The Dexcom G6 has an overall accuracy rate within about 9 to 11% of a fingerstick blood glucose meter, but several common situations can push readings further off. Some are built into how the technology works, others are fixable on your end. Here’s what’s likely going on and what you can do about it.
The Sensor Doesn’t Measure Blood
The G6 sensor sits just under your skin and measures glucose in interstitial fluid, the liquid between your cells, not in your blood. This creates an inherent time lag. Your blood glucose changes first, and the fluid around the sensor catches up roughly 10 minutes later. During stable periods, you won’t notice the difference. But when your glucose is rising or falling quickly, like after a meal or a correction dose, the sensor reading can be noticeably behind your actual blood sugar.
This means if you do a fingerstick while your glucose is swinging, the two numbers may not match, and that’s expected. The faster the change, the bigger the gap. If you want to compare your sensor to a meter, the most reliable time to do it is when your glucose has been relatively flat for at least 15 to 20 minutes.
The First 24 Hours Are the Least Accurate
A freshly inserted sensor needs time to stabilize in your tissue. During the two-hour warmup, the G6 isn’t giving you readings at all. But even after warmup ends, accuracy continues to improve. FDA data on the G6 shows a mean accuracy error of about 10.7% on days one and two, improving to 9.2% by mid-session (days four and five), then settling around 9.6% near the end of the 10-day wear period.
Dexcom acknowledges this directly: with a newly inserted sensor, the gap between your meter and sensor reading may be greater, and the match generally gets closer over the first 24 hours. If your sensor seems off right after insertion, giving it a full day before judging its performance is reasonable. Some people insert a new sensor a few hours before ending their current session to let the new one “soak” and stabilize, though this isn’t officially recommended.
Compression Lows: False Drops While Sleeping
If your sensor shows sudden, steep drops into low territory at night that bounce back up on their own, you’re almost certainly seeing compression lows. When you sleep on the arm or abdomen where your sensor is placed, body weight compresses the tissue around the sensor. This squeezes out the interstitial fluid, so the sensor detects less glucose and reports a falsely low number.
The telltale sign is a sharp dip that doesn’t match how you feel and resolves when you shift position. The fix is straightforward: place the sensor on a body part you don’t sleep on. If you’re a side sleeper with the sensor on your upper arm, try switching to the other arm or your abdomen. Some people also find that a slightly looser adhesive overlay helps reduce the pressure effect.
Medications That Throw Off Readings
Two substances are specifically known to cause falsely high G6 readings. Hydroxyurea, a medication commonly prescribed for sickle cell disease and certain cancers, is a small molecule that can penetrate the sensor membrane and generate extra electrical current. This makes the sensor think glucose is higher than it actually is. The same thing happens with acetaminophen (Tylenol) at doses above 4 grams per day. At typical doses of one or two pills for a headache, the interference is minimal. But if you’re taking maximum-strength acetaminophen around the clock, your sensor may read higher than your actual blood sugar.
If you’re on either of these medications and noticing persistent high readings that don’t match how you feel or what a fingerstick shows, the medication is a likely culprit. Confirm with a meter before making insulin dosing decisions.
Accuracy Varies by Glucose Range
The G6 is not equally accurate across all glucose levels. FDA clinical data shows a clear pattern in adults:
- Above 250 mg/dL: about 7% average error
- 181 to 250 mg/dL: about 9% error
- 70 to 180 mg/dL: about 11% error
- 54 to 69 mg/dL: about 12% error
- Below 54 mg/dL: about 14% error
In children, the accuracy at very low glucose levels is even more variable, with an average error of 26% below 54 mg/dL. This means the sensor is least reliable precisely when accuracy matters most: during lows. A reading of 55 mg/dL could actually be anywhere from the high 40s to the low 60s. If you’re getting low alerts and something feels off, a fingerstick is worth the extra step.
Dehydration, Temperature, and Site Issues
Because the sensor relies on interstitial fluid, anything that changes fluid levels or circulation around the sensor site can affect accuracy. Dehydration reduces the volume of interstitial fluid available, which can make readings drift. Extreme heat (like a hot tub or sauna) increases blood flow to the skin and can temporarily shift readings. Cold temperatures do the opposite.
The sensor site itself matters too. Scar tissue, areas with very little subcutaneous fat, or spots where the sensor wasn’t inserted cleanly can all contribute to inconsistent readings. If you notice one sensor performing worse than usual, the insertion may have hit a bad spot. Rotating your sites between sessions helps avoid tissue buildup that degrades performance over time.
How to Tell If Your Sensor Is Actually Defective
Some amount of variance is normal. A practical way to check: when your glucose is stable, compare a fingerstick to your sensor reading. If the sensor is within about 20% of your meter (or within 20 mg/dL when you’re below 100), that’s within the expected performance range for the device. A sensor reading of 130 when your meter says 110 is within spec. A sensor reading of 180 when your meter says 110 is not.
If your sensor is consistently off by a wide margin across multiple checks, gives erratic readings, throws repeated error messages, or stops giving readings altogether before day 10, it’s likely defective. Dexcom’s replacement policy covers any sensor that fails to meet published performance expectations during its intended 10-day life. Contact Dexcom Tech Support, and they’ll review the case. If it’s a product failure, they’ll replace it at no cost with no limit on how many replacements you can receive. Even if the issue turns out to be a non-product cause, like an application error, you may still qualify for a courtesy replacement.
Practical Steps to Improve Accuracy
Most G6 accuracy problems come down to a handful of fixable issues. Clean and dry skin before insertion gives the adhesive a solid hold and keeps the sensor stable. Avoid inserting near your waistband, bra strap, or anywhere clothing creates friction or pressure. Let a new sensor settle for a full day before judging its accuracy, and avoid calibrating during rapid glucose changes since that can actually make readings worse.
If your transmitter is near the end of its 90-day life, that can also degrade signal quality. Check the transmitter expiration in your Dexcom app. And keep your phone or receiver within 20 feet of the sensor. Bluetooth signal gaps don’t affect accuracy directly, but they create missing data that can throw off trend arrows, which then affects how you interpret your readings.