Thank you all for being here. I am Sue Brown. I am an endocrinologist from the University of Virginia and I Am very thankful that you all here. I'm sure you're all trying to catch your flights and what-have-you so I'm excited today to talk about Practical management of closed-loop insulin delivery devices I do a lot of work in this area for clinical studies And there's a lot of really important advancements have been happening and a lot of new devices coming out and along with those advancements There's a lot of confusion about how to use them, and they all have different nuances so what I hope this Presentation is going to be about is About what some of those nuances are and specifically what buttons can I push on these devices? So I'm gonna kind of orient my talk toward that you know when someone Is before you and in clinic, and you're trying to help manage. How can I optimize some of these systems? What what buttons can I push? so with that Here are my disclosures, I do do a lot of clinical studies. I have no personal disclosures. These are research support to UVA and I have several. I hope this is interactive session, so please everybody take a minute. Let's grab the polls I'll have a bunch of polls throughout this to Have us think together about some of these challenges And the way I'd like to divide up this talk I'll spend the first half or so just making sure we're all on the same page Just in terms of what the devices actually are what how they're designed What buttons you can actually push and then the last half let's go through three different cases Using the three major devices that we have out there right now And we'll just kind of work through a couple of those cases to illustrate. I think some key features Okay, so we'll start right away with our first Question so this is not a this is you can vote for Whichever one of these you think are true, and they could all be true or all false So this is a question whether key features of what we call an auto auto automated insulin delivery device or a ID device The current FDA approved devices are hybrid closed-loop controllers, and I will give a caveat to all my international colleagues I am talking specifically focusing on the FDA approved ones Is that true a ID devices are driven by a CGM and a pump being connected through an algorithm? You think that's you vote for that? Technologically advances are decreasing health care disparities. You think that's true vote for that If you think total daily insulin is not a relevant metric for an AID device vote for that one And if you think basal rate changes are the most effective adjustments that can be made to an AID device to vote for that one Again, there's no one right answer. This is just showing so all these could be correct So, let's see thanks all for voting here It sounds like we have absolute agreement everyone understands an AID device is absolutely connection between a CGM and insulin pump And it's certainly true that the current FDA approved devices are in fact hybrid closed-loop controllers That's just another terminology. We'll go through what that means, but it's pretty straightforward It just means that it partly Automates the insulin and it partly requires that the user does something and in this case the user does something Simply means they should they need to give me a boluses. That's how their design and that's all that term that that hybrid means and In terms of we'll get into what total daily insulin does basal rate changes And both of those bottom two DNA are not true And TDI is relevant and basal rate changes don't do much in most devices So I want to just just know this is just one slide to acknowledge it I think should be acknowledged up front and every talk like this technology advancements with their amazing glycemic outcomes Sadly are increasing health care disparities and this isn't just a ID data This is lots of technology advancements as multiple studies coming out about this So this is a very important area that we have to address. It's not the topic of my conversation I just want to up front state that So just some terminology again lots of alphabet soup my preferred term is automated insulin delivery That can also be called closed-loop control the loop again is the loop between the pump and the CGM Which is which is looped in through an algorithm I do not like the term anymore the artificial pancreas for those of you that were Greg Florent Florenza's talk yesterday. He joked about we only use it for grant writing because it sounds very sexy and appealing But we know this is not an artificial pancreas. There's no this is not alpha and beta cells And then I want to just also just clearly state that all the current devices at the moment Do have what we call a manual mode or you might hear open loop mode meaning it can revert back to just being a regular Disconnected pump where it's working like a pump and doesn't have any automation So then the other thing to just clarify here is I like to think about categorize some of these devices based on the CGM I will say the CGM is the most important technology if you want your person can afford one thing it is the CGM But the devices when they're not driven by a CGM And they're just disconnected some people in the literature use a turd sensor augmented pump I've used that term myself in several studies. It's not my favorite term It's just when that sensor and pump are disconnected then some of the early devices were driven by the CGM But the only thing they could do is to lower the insulin and when you have a device like that We often use the word low glucose suspend or predictive low glucose suspend Common device names for that or basal IQ or the manual form of 670 G when it's used in suspend before low What we're talking about today those devices driven by the CGM that both increase and decrease insulin as I mentioned the Common term currently, but that would be changing is hybrid closed loop And these are a ID devices where the people are expected to bolus for a meal They have three components a pump a CGM and the most importantly the algorithm and You really it's the algorithm you're trying to understand a bit about These systems usually use Bluetooth and they the algorithm itself can either be embedded directly onto the pump It can be an app on a lockdown device or an app on a mobile system And finally we have the first version on a personal mobile system And then there's also all kinds of compute confusing viewing apps that might be helpful to view data or to share data I'm not so much focusing on those today, but those are our part of these devices So what's the CGM technology? Just as again, this is I apologize. This is FDA centric discussion We have the the Libre 2 and the Libre 3 Libre 3 was just approved Which is