Which would you change....Dexcom G5 runs 40 points low


So, unless you were lagging, if your sensor drifts significantly again (30-40) within 1-2 hours, then the sensor is bad and Dexcom will replace it. If this happens with several sensors of the same lot number, the batch is bad. I have seen bad sensors like that before, probably 3-4 in the past 2.5 years. I have never seen a bad batch but I know it can happen.


Thank you, @Michel. I sooo appreciate your intel. I printed one of your posts about G5 transmitters, manufacture dates, etc and brought it home w me to check against my transmitter boxes.

You provide a great service to us lucky FUD tenants.


Three hours later after double calibration: spread of 10. Tested at 139, Dexcom shows 149. Five hours after snack injection.


If the error stays around 10-20 then you are good. The difficulty is to get close when you have drifted far, since, if the CGM is low, it only calibrates to about half the error every calibration. This is when I use a double calibration, as much as I don’t like to force the CGM to a specific number.


The sensor has been within 10 or less points since doing the double calibration last night. Since I’m basal testing right now, I’ve done fingersticks at 1AM, 6:30AM (waking), 8AM, and 9AM.

I never knew about entering a double calibration! Thank you, @Michel and @MM2!


@T1Allison. Just chiming in, and it may not be relevant, but hydration plays a large part in sensor accuracy too. If you are dehydrated it will affect sensor accuracy to the degree of dehydration.

Good luck on your campaign!


Thanks, @elver!! Normally, I probably am a bit dehydrated…but with basal testing and not drinking coffee, I’m the best hydrated I’ve ever been! lol!

A vote for T1Allison is a vote for stubborn hope and sarcasm!!


@T1Allison with that slogan you have my vote. The most forward thinking honest political plank I’ve heard in a long while.


I calibrate anytime there is a > 20% difference. Not doing that causes larger variability and accuracy in the numbers. So, I always calibrate when this difference exists AND when he’s “level”.


So I ended up getting another Faraday restart out of my previously wonky sensor to float me to a more convenient time to do a real sensor change. As long as I did a double calibration once every day or two, it did great. That was one of my older sensors. Not expired (which may not matter), but about four months from expiration.

I switched to a sensor from my most recent order yesterday. It did pretty well for a Day 1 yesterday. I’m on Day 2 and it still wants to read lower than my fingersticks by about 30 points. I just did a double calibration to force it to read 123 to match my fingerstick. Within two data points it had lost another 30 points which did not match a subsequent fingerstick.

So…too soon to really tell…but I think it may be reasonable for me to use Dexcom as an aid in trend direction/magnitude, but to rely on fingersticks for dosing decisions. Which is probably best practice (at least for me for the G5) anyway.


I have occasionally had G5 sensors act similarly with missing a rising BG…it’s almost like they fall asleep and don’t notice a slow rise. Seems to usually happen on a pretty decent in range BG which fools me into not testing with strips. This has occurred for me only on “aged” sensors that are close to or past 2 weeks of use. When this has happened to me, the sensor still seems to track pretty well with BG drops and lower numbers.

The double calibration Michel suggests seems to wake up these aging sensors, producing better tracking of BG rises. I have also restarted sensors If they start with this type of behavior and if I am partly into my second week on that sensor. The restart also seems to wake them up. I look at this “falling asleep” as a symptom of a sensor getting close to quitting but have usually been able to get at least a few more days of use and sometimes more by restarting.


It may be your end conclusion, but it would really be too bad if that turns out to be the case :frowning: I hope you will find out otherwise.

In my experience, there are several types of circumstances when calibration results in bad results:

  • when there is a change of trend direction or a recent variation in slope, the Dexcom eval is off. So, if you calibrate at that time, or within 30 some minutes of it, it will result in a bad calibration. For instance, say you were recently going up by 5 mg/dl/5mns, but, over the past 30 minutes, have been seeing a progessive decrease of the rate of climb, it is possible that the Dexcom is off.

  • when there has been recent (say, last 90 minutes) insulin injection or carb intake, the Dexcom may be off the BG. For instance, when my son takes carbs after a downtrend, we know the Dexcom will be off for the next 35-40 minutes (typically way below reality), so when we are dealing with a long low, we only use fingersticks to decide whether to dose further.

  • There also are circumstances, not very frequent, when, for some reason, the interstitial fluid does not seem to catch up to the BG for one to two hours. Then, eventually, it does suddenly, so, if you were low by, say, 30, and calibrated, you might later be high by 15.

I have also noticed that multiple consecutive lows often leave you a little off. So, after 2-3 hours of difficult lows we often check w a fingerstick.

