2 Hba1cs, 2 Labs, 1 Blood Sample.

[Bluetit1802]

My last HbA1c in June came back 10 points higher than other data suggests.

Personal data from rigorous and frequent finger prick average over 3 months was 5.62. Converted to HbA1c measurement units = HbA1c 33.

My Sugr estimate for 4 weeks prior to test = HbA1c 32

Home HbA1c test (A1cNow) few days before test = HbA1c 33

Last Libre sensor for 2 weeks up to 26th May = HbA1c 31.

My actual lab HbA1c was 43.

This has been happening at every test for about 4 years, causing me to do lots of research on the fallibility of the HbA1c in many people. On this last occasion I’d had enough so queried it with my GP.

She spoke to the clinical biochemist at the East Lancashire Trust lab where my blood samples are normally analysed. The biochemist explained the method they use is high performance Liquid Chromatography. However, they sometimes send samples to Salford Royal Hospital for the HbA1c to be measured by a different method called Affinity Chromatography if there may be anything in the blood sample that may interfere with the result. (eg. Haemoglobin variances).

I was asked to give a new sample of blood, which I did. This was sent to the normal East Lancs lab who analysed it then forwarded the same sample to Salford for their analysis.

Results.
East Lancs HbA1c = 44
Salford HbA1c = 40

My GP has now spoken with the Consultant Clinical Biochemist at our normal lab who has recommended that all my future HbA1c test samples should be sent first to them for analysis and then re-directed to Salford for a second analysis.
This does not wholly explain the 10 point difference between my own data and the actual tests, but it goes some way to show that my haemoglobin does have some variances that make the HbA1c unreliable.

 

[Guzzler]

Not an exact science, it would seem. The difference of four points alone in that range means a lot to us as individuals if we are to strive for the sometimes elusive non D numbers. But the difference of ten points could in theory lead to misdiagnosis if levels are anywhere approaching the lower end of T2.

 

[britishpub]

This does not wholly explain the 10 point difference between my own data and the actual tests, but it goes some way to show that my haemoglobin does have some variances that make the HbA1c unreliable.

It would be interesting to find out whether your are unusual in this, or if it is a common occurence.

 

[Bluetit1802]

It would be interesting to find out whether your are unusual in this, or if it is a common occurence.

It would be a start to find out which method one's usual lab uses. It would also be useful to know how many different methods there are.

 

[Goonergal]

Really interesting @Bluetit1802 but very frustrating for you, I imagine.

 

[Mr_Pot]

It would be a start to find out which method one's usual lab uses. It would also be useful to know how many different methods there are.

That would be interesting but it wouldn't help with understanding the difference between the lab readings and your own. What ever method the lab uses and however accurate their separation of the A1c component they are still measuring glycated hemoglobin and not the average glucose level in your blood. If, as it would seem in your case, the two do not have the usual relationship due to your red blood cell lifetime or whatever, then the HbA1c test will not work for you however accurate it is.
Most worrying would be the case where HbA1c indicates that average BG levels are lower than they really are, which could lead to diabetes not being diagnosed or not properly treated.
In my case I don't measure my BG anywhere near enough times, or at regular enough intervals, to say what my average is but when I do measure it before meals it is about 6.0 and my HbA1c is 38 so at least it's in the right ballpark. I would suggest others who may want to question the accuracy of their HbA1c note that @Bluetit1802 has been careful to calculate her average accurately, averaging a few random finger prick tests is not going to work.

 

[Bluetit1802]

That would be interesting but it wouldn't help with understanding the difference between the lab readings and your own. What ever method the lab uses and however accurate their separation of the A1c component they are still measuring glycated hemoglobin and not the average glucose level in your blood. If, as it would seem in your case, the two do not have the usual relationship due to your red blood cell lifetime or whatever, then the HbA1c test will not work for you however accurate it is.
Most worrying would be the case where HbA1c indicates that average BG levels are lower than they really are, which could lead to diabetes not being diagnosed or not properly treated.
In my case I don't measure my BG anywhere near enough times, or at regular enough intervals, to say what my average is but when I do measure it before meals it is about 6.0 and my HbA1c is 38 so at least it's in the right ballpark. I would suggest others who may want to question the accuracy of their HbA1c note that @Bluetit1802 has been careful to calculate her average accurately, averaging a few random finger prick tests is not going to work.

