• 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.
  • Take a look at our new Type 2 Diabetes Remission section on the Diabetes UK website: https://www.diabetes.org.uk/diabetes-the-basics/type-2-remission
  • 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

Blood Sugar - Fingerstick vs HbA1c

S

saz9961

Member
Relationship to Diabetes
Type 2
Pronouns
He/Him
I was diagnosed T2d in October 24 with a HbA1c of 88mmol/mol. I since measured my blood sugar twice a day (last thing at night, first thing in the morning, and recording, eventually, a 90d rolling average. The conversion of capillary blood glucose from a 90d rolling average was about 10% off my most revent NHS measure (last week) of 37mmol/mol. I think if I only recorded fasting measures, the predicted value would have been too optomistic.
 
@saz9961 - I do not understand your post. Are you measuring your BG by finger pricking? If so, that measures blood glucose in mmol/l, not mmol/mol and cannot be simply related to an HbA1c result which is measured in mmol/mol.

Either way up, well done on the reduction in HbA1c - whatever you are doing has been very effective!
 
Docb said:
@saz9961 - I do not understand your post. Are you measuring your BG by finger pricking? If so, that measures blood glucose in mmol/l, not mmol/mol and cannot be simply related to an HbA1c result which is measured in mmol/mol.

Either way up, well done on the reduction in HbA1c - whatever you are doing has been very effective!
Click to expand...

Yes, daily in mmol/l, then apply a formula, =(((2.59+x)/1.59)*10.9)-23.5, where x is the daily measure, to get an approximate conversion. I just dump that into a spreadsheet, and have daily, 30d, 60d and 90d calculation.

HbA1c has no physiological significance, its just a value associated with an observation of diabetes/no diabetes, and is a proxy for blood sugar. Capillary blood level will vary from venous blood level, and there is a normal flux. Capillary drawn blood can also be contaminated with interstitial fluid.

There are a few variations in the conversion equation, all can be rationalised. Blood sugar is a measure of the number of moles of glucose per volume of blood. In actuality, the tables used by a typical glucose meter are very shakey, but good enough. Technically, you ought to be running a standard curve each time plus controls. But you don't, because you don't need to, because of the purpose of the test.

HbA1c is a measure of glucose bound to haem protein. Glycation is non-enzymatic, so there is no regulation by the body. The body doesn't start increasing haemoglobin is response to increased glucose. The bound glucose doesn't serve as some sort of emergency store. There is no function to it. I see this in lots of other proteins. I did a lot of research back in the day on ricin protein, part of which is quite sticky for lactose. The reason might to go with how these molecule bind to cell membranes, where there are a few sugar molecule moeties.

HbA1c is mmoles of glucose per mole of haemoglobin, which is typically quoted as having a molecular weight of 64,500 Daltons. The conversion has to take account of a few assumptions; how much haemoglobin an average person has per ml of blood, typically 0.12g-0.16g per ml blood, with men tending to be at the higher end. This works out at about 0.0000025 moles per liter, or 0.025mM/l, variation 0.022mM/l to 0.027mM/l.

So you can absolutely carry out a calculation to convert HbA1c values. How do you think the meter, which detects a very small change in current, converts to glucose per volume? The conversion is improved if you know your own Hb numbers, or if you make assumptions based on gender at birth. The conversion though will never be precise because of the difference between venous blood and capillary blood, but the difference is insignificant if all you want to know is "how close am I", for directional purposes. The variance would also be of less importance for those firmly in the non-diabetic camp, and those firmly on the diabetic camp.

The OP wanted to know what their likely fall is, but as implied, thats impossible to determine from a forum posting. Single BG measure will, in my experience, give little indication, hence performing moving average calculations, I would suggest for about a month to gain confidence.

Not sure why congratulations is in order. I did FA to move the numbers. Just applied some logic. I think its diabolical that the NHS will not issue all diabetics a £5 meter, and £10 worth of strips a month. Very cost effective manner of applying nudge psychology. People react well to targets, and feeling in control. There is no control when you are left guessing over food, and being at the mercy of label values. They expect diabetics to manage their own disease, usually through exploitation from any number of lifestyle charlatans and grifters out there.

The proxy HbAc1 value had value in that there was zero surprise or shock at the NHS determined number. The value of routinely recording rolling averages is demonstrated here (I need to update this)

1751618068049.png

Daily noise:

1751618129199.png

So naturally over the Christmas period there were some high values. In the 30d rolling average, there was an extended recovery to steady state. In the 60d rolling average, I see no issue.

If you have a meter, get hold of a cable to pull the data into a laptop. I just called the manufacturer of mine, and they sent me out a free data cable, and a lifetime supply of triple A batteries. The software the meters come with is rubbish, and I just use to to convert to a csv file, and away I go.
 
