The Medtronic Mothership poised for takeoff |
I was particularly pleased to be invited to walk in the footsteps of the legendary Alison and Tim from Shoot Up or Put Up and to needlessly and completely unnecessarily rekindle the rumours that Medtronic have an entire department dedicated to the drowning of puppies in their relentless pursuit of evil. Rumours which are, of course, entirely untrue. And completely made up by Tim. Or are they?
We arrived on Thursday afternoon and were swiftly transported to the impossibly quaint, beautiful and strangely fragrant 'Chalet Suisse' restaurant where we were treated to a slap-up meal composed of cold meats, pickles and industrial quanities of gently bubbling cheese into which we enthusiastically dunked hunks of white bread. Fluffy white bread, a vat of cheese and a little alcohol. What easier meal could there possibly be for a tableful of pancreas impersonators to contend with? As inevitably happens at these events, people with diabetes quickly feel at ease with each other and conversation flowed to a burbling soundtrack of bleeps and buzzes from various bits of diabetes kit. It was a very lovely evening. Thanks Medtronic!
All those friendly faces and barely a functioning pancreas between them. |
It was perhaps no surprise that Medtronic had invited us during this particular week - in which their latest hybrid closed-loop insulin pump system launched in Europe. The launch begins with the UK, the Netherlands, Belgium, Sweden, Italy and Slovenia as of 10th October, with Finland and Denmark following hot on our heels next week. I have been using Medtronic pumps for the past 7 years or so, but others around the table used other brands, pods or MDI injections.
Inevitable device timeline pic. No future dates, naturally. |
The MiniMed 670G adds extra oomph to the mix by also having the ability to increase basal insulin where sensor glucose is rising. Gary Scheiner, who wrote 'Think Like a Pancreas' wrote a review of the MM670G after it launched in the US which has a good deal of detail if you are interested, but which also prompted several questions for me.
What I thought I knew
The MiniMed 670G uses two main targets, neither of which can be changed by the user. Overall it aims to direct your glucose value towards 120mg/dl (6.6mmol/L). Where it is taking corrective action for elevated sensor glucose, it aims towards a target of 150mg/dl (8.3mmol/L). These were the launch values in the US, and will be exactly the same for the UK version.
A number of options that experienced pump users might expect to be able to use (temporary basal rates, dual (combo) wave, and square (extended) wave boluses are unavailable in auto mode.
From what I can tell, things like differing basal patterns for weekdays/weekends do not apply in auto mode either. Nor can you alert the system that you are about to undertake exercise in advance.
In fact there seems to be very little you can adjust and tweak in auto mode aside from insulin:carbohydrate ratios and Duration of Insulin Action. If you are planning to exercise you can temporarily set the pump to aim for the higher target, but that's about it.
All of this may fill you with dread and impending frustration. Or it may elate you with gratitude of being able to hand over responsibility almost entirely.
Users can, of course, use the pump in 'manual' mode whenever they wish. This would give them effectively the same choices/options as the MM640G, but doesn't take advantage of the additional automation on offer.
Is there anything else like this?
No. Well... not regulated, approved and currently commercially available no. But of course the MM670G is not being released into a vacuum. There are other emerging sensor augmented pumps (eg Tandem T:slim) which incorporate predictive low glucose suspend (rather like the MM640G). Alongside which there is a growing community of enthusiastic loopers who have decided #wearenotwaiting and use a variety of homespun, and significantly more tweakable, options. Of course not everyone has the confidence to build their own diabetes device - I am assured that these are no longer cobbled together from bits of string and paperclips, but nevertheless still do require careful configuration, software and in some cases additional DIY hardware. Additionally many loop solutions currently require older out of warranty insulin pumps. This may not be the case for long, nor might OpenAPS and Loop's unregulated status continue, now that Tidepool have committed to developing Loop as a supported, regulated app available for users on a range of in-warranty devices.
What I found out
It's fair to say that I was very enthusiastic about the prospect of the MiniMed 670G about a year ago. But things are moving so fast these days, that this is already beginning to feel like an older model. And as an inveterate fiddler the idea of so few settings being settable by me is a bit... well... unsettling.
What I found very interesting, was the observations Medtronic have made from 3.2 million days of data since its launch in the US. Essentially, they suggest the MiniMed 670G tends to work better in people who don't fiddle and try to come up with their own ways of working around not being able to alter settings. Often people who are new to an insulin pump, who have not already built up a library of hacks, tricks and strategies to bully their pump into doing what they want it to. People who simply give the MM670G the information it requests and heed the alarms, who keep the sensors well calibrated, and who count and supply carb information of meals and then let the pump do its thing, are the ones who tend to get the better results.
I am aware from my own relentless tweakery that at least some of it is frequently down to constant changes in insulin sensitivity and insulin need (and working around times when my pump settings aren't quite right). I suspect that one of the main reasons behind the lack of tweakability on the MM670G is the closely guarded algorithm. This updates up to 17 different parameters at midnight every single day, based on the previous 24 hour's worth of data along with a rolling average of the previous 5-6 days. Essentially the pump is continually adapting and tailoring itself to you. So those unmovable targets may not seem to be what you have chosen, but the intention of the MiniMed 670G is to keep your BGs as stable as possible to increase your time in range, to reduce glucose variability and to reduce hypoglycaemia to almost nothing. At least... that's the idea. Medtronic are quite upfront that this is the first iteration of this algorithm, based on dozens and dozens of research papers, and that it is intentionally cautious. Other more 'aggressive' versions with greater user control are being strongly hinted at.
