30-day trial

I'm a couple days in to a 30-day trial of the Dex-Com, and I'm just not sure what to think. The rep said to expect a typical variance of plus or minus 20% -- that seems like a HUGE variance to me! If it says 200, I could be anywhere from 160 to 240???

I don't understand why something that is injected into my body can't be more accurate.

Does anyone have any insight into this & is it worth putting up with such inaccuracies?

Yes, Chris, I feel the same thing you are saying and I've been at this 3 months now. They say it takes about 3 months to adjust to it, make it work for you, so I'd say don't give up yet. I've been ready to pitch mine many times, but too many people I respect love it and say they couldn't be without it. Be sure to connect with the folks with more experience and I'm going to go to a couple classes offered in my area, too. Still, I'm sensitive to insulin, take very little actually, and that much variance makes a big difference on how I'm going to handle it. I end up testing all the time anyway - nearly as much as I used to. But Dexcom told me I was calibrating it too much and confusing it even more. I have complained to Dexcom that many many times now the Dexcom has been even much higher than the 20 percent difference. It says 260, my meter says 160 = and it gets worse in the very high and the very low ranges. That's when they told me I was calibrating it wrongly. I'm tired of being alerted that I'm low in the night when I'm perfectly fine at 103. Dexcom techs remind me that our meters also have that much variance. I know Dexcom is working on this for the next model to come out later this year (?), i.e. the accuracy factor. I hope so. Even the 10-minute lag is much longer for me, more like 1/2 hour. And all my doctors and friends here keep telling me to see it as a TREND device. They seem to think that the trend picture is very valuable, but I keep saying, how can it really be of that much value when it's based on false information? Back up, they say, get the bigger picture - it's telling you your sugar is rising and you might not have realized it and need to deal with it. And so many here, including my dr who is on it, find it so valuable, so I keep plugging to try to adjust and you may find it takes time, too. My dr wears his transmitter much higher on his abdomen, near his rib cage where there is less fat, and he feels that helps with stability. He has found his amazingly accurate. I have scar tissue I stay far away from. People here say calibrating it ONLY twice a day and especially when it's near the value you ideally want it to be kind of helps it situate closer to your waves. I hope that helps some. I know a guy who was a perfectionist and dropped his Dexcom for a couple months, then tried it again and dealt with it better. Just give it a whirl. It's the best we've got so far.

This is simple to explain. 1) a finger stick BG is ±10% compared to a lab test where the blood cells (eaters of serum glucose) are spun from the plasma almost immediately after collection and then sent for testing by "gold standards". Deviation here. 2) interstitial fluid (IS) [fluid surrounding body cells outside the circulation] is what the Dexcom or any other CGM uses for "testing". With the variation in finger sticks and the variation between blood and IS is nearly the same when BG is very level, however we are back to the finger stick being ±10% which calculates to a CGM reading ±20% compared to a lab test glucose. Dexcom does a great job of explaining the math behind the CGM in the documentation.

Most of the time, my Dexcom is within ±5% or less when I test. Pearl of wisdom: Meters and CGMs are tools to help you regulate your "sweetness". Just like a hammer or any other tool, you are the craftsman and must the quirks of the tool in order to use it effectively.

PM me if you want to discuss further. As of Sept 2012, I have been using a Dexcom 7+ for three years.


My daughter's experience with Dexcom has been very positive. A few weeks ago, she finished her first quarter on the CGM. Her endo was extremely pleased with her A1C reduction. The CGM allowed us to see trends that would not typically show up via finger sticks. One example is that we found out that her biggest postprandial spike occurs after breakfast. Armed with this knowledge, we changed her insulin to apridra and try to give her the breakfast bolus 15 to 30 minutes before her breakfast.

We find overall that the vast majority of the times that the difference in readings is not very large. Just a few minutes ago, she tested herself before dinner and the blood stick showed 104 and the CGM 99. For her, the greatest discrepancy is when she is experiencing a rapid change in BG. In those situations, we know that the CGM tends to overstate the high or low.

It is important to realize that test strips also have an "accepatble variance" of around 20%. Hopefully further advances will occur that low this amount of variance.

What the others have said is exactly right....the sensor technology is not that good yet. it is best used for trend information. It SHOULD NOT be used for dosing insulin. Always do a fingerstick for that. It is fairly accurate in trends - if it is going up , you probably are going up, if it is going down , you probably are going down, and similarly if the line is flat. I found it very useful for understanding what was happening overnight while I sleep and making appropriate basal adjustments. It's also good for seeing how you react to food and adjusting the timing and dosing of the insulin to reduce spikes, as was pointed out above.

The fact that it is testing interstitial fluid instead of blood introduces another factor - delay. The interstitial fluid's glucose lags the blood by 5 to 20 minutes.

Lack of a good CGMS sensor technology is one of the main things holding up a good artificial pancreas - it would be dangerous to close the loop and have the CGMS control the pump if the sensor technology is way off. We are getting there and each generation of sensor is getting better, but it's by no means perfect.

THere is a good table in the back of the manual that is eye opening about how far off the sensor can be and how bad the false alarm rate is at various high and low settings. If you understand the CGMS' limitations, it can be a great tool for fine tuning.

Thanks for all the responses so far!

I'm excited about the idea of seeing trends, but feel like if I can't trust the accuracy I'm not sure I can trust the trend.

Also, it feels like a big deal to me to have to inject something into my body & I want better results if I'm going to be doing it regularly.

I will keep at it I think and hope for the best, but I'm quite disappointed at the gate.

it is a huge difference and annoys most of us. The truth is that although the difference is big, most of us have much greater control with the Dex than without it. I would not give it up.