FDA approves Medtronic "artificial pancreas"

Approval for a Minimed 670G system in tandem with a CGM was given FDA approval six months sooner than had been expected. It will be available in the spring of next year. It will deliver basal insulin according to readings taken from the CGM. Patients will still have to bolus for meals and adjust insulin delivery for factors like exercise.

1 Like

Thanks @Jim2. I think this is already under discussion here though:

1 Like

It is a source of some significant irritation to me that they keep calling these half-way (maybe even less) systems an “artificial pancreas”.

While the advance is laudable, and certainly T1s being largely freed from having to fiddle with basals all day (if the system actually achieves that – we’ll see), it’s light-years from being an “Artificial Pancreas”.

I reserve that designation for the day I can, well, slap something on my body (or take an injection – smart insulins) and then forget diabetes… period. Confidently eat, exercise, get sick, whatever – and not be concerned at all about my BG. Nirvana is achieved when I can so completely not worry about it that I can ditch the CGM, permanently.

THAT would be an “Artificial Pancreas”.


Do you think if such a treatment ever became available it would be allowed to reach consumers? And all the billions spent on test strips and other diabetes paraphernalia would just be allowed to go up in smoke?

I used to have that kind of faith in the system. I lose a little bit more of it every day.


I wholeheartedly agree with you, @Dave26.

Would you mind reposting your comment in the main thread dealing with this topic?

thanks @DrBB we try to keep comments on one thread, but it can be hard. Discourse does have what’s called a “duplicate thread prevention” - you might notice while you are commenting, it pops up on the right hand side, saying “your topic is similar to…” and then a list of similar topics.

I do. I don’t abide conspiracy theory too much on stuff like this.

There is vigorous work in this area. Also, I don’t think it will happen in one fell swoop – it will gradually develop incrementally, as we’re seeing. First, automating basal.

Long before the sort of “Artificial Pancreas” I’m talking about, there will be increasingly automated closed-loop systems. Just like buggy-whips, current products will become obsolete, and fade from use.

I don’t think insulin analogs were delayed out of concern for preserving the Humulin and NPH markets, or their producers.

Insulin pumps and analog insulins are 1990 technology. id argue that hey are no longer the most effective treatment or even the most effective avenue to pursue advancements. My A1C supports that argument. Yet they are not fading from existence-- and the superior technologies such as afrezza are being suppressed in the marketplace in my view. I guess that makes me a conspiracy theorist.

1 Like

That was my first reaction to the Med-T 670G system. I’m ready for a much more advanced system. I am also put off by the unadjustable 120 mg/dL target that’s hard wired into the code. In fact my comments yesterday were cautiously positive referring to this incrementally improved system as a good “baby step.”

This morning I read the comments of @Dragan1, a member here who has lived with, for more than one year, an OpenAPS and now the Loop do-it-yourself systems. These systems use temp basal rates to control BG levels while the user still programs all boluses. @Dragan1 disagreed with my characterization of the 670G system as merely a baby step. He has lived with the benefits of a basal adjusting system and has seen incredible results. For those of you who don’t regularly visit the Flatliners Club, I’ve copied and pasted a recent graph that Dragan posted there. These are the best 90-day results I’ve ever seen posted by someone with diabetes:

This is the 90-day Dexcom Clarity standard-day report. A standard day overlays BG results one day on top the previous so that the entire period is displayed as if it was a single day. It’s one thing to trace a graph like this for 14 or even 30 days but 90 days is superlative.

I think it’s a big deal getting through the night within a relatively tight range. I found good nights generally lead to good days BG-wise. Now I know that Dragan must put a good deal of effort into his daily BG program and he likely exerts good discipline with eating and exercising. That line is a thing of beauty to me. Perhaps the 670G is a bigger deal for us than I at first thought.


What was your A1C before and after Afrezza? I thought it was virtually the same and was really low to begin with even on novolog.

How do they get such a flat line when the Medtronic trial ones I’ve seen are a roller coaster? Whats the secret?

It hovers in the low 5s both before and after but that’s meaningless without the context that I’m now barely trying and before it was consuming my entire life

I don’t have a complete answer here. Perhaps @Dragan1 can answer more fully. My understanding of the 670G system is that it will turn off basal when BG drops below a low threshold or when the system predicts dropping below a low threshold. My limited understanding of the OpenAPS and Loop systems is that it continuously assesses whether to set (+ or -) or cancel a temp basal rate.