a even smaller device with really amazing what we call M a RD or accuracy measures less than 8% and That is currently approved. These are not device CGM's that are currently running any a ID devices at the moment I hope they will but they don't at the moment Guardian 3 is the one that you're familiar with with the 670 G Guardian 4 is under FDA approval right now And this is finally a game-changer in this suite of sensors because it will be the first one. That's not factory That will be factory calibrated. So you don't have to do finger-sticks, but It's not approved just yet Under review and then I'm sure you're all familiar with the g6 This is a common one used for a ID devices and the g7 is under FDA review as well That's where the transmitter and the CGM or one one device It still is a 10-day sensor and I think they're working on longer sensors And then there's the ever since implantable CGM sensors And I won't be talking about those because those at the moment don't connect with an ID device So what are the ID systems? I'm gonna discuss there Technically for but all lumped 670 G and 770 G together as if they're one sense system They're they're two different algorithms Let's just say they're similar algorithms, but they're different Softwares and then we have control IQ and Omnipod 5 I mean 5 5 is the one that was just approved in January and they just released the limited market release broadened it about maybe about a month ago or so and Then the ones that are around the corner that I will be mentioned because you really need to know about them and by this time Next year, I am sure though. I will be shocked if they're not approved and that is the 780 G Which is under FDA review right now And again, that's the one I was saying uses the Guardian for sensor which is a factory calibrated one and then an eyelet and that just That that group just presented their pivotal trial results, so I anticipate that Hopefully we'll be in our hands soon. And then we have the tidepool DIY loop program, which we I have Several people who do really great with that particular Configuration and that also has some data actually under FDA review Then there's some other devices outside the u.s. That eventually I hope come here, but aren't here at the moment So Let's take a minute to think about who do you think could benefit from an a ID and again? This isn't one correct answer. Just vote for all that you think might be true. So type one only Any voters for that you can't use in a type two Do you think only tech savvy individuals can use an a ID? Do you feel like you really need to have your People with diabetes be able to do carb counting to use these devices And how about those that have really low already are meeting all our targets are really low a1c is less than 7% and then of course Anyone living with diabetes that it can afford the technology and clearly Most of you will check that box. That is my favorite answer If they're willing to wear a device and they can't afford it I Think they could benefit In terms of type 2 diabetes I certainly have several People who have type 2 that are using an a ID device and I think the published data at least for some of the real-world Data suggests even up to 10% of some of the groups have type 2 I feel very strongly that you do not have to be tech savvy to use these devices I also feel quite strongly that you don't have to be a carb counter I've been able to manage it with some individuals by By just giving them kind of what would be a small medium a large meal But I will say these devices at the current time the current ones Do better if you do carb count that you absolutely will get more out of it But you don't have to and then there's plenty of data showing how even people who have amazing a1c is less than six and a half Percent will improve their time and range and decrease their hypoglycemia the hypoglycemia probably the most important part The people who haven't a1c is much higher than that stand to gain the most in terms of you know percentage change But I think all a1c is what could potentially benefit If you can afford it, so just not to go into data so much, but pretty much all of these devices In their pivotal trials will show that that individuals the average person can see can achieve the time and target range of 70% It's not a little unfair to put them all on one graph because they're different study Populations and they're different study designs, and I think that's just really the reason why there's slight differences between them but it's quite remarkable how consistent the results across are across very different platforms and Clearly a1c is improved And so so they consistently increase time and target range they lower a1c So that most people can meet their 7% target on average and it decreases hypoglycemia So let's transition a little bit more into what what are we going after what are the buttons we can push well the most important Element of course in all of these devices is the target of the target range that's set by the AID system So I want to start off with having you recognize What the targets are and what targets can you change or not change and some of the systems can be adjusted so what? What buttons can I adjust in the target? I went ahead and put on here the 780 G in the islet since these will be around the corner, but I Will just line them up so you have a sense the targets used during AID use for 670 G 770 is 120 the control IQ uses 112.5 to 160 and then the OP 5 It has selectable targets So you an individual person can change from 110 up to 150 at those increments you see there 780 G will have selectable targets when that's available Even down to as low as a hundred and then the island is really designed to be very user-friendly and it basically Allows the person to select where they want their usual target when they want to hire a lower target without defining precisely what that target is and I think another element that that I have taken advantage of particularly with OP 5 is Whether you can adjust it by time of day So if you wanted a different target either during a high exercise time period or an overnight period or what have you? You can adjust those in OP 5 and and I think the islet might have to two times a day you could adjust Each one of these also has additional targets So if someone has in a high-risk situation where there might be a higher risk for hypoglycemia That would include you know most obviously exercise Then you have what we call a temp target or some kind of activity change that they all have for the most part they tend To pull it up to about 150 And some make it easier than others to either set a duration or what have you so there's a couple little Nuances there, but just remember that there's a an activity setting. That's