• Guest, the forum is undergoing some upgrades and so the usual themes will be unavailable for a few days. In the meantime, you can use the forum like normal. We'd love to know what you think about the forum! Take the 2025 Survey »

This is insane... (FreeStyle Libre inaccuracy)

Ellenor2000

Well-Known Member
To clear the air: I don't think Abbott can be blamed because I put the sensor in my left pectoral, not on my arm as you're meant to.

However, it seems like ISF glucose is, at least on this sensor and this site, reading unreliably between .7 and 1.8 mM-equiv below fingerstick blood glucose on an Accu-Chek Aviva. Additionally, it won't let me set a hypo threshold that actually corresponds to my experience, instead displaying low at 3.9, which on top of the erratic low read, means that Libre may show me low (3.6, say) when I may be 4.0, 4.8 or even in the 5s and 6s, or it may "correctly" show a low when I'm not low enough to worry (3.7, for example).

Six words that keep coming to mind are "This would kill an actual diabetic." I've heard (through YouTube) of exactly this problem coming from actual diabetics (including one French man, Hidgi Chuan, who had two sensors and two meters show two massively different numbers: 240++ on one arm and 183++ in another arm - the ++ means a straight up arrow) who I must presume actually followed the instructions, which is why I keep saying that to myself.

One hypothesis I have other than the obvious "wrong site, problem exists between keyboard and chair" is that even though I'm not scorbutic, my low vitamin C intake (carnivore diet) may be making the sensor read erroneously low (since hypervitaminosis C can cause you to read a false high). This shouldn't be happening, but it seems to be. When I ate strawberries for a bit, Libre and fingerstick lined up perfectly at 4.6 and then diverged a bit again. It didn't work the second time I tried.
 
Hello @Ellenor2000

The libre is a flash monitor and is much different device to a blood glucose monitor which checks blood and is more accurate, the libre is testing interstitial fluid and is well known for having up to a 20 minute delay, hence the big difference. I would never use a libre for decisions in regards to insulin dosing and if having a hypo I would confirm it with a blood glucose meter.
 
There are many many many many threads on Libre (type "Libre" in the field on the top right to find them).
Quite a few talk about inaccuracies but still they are very popular.
My initial reaction was similar to yours regarding inaccuracy but I soon learnt
  • Libre uses interstitial fluid to calculate Bg which is 20 minutes behind finger prick reading
  • If you BG is changing quickly (e.g. when you have just eaten or whilst doing exercise), the delay will make a big impact
  • Libre is most accurate between 4 and 8 mmol/l
  • If you apply pressure to Libre it will read low
  • Libre is factory calibrated and not everyone is the same as the factory. I use Glimp app (only available on Android phones) which uses a different algorithm to the LibreLink app to convert interstitial fluid reading to to BG. This takes into consideration calibration from finger pricks
  • The value of Libre is not the moment in time readings which we compare with finger pricks: it is the trends both in terms of "I am now 6.3 but my BG is rising quickly" and "every morning my BG starts rising at 6am" information.
  • You can "pimp" your Libre with a device such as Miaomiao and convert it into a CGM.
If you are using Libre to replace finger pricks readings, it is like using a mobile phone to make phone calls and nothing else - you may as well have an old school Nokia; at least you can play snake on the phone.
 
To clear the air: I don't think Abbott can be blamed because I put the sensor in my left pectoral, not on my arm as you're meant to.

However, it seems like ISF glucose is, at least on this sensor and this site, reading unreliably between .7 and 1.8 mM-equiv below fingerstick blood glucose on an Accu-Chek Aviva. Additionally, it won't let me set a hypo threshold that actually corresponds to my experience, instead displaying low at 3.9, which on top of the erratic low read, means that Libre may show me low (3.6, say) when I may be 4.0, 4.8 or even in the 5s and 6s, or it may "correctly" show a low when I'm not low enough to worry (3.7, for example).

Six words that keep coming to mind are "This would kill an actual diabetic." I've heard (through YouTube) of exactly this problem coming from actual diabetics (including one French man, Hidgi Chuan, who had two sensors and two meters show two massively different numbers: 240++ on one arm and 183++ in another arm - the ++ means a straight up arrow) who I must presume actually followed the instructions, which is why I keep saying that to myself.

One hypothesis I have other than the obvious "wrong site, problem exists between keyboard and chair" is that even though I'm not scorbutic, my low vitamin C intake (carnivore diet) may be making the sensor read erroneously low (since hypervitaminosis C can cause you to read a false high). This shouldn't be happening, but it seems to be. When I ate strawberries for a bit, Libre and fingerstick lined up perfectly at 4.6 and then diverged a bit again. It didn't work the second time I tried.

Are you not diabetic yourself then? I'm confused by you talking about 'actual diabetics'.

Many of the differences you're talking about are tiny - 0.7 mmol out or, showing 3.6 when you're 4.0 - and suggest the sensor is performing fairly well. Every glucose sensor or blood glucose meter has a MARD stating how accurate you can generally expect it to be. If you're not diabetic, it will still give you all the trending information you are probably looking for.
 
hello im having inacurracies wirh my senor the libre saying im 2,5 but x drip and blood test saying 7.5 is this what is known as a dead sensor tia
 
Are you not diabetic yourself then? I'm confused by you talking about 'actual diabetics'.

Many of the differences you're talking about are tiny - 0.7 mmol out or, showing 3.6 when you're 4.0 - and suggest the sensor is performing fairly well. Every glucose sensor or blood glucose meter has a MARD stating how accurate you can generally expect it to be. If you're not diabetic, it will still give you all the trending information you are probably looking for.

Yeah, non-DM case.

I put the sensor in my arm this time (explanted the old sensor early) and the numbers are more CONSISTENT in their inaccuracy. They're still inaccurate if you compare to my old fingerstick meter, but now they're consistent, and as a result of that consistency they told me why I sometimes get cold sweats and jitters after having coffee - turns out coffee gives me possible hypoglycaemia indicated by low ISF glucose (who knew?!)
 
Last edited:
Back
Top