Medtronic's Closed Loop System

Type 1 and LADA
1

It is the closed loop system that would be adjusting your insulin delivery, without your doing it manually. I do NOT like that because I know the sensor is not always accurate. If the sensor says your BG is 175 and your meter says 120 then the system will deliver insulin that you do not need, and then you may have a hypo. That is NOT what we need!

http://diatribe.org/medtronics-pivotal-trial-hybrid-closed-loop-system-now-enrolling-us?utm_source=diaTribe&utm_campaign=09dae2aa9e-diaTribe_Issue_86&utm_medium=email&utm_term=0_75cdadd67f-09dae2aa9e-411810781

2

I totally agree! The delivery of insulin when it’s not needed is a big worry but for me an even greater worry is the system shutting delivery off. I know that there are lots of customers who have success with the CGMS product and this might be a good fit for them. For those of us who can’t trust the sensor it’s downright dangerous. I am very surprised to read that 14 year old kids are being considered for the trial.

1 Like
3

I wear a Dexcom CGM and it is usually pretty close to accurate. I have episodes fairly regularly, however, when the sensor is not close enough to administer an insulin dose. Reading on this site, I’ve concluded that the MedT CGM is not as accurate as the Dex. Perhaps the accuracy of this system is much better than the current MedT CGM system.

I guess this trial intends to see if the system can safely handle real world conditions. I’ve found that the various closed loop trials do not target control anywhere near to my personal targets. This trial, for instance, targets 120 mg/dl for corrections; I target 83 mg/dl. I guess they are erring on the side of caution.

I’m thinking that I will not be interested in any closed loop system until it can demonstrate control better than I have now running in manual mode. It will probably the third or fourth generation of the artificial pancreas that will finally entice me to cross over from manual to automatic.

3 Likes
4

I understand the Dexcom algorithm is the same as what’s being used in the so-called artificial pancreas. I find it pretty accurate, like within 10 points of what my glucometer shows, about 70% of the time. But some of those other readings are far enough off to create some pretty serious problems if my pump were bolusing or correcting based on 'em.

2 Likes
5

  1. modern day insulins, despite being called “rapid”, last for up to 5 hours or so. that’s not going to work with a TRUE “closed loop” pump.

  2. Sensors measure interstitial fluid, so there is about 10-15 minute delay.

  3. MM sensors aren’t accurate enough for dosing, at this time

  4. I don’t look forward to a bulkier, SUPER EXPENSIVE pump that also uses glucagon to correct lows. I’m not sure that any exhaustive studies have been done on daily, repeated dosing of glucagon over the long haul.

  5. If they do decide to use glucagon, anyone priced that stuff lately??

  6. Give me another minute or two and I’ll think of at least 3 more “gotchas”.

1 Like
6

It’s a trial. How about we wait for the results of the trial before we shoot it down? I applaud Medtronic for using technology to find improved treatments for all of us.

2 Likes
7

That’s not what we are doing. We leave that up to the FDA.

8

FYI - I just learned that the 670 does not automatically deliver insulin to correct a high bg, it simply alerts and will prompt the user to take a correction bolus and recommends the proper amount of insulin to give.

1 Like
9

Karen, maybe I misunderstood the article?? The section saying “…the 670G pump’s software automatically increases/decreases insulin delivery to target a blood glucose of 120 mg/dl,” seems to mean the system can cause a hypo. What if the system indicates the BG is 190, but the glucometer gives a 100. Then the system automatically delivers enough insulin to lower the BG 70 points (190 - 120 = 70). Since the actual BG is 100, the BG would be lowered to 30. My use of the Dexcom did give me discrepancies as large, and even larger, than that.

10

Yes, that is what I read also. However, the pump does not automagically bolus a correction. It does adjust basal insulin according to the target, I understand that the target bg can be personalized. In my search it seems the hypo risk was reduced with this system. I understand the concern with the sensor being inaccurate and would not trust my closed loop gig for that reason.

I looked at the trial participant link in the article which is where I learned (among other things): [Editor’s Note: The 670G, unlike other hybrid closed loop systems, does not give automatic correction boluses. When a meter blood glucose value is above target, the system will prompt the user to take a correction bolus. In this version of the algorithm, the correction bolus is based on a target of 150, and the algorithm recommends the proper amount of insulin to give.]

