Does the performance of CGMs degrade during their usage?

[mysorian]

For example, I have used a CGM which should expire after 14 days of measurement. The question is, is the performance uniform and constant during this 14-day period? I am towards the end of this period, and I am seeing lower and lower values but not alarmingly low. I have not made a detailed analysis, but it looks like this. I had no lows crossing to the red zone in the first 10 days and I have seen a couple of border crossings lately. After two more days, I am going to take a detailed look.

 

[EllieM]

Hi @mysorian

As a T1 I usually check with a glucometer when I'm uncertain of cgm readings (dexcom). I generally find it is least accurate in first 24 hours (tends to read low for me) though I do rarely notice it reading off on the last day.

To be honest, if in doubt you need to use a meter...


But are you on any medication which might cause a hypo? Also, be aware that the libre is known for occasionally giving false compression lows at night if you lie on the sensor...

 

[mysorian]

I come from an academic background. I tend to look at all possibilities. It is just an observation for now. I am not too concerned about its performance but just using it for the trend to see how I could average out the glucose during the day. The trend during the night appears to be within safe limits.

The larger question is 'How believable, and repeatable are these expensive measurements*' and whether a day will come when with certainty we can put this matter to rest.

* include both point-in-time measurement and CGM

 

[sgm14]

​

Does the performance of CGMs degrade during their usage?

My view is that only you can answer that question as your experience may not be the same as other peoples. You are not the first person I have seen to say that, but it is not something that I have experienced.

How believable, and repeatable are these expensive measurements*

Not very. As I understand it, a finger prick meter is considered "accurate" or working correctly if 95% of its readings are within15% of your "real" blood glucose and I think cgms only have to be within 20% of the real values.

So whilst Abbott state that the libre 2 is accurate across the whole 14 days, that statement would still be true in the hypothetical situation of a libre which was returning perfectly accurate results for days 1-13 and was 19% out on day 14, but if you had such a device, you would probably consider that a significant degradation.

So whilst I don't think of my Libre2 or even my finger prick reader as being particularly accurate, I do consider the Libre to be extremely useful.

 

[Juicyj]

When I was self funding the Dexcon G6 I would regularly extend the sensor for a further 10 days and noticed no impact on readings when cross checked against a BG meter (so 20 days total), once I managed to extend it twice, however the sensor adhesive had deteriorated greatly so took if off after about 5 days.

 

[SimonP78]

I find my libre 2 calibrations are often quite severely off both for the first day or so, and the last couple of days (looking at the XDrip+ calibration curve changes/inability to generate a good fit.)

I've done it less often recently, but I'd often end up changing sensor a day or more before it had ended as the calibration offset was annoying.

I do have lots of historic data logged in XDrip+, I will try to have a look at it to determine calibration changes during the life of a given sensor....