Advice regarding discrepancies between Libre 2 sensor readings and blood glucose monitor readings

[Introverted_And_Proud]

Okay, so, I thought I should post an update here regarding what I've been testing with the sensor for the past two weeks and how it's affected the discrepancies between my ISF readings and BG readings.

Per the advice from this thread, I put a new sensor on my right arm on September 30th, because the activated sensor on my left arm was giving me ISF readings that varied over 1mmol to 2.5mmol higher than my BG even after two days passed. I suspected it might have been related to where I applied the sensor, as it was quite low on my upper arm and a little too close to my elbow for my liking. I applied the new sensor higher up on my right arm and then waited a little over 24 hours before officially activating it.

Within the first 48 hours, the new sensor readings were a little all over place, but the majority read under 1mmol difference between the ISF and BG (I should mention this was regardless of how high or low my BG was), so I stopped finger pricking after the two days were up. I didn't finger prick again until the sensor reached the final three days of its life, where I finger pricked to double check the sensor was still accurate. To my surprise, the ISF readings remained under 1mmol, right up until the end of its life. It seemed like activating the sensor 24 hours after application had improved its accuracy and lessened the discrepancy.

Yesterday, however, I activated my most recent sensor. As with the previous one, I applied it to my left arm and made sure the placement was similar to the last one, waiting a little over 24 hours before activating it. So far, the ISF readings have continually read 1.1mmol - 2.2mmol higher than my BG, with only one reading showing a discrepancy of 0.2mmol when I tested early in the morning. Not sure if the ISF readings can potentially vary per arm (a little like how BG readings can vary per finger) or perhaps it was because I activated the sensor fifteen minutes earlier than the previous one. But as it's only been a day, I'll have to see how the discrepancies vary after another 24 hours. But so far, it does seem like the discrepancies aren't as large for me if I let the sensor have a day to settle into my arm, which is reassuring and gives me less anxiety.

 

[Jasmin2000]

Just want to add that, in addition to all the effects you may see from the placement of Libre 2, the time since application (that first 24h warm-up), and a host of other possibilities, there is a purely technical issue. Abbott have published papers showing that Libre 2 reports lower that actual BG at levels of about 4.5 or less, and higher than BG at levels of about 8 and above. This skew is apparently intentional to give earlier warnings of hypos or hypers.

In addition to the technical skew, there is an algorithm skew. Libre 2 doesn't report the ISF glucose levels directly, but it uses these to predict what your BG is (this mostly overcomes the fact that ISF levels are 5-15 minutes behind BG levels). For most folk, the algorithm is pretty good and we use the Libre 2 value as a close estimate of actual BG, but for those tracking a rapidly changing BG, for example when exercising, the algorithm creates lows that simlpy never occur in the BG readings.

I've mapped the drop and rise in BG when exercising and eating by simultaneously using Libre 2 and doing 5 minute finger-pricks (about 1800 so far) and I get 2 curves, which start and finish at about the same point, but bottom out or peak with a difference of 0.5-2.5 mmol/L - Libre 2 always being lower than BG on the drops and higher than BG on the spikes.

 

[Introverted_And_Proud]

Just want to add that, in addition to all the effects you may see from the placement of Libre 2, the time since application (that first 24h warm-up), and a host of other possibilities, there is a purely technical issue. Abbott have published papers showing that Libre 2 reports lower that actual BG at levels of about 4.5 or less, and higher than BG at levels of about 8 and above. This skew is apparently intentional to give earlier warnings of hypos or hypers.

In addition to the technical skew, there is an algorithm skew. Libre 2 doesn't report the ISF glucose levels directly, but it uses these to predict what your BG is (this mostly overcomes the fact that ISF levels are 5-15 minutes behind BG levels). For most folk, the algorithm is pretty good and we use the Libre 2 value as a close estimate of actual BG, but for those tracking a rapidly changing BG, for example when exercising, the algorithm creates lows that simlpy never occur in the BG readings.

I've mapped the drop and rise in BG when exercising and eating by simultaneously using Libre 2 and doing 5 minute finger-pricks (about 1800 so far) and I get 2 curves, which start and finish at about the same point, but bottom out or peak with a difference of 0.5-2.5 mmol/L - Libre 2 always being lower than BG on the drops and higher than BG on the spikes.

I'll be honest, I was not aware that the Libre 2 exaggerated ISF readings for BG under 4.5mmol. I knew it did this for readings above 10mmol, but not the other way as well. Though, it does explain why my BG readings would sometimes still read above 4mmol despite my sensor claiming it was below. XD

But thank you for letting me know this information. It's helped to ease my anxiety over the above 1mmol discrepancies a little more. And also made me realise that the diabetes consultants at my clinic are not very well informed on the Libre as I initially thought. At my last consultation appointment, the consultant I was seeing questioned some of my readings from the week before because of how high they seemingly were, and I had to explain that the sensor had overestimated the ISF readings from the BG readings by 1-2mmol, making them look worse than they supposedly were. Their response was to ask if I'd changed my sensor over yet, and when I said 'no', they advised I do that and see if that helps. Spoiler: it did not. XD The readings were still a little exaggerated. At least now I can see a probable reason for that outside of the sensor potentially being faulty.

I might mention this at my next appointment and see if that changes how the consultant approaches my case. But if not, per Nicola's advice, I'll try to see if they can get me in touch with the DSNs again. The DSNs at my clinic seemed better informed on the Libre 2 and actually dug deeper into my case through the reports, so I feel like they could help me to understand the Libre 2 more and how I can use its data to better manage my condition.