When n=1 is not enough. An initial take on the Libre2 and Libre2 Plus comparison

Libre2 and Libre2 Plus side by side. With helpful labelling.

As I mentioned a couple of weeks ago, I received a number of Libre2 and Libre2 Plus devices to try. So far, I’ve only used one pair, but this is an initial view of what I saw.

The method used here to compare looks at the differences between TIR, TBR and TAR over the latter two weeks of each sensor, compared with Dexcom data from overlapped sensors for the same period.

There were also spot checks throughout the period, as well as a number of days with 1000mg vitamin C dosing to see if that affected the results of the Libre2 or Plus.

Initial Results

In this case, I’m using the Dexcom One as my datum for comparing the Libre sensors, as the device is calibratable and in my use of it has tended towards being the most likely to match glucose from fingerpricks (usual caveats about sources of glucose data andd inaccuracies apply).

In terms of variation from the Dexcom, the Libre pair provided some interesting results. The numbers in this grid are the relative percentage (positive number means greater amount of time, negative number means less time) of time either above, in or below range compared to the Dexcom device.

Relative performance compared to Dexcom One.

What does the above show? It shows that with these two sensors, the Libre2 read below 3.9 mmol/l nearly 50% of the time more than the Dexcoms, and the Libre2 Plus managed this more than twice as much, with corresponding reductions in other metrics.

Point tests on the two sensors tended to back up this variation as well, although the Dexcom was not without its faults, reading nearly 30% higher than bloods did on one occasion.

The difference being that when the Dexcom was in that state, it could be calibrated.

This is a challenging set of data, because anecdotally, I’ve heard from others that they saw significantly higher readings from the Libre2 Plus.

Effects of Vitamin C

Given the FDA concerns about doses of more than 500mg of Vitamin C supplementation, I expected a significant variation to glucose levels displayed by the Libre2.

The warnings all suggested that there would be elevated glucose values with 1000mg supplementation.

Libre2 Vitamin C warning screenshot

Expectation, in this case, did not meet reality. Four days of supplements in the middle of the sensor life showed little to no effect on either of the two sensors. Both continued to sit generally lower than Dexcom and capillary fingerpricks.

What can we take from this?

Sadly, not very much. Neither of the two Abbott offerings covered themselves in glory on this occasion.

What it shows is the constraints of a single sensor n=1 comparison, why larger scale trials are required, and perhaps most importantly, that there is clear sensor to sensor variability, even with the updated sensors.

I’ll continue to try the Libre2 and Libre2 Plus side by side and see if anything changes.


  1. Do we think libre2 and libre2+ are actually different or is this attributable to natural sensor to sensor variation

    • We know there is a difference due to the changes associated with Vitamin C. I would attribute this to sensor to sensor variability though, without further tests.

  2. Very interesting start and only one possibly irrelevant comment that in the ’70’s we were taught vitamin C can’t be stored in our bodies but must be taken each day so I still assume this becomes unusable (akin to cutting into fruit destroying vit C ) and would need to be taken daily to affect a sensor?

  3. I was using Libre with Maiomaio and XDrip which allowed the app to handle calibration. I now the Libre2 and its app only as a guide and check blood values more now than ever before because of the variation of values.

  4. The raw numbers may be more useful to see in this case rather than the relative percentage. 3min TBR vs 1min TBR is a large relative percentage but not a meaningful difference.

    • To put it in context, Dexcom One showed 5% TBR over the duration, Libre2 7.5% and Libre2+ 10.8%.

Leave a Reply

Your email address will not be published.