Interesting @saz9961. Where did you get the equation from? I have developed a predictive equation from my data which now extends over the best part of 6 years simply by correlating the dozen or so HbA1c results I have had with the mean blood glucose over the 90 days preceding the HbA1c result. The correlations are very good (R^2 +.98) and have quite accurately predicted my last couple of HbA1c's. I am currently predicting an HbA1c of 46 mmol/mol and from previous experience I would expect that to be within +/- 1 unit of a lab result.

Plugging my current 90 day average into your formula predicts an HbA1c of 40 - somewhat lower than the prediction from my data.

Sadly my AccuChek Performa Nano (provided by my GP with a generous supply of strips) does not have any external connectivity, wired or bluetooth so much of my data transfer has been painstakingly manual. It finishes up in a csv file and is interrogated either by a spreadsheet or more recently by an app I have written to run on a SBC.

Fully appreciate all the nooks and crannies when it comes to dunking a strip with a chemical on it into a drop of blood and then magically displaying a number on a screen. Also the many more nooks and crannies when it comes to drawing conclusions from those numbers. My advice to anybody worrying about small changes in readings is to round the reading on the meter to the nearest whole number and then only bother with differences of 2 or more units.

Finally, I am getting more and more comfortable with the notion that one's 30 day waking average is a very good guide to overall glycemic control. If it is below 7, then it is probable that HbA1c will be below the diabetes diagnosis limit.
 
Docb said:
Interesting @saz9961. Where did you get the equation from?
Click to expand...
There’s plenty of equations available. You can use the one in the libre manual or dexcom manual if you want an official one. All they do is manipulate average glucose levels.
 
Well blow me down @Lucyr . The meter came from my GP surgery without a cable. Since it had neither usb socket or any sign of bluetooth I assumed that there was no connectivity. It never occurred to me that it might have some odd system. I will investigate.
 
Interesting observations @saz9961. I’ve split this more technical discussion away from the earlier thread so that it doesn’t overwhelm the original topic.

saz9961 said:
So you can absolutely carry out a calculation to convert HbA1c values. How do you think the meter, which detects a very small change in current, converts to glucose per volume? The conversion is improved if you know your own Hb numbers, or if you make assumptions based on gender at birth. The conversion though will never be precise because of the difference between venous blood and capillary blood, but the difference is insignificant if all you want to know is "how close am I", for directional purposes.
Click to expand...

I’ve never seen these as a conversion. The two values are measuring completely different things, and there is no numerical/mathmatical conversion between them. There is only an estimate of what value x often leads to in terms of value y.

Almost certainly not as scientific as your understanding, but in lay terms - HbA1c measures glycosylated haemoglobin. That’s a one-way change in red blood cells based on their exposure to circulating glucose. The higher the level of circulating glucose, the more red blood cells are changed. So you can use Haemoglobin A1c as a way of observing glucose concentrations over a 3-4 month period (112 days being the approx life of red blood cells).

Fingerstick results (I think) are based on electrical measurement following an enzyme reaction in the strip and provide a plasma glucose (not whole blood glucose) value with reasonable accuracy. The ISO standard required around +/- 15% of a lab value for 95% of results.

The Glucose Management Indicator / HBa1c estimates offered by some software packages / CGM, are an empirical ‘best fit’ based on paired 90-day averages and associated HBA1c between either sensor glucose, or fingerstick glucose values.

@Docb has developed his own from fasting values, which works well for him.

I adapted a standard one back in the days when HbA1c were given in %
HbA1c % = (BG Avg+2.52 (I added 0.5 or 0.8 here))/1.583

Dexcom’s empirical GMI (Glucose Management Indicator) is
GMI mmol/mol = (4.70587 x CGM Avg)+12.71
 
Docb said:
Well blow me down @Lucyr . The meter came from my GP surgery without a cable. Since it had neither usb socket or any sign of bluetooth I assumed that there was no connectivity. It never occurred to me that it might have some odd system. I will investigate.
Click to expand...
You had to order the infrared cable from
Accu chek when you got it. Always read the info that comes with these things!! Otherwise you miss stuff like this.
 
@Lucyr - I'm a bloke of a certain age - reading instructions is not high on my priority list!😎

@everydayupsanddowns you make the important point that all the measurements are estimates and do not have the precision implied in the way they are quoted. In a perfect world all the values would be quoted with an error to reflect that.

@saz9961 also makes the important point when it comes to inferring an HbA1c from a series of spot readings that there are differences between individuals both in metabolism and record keeping which will increase the error when using a general equation. My model is based entirely on my data and works very well for me and I am careful to make sure that is understood when I refer to it.
 
The key point here is that this converter "indicates" a conversion.
 
You must log in or register to reply here.
Back
Top