Agreed Absolutely.

It is very frustrating for me. I know, with hand on heart, that despite finger pricking many times a day and catching my peaks as often as I can, my averages are under 6mmol/l week on week. I use a Libre on a part time basis, which shows me what is happening over night and between finger prick tests. I did my own home HbA1c test using A1cNow (which research has shown is as accurate as a lab test using venous blood) and which several other folk on this forum have used and found it to tally with their normal HbA1c. My HbA1c was 10 points higher more or less.

This double testing the same sample at 2 labs using different methods has at least shown me how unreliable the whole process is. There may only have been 4 points difference, but those 4 points change my mmol/l average from 6.6mmol/l to 7.3mmol/ and that is a huge difference to my way of thinking, considering it is an average (of sorts).

Anyways, I just wanted to draw this to the attention of other forum users.

By the way, I am not anaemic in any way, and do not have any thyroid issues, nor am I on any medication for anything.

 

[Pipp]

It is of interest to me @Bluetit1802, how you managed to persuade the GP to contact the labs for all that info. I think mine would not have a clue because once the blood form has been handed over, so it seems is their responsibility. Well done for being so persistent.

Also interesting how blood from the same sample showed different results.

 

[Bluetit1802]

It is of interest to me @Bluetit1802, how you managed to persuade the GP to contact the labs for all that info. I think mine would not have a clue because once the blood form has been handed over, so it seems is their responsibility. Well done for being so persistent.

Also interesting how blood from the same sample showed different results.

I sparked my GPs interest during a phone conversation after my last lot of blood tests when she wanted to stress that my cholesterol had gone up! I changed the subject to my HbA1c pretty quickly. I told her about my own data etc and how disappointed I was that my HbA1c doesn't match it. I asked about a fructosamine test, which she had never heard of so I abandoned that one and went on about haemoglobin variances, which she also knew nothing about, but she was concerned about the huge difference between my HbA1c's and my own records. That's when she decided to contact the biochemist for her opinion. It was the biochemist that suggested they could send my sample to Salford, not my GP.

 

[Guzzler]

I sparked my GPs interest during a phone conversation after my last lot of blood tests when she wanted to stress that my cholesterol had gone up! I changed the subject to my HbA1c pretty quickly. I told her about my own data etc and how disappointed I was that my HbA1c doesn't match it. I asked about a fructosamine test, which she had never heard of so I abandoned that one and went on about haemoglobin variances, which she also knew nothing about, but she was concerned about the huge difference between my HbA1c's and my own records. That's when she decided to contact the biochemist for her opinion. It was the biochemist that suggested they could send my sample to Salford, not my GP.

Great dodging skills there, well done!

 

[Mbaker]

My last HbA1c in June came back 10 points higher than other data suggests.

Personal data from rigorous and frequent finger prick average over 3 months was 5.62. Converted to HbA1c measurement units = HbA1c 33.

My Sugr estimate for 4 weeks prior to test = HbA1c 32

Home HbA1c test (A1cNow) few days before test = HbA1c 33

Last Libre sensor for 2 weeks up to 26th May = HbA1c 31.

My actual lab HbA1c was 43.

This has been happening at every test for about 4 years, causing me to do lots of research on the fallibility of the HbA1c in many people. On this last occasion I’d had enough so queried it with my GP.

She spoke to the clinical biochemist at the East Lancashire Trust lab where my blood samples are normally analysed. The biochemist explained the method they use is high performance Liquid Chromatography. However, they sometimes send samples to Salford Royal Hospital for the HbA1c to be measured by a different method called Affinity Chromatography if there may be anything in the blood sample that may interfere with the result. (eg. Haemoglobin variances).

I was asked to give a new sample of blood, which I did. This was sent to the normal East Lancs lab who analysed it then forwarded the same sample to Salford for their analysis.

Results.
East Lancs HbA1c = 44
Salford HbA1c = 40

My GP has now spoken with the Consultant Clinical Biochemist at our normal lab who has recommended that all my future HbA1c test samples should be sent first to them for analysis and then re-directed to Salford for a second analysis.
This does not wholly explain the 10 point difference between my own data and the actual tests, but it goes some way to show that my haemoglobin does have some variances that make the HbA1c unreliable.

