Just started Dexcom G6 - all readings incorrect

I have never heard that ketoacidosis, a life threatening event, occurs in Type 2 diabetics on or off a low carb diet. Getting into ketosis, yes, that is what enables you to burn energy from fat rather than from carbs and shed several lbs every week. I am on about 20-25 grams of carbs a day, so am in that category and initially my bg dropped like a rock as well and that allowed me to discontinue my use of Lantus as a basal insulin and I was able to cut my Humalog by over 70%. So in short, I got the weight drop I was looking for and dramatically reduced my insulin use which made me feel much better as well. I am now slowly increasing my vegatable and fruit intake to have a more varied diet and to stop my ever slowly increasing LDL from eating way too much saturated fats.

Your way of eating, I believe, will allow you to keep your BG’s well in control but I question how you will be able to lose weight on that diet. The 2 experts on this forum on your way of eating are Marilyn6 and Marie20 who posted a little higher up in this thread. I am sure they would either or both be happy to share tips if you want or need them.

In the meantime have a wonderful holiday!!!

@CJ114 Oh darn I used the wrong word didn’t I? Gahhh. Not ketoacidosis I meant ketosis so sorry!!! [headdesk] I made an update to my post.

Wow that is an amazing difference. Can you give me an example of a daily meal plan? I just don’t know if I could stomach the very low carb again. Maybe things have changed since I tried it.

I like you, @margiek. Your a high achiever! A photo album? Like a scrap book? That sounds fun.

I eat items like these and then supplement with Whole Almonds and Chopped Walnuts, and peanut butter combined with nuts and whipped cream to keep my calorie count where I want it. Just a few to whet your appetite. If you like what you see, there a tons of options on several different websites. I look at recipes I can make a 2 week batch to freeze and clean the pots, pans and dishes all in under 45 minutes. Cleanup usually takes longer than meal prep.

Someone (I think @YogaO ) told me to call Dexcom Care line and talk to the nurse about sensor inaccuracy. We talked about the 20/20 rule. It made a world of difference and I highly recommend it.

Dexcom wants their sensor readings within 20% of manual finger sticks (above 80) and within 20 points below 80.

They call this the 20/20 rule .

If it fails that condition, wait until its above 80 and achieves a stable reading (that is not rapidly increasing/decreasing), then calibrate.

We should calibrate ONCE. If we over calibrate (too many times), then the sensor is designed to fail.

They asked me to read them the last three data points. Of which, one failed and I should have re-calibrated at that point.

Dexcom CARE

Dexcom CGM training, software downloads, and tutorials


Monday - Friday 6 AM - 5 PM PST

Saturday 7 AM - 1:30 PM PST

(All hours subject to change)

Bacon Cheeseburger soup

If you live somewhere cold, you can’t tell me this doesn’t sound like great comfort food :yum:

Just over 3 grams of carbs per serving

1 Like

Sounds exquisite except for the saturated fats I am trying to avoid these days. Those are probably what makes this taste soooo good.

1 Like

Well this is just unbelieveable. I contacted Dexcom because after over a week the overpatches did not arrive. We also had to go to the endo office to pick up some for my remaining sensors. Dexcom refused to overnight them and said that USPS takes up to 16 business days. I believe they were sent on a Friday which means they would take A MONTH to arrive. And they are an essential part of putting the sensor on your body so why would they not even include them with the sensors to begin with? So I am going to use the very helpful advice given by @CJ114 and purchase the OpSite on Amazon and also the skintac. I will make my own overpatches and keep them in place using the skintac on the edges.

The advice and help you all are giving me is just invaluable.

I am disappointed in the Dexcom, first because they overstate that it eliminates fingersticks in their advertisements but then in the fine print say you have to fingerstick for dosing, contradicting their advertising. Also after 14 days it has not found any “patterns” in my readings even though the overlay shows a daily pattern. The summary information it gives me about what would be considered my equivalent A1C and my percentage “in range” are way better than my last A1C and what my fingersticks are telling me. If you go by that data I am already hitting the goal, which is very far from the reality. So I am wondering if I even want to get another set of sensors and run out the transmitter life or just bag it.


Some people need skintac and others never do. I have never had to use skintac so don’t keep it around as that is just an extra item for me to worry about. Diabetics do not need any extra worries!!! The secret for me has been to shave the area where the sensor goes and then take a shower, dry area with a hair blow dryer and apply the sensor. During many months of the year I Inline skate 25-30 miles on nice days and when it gets really hot can get pretty sweaty and still don’t need the skintac. If I really need extra firm adhesion, I make an under patch first with OpSite and place on skin, put the sensor on the under patch and then make an over patch to go over the transmitter. I rarely needed that much adhesion, but that worked the best without any additional products such as skintac. Each person needs to find what works best for them.

@Marie20 Hi I have some questions. How can you wear a sensor longer than 10 days (is it a different brand than the G6?) - my receiver is telling me that my sensor session is going to end because it is at the end of the 10 day period.

Do you put the Skin Tac on the adhesive that is on the Dexcom sensor applicator after you peel off the two strips that protect the adhesive? Or did you mean that you put the Skin Tac on the overpatch adhesive? If I make my own overpatches with the product that @CJ114 recommended do I put the Skin Tac on that before I apply them over the sensor, or just if the edges peel, and will the Skin Tac work with that product too? We will go to the drugstore tonight to try to find the Skin Tac and I will first test it on my skin to make sure it does not irritate my skin.

