A bit of Baz Luhrmann seemed an appropriate title, as his advice, “Wear sunscreen”, was obviously oblivious to the issues connected to sunscreen, swimming pools and being on holiday in general, whilst trying to closed loop.
Having spent a week away in a location that was warm and sunny (average daytime temperature, 33 celsius) and somewhat humid, I thought that reviewing how you live with a loop in a scenario like this would hopefully be of use to somebody out there.
As always, when going on holiday, we take bags full of stuff with us, and I’m no different, with spare cannulas, pumps, CGM, insulin, in fact everything I need plus one of at least everything, just in case. And this was no different. We were off to Ibiza, so I also knew it would entail a lot of time in the sun and pool, so maybe n+1 should be n+2, well, just in case.
So let’s take a look at what typical days on holiday entailed, and what the end result was.
An average day…
Essentially, each day entailed getting up at some late hour, going to bed at some equally late hour, and in between times, covering oneself in factor 50 and hanging around both in and out of the pool, being in the water for an hour at a time, as well as being present for long periods of time in what could be humid and sweaty establishments.
I’d put a new Dexcom sensor on two days before leaving, with the intention that this would last until I got home.
Likewise, I had expected to need three or four cannulas.
I’d packed a spare dexcom sensor and plenty of spare cannulas. And it turns out that the heat, suncreen and water were to prove too much for all of the above, and I’d need to change them more regularly than I expected.
Starting with the Dexcom, after 7 days, I was reduced to adding overtape on the sensor, which was becoming unstuck everytime I went in the water. The additional grease of the non-greasy sunscreen simply destroyed adhesion, and I just about managed to get seven days out of it.
What you can’t see in this image is the loose Dexcom adhesive pad all the way around the sensor, but hopefully the overtaping shows some of the issue.
Sadly, it had to be replaced at this point, however, I was fortunate to have identified that there was an issue and put a spare in place, as every shirt removal was a bit like Russian Roulette. Would the sensor stay in place?
Of course the other piece of the puzzle when jumping in and out of the pool was the time to recovery of the sensor data to the CGM receiver, which in the case was both xDrip on my phone and xDrip-JS on the OpenAPS rig. In general, with the G6, this turned out to be remarkably quick. After 45 mins to an hour in the pool, where the bluetooth signal is stopped by the water, in general, both systems would pick up the next or next-but-one scheduled signal, which I was impressed with. Having used the Medtronic Guardian 2 system jumping in and out of the pool previously, the G6 was much quicker at recovering and needed no intervention from me to resume.
So far, Sunscreen, water and sweat 1.5 : Diabetes devices 0.5.
Pump cannulas and insulin
As you might expect, the adhesive on the pump cannulas didn’t stand up as well to the conditions and soaking as I might have liked them to. At between two and three days, I was seeing a loose, unsticky adhesive pad. I found that the Mio sticky pads lasted longer than the Sure-T ones, with Sure-Ts ready to fall off at just about two days, where the Mios seemed to make it to 2.5 days.
Sunscreen, water and sweat 1.5 : Diabetes devices 0.5
On the other hand, insulin in the pump wasn’t such a big deal. Both the Humalog and Fiasp that I used through the holiday managed to maintain efficacy and I didn’t see any issues towards the end of reservoirs (at about 4 days with Humalog and 3 days with Fiasp due to increased insulin needs with the latter).
Whilst I took care to keep the pump out of the sun, either in the shade when off my body, or under the side of T-Shirt when wearing it, the surrounding air temperature was pretty warm, but not dangerously so, and this seemed to reflect in the insulin life.
Looping whilst in and out of water
With the knowledge that I was going to be taking the pump off a lot, I changed my usual meal management approach using the loop.
Normally, I’d take less insulin and then leave OpenAPS to SMB. Knowing that post meals I was going into the water, I operated on more of a “Superbolus” type model, knowing that when the System was getting data, it would likely suspend and when it wasn’t, it probably wouldn’t matter.
Typically, I’d do a full meal bolus at a sensible pre-bolus time, check IOB and glucose values before getting in the pool, and then either use a single “easy bolus” click before getting wet, or depending on the OpenAPS output, decide to leave it until I next got out.
The point here is that the data you have from the loop system you use gives you much more ability to fine tune your approach to suspending the pump and when to give additional insulin. The data is invaluable in popping the system on and off and generally deciding when you need to top up.
Obviously, OmniLoop users have a different thing to contend with, as the Omnipod will revert to programmed basal at the end of the TBR, but I’ve not had experience of this.
Overall though, this made managing my time jumping in and out of the pool a great deal easier.
In this particular scenario,
Sunscreen, water and sweat 0 : Diabetes devices 2
Looping in clubs/sun/beach/etc
First up, I’ll put my hand up, I;m not one for spending a lot of time on the beach, so sand ingress into my OpenAPS rig wasn’t something I was overly concerned about.
We did go to clubs, walking in the sun and as I’ve already mentioned, the pool. In fact, this is my rig enjoying a bit of Pacha:
As I’ve already mentioned, I was using OpenAPS with xDrip-js, including the FakeMeter function, which pushes the reading from the rig to the pump. The xDrip-js link was on the alternative bluetooth channel, while xDrip on my phone carried the standard one.
In densely populated areas, like the picture above, I found that Dexcom-to-phone/xDrip data wasn’t always that reliable, whereas the Dexcom-to-rig performance was considerably better and there was never an issue with the fakemeter functionality, presumably as a result of a large number of phones in the area broadcasting on similar bluetooth channels causing interference, but not on the alternative channel.
The other benefit of xDrip-js is that when the sensor is changed, OpenAPS still has a source of data while the system is warming up, and if you’ve put the calibration code it, aside from two or three jumpy, noisy readings at the start, on a sensor that’s been embedded for a few hours, it seems to be similarly accurate as once it’s through the warm up period, or at least, based on my testing, accurate enough. This is one of those areas that you should consider with care, as your experience may vary.
As the xDrip image below shows, OpenAPS was able to dose insulin while xDrip considered the sensor to be warming up.
Sunscreen, water and sweat 0 : Diabetes devices 2
So the holiday worked out alright then?
It did. The limitations that I faced were really all physical and are generally faced by anyone in a warm, extensively sunny environment. Having a self-contained OpenAPS system with the use of xDrip-js was extremely convenient and meant that I had very little concern about the loop continuing to operate, and it meant there was always a conduit for sensor readings if I wanted to check.
As it turned out, the only time in the entire week’s holiday that I did a fingerprick was when I changed the sensor, just to check that the interim data, when warm-up was still going on, was aligned. I hadn’t realised how little I was now fingerpricking when using this.
And the clinical outcomes? Well with an average daily carb amount of 206g and TDD of 66.6U, I’ll take the following stats. Especially given the approach I’d settled on to manage the pool.
And overall, based on the scoring, the totals are:
Sunscreen, water and sweat 3 : Diabetes devices 5
As you can see, the human/system mechanical interface is the biggest issue, but if that’s well managed, certainly the devices that came with me on this trip survived very nicely indeed, and I’d far rather be holidaying with them and the benefits they provide than going back to basics.