How useful is an A1C test less than 90 days after the preceding one?

Continuing the discussion from Starting insulin . . . so far, so good:

I am hoping that someone who has actually looked into (in some depth) what factors influence an A1C test result will comment. I just wanted to mention that an A1C result may not be as simple as whatever I always thought it was.

I say this because a few months back I tried to schedule an A1C sooner than 90 days after the previous test. That requested A1C was ~85 days after the test before it, but the VA’s software rejected the attempt to schedule the test because it was less than 90 days. Apparently someone in a position of influence for some ostensibly good reason(s) feels strongly that getting an A1C more frequently than every 90 days is a waste of resources.

FWIW, the test was able to be scheduled and, IIRC, the results were identical to my previous A1C, 6.7%. :disappointed: I expect my next A1C will also be in that general vicinity … and hopefully not worse. Though from looking at my CGM numbers I’m not sure I can count on that.

But I digress … :blush:

Because the test measures the previous 90 days’ average, most insurers limit them to once every 90 days. Mine did as well. When I began eating LCHF, I was anxious to see the effect it was having, so I purchased my own self-kit from Amazon (most drugstores also carry them now). I’m glad I did, because it showed that in 45 days, my 8.7 reading went to 7.0. At the end of the 90 days, the test at the doctor’s office was 6.3. The 7.0 reading gave me the impetus to keep going and to really stay the course.

Good luck!


I can’t cite any sources, but I’m under the impression that the most recent 30 days before the A1c draw has more weight than the initial 30-day period.

I’m interested in this issue since I asked my doctor for another A1c draw at 45 days. I may run into insurance payment problems from what you write. I requested this quick retest since my last several A1c’s are higher than my meter and CGM numbers suggest.

My hypothesis is that due to my considerable coffee habit, I have been chronically dehydrated. I upped my water consumption just after my last A1c draw and think that a 45-day test will confirm or reject my hypothesis. My last A1c came in at 6.0% but my meter numbers say I should be below 5.5%. We’ll see.

Good luck getting approval, Terry - I’m interested to know if they will do it.

As for your coffee habit, I thought that was laid to rest last year - here’s the NPR article about the study:


I agree it’s a waste of time and money. Some blood tests aren’t very useful if done too soon. Others can show meaningful change in just as week or two (liver tests results can change dramatically in under 3 weeks. been there; done that).

I’ve heard the same about the most recent 30 days having more weight than the first 30.

Not a direct answer to the OP, but somewhat to your question @Terry4 can be found here

You are correct the results are weighted toward the more recent history. This is because a higher percentage of the “older” glycosylated hemoglobin cells (not sure if they’re cells or molecules, actually) have died off than the younger ones… In layman’s terms

exactly. the results are skewed towards more the more recent period, rather than showing a “fair” comparison to an A1c taken in the more distant past. someone who has “gotten religion” and wants to see if they have improved their A1c by too quickly getting it tested, is going to get a falsely better (compared to their older result) A1c result that won’t be sustained if they backslide, and conversely, if they maintain better control in the future, will improve (the A1c result) even further.

Thanks for the link. My initial reasoning about my “skewed high” A1c was mentioned in this article. Not everyone’s red blood cells live the same length of time. But then I read some studies that cited dehydration as a reason for a false high A1c. It made sense since the lower blood volume concentrated the red blood cells in a solution of relatively higher glucose concentration.

From your link, this impressed me:

In a person with normal blood sugar, hemoglobin will be around for a lot longer, which means it will accumulate more sugar. This will drive up the A1c test result – but it doesn’t mean that person had too much sugar in their blood. It just means their hemoglobin lived longer and thus accumulated more sugar. The result is that people with normal blood sugar often test with unexpectedly high A1c levels.

So, if my BG control markedly improved, which it did three years ago as measured by my fingersticks and CGM time in range numbers, then my red blood cells live longer when compared to my relatively worse control in the preceding period.

Bottom line, I think the clinicians put too much stock in the A1c and have swayed me somewhat in their interest in it. They’re also unduly hypo-phobic and I definitely don’t share that value. Perhaps I should just wait for the 90 days to elapse and get my next A1c then. Drinking more water will not hurt me.

