• Guest - w'd love to know what you think about the forum! Take the 2025 Survey »

Can Libre2 readings "drift" out the longer you wear it?

P

PlutoIsAPlanet

Member
Messages
8
Type of diabetes
Type 2
Treatment type
Tablets (oral)
Does anyone know if it's common for readings from Libre2 to drift out from readings taken from a blood prick the longer you wear it?

Basically - In the first week-ish of me wearing mine I was checking my blood sugar readings by finger prick several times a day and comparing to the Libre. For the most part it seemed accurate - within 0.5mmo/L to 1mmo/L.

Noticed it 2 days ago (10 days into wearing it) that the readings were off consistently by over 1.5mmo/L.

This is with me taking into account how there's a lag. So today - for example this evening I've been getting some readings as low as 5.5 from my fingerprick. However according to my Libre2 graph I've not once any point during the past few hours when I've taken these finger pricks gone below 7.2.

Putting a new one on tomorrow so will see how that goes..
 
Last edited:
I don't have this experience, for me the first day is different (even when applying a day before activating) but the rest of the 13 days my sensors are pretty consistent (usually off by 0.8 to 2.0 mmol/l on the low side). I take fingerpricks every day too.

Which does not mean that others can see a difference towards the end of a sensor of course, you're not the first to report this.
 
I read this study a while back: https://pubmed.ncbi.nlm.nih.gov/31613140/

Note that the study was undertaken on adults with type 1 diabetes so the sample population did not include those with other types of diabetes or on children. That said the results can likely be applied to other people including those without diabetes and children (at least it’s not a study done on rats). So as long as you keep that limitation of the study in mind it’s an interesting read. Summary least accurate first and last days. Also check the author affiliations for any potential bias — I’m on my phone at the moment and for some reason the link won’t open.

My personal experience was that on the first day the Libre 2 results were sometimes almost semi-random numbers, except the general trend was ok (i.e. I could still see when I’d eaten on the graph). I tried applying it, without activating, 24 and 48 hours before activation but for me it didn’t seem to make a lot, if any, difference. As for the last day of sensor wear I can’t recall there being anything particularly unusual but many people have had such an experience as @Antje77 mentioned And @Lakeslover at least has personal experience

Slightly off-topic but I have the same experience now that I’m on the Dexcom G6: the first day is so semi-random I rely on finger pricks; the graph looks like a scatter plot rather than the normal fluctuations… it’ll go from 2.5 to 8 and back for each 5 minute interval. When I’m back at my main computer I might post a photo. Anyway the reason I mention this is that the topic, in my experience, isn’t restricted to the Libre 2
 
First day is off, though as @plantae says the trends are correct. I use XDrip+ and find I have to sometimes completely clear the calibrations on the 2nd day and effectively start from scratch again (XDrip+ recommends a couple of calibration finger prick readings a day, ideally at different levels to update a calibration graph).

I started using XDrip+ for access to the data rather than the ability to calibrate, and only occasionally calibrated in the early days (i.e. every few days, with a couple more over the first day or two). That was up until the point where I had a sensor that was reading much lower than my actual blood glucose (libre was saying something like 5, finger prick was saying something along the lines of 11). I was really surprised that the sensor was so far off, so much so that I had to go to the GP and ask the nurse to do a finger prick test as I wasn't sure whether my finger prick device was malfunctioning (I'd obviously been running high for long enough that actually dropping back to normal numbers made me feel like I was going low). That was rather a wake-up call, and I was quite annoyed as I had thought I'd been running along with good control.

From that point on I've made sure I do calibration fingerprick tests whenever there's a vaguely suitable steady state period, and even if there's not that I do a test and check by eye that the calibration isn't way off (i.e. sensor has failed). So at least 2 calibration tests per day.

I did get Abbott to replace that sensor fwiw and they had no issues with it. I almost always find that failed sensors read lower values than the actual blood tests, which is afaiu a safety measure as it's more dangerous to go low than high (short term at least). With XDrip+ I can often keep running a sensor that otherwise would have such an annoying offset (I do sometimes also test with the Libre reader to see what it thinks) that it would need to be replaced.

However, XDrip+ has a limitation on its calibration abilities such that it cannot calibration a sensor that would not be able to read low (i.e. it cannot pull the values from the sensor up so much that the minimum value produced by the sensor is above some low value, e.g. low alarm, though I think it's hard coded in fact) - this is fine, it's avoids the sensor going off-scale low with low being above the alarm level so you never get an alarm. At this point I also contact Abbott for replacements, thankfully I've not needed to do this for a while, but I have had runs of sensors where they have all failed early.

It's another one on my todo list, to look at the historic data from XDrip+ and look at how the calibration curve changes over time and the error between the Libre values & fingerprick values and see how far off they are, how long the sensors last, etc. When I finally get round to doing it I'll share the graphs.
 
Last edited:
PlutoIsAPlanet said:
Does anyone know if it's common for readings from Libre2 to drift out from readings taken from a blood prick the longer you wear it?

Basically - In the first week-ish of me wearing mine I was checking my blood sugar readings by finger prick several times a day and comparing to the Libre. For the most part it seemed accurate - within 0.5mmo/L to 1mmo/L.

Noticed it 2 days ago (10 days into wearing it) that the readings were off consistently by over 1.5mmo/L.

This is with me taking into account how there's a lag. So today - for example this evening I've been getting some readings as low as 5.5 from my fingerprick. However according to my Libre2 graph I've not once any point during the past few hours when I've taken these finger pricks gone below 7.2.

Putting a new one on tomorrow so will see how that goes..
Click to expand...
As your sensor comes to the end of its life it can be a wee bit less reliable but research has proven it is still reliable. Things should be normal once you change your sensor.
You should confirm with a finger prick reading if your symptoms don't make the readings. Otherwise relax!
 
I definitely found Libre drifting through a sensor’s life.
like @SimonP78 imused xDrip+ which allowed me to calibrate.
 
Personally, I only found it becomes unreliable on days 13 or 14 if the sensor has started to work loose I.e. visibly becomes detached from my arm.

Day 1 reading are always suspect, despite me (usually) inserting a new sensor up to 24 hours before activation.

(I use Bubble plus Glucosedirect app to calibrate.)
 
Back
Top