Just a thought/question.

According to the old logic HbA1c reaches a steady state in 4 to 12 weeks depending on if you are a high or low glycator.

However, as an excellent post by reidpj says, these timescales were derived from the mistaken assumption that the glycation reaction was irreversible:

Glucose binds slowly to hemoglobin A, forming the A1c subtype. The reverse reaction, or decomposition, proceeds relatively slowly, so any buildup persists for roughly 4 weeks. Because of the reverse reaction, the actual HbA1c level is strongly weighted toward the present. Some of the HbA1c is also removed when erythrocytes (red blood cells) are recycled after their normal lifetime of about 90-120 days. These factors combine so that the HbA1c level represents the average bGlevel of approximately the past 4 weeks, strongly weighted toward the most recent 2 weeks. It is almost entirely insensitive to bG levels more than 4 weeks previous.

Click to expand...

viewtopic.php?f=1&t=25457&p=233638#p233636

So according to this, it means that your HbA1c is only good for the past 2-4 weeks (I recommend you read the whole post).

He also makes the point that HbA1c can vary by as much as 1% between individuals with identical average BG scores, such that comparing your own HbA1c with other diabetics isn't entirely meaningful. However, he notes that:

Although HbA1c varies among individuals with the same average bG, it is very stable for any given individual. Thus a change of 1.0% in your own HbA1c is definitely meaningful.

Click to expand...

I also really like this review of the science behind HbA1c, its history and its practical use:
http://professional.diabetes.org/Content/Editorial.pdf
It's not too technical and very readable.

From my own experience, all my HbA1cs have been significanty lower than my BG readings would suggest. To be honest, I think that the correlation is a bit dodgy for very low or very high HbA1cs. For example a HbA1c of 4.5% is supposed to equate to an average BG of 4.6mmol/l, when I usually measure in the 5-6mmol/l range (although it could be that the meter readings are inaccurate too).

Grazer said:

AVG BG = (HbA1c x 1.98) - 4.29

Click to expand...

Yes. A 0.5mmol/l change in your average BG should result in a 1% change in HbA1c.

It doesn't make any sense to directly equate BG readings (in mmol/l) to HbA1c readings (mmol/mmol). The fact that the numbers are quite close in the 4-6% range is just a numerical coincidence.

jopar said:

Now because of the frequency of testing I do, I know it's unlikely that my BG is spiking scarily high, due to very limited times I go below 4mmol/l so I'm not having hypo's masking high's etc..

Now it's my opinion that the HbA1c's can tell more and is more reflective towards T2's who are either following diet controlled or non-hypo inducing medical such as insulin.... Than it can for a T1, who's hypo's masks the high's or T2's who are taking hypo inducing medication.

Click to expand...

I think that you're misunderstanding what HbA1c actually is. Just because there is a correlation betwee HbA1c and average BG, it doesn't mean that the two things are equivalent. Glycation occurs when your BG is high, your blood doesn't become "unglycated" when you hypo. As I understand it, HbA1c is only an average in the sense that it accumulates every time you spike, it doesn't go back down again when you "hypo". For this reason: hypos do not mask spikes in your HbA1c reading.

It's a big like dust on your mantlepiece. It accumulates when there is dust in the air, and doesn't "disaccumulate" when the air is clean.

(This is how I got an HbA1c of 4.5%, I never spike (or I spike even less than a non-diabetic) so my HbA1c never increases).

