insulin dosing decisions are minimal. For acetaminophen, switching to ibuprofen or naproxen (which do not interfere with CGM sensors) is an option when clinically appropriate, though NSAIDs carry their own risk profile.

For inpatient settings, the Endocrine Society recommends that hospital CGM protocols include a drug-interference checklist at sensor placement and at each medication reconciliation 11. Patients receiving dopamine, mannitol, or hydroxyurea should have point-of-care capillary glucose as the primary monitoring modality, with CGM used only for trend arrows and alerts.

Confirm CGM glucose with a fingerstick before administering a correction bolus of 2 or more units of rapid-acting insulin in any patient taking acetaminophen, vitamin C above 500 mg, hydroxyurea, high-dose salicylates, or IV dopamine.