Just started Dexcom G6 - all readings incorrect

Hi I just joined the forum. Long-term Type 2 on insulin for about ten years now and may be LADA now, endo is going to check at follow-up. In my 60s. I started the Dexcom G6 and what I found what that while it may be good for patterns, as far as actual bg readings it is always way off from my meters, over the 20%. It didn’t even record some postmeal high curves but just flatlined. All the Dexcom readings are much lower (for instance, this morning delayed fasting - in which my bg will climb considerably until I eat breakfast - fingerstick was 183 while Dexcom was 139. This was very disappointing to me because it was the data I was most interested in, not reports and so forth. My POC at the endo clinic used the code to calibrate it and I think this was the issue. Next sensor I am going to use manual calibration and see if that makes a difference,. Has anyone else had this issue?

I did some research online and found that generally alternate sites do not do well at determining bg, so I have to assume their code-based calibration algorithm is crap. I would really like to get some use out of this, it was expensive, and hope that manual calibration works better (I will post results when I have changed sensors).

Also I find it ironic that Dexcom advertises no fingersticks when they then say don’t rely on it for insulin dosing, fingerstick for that. (I don’t have an issue with fingerstick, I use the sides of my fingertips which causes no issues)

1 Like

Are you well hydrated?
Where are u wearing it?
What meter are u using?
Have you tried checking CGM readings 20 minutes AFTER a fingerstick, to see if it has caught up?
Ever worn any other CGM?

Hello. Yes am well hydrated. Am wearing on back of my left arm. Using One Touch Ultra Mini (I have four of them and sometimes I will use more than one to check - they are calibrated fine). Yes the CGM seems to flatline or stay “behind” my fingerstick readings. No this is my first CGM.

Would you recommend manual calibration? That was the first thing that occurred to me.

I’d hold off on calibration.

First of all, do u have a Contour Next meter (any model)?

Try the tummy area, at least 3" away from belly button.

Do not be discouraged by your first sensor. It takes a while to get used to how they work. They have a delay of 10-20 minutes, so keep that in mind at all times. When the readings match a finger stick, it is because your bg’s are relatively stable, you have a good sensor in a good location and are using a Contour Next meter. One Touch meters are NOT accurate across the entire bg range.

Don’t take my word for it re: meters. there have been studies done on many meters and the Contour Next came out on top. Thankfully, that is what Dexcom and Medtronic both supply.

My new insurance will only cover Accucheck so I was going to change from One Touch anyway when my strips ran out. They will not support Contour Next meter so I will look into the cost of the strips.

I have tested all four of my One Touch at once a number of times and they were all very close.

Location can make a big difference.
I cannot use my arms at all.
I put mine out towards my love handle area of my abdomen. This area works GREAT for me.

I would resist doing calibrations, especially in the first two hours.
Also, try ‘presoaking’ the sensor to help moderate your first 24 hour readings being all over the place. Presoak just means insert the sensor a set time ahead of when you actually insert the transmitter into the sensor. I put a sensor on the night before, and swap my transmitter over to that new sensor in the morning.

Also, I know you said you are hydrated, but drink MORE. Sometimes it takes a bit more to keep your interstitial fluids moving the way the sensors want. It won’t hurt, and is easy to try. I know I have to stay very well hydrated to get my get results.

Thank you I just did read and learn about soaking the sensor. I will try the abdomen location for my next sensor. I will also hydrate more and see if that helps.

Should I use the code for calibrating the next sensor or should I do manual calibration?

After the first 24 hours, do not be afraid to calibrate. Calibrating is not incompatible with the initial 4-digit code but an improvement on it.

The clinic put in the initial code BUT 20% variances after that can be calibrated in the coming days. On the G6 receiver, hit the button, 1, 2, Main Menu, Calibrate.

1 Like

Hey that is great to know! Ok so is there a specifc time to calibrate, I am still post-meal, I would guess when readings are close to flat?

That is the only time to calibrate ANY type of CGM. The flatter the better. Otherwise you will throw off the calibration.

Thanks so much - you and everyone - for the feedback. The point of contact at my endo’s office thinks she knows the Dexcom but does not. She told me entering the sensor code has nothing to do with calibration. I am having to bootstrap myself on this and I really appreciate the feedback.

Entering the sensor code gives you a starting calibration.

If after the first day you find it is consistently off then you should follow up with manual calibration.

20% off isn’t too bad.