I understand the 670G system to be less active than the OpenAPS or Loop systems. But eliminating lows, especially overnight, will also eliminate the rebound highs that resist response to correction insulin. It would, in my estimation, reduce its propensity to roller-coaster BGs.

I also think that studies like the one that supported the 670G used a population with a wide range of existing control. I think people in the higher A1c ranges of > 7.0% tend to have more variability in their glucose profiles.


Right I totally understand. I try very hard but still dont get those numbers, its very frustrating. I think I have a plan now so hopefully it works.

Thanks for sharing your numbers Sam.

As far as I’m concerned, this is NOT an artificial pancreas. If you are expecting it to function as one, you are in for a world of disappointment.

Well, as with anything D related, you can’t assume that individuals with access to the same tool will observe the same outcomes. Large differences among individuals can be explained by three groups of factors: (1) objective factors, primarily the remaining endogenous insulin production, (2) diet and lifestyle (e.g. exercise), and (3) motivation, attitude, aptitude, knowledge.

I can only speak for myself. First, thanks to CGM, I’ve learned how to (manually) bolus for pretty much everything I like to eat, with minimum or no spikes. My diet is not very restrictive at all - about 100g of carbs per day on average (but I do not eat junk carbs). Second, I run a DIY AP system (now Loop, earlier OpenAPS+oref0), which continuously adjusts basal rates on my pump. In essence, every 5 minutes AP looks ahead and predicts eventual bg in the DIA (duration of insulin action) interval. It adds to or subtracts from the scheduled basal rate to drive the eventual bg to a preset bg target - just like making a correction every 5 minutes. I have no idea how 670G works, but I’d surprised if it does something radically different. The system is particularly effective during night, which is when it matters the most. Some nights I do not see much AP action at all, which reassures me that my pre-programmed basal rates are not too bad. But, there are nights when AP adds or subtracts significant amounts, for whatever reason - I no longer even try to explain these variations. As @Terry4 noted, consistently waking up to a non-D bg is a good way to start a day. I used to put lots of effort into my bg control, and had A1c hovering around 6 for many years. Over the past year, my A1c’s have been 5.3, 5.0, and 4.9, with much less effort. Yet again, experience of a single person means nothing. Just find whatever works best for you. Aside from pumps or emerging AP systems, there so many excellent options available now, e.g. phenomenal basal and inhalable insulins, and/or more restrictive diets …


[quote=“Terry4, post:9, topic:56434, full:true”]I think it’s a big deal getting through the night within a relatively tight range. I found good nights generally lead to good days BG-wise. Now I know that Dragan must put a good deal of effort into his daily BG program and he likely exerts good discipline with eating and exercising. That line is a thing of beauty to me. Perhaps the 670G is a bigger deal for us than I at first thought.
[/quote]That graph is amazing! Surely there isn’t much in the way of carbs in there!

In any case, I take no issue with the significance of the “step forward” for the 670G. I think it’s a wonderful advance.

My issue is with calling it an “Artificial Pancreas”. It’s nothing of the sort.

Like I said, the device that allows me to forget I’m a diabetic, or at least an otherwise healthy T1 forget, will be an “Artificial Pancreas” to me.

He writes above:

I agree that the 670G is not what we had in mind as the the idea of an artificial pancreas dawned on us over the last 10 years or so. This will be an evolutionary process but one that will likely unfold more quickly than the pace of D-tech advances even in the recent past. The DIY movement will serve as a productive spur to keep things moving in the patients’ interest.


My system simply isn’t capable of that, unless I dribble carbs in in tiny amounts through the whole day.

I can’t have more than 25g in a meal without spiking at least to 135-140, assuming I’m starting around 100 (where he seems to be), no matter what sort of bolus strategy I use with my pump.

But then, I’m fighting IR with an IC of 1:5… that makes a bit of a difference in terms of processing glucose.

We each have a unique hand to play. I have a breakfast I:C of 1:4 and a dinner I:C of 1:6. But my TDD (total daily dose) of insulin is 35-40 units.