Yes, my range is 65-120 nighttime and 65-130 during the day. But that’s just for my own assessment purposes. Not to be confused with my pump (and Loop) settings, where I have 80-90 as the target range.
May look like this: skip breakfast, 15g morning snack, 40g lunch (with lots of protein+fat), 15g afternoon snack, 20g dinner, 10g evening snack. But this varies a lot from day to day - I am not into any prescribed patterns or diet rules, with one exception: skipping breakfast works well for me.
I have always had a higher A1c than my meter and CGM suggested. My average blood glucose shown by my meter and my CGM usually predicts an A1c of 5.5% or less. My last A1c was 6.1%.
I’m thinking that the A1c number is not as precise as we think it is. One study that examined how an A1c value could be translated to an estimated average blood glucose number actually showed that that number was derived from a range of numbers. Here’s a part of the table posted in the study:
As you can see, an A1c of 5.0% translates to a range from 76-120 mg/dL. The average of that range is 97 mg/dL. This is usually the number that the doctor or other medical professional will tell us about. But that 97 is actually an average of a range of people. You can see how the 5% A1c range significantly overlaps the 6% range and so on with the other ranges.
I have long thought that we as patients should reject the use of the A1c as a legitimate precise proxy of our overall blood glucose control. The A1c does not reflect BG variability or time in range, much better descriptors of the quality of glycemia.
In diaTribe’s article, Going Beyond A1c – One Outcome Can’t Do It All, they write this about the distinction of the A1c of a population of people versus the A1c of one person:
A1c is a valuable measure for how a population is doing with diabetes, but for individuals themselves, knowing what goes into a person’s A1c is really important.
I do not have a complete understanding of the A1c number but I get the distinct impression that we should not invest in it an accuracy and precision that we do, both the patient and medical community.
Unless some other factors are affecting BG, the A1c is useful when it goes up or down, rather than indicating an accurate BG average.
1 Like
A1c also varies by lab. I’ve used lab where the range is 4.8-6.2, so your 6.1 would be in the normal range. The other lab I use lists 4.5-6.0 as normal.
I agree that the A1c is a good indicator of long-term BG trends. I find that the data that my CGM outputs gives me a better picture of my glycemia than the A1c. I follow time in range, time spent hypo, standard deviation (variability), and average blood glucose.
I find time in range to be a comprehensive statistic that reflects all the measures I follow. A high percentage time in range correlates reasonably well with a low time spent hypo, low variability, and a good average. Of course, CGM use is needed to follow time in range.
But which really matters? Overall average(as roughly indicated by A1C) Or variability and time in range? As far as I can tell, nobody really knows…
Also, while time in range may be a good variability may be more accurate statistics, they are hard to document for people who don’t use CGMs. And CGMs are used by a small minority of people with diabetes. So I don’t think the A1c will be going away any time soon…
1 Like
As a stats guy myself, I would definitely choose (if I had a CGM) average reading, variability, and time in range over A1C (which gives a median point of an estimated range). Even if your meter, your CGM, and your A1c match up exactly, there is more information encoded in the average, variability, and TIR from a CGM than a single-point median from an A1c.
For those of us not on insulin, who don’t have access to CGMs, the A1C is the measure I can use to judge long-term progress. I test ~3 times a day, and I do have pretty good statistics from my meter over a 3 month period (and the EAG from my last A1c matched my meter’s 3 month average BG exactly), but with only 3 measurements a day I miss a lot of information. A1C seems useful, in my case, for judging how relative control changes over time (in response to changes in meds or diet/exercise, etc.).
I’d love to have a CGM, not because I really need it (I don’t go very high or very low), but because I love data and would really like to see what happens overnight (right before my DP kicks in) and have a better resolution of readings for when I exercise or go low-ish during the day (60s).
*edit: Realized @Jen basically just wrote exactly what I did 
Yeah I get that and my first tendency would be to think the same— but when I take a step back and look at the bigger picture I can’t help but notice that every study ever done has correlated complication risks directly with A1C, not with time in range and variability…
So I think maybe things like time in range and variability may be phenomenal measurements and tools for tracking in an individuals toolbox and tremendously helpful for management, but in the bigger picture I think it’s hard to make the case based on evidence that they can or should outweigh A1C in the context of complication risk-- which is what really it’s all about isn’t it?
