The Great Insulin+Glucagon Dual Pump Experiment

Alright, but let me throw the idea out to the Unlimiteds. Work with me on developing a useful test of a pump that has insulin and glucagon available. What would you like to see? What things should I test?

It will actually be 2 different pumps, but just as a proof-of-concept of the idea of a single pump delivering either insulin or glucagon.

Maybe someone can post some use-cases to run through.


@Eric I know I mentioned this before, but you should take a look at these guys

One of the modules they use in their artificial pancreas is a glucagon pump.


Here is a 2016 write-up in Diabetes daily:

But I get the feeling that they are stalled. In 2016 they hired Wilson Sonsini, which means they needed help raising funds. A year later, no fundraising – bad sign. I don’t think they are going to come out with anything. I hope I am wrong.


Currently available dissolved glucagon is unstable, isn’t it? (should be administered immediately according to the instructions). So, you may have very little time available for an experiment, not sure how long. But, if you are up to it, I guess a use case would be an attempt to flatline a super-high carby meal using a super-super-insulin bolus followed by a carefully timed glucagon bolus some time later. (I hope you are kidding about the experiment :wink:)


I think I see where this is going… (or maybe I am just fantasizing about such a system being available.)

-Two pods - one with Fiasp and one with glucagon.
-Microbolus glucagon and record blood sugar response over time
-Overbolus with Fiasp and drive blood sugar low. Bolus with glucagon to stop low.

1 Like

I was told by the hospital it is stable for about 24 hours but they suggest putting it in the fridge between doses. Here is the handout they gave on how to deal with it:

EDIT: In all seriousness, this handout is actually pretty useful for doing mini-doses of glucagon. It gives some guidelines on how much is required to raise BG a certain amount and provides the conversion from micograms of glucagon to units on an insulin syringe.


I did a demo on this a while ago. I can easily get 3 days out of a vial of glucagon. I posted about that. Since pods only last 3 days, the viability of glucagon is not a problem.

Here is the post on it. The whole thing thread is informative, but here is the part on how long it lasted after mixing.

No, I am totally serious. This is really nothing too crazy.


I have no doubt.


Oh, that’s interesting!? Well, it would be fairly straightforward to modify Loop algorithms and setup to work with two pumps (or two pods, once they develop necessary drivers). In the algorithm, dosing calculations routinely produce negative numbers, which are then clamped to zero, because that’s the best we can do with an insulin-only pump. Instead, we would just divert the negative micro doses (with an appropriate scale factor), to the glucagon pump - done!


I agree with @Eric. This makes sense. He already reported on glucagon stability. The only difference here is that it will be on the skin, so possibly a bit warmer.

The biggest issue is the mini-glucagon dosage. But @Eric also did some experiments with that. @Eric, do you still have the data on dosage and mg/dl change?

1 Like

@Eric - What is the oldest Glucagon that you have used and determined that it still works. Similarly, have you ever used an old Glucagon that you determined either didn’t work or was significantly degraded?

(In terms of dry unmixed - years expired. Not about how many days of a mixed solution.)

I gather from some of your “feeding frenzy” posts that you only use Glucagon for testing purposes but not for everyday treatment purposes? (I still have the mental image of the chocolate cake - or what was left of it…)


A lot of different comments, trying to put them all together in a single post

I have been interested in this idea for a while. That was the idea I saw a while ago, it had sections you could remove and replace for insulin and glucagon.

Thank you, I will check it out.

I didn’t do them very “micro”. I did them as 5 unit doses a few years ago, and it didn’t work, but that’s because I was running. I kept doing them until I had taken the entire vial, and it never did a thing. But a while ago (it was July) I did 25 unit doses, and that spiked me like crazy. So I definitely need to fine-tune this thing.

The one I did from July was 3 years past expiration. I really don’t think degradation is an issue.

I have mainly done self-dosed glucagon as an experiment. Tried it to see if it could be used to treat a low after long distances of running (it can’t, your liver glycogen is depleted!), and also the expired glucagon testing. But for practical purposes, if I am awake, I will opt for the chocolate cake type of treatment every time!


Here are some thoughts I have for the experiment:

  • Make my PDM screens obvious. The words “INSULIN” and “GLUCAGON” as the PDM identifying screens, along with different color gel skins and different background screen colors, so I know which PDM I am dosing from.

  • First day getting good baseline levels for micro-dosing of glucagon

  • A test where I can have a programmed increased insulin basal from, say 1-3, and then a programmed increased glucagon delivery from 2:30 - 3, and see if how it counters the low that would be in progress.

  • An increase in my normal insulin basal rate that coincides with an increase in the glucagon delivery rate. Personally my liver doesn’t spit out much glycogen during the day unless there is a stress stimulant. I am interested in seeing how a glucagon stimulant that causes the release of liver glycogen, and having it matched with a corresponding insulin delivery would feel. A constant source of fuel and insulin so it can be used! Something I suspect diabetics with a lot of years never get to feel.

What else?

1 Like

:blue_heart: Our own FUD :guinea_pig:


We’d really need to determine a glucagon absorption curve, so we can perform glucagon on board (GOB) calculations. I’d suggest an experiment during a fasting period when you are in range, and when you’d expect bg to remain approximately flat. Take no extra insulin, just keep standard basal rate. Take a moderate glucagon bolus in the amount you’d expect would increase bg by say 50-100 pts and record the bg response - download actual bg data points from Clarity or whatever tool you may be using. Based on this experiment, we could then construct approximate G activity and GOB curves for future use, either manually or in dual-hormone closed-loop systems.


From what I found over the summer, there isn’t much of a curve to it. It’s pretty much a line pointing straight up.

25 units bumped me 106 points in 35 minutes, and a different time it was 89 points in 30 minutes.

So I will be using much smaller amounts this time around.


Those are good data points - thanks. So, it’s all gone after about 30 minutes, with no further impact on bg observed? (that’s pretty amazing, wish insulin worked like that)


I can’t really say it was all gone because I countered with insulin at that point. I don’t like high BG, so I took insulin to bring it down soon after I saw the spike.

The expired glucagon test was just for the purpose of testing old glucagon.

With the pump test, I will be using smaller doses of glucagon, so I will be letting it ride a bit longer.

And I will try out smaller increments - like 1 unit, 2 units, 5 units…

I may try something like a flat start and then 5 units of insulin to bring a drop and then 5 units of glucagon so see where I level out.


@Eric - How much can you get out of a liver? Do you have any idea if that would be similar person to person of a similar body weight? How do you quantity how much you can get from a liver? Would it be in total BG units raised? Or carb equivalents?

Assuming a healthy liver, fully loaded, no extreme exercise.

1 Like

It is measured in grams of glycogen. The amount of glycogen an adult has stored in their liver is about 90-110 grams of glycogen.

It varies a bit based on activity, metabolism and a person’s eating habits. But it is very similar from one adult to the next.

The amount of glycogen a liver stores is much closer from one adult to the next compared to their muscle glycogen storage! For adults, muscle glycogen is about 350-500 grams, depending on muscle mass. That is a much bigger range than the 90-110 grams in the liver.

The two main storage sites for glycogen are liver and muscle.