The Fallacy of Average: How Using HbA1c Alone to Assess Glycemic Control Can Be Misleading

Brunneria

Guru
Retired Moderator
Messages
21,889
Type of diabetes
Type 2
Treatment type
Diet only
My experience is very similar to @bluetit 's
Like her, my finger prick testing using two different meters, is rarely in the 7s, and like her, my HbA1cs place my average readings at 7mmol/l or a little below. I usually run between 4 and 6.5 in my prick tests.

Since I have a part time Libre, I can see that my Libre graphs perfectly track my prick testing, but run lower.

My conclusion is that the Libre reads low (but is very useful for trends and showing sharp rises, falls and timing peaks).
The prick testing is reasonably accurate but only shows snapshots, so is of limited use.
and my HbA1c reads high for whatever weird reason and is pretty useless for anything except making me scratch my head and worry.

Having said that, I would be much more wary if my HbA1c result ran lower than my prick testing. That could have an impact on diagnosis, treatment and future complications, with health care professionals not recognising that there was a problem.
- so I completely understand your concerns @phdiabetic
 

Bluetit1802

Legend
Messages
25,216
Type of diabetes
Type 2 (in remission!)
Treatment type
Diet only
I should add to my previous post:

Like @Brunneria my Libres read low, but are consistent. I do lots of calibrations with my two meters. Because the Libre reads low I always add 1mmol/l to the averages it comes up with (and it never runs as much as 1mmol/l lower) and still the averages are nowhere near the 7.3 the HbA1c gives me. In fact, the Libre I was wearing up to a few days before my last HbA1c gave me an average of 5.1mmol/l. So in my head that was 6.1mmol/l, which was more or less the same as my finger prick averages. Where did the HbA1c of 44 come from?

Also like Brunneria, I much prefer them to be higher than lower. Lower would lead me to a false sense of security, so I do sympathise with @phdiabetic . Wouldn't it be lovely if one size did fit all?.
 

phdiabetic

Well-Known Member
Messages
880
Type of diabetes
Type 1
Treatment type
Insulin
I should add to my previous post:

Like @Brunneria my Libres read low, but are consistent. I do lots of calibrations with my two meters. Because the Libre reads low I always add 1mmol/l to the averages it comes up with (and it never runs as much as 1mmol/l lower) and still the averages are nowhere near the 7.3 the HbA1c gives me. In fact, the Libre I was wearing up to a few days before my last HbA1c gave me an average of 5.1mmol/l. So in my head that was 6.1mmol/l, which was more or less the same as my finger prick averages. Where did the HbA1c of 44 come from?

Also like Brunneria, I much prefer them to be higher than lower. Lower would lead me to a false sense of security, so I do sympathise with @phdiabetic . Wouldn't it be lovely if one size did fit all?.

I would tend to put this down to libre/meter inaccuracy, but it's also very suspicious! It's especially weird that the two of us seem to have the opposite problem haha, I would at least expect this sort of thing to be consistent.

I wonder about this sort of thing because my last HbA1c was the lowest my doctor's ever seen, including non-diabetic people, and my current number is still lower than pretty much all the non-diabetics she's seen. Recently I struggled with getting my basal right, as well as stress, and spent quite a bit of time in the 10's-13's. I also didn't have any hypos. So if normal people also don't have hypos, and they don't go as high as I did, how is it possible that my HbA1c is better than theirs??
 

Bluetit1802

Legend
Messages
25,216
Type of diabetes
Type 2 (in remission!)
Treatment type
Diet only
I would tend to put this down to libre/meter inaccuracy, but it's also very suspicious! It's especially weird that the two of us seem to have the opposite problem haha, I would at least expect this sort of thing to be consistent.

I wonder about this sort of thing because my last HbA1c was the lowest my doctor's ever seen, including non-diabetic people, and my current number is still lower than pretty much all the non-diabetics she's seen. Recently I struggled with getting my basal right, as well as stress, and spent quite a bit of time in the 10's-13's. I also didn't have any hypos. So if normal people also don't have hypos, and they don't go as high as I did, how is it possible that my HbA1c is better than theirs??

I doubt very much it is Libre/meter inaccuracy when averaging over a decent period of time, and 2 to 3 months prior to an HbA1c is a very decent period of time. . Meters aren't all that inaccurate on the whole, and any "rogue" readings are evened out and become insignificant. It sounds to me like your red blood cells die off and renew a lot faster than the standard 120 days. Whereas mine live a lot longer than that.

The other point is that the HbA1c average is calculated with a lot of emphasis on the previous couple of weeks, so any sudden improvements or deterioration in those 2 weeks will have an effect. Another thing is blood being donated - if a person donates blood in this period they will have less glycated haemoglobin and more brand new red blood cells with hardly any glycation.

It is a mine field and a mystery yet to be solved.
 

phdiabetic

Well-Known Member
Messages
880
Type of diabetes
Type 1
Treatment type
Insulin
I doubt very much it is Libre/meter inaccuracy when averaging over a decent period of time, and 2 to 3 months prior to an HbA1c is a very decent period of time. . Meters aren't all that inaccurate on the whole, and any "rogue" readings are evened out and become insignificant. It sounds to me like your red blood cells die off and renew a lot faster than the standard 120 days. Whereas mine live a lot longer than that.

The other point is that the HbA1c average is calculated with a lot of emphasis on the previous couple of weeks, so any sudden improvements or deterioration in those 2 weeks will have an effect. Another thing is blood being donated - if a person donates blood in this period they will have less glycated haemoglobin and more brand new red blood cells with hardly any glycation.

