My G5 Clarity report is way off my most recent A1C

Good feedback, thanks. It sounds like there might be regions where the estimate is closer and areas where it is not as close, so that type of information is useful to people who possibly get surprised by lab results, like the OP.

Actually my curiosity about it is simply why it seems to always underestimate. I am not saying I am disappointed, and I know it is an estimate, and should be treated as such.

But it just seems very interesting to me that Clarity estimates are either right on, or under. I just wanted to hear if anyone had an estimate for 7.0 and came in at 6.5, or something like that. It just seems we never hear that. So for me, it is just a curiosity!

Not that is is really where this discussion is headed, but for 3 years my clarity report has bene within 0.2 of my A1C. I tend to test a lot however and I tend to correct my Sexi Dexi when needed or even more than needed. I also try awfully hard to keep my blood sugar as close to 100 as possible. Yes, some days (like today) I fail, other days, I am fairly successful.

I suspect one of the reasons so much weight is put on A1c numbers is because most diabetics seen by doctors (including endos) are Type 2, not on CGMs, possibly not on insulin, and often testing at most three times a day. As such, there isnā€™t really enough data to get a picture of what their average blood glucose profile is like. Time in range, variance, even average of the readings all becomes less useful the less data is available.

Iā€™m not saying that makes it right, but itā€™s certainly easier for a doctor to infer level of control from an A1c, a fasting BG, and other metabolic panel diagnostics than it is to interpret meter readings for those that donā€™t test very frequently. And, to my understanding, the vast majority of diagnosed diabetics test only very infrequently. One of the reasons CGMs should be standard care for any diabetic, from a medical effectiveness standpoint.

The A1c test is not perfect. The current lab standards are that to be certified by the NGSP the A1c has to be accurate to +/-6%. This means that a test result of 6.7% could be anywhere from 6.3% to 7.1% and still be within the proper accuracy. And as others have noted there are things that can cause systemic interference which results in ongoing biases in the A1c. The NGSP lists a number of those things.

My general feeling is that the CGM provides a much better measure of blood sugar control. It captures many more readings, the average can be trusted just as much as the A1c and it gives a complete picture of time in range. Since you calibrated every day, as long as your meter is pretty accurate your readings will be pretty accurate.

1 Like

I certainly understand the argument that the A1C is not a perfect indication of BG for the previous three months. However, if I have been pretty consistently at 6.1 (whatever that actually means) and all of a sudden I pop up to 6.7, I think that does give room for pause. Obviously, something has changed, right?

Consider this:

There were studies that showed this, Iā€™d have to hunt for them. How was your control 1-2 weeks before the test? That can bump it up or down more than you might think.

According to Dexcom, my control for the final few weeks prior to the test was pretty consistent with what it had been for the entire three months.

I agree that the A1C test is not a perfect test. So many things going on at any moment in a personsā€™ body that could throw something off.
Also it depends on the lab running the test. Every three months I see my endo and have labs drawn before that visit and at the same time, I go into UCSD for labs for a trial Iā€™m in. The numbers are never the same and pretty big differences. Last trial test 7.4 and endo was 6.8.
My dexcom said 6.4.
Keep in mind I donā€™t trust my dexcom for dosing insulin. It is just not as close as it needs to be. But would still never give it up.

Just wanted to quickly put in an update to this.

Iā€™ve had very, very good luck with my a1c being very close to my Clarity report. Last week my Clarity read 6.0 and my a1c was 5.8. So very close again.

1 Like

We have similar results, although flipped. Calebā€™s clarity was 5.8 and A1c 6.0. Clarity has always been a smidge higher.

Nice a1c for Caleb!!!

I actually attribute the slightly higher Clarity to my One Touch Verio IQ meter as Iā€™ve found it tends to run about 10-15% higher on test results.

1 Like

My clarity is always lower. I was 6.1 on clarity, 6.2 on mySugr and my A1c was 6.6. Donā€™t understand why such a difference.

Well, the Dexcom Clarity estimated A1c is now, as of February 11, 2017, ā€œtemporarilyā€ offline until Dexcom decides how to change the calculation or provides better e-A1c citations. Iā€™ll be curious if the actual calculation changes as I was an outlier on their previous estimate.

1 Like

My theory:
I think the reason it is off is because of the 20 minute lag between actual BG and interstitial fluid. If you see your Dexcom going up and it is at 150, you will correct. But your BG may already be at 180 or 200 by then. So the Dex report will not reflect the time you spent at 200. Then after your correction, your BG will come down, and the Dex will reverse. So it never gets up to the high that you actually were.

Along the same lines, when you are low, you correct that quickly. So I think in general the Dex has a tendency to be more centrally biased to actual BGā€™s, and thus gives people a more favorable A1C reading.

What would be interesting, maybe one day I can do this, is to plot several days of Dex against my BG tests. And give a comparison. I would predict that the Dex numbers would always be more central and appear flatter than the actual BG tests.

Problem with such an experiment for me is that I use 5 different meters, and the fact that none of them are integrated with anything other than paper and pencilā€¦:open_mouth:

Your theory makes sense to me. My thought was always that if my reads (mySugr) and the dex reads were close , the A1c should be too. I have wondered if it could have anything to do with the lab procedure. Would all labs give the same A1c? They should.

Support First Responders

The A1C test is performed in a standardized way in the lab. From one lab to the next, there is not great variability in what it measures. BUT, there is great variability in how that marker determines a personā€™s control. The test simply measures how much glucose is attached to the hemoglobin. But there are factors that can change that.

So while I donā€™t think there is variability from one lab to the other, I do think there is variability for each person. For example, if you have severe lows, your body can try to strip away glucose from the hemoglobin and use it. That would lower your A1C. If your red blood cells donā€™t last as long as those of someone else, then your test may reveal your average blood sugar over a shorter time than 3 months.

More recent blood sugars also have a higher impact on the A1C than those from 3 months ago, because you have more red blood cells alive from recently than you do from 3 months ago. What happened with your blood sugar 3 months ago, a lot of those red blood cells are gone.

I consistently have lower HBA1C measures than what Clarity suggests. A1C is always in the 6.0-7.0 range and average blood sugars in the 150-190 range.

Same here. The CGM if not calibrated frequently will have a clarity reating different then blood/lab draw amounts. I have found this to be true often when reading my clarity # with my Endo vs actual 3 month A1C.

My endo, my CDE, and I trust my CGM data more than my A1C. Why?

Here is a reference to the original, frequently referenced paper that came up with the ā€œmagic equationā€ that most systems use to convert eAG to A1C and vice versa:

If you look at that reference and scroll down to Figure 1, you will see a cloud of data points and a single, solid dark line running through the cloud. That solid line is the ā€œmagic equationā€ that is still largely used today.

I donā€™t think that you have to be much of a statistician (I am not ā€¦) to appreciate that there is a significant margin of error when applying this equation.

In my case, my A1C is pretty consistently 0.6-0.7 HIGHER than predicted by my 30- or 90-day CGM averages. Of course, I never realized that until I had a CGM ā€¦

My advise is to not rely too heavily of the equations that convert eAG to A1C and back without considering the accuracy of that equation for any particular individual.

Stay safe out there.

John

2 Likes