CGM’s direct comparison (g6 versus FSL3) - what is going on here? Please advise

Hi, all. I am trying to learn and interested in your opinions. After lots of g6 “bad” sensors (despite multiple calibrations) and same careful application techniques, etc, I finally ended up with 3mbad ones in 3mdats (having suffered too long with previous poor performers and having lows due to dexcom being 44+ off, and actually not even always in the same direction. So I began to put on Freestyle Libre 3’s, same relative location, but sometimes same place but other arm when not on same arm. From June 3rd to July 11th I did this. I use a Contour Next One meter and follow protocols on using it. In a nutshell, for 12 finger-sticks during this time range (note: once I felt I could trust one or both or knew which one, I did not fanatically test, so these are 100% of my finger-stick data comparisons): 4 of these, the FSL3 was dead-on (which is confusing since interstitial blood is supposed to be 15 minutes behind - but I am shocked about it being absolutely dead-on, in a good way for me if going low). 2 readings were insignificant: (dex +3 and FS -8 from finger stick). BTW - most checks were when BG’s had been steadty, but a few for necessity of checking. For each if the dead on FS readings, dex was: 11 lower than finger stick, 37 lower, 39 lower, 20 lower. One comparison had FS 6 higher, but dex was reading 55 lower (after 2 prevous calibrations). And, of note or not, Dex gave low alarm of 74 and diagonal falling during lunch on July 11. Libre 3 was 107 and Finger stick was 99. Another time dex was 171 and FSL3 was 150, but I did not do finger stick, just showing it can also read higher. The worst was when dex dropped to 44 and falling diagonal, then immediately to “LOW” which I had never seen before, only to find I was 99 with Libre reading105. So, my questions:

  1. Could it be that because the Libre 3 is inserted with the cannula or whatever it is called, straight in at a 90 degree angle and Dexcom is very angled, that my body reads better with 9o degrees? Sometimes upon removal, Dex seems damn near parallel. But, I should point out that Dexcom g6 has worked for me in the past.

  2. Related to above - lately, I have noticed the two black dots where the transmitter sits, that the pieced is dangling prior to my putting in transmitter, as if unglued. Maybe it was always not firmly attached, but I mean literally dangling and raised a concern from me.

  3. Just what is going on here? I know some people don’t believe in certain codes being problematic, but I really have one that is and in my efforts to get over the superstition, I would still use them (and because my boxes of three might have 2 or all 3 with the same code), but seriously is real to me.

  4. And now, the replacements with longer expiry dates are just as bad as the ones they replaced.

  5. So now, if I could tolerate not having the slider to check my hourly overnight sugars and such that the dexcom has, and if I could tolerate the awful alarms on Libre 3, and if Libre 3 worked with my pump Tandem t-slim or OP5, it would be a no-brainer for me to just use Libre 3 and call it a day. But I am addicted to that slider feature. Still, while on MDI, do you think I should just go to Libre for peace of mind on treating lows and correcting highs and the benefit of one-hour start-up which matters to me, and 14 days of accuracy, in my opinion, if sometimes a day to get going right? My fear is I will want to put on my pump and be sort of out of sync with even the errors of Dexcom and have a rough time of it on CIQ or automated OP5. That is primarily why I have tried to stick with Dex, and even while wearing both, once I thought Dex was fine I turned off Libre 3 app to just use the Dex and reopen if readings seemed totally off (and Libre 3 fills in missing data and does start right back up with info. when you do this). I am very lucky that my endo in Feb. suggested I use Libre 3 if going back to MDI since they have seen so many problems lately with Dex (and those were her words, and this is the premier endo office in the capital city of my state which people try to get into when they have diabetes and other issues, according to my nurse anesthetist neighbor).

What say you, my wonderful team? I am interested to know your views as I make decisions and ponder what is going on. Thanks for hearing me if you made it this far in my post!

The job of the CGM algorithm is to tell you your current BG. It’s true that changes in the glucose content of the interstitial fluid lag the changes in BG, although I think that “15 minutes” is a very old estimate and the current estimates are less than half that.