Well, there is a good reason for the studies: A1c has been around since the late 70s as a measure, and has become the standard (and cheap) method of assessing a 2-3 month average blood glucose for patients/study subjects. It’s also a single-point measure (i.e., you take blood draw). From a technician’s viewpoint, A1c is markedly superior to CGM (cheaper, easier to do, easier to do right).
CGMs are relatively new (compared to the A1c), and to get very high quality data you’d need to know that the sensors were placed properly, that the CGM was calibrated consistently, and that there is no wonkiness in user or tech error. I think what we’ll see is that more studies will start using CGM data as the technology becomes more standardized, more common, and scientists become more used to estimating accuracy of individual CGM records.
So you’re exactly right. CGMs are (currently) far more useful to individuals than an A1c, but in terms of scientific assessment of average + variability + TIR vs. median BG as a predictor of complication onset, we probably don’t have the data sets to make good statistical analyses. I hope (as more people get CGMs) that that will change. Just imagine a study 20 years from now where a cohort of diabetics has used Dexcom CGMs consistently for that 20 years. You’ll be able to do look at that question in great statistical detail: how does median BG (measured by A1c) compare to 3-way data from a CGM for predicting complications. Even then, the data will be noisy because so many other factors contribute to complications (from genetics to diet to trauma to…etc.).
Yes but 20 years from now we could just as easily look back when we have adequate data and see that variability and time in range don’t matter one bit and that A1C really was the gold standard all along
I don’t think that’s necessarily a likely scenario either… but just pointing out that we’re kinda evaluating ourselves by metrics the value of which aren’t really known, but rather assumed, other than A1C
From my personal experience, I have lived with both poor and very good time in range and variability. For me it’s not even close, when I spend 80-90% time in range and a low standard deviation, I feel more energy and think more clearly when that’s the case. When my BGs are only 40% time in range with high variability, I am tired, grumpy, and have little patience for even mild challenges.
Perhaps you’re saying that we don’t know in a scientific way. I agree with that. But I know well enough living with diabetes and watching my CGM data.
1 Like
While we may not yet know in a scientific way, the fact that my body continues to feedback to me a sense of well-being, I know that a high percentage of time in range and low percentage hypoglycemia are highly correlated in me. Failure to prove something scientifically does not make it untrue; it’s just an untested hypothesis.
My time in this world is finite. I find it valuable to act on personal health observations because it’s likely that science will be chasing these issues long after I’m gone. Sometimes you’re forced to go with your best guess.
Not dismissing the value of your quality of life observations… and I agree with you that I feel best when by bg are tightly controlled within a narrow range and with a low A1C. However to say that variability is more important than A1C-- in the context of avoiding complications, seems unsubstantiated in my mind… of course in the perfect world we’d all have healthy A1C and low variability.
To imply that low variability is more important than a low overall average could be extrapolated to the point of saying someone is better off with their BG set in stone at 300 with zero variability than fluctuating wildly between 80-100 with a low and healthy overall average (A1C)… that’s an extreme example and a bit of a silly comparison, but I think you see what I’m trying to say… variability and time in range are important in management, but only meaningful in the overall context that you’re already in tight control… imo
1 Like
I said high % TIR and low variability are key stats associated with me feeling well.
1 Like
1 Like
Yep, think we’re talking about different things. If your estimated average BG is 100 mg/dL, and your A1c is therefore 5.1%, that doesn’t tell you anything about TIR or variability. You could be regularly swinging between 40 mg/dL and 160 mg/dL, or right on 100 mg/dL but only +/- 15 mg/dL. A1c doesn’t tell you anything about TIR or variability, just that median value. The comments about variability aren’t saying minimizing variability is most important, even when at the expense of average BG. I suspect most of us want to live in the “normal BG range” if and when possible, and we want to minimize variance to the point where we don’t spend any time out of range.
And I think there is enough data out there on the cost of excursions (even brief ones without really highly associated A1Cs) to make a reasonable conclusion that reducing variability is very important in the long run. The only thing we’re missing is long-term studies demonstrating the relationships between the different measures, and that will come as CGMs are more widely used and the data is of sufficient quantity and quality to test the hypotheses. Til that point, there is no question that to an individual, CGM data is far more useful than A1c can be.
1 Like
On the other hand, all clinical trials have only been able to tie A1c to complications, and measures like average blood glucose have never been validated in the same way. So it’s possible that some common factor leads people to hang onto RBCs longer AND causes complications, in which case it’s premature to scrap the A1c.