Hi there I am know much less than you guys, but I have wondered about my HB1AC, it has been 5.7% the last twice and has baffled the docs and the consultant at the diabetic clinic.They are thinking LADA or MODY with me so I dont know.When it all started I had massive spikes and massive lows, range of 2.2 to 16.8.The last few weeks I have been checking my bs about 10 - 12 times to get an average picture as its staying high all the time and hypos are more or less gone(which I am glad of )My readings yday were-:
7.2 fasting
11.3 after breakfast
11.3
12.2
13.6
15.4
14.8
11.4
11.3
11.3
9.3 before bed
It has been like this for a few weeks surely my HB1AC would have gone up, its 2 months since my last x

Estimated average glucose Calculator
The calculator on this website works to an older algorithm. The average glucose results are higher than those derived from the IFCC validated formula . ie it's out of date so I'd rather use the calculator on the ADA pro site
eAG (mmol/L) = 1.59 x HbA1c – 2.59
This site says that an HbA1c of 7% reflects an average of 9.6mmol/l
The ADA calculates 7% to be an eAG of 8.6mmol/l
http://professional.diabetes.org/GlucoseCalculator.aspx

It is an estimated average, in the US they are supposed to be starting to give the estimated glucose levels as part of the lab test result. In the UK and elsewhere they have decided not to. The reason is 2 people with the same actual average readings can have widely different HbA1cs. ( some people may be higher or lower glycators, some people may have conditions such as anaemia which can affect results. )

So person B is "hiding" a dangerous spike to 11.0 each day for 120 days while appearing to be as good as person A...

To me this is the primary reason why self testing is just as important as a quarterly HbA1c. The HbA1c would not pick up Person B's dangerous eating pattern

Click to expand...

HbA1c may be affected by variability
Some recent research has shown that variability (ie wide ups and downs) does affect the HbA1c in T1 . People who had more variability ie wider variations , had higher HbA1cs than people with similar average but lower variability. Variability caused more of a difference at higher HbA1c levels than at lower ones. They didn't find the same tendency in T2 but also say they might not have had sufficient T2s with wide variations in the trial to show such a difference.
http://www.ncbi.nlm.nih.gov/pubmed/21700921

Dangerous spikes? perhaps, the juries still out
Whether someone with more variability (ie someone with a more spiky glucose profile) has more risk than someone with the same HbA1c and a flatter profile is very much a matter of hot debate. Many researchers think that the spikes must cause more damage;.. cells are certainly damaged by glycation in a test tube and some of these changes can be demonstrated when they analyse blood after a high spiking meal.
However, at the moment there is very little evidence to show it makes a long term difference. The limited long term evidence we have( ie the DCCT) suggests that it doesn't matter how the HbA1c was achieved.(ie spiky or flat), or whether someone was a high or low glycator. It was the HbA1c that was the risk determinant for microvascular complications and the average glucose levels for macrovascular complications (google glucose variability Kilpatrick or Monnier for 2 major players in the debate)
What did make a difference was variations between HbA1cs ie changing between periods of good control followed by periods of poor control . Rapid lowering of HbA1c was also found to cause rapid progression of retinopathy (indeed the DCCT was almost stopped because of this; eventually eyes stabilised and the trial continued)
.

I have always used HbA1c as a tool to confirm the trend of my meter readings. I have in excess of 4 years readings now (nearly 20,000) and my own personal calculation is normally within 0.2% of the actual.

Stephen your point about spikes is interesting.. because I do get lower HbA1c's when I lowcarb.. even if my average blood glucose is not vastly different.. for instance my daily average blood glucose is 99% of the time somewhere between 5.0 mmol/l - 5.5 mmol/l and from this I get HbA1c's back between 4.8% - 5.4%..

Pneu said:

I have always used HbA1c as a tool to confirm the trend of my meter readings. I have in excess of 4 years readings now (nearly 20,000) and my own personal calculation is normally within 0.2% of the actual...

Click to expand...

20,000 :shock: You must have finger tips like pin cushions...

Just shows that if you test 14 times a day, most of the error disappears. I'd guess some of the remaining difference is due to unrecorded night-time variations.

Pneu said:

for instance my daily average blood glucose is 99% of the time somewhere between 5.0 mmol/l - 5.5 mmol/l and from this I get HbA1c's back between 4.8% - 5.4%..

Click to expand...

That is an amazing achievement for a T1. I can see why it took 20000 meter readings to achieve!

