False lows on libre

[curlygirl]

I have had two lows on the libre recently that did not exist. I know from experience some sensors are more accurate etc but this was really reading way off. The one today said 3.2 and a straight down arrow but when I tested blood on the libre and on my other meter both said 6.4 and 6.7. This is the second time such a low has shown but blood fine in last week or so. Hoped it was dodgy sensor but have changed it since the last time. This has happened before but with both of these the strangest thing is that the false low appears in the logbook but not on the graph. Has anyone else had this or know why the graph does not show the false low.

 

[helli]

Given the night time (ish) and sudden dips and rises, they look like compression lows.
This happens with all CGMs and are not specific to Libre and not a fault of the sensor.
It is physics - if you apply pressure to a sensor (e.g. lying on your sensor at night), they will report false lows.

Regarding lows in the logbook which do not appear in the graphs, these may be due to a change in direction of the trend in the last 15 minutes.
There is a 15 minute delay in interstitial fluid readings (read by CGMs) and finger pricks. To account for these, Libre will predict the current value by extrapolating the current trend. When the true data comes in, Libre will correct the prediction if the trend changes in the last 15 minutes.
Again, this is not a fault in the sensor - it is the way they are designed to work.

 

[rebrascora]

Was there a reason for your levels dropping at that time? ie. Had you injected insulin or done exercise/activity which would cause your levels to drop. Your levels look to go up significantly afterwards from what I can make out on the graph.
Did you eat carbs perhaps to treat the reported low or did you double check with a finger prick first?

Could you have been leaning or lying on the sensor and causing it do have a compression low?

 

[curlygirl]

The false low was after lunch. The actual blood reading was over 6 but the libre said 3.2. It was not a compression low as I was not in bed at 1.45pm. The levels did not drop that is the point really. The level was over 6 but the libre was showing 3.2. It was more that it does not show on the graph that confused me. It should not be on the graph because the 3.2 did not happen but the libre thought it did so I would expect it to show on the graph. Anyway maybe just a strange thing the libre does.

 

[helli]

The false low was after lunch.

I was referring to the dips at about 6am

 

[Spathiphyllum]

The false low was after lunch. The actual blood reading was over 6 but the libre said 3.2. It was not a compression low as I was not in bed at 1.45pm. The levels did not drop that is the point really. The level was over 6 but the libre was showing 3.2. It was more that it does not show on the graph that confused me. It should not be on the graph because the 3.2 did not happen but the libre thought it did so I would expect it to show on the graph. Anyway maybe just a strange thing the libre does.

About false readings in the Logbook which later do not appear on the graph-- yes, that is "just a strange thing the libre does"!

Puzzled me too at first, but the explanation is essentially as given above by Helli. Remember that the Libre sensor filament is not actually in your blood; it's in the interstitial fluid between your cells. Glucose from your blood seeps through the walls of your capillaries into your interstitial fluid, but it takes a little while--about 15 minutes-- for a change in the concentration of glucose in your blood to be reflected in a change in the concentration of glucose in your interstitial fluid.

Every time you scan the sensor, what you are getting is not the 'raw' result-- the concentration of glucose in your interstitial fluid. The result you get, flashing up and then kept in your Logbook, is what the Libre algorithm thinks the result from your interstitial fluid means in terms of what your actual BG is likely to be.

Then, when later results come in, Libre may realise this was wrong-- and the graph will show a re-adjusted result, not the one you were given at the time and which remains recorded in your Logbook.

This phenomenon was discussed on a previous thread, and someone-- can't remember who!-- said it was like someone getting to redo their exam after it was marked. ; )

And-- all of the above is a separate issue from whether either of those Libre results-- either in the Logbook or on the graph-- is correct when compared to your actual blood-test result! Hours of fun, eh? But, I know, most of the time the Libre is very useful and we should be grateful ...

 

[everydayupsanddowns]

I have had two lows on the libre recently that did not exist. I know from experience some sensors are more accurate etc but this was really reading way off. The one today said 3.2 and a straight down arrow but when I tested blood on the libre and on my other meter both said 6.4 and 6.7. This is the second time such a low has shown but blood fine in last week or so. Hoped it was dodgy sensor but have changed it since the last time. This has happened before but with both of these the strangest thing is that the false low appears in the logbook but not on the graph. Has anyone else had this or know why the graph does not show the false low.

Hard to tell from those graphs, but it looked like your BGs had been wobbling quite a bit around that time, so it could have been sensor lag. Sensors generally lag behind by 5-10 minutes, and I find mine are laggier if I am not as well hydrated. Apparent differences are likely to be larger when BGs are on the move.