Do read the Interview with 17-Year-Old Trial Participant (Josh), it is very interesting. And he said: "The 670G was amazing and I honestly can’t think of something that was bad about it. The Enlite 3 Sensor could definitely use a little work."

1 Like
11

Yes, that’s my impression from the article. It’s basically a way of dynamically fine-tunning your basal, with some refinements in how it alerts you about corrections. Instead of programming in 4 or 5 rates or whatever for a 24 hour period, it’s just setting those rates as you go along based on CGM readings. I wonder how effectively it can do this when there’s a significant lag between your current BG reading and when any insulin you take for it (or don’t) takes effect. I know a big part of the artificial pancreas project is to develop faster acting/clearing insulin to get around that lag-time problem. Still, it is kind of what we’re stuck doing now ourselves anyway, so I guess it’s not so crazy.

12

That seems promising! I use basal adjustments all the time as for it seems to work great for moderate highs/lows/ sketchy sites and other “fiddling around” with my BG. I kind of like it in that if I do a correction bolus, it’s on board and if I get a late rush of the last bolus or whatever, it can overload things a bit whereas a basal “nudge”, I can turn off when the wave hits and perhaps “untake” some of the correction, although the downside might be that the basal boost isn’t accounted for by IOB. To me, fine tuning basal would have to involve adjustments to the basal rate, rather than a basal correction and I suspect that the gizmo wouldn’t do that.

Besides having to elevate my targets with the AP, my other big question about it is how having an AP would affect the skills we have developed to manage diabetes. I have a pretty refined skill set that seems as if I’d risk losing if I put all of my eggs in a computerized basket. I can totally see that the parents of kids would benefit from being able to sleep through the night, etc. but those kids seem like they’d perhaps be underdeveloped in terms of developing a diabetes skill set they might need in the future , should the medical system fail to provide fancy gizmos for them/ us. And, of course, this leaves out everyone outside of the first world other than those rich enough to cover the expen$e of these things.

1 Like
13

Allowing for the fact that the Medtronic isn’t a full-blown AP, it’s an interesting question and relevant to both devices to different degrees. I’m one of those who got to pumping by way of a longish stint on basal-bolus MDI (and a longer one on R/NPH before that), so the increased bionicization of my treatment hasn’t bypassed all those fundamentals. I somehow doubt that even the AP will get you out of having to understand the principles of T1 management, but the basic desire it’s answering is for a set-it/forget-it treatment–that’s what a REAL fully functional pancreas is like, after all–and that “forget it” part always comes with a cost. My own resistance to AP hoopla comes from a recognition that it’s going to be expensive as hell, and given what you have to do to get health insurers to cover an adequate number of test strips, let alone god forbid a CGM, the number of people who are going to qualify for coverage for one of these gadgets is going to be small, and probably won’t include those of us who’ve been managing ok with existing technologies lo these last several decades.

14

When I experimented with Afrezza as my mealtime insulin, I stopped using my pump to dose for meals. My pump mealtime regimen was more involved than a simple insulin to carb dose. I also count protein and fat and then administer an extended dose to cover that. In addition I use a super-bolus where I borrow 1.5 hours of basal, deliver that with my meal dose and then pay it back with a 1.5 hour temp basal of zero.

So I used Afrezza for a month but finally concluded that my overall control was better using the pump than using Afrezza. I stopped taking Afrezza except for corrections, which I still do. Finally, to AR’s point, my memory of my more complicated pump meal dosing was already plenty rusty. It took some effort to recall my old routine and then to re-establish it. If I had gone a year or two on “automatic,” it would have been very difficult to restore my old routine. And in that time my body would have morphed to a different pattern and even if I did restore the exact same routine, it would have been wrong.

I think I would write a detailed explanation of my insulin dosing routine and store it in a safe place so I could refer to it if needed if/when I ever go onto an AP system.

Separately, I’ve been following the various artificial pancreas trials and I was a bit surprised that the targets they use are not nearly as ambitious as my current regimen. One trial uses 120 mg/dl as a BG target. I target 83 mg/dl. I would not be happy controlling my BG range to between 80-180 mg/dl. So I’m thinking that I will probably not jump to an AP system until its performance can exceed my manual one.