a common thing. I have to remind My patients that they they could Set and then not to make it super confusing, but one thing I get asked about is the difference between the algorithm target and The target for an actual user derived bolus and those can be separate things And they often are in the algorithm, and I think it often gets confused So 670 G. That's a fixed number. It's 150 whenever the person is requesting a bolus is calculating down to 150 Control IQ it's fixed at 110 and you can't change those that is just the the target That's there op5 is different Whatever selected you target if you targeted, let's say 110 that's also gonna be the same target used for the bolus calculator so many G has a fixed one at 120 and Eyelid isn't automated So which ones can you change if you see the more modern ones coming out? That's one feature. We're gonna see that the targets you connect for the a IDs part of it. You can start changing So What's another important element an element of all of these systems the algorithm has in common is that they're they're geared around total daily insulin It's a common driver of all the solutions, and it's generally tracked over time automatically in these devices typically They're averaging it over a six-day period and the the reason there's different startup procedures There's some there's different reasons for that But one key reason why there are is because it's trying to ascertain what a TDI is when it's when you're first putting on The device it needs to know what where are we starting with with this person? And so the the next series of slides? I've tried to Show you the same way with this the 670 G and orange to control IQ and green and Omnipot 5 and blue and just briefly these are these are running TDIs for 670 G over six days And it's that parameter that really derives What it's deciding in terms of what the autobasal rates going to be an insulin sensitivity factors these are factors You cannot adjust and they're driven off for the most part the TDI That's also true of Omnipot 5 so it uses the TDI that gets updated with each pod change And it's using that to develop what's calling an adaptive basal rate not one that you set, but it's it's Working in the background control IQ. I guess is a little bit different in that regard it does it does use a rolling 16 average as well, but it has a smaller role in modulating what some of the Boundaries are for the micro boluses. It doesn't have so much of a role in determining what the parameters themselves are So one other element That I do want to make sure That is put out there is that you know this idea about basal versus bolus split It's not really how you should think about an AID system I like to think about them as two types of insulin delivery one is What can the user decide the user directed bolus how much insulin can requested by a person? And then what is the algorithm doing the algorithm directed insulin delivery? and So let's look a little bit about how this relates to what? What you can adjust I do want to say this first line here is about system initialization and again both the 670 G and the upcoming 780 G require a little bit of information Where it needs to learn what the TDI is before it can actually go into auto mode So that's why it has a little ramp up for that Control IQ you just type in the very first TDI. I get a lot of questions about that That's just to start it it becomes an irrelevant number as time goes on Unless you had a unless they really rebooted the entire system, and that's not very common and Op5 only needs some basal rates it'll use just for the very very very first pod and after that it doesn't that it doesn't have a Bearing and the weight the eyelid is really Streamlined to make it super easy and the only thing you put into that system is the weight the weight of the person and it calculates everything from them So user defined parameters, I'm going to go eat through each one of these for the three main devices But I do just this is the only place I mentioned the islet you in there You don't do carb counting you just put whether the meal size is usual less or more And Then the last thing that I I just this is kind of a little more complicated, but I'll show it to you in the in the The examples as whether or not there's something called an automated correction bolus. I think frankly this is turning into kind of a marketing tool that I don't really understand all of these devices every five minutes will make an adjustment to insulin so they all do that and Control IQ has a particular Element extra element where it can give a more discreet bolus every hour and you can call that an automated correction bolus But it's adjustable because it's affected by the correction factor, and I'll explain that the 780 G and their data They'll show that they do have what they call kind of an automated correction bolus But it's just it's just the algorithm can be more aggressive at these five-minute intervals. So I think it's really a reflection of, what are the methods that an algorithm can use to be more aggressive? And the outcomes look quite similar across the devices. So it's just a difference in how they work. And I think, as a clinician, understanding how you can manipulate that. So with that, let's look specifically at carb ratios. Carb ratios matter in every single one of these three devices, because they are hybrid closed-loop controllers. It will not matter in the islet, because there's no carb ratio. But in the current ones, it will matter. And it will matter in the 780G. Correction factors are a little complicated. And correction factors for 670G, it has absolutely no effect on the automated insulin delivery. They are extremely relevant. Just remember, these devices fail, right? And someone can, especially the 670G, get kicked out of that quite a bit. And so when it gets into what we call manual mode, it's not operating as AID. Whatever you put in there as a basal rate is what the device would use. So the relevance of some of these starting parameters have more to do with, what if the device fails? When the device is working in AID mode, it doesn't do anything. In terms of OP5, just to jump to that one, similarly, the correction factor has no effect on automated insulin delivery. It is used in the user-directed calculations. So that's where it will, when they're going to do a correction themselves and request something, then you'll see it there. And again, it's always relevant in manual mode. Control IQ is, I think, probably more traditional in the sense of the use of correction factors. Certainly, it matters for all user-requested corrections. But it does matter for these discrete one-hour automated corrections. And you can see that's the actual formula right there. It's about 60% of