Conversion between Freestyle Libre and NHS HbA1c measures

A

Andy261024

New Member
It looks like Freestyle Libre and the NHS use different definitions of HbA1c. Please can someone explain this to me?

I used the Freestyle Libre sensor for a couple of weeks as part of the ZOE project (not for any diabetes check). Based on the glucose levels this measured, it calculated my HbA1c as 6.2% = 44 mmol/mol. The NHS test however measures "HbA1c level - IFCC standardised" to be 32 mmol/mol. I guess the difference is because of the "IFCC standardised" bit, but how and why? Os is the Freestyle calculation just wildly inaccurate? - though based on the graphs it produced, 6.2 looked like a reasonable estimate.

Also, since Freestyle Libre is calculating this from the glucose levels it measures, if the standardisation is different then it that case it looks like even basic glucose levels may also be measured differently from the NHS standards. One way or the other it all gets to look a bit odd and I'd like to understand things properly.
 
The Libre just gives an estimate. As it’s not that accurate, it can be out. Go by your NHS HbA1C as that’s actually testing your HbA1C rather than making a guess. 32 is a normal HbA1C- ie non-diabetic.
 
The definitions of the HbA1c are the same, but both the Libre and/or the NHS blood lab test can be inaccurate, for various reasons. Some people find that the Libre measures low, or high, for them (I personally find it accurate, but I know some people don’t) and the estimation of HbA1c (and it is just an estimate) will depend on what the Libre sensor has been reading, The NHS lab test can also be inaccurate if you don’t have a 'normal' or 'average' number of red blood cells, as the test measures how much glucose has stuck to your red blood cells, and calculates the answer based on how many red blood cells an average person has.
My money would be on the NHS blood test being the more accurate, though.
 
Libre doesn't actually measure HbA1c, it uses your average glucose level and a conversion factor to try to predict what your HbA1c will be, but average glucose levels and HbA1c are not directly related, so it is only ever going to be a "ballpark" estimate.

For me Libre usually "predicts" a lower HbA1c than my actual measured HbA1c, so my last blood test gave an HbA1c result of 46 and Libre was predicting 43 from the previous 90days of monitoring. This has been a pretty consistent discrepancy over the years that I have been using Libre and ties in with Libre typically reading about 1mmol lower than my actual BG level from a finger prick.
 
I found the Libre estimate lower that the”real” hba1c blood test by more than 20%.
It was so far out that I decided it was just another measurement. Useful to track progress but not as a HbA1C indicator.
 
Andy261024 said:
It looks like Freestyle Libre and the NHS use different definitions of HbA1c. Please can someone explain this to me?

I used the Freestyle Libre sensor for a couple of weeks as part of the ZOE project (not for any diabetes check). Based on the glucose levels this measured, it calculated my HbA1c as 6.2% = 44 mmol/mol. The NHS test however measures "HbA1c level - IFCC standardised" to be 32 mmol/mol. I guess the difference is because of the "IFCC standardised" bit, but how and why? Os is the Freestyle calculation just wildly inaccurate? - though based on the graphs it produced, 6.2 looked like a reasonable estimate.

Also, since Freestyle Libre is calculating this from the glucose levels it measures, if the standardisation is different then it that case it looks like even basic glucose levels may also be measured differently from the NHS standards. One way or the other it all gets to look a bit odd and I'd like to understand things properly.
Click to expand...
The NHS blood test measures the average bloods over a 3 month period. You can get rough trending from the interstitial fluid levels from a sensor. You just done 2 weeks out of 12? Your official laboratory blood test of “ 32 mmol/mol.?” Seems good to me. What does your diet consist of?
 
There’s a little bit about these GMI (glucose management indicators) which make an estimate of HbA1c from average capillary (or sensor) glucose here:

J

Post in thread 'Hi it is JOV'

What are peoples experience of their average morning BGM test in mmol/L compared to their HbA1c.

Just curious.

Including some of the formulae involved
 
Andy261024 said:
is the Freestyle calculation just wildly inaccurate?
Click to expand...

Yes, it is, especially if your BMI and IR are low.
But to fix this problem they simply changed the rules, let me explain with an example.

Imagine your CGM gives a glucose reading of 7 mmol/l and the fingerprick a reading of 10 mmol/l.
A#$%t says the fingerprick may be 15% off and the CGM as well and 15 plus 15 makes 30,
so in this case you will not get a replacement sensor, because it is within this 30% margin. Sounds logical.

But according to ISO 15197* the fingerprick result in this example may not be higher than 9.4 mmol/l.

Your sensors are also within A#$%t’s margins, an HbA1c of 44 means an average glucose of 7.29 mmol/l, so (according to their rules) the A1c may be as low as (0.7 x 7.29 = 5.1 average glucose) = 29 mmol/mol or as high as 59 mmol/mol.
Small acceptable difference between 29 mmol/mol or 59 mmol/mol.

The NHS test you have by the way is probably 10%, 4% or 0% from a reference result, depending on the lab.


* ≥95% of meter results fall within ±15 mg/dL or ±15% of the laboratory reference result at blood glucose concentrations of <100 mg/dL and ≥100 mg/dL, respectively (so you must divide 7 through 0.85 and multiply by 1.15, not add 15 plus 15).
 
You must log in or register to reply here.
Back
Top