A1c issues

Hi, I have had T1 diabetes for almost 45 years. I use a tandem pump w control iq and fiasp insulin. My A1c is usually around 6-6.5. My last A1c was a 5.2. I was surprised because I use sugarmate app for analytics and it said I should be 6.4 and usually the sugarmate number is very close to the blood test result. And I looked at the detailed data and came to almost the same #. So I asked my doctor to re-test. This time it was 5.3. I don’t have more than 1 low blood sugar a month, if that. I’m wondering if there are other things that impact the A1c that could be causing this? My internist suggested that maybe since I’ve had T1D for so long it’s no longer a valid test? Could that be right!? My fructosamine was a 283. I am on a fasting diet 16/8 and sometimes have slight ketones because of that. Not sure if any of that matters. Any thoughts appreciated

1 Like

Changing labs can cause an A1C shift often in the 0.5% range but potentially in the 1.0% range.

If for any reason your body is replacing red blood cells faster than before then your A1C will drop even though your bg average is the same. For example, blood donations or surgery.

There’s a substantial amount of scatter in the correlation between average bg and A1C, way more scatter than any of the “calculators” or “conversion tables” would lead you to believe. See scatter in chart before.


Your A1C is affected by how long your red blood cells last in your body, as well as what your glycemic control was in the months preceding the test.

  1. It is possible this test was just a one-off, like a bad test (as @Tim12 referenced above).
  2. It is possible that your control was much better in the months preceding this test (so you dropped from the 6’s to 5.2).
  3. Or it might be that your red blood cells are not lasting as long as they used to.

Of this list, #1 and #2 are not a problem. But #3 could possibly be a health issue. So for that reason, I suggest getting it re-checked in a few months and then exploring if there are other issues with your red blood cells.

A shortened red blood cell lifespan may not be a health concern. For example, very heavy activity can shorten the life-span of red blood cells too. So it just depends on the cause.

This is a good article to read:



My endo switched to a fingerstick A1c device that seems to read lower than my “best guess actual” using Dexcom Clarity. She’s been using it for about 2 years and the variance has been consistently in that 0.5% to 1.0% range. It does not really matter to me what her device shows so I have not had a recent lab A1c for comparison.

1 Like

Thank you! I have gone to the same lab multiple times with no issues, and a repeat test a month later produced nearly the same results. My control is consistently good but not that good. And I am a data geek and pay attention. But…that article you sent is so helpful. Especially since about a month ago I stopped taking the vitamin B12 supplement I was taking (suggested because of tingling sensation in my hands). So I’m thinking the pernicious anemia (suggested by the article) might be my issue. I will try the vitamin B again and retest in a few months. Appreciate you Eric2.


Thanks everyone


Your a1c is weighted to your night time blood sugar. So if your nights got better, your a1c will improve by a lot more than you might think.

Most red blood cells are produced while we sleep. Red blood cells are glycated at the time they are released into your blood stream. That’s why they are a good indicator, they live 3 months so we get this 3 month view into our sugars over thst time period, but obviously it’s not perfect and only a rule of thumb since so many factors effect it.

I depend much more heavily on Time in Range


Same thing happening here, I’m pretty sure my iron is low which is known to cause a difference in A1C vs BG averages. I need to get it checked but haven’t gone to my primary yet to have the test ordered.


I’ve also had T1D for 45 years , the same equipment for a year and use Novolog - and had the same experience. From 6.7 it dropped to 5.3 and then 5.2. It didn’t make any sense to me because the stats from the T:connect portal and Xdrip+ both indicate my A1C should be closer to 6.2.

I recently discovered that I had been experiencing a severe vitamin B(s) deficiency.

Iron and folate (B9) and B6 and B12 are needed for hemoglobin production and its health.

I’ve got a endo visit next week and am interested to see whether my A1C has risen since I started taking a B-complex supplement.

[update 6-11-24] My A1C dropped again to 5.0 . My CGM stats all say it “should be” 6.1.