what the calculation would be. So if you're trying to lower or increase those, and I'll give you an example of that, you can do that by changing the correction factor. It has a little bit of influence, without getting into the weeds, a little bit too much on the basal rate increases. But that's, I think, minor and a nuance. What about duration of insulin action? Well, it's a button that I think they created on the first generation 670G. And it affects insulin delivery. You can make it more aggressive than that. But it doesn't, it's not, in the 670G, it's not at all what I call a physiologic value. Usually, I'm trying to set that thing a little aggressive, more like two and a half or three hours. But it doesn't have anything to do with control IQ. And it has barely anything to do with OP5. It does affect a little decay when you have a user-directed bolus. But not during the AID use. So OP5, I said how I would typically set that at three hours is a common thing that I tend to choose. But it's not really physiologic. And it just has to do with the decay when someone does a correction, how long till it decays. So what about basal rates? Well, basal rates don't matter at all in AID for the 670G or OP5. And the OP5 is only, if you guys are just beginning to have experience with that, you should just know that the very first pod is an anomaly. That is a time where it's learning, really, what kind of TDI this person has. And it can get started immediately in automated mode. But it's a really restrained pod. And by the time you get at least 48 hours in the first time and you change the pod, then it's kind of up and running as usual. And after that, the basal rates don't do anything in OP5. You can change them, and they will not affect the AID. Control IQ, it does have an influence. You can use it. It does have to do with where the system tends to settle. However, I'll tell you, once I kind of get what I think the basal rate should be, I don't change them a lot in Control IQ. Now, I'm an adult provider. I'm not usually doing kids or adolescents. So there are other scenarios where their needs can change significantly. But that's because the basal rate can really change between zero up to roughly about four times that preset basal rate. And my goal isn't to try and get the AID not to do work. I don't think that's really what I'm trying to achieve. But I am trying to get a reasonable amount of basal rate modulation. But changing the basal rate in Control IQ will have an influence. So with that, let's think about three different cases here. And this first one is 670G. And this is a patient of mine who's a 55-year-old gentleman with T1D who has no complications, had T1D for long decades, and has an A1C of 6.7%. And he does phenomenally well on 670G. He does estimate his carbs pretty carefully. And he administers meal bolus on time. Although he's clearly meeting glycemic targets, he's asking me, really wants to know, look, I'm not happy about my prospangyl hyperglycemia after dinner. We need to do something about that. So that's the question he's asking here. And you all, I'm sure, are familiar with these printouts. You can see here, he's doing amazingly well because his auto mode time is around 98%. So my people in my clinic who do really well on 670G have figured out a way. They have no problem doing all those finger sticks. And they just are on it. And they stay connected. And every single AID device, if you stay connected to it, you will do so much better if you do not. And the more modern AID devices, the connectivity rates are generally around 92% to 95% or 96% in contrast to the original studies with 670G, which was just, frankly, the first one out of the gate. So he has time range 80%, time less than 70% is just 2%. And you can see it's an amazing coefficient of variation at 30% with that really narrow, narrow, if you look at the blue, focus on the blue curve overnight. But he's not happy about this little bit going on here. Let's see if I can, in his nighttime, let's see. Oh, there we go. That part there, that doesn't make him happy. And so he's asking, what can I do about that? So this does have what I think is one correct answer. So select one of these. If a parameter change is needed, and we can argue whether any parameter change is needed at all in this individual, which of the following changes do you think in this device is the most direct way to address this preshpandial hyperglycemia at night? Do you want to decrease insulin action time? You want to decrease the carb ratio? You want to decrease the correction factor to allow more automated insulin delivery? Increase the basal rate as one method, or maybe lower the target? And everyone is not being swayed by some of the lower ones. So absolutely, decreasing the carbohydrate ratio is the one intervention that you can do with this device that's going to address it. So very specifically, and in fact, I would say it's the only intervention you could do with this device that's going to directly adjust insulin for a meal. Sure, as I just mentioned, you do have this other button you can use, which is decrease insulin action time. And that is absolutely not my preferred parameter in this situation. And it's for several reasons. One is it's not that powerful. And all it's really doing is looking at what the previous corrections were and allowing more corrections later. It is allowing it to be more aggressive. So that's true. But in his scenario, he is in amazing control pretty much all day till we get closer to lunchtime. And you cannot really change this by time segment. So even if you thought it was something that would be helpful, it's going to affect the whole entire day. Basal rate and correction factors can't be adjusted in automated mode. These are system-determined settings based on the TDI. And then the target of the algorithm for the postprandial dinner overnight period cannot be lowered below the fixed targets. In this case, it's 120. And again, the 780D, when it comes out, that will be a game changer because it has a much better sensor. And that's got targets that are 100, 110 that are lower that can be set. And then remember, this one has a temp target of 150 that can be used. I don't have, I think a lot of people just forget that and don't use it very often. And in this scenario I presented to you, it's not relevant. But I just want to mention that again. So carb ratio is the main thing you can do with this device. That's interesting. OK, there we go. So this is someone on control IQ. So this is our second scenario. And this is one of my college students. He's 21 with T1D. Has no complications. Used to have an A1C that was way up there. And the 8% didn't want to end higher and didn't really want to do boluses. Once we got him on a AID, then