Would be interested in what variation of grams of carbs / day that spread of HbA1c%'s covers. What are you thinking is low carb compared to "normal" carb in your regime?

phoenix said:

Estimated average glucose Calculator
The calculator on this website works to an older algorithm. The average glucose results are higher than those derived from the IFCC validated formula . ie it's out of date so I'd rather use the calculator on the ADA pro site
eAG (mmol/L) = 1.59 x HbA1c – 2.59
This site says that an HbA1c of 7% reflects an average of 9.6mmol/l
The ADA calculates 7% to be an eAG of 8.6mmol/l
http://professional.diabetes.org/GlucoseCalculator.aspx

<snip>

Click to expand...

Very useful link - I hadn't realised that this had changed!
My last HbA1c was 7.0.

According to this site, this gives an average of 9.6 mmol/l
According to the one in your link it is an average of 8.6 mmol/l
[Couldn't even find a calculator on the diabetes.org.uk site and Google is not my friend]

I am far more comfortable with an average of 8.6 mmol/l because that ties in with my morning (so probably overnight) readings.
I was concerned with an average of 9.6 mmol/l because I just wasn't seeing readings in this range when testing, and an AVERAGE suggests that a goodly portion of the readings can be significantly above this.

Has anyone PM'd the moderators to suggest that the algorith should be updated?

Cheers

LGC

LittleGreyCat said:

Has anyone PM'd the moderators to suggest that the algorith should be updated?

Click to expand...

If it's generally accepted that this is more accurate then I think we should. But does the UK and other countries accept the new formula? Although that doesn't neccessarily mean we shouldn't go to it!

Phoenix, it is also interesting that the study says there maybe a difference between T1 and T2 in all of this. In the underlying study it would appear the T2's included where non insulin using T2's. This is quoted as part of the conclusions

In general, GV is higher in patients with poor glycemic control and in type 1 diabetic patients compared with type 2 diabetic patients, which can be attributed to insulin therapy and higher insulin sensitivity. High GV may affect glycation because of periodic exposure of the erythrocyte to high glucose levels and therefore to faster irreversible glycation.

Click to expand...

So in their sample it looks like the T2's where generally less spikey than the T1's. That seems a pretty reasonable conclusion as intuitively spike control in T1 is a harder thing to achieve compared to say the average diet or diet + meds T2 who has just learnt what is going to spike them and therefore avoided consuming stuff that would spike to start with.

I also found the earlier post I mentioned you made on the subject which contained a number of interesting links...

Phoenix said:

Intuitively I would agree that damage is being caused by those spikes but the matter is hotly debated by researchers.

The role of glycemic variability in the
pathogenesis of vascular complications
is still not clear from the available data

On the one hand there is no data to suggest that it matters a jot whether the HbA1c was gained through a flat line or a bumpy sequence of peaks and troughs. Post hoc analysis of available data from the DCCT suggests that it doesn't matter.(In T1)
On the other hand there are studies that show oxidative stress is highest when glucose levels peaks... and this is thought to be a culprit.
If you want to research this further here are a few starters but be warned I once started with the idea of writing a blog post about it and 3 years later still haven't attempted to write it .
Glycemic variability: a debate still not settled
http://www.internationaldiabetesmonitor ... Fs/836.pdf
Glucose variability debate, halfway through paper.
http://www.diabetes.procampus.net/cours ... _Nov07.pdf
Postprandial Hyperglycemia and Glycemic Variability Should we care?
(this last looks at recent T2 evidence)
http://care.diabetesjournals.org/conten ... 0.full.pdf

Personally as someone who can have occasional very high (and very low) readings... I have a a horrible 'bug' at the moment which makes control a bit unpredictable . I take comfort in the fact the long term evidence in T1 (so far) seems to show that it's the overall 'average' that matters and not the way that HbA1c is achieved.

Click to expand...

xyzzy said:

Pneu said:

for instance my daily average blood glucose is 99% of the time somewhere between 5.0 mmol/l - 5.5 mmol/l and from this I get HbA1c's back between 4.8% - 5.4%..

Click to expand...