1 Like
15

<sigh /> Humans, they be some crazy ask primates. :confounded: :roller_coaster: :scream_cat:

Yes, there are a number of potential problems having the pump adjust even only the basal delivery rate based on the current CGM Sensor Glucose value. What amazes me is when people appear to assume that these risks would never occur to the people running the trial or be something which they would take into consideration when running the trial.

It also amazes me that folks often appear to assume that the technology they have tried is what will be used during the trial. I was partly expecting that someone would throw out that they’d tried the Medtronic Sof-Sensor back in 2008 and had horrible results using it so therefore this clinical trial was also a horrible idea. :smirk:

This trial is using Medtronic’s Enlite 3. I know nothing about that sensor. I’m not sure who is familiar with it or what changes it includes. I would not be surprised to learn that the Enlite 3, as one quote above put it, “could definitely use a little work”. But I also would expect that it is still better than the Enlite and Enhanced Enlite currently in use inside and outside of the US, respectively. And those sensors, when they work, can be very accurate.

One of the requirements of the study is a “two week run-in period” where folks will be essentially using the 670G pump as a glorified 530G. That is, both the SmartGuard and Hybrid Closed Loop functions of the 670G will be turned OFF. The only action the pump will be able to perform if the SG goes out of range is to sound an alarm.

It seems obvious to me that the point of this period is to be able to screen out from the study anyone who is not able to routinely get consistently accurate results from the Enlite 3. In other words, a serious attempt will be made to exclude those folks who could get an SG of 175 when their actual BG is 81 or whatever.

Another thing to consider. From the first paragraph of the Wikipedia clinic trial article (emphasis added):

[Clinical trials] are conducted only after they have received health authority/ethics committee approval in the country where approval of the therapy is sought. These authorities are responsible for vetting the risk/benefit ratio of the trial - their approval does not mean that the therapy is ‘safe’ or effective, only that the trial may be conducted.

In my opinion, one of the basic goals of this study is to find out just how bad things can go when the 670G is used in the “best of all possible worlds”. If you attempt to rule out almost all of the many things which can go wrong, just how much worse can this pump make it?

Obviously Medtronic is hoping the answer will be that their pump does not make things worse, but in fact makes them better. But I think it is also obvious that they are aware of potential risks and are, as a first step, very actively attempting to minimize them at this time.

You might want to also take a moment to look at the list of the 17 inclusion criteria and 25 exclusion criteria listed in the description of the study on http://clinicaltrials.gov.

I also feel that this study is anything but “pivotal”. Given the relatively small number of participants and the restrictions on who qualifies to enter the study, it seems unlikely to me that Medtronic expects the next step after this study would be approval for use of the 670G in the US. This study seems more like a study conducted in order to receive permission to conduct yet another study.

So, let’s wait and see, no? :sunrise_over_mountains: :question: :scream:

#####Appropos of nothing:
A link to a list of descriptions of the phases various drug/device trial process which I just wanted to stash away someplace where I could find it later. Clinical Trial Phase Terminology

16

[quote=“irrational_John, post:15, topic:46632, full:true”]

So, let’s wait and see, no? :sunrise_over_mountains: :question: :scream: [/quote]Sure, but it’s an interesting topic for discussion, no?

17

@DrBB, by saying AP, I was more stepping back and thinking “any” APish device. I’ve followed the AP stories pretty closely, partially for personal reasons as my daughter’s first pediatrician is Ed Damiano’s wife. I don’t see any of the AP models as coming close to “OEM” pancreases. We had a presentation from a gentleman who tried one and was in the process of going to work for the project at the Diabetes Unconference. He reported great results, eating all kinds of stuff and running up to 180 which is very good for a no-input device but not exactly the way I play the game.