his A1C is finally around 6.9%. So what I selected here is just one single day to just illustrate a couple of things. This was a day when he was just, he wasn't doing a full on move. But he was moving a couple items between apartments for the new semester. And he was frustrated because he had to do a lot of hypo treatments during the day. And he felt like he was eating throughout the day. He does definitely miss meal boluses. And including on the days shown, there's two little boluses in there. And he's not using exercise mode. And he's asking for my advice on what can I do in a day like this to avoid hypoglycemia. So let's just look at this a little bit more closely just to walk you through. With this device, you see the setting here. This is his sleep time. And this device is attuned. I didn't really actually in the target slides point that out as much. But what it's trying to do during sleep is by the end of sleep, get you to 112.5 to 120. So it's a very narrow range. It's trying to have everyone wake up at that every morning. It's much more aggressive in the second half of the night. And you want to make sure when you set that sleep setting that you have a minimum of five hours. I'm always going for eight or nine hours. And then I have a couple, more than a couple, that want to do it all day long. So that's another method that people can use. But the sleep setting here. And he wakes up right in the target range. That's around 110 or so. And what you can see here is this brown bar, I guess. That's an override that he is doing. And the rest of these kind of crosshatched ones here that you see here, these are all those automated corrections. So this is the system doing it, giving 60% of a calculated correction factor based on his predicted glucose and doing it periodically. And one thing that I'll just point out that I don't love about AID systems, in particular, you can see it illustrated quite well here. You see how he's hypoglycemic and he's recovering and is shooting right up because he clearly is taking glucose. And he gets a chunky automated insulin delivery in the rebound of the hypo. And that likely contributed to the second hypo here. Now granted, he's moving and I don't know how many grams of carbs he took and what exactly is happening here. But I don't love that. In this case, it fires again after he's increasing. And that might be the reason why it limited it. The idea is trying to limit a little bit of that rebound. But you will get a response for hypoglycemia. So the question here is, and again, these are one best answer, I think. Which of the following would be a reasonable initial adjustment? Obviously, there's several things you could do. But what would be the first thing you might try? Would you decrease the basal rate? This is the setting of recurrent hypoglycemia. Would you increase insulin action time? Would you increase the carb ratio to give less with meals? Would you increase the correction factor? And that will decrease some of the automated insulin delivery. Or increase the target of the algorithm during activity? I'm getting a little bit of a spread here, which is perfect. And it's always, you know, it's the question of the initial adjustment. There's a couple things you could do. But I agree with the majority here. I would increase the target of the algorithm during the activity. And very specifically, it's gonna, when you engage exercise in this activity, it's gonna bump up that algorithm target up to 140 to 160. Instead of down around 112 and a half to 160. It also has, when it does that, it also will decrease the, it has a higher threshold for when it's gonna lower insulin. So instead of targeting kind of a predictive value of 70, it's gonna target a predictive value of 80. So it'll have a little quicker action. And it also fires that control IQ low alert a little bit earlier. So there's a couple settings that are helpful. I find it helpful when people are doing more sustained light activity. This gentleman is a runner. When he runs, it is not powerful enough to deal with, in his case, quite sustained aerobic heavy activity for him. But I think it works pretty well. It has to be individualized, but it works pretty well for that. I usually start it at least an hour before the activity, duration of the activity, and about an hour after. I start with that, and then we kind of adjust it from there. So you could increase the correction factor. I think that was the next best option, and that will definitely lower some of those periodic correction boluses that you saw. And I think he had five or six that day. So that would have been an option. Decreasing the basal rate's certainly an option because in this system it can do that, but it has less of an intended effect than you think. I think it is still, when it thinks it needs more, it's gonna increase. When it thinks it needs less, it's gonna decrease. So the impact is not as great as usually I would like. Increasing the carb ratio, you can do that with this device, but it's not very effective because he's not at all using, he's not doing, his issue that day wasn't doing meal boluses. So and that's the only reason why that wasn't a key thing. And then duration of insulin action is not adjustable for this system. You'll see it fixed. When you see it pop up, a lot of people have asked me this, it'll say correction factor's fixed at 110 because that's what it's correcting. I'm sorry, the target is fixed at 110 because that's what it's correcting down to when you're doing a user-directed bolus, and then it says five hours on the duration of insulin activity. And it's not, these things are not, I mean the insulin action curves, a lot of them are proprietary so they don't always publish them. But I think of them as pretty similar and most of that insulin action is occurring much early on. I mean there's just a tail that goes out. So let's see. Okay, so exercise is a glycemic challenge. I absolutely think these devices have not solved this issue. And one thing that I think you should know about is a lot of times people will try carb loading, right? That's a very common, I wanna start my exercise at a higher load. That tends to backfire an AID system. So if you carb load too much, what is going to happen is that system will absolutely respond. And then there you are about to exercise and you've got an insulin bore that is just too high. And it just accelerates an inevitable hypoglycemia. Now that's not to say that do I have people that will do smaller amounts of carbs and they're not trying to get themselves up to 200. You know, they're just trying to make sure, especially with an insulin target if you use in this example, I'm using control IQ that goes up to 160. If they're sitting there at 150, it's not gonna be, it's not gonna