Ya, One Touch are crap. It took me a few decades of using them until I went to CGM and could not get close to same results. One Touch always read way off from CGM. Contour Next meter dead nuts on most of the time. Read the accuracy comparisons (MARD) between meters and you will see this is not my opinion but proven out as well when the meters are tested against one another.

1 Like

Thank you I will see about getting the Contour Next meter and strips.

I am one of the few that think my One Touch Verio is okay! The one I use the most is within 5 points usually of my Contour Next One.

If you want a Free Contour Next Meter you could call Contour and tell them your usage and they used to give you a free meter. Or on ebay off and on they have someone that sells the strips at a good price and you get a free meter with it.

But I do believe you will benefit from calibrating your Dexcom. Just don’t overdo it, it messes with it. But I have found more success at calibrating at the same numbers I am at most of the time or want to be at. For me that is between 95-110. I have found if I calibrate it at 100, it will be off by a lot more at say 160? Since I rely on most of my dosing at the lower number I want it more accurate then. If I am too high I know that I need to be “fixed”, the same with too low. So I’m not sure I care about exacts as that can vary anyway. I like my Dexcom to be within 5 points of my actual readings on a meter at that range.

I only notice the Dexcom missing a few numbers if there was only 1 or 2 that were higher or lower? It then levels out the readings. So I’m not sure if you mean it’s missing a bunch of readings or just a couple? But I have found it invaluable! Setting the alarms before I go out of range has helped tremendously.

And always keep in mind the Dexcom is behind the meter anywhere from 5-20 minutes. I like to use 10 minutes as my guideline. When you have a sideways arrow and are close to your average BG, take a meter reading and set a timer for 10 minutes and then another 5 minutes and that will give you a better picture of how accurate it is.

If cost is an issue, Walmart caries their own brand of meter called the Relion PRIME which is about $9 for the meter and $9 or so for 50 strips. They are not as good as the Contour Next but pretty close and definitely better than any of the one touch line. No prescription needed, but make sure you get PRIME and not any of the other Relion meters/strips. See https://www.walmart.com/search/?query=relion%20prime first 2 entries. Note the other relion versions. Only prime is the one that scored well.

The Diabetes Technology Society tested 18 blood glucose meters in 2017 and only 6 meters passed. Here are the six.



I have the Accu-check Aviva plus. It does a good job.
I have thought about changing to the Contour Next, but I am not sure it’s worth it.

I read the MARD report which was very helpful. It appears long-term if I wanted strips my insurance would cover I could go with Accu-check Nano or Aviva Plus. Also I went ahead and purchased a Contour Next with free next-day shipping from Amazon (I have Prime) that was a complete package, including control solution and a storage case, lancets, lancing device, meter and also had 100 strips for around $50, to try it out, so I will be able to try it starting Saturday. And this will be a good way to check against my One Touch meters as well.

Also I calibrated the Dexcom against a fingerstick and after it settled down it was within 5% of the fingerstick reading I used for calibrating, which is a huge difference from being 25-30% off.

Thanks everyone again for all the help. I am excited to be able to have the Dexcom and see what it can tell me about how to manage the diabetes. I can see I have a lot to learn.


OK, I didn’t read the thread, but am going to attempt to get down to biz on this with you.

I rarely see a Dexcom reading that is higher than the finger stick. It reads ‘low.’ Often by 50 points. But, my BG machine also tended to read a little higher than my G4. Although, not to this degree. So, I imagine actual BG is not as high as the machine’s value, and not as low as the sensors. I favor the machine reading.

That being said, I don’t think that there’s a gigantic difference between 183 and 139. You want to be on the lookout for errors that are much, much higher than that. When you see readings that are 50% different than one another, or higher, consistently, I find that is a good indication of sensor failure.

I believe that the complexity they introduced into the device, to accommodate no calibration, is working against its accuracy. I have also hoped to test it by not using the code and perhaps get around that subset of functionality. There have been some suggestions that might bypass the code, but I have not tested it. In fact, calibrations have made the thing read pretty poorly, for me, in the past. So, I was hoping to try not entering the sensor code eat the start of the process and see if that helps.

Please post data. Please message me directly if you find anything out. I have too much data to post here. But, my findings have been consistent with yours.

You may have missed the ‘good old days’ of Dexcom. It really used to be an extraordinary company with an exceptional product.

This is the type of error indicating sensor error:

I do not see error consistently. It is somewhat random. That makes it hard to identify. I recommend testing over multiple sensors, over the course of several weeks. You will see it.

Calibrate correctly, or it will mess up the sensor.