CGM Sensor Lag and Artifacts

Why sensor readings can mislead

A continuous glucose monitor (CGM) reports an estimate of blood glucose from the fluid under your skin. [1] That design creates two well-known sources of error: a delay (lag) and false readings called artifacts. This page explains both. It describes research on specific sensors and does not rank devices or claim any hands-on testing.

Evidence labels: established (consistent findings), insufficient evidence (too little research), and Proper evidence not found (we looked and found nothing usable).

Sensor lag

There are two parts to the delay. The first is biology: sugar takes a few minutes to move from blood into the fluid under the skin. In fasted adults, direct measurements gave about 5.3 to 6.2 minutes in healthy volunteers [2] and a median 6.8 minutes (range 4.8 to 9.8) in adults with type 1 diabetes. [3] The second part is the sensor and its software. A study that compared two commercial sensors with a laboratory analyzer reported a varying distribution of lag, from about minus 10 to more than 10 minutes. [4] That is one abstract, about those models. Do not assume the same numbers for other devices or newer versions.

What it means: when glucose moves quickly, for example after a fast-digesting meal or during exercise, the sensor line can trail the true value. For adults with type 1 diabetes, NICE guidance keeps fingerstick testing as a back-up when levels change quickly or the device stops working. [5]

Pressure artifacts ("compression lows")

When you lie on a sensor, pressure on the area can make the reading drop even though your blood glucose has not. A modeling study of 72 people wearing one older sensor model (a Dexcom G4 Platinum) isolated these events. [6] Median compression artifacts lasted 45 minutes. [6] They happened more often at night, while disconnections, which are gaps in the signal, were more often during the day. [7] About 90% of disconnections lasted under 20 minutes. [8]

Proper evidence not found: we searched PubMed and Europe PMC and found no readable study measuring how often pressure artifacts occur with current consumer sensors in people without diabetes. We also did not find a source showing whether people with normal glucose are more likely to see false lows. Healthy people do have genuine brief low readings, so a low line is not automatically an artifact. In 60 healthy adults, median time below 70 mg/dL was 1.6%. [9] [10]

Other practical issues

• Sensors are replaced on a schedule, for disposable models every 7 to 14 days depending on the model. [11]
• Skin redness or irritation from the sticky patch can occur. [12]
• The first US over-the-counter CGM clearance listed local infection, skin irritation and pain or discomfort as adverse events. [13]

When to double-check

NIDDK says people may sometimes need to compare CGM readings with a fingerstick on a standard meter. [14] NIDDK lists doubting the accuracy of a reading as one such reason. [15] If you have symptoms that do not match the sensor, such as shakiness, dizziness or confusion, treat the symptoms as real and seek care. [16] In a US emergency, call 911. If you use diabetes medicine, follow your care team's plan and do not change doses based on one sensor reading. See Diabetes Medicines and Consumer Glucose Data.

How to think about it

One odd reading is not a finding. Look for a pattern, and ask whether a lag, pressure or a loose sensor could explain it. Related pages: Interstitial vs Blood Glucose, CGM vs Finger-Stick Glucose Measurement, How to Read a Wearable Validation Study.