But it is not, I think, normal, for your sensor to always be off by 40. What I don’t get is, if it trends down, then why doesn’t it keep on trending further down?

I think something else is going on. It would be worth, I think, assuming that there is a cause to these issues that deserves investigating. One possibility, of course, is that you have a bad batch. Have you discussed your issues with Dexcom?


Thank you for your experience and feedback. I always appreciate it and use it!! Always!!

Not yet. I’m feeling like I’m back in the driver’s seat (finally) of my blood sugar these last two weeks. I am sure I’m not done acclimating to that identity change. Which is to say that I’m getting used to redefining where is “good enough” for my own balance between my Atari lines (high and low bg thresholds), how many times I want to be doing fingersticks (I’m averaging 15 a day right now…want to get eased up on that, if possible), how hardcore I’m going to be on pre-bolusing timing since Novolog tends to have better endurance than Humalog, so it’s still critical but not AS critical.

Once I get through another hormonal cycle, get acclimated to my own definition of acceptable, and get used to this paradigm change…THEN I’ll have a better feel for how Dexcom lives up to its changing role in my day. I’m currently (and hopefully that doesn’t change again on me anytime soon) feeling like I have a handle on the cause and effect of my day, and the up and the down of my bg. That’s been missing for so long…I was living an UP and STUCK mystery and misery. But now that it’s back, I’m not quite ready to start asking Dexcom about performance bc (1) I unfortunately do not have too much faith in industry people outside of myself (tech support, endo’s, nurses, etc) once I get to this point of having my own blinders on about staying in my lane for what’s working for me, (2) I don’t want to get sidetracked from where I’m at, and (3) I feel like I need more data before I can conclude anything and whether it needs to be solved.

When I was well controlled on MDI, I didn’t really want to go looking too far for input on how to change things bc diabetes care can be a set of dominoes, in my experience. And then postpartum, going off of birth control, everything went to heck and I HAD to ask lots of people for help and start on the pump and everything was new territory. That process lead me to making changes that I didn’t realize how badly they would negatively impact the things I was doing that WERE working for me. Bc when everything is chaos, how can you tell? But now that I have pinned down the basic framework of female hormones with help from exactly no one (but LOTS of help refining and exploring it on FUD)…and have pinned down the basic idea of my random “feet on the floor” guesswork in the mornings…and have pinned down leaking cannulas from large boluses…and pinned down improved pump performance with Novolog…I’m going to do my thing, keep observing, and block out what I don’t need at the moment.

THEN I might call Dexcom tech support. :grinning:

(Do I win an award for giving longer answers to simple questions than @Nickyghaleb??)


Maybe. That one was a doozy.


You get me. You live it ugly and I love livin’ it ugly right there with ya!


So, I worked through those emotions quickly…and relented and called Dexcom after another 40 point difference. (Dexcom was showing an awesome flatline after breakfast. Fingerstick disagreed by 40 points.)

Dexcom said:

  1. Rotate your sites more. Scar tissue can cause this.
  2. It should read within 20%. If it doesn’t, calibrate up to three times in a row 15 minutes apart.
  3. We cannot say to you any of the fixes you’ll find on the internet. We can only recommend what is in the user manual. (I feel like that was a subtle, “Go see what FUD says!” prompt.)
  4. If it is off by more than 20% and you can’t get it fixed, call us and we’ll replace it.
  5. It could be that mishandling by your supplier is adding to the problem. These are temperature sensitive but we cannot say for sure if that is the problem here.

Last shipment arrived in a normal box. The one before that came in coolers with dry ice. That variation in shipment method is interesting to me…but who knows if that is contributing or not.

I have used my right and left back for sensors for years. But I move them around and alternate sides. I wouldn’t have thought that would lead to scar tissue since Gary Scheiner told me not to worry about it…since it doesn’t have the same effect on skin as a pump that’s pumping insulin into the same spot for days on end.

Last year, when I rotated Dexcom sensors to my arms and legs and abdomen per recommendations on FUD, all spots other than my back were inaccurate. I don’t think I have scar tissue in all of those spots…sooooo…


This thing is a drunk nightmare. There won’t be any bg techniques with great finesse happening with this trainwreck.

Just aiming in between the goalposts from now on. Inside of that is anyone’s guess.


@T1Allison, this looks like a noisy track: this is, for us, a sign that the sensor is on its last legs.


Which is crazy! Because it is from my newest shipment, it is day three of this sensor, and it has never been restarted.

And I inserted this one in fresh skin…further around to my side than I ever normally go.

Random number generator, right now.


Yes but… some sensors turn noisy quickly for us too :frowning: Those are the ones that end up giving us trouble.