You are using Fiasp in your Tandem pump w no problems? How long has this worked for you?
(I had heard that Fiasp doesn’t “agree” with the Tandem cartridges. I’d much rather use Fiasp!!)

Interesting. I’m a vegetarian (that still eats fish and dairy), and my A1C lab result always comes out lower than what my Dexcom averages. My last A1C was 5.8, but my Dexcom averaged 6.2.

Sugarmate displays a GMI not an A1c, different number.

Glucose Management Indicator (GMI): A New Term for Estimating A1C From Continuous Glucose Monitoring


GMI vs A1c

1 Like

Correct but if you aren’t having lows the numbers should be close. Assuming your Dexcom is accurate:)

Sorry Timothy but the A1c is not weighted to the night time blood sugars. Overnight sugars count as much as day time— A1c is basically an average, assuming no other interference (see above). It’s not perfect—low blood sugars can offset high blood sugars all else equal. Night time control can make a lot of difference though, since that’s about 1/3 of that average—sometimes fixing the overnight numbers can move the dial quite a bit.

I am using fiasp. It works great as long as I change my infusion set timely. If I get past the 2 day mark, I start having issues after that time period, especially with the autosoft 30 infusion set.

It is weighted because most red blood cells are produced while you sleep. Red cells become glycated as they mature, and are released by the bone marrow. There is a window for glycation.
After they are in the blood stream, they do not be come glycated and they can not lose their glycation. That’s why they are an indication for glucose control. They lock in the glucose levels at the time they were released.
If red blood was produced evenly all the time, you would be right but it’s more like 65% of red blood is produced while you sleep. So 65% of your red blood cells will lock in the glycation rate comesurate with what the glucose level was at that time.
Therefore it is reflecting your night time sugars more than daytime.

Thank you very much… yeah I’d heard that it didn’t do well with the cartridges … n
ow to decide Fiasp (& change cartridge frequently) OR Lyumjev (and deal with a little sting on occasion).
I really appreciate your reply!!

AND how does one make a personal reply without posting it???

1 Like

GMI is just an updated correlation between HbA1C and average blood glucose. Assuming your CGM is Dexcom, you can look at/compare Time In Range and (CGM) Average Glucose. If Time In Range and Average Glucose are close, your control hasn’t changed.

Tidepool gives a more precise view of TIR, showing hours/minutes per day in range. Dexcom dropped tenth of % from TIR in their reports.

The nighttime weighting isn’t an adequate explanation for my nightime level being kept between 110 and 120 mg/dL (averaging closer to 120) by Control IQ, and my A1C being 5.0. All my bloodwork is so normal it’s abnormal.

I think that it’s past time we stopped treating the A1C as an objective metric, and looked at it as what it is - a statistical correlation with a large standard deviation. The scatter graph in Tim12’s post shows it for what it is - a lame attempt to fit a straight line to measurements from a group.

The line has no objective reality or value because ninety plus percent of the values are NOT on that line. Taking a group statistic mean, applying it to an individual and saying that any individual in that group is wrong and needs to change their numbers to “toe the line” is silly. The total statistics from the DCCT have a much wider range than what’s shown on that graph, and if you include people who are using modern technology, the distribution is wider and lower. 25% are more than +1/-1% off that straight line.

It made sense to use it with the DCCT which was a large cohort study in the 1980-90s. But it never made sense to apply absolute numbers from that best fit straight line to managing an individual case of diabetes. It’s science misunderstood and misapplied.

Choosing to use one metric in the wrong context, doesn’t make it gold, it makes its believers foolish.


That seems contrary to the following-

Glycated Hemoglobin or Hemoglobin A1c

HbA1c is a minor fraction of adult hemoglobin, which is formed slowly and nonenzymatically from hemoglobin and glucose. Because erythrocytes are freely permeable to glucose, HbA1c is formed throughout the lifespan of the erythrocyte; its rate of formation is directly proportional to the ambient glucose concentration. The concentration of HbA1c, therefore, provides a “glycemic history” of the previous 120 days, which is the average lifespan of erythrocytes.


1 Like