• Please Remember: Members are only permitted to share their own experiences. Members are not qualified to give medical advice. Additionally, everyone manages their health differently. Please be respectful of other people's opinions about their own diabetes management.
  • Diabetes UK staff will be logging into the forum at various times throughout this Bank Holiday weekend, however, if you require emergency medical assistance or advice please call 999, or if it is less urgent then please call the 24 hour NHS 111 service on 111. Alternatively, please speak to your GP or healthcare team.
  • We seem to be having technical difficulties with new user accounts. If you are trying to register please check your Spam or Junk folder for your confirmation email. If you still haven't received a confirmation email, please reach out to our support inbox: support.forum@diabetes.org.uk

Libre HbA1c v Actual HbA1c

Thanks @100cas - added to the list

Libre Actual Difference
45 47 -2
51 47 4
41 39 2
35 45 -10
38 47 -9
48 46 2
39 44 -5
48 56 -8
46 58 -12
33 38 -5
45 54 -9
34 37 -3
42 36 6
 
Portugal1000 said:
Libre has been predicting 40 for me for last couple of weeks which i was super excited about until i read this thread i had blood tests yesterday so will report back when i get result’s back.
Click to expand...
So I got my results back today. Libre was predicting 40 but sadly results were 46. Still big improvement on the previous 55 so I shouldn’t be too disappointed. Still a way to go.
 
Had blood tests in mid-September- not diabetes related but nevertheless HbA1c was included: 49.

Blood test at end of October for annual review - HbA1c 59.

Libre predicted 49 throughout the whole period
 
Thanks @JJay added to list.

Libre Actual Difference
45 47 -2
51 47 4
41 39 2
35 45 -10
38 47 -9
48 46 2
39 44 -5
48 56 -8
46 58 -12
33 38 -5
45 54 -9
34 37 -3
42 36 6
49 49 0
49 59 -10
 
Just had my most recent Hba1c results 50 this was from about a month from I remember libre said 46 or 48
 
Added to list
Libre Actual Difference
45 47 -2
51 47 4
41 39 2
35 45 -10
38 47 -9
48 46 2
39 44 -5
48 56 -8
46 58 -12
33 38 -5
45 54 -9
34 37 -3
42 36 6
40 46 -6
49 59 -10
48 50 -2
51 48 3
 
I had my blood test last Friday. I downloaded my reader to LibreView when I got back and it gave me 44mmols, although I am still convinced that is only based on the last 2 weeks of Libre data because that is the AGP report period and I have tried changing it to 90 days, but it changes back when I go to look at the prediction.
Libre usually reports 2-3mmols lower than actual HbA1c for me but results came through last night and my blood result is 44mmols, so a gold star for Libre for getting it spot on.
 
Added to list
Libre Actual Difference
45 47 -2
51 47 4
41 39 2
35 45 -10
38 47 -9
48 46 2
39 44 -5
48 56 -8
46 58 -12
33 38 -5
45 54 -9
34 37 -3
42 36 6
40 46 -6
49 59 -10
48 50 -2
51 48 3
44 44 0
 
On the phone app on my phone it says 90dats with the prediction.
 
My estimate was 38
Actual result 44
 
grovesy said:
On the phone app on my phone it says 90dats with the prediction.
Click to expand...
I don't use the phone app. so I have to go into LibreView to see the prediction. Maybe I am doing something wrong when I try to change the time period.

Surprised to see I am the only one so far who's Libre has got it right out of 18 results. However I was also very surprised that Libre got it right for me!

Are people with -10s and a -12s getting a lot of compression lows?
 
rebrascora said:
I don't use the phone app. so I have to go into LibreView to see the prediction. Maybe I am doing something wrong when I try to change the time period.

Surprised to see I am the only one so far who's Libre has got it right out of 18 results. However I was also very surprised that Libre got it right for me!

Are people with -10s and a -12s getting a lot of compression lows?
Click to expand...
The number in LibreView is GMI, not eA1C, a different formula, a different number. GMI should, usually, be more accurate. I think the list @Docb is creating is the eA1C number from the app? Uncertain. Either way they're not comparable. eA1C will pretty much always be lower than GMI for the same set of BG data.

The people with the -10s and -12s are perhaps posting the eA1C from the app (which typically provides an estimate of HbA1c which is lower than reality, which is why the newer GMI formula was created) and wearing sensors that auto-calibrate on the low side for them (unless they really are compression lows? Would that not set off the alarm all the time and wake a person up?)
 
PerSpinasAdAstra said:
The number in LibreView is GMI, not eA1C, a different formula, a different number. GMI should, usually, be more accurate.
Click to expand...

@PerSpinasAdAstra

For information. 'GMI should, usually, be more accurate. As a measure of average glucose?

1741097291109.png

The hashed red line represents the GMI equation. The solid black line is the least-squares regression line fit to the observed data, providing predicted HbA1c for any given CGM mean glucose. CGM, continuous glucose monitoring; GMI, glucose management indicator.


