Dexcom G6 underestimates HbA1c

[briansolar1]

Hi. I am T1 and take multiple daily injections and use a Dexcom G6 sensor.
On the G6, my 90 day HbA1c estimate is consistently around 6.4-6.5%.
However when I get a blood sample HbA1c done from the doctors it has been 7.1-7.2% the last year or so.

Has anyone else had that type of discrepancy or thoughts on the reason for the difference?

Thanks for your help.
Brian

 

[urbanracer]

I don't use Dexcom but these types of devices can only work out the arithmetic average based on your readings.

Plus, the device measures the glucose in interstitial fluids , not in actual blood. Then it does a fancy calculation to interperet the readings and approximate A1c results. If your body does not match their standard model (who's does?) then it will be a bit out.

It is just an "estimate".

Edited to add:- do you still do finger pricks? How close are the Dexcom readings for you?

 

[Juicyj]

Hello @briansolar1 As per what @urbanracer has said, it's not an accurate measurement.

I also use the G6, but never use their estimate as a guide - instead I focus on time in range which for me is a better indicator as to how I'm doing, I also use their Clarity app to drill in on time ranges and see where I may need to make adjustments too.

 

[In Response]

Another thing to add to @urbanracer 's comments is that the estimate from G6 and Libre will be taking an absolute average over the last 3 months. However, I understand that a true HbA1C is weighted towards more recent levels.
This was very clear for me last year when I was unwell, fighting high blood sugars just before last year's test and my HbA1C was significantly higher than usual and my Libre estimate.

 

[searley]

I use nightscout and that has very good hba1c calculations my last one was almost the same

Mind you the Dexcom wasn’t far off

 

[briansolar1]

I don't use Dexcom but these types of devices can only work out the arithmetic average based on your readings.

Plus, the device measures the glucose in interstitial fluids , not in actual blood. Then it does a fancy calculation to interperet the readings and approximate A1c results. If your body does not match their standard model (who's does?) then it will be a bit out.

It is just an "estimate".

Edited to add:- do you still do finger pricks? How close are the Dexcom readings for you?

Hi,
Thanks for that.
Yes I still do finger pricks- useful for the times that the G6 fails, and I take a reading every morning when I wake up to calibrate or check the G6. At that time the G6 is pretty accurate when my blood sugars are often steady- maybe +/- 0.5 mmol/l between the test strips and the G6. But of course the test strips can be inaccurate as well...

 

[briansolar1]

Hello @briansolar1 As per what @urbanracer has said, it's not an accurate measurement.

I also use the G6, but never use their estimate as a guide - instead I focus on time in range which for me is a better indicator as to how I'm doing, I also use their Clarity app to drill in on time ranges and see where I may need to make adjustments too.

The stats the Clarity app provide are very informative, particularly where blood sugar is worse at particular times of day. I am finding my 'time in range' is between 75% and 85%.

 

[briansolar1]

Another thing to add to @urbanracer 's comments is that the estimate from G6 and Libre will be taking an absolute average over the last 3 months. However, I understand that a true HbA1C is weighted towards more recent levels.
This was very clear for me last year when I was unwell, fighting high blood sugars just before last year's test and my HbA1C was significantly higher than usual and my Libre estimate.

Thanks for that, I hadn't realised that was the case, so in that respect at least the G6 and Libre might be providing a better estimate.

 

[briansolar1]

I use nightscout and that has very good hba1c calculations my last one was almost the same

Mind you the Dexcom wasn’t far off

Interesting, I wasn't aware of Nightscout and will have a look....
Brian

 

[EllieM]

The dexcom does allow you to calibrate it against a blood glucose monitor, but if it seems to match your meter then I'd be more inclined to trust it than the hba1c. There are a number of reasons why hba1cs can be inaccurate (ag anaemia) because they assume that you have a certain number of red blood cells that live roughly 3 months. If your body is slightly off the standard model, then the result can be inaccurate. You could try asking for a fructosamine test if you think that might be the case.

Personally I've found the dexcom hba1c estimates to be slightly less than my actual result, but only by a small amount (eg latest was 51 or 6.8 while dexcom was 6.7).