A consequence of the 'machine learning' of the pump adapting to you as an individual though, is that it does take a few days of sensor wear for auto mode to become available (48 hours from midnight following sensor start). I'm not exactly sure how that would impact occasional-sensor-users like me, who cannot afford full-time coverage, but it's clear that only a portion of each sensor for me would be auto-able. In theory sensors can be restarted just as they can on earlier enlite-using pumps (sorry Medtronic!), but this newer pump uses Guardian 3 sensor technology, and because the MiniMed 670G has more control over insulin delivery it makes more checks that the sensor data being received is of high enough quality. Without having used it, it's impossible to tell whether it would get twitchy and drop restarted sensors earlier than I am used to.
Case study
It was a delight to meet Rob Howe a Texan ex-pro basketball player and MiniMed 670G user. Tall, slim, athletic, and with that easy, effortless charm that seems to be the gift of so many Americans. He shared his experiences and took part in a Facebook Live broadcast after which we were able to ask more questions.
Rob was an enthusiastic MM670G poster-boy, and having moved from MDI/CGM to an insulin pump with an impressive HbA1c of around 6.2% (44mmol/mol) he was obviously no slouch when it comes to BG juggling.
When asked about changes he had noticed during his year on the pump, one of the most striking was that his regular weekly order of two 1 gallon containers of orange juice he kept in his fridge to treat overnight hypos started to go off before he could use them. They simply weren't needed any more. As a basketball player and all around active type, he also seems to have put the MM670G through its paces exercise-wise, from high intensity BG boosting sprints/weights to longer distance runs of 6-10 miles. Rob was able to keep the pump connected throughout exercise and allowed it to manage his activity, with great success it would seem. Initially Rob said his HbA1c rose a little, perhaps up to the mid 6's (48mmol/mol), but seems to have settled back to the low 6s since. So while the fixed settings on the MM670G algorigthm seem to target an HbA1c of approx 6.7% (50mmol/mol), it looks as though, once again, Your Diabetes May Vary.
What next?
It will be fascinating to see more accounts from real users 'in the wild' across Europe. Several people in the group were about to start with MM670G in November. There was an extended discussion during the end of the afternoon where bloggers raised concerns and questions. There are plans to subtly alter the materials and structure of the MM6xx cases to help avoid those battery compartment cracks (which in turn will mean Medtronic don't have to replace those pumps), plus small alterations to clips, and clip rails. Medtronic are planning a version of the MiniMed 670G which will be able to communicate to phone via Bluetooth. It is notoriously difficult to predict how long it might take to get devices through regulatory approval (and goodness knows what additional chaos Brexit might add to the mix for us in the UK), but reading between the carefully veiled lines it seemed that they were hoping it might be in the next year or two. No promises etc etc. Perhaps even more interestingly... this version of the MM6xx series would have the ability to receive updated software/firmware after distribution. This is something that Medtronic have been inching towards for some time, and could significantly accelerate the takeup of new and better pump versions where the hardware is essentially identical, and it is only the algorithm which needs updating.
One thing is for certain. The rate of change in diabetes devices doesn't look to be slowing down any time soon.
Disclaimer. Medtronic Diabetes invited me to attend the MiniMed 670G European launch in Switzerland. They paid for my economy travel, transfers, accommodation, the welcome meal and lunch at their offices. I wasn't asked or paid to write this blog post or any other social media posts connected to the event. They also included a nifty little Medtronic branded battery power pack thing in a 'welcome' bag. Oh and a pen too.
Hi Mike,
ReplyDeleteBrilliant post - thank you so much for the update :)
How wonderful to have been invited to Switzerland for the launch too :)
It's great that commercially available hybrid closed loops are finally getting closer. The Open APS group have some great ideas about coping better with meals - I listened to a podcast with Dana Lewis recently. She was talking about the importance of timing the meal bolus (20 minutes before eating) and instead of specific carb counting, simply estimating whether it was going to be a low, medium or high carb meal - which sounds more practicable. Her initial results with it sounded excellent.
This podcast from Diabetic Tim is pretty cool too:
https://www.diabetictim.com/028-type-1-diabetic-for-over-40-years-no-complications-michael-griffith/
It really illustrates the importance of regular exercise in maintaining good control.
I've been using the Dexcom G6 for several months now, and it's brilliant. The predictive low alarm is great :)
I'm still on levemir bd and humalog, HbA1c 41, no hypos.
Well done with all the work with CCG funding for CGMs. I gave a lecture on optimising insulin with CGMs at a national pharmacy meeting earlier this year and there was a huge interest in CGMs. Several of the pharmacists pointed out how CCGs are often hampered by short term thinking, and that the cost of complications will come out of somebody else's budget at a later date! Many of the US 'birth to grave' insurance companies do seem better than the NHS at long term preventative thinking at the moment.
With best wishes,
Ian