have that same response. But just know that that's not always the easiest thing to do. And you can mitigate some of this by adjusting the targets. But you know, I certainly have plenty of people that just have to disconnect from the system as their method of dealing with it. So we do a lot of research in this area because it is not solved. So remind all of these three systems do have temp targets that generally hover around 150 or so. Okay, so how about OP5? This is a 45 year old woman with T1D who doesn't have any complication. A1C is 7.6%. She's currently just using an Omnipod system without the AID part of it, just on an insulin pump. They're not connected. And she wants to transition to an AID device. And she is someone who does meal bolusing. But she ends up waking up high at night because, or when she wakes up in the morning because she's always been so fearful about a low overnight. And so as a result, she has these basal rates that are really too low overnight. And her main glycemic issue is that she tends to wake up high. And so you're getting ready to transition her. So the question is, you know, how do you think about setting those new settings with this new device? This is not her actual glucose. I just tried to grab a, I only had really the, because I participated in the OP5 trials, I've been using the device the last two years plus. And this is just a glucose trace from one of my regular patients who is on it. So they don't have the best display when you're trying to look at what the system is using. But this is what you will see on gluco. And it has, this is similar to what's on the actual device for those of you who have worked with it so far. The purple indicates it's an AID mode. And these red bars are when it's really maximizing, it's dropping to zero, basically it's maximizing its low hypo protection basically, or decreasing the basal rate. And then sometimes there'll be a yellow bar where it's really maxing out the increase in insulin delivery. So the question about this device is which of these parameters that could be adjusted to improve a glycemic outcome? which is the one that you'll kind of prioritize when you're initiating the OP5 device. So you could, again, this is just setting up the scenario where she's got, she wakes up too high in the morning and has a high glucose overnight and after dinner. So you could decrease the carb ratio at dinner. You could decrease the correction factor that might increase, and the question is whether it increases automated insulin delivery. Decrease the duration of insulin action. Increase the basal rate, because that's the most direct way to affect automated insulin delivery by this device. Or you could lower the target in the overnight period to try and get better glucose control when they wake up. Okay, excellent. So this is kind of a, it's not a trick question exactly, but I do think a lot of you selected the basal rate. I would say it's the target. So what you really should know about this device is that you want to be setting an aggressive target. And the published data that we've published about it will show that those that do the lower one, about 75% in the pivotal trials. Now my pivotal trial participants are not always who exactly I'm working with in clinic. They're often much better controlled, but they were on the 110. And the best glycemic control was on those lower ones. So most of my adult patients, I'm 110 to 120. Some of the teenagers we're working with, and certainly the young, young, young kids, because we go down to two years old in these trials, tend to be a little bit higher target. But lowering the target is the most effective way you can get the thing to be more aggressive. Increasing the basal rate, which is what I think most people selected, will affect automated insulin delivery, but absolutely only for that very, very first pod. So it's the first 48 to 72 hours of use, and then after that, it's irrelevant. So you can't, adjusting that is only gonna be helpful if they get kicked out of closed loop, which is not common with this system, and they have to be in manual mode. So decreasing the correction factor, you can. That is an adjustable factor for this system, but it's not really preferred because, and I wrote it specifically that way, because it doesn't affect automated insulin delivery at all. It won't affect anything overnight. It only affects the user-directed boluses. You can decrease the carb ratio at dinner, but I don't think, I would think of that as a secondary effect once I set the target. And sure, you can decrease duration of insulin and action time, but I don't usually change that much. I think it's so subtle in this system. I don't usually, I'm not looking to adjust that one. And one thing I didn't mention about this system in particular is that the bolus calculator actually uses CGM trend, which is one of the only ones that does that. So if CGM is trending up or CGM is trending down, it's gonna add a little insulin if it's trending up and take away a little insulin. You can see the actual calculations if someone wants to see the detail. They're available with a couple of button pushes to see it. But I would say in general, it has a bit of a conservative bolus calculator. I'm often trying to set a little bit more aggressive carb ratios to get the coverage I need, especially in the breakfast period. So to summarize, 678G and 778G, it's really the carb ratio is a key thing you can adjust. You can adjust duration of insulin action. I most common set mine around two and a half or three hours, but individualize it a little bit. It's not physiologic, don't think of it that way. It's just a setting to make it more aggressive. Control IQ, carb ratios for sure, correction factor, just remember that's the one that does affect those automated corrections. So sometimes I find that very helpful to adjust that, especially a lot of times sometimes people have their correction factor set super high overnight and that's not, I'm usually lowering that to a more usual thing since it has something to do with the automated insulin delivery. When they're in sleep mode with this device, those automated corrections do not happen. So for those people who are choosing to do sleep mode for 20 hours out of the day, they don't get the automated corrections. And that may be just fine, for some it may be perfect, but other people it doesn't always work so well. You can set basal rates in that and set the sleep schedule. OP5, carb ratio, and I probably should've, I didn't really, setting of the AID targets, the key thing there, these aren't really done in order of priority, but just to remember, carb ratio is important. And then correction factor is only marginally important because of the user requested corrections, but not the