HbA1C vs CGM: Some clarity at last

#1

As a few others I have a stubborn HbA1C that hovers at 5.9% since I measure it.

For the record, I’m eating low carbs and I’m currently taking Empagliflozin 25mg/day, Tirzepatide 5mg/week, Plioglitazone 7.5mg/day and Acarbose 100mg/meal with a little carbs.
My fasting glucose is 90mg/dl and my insulin is 2.7 uIU/ml which gives an HOMA-IR of 0.6.

So why is my HbA1C still at 5.9%?
The eAG (estimated Average Glucose) is 124 mg/dl which is very far from the fingerprick glucose measurements which are around 100~110.
The CGM average at 104 is also very far from that 124 eAG.

In that case there are various dubious theories that my RBC might live longer because I’m on low carbs or an endurance runner so they would accumulate more glucose. Other theories are that the glucose spikes are more4 glycating than the average glucose.

It took me a while but I finally did on deep dive on how they found out that eAG calculation which is all over the internet.
That formula is eAG=(28.7×A1c)−46.7
The problem is that the 2008 study that determined it is incredibly mediocre at best!
Here it is: Translating the A1C Assay Into Estimated Average Glucose Values
Basically they asked 507 persons to get 7 fingerprick glucose measurements:

AG was calculated by combining weighted results from at least 2 days of continuous
glucose monitoring performed four times, with seven-point daily self-monitoring of capillary
(fingerstick) glucose performed at least 3 days per week.

The 7 points where supposed to be premeal, 90 min postmeal (3x), prebed.
In the end they got:

36% of the seven-point profiles were complete, with the mean number of tests being 5.1

Then they somehow decided that those shady measurements were representative of the average glucose while it’s obvious that this is going to over exaggerate the average glucose by quite a lot.

Basically the eAG is totally worthless.

In 2018 another study was done. This time more seriously, using CGMs: Glucose Management Indicator (GMI): A New Term for Estimating A1C From Continuous Glucose Monitoring

The formula they got is quite different: GMI = 3.31+(0.02392×Mean Glucose)
Where GMI = Glucose Management Indicator = A1C estimation from CGM data.

If I use that with my average glucose of 104 I get 5.8% which is almost exactly the 5.9% HbA1C measured by Labcorp.

BTW 100 mg/dl is exactly 5.7% with that GMI formula.
The conclusion of all that is that to have a normal A1C below 5.7% the average glucose needs to stay below 100 mg/dl which is very very far from 117 mg/dl as given by the old eAG.

Next steps for me: Replacing Plioglitazone by Imeglimin and cranking up TZ.

1 Like
#2

The couple of things I’d think about is … if we think your RBC’s are living longer than anticipated, check a reticulocyte count. However, I think the fructosamine is a reasonable approach here.

I’ve personally found CGMs to be incredibly valuable on seeing a pattern of glucose in response to what I do. I’ve also found them to be quite unreliable compared to blood testing. For example, the last time I had a blood test for fasting glucose and was wearing a CGM at the same time - CGM = 90 mg/dL, lab test was 64 mg/dL. This was with a freestyle libre 3.

I find these devices to be incredibly valuable in regard to patterns of response to food, but their absolute values, at least in my case (and my wife’s) have been different than formal blood draws, with the CGM being 20+ mg/dL higher than actual.

1 Like
#3

I hear you on swapping ime for pio, based on safety. But I wonder if a more nuanced conversation is necessary. To me, pio safety is much enhanced (modulated) by the addition of telmisartan and a SGLt2i (plus keeping pio at no more than 7.5mg/day). As I take both, I am not as concerned by pio safety (for extraneous reasons - ACDF surgery - I will not touch pio for at least the next couple of years). However I don’t know if you take telmi, so that might alter your calculus otherwise I’d wonder why not add ime rather than swap for pio. Btw., I’m in a similar A1c situation to you (5.7-5.9), though my HOMA-IR is not as good (1.7).

#4

Indeed, they are not very accurate out of the box but can be calibrated with fingerprick measurements. The Dexcom CGM app can do the calibration and once calibrated over the first few days it’s generally only off by a few mg/dl.
Sadly the Libre 3 (which I will have to use soon due to the health insurance) do not.

1 Like
#5

I take Telmisartan 80mg/day.

Plioglitazone makes me gain fat. BTW it’s not a side effect but rather the intended effect (store the glucose in normal fat cells). My DEXA fat went from 14 to 16%.
Also it improves my insulin sensitivity but the problem is more than my set point for the blood glucose is too high and the liver happily makes too much of it.

I’m going to discuss that with my physician soon to decide what to do next but I do what to reduce that glucose! :slight_smile:

#6

Right. But keep in mind, pio provably lowers hepatic neoglucogenesis, one of the only agents I know of which does so reliably - metformin has that reputation, but closer studies prove that to be a mistaken notion. I posted studies to that effect in the big pio thread. Anyhow, it’s all somewhat complicated😅.