Dear Dr. Edelman,
I love my CGM, but sometimes it’s considerably off from my finger stick number and I was wondering if there’s an acceptable variance between the two? How accurate should CGM values be compared to values from blood glucose meters?
Dr. Edelman: A lot of us folks with diabetes who use CGM devices wonder how accurate they are, and also how accurate they are required to be. There are several schools of thought, but most feel the following national standards are reasonable:
- For glucose values less than 100 mg/dL, the CGM value should be within 15 mg/dL of the blood glucose meter number. For example, if the blood glucose meter number is 90 mg/dL, then the CGM value should be between 75 and 105 mg/dL.
- For glucose values greater than 100 mg/dL, the CGM value should be within 15% of the meter value. For example, if the blood glucose meter value is 180 mg/dL, then the CGM value should be between 153 and 207 mg/dL.
There are still lots of other variables that come into play. Were your hands clean and dry before testing with your meter? Were the glucose strips exposed to air for an extended period of time, or were they expired? Was it the first day of a new sensor? What brand of CGM are you using?
Remember that CGM devices continuously auto-calibrate, and because you get a new value every five minutes along with a trend, having a CGM number that’s not exactly perfect isn’t as detrimental as a single finger stick that is way off.
Finally, remember to test your glucose using a blood glucose meter when your symptoms do not match the CGM value.
Fifteen percent tolerance isn’t good enough! If my son’s Dexcom were only accurate to within that standard, we’ve have never bothered to buy it. In our experience (after the first day of a news sensor) the G6 is usually the same as a finger stick within about 5 mg/dl, for just about any reading from about 60 to over 200. I think the +/- 15% standard is too lax, too outdated. Since today’s CGM’s are typically much more accurate than the standard, the standard should be updated to reflect reality.
Wife freestyle 1 is now, after 3 days try #2, 30 points low compared from the finger. the freestyle 2 failed on launch, they sending us another. Dexcom was 10-20 points off the whole 10 days. now that the freestyle 1 sensors (x 2) are acting the exact same, I believe it just that way. But one finger stick compared to 6 or more fingers sticks a day, are worth just going along with the addition.
My dexcom 7 measurements are 40 plus points higher than my BGM . Ive changed sensors, somethig is wrong. It goes off all day and night. Frustrating.
Most people do not experience those wide differences all the time. Typically if you’re getting symptoms that don’t match your blood sugar reading, then you should test with your meter. Other than that, there’s no need to test. The G7 calibrates itself over time. But if you’re taking insulin, and you’re concerned, then yes, prick your finger.
My free style Libre is 30 points lower than my glucometer with test strips two months old. Oddly, the sensor I was using directly before this new one showing the much lower readings was within 10 points of the glucometer. I can’t figure out which reading to use for insulin. I am a type 2 diabetic. Very confusing with this new sensor.
You always have to err on the side of caution, and use the number that give you the lower dose. Remember that when your numbers are higher, let’s say around 200, a difference of 30 is within the realm of normal.
My meter runs very low difference has been off as much as 90 points or can be as little as 2 or 3 points difference from my finger stick.I can’t use libera 3 This meter is going to kill me. If this is normal I can’t I can tell any one to ude Libre 3.
Complaints about any CGM system do seem to be fewer and far between, but that is a big difference, and you might have a bad sensor. If it continues to happen with other sensors, you might try the Dexcom because you can calibrate it.
ICF fluid in the CGM will always be different than a blood finger stick . Normal to be different! Err on the sise of lower insulin dose.
Good advice!
I have the Dexcom 7 on my second sensor and my cgm says 400 but finger sticks are saying 220. I have type 2 already went into diabetic coma several years ago need to have accurate numbers so I know how much insulin to take. Decided to go with cgm cuz I work construction. Fingers hurt testing 6 to 7 times a day. Have had this sensor on for 3 days.
If we’re assuming the finger stick is correct, the CGM should be within 20% above or below it, so that is a significant difference. Sensors are not as accurate the first 24 hours and sometimes that can be a little longer, but you can always call Dexcom customer support and get a new sensor. If you’re worried about the correct insulin dose, use a finger stick when you’re giving insulin, and calibrate the CGM once a day at most.
A friend said that the finger stick reading is from what your blood sugar was 15-20 minutes previously – is that correct? Where the CGM reading was from 5 min previously.
Not correct…the finger stick is pretty close to the blood in circulation and CGMs are usually 5-10 minutes behind. It’s called the “lag time”.
on first set of Frestyle GCM. first sensor was close to fingers stick readings. Second went low low the first day and then back to the ranges i had been seeing. Then 9 days in, its going low difference of 50, finger stick being higher and the CGM riding the low limit. I’m going to get the second set, but have questions about the accuracy of the second monitor.
on first set of Frestyle GCM. first sensor was close to fingers stick readings. Second went low low the first day and then back to the ranges i had been seeing. Then 9 days in, its going low difference of 50, finger stick being higher and the CGM riding the low limit. I’m going to get the second set, but have questions about the accuracy of the second monitor.
autocorrect got me. re submitting to correct website
Stop comparing the numbers – they’re not comparable because your finger stick values do not have trend arrows.
I have been monitoring blood glucose using an Accu-Chek Guide Me glucometer that I read in tandem with my Freestyle Libre 2 (now FSL2+) CGM for my first morning reading. I average an 11.94% variance in the past 3 years of CGM readings over BGM which is within the variation range reported between capillary and interstial glucose readings. I also monitor variation by sensor serial number and lot number averaging 12.4% variance. There can be a rare widespread variance with particular sensors. (One sensor was just over 30% variance) Also, since I used to work in hospital laboratory, I started BG sticks at the top of my knee instead of creating an infection control issue with holes in my fingers. This alternate site testing site average has been fairly consistent with my 3-month A1C measurements. Most importantly, the site has less nerves than fingers and is tolerated much better, ie. improved compliance.
Thanks for sharing what works for you, Paul.