AID part of it. So carb ratios are relevant for these hybrid closed loop systems. Basal rates adjustment's not critical. Correction factor is relevant for some, but not all. Remember the IOB upon awakening for use of an AID device, a lot of times is a lot lower than a standard pump therapy, therefore they tend to have an exaggerated response to breakfast, so I'm usually very aggressive across all of these devices about setting the carb ratio at breakfast to really be high. And breakfast, we know in general, has to be covered for a lot of people, but I'm aggressive about the breakfast. And then IOB is often lower prior to a hypotreatment, so sometimes they get an exaggerated rise and they tend to get away with a lot less glucose required when they're actively using one of these devices. Limitations, cost and access, absolutely postprandial hyperglycemia, we're still stuck with very slow subcutaneous insulin action which is frustrating, exercise has not been solved, and then some of them, not great about the hypoglycemic response at times. So with that, thank you, and we have time to take some questions. Hi, Tina Cater from Montreal, and I also happen to work in Plattsburgh, so I have used, I see the two different views of the pumps, first of all, thank you. This was fantastic, you did a fantastic summary of the difference of all the pumps. The only negative is, I think my colleagues that are just getting into pumps, if they see all these nuances, they're gonna go, no, I'll leave it to Cater, I'm not touching anyone on the pump. Because each one, it's sad, it would be nice if they all had the same sort of background, because each one has a special thing. I'll just give you two caveats of what I've learned with this. One is Medtronic, as you said, first out of the gate was fantastic, but the only thing I've learned, you really have to look at the insulin use. When in auto mode, I had a guy that lost 50 pounds, didn't come to clinic, and every time he got kicked out, he was going down to sugars of 30, because you had to adjust the basal rates for the auto mode, and this pump, as you know, they're not like your fantastic patient, many patients are going in and out, and I know it's gonna change. And the second thing I do, I'm a big athlete myself, so all the athletes come to me, and I do a terrible job on the pump, it's very difficult. So I was excited with the tandem, and what I've learned, I'd like to hear your experiences. From my athletes that do the marathon, I've chosen a different basal rate, different correction factors, I have an exercise profile that they switch to an hour, even two hours before they start, so they won't get those big correction boluses, and even though, as you said, the basal rate isn't used, but it's a little used, and I find that's been helpful. And then just two, if you can comment on, sorry, I could be here all day with you, gastroparesis, or patients that are going into these new pumps with retinopathy, what do we do, do we aim for higher targets, worrying about that first year progression, and the last one is pregnancy, and I've taken too much time, thank you very much. Those are all outstanding points. So I do want to point out, and I didn't, I'm glad you brought up about the different profiles, I absolutely use that, lower exercise profiles and menses, I mean, that's the other two things that I sometimes will use, and if you're using Kontrol-IQ, you can clearly set those basal rates low, and if you set them inordinately low, knowing you've got the marathon runner, then that is definitely gonna help. So I appreciate the clarification on profiles, and I think, I have some concerns, it's certainly true that these devices, within, this has been shown over and over, within one week, they are right at their targets, and usually we can do it within one or two days. So to your point about retinopathy, where we've been concerned about where people have sudden, you know, all sudden they're in range, I do think there's a little bit of, I haven't seen data that people have shown that to be a problem, but remember, most of these studies, they're already, you know, a special population that's being studied. I do think what I'm trying to do is get individuals to be able to trust the system and to work with it, so I absolutely will be a little more conservative, you have time to get them where they need to be, just using the OP5 as an example, you can easily start at 150, and just kind of work your way toward where you wanna be. So I think there's some flexibility with the newer systems with those targets. Pregnancy, sadly, it's just ridiculous that these CTMs are not approved in the use of pregnancy. I will individualize it, because I have some people that do amazingly well on their own during pregnancy, and the device I actually think might, you know, sometimes in some cases, make a worse scenario, but I don't think that's the average person. The average person struggles, and so the average person, I think, actually does okay with that, but clearly, off-label, an individualized thing. There's an amazing algorithm out of Cambridge that I hope comes to the US someday that's the one that's really been studied most in pregnancy, and I'll be scooping that one up for pregnancy. Gastroparesis, that's just a tough thing, right? I mean, it compounds a lot of what we're seeing here, and that's just where, in some ways, you can, if you don't have too aggressive of a CARB ratio, you let the system that's more gentle deal with what happens. So in some ways, that's a really nice population that you can increase the flexibility using these devices. I'm not sure I got them all, but maybe the next question. Hi, Jessica Schill, Henry Ford, Detroit. Question for you. I was intrigued by what you talked about regarding those who do not CARB count and making suggestions regarding small, medium, large meals. How would that translate to the non-islet pumps, the ones that are currently available for AID? Yep, thank you, because that's another point I wanna expand about that. You guys are awesome with the questions. One thing that's really important with this device is that I tell all my patients, you have to use the bolus calculator. And the reason for that is you have to get them informed with the IOB. And you run a risk, if they're just overriding it, or they're just doing five units, that they're gonna avoid the IOB. You take the Control IQ one, it could have just given them a three-unit bolus, you know, depending on their settings, with the automated insulin correction part of it, and they wouldn't have realized that. So I really have them use the bolus calculator, and that will protect with the IOB. And then I'll choose like, you know, 30, 45, 60. I might start something like that, that that's a