So the CGM algorithm is basically looking at the trend in the interstitial glucose that it actually measures, and predicting into the future to compensate for the lag, and that’s how it estimates the actual BG right now, which is the number that it shows. This should work well with a steady trend, but a rapid change will make the algorithm estimate wrong until the new trend starts appearing in the interstitial fluid.

As far as the trouble with G6, I just hope the G7 works better in your body. I think it has a shorter, more vertical sensor wire, so maybe that will help. And they say that the sensor works better in the arm than in the belly. For me, the G6 worked well and the G7 (in the belly) is more likely to be off, but I like the fact that with the G7 there is no warm up time at all after the first sensor: Put in the next sensor to pre-soak for a couple hours but don’t start it, it will warm itself up, and then start it and the Dex app will show the data from both the old and new sensors; remove the old one.


With Dexcom, have you ever tried just doing the calibrations once-a-day instead of using the code?

I mean, at the sensor start, not entering any code, and saying “yes I will do calibrations instead…”

I think doing calibrations instead of the code works better for me.


I found this works better for, too.


Thanks, @bkh! That explains alot, and so simply. I do look forward to the g7 and glad to hear those things you mentioned that are pertinent to my situation. Thank-you! I am glad to know it is not a full 15 minute lag time - that was really worrisome with a recent low. I actually have a good dexcom in now, but oddly this morning it had no directional arrows for an hour or so. And this one was good at the getgo - no calibration needed. It was also a code starting with 7 that I have only previously used one time and made a note of it because it was so good. Anyway, thanks for the hope for the g7. Love the no warm-up time and such you described. That really limits my safer times to change on the g6. Very much looking forward to it.

1 Like

Thanks @CatLady and @Eric. No, I have never done it that way, but may consider it. I am becoming a believer that it may work best if never calibrated now that the one with the most calibrations just went off the rails. But thanks for the suggestion.


One of the common things I’ve heard is that if you enter the code and calibrate, it’s a mess.

So if you ever want to try it with calibration, make sure you did not enter the code at start.

1 Like

Wow, so many misunderstandings about the function and facts related to the CGM. It does really make wonder how much negative impact this poor info that is being posted has on the readers misunderstanding and incorrect interpretations of CGM readings. It’s a bit scary. Read the science before posting misinformation, please.

I wouldn’t worry too much about switching. Maintain a prescription for both so the switch will be straightforward? Probably want twoish weeks before switching to a pump, but IME it doesn’t take too long to remember how things work. I switched from medtronic to dex while on a medtronic pump and the changeover was pretty painless.

FWIW, G6 has gotten progressively better for me over the years. At first I only trusted it a little over half the time but now it works so well I forget to second guess the number. Maybe some of this is due to changes in the hardware, but I think I’m also better about site selection and putting it on. Obviously everyone is different.

@Eric Thanks for that. I am glad it is not just me, but a known thing to happen. I will definitely not use the code if I plan to calibrate. Great to know this stuff!

1 Like

@Sjwprod. I will read first next time but was just posting my anecdotal evidence, what my endo told me and reaching out for help in making quick sense during a very trying time. Sorry to have created an issue.

@Karl.n Thanks for your experience. I also had a long spell of rarely questioning it. I plan to keep this string going now that I have one in that felt right and has shown itself to be accurate. And I let my Libre 3 run out the first day of this Dexcom so am using just it now.

1 Like

As far as the slider for the Freestyle Libre 3, there is a workaround. Install the LibreLinkUp app on your phone and then in the Libre3 app settings, share your data with an invite to your own email. After setting yourself up as a follower, the LibreLinkUp will show your glucose values and you can slide the vertical line to see previous values on the graph. I learned this because my spouse follows me on that app. Don’t understand why Abbott gives that capability to the follower but not to the diabetic! But it works until the primary app gets better.

@rdh76 Wow! I never would have known this! Thank-you so much!