It is a mine field and a mystery yet to be solved.

I wish there could be a test for red blood cell lifespan, I am quite curious now. However, if mine died faster than everyone else's, wouldn't my recent troubles cause a much higher HbA1c than my previous one (which was taken during uni break when I was not stressed, and was able to manage my numbers a lot better while I was at home)? I went from 27 to 29 (or 4.6 to 4.8), which are really almost the same. Based on the actual finger pricks/CGM readings, I expected a large increase. I also had the worst numbers in the days leading up to the blood test. (And I didn't donate or receive blood).

I guess there's no use worrying about it, I'll just try my best and hope that it'll be ok.
 

Bluetit1802

Legend
Messages
25,216
Type of diabetes
Type 2 (in remission!)
Treatment type
Diet only
Thanks for that, it will take me a while to study.
Have you read that thread @Bluetit1802 ?

Yes, I have read it. I am a little sceptical because they are based very old studies from the 1980s, and I'm not sure about their conclusion that glycation is reversible, and that earlier haemoglobin from the first part of the 2 to 3 month period has "lost" its glycation. This research was done in a test tube. Maybe it is different when the blood is still in the human body?

As the HbA1c is a very important tool in diagnosing diabetes and also monitoring control, I cannot understand why this research has not been taken on board by the powers that be, and why every one of the powers that be continue to believe that red blood cell survival times are so significant. However, I am willing to learn. Whatever, it is clear that the HbA1c is not as reliable as our doctors seem to think it is.

http://www.diabetes.co.uk/forum/thr...due-to-red-blood-cell-age-variability.110793/

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2581997/

https://patient.info/doctor/glycated-haemoglobin-hba1c

http://care.diabetesjournals.org/content/40/8/994

http://www.gpnotebook.co.uk/simplepage.cfm?ID=x20070614192625295600&linkID=70250&cook=yes

https://pbs.twimg.com/media/DG4FKXNUQAA2rMo.jpg
 

tim2000s

Expert
Retired Moderator
Messages
8,934
Type of diabetes
Type 1
Treatment type
Other
On the topic of the 50 year medallists study, I think it's a Joslin study, and has been going on for some time. This is one of the publications: https://www.ncbi.nlm.nih.gov/pubmed/20699420

There are quite a few others linked to it.

Answering the question on Hba1C - the reason this is used is:
  1. The DCCT trial broadly shows that across a population, Hba1C has a direct correlation with the risk of complications, in spite of it's limitations.
  2. The alternative methods of providing an indication of overall "diabetes health" are all rather expensive, and as a result, nowhere wants to pay for them. Imagine CGM for all T1s in order to use GLycaemic variation measures. 1,000s of times more expensive than a three monthly hba1C.
 

Mr_Pot

Well-Known Member
Messages
4,573
Type of diabetes
Type 2
Treatment type
Diet only
Yes, I have read it. I am a little sceptical because they are based very old studies from the 1980s, and I'm not sure about their conclusion that glycation is reversible, and that earlier haemoglobin from the first part of the 2 to 3 month period has "lost" its glycation. This research was done in a test tube. Maybe it is different when the blood is still in the human body?

As the HbA1c is a very important tool in diagnosing diabetes and also monitoring control, I cannot understand why this research has not been taken on board by the powers that be, and why every one of the powers that be continue to believe that red blood cell survival times are so significant. However, I am willing to learn. Whatever, it is clear that the HbA1c is not as reliable as our doctors seem to think it is.

http://www.diabetes.co.uk/forum/thr...due-to-red-blood-cell-age-variability.110793/

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2581997/

https://patient.info/doctor/glycated-haemoglobin-hba1c

http://care.diabetesjournals.org/content/40/8/994

http://www.gpnotebook.co.uk/simplepage.cfm?ID=x20070614192625295600&linkID=70250&cook=yes

https://pbs.twimg.com/media/DG4FKXNUQAA2rMo.jpg
You said in post #24 that HbA1c is calculated, did you mean measured?
 

Art Of Flowers

Well-Known Member
Messages
956
Type of diabetes
I reversed my Type 2
Treatment type
Diet only
I saw an article the other day where it said that high glucose levels don’t raise HbA1C levels if the glucose spike is short. So you can have the same HbA1C where glucose is always in range and where there are short term glucose spikes with hypers and hypos. HbA1C does give a good idea of where you average sugar levels are, but life expectancy is intimately connected with the number of spikes and hypos which causes arteriosclerosis.

If diabetes life expectancy for a given A1C is inversely proportional to the amount of medication needed to achieve this, then perhaps people should take more notice of “the law of small numbers” as advocated by Dr Bernstein. You don’t hear this mentioned much on the forums.
 
Last edited:

OdinsBeard

Member
Messages
21
Type of diabetes
Type 1
Treatment type
Insulin
I saw an article the other day where it said that high glucose levels don’t raise HbA1C levels if the glucose spike is short. So you can have the same HbA1C where glucose is always in range and where there are short term glucose spikes with hypers and hypos. HbA1C does give a good idea of where you average sugar levels are, but life expectancy is intimately connected with the number of spikes and hypos which causes arteriosclerosis.

If diabetes life expectancy for a given A1C is proportional to the amount of medication needed to achieve this, then perhaps people should take more notice of “the law of small numbers” as advocated by Dr Bernstein. You don’t hear this mentioned much on the forums.
Does arteriosclerosis cause a low heart rate? I used to run around 75 bpm when I was fit, now I am resting at around 52 bpm, could this be linked?