Dexcom g6 sensor accuracy drops ~day 10


Could the restart impact accuracy?
Yes. IMHO it is definately possible that the restart could impact accuracy.

DOES the restart impact accuracy? That is a good question. But I would absolutely consider that a possibility.

We have only done a G6 restart I think three times? I did not record good data on those. But I do recall on one of those, I checked the G6 against a meter prior to restart and it was good. After restart, it was bad. I didn’t record numbers so I don’t want to guess.

The other two restarts, I don’t recall enough to discuss even.

I would suggest if you are planning a restart, take a couple of meter checks during the day prior to the restart then a couple meter checks after restart. Any change of significance? I would find that interesting data. Obviously one sensor and one session is just that - one. But like anything else if you record the data over a number of times then it is possible you could start to see a particular pattern which then becomes even more interesting.

1 Like

Will do :+1: (also I just ordered a contour next meter in the mail so I’ll check against 2 meters in days 9, 10, and 11 (before and after restart))


This sensor I’ve had in for the past 3 days (that I have not manually calibrated) has been consistently higher than fingersticks by ~15 (when I’m flat and not trending up or down).

I’ve been staying hydrated and willing myself not to calibrate just once. Any suggestions on pushing the sensor towards getting tighter with my BG since you’ve had pretty good luck with BG matching SG, @Thomas ?

The 15 point difference doesn’t bug me too bad, just when I’m chilling in the lower range, dex is telling me I’m 80 and flat, and am starting to sweat and get low symptoms

1 Like

My question was about to be whether the meter was accurate. But if you have symptoms that confirm the meter - that is the answer IMHO.

Yeah - I would calibrate. That makes sense.

Obviously picking proper time for calibrate when you are straight and level on the cgm is critical.

If it was me, I would try:
Leave it in factory calibrate mode - no need for a restart.
Give it 4 calibrations and see if that pulls it into alignment with your meter and your symptoms. The BG number is not as relevant (assuming between 400 ~ 400 of course) as the cgm being straight and level.

Today, 2 calibrations approx 12 hours apart.
Tomorrow, 1 calibration
Next day, 1 calibration

One would hope that each calibration would pull it closer and after the 4th calibration, it would be looking good and no need for anything further.

Obviously (in case anybody else reading) this is VERY specific to a Dexcom G6 cgm system.


With my shiny new contour next one meter that arrived today, next to my freestyle flash that is 16 years old I think, LOL! Dexcom sensor was reading flat at 123 at this time.

What’s the reasoning behind the 4 calibrations? I think I’ve heard that elsewhere but have never done it before. I’ll try with this sensor :slight_smile:


I’m curious, too, @Thomas. Tandem describes a different way to use calibrations in order to tighten the spread. They recommended calibrating up to 3 times in a row, 15 minutes apart, using fresh meter readings for each calibration. I’ve never done anything like it as there’s been no need, but what do you think about that approach?


I think that is for a different scenario. I would have to read the manuals (Dexcom and Tandem) but I think that is more for when the cgm reports an error and demands the 15 minute calibrations. Twice I had that happen but neither time was successful and both of them (G6 Sensor) failed and were replaced.

The situation with @LarissaW sounds more like a well functioning cgm that only needs a bit of a bump. Not a complete calibration reset.


Interesting. I don’t doubt you at all, but the rep was suggesting this in response to a situation like what @LarissaW is describing. We were just talking about general tips and tricks… nothing about any kind of complete calibration reset. With all that being said, I’ll take something you recommend over a rep’s recommendation any day, so what do you think about doing this just to improve accuracy?


Depends on how far off it is and if it is consistently off by a linear amount.

If the thing is jumping up and down then IMHO a bad sensor is a bad sensor is a bad sensor. No amount of calibration will that.

But something like what @LarissaW describes really sounds like a relatively minor calibration issue. But also on the edge of what can be corrected with calibration. You (@Nickyghaleb) had mentioned a 10 point difference. That sounds potentially too small to be resolved with calibration. Trying to correct for that may have you chasing your tail to no avail.

Doing a triple calibration reset for 10 or 15 points would IMHO be like going after a possum with an elephant gun. All you really end up with is a mess. Strictly my opinion.


