New Generation Enlite CGM: Real World Performance So Far

We're coming to the end of our second New Generation Enlite sensor (aka Enhanced Enlite).

Whilst there's nothing new with the sensor on the Medtronic 640g (it's the Guardian 2 Link transmitter that's new) I thought it might be worth sharing our early experiences, focusing on the first sensor performance, particularly as I'm aware some other new 640g users haven't had a smooth start.

If you don't want to scroll all the way through, here you go :-)

Insertion

ULCH have a good video on how to prepare and then insert the Enlite sensor.

Inserting the first sensor was not pain free for Janki - she made that clear to us... However, she settled down within a minute or two and barely mentioned it whilst wearing it. On the advice of our DSN, we plumped for placement on the upper buttock. So far it's worked very well, using an additional adhesive patch (supplied with the Enlite) to tape over the edges of the transmitter and help stop it getting snagged (or worse) when Janki's getting herself changed.

After a soak in the bath and with the aid of some silicone adhesive removal spray, we got the first sensor off without incident. The second sensor went on without any grumbling (Janki was busy playing on her Mum's phone at the time as a distraction).


Transmitter

For the seven days Janki wore the first sensor, the Guardian 2 Link transmitter didn't cause us any problems and we didn't have any low battery warnings. Recharging between sensors took around 40 minutes in its tiny AAA-battery powered charger (good for around 40 charges apparently, which gives you some idea of just how efficient this transmitter is, with its 802.15.4 RF protocol no doubt helping to achieve that. The manual suggests a full recharge (from zero) takes around an hour, so I guess that means the transmitter's battery probably had another day or two in it before a recharge would be needed.

As I mentioned in my post on connectivity with the Guardian 2 Link, we've not had any real sensor data connection issues. Over the week with the first sensor, reviewing the data log, I think the only sensor loss errors we've seen are connected with bathing and swimming (when I don't expect the signal to make it out of the water).

Lifespan

We didn't set out to 'extend' the lifespan of our first sensor, but logistics (i.e. preschool, work and wanting two of us around for this first change) meant that we did push the sensor (and transmitter) out to seven days. We had no issues in doing this. We didn't recharge the battery at the extension point, we just disconnected it from the sensor and immediately replaced it on the sensor, so that the pump thought a new sensor had been popped in. The Enlite Sensor pack insert confirms that there are "no moderate or severe ... adverse events or unanticipated adverse device effects after seven days of use."


Calibration

Warm-up takes a while, up to two hours. In our experience with two sensors, it's been pretty close to two hours. After that you need to bank a calibration point and then another one within six hours.

Regular calibration is then at least every twelve hours. Up to four calibrations in a 24 hour period were recommended during our introduction to the pump. There's no leeway in that twelve-hour timing. Come on! Bringing up a three year old is all about flexibility: we've tried to predict ahead and grabbing a calibration at a reasonably stable time (usually before a meal) but sometimes, that's just not been possible (particularly when her levels are so up in the air at the moment) and we've had to go with a calibration at a time of rapid change when we know we're going to introduce bias (i.e. it'll be a rubbish calibration point). But it's either that or you lose further sensor readings. No warning text after that twelve hour period, just no data. I imagine this feature is in there to smooth the path through the Regulators, but on a bad day when you've not been able to bank a stable calibration, particularly when you're having trouble over-night, why force me to put in a dud calibration point? Hopefully there won't be many bad days going forward :-)


Sensor Performance

Medtronic quote (in their pack insert) a Mean Absolute Relative Difference (MARD) for the New Enlite sensors as being 14.2% (Median is 10.5%). This is when compared to the "Gold Standard" YSI (Yellow Springs Instrument) BG reading. You could think of the YSI as a Rolls Royce BG meter: several orders of magnitude more expensive to buy and run, completely impractical for daily use but delivering the ultimate experience when compared to your Bayer Contour Link 2.4 or Accu-Chek Aviva :-)

I thought I'd take a look at MARD for our first Enlite vs our Bayer Contour Next Link 2.4's BG values (i.e. not a Gold Standard, but the one we rely on for treatment decisions nonetheless). I looked at nearly one hundred finger prick readings over the seven day period the sensor was worn (maxing out at 18 readings on our 2nd day) and compared each BG with the nearest sensor reading (i.e. always within five minutes of the BG time). I didn't feel I had enough data to look very deeply, for example, looking at performance at different BG ranges, but here are the mean  and median absolute differences:

So here are the headlines:

• Overall MARD was 14.5% +/- 14.9% (remarkably close to the quoted 14.2% +/- 13.0% MARD cf the YSI BG)
• Day 1 was pretty good (MARD 16.0%)
• Day 2 was our worst day (MARD 24.5%), thanks to having to calibrate at times of changing BG. Day 2 was not a Little D day...
• Day 7 was our best day, with a MARD of 7.3% :-) (i.e. the sensor didn't show signs of running out of steam at that point)

Please remember this is just one sensor with almost a hundred BGs, compared to the Medtronic dataset of 3818 comparison points. But, as parents who want to have confidence in the Low Limit Alert and Low Predicted Alert, it's good to see our sensor performance is comparable to the data that was good enough for the Regulators (i.e. 97.5% of Low Alerts were correctly called by the 640g).

I'll keep tabs of our sensor performance going forward, as accuracy is important for the effectiveness of the SmartGuard system, but also because we're having to pay for the CGM components out of our own pocket. After nearly two weeks with the system, I'm currently convinced its money well spent - we have definitely avoided hypos as a result of having the complete system and treated others at a more manageable level (i.e. above 3.4 mmol/litre).