A nurse told me that her new Dexcom sensor was considered to have better accuracy than a traditional BG machine. Can this rumor be confirmed or denied? Cited somewhere? Its very interesting.
1 Like
Interstitial fluid will NEVER be as or more accurate than blood.
1 Like
We may use both to get information, but comparing their accuracy is not an appropriate question, because they use different technologies and are actually not even measuring the exact same thing.
The nurse is wrong. Rumors are a horrible thing. 
Uh, then what exactly is a CGM measuring if not glucose levels?
There are 2 “kinds” of glucose that we can measure: Blood Glucose, which is what we measure when we do a BG with our meters, and Sensor Glucose (some people use different names for this), which is (as DanP said) a fluid in the blood. I am probably not describing sensor glucose in a completely technically accurate way, but the point is that they are not the same thing. The one that is of most significance to us is the Blood Glucose, we look at sensor glucose more as a way of estimating or looking for patterns in Blood Glucose.
Really gets into what is meant. Certainly many people have strong opinions.
I do feel that a cgm that works well for an individual can provide a vast amount of highly useful information.
We no longer use a meter routinely with our G6. Periodically we will check the G6 against the meter just to be sure we are not being fooled. Maybe every 2 or 3 days or if the BG doesn’t seem reasonable particularly at nighttime before bed. But so far the meter has agreed with the G6 each time to within single digits.
So for us, we are finding the information provided by the Dexcom G6 cgm to be huge with the accuracy to be on par (as far as we can reasonably tell) with our Contour Next One bg meter.
1 Like
I don’t doubt that next gen sensors will be every bit as reliable. My Guardian (minimed 670G) sensors are consistently spot on. Never outside of a 10 point differential. I completely trust it. I feel my lows after nearly a decade of hypo-unawareness…because it just never lets me go low. And that’s crazy because I hated the minimed sensors, Gen 1, Enlite, until this one. It’s really great. I can’t be enthusiastic enough about Guardians! (And I know some swear by Dexcom over Minimed)
1 Like
I guess the gold standard is hospital lab for blood glucose. That’s how you read BG’s with four digits. So, I think we would compare various readings to that. I have an older sensor. It was a very surprising comment, so I just thought I would verify. Traditional BG machine has a high amount of inaccuracy built into it, so I was super hopeful that newer sensors might have beat the old, gold standard. Soon, maybe.
Let me know when you are calibrating your blood glucose meter from your CGM reading, instead of the other way around…
1 Like
I don’t calibrate beyond the initial calibration. Haven’t for years. No need. But, let me clarify.
I believe she was using Dexcom G6. I use a G4.
Here’s what I found…now, we can argue about it. I wish someone other than Dexcom had published on this. They probably will soon. Its a relatively new device. But, this must have been what she was referring to,
"Dexcom showed new data from a 49-person trial testing use of the new G6 sensor over ten days of wear. The average error compared to lab measurement was just 8.1% with one fingerstick calibration per day (after startup), and an expected 8.8% error with no fingerstick calibration at all. Notably, 96% of G6 readings were within 20% or 20 mg/dl of the true glucose value, the strongest data Dexcom has ever released. Improving accuracy – and especially minimizing the error percentage – is important because a difference of plus or minus 20 mg/dl can have big implications in the hypoglycemia range. "
https://diatribe.org/dexcom-g6-sensor-shines-early-accuracy-study
So, we might say that 4% of Dexcom readings are less accurate than what BG machines publish for accuracy in order for FDA to put them on the market (although, many of the traditional BG machine claims regarding accuracy have been called into question by the Diabetes Technology Society.) So…I guess I kinda buy it. I would say that her claim was true, or at least backed by what they publish.
@Eric2, Apparently they are not using traditional machines for calibration anymore. Its pretty shocking news. FDA Authorizes Dexcom G6 CGM - Eliminating the need for fingerstick blood testing for people with diabetes | Dexcom
I am embarrassed to admit but hoping to inform how important to double check the code on the strip bottle against the meter screen. I was so disappointed with the wide difference in comparison. I read a post from a Facebook group. Noticed that changing codes. Sure enough my codes did not match 
. Entered the correct code and Shazam an significant improvement. I should have known this 
1 Like
I wore the Guardian 3 and G5 concurrently for a couple weeks and I agree, they were pretty indistinguishable accuracy-wise, except the Guardians consistently wanted to run low overnight. Meant that on auto, I was continuously waking up too high b/c the pump had been underestimating my insulin needs for the previous several hours. Couldn’t figure out a fix for that. Glad you’re having better results!
Sounds like you’re speaking of a OneTouch meter? I used those for years–decades?–but I noticed for the last several years that the strips were always coded 25, where it used to change periodically. That coincided with other meters touting “No strip codes!” so I figured OneTouch had just figured out a way to standardize on one without making a big deal of it. I still have one as an emergency back-up meter, with a vial of strips. They’re coded 25.
Given that home use meters comprise a wide range of manufacturers with a wide range of accuracy, that might be kind of a low bar. The Dexcom article @mohe0001 cites specifies that they’re comparing against “lab measurement,” which as several others note is much more accurate than even the best OTC meters. I believe the FDA standard for domestic-use meters is ±20% (pretty lax, really). So when they say this–
–they’re saying that 96% of the time they’re meeting the standard required of finger-stick meters. That’s good, but again maybe not quite as high a bar as the nurse’s remark in the OP implies.
As many have been saying since back when hybrid closed-loop was just starting to be a discussion topic here, it’s not like the standard for standard BG meters is all that tight, so who’s to say that it’s your CGM that’s wrong when they diverge by > 20%. But like the Dexcom article says, “a difference of plus or minus 20 mg/dl can have big implications,” and it’s hard not to favor the BG meter in those circumstances.
Dexcom G6 CGM measures “glucose equivalent” contents in interstitial fluid and correlates the readings to the laboratory blood test results, while the traditional BG Meter measures the actual glucose contents in the whole blood. The traditional BG Meter readings are approximately 10% lower than the laboratory test results due to the removal of blood plates, etc. in the preparation (i.e. centrifugal of blood sample).
The interstitial fluid readings have 10-15 minutes lag time, as compared to traditional BG meter’s readings. Make sure to take “stabilized” CGM readings to compare the traditional BG Meter with proper lag time in consideration.
Although CGM accuracy is important, the repeatability of sample measurement is important as well. The expected G6 sensor life of 10 days is probably the limit due to sensor being somewhat degraded and the results are not repeatable or reliable.
FDA should request all CGM manufacturers to report factory data for a longer period of time beyond the guaranteed duration, so that all CGM users are informed.
They are claiming 8% error on average, against lab. Now, we would need the error from the other 4% of data points in order to do a hard and fast calculation of which is better. But, I think its pretty safe to assume that Dex has 'em beat. We know theres gonna be higher error at the extremes - super high or super low BG. So, its partially dependent on how much time you spend in the extremes, like below 40 or above 400. I know the old model G4 doesn’t profess to read those numbers well.
Now, I know that most of us have seen this, but just to review meter accuracy: https://www.diabetestechnology.org/surveillance.shtml
QUESTION: Could I say that Dex has beat the manual BG machine standard in accuracy? Would you feel comfortable stating that in front of a room full of people, given the evidence? If not, how would you phrase the results so as to be perfectly honest about the state of blood glucose monitoring tech?