Thanks for referencing this article. As a die-hard coffee drinker, I’m happy to read that there’s some science to support that coffee does not dehydrate. If that’s true, and the article is persuasive, then my hypothesis that adding water will lead to a lower A1c will be disproven. (I know, with an n=1 experiment like this, it is not statistically significant!)

Following the other link that @YogaO cited, I now doubt the ability of the A1c number to give me solid precise feedback on my glucose control. In fact I’m coming around to the conclusion that the A1c can only be used to paint with broad brush strokes. Like reducing your A1c from 9% to 7%. Perhaps as the number approach the more normal 4% range the statistics produce more outliers.

there is a range of normal and it isn’t 4. it is 4.5 to 6

You can screw up an A1c with one really bad day right before the test. I did that once. I had been almost perfect for months, back when I was eating Bernstein’s diet. But my ob/gyn gave me a compounded hormone pill I had an extreme reaction to. My BG shot up into the high 200s regardless of food for one whole day. Then it came back down. The A1c was scheduled for the next day and was 1% higher than the previous and next tests. I think it was my highest ever. But one day of highs would have zero impact on actual health.

Other data I have seen supports the idea the A1c mostly reflects your most recent 2 weeks. But it is marketed with that 3 month claim and docs are too busy to read up on the subject.

Test with a meter, 1and 2 hour after meals and first thing in the morning. That is a much better predictor of long term health.


testing more often than that is even a better indicator of your overall bg’s and therefore should be a better indicator of how your diabetes is under control. I used to test more than a dozen times before getting a CGM. Testing often is OK as long as we don’t succumb to the temptation to over-correct; something I plead guilty to, but I’ve gotten better about it since going on the CGM, thanks to the trend data.

So, to this end (to see how average glucose varies from tested A1c values), please post what Dexcom says your average glucose for a 90 day period is and your A1c from that same 90 day period. I am interested to see if there is much variance.

Thank you

My A1c last week was 4.7 and my pump/meter average was 133 and my SG average was 128.
Sometimes my A1c is just worthless info. They re-ran my sample at the Mayo clinics lab and confirmed that my Glycated Hemoglobin was around 4.7%. They are going to re-run my A1c again in two weeks. My Hematocrit was a little lower than the cutoff.

Remember that your A1c is just the % of glycated hemoglobin in your blood sample at the moment
in time when they draw the blood out. I have had many A1c tests that my doctors said where great for a diabetic…but I still suffer from high BG, nerve damage, Gastroparesis, and have a GFR in the low 50’s. Having good A1c’s for many years has not changed the fact that I’m a diabetic.

1 Like

My recent blood draw for my A1c showed my 90-day Dex CGM average of 107 mg/dl. That would predict an A1c of about 5.3%. My A1c clocked in at 6.0%, an equivalent 90-day average of 126 mg/dl. My meter showed a similar average as my CGM and my fingerstick taken at the time of the blood draw came very close to the lab value.

I’m beginning to appreciate that the A1c metric is flawed and probably over-relied on by busy doctors.

1 Like

I don’t know if that’s the conclusion we need to jump to… I don’t see it as flawed, just limited in its own right just like any other single measure. I’m not sure if we actually know which it is that harms our body over time-- is it transient and fluctuating glucose levels? Or is it the amount of our hemoglobin that is glycosylated over the course of our lives? Or a combination of the two? We can speculate all we want but the reality is that we probably really don’t know…

I understand the A1c test uses the hemoglobin as a proxy to measure the average concentration of glucose in the blood. It’s this concentration of glucose that is exposed to all the tissues in our body. I don’t know if the hemogloblin itself is damaged but we do know that kidneys, eyes, peripheral nerves, autonomic nerves, and vascular systems do not like concentrations of sugar that are too high.

I’ve also learned that glycation occurs in the fascia in the bottom of my feet that causes adhesions that lead to plantar fasciitis.

I’ve read speculation about fluctuating BG levels as also causing stress. I personally have more energy when my BGs stay in a more limited range.


Unfortunately, not everyone has insurance that pays for unlimited strips. So testing ten times a day, though very informative, may not be possible.