So on day 5 now. I did the 1st calibration last night with the intent of doing another this morning. However, it seems the 1st calibration was enough. Now the dexcom is only off by 2 or 3 points compared to several meters. Still recommend following through with the rest of the calibrations that you’d recommended above ?

1 Like

Probably not. If running factory calibration, I think less is more. You are adding a calibration to the factory calibration, not replacing the factory calibration.

If doing manual calibration from the get go (ie - NOT putting the sensor code in) then yes, you should at least follow through with the calibration schedule for the first two days including a minimum of four calibrations.


I’m on day 13 of first G6. Used the calibration code for first 9.5 days, then restart. On restart, decided to use with calibration. As suggested, did enter several calibrations the first 2 days.

Also wearing G4 (now day 4), and interesting to see often the G4 and G6 might be 20-30 points apart, and meter BG is in middle!. Also, the trend arrows on G4 are more sensitive, eg, G4 shows single slant, G6 shows flat, but then goes slant. Or G4 is double up, and G6 is single slant. With corresponding numbers, but eventually they match up. (G6 left thigh, G4 right thigh).

Today G6 had a 2 hour no data, while G4 was fine. But G6 back on target.

To be continued…


Here’s my results (sorry they’re not pretty I just kept track on a note on my phone:

Day 8

Dexcom g6: steady at 73
Freestyle lite: 79
Center health nano: 73
Contour next one: 76

Freestyle flash:114
Dex:108 and steady

Dex: 116 and steady
Freestyle lite: 118

Day 9

Dex: 86 steady
Freestyle lite: 83 and 87 (tested twice)

Dex: 78 steady
F lite: 83

Dex: 97 steady
F lite: 88

Day 10

Dex:137 steady
F lite: 152

F lite: 86
Dex: 85 steady

2:07 pm
F lite: 157
Dex: 147 steady

Expiration should be 8:06 pm


5:05 pm
F lite: 55
Dex: 89 steady

Dex 142 steady
Freestyle lite : 79

Dec 101 steady
F lite: 65

Day 11

Dex 105 steady
F lite 93

I have NOT manually calibrated this sensor since day 5 and I didn’t take Advil and I’ve been staying hydrated. Sensor accuracy was great from day 8 and 9 and beginning of 10 but after the restart was wonky, then returned to normal day 11. It seems like the dexcom restart messes with accuracy temporarily. Sorry I would’ve done more finger checks for the time frame after restarting, but I was out with friends. I’m happy to keep a record log again for next restart/next sensor


We don’t often get to do a G6 restart due to not typically having good solid data at day #10. (More often then not, we don’t even get to day #10 period.)

This sensor looked solid and a meter check during the day (#10) showed it still providing accurate data.

So, we have restarted and will of course do a meter check as soon as the 2-hour warm-up finishes to see how it is doing. We restarted with factory (code) calibration but will enter calibrations if the cgm is different from the meter (Contour Next One) by an unreasonable amount. (Not sure - maybe > 15 points different?) I will wait to see what the number are.

EDIT: After warm-up. CGM=94. Meter=73. Calibrated.
EDIT2: Next morning. CGM=74 Meter=85. Calibrated. (closer but flipped)

For 11 complete days - The graph seems more solid than typical. Still no jitter/shake. (This is on the non-smoothed graph from old Follow app). This is the same sensor from our previous order with exp date end of October '19 so not a new lot or anything. Just a lucky sensor/location.


Thanks for following up. I’ll be interested to see what your numbers look like after restarting. And to be clear, I didn’t calibrate after restarting and the accuracy came back. So I’m not sure if waiting it out is an option for you (or one you want to do since I know it’s personal and you may be using the dexcom for suspends etc)

1 Like

Yeah - you are correct that we use it for suspends. With the Basal-IQ, more than 20 points off is a pretty big deal. Even more so with the meter reading lower than the cgm. If that 21 point spread were consistent and the Basal-IQ had the cgm at 75 overnight (which is typical to be in the 75~85 range per cgm consistent through the night), that could mean that the meter would be reading 54.

Too much to let it go.

1 Like

And before the restart you said it was accurate (which for you guys is within a couple points, right?)

So I’m not alone in the restart causing//leading to a correlation in a substantial deviation in BG vs SG?

1 Like

Within 10 points. At any range. But not at any slope.

I certainly don’t understand the details of the technology involved. But this totally makes intuitive sense to me.

1 Like