Sick Days and CGM

Janki has spent the last week or so trying to shake off a cold and cough. Everyday stuff for three year olds. Nudge high temperatures down with Calpol and follow them around picking up tissues. Job done.

Well that's the way it used to be, but adding T1D into the mix and you get chaos! Those familiar with the impact of a busy immune system and T1D, particularly in children, will know the routine... You'll usually need more insulin, but sometimes less... Maybe 10% change. Or 20%, Or even 50%. If your PWD spends too long above 13.9 mmol/L without enough insulin sloshing around you'll see ketones. And that can lead to diabetic ketoacidosis (DKA). Not good.

So before we had Janki on her Enlite CGM with the 640g, we relied on lots of blood glucose (BG) tests and, initially, urine ketone test strips.

The first significant leap in management of sick days for us was getting hold of a GlucoMen ketone meter: pop a test strip in and take a reading from the same finger prick you've just used for a BG and you have a real-time, accurate ketone reading. Fantastic. Whilst the urine strips were helpful in guiding us and Janki's clinical team as to when we might be heading into trouble, they show what was happening a while before the test, and accuracy is not fantastic (low / medium / high was about what we managed).

Add Enlite CGM into the mix and we've found we have the potential for much better control. Whilst you still need to BG to confirm the sensor glucose (SG) readings, we can (usually) rely on the CGM to tell us when highs (or lows) have turned and track correction boluses more frequently.

What does that mean in reality? Well, for us it's meant we have been able to treat hypos more "gently", trying to avoid the rebound of hyper and hypo bouncing through the day (or night). With alerts on for approaching (or at) high SG, we can correct earlier (again confirming with a BG first).

Knowing the general BG position (low, medium, high) and the rate of change of sensor glucose readings has allowed us to find Janki's sick-day rates quicker this time than on previous occasions.

Our experience with the 640g and Enlite sensors has been generally very good, as I've written about already. But occasionally we've noticed things trip up. Last night, for example. we had a suspend before low, monitored with a couple of BG's showing we'd actually rolled level at 4.1 - 4.2 mmol/L. Great. I decided to sit and watch the SG (which was tracking closely) to ensure Janki's glucose levels turned round. Five minutes later we lost signal - "SG Unavailable". Okay... Five more minutes and the reading's back - 2.8, two arrows down, low alarm pinging. Time for another BG; 4-point-something. If we'd woken up to that alarm and seen 2.8, two down arrows, we'd be pushing glucose in, maybe even before or at the same time as a BG depending on how clearly we're all thinking (waking Janki up for Ribena at 3am is no fun for anyone). But in this case, no glucose was really required and we would have most probably ended up trying to correct out a high bounce later. To me, it looks like the sensor algorithm may be hoovering up the missing SG (ISIG) reading as part of it's smoothing process, which results in an underestimate of SG and an apparent downward trend?

More on that in the next post...