Moderator Note: This helpful reply was copied from another thread as it details some of the commonly experienced limitations of continuous glucose sensors.
If these are very short periods of lows and occur overnight, they are more likely to be compression lows - when pressure is applied to a sensor, it will report a false low. The reason I say this is more likely to happen at night is because that is when we could lie on our arm and squash the sensor. This could happen more in the hot weather because we don't sleep as well and toss and turn ... at least I do.
I am not sure if anyone shared with you the limitations of CGMs. As Libre is the most common one, these are often reported as "Libre limitation" but they are problems with physics, The only difference is how the manufacturers have chosen to deal with them.
Anyway, enough waffle, here they are with apologies if you already have this
- Some of us find that our bodies do not like have an alien object inserted into our arm. It take a day or two to "bed in" a new sensor. As a result, the first 24 to 48 hours after inserting a sensor could be more random. Some of us insert a new sensor the day before activating it to overcome this.
- Compression lows. Take care where you place your sensor and try to avoid the part of your arm that you lie on. If you get a low alarm in the middle of the night, check it with a finger prick before treating.
- "Normal range". CGMs are designed to be most accurate at "normal" BG. This is around 4.0 to about 8.0 and, if you see a value outside of this range, it can exaggerate the high or low. Again, the advice is to check with a finger prick before treating.
- Extrapolation. CGMs read interstitial fluid which will react to changes in BG about 15 minutes slower than blood. This is where I know some CGMs differ. Libre handles this by extrapolating the current trend to "predict" the current reading. If your trend changes direction in the last 15 minutes (e.g. when treating a hypo), the prediction could overshoot. Libre will correct this when it has "caught up". But at the time, it will seem like you are going higher or lower and taking longer to recover. Again, another reason for double checking with a finger prick.
- Faulty sensors. Not every sensor will be checked in the factory so there are some faulty ones in circulation. If you read social media, it will seem as if all sensors are faulty but human nature is to complain when things go wrong and say nothing if things are ok. The most common "fault" is inaccuracy. Therefore, it is a good idea to check the accuracy against .. .yes, you have guess it ... a finger prick. I tend to check once a day when my levels are stable and in that "normal" range I mentioned above. Don't expect exactly the same numbers (meter standards allow 15% inaccuracy and both could be out by 15% in opposite directions) but it is useful to give yourself confidence your current sensor is in the right ballpark ... and staying there each day. If it is out by a lot or if you get a "sensor failure" reported, in the UK, Abbott are pretty good at replacing sensor either via an online form or by calling. I do not know what that is like in Germany.
- Third party apps. Libre is "factory calibrated". Some of us find "factory man" does not represent them well. There are third party apps like Juggluco, Shuggah, xDrip+, Diabox, ... which allow calibration against a finger prick. These were very popular when Libre required scanning as they also converted the sensor into a rtCGM (real time CGM which did not require scanning). Some people still prefer them as they also allow things like integration with a smart watch. However, they may not update LibreView which could be an issue if you share that with your endo.
I think that is all and I have probably waffled for too long on something you may already know.