@Guzzler thanks for directing to this post and @Bluetit1802 your experience is very similar to mine; we are both experienced at managing Type 2 and know through rigorous testing that something is up with our official tests, you of course proving your case. I would also like to get a cross referenced test to see what on earth is going on, I know I am much better controlled now than ever. So in summary:

In April 2017 my HbA1c was officially measured at 36, below are the numbers for Freestyle Libre at this time (weight 87.2 kg):

On 28th June 2018 my HbA1c was 38, below are the numbers for Freestyle Libre (weight 85 kg)

So a drop of 16% in Libre readings (and 2 kg weight loss, better low carb|) equates to a 2% increase in HbA1c? I don't think so, something is fishy.

On 30th June 2018 A1CNow+ was 31 (my non-diabetic wife on the same unit 33, two other in-laws 34 and 38). How is there an 18% difference between my official HbA1c and A1CNow+ results. I have tested around 8 further times between March and July 2018, with an average of 31.5 (latest on 30th July 2018 was 33 and 30):

I have used the following meters over the last year (some times 3 at a time to ensure accuracy):
Select One Touch Plus, Accu-Chek Mobile, Freestyle Optium Neo and GlucoZen.auto

Two Meter Averages (not enough data for the other 2 devices, but on individual reasons, always low 4's or less FBG, and expected up to mid 5's post prandial):

Select One Touch Plus 90 day average combined is 4.3 mmol (4.4 before meal, 4.6 after meal)
Accu-Chek Mobile 90 day average combined 4.3 mmol (4.3 before meal, 4.8 after meal)

I know my measures have been improved over the last year and a half as verified on 4 separate meters, the Freestyle Libre (used to see if spiking during the night - which I am not) and the A1CNow+. In my view it is not possible to have so many devices, 6 be wrong, I also have lost weight, have body fat around 10% and visceral around 6%, I know I have been better at low carb, especially at the last A1c, I had been OMAD for over 60% of the last 6 months.

I will be seeking answers from my surgery and asking them to verify the machine which tests A1c's; and for the sake of other diabetics (and myself) have results tested elsewhere on another competitive unit.

Well done @Bluetit1802 for your follow through. I hope to also get answers.

 

[Bluetit1802]

@Guzzler thanks for directing to this post and @Bluetit1802 your experience is very similar to mine; we are both experienced at managing Type 2 and know through rigorous testing that something is up with our official tests, you of course proving your case. I would also like to get a cross referenced test to see what on earth is going on, I know I am much better controlled now than ever. So in summary:

In April 2017 my HbA1c was officially measured at 36, below are the numbers for Freestyle Libre at this time (weight 87.2 kg):

View attachment 27828

On 28th June 2018 my HbA1c was 38, below are the numbers for Freestyle Libre (weight 85 kg)

View attachment 27829

So a drop of 16% in Libre readings (and 2 kg weight loss, better low carb|) equates to a 2% increase in HbA1c? I don't think so, something is fishy.

On 30th June 2018 A1CNow+ was 31 (my non-diabetic wife on the same unit 33, two other in-laws 34 and 38). How is there an 18% difference between my official HbA1c and A1CNow+ results. I have tested around 8 further times between March and July 2018, with an average of 31.5 (latest on 30th July 2018 was 33 and 30):

View attachment 27830

I have used the following meters over the last year (some times 3 at a time to ensure accuracy):
Select One Touch Plus, Accu-Chek Mobile, Freestyle Optium Neo and GlucoZen.auto

Two Meter Averages (not enough data for the other 2 devices, but on individual reasons, always low 4's or less FBG, and expected up to mid 5's post prandial):

Select One Touch Plus 90 day average combined is 4.3 mmol (4.4 before meal, 4.6 after meal)
Accu-Chek Mobile 90 day average combined 4.3 mmol (4.3 before meal, 4.8 after meal)

I know my measures have been improved over the last year and a half as verified on 4 separate meters, the Freestyle Libre (used to see if spiking during the night - which I am not) and the A1CNow+. In my view it is not possible to have so many devices, 6 be wrong, I also have lost weight, have body fat around 10% and visceral around 6%, I know I have been better at low carb, especially at the last A1c, I had been OMAD for over 60% of the last 6 months.