That is an amazing achievement for a T1. I can see why it took 20000 meter readings to achieve!

Would be interested in what variation of grams of carbs / day that spread of HbA1c%'s covers. What are you thinking is low carb compared to "normal" carb in your regime?

Click to expand...

'Low Carb' would be 60g - 80g carbohydrate per day.. I would typically sit somewhere around 100 - 120g carbohydrate..

The lower HbA1c's generally correlate with sustained periods of 'low carb'.. whereas for the last 6 months I tend to eat a mixture of low and higher carbohydrate days depending on how busy I am.. if I am doing a lot of travelling or staying away from home then low carb is pretty difficult to maintain.

Grazer said:

LittleGreyCat said:

Has anyone PM'd the moderators to suggest that the algorith should be updated?

Click to expand...

If it's generally accepted that this is more accurate then I think we should. But does the UK and other countries accept the new formula? Although that doesn't neccessarily mean we shouldn't go to it!

Click to expand...

If you are going to use an equation it's better to use one that has some validation.
It isn't really supposed to be used to predict an HbA1c, it's supposed to be used the other way round to show what type of average glucose may have resulted in an HbA1c.
In the UK Kilpatrick (again,he writes a lot of papers ) and Leslie have argued

• that the sample was not fully representative;
  that it hasn't been validated on a second phase group (though I have to say they used samples from several study centres worldwide)
  that various disorders can cause lower or higher HbA1cs
  that there is great individuality so that one person with an average glucose level of 170mg/dl ( 9.4mmol/l)could have an HbA1c varying from 6.5%-9%

http://care.diabetesjournals.org/content/32/1/e11.full
These are all very good reasons not to give a person an estimated average glucose reading in place or even in addition to an HbA1c result. That's why the UK , Australia and Sweden have decided not to use these figures at the present time.
The link below summarises the research that led to the formla (joint funded by the American and European diabetes asscociations) and discusses the why countries will or won't use it.
http://www.aacc.org/publications/cln/20 ... _1008.aspx

The reasons for not using the estimated average glucose formula are good reasons not to use any formula or converter, especially in an to attempt a prediction. If you are going to use one, its surely better to use one with some research to back up the calculations. I've no idea where the old algorithm actually came from.

It looks though that not everyone in the UK is against giving patients an estimated average glucose.
http://www.pathology.stgeorges.nhs.uk/p ... ----(hba1c)

Defren posted ;

. . . my question was about counting our diagnostic test as moot, as we would not have know we were diabetic. If we want to have a benchmark for our testing and ultimately our HbA1c we should used the ones following changing our diet, where we can then have a clearer idea of what we are looking at, post lifestyle changes.

Click to expand...

I agree, Defren. Our diagnostic test was a record of an "abnormal" HbA1c, as it led to a diagnosis of Type II diabetes. If you've lowered it by the time of the next one (I was 6.5 in the April, at diagnosis, 5.6 in the September), that is a better benchmark or baseline for all following tests.

The following September, 12 months later, was 5.2 for me, showing I was well on track to good control. The following December it was 6.0 - this showed my BG had been running higher. In my case it was not dietary, but due to a chest infection (treated by two 7-day courses of antibiotics) followed by a tooth abcess (treated by 2 7-day courses of 2 different antibiotics at the same time) until the offending tooth was removed. So the "abnormal" HbA1c of 6.0 showed not a loss of diabetic control, but general 'bad' health in the preceding 4 weeks. (Tooth was removed 2 days before blood was taken.)

I wasn't doing a lot of testing at the time, but interestingly my daily fasting tests were never above 6 - 5s were 'normal' for me then - so there was no real sign of the higher blood glucose in those tests, though I was expecting a higher HbA1c because of the infection.

Any thoughts on that, anyone?

Viv 8)

My ex boss was type 2 and after each blood test would have his standard lunch of 1 large packet cheese & onion crisps and a snickers bar. About 3 months before his next test he would "behave" and eat sensibly so that the nurse didn't nag him!