How wrong could "estimated" HBA1C be / Do I need a kick up the backside?

[Tac0caT]

Hi All,

It has been a few months since I last posted - I *think* I am still doing well, but I always have this niggling feeling in the back of my mind that I am going to get a nasty surprise at my next HBA1C test or something like that. Either way I am asking for you to keep me accountable if you think I need it.

I will cut to the chase I have been using Libre 2 CGM since end of Jan. My eating habits changed over Christmas, all of that tempting food and what not kind of kick started a "test" if you like. I had been as little carbs as possible/Keto until Xmas but since then I am kind of trying to find a happy medium, though I will admit I do not really track carbs.

My estimated HBA1C on the Libre app is still showing as 37, average glucose 6.1mmol/l - but I do indulge in naughty things more often than I had been, for example sometimes I might have some breaded chicken pieces for tea, or I might have a Kitkat chunky (because I used to LOVE them) and I will spike to say 10mmol/l, so I go for a walk or workout to bring that back down ASAP, I am not seeing levels at that level for prolonged amounts of time, an hour maybe 2 at most, generally after I have been for a walk or worked out I might see levels hang around 7 or so.

How my daily diet looks at the moment:

Breakfast - generally just a coffee with some sugar free vanilla syrup with either Cream or Blue top milk, this is followed by a flat white when I get to work with sugar free vanilla syrup too. BS usually goes to around 8.0mmol/l around 9am.

Lunch - Usually two Chicken thighs with a salad, or maybe some salt and chilli pork bites.

Snacks - I have recently been keeping a bag of Dry Roasted peanuts nearby and have been snacking on them as and when I feel hungry

Dinner - This could be anything really, something out of the Freshwell recipe collection, or even chicken wings. Every Friday I cook a Chicken Balti (No Rice, max of 4 Poppadoms at 3,3g Carbs each), sometimes I'll have 2 slices of Brown Sourdough or Multiseeded Brown Sourdough topped with Egg Mayo, or No added Sugar Jam, or use it for some cheese on toast if I am not that hungry.

Pudding - I will opt to have a pudding most nights directly after tea, this will generally be a slice of sugar free cake with some cream or extra thick cream, or I will make a chocolate mouse thing using 10g of Dark Chocolate chips, Natural peanut butter and Cream.

My weight has held consistently around 80 / 81kg too.

So tell me, could my estimated HBA1C be massively out and do I need a kick? I am not due my next HBA1C until December as they moved me to yearly, but I think I will call my Doctor on Monday and just ask for 6 monthly instead.

 

[JITR]

You have held your weight down so your HbA1c should be quite stable. A 6 month test would be helpful.

 

[mashedupmatt]

The estimate on Libre can differ a lot on occasions - see this thread for more info on that:

Thread 'Libre HbA1c v Actual HbA1c'

Jan 30, 2025

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...

• mashedupmatt
• Replies: 61
• Forum: General Message Board

• 

 

[JITR]

The estimate on Libre can differ a lot on occasions

Isn't it the trends that matter? Of CGM readings on the one hand and of HbA1c on the other.

 

[Tac0caT]

The estimate on Libre can differ a lot on occasions - see this thread for more info on that:

Thread 'Libre HbA1c v Actual HbA1c'

Jan 30, 2025

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...

• mashedupmatt
• Replies: 61
• Forum: General Message Board

• 

Ah yes, I have just been on Libreview and for both 60 & 90 days my GMI comes out at 42 - which IF correct is obviously higher than where I want it to be ideally, but is not necessarily the end of the word as I believe that is at the lower end of what is considered ideal control?

 

[Docb]

Just to build on @mashedupmatt 's reference to an earlier thread on this topic. I asked if anybody would like to report their libre predictions of Hba1c with a at the time they had a lab result. Theses are the results that came back.

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

As you can see there is a bit of a pattern, the results from the Libre tended to underestimate the lab HbA1c by anything up to 10 or 12 mmol/mol. Even so, there were occasions where the Libra over estimated Hba1c. I appreciate it was a far from rigorous study but there is a trend.

So, my answer to your question is don't be surprised if your actual result is higher than what your libre is predicting. You will not be alone. Equally, if it is bang on or your Lbre prediction is over the actual, you will not be alone, just be in a less crowded place.

 

[Tac0caT]

Just to build on @mashedupmatt 's reference to an earlier thread on this topic. I asked if anybody would like to report their libre predictions of Hba1c with a at the time they had a lab result. Theses are the results that came back.

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

As you can see there is a bit of a pattern, the results from the Libre tended to underestimate the lab HbA1c by anything up to 10 or 12 mmol/mol. Even so, there were occasions where the Libra over estimated Hba1c. I appreciate it was a far from rigorous study but there is a trend.

So, my answer to your question is don't be surprised if your actual result is higher than what your libre is predicting. You will not be alone. Equally, if it is bang on or your Lbre prediction is over the actual, you will not be alone, just be in a less crowded place.

It certainly makes for interesting reading, I would also be interested to see what each users AVG Glucose was as well to see if there is any correlation.
If I can get a HBA1C test and you're still collecting figures, I will provide mine for the thread for you too.

Looking at those figures though and seeing that in some cases there was a difference of 12 points higher it makes me panic a bit.

 

[BobbleHat]

I always wonder what calculations Libre uses to get their estimate from. I know Dexcom do an estimate too but they don't call it an estimated hba1c, in acknowledgement that its not a fair reflection.

Right now, I've been back to using a sensor for 3 weeks (I can't afford to continually self fund so they're an occasional management 'treat' for me), Mine is predicting 31, whereas my daily average is 5.6 for 7 days and 5.4 for 14 days.

That's not in anyway comparable of an hba1c of 31. I know I've been a tiny bit more relaxed in the run up to my next hba1c so I would expect it to be the same or even slightly higher than my last one of 39 in November.

 

[JITR]

Just to build on @mashedupmatt 's reference to an earlier thread on this topic. I asked if anybody would like to report their libre predictions of Hba1c with a at the time they had a lab result. Theses are the results that came back.

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

So, my answer to your question is don't be surprised if your actual result is higher than what your libre is predicting. You will not be alone. Equally, if it is bang on or your Lbre prediction is over the actual, you will not be alone, just be in a less crowded place.

Isn't it the trends that matter [for each person]? Of CGM readings on the one hand and of HbA1c on the other.

I have read that both CGM readings and HbA1c results are quite consistent over time for each person. That means everyone could work out a rough and ready 'personal conversion factor'. For example, for the top four readings in Docb's list:

Libre	Actual	Difference	Conversion factor
45	47	-2	1.04
51	47	4	0.92
41	39	2	0.95
35	45	-10	1.29

 

[Docb]

It has often occurred to me that if I were designing a prediction algorithm I would incorporate a routine for inputting lab HbA1c results and using them to refine the prediction for any individual.