You are supposed to only wear it 10 days, but I find it very accurate to wear it as long as possible until it goes wacky. That’s usually about 26 days for me, the longest being 38 days.

But it’s a hack people have figured it out. Dexcom keeps trying to stop it but so far someone keeps figuring out how to keep doing it. It can be really easy but it depends on your transmitter number. The latest transmitters require you to take out the transmiiter for at least 30 minutes and then restarting.

Easy way with older transmitters, doesn’t work a lot now though

Newer method working well.

I don’t use an overpatch, they didn’t work well for me, they came off right away. And once I used the Skin Tac it worked so well there was no need to use or try to use an overpatch.

This is what I do

  1. Unpeel white paper from sensor
  2. Apply Skin Tac all over the adhesive on the sensor making sure to get to the edges, but avoid the middle area that’s outlined that has the filament
  3. Put the sensor loader on the skin gently without pushing on it, hit the button without pushing into skin. If you push it will crumple the adhesive
  4. Let it set just a little bit to make sure it is adhering to skin
  5. Run your finger on top of the adhesive to make sure it is nice and flat and adhering to your skin
  6. If you applied a lot of Skin Tac you might want to hold it on for a few seconds to make sure it’s drying and adhering okay
  7. once the adhesive is nicely dry insert the transmitter

Then I touch up when it starts to peel up under the edges with a Q tip and more Skin Tac, making sure I try to only put the extra Skin Tac under the edges


just a plus one on the Alcohol Prep and SkinTac combo, although I’m currently using the 14-day Libre. Using it directly on skin, not leave any clear spot - just daubing a circle the size of my sensor. Used to knock them off regularly from the back of my upper arm (because I’m broad shouldered and tend to knock into door jams) trying other things and the SkinTac has been a savior.

1 Like

I have had very good correlation with my dexcom.
I wear it on my arms.
I check it with a finger stick now and again but usually don’t have to. It is pretty spot on.
You might want to check your meter to be sure you are calibrating to the correct readings.
They sell meter check solutions at the pharmacy. A low 60 and a high 200. Then you will get an idea how well it is performing.

Are you supposed to calibrate if it’s within 20% or outside 20%. 20% can make a difference between perfectly in range and being in hypo state and the alarms blaring.

So far my readings are roughly 36 mg (2mmol) different. Slightly outside the 20% for some readings. I’m finding its trending in the right direction (so moves with my BG rises and drops) but the discrepancy is enough for me to turn off my basal iq.

My DE says avoid calibration was the instruction given to her from tandem reps.

Just to prove my point it’s 1 am and my alerts are going crazy. The sensor reading is 3.4 mmol (61mg) and my meter reading is sitting nicely at 4.9mmol (88mg). I had to get up to get something to eat even though I’m in a good place just so the the sensor reading creeps back above the line and the alerts can be quiet so I can sleep! Grrrrr…

I’ve had my sensor on for around 40 hours. Is it time to call support?

I refuse to treat a phantom hypo with carbs just to silent a false alarm, but that’s me and I’m not using a Tandem pump as my Dexcom receiver. In your situation, I would use 1.1 mmol/L (20 mg/dL) to calibrate and then turn off the audible alarms, if possible. Is there any way for you to quiet the non-urgent low (55) alarms with the Tandem pump?

I find that the G6 does not like any calibrations that attempt to make a large correction. If I keep it to a maximum of 20 mg/dL (1.1 mmol/L) and maybe do two calibrations to step the sensor number up to a correct value. It’s more of a nudge tactic than a jump one.

Since the G6 sensor is notorious for first-day inaccuracy, I’ve taken to inserting my sensor 12-24 hours before I mount the transmitter and start a new session. In this situation, the best time to start a new session is in the morning so that any inaccuracy drama occurs during non-sleep daylight hours.

If I was faced with a pump that wouldn’t let me sleep due to repeated false alarms, I would be tempted to turn the pump off and replace the missing insulin with Regular or NPH.

This is a more drastic solution and would require some personal testing to determine safe doses for you. You could simulate this by doing it during the day so that you could monitor closely to observe what happens. One added benefit of trying this is that it will give you a basic tool to use when your pump fails. It’s survival education.

1 Like

I have no idea why some people refuse to calibrate. I calibrate when necessary and it behaves beautifully after that. You just have to make sure you’re in range and flat lined (BG, not EKG!) when you do so.


In this case I would just calibrate it, following Dexcoms 20/20 rule.
Possibly do another finger bg check based on symptoms if feeling it is lower.


Based on the sensor reading of 3.4 mmol (61mg) that’s considered urgent , so to my knowledge cannot be silenced below 3.9mmol (70mg). This is annoying when it’s a false alarm. I’ve acknowledged the alarm on the pump but I’m actually not sure whether it will continue beeping if my sugar doesn’t rise after the acknowledgment. Does anyone know?

I must say the dexcom sensor is already far better than the Medtronic. Although I still haven’t achieved accuracy, the direction of BG seems correct. It also feels a lot more comfortable and has very little pain. Just wish the accuracy will snap into place.

I’m actually not sure but my DE said that was the training she got. I read somewhere that the calibration mucks up the algorithm on the sensor. But if it’s not accurate in the first place I don’t see what can go wrong? Perhaps affect the sensor life? I don’t know