Libre 2 not registering hypos

MarthaD

Well-Known Member
Messages
47
Type of diabetes
Other
Treatment type
Tablets (oral)
I'm not yet formally diagnosed with RH, but currently having tests for my hypos. Meanwhile, I've a question about the libre 2 and how it records hypos. I had very bad hypo symptoms the other day that took me a while to come out of. While I was experiencing the symptoms, my libre 2 wavered between 3.9 and 3.8. This continue for about 20 minutes. I couldn't do a finger prick at the time, but I think if I had it would have confirmed the hypo and I might even have read lower than 3.8 (I find the libre sometimes will say 4.1 when a finger prick reads at 3.6, for example). The problem I have is that I'm wearing a libre 2 so I can demonstrate to the endo that I'm having hypos, but it rarely registers them on the app as a low glucose event. In this particular case it shows a very rapid decline from 11mmol to about 4mmol, but the hypo itself didn't register, despite reading under 3.9 for ages. This has happened to me loads of time and so it's very difficult to "prove" my hypos through the libre, even when I'm very symptomatic and finger prick confirms. Has anyone else experienced this?
 
D

Deleted member 527103

Guest
There are a number of limitations of all CGMs but doctors, DSNs and manufacturers either don't know or don't publicise them.
Therefore, it is down to users to spread the word which you will find repeated again and again on forums such as this one.

Off the top of my head the common limitations are
- insertion trauma. Inserting a foreign object into your body can affect the readings near the Libre filament. Our body can take up to 48 hours to get used to this foreign object resulting in apparently random numbers for the first day or two. These affects vary from person to person but those of us experiencing it often insert our sensor a day or two before activating.
- compression lows. If pressure is applied to a sensor, the flow of interstitial fluid (what is used to measure BG) is restricted. Therefore, the sensor may report a false low. This is most common when sleeping and we turn over onto the arm with the sensor. You can spot these on the graphs as they are typically very short dips and returns to the previous reading, likely to be when you turn over again.
- reading delays. CGMs do not get their readings from blood, they use interstitial fluid. However, changes in BG are reflected in interstitial fluid readings about 15 minutes later. CGMs such as Libre use an algorithm to convert from interstitial fluid readings to BG. This algorithm takes the delay into consideration by extrapolating the current trend and effectively "predicting the current reading. This approach works most of the time but can exhibit some weird behaviours when our BG trend changes direction. For example, it can predict higher (or lower) than our BG actually goes so you may see a high (or low) reading which is later overwritten when the interstitial fluid "catches up". But the most common unwanted behaviour is a delay in showing a recovery from hypo.
- limited calibration - the relationship between BG and interstitial fluid is not linear. Therefore CGM manufacturers have chosen to use a conversion algorithm that is correct at "normal" BG. As a result, high readings (typically above 10 mmol/l) and low (typically below 4 mmol/l) can be inaccurate.
- meter accuracy. In the UK, all glucose meters (both finger prick and CGMs) are required to be with 15% of "true" BG value 95% of the time. If we have a finger prick meter which is, say 10% under reading and a CGM that is 10% over reading, the difference between the two can be significant. I believe this is why people often report that "Libre is always high". It is likely to be that it is higher than their finger prick meter which they have relied on for some time.
 

MarthaD

Well-Known Member
Messages
47
Type of diabetes
Other
Treatment type
Tablets (oral)
I understand all of this, thank you. My question was more about why I can read as 3.9/3.8mmol for over 20 mins, but it doesn't get captured as a low glucose event in the data or on the graph of the app.
 

ATB123

Well-Known Member
Messages
139
Type of diabetes
LADA
Treatment type
Insulin
I understand all of this, thank you. My question was more about why I can read as 3.9/3.8mmol for over 20 mins, but it doesn't get captured as a low glucose event in the data or on the graph of the app.
I've just had a look at my graphs I've had 3x hypos the last week all when I'm sleeping and all have registered. The shortest one was an hour, the longest 2 hours. I've had other mini hypos during the day, less than half an hour, maybe 15 mins toll I've fixed it, and they don't show up on the log. Wonder if it's something to do with the length of time, or perhaps needing the blood sugar to drop below what yours was? My recorded lows were all around 3.
 

TheSecretCarbAddict

Well-Known Member
Messages
298
Type of diabetes
Type 2
Treatment type
Tablets (oral)
I understand all of this, thank you. My question was more about why I can read as 3.9/3.8mmol for over 20 mins, but it doesn't get captured as a low glucose event in the data or on the graph of the app.
There are a few things. Have a look at your settings and see what you have set as a target range. If the low end is 3.9, it will still count as normal level, and only when it goes below this, it will start tracking this as a low glucose event. You can change this to 4 if you want 3.9 counted as below mormal range. Also, the reading needs to be below your target range for a period of time before it gets recorded as low glucose event - I'm sure I've seen some rules explained around this, but don't seem to find this reference anywhere. Finally, I haven't yet fully investigated the downloadable data set, but it seems that there is very detailed by minute data for very recent periods, but then it gets averaged and compressed into 15 min intevals and this averaging can potentially make hypos 'disappear' if they were borderline to begin with and lasted under 30 min.
 

TheSecretCarbAddict

Well-Known Member
Messages
298
Type of diabetes
Type 2
Treatment type
Tablets (oral)
Another thought that just occurred. I self fund Libre 2 and also BG test strips. I use Libre 2 reader to do my BG and ketone fingerprick tests. The reader connects to PC and syncs records to LibreView, which also has my CGM readings. When creating my BG analysis reports, I seem to remember to be able to specify which devices I want to include data from, and the report combines it all to show both CGM and fingerprick tests.