I perceived some of the hostility towards Medtronic in the DOC which is not necessarily unusual. It seems that many doctors “push” them so they seem to come across as the big, bad corporate wolf selling less than optimized pumps that don’t have color screens. I agree I’ve seen the data that the Dexcom offers 5 better MARD (8 vs 13 I think? I could be wrong?) in comparable tests but 5 mg/dl to me is nothing and I get great results from Medtronic and, as far as corporatism or anticorporatism, Johnson and Johnson, who run Animas, dwarf Medtronic but, it seems, aren’t as interested in penetrating the insulin pump market, for whatever reason. I don’t mind calling even the 530s “kill switch” an artificial pancreas, even if it’s a primitive one as it’s the most artificial pancreas I can go buy now however I hop through extra hoops to keep my 523 instead, because I don’t want my pump turning off unless I tell it to.

I don’t know that many people using other systems who are kicking my ■■■ at BG management and I sort of think that success might be easier to achieve by thinking about what we do with these gizmos than by the this gizmo or that gizmo debate. There’s a lot of hay to be made with tactics and the gizmo debating sort of gets in the way of that. And the AP totally unplugs the “thinking” part which concerns me. The big target and a lot of passion is poured into pro-artificial pancreas advocacy by parents who are ground down by the brutal grind of loving their T1 children but, in “going live” the senses being dulled will be the kids. And where will that leave them when they’re off their parent’s insurance? Obviously, there will still be crafty veterans hanging around Tu with tips and pointers…

1 Like
18

My personal feeling is that while I don’t mind criticism of Medtronic, I am (obviously) annoyed by what I perceive as the sloppy sort of “I had (or read of) a bad Medtronic experience so everything will always be bad both for me & anyone else from now on” criticism. This is actually a very human way to think and an evolutionary biologist can make a compelling argument why “superstitious learning” like this can be so very hard to avoid. But, when I’m not the one doing it, it tends to press one of my apparently many buttons. :confused: Oh, well. :flushed:

Yes, but … The Medtronic Corporation (as opposed to the people who work for Medtronic I have met so far) tends to exaggerate very easily and very, very often. This makes it hard (for me) to root out the truth from their text.

Referring to the 670G as a “Hybrid Closed Loop (HCL) System” is yet another example of their dogged hyperbole. The 670G is clearly not “closed loop”. Incrementally more closed than preceding pumps, yes. But clearly there is still a long way to go to reach the point of a plug it in and forget about it Artificial Pancreas. Oh, well (again). It’s apparently just how they roll. :flushed:

My perspective is that many (most?) of us are sheltered from perceiving just how large a percentage of PWDs already unplug the thinking part even without being encouraged to by having an AP. At least that is my explanation for why the stated target goals the 670G sound so inappropriately lax to some of us. The pump is not targeted towards people who already achieve good control with existing methods. It’s more something for just about everyone else, no? :expressionless:

19

Yup, understood and I wasn’t trying to imply otherwise, just trying to be clear since the thread wants to pull in a couple different directions.

Speaking as a “Snap orphan” who is this very night going back to my Minimed, I looked at the other options very intensively these last few weeks, and Animas was first out of the box in trying to appeal to those of us in this situation. One thing that was clear to me is that the current state of pump-CGM integration is really not making it a huge selling point for me. I’m perfectly happy with the hybrid approach, using a Dexcom CGM and a Medtronic pump, since just being able to look at the CGM results in the same box I’m using to program boluses and whatnot doesn’t seem that big a deal. For me, being able to echo the CGM to my iPhone with Share is a better than adequate substitute for that–I don’t mind carrying another little thing around in my pocket if I basically never have to take it out to see the results. Just my preference; other peoples’ mileage may differ, of course.

Perhaps I should point out that I prefer all-tube guitar amps while I’m at it… :wink:

20

The you are going to LOVE the new Marshall-Med JCM830G the first all-valve insulin pump. Scheduled for release in Europe in 2016 (the FDA delay-pedal means about 2019 in the USA).

The pump features a black Tolex cover with optional plexi labels, a GEC NOS large-bottle KT66 reservoir driven by 4 Mullard NOS EL34’s. A pedal-controlled overdrive Bolus button - dial in 10 units and it delivers 11 (for those chocolate cake moments). 230/115V switchable mains supply with authentic Partridge output transformer (20 metre extension cable supplied as an optional extra). New 4 x 12 mm infusion set (2 angled and 2 straight in - connects to the pump with a detachable standard jack plug).

Comes with an optional set of on-stage monitors (not NHS-funded in the UK).

Joel

2 Likes