Were you getting the same sort of discrepancies when your levels were pootling along fairly flat?

If you call-in to Abbott they may offer to replace sensors that are not performing as they expect - and they will ask for pairs of readings from various times. And the sorts of differences you saw do seem unhelpfully large.

 

[curlygirl]

Were you getting the same sort of discrepancies when your levels were pootling along fairly flat?

Thanks for replies. The false lows that have not happened do not appear on the graph. That is really what I was asking about. I always check blood against the libre if the scan says something unlikely or low etc and the libre scanner said 6.7 vs 3.2 for the scan. I thought all scans appeared on charts but now I know this is not necessarily true. The line on libre graphs looks wobbly... but actually my blood had been between 5 and one quick rise to 8.9 after breakfast that settled straight back to 6.
I have seen quite a few scans recently that are not actually true. I think the libre is really helpful and I appreciate having it.
The difference it makes generally is huge...less sore hands due to many less blood tests than before had libre and the best feature is the arrow as before Flash GM a blood test was a single information of what it is but not what it has been or what direction it is going in. So they are very helpful. I personally would have thought it a better idea if abbott had decided adding calibrating it as an option. That way when a sensor is reading out the blood test done to confirm that would actually help to get the scan and blood more the same...

 

[Spathiphyllum]

I personally would have thought it a better idea if abbott had decided adding calibrating it as an option. That way when a sensor is reading out the blood test done to confirm that would actually help to get the scan and blood more the same...

Me too!! I think some CGMs do allow, or even require, calibration against blood test? But not sure which ones-- or whether we can get them on the NHS.

I mean, I know we can ring Abbott and get a replacement-- did that for the first time recently, having prepared a whole list of pairs of readings showing the Libre was way off-- and then when I rang found they only wanted three pairs of readings! So I've now rung them and got replacements twice ... But would much rather not have the bother in the first place.

 

[helli]

I totally agree about the need to calibrate a CGM.
When I first had a Libre, I hated it because it was so far out I tried another and had the same problem. I nearly gave up until I realised there were unofficial apps that can read Libre, like xDrip and Diabox, which allow calibration as well as streaming the data over Bluetooth so no need to scan. I used xDrip for years.
I was lucky that my DSN was happy not to have my data available to review. I guess I could have scanned just for her but there seemed little point sharing data I knew was inaccurate.

 

[Spathiphyllum]

I totally agree about the need to calibrate a CGM.
When I first had a Libre, I hated it because it was so far out I tried another and had the same problem. I nearly gave up until I realised there were unofficial apps that can read Libre, like xDrip and Diabox, which allow calibration as well as streaming the data over Bluetooth so no need to scan. I used xDrip for years.
I was lucky that my DSN was happy not to have my data available to review. I guess I could have scanned just for her but there seemed little point sharing data I knew was inaccurate.

Ooh! I did not know about xDrip or Diabox, will look them up.

You say you "used xDrip for years"; why did you stop?

 

[helli]

You say you "used xDrip for years"; why did you stop?

I stopped because I decided to self fund a different CGM that talks to my pump.
BTW neither of these apps are available on the PlayStore as they are not official. They will need to be downloaded from github. xDrip has a very helpful Facebook group which includes documented instructions for getting it working. I say "very helpful" but they are volunteers and have limited patience for someone unwilling to read the instructions which I think is reasonable.

 

[curlygirl]

I know dexcom g6 and g7 are callibrated. I was thinking of trying one g7 to see if better accuracy than the libre 2 but not sure if adhesive would be as good. The adhesive on L2 is ok but very sensitive to some dressings etc so not sure. Be helpful to be able to calibrate though. Do not self fund though so would have to talk to nurse about whether would be able to get g7 if did find adhesive was gentle enough. I know L3 is supposed to have vetter accuracyvandvitbis real time but not got option to calibrate still...

 

[Tdm]

I had real trouble with libre. I got my levels steady and nicely low overnight but libre couldn't handle it- it gave repeated false lows, one night it was 10. They weren't compression lows as i took measures to ensure i did not sleep in that arm. It was interesting to run a dexcom alongside the libre - occasionally they agreed but in most case it was a false, sudden libre low.
Now on dexcom g6, i can calibrate it ( though generally don't need to ) and its accurate pretty much out of the box. Plus its authilorised to use on the belly
libre is great, and has a lot of benefits, but dexcom, for me, is better. Things may be better with libre 3 but my nhs team said that wasn't an option