Reliability of Libre 2 sensors

My wife’s FL2 sensors have had a 75% failure rate during the last year or so. Previously, failures were rare. By failure I only mean consistent reading errors large enough (>20%) to qualify for a free replacement by Abbott. Sometimes the error is >50%. (I do not count the rare failures of new sensors). She compares the sensor BG reading with a trusted finger prick device whenever she’s suspicious of the FL2 reading. They seem to either last 2-3 days, or about 10 days before the large error occurs.

How often do you test the sensor readings? Have you had many free replacements for this reason?

When I used the Libre 14 day I had a failure rate of 52% tracked over a year where failure was either the app saying replace sensor early or the sensor falling off.

When your wife is comparing CGM to a BGM how many measurements is she comparing over how much time before she replaces the sensor? Is she using the freestyle strips and the Libre reader or something else?

Thanks, 52% was high also.
She knows to wait 10mn to compare with the BGM, and only decides to replace it when the errors have lasted a day or 2, perhaps 4+ comparisons. Once the sensor starts underreading by well over 20%, it never gets better. Her strips are Contour Next which have the best accuracy on the US market.

Yeah. At the time the Libre 2 was new on he market but I switched to the Dexcom G6 and it was bad much less often.

I use the Contour Next also. From the manual

Table 1 - System accuracy results for glucose concentration < 75 mg/dL
Difference range in values between YSI laboratory reference method and Contour Next meter Within ± 5 mg/dL Within ± 10 mg/dL Within ± 15 mg/dL
Number (and percent) of samples within specified range 74 of 78 (94.9%) 77 of 78 (98.7%) 78 of 78 (100%)
Table 2 - System accuracy results for glucose concentration ≥ 75 mg/dL
Difference range in values between YSI laboratory reference method and Contour Next meter Within ± 5% Within ± 10% Within ± 15% Within ± 20%
Number (and percent) of samples within specified range 390 of 522 (74.7%) 513 of 522 (98.3%) 521 of 522 (99.8%) 522 of 522 (100%)

So if the CGM is saying something that doesn’t match what I’m feeling I’ll treat based on the BGM without a second thought. When I suspect a problem CGM to the BGM I have to take the error margin of the BGM into account. The Libre2’s performance is in the manual starting on page 197. Error margins are in Table 1d on page 202. For 70-180 mg/dL only 86.9% of Libre2 readings were within +/-20%. Out of a 14 day sensor session that is only 12.2 days of readings that will be within +/-20%. Nothing says those readings that are within +/-20% have to be contiguous.

Table 3c on page 209 give more detail. Find the bins for each of the min, median (BGM reading and max and read across the row to see how often the CGM might agree. Let me know if you want me to do a couple of example blood glucose readings, this is already a long post.

All that to say not matching a BGM reading by 20%, even for a day, isn’t a sensor failure. It is nice of Abbott to replace sensors when you document reading outside of 20% from a BGM but they won’t do it forever. I’ve heard of people being denied replacements by Abbott. You also say the Libre2 reads lower than the BGM. That is considered safer because reading high could lead to dosing too much insulin.

Ask your wife to join JDRF. There are lots of people here she can talk to about strategies to handle CGM readings and tips to try to get better numbers.

Thanks, those numbers are useful.

My wife is fairly forgiving and always waits at least a day before concluding on a bad sensor, and the fatal errors are usually in the 50-90% range (under-reading), WELL above the expected range.

I used to use the FL2 & had A LOT of sensors that didn’t work (message saying replace sensor now within the first day). So I got fed up & switched to the Dexcom G7. It’s MUCH more reliable for me. Give it a try.