Maybe that's just a fancy way of saying that your doctor should be considering CGM evidence as well as whatever you can support via a blood glucose meter. Just make sure you set up date/time correctly so that the test results match. Borderline normal result from CGM confirmed with fingerprick as hypo should be good evidence.

The attached pic was me testing the difference between CGM and a fingerprick. As you can see, for me, CGM also runs higher and highlights the importance of setting the correct time.
IMG-20240422-WA0020.jpg
 

EllieM

Moderator
Staff Member
Moderator
Messages
9,797
Type of diabetes
Type 1
Treatment type
Pump
Dislikes
hypos and forum bugs
Ideally, you'd manage to do a fingerprock test when hypo so that you can demonstrate how much the libre is over reading. Failing that, you could try getting a software add on that allows you to calibrate it (xdrip???) Tagging @Antje77 who I believe uses this.

You could also consider using a cgm that permits calibration (eg dexcom g6 and g7 and dexcom ONE+ but unfortunately not dexcom ONE.

Good luck
 
  • Like
Reactions: Antje77 and MarthaD

MarthaD

Well-Known Member
Messages
47
Type of diabetes
Other
Treatment type
Tablets (oral)
Another thought that just occurred. I self fund Libre 2 and also BG test strips. I use Libre 2 reader to do my BG and ketone fingerprick tests. The reader connects to PC and syncs records to LibreView, which also has my CGM readings. When creating my BG analysis reports, I seem to remember to be able to specify which devices I want to include data from, and the report combines it all to show both CGM and fingerprick tests.

Maybe that's just a fancy way of saying that your doctor should be considering CGM evidence as well as whatever you can support via a blood glucose meter. Just make sure you set up date/time correctly so that the test results match. Borderline normal result from CGM confirmed with fingerprick as hypo should be good evidence.

The attached pic was me testing the difference between CGM and a fingerprick. As you can see, for me, CGM also runs higher and highlights the importance of setting the correct time. View attachment 67355
Oh, thank you, that's super helpful! I'm going to try this.
 
  • Like
Reactions: Lamont D

sgm14

Well-Known Member
Messages
258
Have a look at your settings and see what you have set as a target range. If the low end is 3.9, it will still count as normal level, and only when it goes below this, it will start tracking this as a low glucose event. You can change this ...

These settings affect your Time in Range, but I don't think they affect how an LGE is defined.
 
  • Like
Reactions: TheSecretCarbAddict

TheSecretCarbAddict

Well-Known Member
Messages
298
Type of diabetes
Type 2
Treatment type
Tablets (oral)
It turns out that the user manual found via the Help menu in the app is quite a good overview of how it all works. Found the reference to what counts as a low glucose event.
Screenshot_20240426_235200.jpg
 
  • Like
Reactions: MarthaD

Antje77

Guru
Retired Moderator
Messages
20,278
Type of diabetes
LADA
Treatment type
Insulin
Ideally, you'd manage to do a fingerprock test when hypo so that you can demonstrate how much the libre is over reading. Failing that, you could try getting a software add on that allows you to calibrate it (xdrip???) Tagging @Antje77 who I believe uses this.
I use the alternative app DiaBox (not available in the app store) to calibrate.
More information about DiaBox here: https://www.facebook.com/groups/974278412920319
 
  • Like
Reactions: EllieM and MarthaD

DawnBenson

Newbie
Messages
1
Type of diabetes
Type 2
Treatment type
Tablets (oral)
There are a number of limitations of all CGMs but doctors, DSNs and manufacturers either don't know or don't publicise them.
Therefore, it is down to users to spread the word which you will find repeated again and again on forums such as this one.

Off the top of my head the common limitations are
- insertion trauma. Inserting a foreign object into your body can affect the readings near the Libre filament. Our body can take up to 48 hours to get used to this foreign object resulting in apparently random numbers for the first day or two. These affects vary from person to person but those of us experiencing it often insert our sensor a day or two before activating.
- compression lows. If pressure is applied to a sensor, the flow of interstitial fluid (what is used to measure BG) is restricted. Therefore, the sensor may report a false low. This is most common when sleeping and we turn over onto the arm with the sensor. You can spot these on the graphs as they are typically very short dips and returns to the previous reading, likely to be when you turn over again.
- reading delays. CGMs do not get their readings from blood, they use interstitial fluid. However, changes in BG are reflected in interstitial fluid readings about 15 minutes later. CGMs such as Libre use an algorithm to convert from interstitial fluid readings to BG. This algorithm takes the delay into consideration by extrapolating the current trend and effectively "predicting the current reading. This approach works most of the time but can exhibit some weird behaviours when our BG trend changes direction. For example, it can predict higher (or lower) than our BG actually goes so you may see a high (or low) reading which is later overwritten when the interstitial fluid "catches up". But the most common unwanted behaviour is a delay in showing a recovery from hypo.
- limited calibration - the relationship between BG and interstitial fluid is not linear. Therefore CGM manufacturers have chosen to use a conversion algorithm that is correct at "normal" BG. As a result, high readings (typically above 10 mmol/l) and low (typically below 4 mmol/l) can be inaccurate.
- meter accuracy. In the UK, all glucose meters (both finger prick and CGMs) are required to be with 15% of "true" BG value 95% of the time. If we have a finger prick meter which is, say 10% under reading and a CGM that is 10% over reading, the difference between the two can be significant. I believe this is why people often report that "Libre is always high". It is likely to be that it is higher than their finger prick meter which they have relied on for some time.
Thank you. This is really very helpful