1741097394725.png

As a means of visualizing the discrepancy between GMI and measured HbA1c, Figure 2 [above] shows a Bland-Altman plot with difference plotted as a function of the average of these two measures. The difference between predicted (GMI) and measured HbA1c is shown on the y-axis. The average of the predicted (GMI) and measured HbA1c is given on the x-axis. The bolded line least-squares regression line fit to the observed data.This plot shows a negative correlation between difference and average such that GMI is higher, on average, than HbA1c when the average value is lower, and GMI is, on average, lower than HbA1c when the average value is higher.

Ref: https://pmc.ncbi.nlm.nih.gov/articles/PMC8255314
 
PerSpinasAdAstra said:
The number in LibreView is GMI, not eA1C, a different formula, a different number. GMI should, usually, be more accurate. I think the list @Docb is creating is the eA1C number from the app? Uncertain. Either way they're not comparable. eA1C will pretty much always be lower than GMI for the same set of BG data.

The people with the -10s and -12s are perhaps posting the eA1C from the app (which typically provides an estimate of HbA1c which is lower than reality, which is why the newer GMI formula was created) and wearing sensors that auto-calibrate on the low side for them (unless they really are compression lows? Would that not set off the alarm all the time and wake a person up?)
Click to expand...
Thanks for the explanation. The GMI is only based on 2 weeks of data though, so I would have thought that would have made it less accurate rather than the eHbA1c on the app which seems to be 90 days of data.
Yes, the alarm goes off when low either due to compression or genuine but there are plenty of times when I sleep through it or it doesn't activate for some reason.

I don't think @Docb was aware there was a difference when he started collating info as he doesn't use a CGM himself. That said, I have been using it for over 4 years and wasn't aware there was a difference, probably because I mostly use the reader.
 
mashedupmatt said:
Interested to know how accurate everyone's Libre guesstimates are against actual blood results - mine was 5 mmol/mol higher on blood test to what Libre was saying - admittedly I don't pay too much attention to what it calculates but would be interested to know others experiences...
Click to expand...
I have used the Libre thing because the the Specialist wants it, but I also do the finger fricking as well and get the same problem. I have used the Decom but but it gives a worse readings. Also the Diabetes Specialist and the Diabetes Nurse when they contact me only look at the last 2 weeks results. They go by this to control your Diabetes but this is useless. I have had Type 1 55 years now and have told the Specialist he must look back at least 2 months to control Diabetes. When I was first diagnosed the Specialist had Type 1 Diabetes and in those days he had to read a set of charts. This new Libre thing does not give accurate readings and this is confirmed by the makers of Freestyle Libre and they said that these Libre things are not always accurate. That is from the makers of Libre directly. So as you say is use both the Blood Test pricking using the Accu-Chek machine as well as a Libre testing thing.
Hope this will help.
 
JITR said:
For information. 'GMI should, usually, be more accurate. As a measure of average glucose?
Click to expand...
GMI uses average glucose in the formula - it is an estimate of what a future HbA1c blood test result might be if the average blood glucose level in the CGM data were consistent in the months leading up to the blood test.

diabetesjournals.org

Glucose Management Indicator (GMI): A New Term for Estimating A1C From Continuous Glucose Monitoring

While A1C is well established as an important risk marker for diabetes complications, with the increasing use of continuous glucose monitoring (CGM) to hel
diabetesjournals.org diabetesjournals.org

The older eA1c formula appears in the app, the newer GMI in LibreView. GMI is thought to be more accurate than the old formula but as the paper you linked demonstrates it's still not truly accurate. I saw a Chinese paper which proposed a third, newer formula that was reputed to me more accurate again but I don't think that went anywhere. Other metrics - time in range, time in tight range and some other one I can't remember are seen as more useful. The more I've learned about how CGMs work the less confident I am that any such formula could work well for a large number of people using auto-calibrated CGMs - the readings would be all over the place, as the list in the posts above seems to demonstrate.

@rebrascora - You can only generate a Blood Glucose Report with a GMI for a two week period, I think, but the GMI is based on the average blood glucose level for a period of time. On the main screen in LibreView you can select say 90 days at the top-left to get your average blood glucose number for each 90 day period, and use the site below to calculate GMI. It's a tricky one because the weeks right before the HbA1c blood test count for more than say 10, 11, 12 weeks ago.

Average BG to GMI conversion calculator: https://www.jaeb.org/gmi/
 
@PerSpinasAdAstra Thanks for that. My 90 day Av. Glucose for that period was 6.4mmols/l which, when put into the GMI calculator you linked, gives me 42.8mmols/mol so just a shade below the 44 I actually got.

I am not convinced that calibrating CGMs is a great advantage because of the error margin on BG meters and the difference between one make of BG meter and another and I generally find that Libre consistently predicts only 2-3 mmols lower than my actual result, which is pretty impressive in my opinion over 4 years worth of 6 monthly testing. Especially if you consider that there is likely an error factor of 1 or 2mmols on an HbA1c result. I don't know that for sure but it seems reasonable.
 
You must log in or register to reply here.
Back
Top