The Dexcom G6 is one of the more accurate continuous glucose monitors available, with a published overall accuracy rate (called MARD) of 9.0% in its FDA submission data. In practical terms, that means the sensor’s readings typically fall within about 9% of what a lab-quality blood glucose measurement would show. That’s accurate enough for the FDA to classify it as an integrated CGM, meaning it can be used to make insulin dosing decisions without confirming with a fingerstick.
What the Accuracy Numbers Mean
CGM accuracy is measured using something called Mean Absolute Relative Difference, or MARD. A lower MARD is better. The Dexcom G6’s published MARD of 9.0% comes from its controlled clinical trial. In a real-world study conducted during the Bionic Pancreas pivotal trial, the MARD came in at 11.0%, which is typical since real-world conditions are messier than a clinical setting. The median in that study was 9.3%, meaning half of all readings were even closer to the reference value.
In a pediatric study of participants aged 6 to 17, the G6 performed well, posting a MARD of 7.7%. In that same study, 96.2% of sensor readings fell within 20% of the laboratory reference value (or within 20 mg/dL for readings at or below 100 mg/dL). That’s a strong result, especially for a device worn by active kids and teens.
How Accuracy Changes Over 10 Days
The G6 sensor is designed for 10 days of continuous wear, and accuracy is not identical across that entire period. On day one, about 88.6% of readings fall within 20% of the reference value, with a MARD of 11.5%. By day 10, those numbers actually improve slightly: 90.6% of readings within 20%, and a MARD of 10.6%.
The first 12 hours tend to be the least accurate window. Real-world data from the Bionic Pancreas trial showed a MARD of 13.6% in the first 12 hours, dropping to 10.5% during hours 12 to 24, and settling at 10.1% after the first day. This is why many users notice their readings seem a bit off right after inserting a new sensor. The sensor needs time to stabilize in the interstitial fluid beneath your skin.
The Built-In Lag Effect
The G6 doesn’t measure blood sugar directly. It measures glucose in the fluid between your cells, called interstitial fluid. Glucose moves from your blood into this fluid with a delay, generally estimated at 5 to 15 minutes. During stable blood sugar, this lag is barely noticeable. But when your glucose is rising or falling quickly, after a meal or during exercise, the sensor reading will trail behind what a fingerstick would show. This isn’t a flaw in the device; it’s a physical limitation of where the sensor sits in your body.
FDA Requirements the G6 Must Meet
The Dexcom G6 holds an “integrated CGM” (iCGM) designation from the FDA, which requires meeting a strict set of accuracy benchmarks. Among the key requirements: at least 87% of all readings must fall within 20% of the true blood glucose value. When blood sugar is low (below 70 mg/dL), at least 85% of readings must be within 15 mg/dL of the actual value. And critically, the device must never show a reading below 70 mg/dL when the true value is above 180, or vice versa.
These safety guardrails matter because the iCGM classification allows the G6 to connect with insulin pumps and automated dosing systems. A dangerous misread could trigger the wrong insulin dose, so the FDA sets hard limits on how far off a reading can be at the extremes.
Where the Sensor Goes Matters Less Than You’d Think
The G6 is FDA-approved for placement on the abdomen (or the upper buttocks for children). Many users also wear it on the back of the upper arm. A clinical comparison of abdominal versus upper arm placement found no statistically significant difference in accuracy. The abdominal site showed a MARD of 12.3% while the arm came in at 12.0%, and roughly 85 to 86% of readings from both locations fell in the most accurate zone on a standard accuracy chart. One area where the arm may have a slight edge: detecting low blood sugar. Hypoglycemia sensitivity was 71.1% for the arm compared to 60.0% for the abdomen in one analysis, though this difference should be interpreted cautiously.
No Fingerstick Calibration Required
The G6 was the first real-time CGM approved with factory calibration, meaning it comes pre-calibrated and doesn’t require fingerstick blood tests to stay accurate. This is a meaningful convenience improvement over older CGMs that needed two or more calibrations per day. The tradeoff is that factory calibration makes on-label use more important. The G6’s algorithm is optimized for its approved 10-day wear period. Extending the sensor beyond 10 days by restarting it, a common practice among users, can compromise accuracy because the calibration assumptions no longer hold.
Acetaminophen No Longer a Major Concern
Older Dexcom models were notorious for giving falsely high readings when users took acetaminophen (Tylenol). The G6 addressed this with a redesigned sensor membrane. In Dexcom’s clinical testing, a 1-gram dose of acetaminophen produced an average interference of just 3.1 mg/dL, well below the 10 mg/dL threshold the study was designed around. That level of interference is small enough to have a negligible effect on insulin dosing decisions. This was a significant improvement, since acetaminophen is one of the most commonly used over-the-counter medications.
When Accuracy Drops
Several real-world factors can push accuracy below the numbers seen in clinical trials. Compression lows are among the most common: if you sleep on the sensor, the pressure can restrict blood flow to the area and cause a false low reading, sometimes triggering alarms in the middle of the night. Dehydration can also affect readings, since the sensor depends on adequate fluid in the tissue. Scar tissue from repeated insertions in the same spot, extreme temperatures, and significant swelling at the insertion site can all degrade performance.
Rapidly changing glucose levels also reduce accuracy, not because the sensor is malfunctioning, but because of the physiological lag between blood and interstitial fluid. If you see a number that doesn’t match how you feel, especially during or after exercise, a confirming fingerstick is reasonable even though the G6 doesn’t routinely require one.