I will be seeking answers from my surgery and asking them to verify the machine which tests A1c's; and for the sake of other diabetics (and myself) have results tested elsewhere on another competitive unit.

Well done @Bluetit1802 for your follow through. I hope to also get answers.

Please let me know what your surgery says. I am very interested. My GP knew nothing about the subject but did believe my own data and realised something is amiss. Thats why she contacted the biochemist at the lab. The biochemist at my surgery's usual lab has recommended all my future HbA1cs are tested twice from same blood daw, once at the normal lab and once at the other lab. The other lab uses a specialist measuring method that is intended to take account somehow of Hb variances, which are generally genetic, but do not take account of people whose haemoglobin cells live longer than the standard 120 days. I think I fall into this latter group, but must also have some variances to produce an HbA1c 4 points lower. I was pleased about that, but still 6 points too high according to other data.

 

[Bluetit1802]

I will be seeking answers from my surgery and asking them to verify the machine which tests A1c's; and for the sake of other diabetics (and myself) have results tested elsewhere on another competitive unit.

I know a long time has passed since this thread. Did you take this up with your surgery, and what was said?

 

[ringi]

A HbA1c's lower then expected could be due to internal bleeding.

 

[Alexandra100]

I would suggest others who may want to question the accuracy of their HbA1c note that @Bluetit1802 has been careful to calculate her average accurately, averaging a few random finger prick tests is not going to work.

Agreed, and unless one is using a CGM of some sort, one could be running high bgs during 8 hours of sleep without knowing.

 

[Alexandra100]

Please let me know what your surgery says. I am very interested. My GP knew nothing about the subject but did believe my own data and realised something is amiss. Thats why she contacted the biochemist at the lab. The biochemist at my surgery's usual lab has recommended all my future HbA1cs are tested twice from same blood daw, once at the normal lab and once at the other lab. The other lab uses a specialist measuring method that is intended to take account somehow of Hb variances, which are generally genetic, but do not take account of people whose haemoglobin cells live longer than the standard 120 days. I think I fall into this latter group, but must also have some variances to produce an HbA1c 4 points lower. I was pleased about that, but still 6 points too high according to other data.

I believe you use a Libre at least some of the time, and I know you test assiduously around meals, so in your case I think I would give up worrying about A1c results. Surely your home testing is far more revealing than the A1c number, which at best is only an average and could theoretically be concealing huge harmful spikes?

 

[Bluetit1802]

I believe you use a Libre at least some of the time, and I know you test assiduously around meals, so in your case I think I would give up worrying about A1c results. Surely your home testing is far more revealing than the A1c number, which at best is only an average and could theoretically be concealing huge harmful spikes?

I gave up worrying about my HbA1c a few years ago! I do know it is skewed and that my other data show this. However, it would be nice to have a genuine one showing on my medical records. I couldn't agree more about the HbA1c being liable to conceal any roller coaster swings, which are very damaging. The sooner a better test is invented, the better it will be.

 

[Antechinus]

In a past life I was a laboratory manager in a soil testing lab. Every couple of years Australian Standars sent out test samples to a whole bunch of labs to check for inter laboratory accuracy. Well I can tell you it was not unusual for results to be out by a factors of 2 and higher. There was huge variance, On some tests it was possible to get concensus on what was the correct result. So I'm not surprised that you got different HbA1c from 2 different labs using 2 different procedures. Its just the nature of analysis. Actually, getting 40 and 44 is pretty good. I think the important thing is to watch trends in the results. There is a lot of fuzziness in analytical testing.

 

[Mbaker]

I know a long time has passed since this thread. Did you take this up with your surgery, and what was said?

Hi @Bluetit1802, in general I have a good surgery who work by standard methods. My surgeries main Doctor kindly gave me the time to explain in the nicest terms that I am not a special case. This was meant in the best possible sense in so far as the Doctors having a significant workload with persons who are sicker than me - this I totally understood. However as the maths do not add up and I am a Computer guy of 30 years plus, this is hard for me to set aside.

