Losing my mind with CGM Inaccuracy!

Hi All-

This is not new to any of us, but am losing my mind with frustration. For background, I’ve lived with T1…specifically since the day Spiro Agnew resigned as Vice President…obviously a long time. I’ve grown up with the advances in techniques, treatment regimens, and technology. My frustration is simply this. I’ve used both the Dexcom 6 and mores the Medtronic CGM as part of my 670G. I’m educated, I’m experienced, and I’m also a certified Critical Care Registered Nurse. I understand technology and I am experienced with trouble-shooting as well as following “rules” and understanding “work-arounds.” My experience with both CGM versions is simply that in roughly 3 out of 4 sensors, I experience sensor failure…always due to absolute inability to be accurate with variance of between 50-300%, and often trending the wrong direction within 6-12 hrs of placement. I’ve remarked to my endocrinologist, who I frequently see at work for curbside consulting, that I would not allow this technology to operate in “Auto-Mode” to care for a dead animal let alone a human being.

Okay, I’ve introduced the origin of my frustration. Does anyone have any solutions for the following general scenario…solutions that are not the obvious as presume I have used any obvious solutions already. The General scenario is this…I insert the CGM (assume Medtronic version for this discussion), The link to the pump is flawless, I usually track my BG without calibrating for the first few hours (although I have also experimented with calibrating more after the warmup period without success), I calibrate according to instructions and pump indicators appropriately, and within the 6-12 hr period it is more often than not the SensorG is trending the wrong direction or simply errant by extreme percentages based on actual BloodG measurements. This is of course followed by Calibration error and sensor change instructions and sensor failure. I’ve experimented with starving myself for an entire day to avoid BG spikes etc to create and induced BG regularity for the sensor to learn from but this too is pointless.

Again, can anyone help with anything that is not just typical suggestions…e.g. rotate sights, follow the instructions, etc.?

Thank you in advance for any new help:)!

I’m going to assume that proper insertion techniques and site selection are part of the obvious things you’ve checked. You’re cleaning the site properly and letting it dry, you’re aiming for fatty areas while trying to avoid blood vessels. And that the transmitter you’re using has the latest software update. (It is possible to get overconfident about these things and miss what should have been an obvious detail. It’s happened to me.)

Any medications that could be interfering with the CGM readings? Tylenol can, among others.

A Medtronic tech told me that lack of sleep and proper hydration can affect body chemistry and screw up the sensor readings. Starving yourself could actually make it worse, especially because I’ve noticed the Medtronic sensor is confused by readings that are too stable. It’s possible that drinking something like Pedialyte could help.

Calibrating too much or too early can also be a problem. For some reason, the Medtronic pump is programmed to set off a “calibrate now” alarm much sooner after a failed calibration than Medtronic instructions say it should. Make sure to give it at least an hour, preferably two.

Here’s one to try: In the pump settings, turn the sensor off. Insert the sensor at bedtime. Don’t connect the pump and start the sensor until morning. The extra time “wetting” the sensor when your sugar levels should be fairly stable might help it be ready for calibration.

Is the adhesive irritating your skin? If you’re having a reaction to it, that could theoretically throw off the readings. I know the Medtronic oval tape always irritated my skin enough to make the whole site red for days.

Is there a way you can try a different CGM? Medtronic sensors never worked properly for me on the first day. You’re supposed to calibrate 2 hours after insertion and then again at 6 hours. I followed that and it never worked. I always had to calibrate again at the 12 hour mark before it would go into auto mode and start functioning. But I switched to Dexcom, and that’s almost always been spot on. (Obviously, it wouldn’t work with the pump. But just a way to test if something specific to the Medronic sensors is somehow incompatible with your body chemistry or something, and if you can get reliable readings from an alternative.)

I’ll keep thinking. But I hope you find an answer. That sounds incredibly frustrating.


hi @bogey2300 you know, CGM just doesn’t work for some people. The sensor is supposed to sit in interstitial fluid, and so I believe that certain body chemistry is just going to be bad for CGM.

as stated above, calibrations for any Medtronic sensor is a trick. I would only EVER calibrate after 4 straight hours of nearly constant blood sugar and it would work “ok” for me thereafter. You may realize that there is a 5 - 25 minute lag between a finger stick blood sugar reading and your interstitial fluid, so if you calibrate when you bs is changing you are in big trouble.

DEX just plain reads low for the first 12-20 hours. I ignore it, sometimes even shut off Bluetooth so the darn thing doesn’t scream at me “YOU ARE ABOUT TO DIE” with that awful alarm. after the first day, I am getting exceptionally good results through to day 10. again, and as with Medtronic, I would never calibrate it when my blood sugar is changing. comparison with a finger stick HAS to be with an understanding of the interstitial fluid lagging reality, AND that the difference between a finger stick and a CGM can be 25% of reading (fingerstick 200, CGM could read 150 mg/dL and STILL BE GOOD. many people do not grasp accuracy as a concept, and it applies to BOTH home devices.

sorry but the answer you are searching for may not be whet you want - a method or instruction to suddenly make it work for you. still I wish you good luck!

Have you tried to use a calibration that is half way between the BG and SG if the readings are more than 30 percent different? Watch and half way calibrate again in 8-12 hrs, if it is more than about 30 percent off and it may follow your actual BG more accurately from then till the end of session. It has worked on occasion but mine usually follows closer than that.

This probably won’t be helpful, but I just wanted to say I too had a lot of the same frustrations with Dex G6. I didn’t find any solutions, I just gave it up. Finger sticks are working even better for me, in terms of frustration, anxiety, and hypos. I will say that for the six months I did use the CGM, it was totally worth it, because even though it was too-requently inaccurate, including often to large degrees (including direction trends), I used the 70% (or whatever) time it WAS fairly accurate to learn a lot about how stuff worked for my body, lifestyle, food, Lantus, and carb ratios. So glad I used it for the time I did, and so glad I was able to give it up!

1 Like


I have not tried this idea. I understand why it may work so I may do that just for experimental state. Also just for the sake of discussion in the real world, isn’t it sad…truly sad… that a company as big as Medtronic can mislead the public and the healthcare system so openly with the reliability and accuracy of a product​:smirk:. (remember I’m a critical care certified RN with two educational degrees and 30 years experience in healthcare as well as 48 years experience with my own disease). I appreciate you suggestion. It’s just pathetic that it’s the equivalent of clutch starting a car with a dead battery…:smirk::disappointed:.

Thank again though.


Anyone’s suggestion and experience and insight is valuable. Yours reinforces that it not just my problem and that the system is horribly flawed and unreliable to a much greater degree than is represented by either Dexcom or Medtronic. They even have endocrinologists now stating finger sticks “aren’t accurate”. REALLY… they may be different than a blood draw because they are capillary and not peripheral blood, but they are never in accurate nor off by more than 100% when done correctly.

I appreciate your comments:)

Definitely feel the pain. I feel like I get one solid sensor per box. I then get replacements and the replacements work well, but the actual hassle of it when they don’t work really bothers me. I feel like I finger stick more than I did without one because when they’re off, they’re off. Today I had one that did not show a blood sugar climb, which caused some ketones. Thankfully I caught it fairly early since it was after a workout, but was kinda annoyed that my cgm showed me at 150. It also didn’t register my correction bolus and stopped again at 160 when I was really at 119. Was able to get a replacement but not fun!

I used the medtronic sensors with a 670g for roughly 3 yrs. I found the performing a calibration outside of bg levels of about 90-180 caused me the same issues you have. I also avoided calibration when I knew I was either falling or climbing rapidly. I also personally sweat a lot so in summer time I really struggle to consume enough water to stay hydrated and found even slight dehydration caused errors. Good luck and I feel your frustration. I did get kicked out of auto mode frequently because I knew a calibration when the pump asked would screw up the sensor readings.

Thank you Jason! I appreciate you taking time to give me feedback. All who have I greatly appreciate your help

1 Like