I have found out that the endo who cancel an appointment with me was because my diabetes status is in remission so again I would be a waste of time. My latest December HbA1c was the same as in the summer at 38, which at my diabetic yearly assessment last week could not be explained when 2 years earlier it was 35. When I said how is this even technically possible considering fasting and post prandial numbers were over a full mmol/l higher 2 years ago, no explanation could be provided. I had a random blood glucose taken at 11.30 at the surgery with their synchronised professional unit, this read 4.7 and my Accu Chek on the same blood sample 4.8 (4.5 on waking which is high for me), so again proving my home testing has been reasonably accurate i.e.no where near a 6.5 mmol/l; even at both units being 10% wrong this would still be wildly inaccurate.

I think my surgery will not go any further. They were happy with my Cholesterol / HDL ratio of 2.52 (trigs 0.28)

I had a private test with Medichecks which returned a random blood glucose of 5.00 mmol/l (4.1 on waking on my Accu-chek) and a HbA1c of 39. They have agreed to test on an alternative machine for free. My other tests were:

FRUCTOSAMINE: 268 umol/L
Ranges:<285
Satisfactory stabilisation 280 - 320
Poor stabilisation 320 - 370
Insufficient stabilisation > 370

Looking at the conversion charts of Fructosamine to HbA1c these don't correlate, but 268 seems to be an around 6.1 to 6.5% mmol/l average, again I never see 6's on my in sync Accu Chek.

Fasting Insulin: 5.2
Ranges : circa 2.6 - 24-9

Fairly happy with this. I did the HOMA IR score which should be below 1
Formula: Fasting insulin x Fasting glucose (mg/dl) / 405
5.2 x 75.6 / 405 = 0.97 So I may have broken insulin resistance or be on the verge of it.

HS CRP: < 0.3
Range: 0 - 5.0
Quite pleased with this low inflammation marker

IGF1: 31.3
Range 6.5 - 31
Happy with this as my regime promotes muscle building, so as long as I keep my inflammation low I am not concerned about
promoting cancer growth faster.

A consultant Doctor at Medichecks wrote the following:

"Dear Moyle,

Many thanks for your enquiry which has been elevated to me. In answer to your first query the laboratory has informed me that Fructosamine is analysed on a Roche Cobas 8000 analyser, on a C702 module using spectrophotometry.

I understand that you are concerned about your HbA1c level as this is higher than expected given your glucose readings. There are some potential causes of a higher than expected HbA1c reading. You have great knowledge about your diabetes so I apologise if I go over ground that you are already familiar with. Glycosylated HbA1c is formed by glucose binding irreversibly to haemoglobin, so the HbA1C level can be affected by alterations in glucose level, by factors that increase or decrease the lifespan of red blood cells (and therefore the persistence of HbA1c) and by alterations in the structure of haemoglobin.

Anaemia due to either iron deficiency, or to pernicious anaemia can lead to decreased breakdown of red blood cells, which can lead to a higher than expected HbA1c level. Kidney problems can sometimes increase HbA1c levels. Some haemoglobinopathies can also lead to increased HbA1c levels, although they will typically lead to a decrease, this includes issues such as sickle cell disease and less well known issues such as Haemoglobin F and S.

The other factor to bear in mind is that HbA1c exhibits a surprising amount of variation between laboratories and laboratory analysis platforms, this is despite attempts to standardise its measurement on an international basis. I've attached a recent paper which may interest you which outlines the not inconsiderable variation in HbA1c measurement between laboratories.

I suggest that in the first instance we recheck your HbA1c level at the other main laboratory that we use. Our customer services team will be in contact with you to arrange this on a complementary basis."

The paper referred to is attached. So where I go from here is to keep doing my n of 1 testing, which is Keto, heavy weights and the like.

Summary
My surgery are kind, but do not have the resources to deal with my specifics (I understand this). My home testing kit has been verified (again) to be inline with professional kit, so my fbg of 3.6 mmol/l this morning can be believed as can my 90 day average of 4.2 mmol/l (147 tests). My other markers suggest a low inflammation within my body, with proxy markers suggesting good pattern A LDL. Some further potential explanations of the differences between my expectation and actuals have been suggested. Overall I am happy due to the inflammation markers and a QRISK2 of 2.1 at the end of 2018, compared to 2.0 in 2012 when I was non-diabetic.

 

[britishpub]

A HbA1c's lower then expected could be due to internal bleeding.

Thanks for that