small. Just know these three numbers, or even, I even have some, I have two numbers, because we just, you know, and we're just trying to get them started, and get them in a much better place if they're willing to at least do the boluses. Thank you. Thank you. Hello, my name is Jennifer Jessel. I'm a nurse practitioner out of Newton, Georgia, locally. Thank you for coming, everyone. My question is with the Omnipod 5, and the lack of ability to change basal rates after that first pod. That definitely piqued my interest. And so what, so what about the, I'm wondering about the adaptability with illnesses, menses, you know, changes in their physiology. How quickly can the Omnipod 5 adapt to these changes in metabolic stress? I have sometimes, so, let's see how I answer that. It can, like if someone suddenly gets put on steroids, right, and all of a sudden their TDI is a lot higher, I have to get them to do a lot more corrections, and user-derived corrections, and being aggressive with the carb ratio. I mean, that's the most direct way that I know to make sure their TDI is informed about that. And sometimes I can do, you know, maybe more aggressive pod changes, because the changing of the pod resets the TDI, but it's a rolling average still. It's not, so, you know, we're trying to use these devices in studies in the hospital, and that's part of the trick, is trying to figure out how to inform it when someone really changes their TDI. So, there's still hybrid closed-loop controllers, and therefore, you know, changing the, getting them to put in their carbs more aggressively, and do more corrections if that's allowable in the system is one method. And so, they're not great at it. So, in real life, so, a menstruating female, would they be a little more aggressive with their, would we change their carb count during their menses? I mean, I normally change basal rate. You know, I usually give them a different profile. Yeah, no, you're gonna have to get them to probably do, change their basal rate or their user-derived correction factor. That will be the challenge. And unfortunately, you don't have, for that one, you don't have a lot of profiles. So, it's not like the Tandem, you could put a, the controller, like you could put a bunch of different profiles in there. And you can't, you know, unlike the Medtronic, sometimes I've had where I've had to kind of stop that system and then restart it, just to kind of reset things in pretty extreme scenarios. You can't do that with the OP5. You actually have to get a brand new app, I guess, in order to reset it. Thank you very much. Hi, I'm Amanda Sheehan. I'm a nurse practitioner at Joslyn in Boston. And I had two questions. One was, the gastroparesis brought up a thought. I have three patients, one with a gastrectomy due to a prior history of gastric cancer, and two with a Roux-en-Y. And all three of them, we've been working to try to get them on a pump. And just with your knowledge of the various systems, like the new Omnipod, is there any algorithm or preferential choice you might have for someone with a glycemic profile such as this? Because with the Type 1 and the gastric surgery, the remodeling, it's just the absorption is so different in their profiles. Yeah, I think that's the beauty of the AID. They all really work in a similar way. Once you understand a little bit of buttons, you can push to adjust them. But they really are, every five minutes, changing their ratio. And they're still hybrid-closive controllers. So I still look at which device feels better on your body. You want tubes, you don't want tubes. Do you want, you know, the kind of the living with a device that's attached to or on your body, I think is more of the question, are they a swimmer, not a swimmer, that kind of thing. I think they all, if you can keep it in automated mode, right, if you successfully do that, then I think they all could handle that reasonably okay. I don't know of studies that have tried to differentiate that level. And then my second point actually kind of gets to, it's more of a, I guess, comment. But I've started some, like when the T-Slim came out, I started a few of my patients on it who were more arthritic or had some manual dexterity issues or were a little bit more in advanced age. And two or three of them just could not continue because of the difficulty in just readying the pump for use. And so one of them, I was able to get the Omnipod approved after the insurance did cover it. I looked at a video on YouTube, like of the start of the T-Slim versus the start of the Omnipod. One was like 15 minutes and the Omnipod was like two minutes. So it's just such a huge, I never would have thought of that before. It was just so. Yeah, that's a great point because especially some of the arthritic, I mean, the infusion sets itself, they're challenging besides the cartridge for some people. And so I don't think you can beat that Omnipod, stick in the needle, peel the thing and slap it on. Yeah, she's so much happier. And I will say some nice advancements, some of the infusion sets are just getting better and longer. There's one that's gonna be maybe a week that I'm super excited about. So I think we're looking for advancements in infusion sets as well. Thank you. Sorry, just one more question, even though I can have 10. Just type two. So you mentioned, because we've been reluctant to put our type twos on the automated delivery. Is there a maximum total insulin dose that you would say no? Because they all have this maximum delivery, number one. And then number two, what do you do when you add on GLP-1 agonist and SGLT2? Do you change the whole profile? Yeah, so I don't have a lot of experience with type two, because frankly, I think it's an expensive solution. And so, you know, sometimes that's part of the limiting thing there. But, and then you've got a TDI, right? So the cartridges for the Omnipod is, or not the cartridges, the pods are 200 units. The Control-AQ and 670G are 300. So there's a practical reason in terms of how often they're gonna have to be changing all those devices. But for those who want to use it, can afford it and have some issues, you absolutely have to adjust. There's no question about it. And to lower the rates, I have been using GLP-1 or subdiagonists a bit more because of the, especially even in T1D for me, because of weight issues. And usually that type of individual is someone whose TDI is already relatively high. And, you know, we might otherwise have them on Metformin or something like that. So I have a tendency to use it in that scenario, and they're not always very insulin sensitive. So back to how that relates to type two, you know, you can get away with, there's not as much variability, and therefore you can, I find that I can be a little bit more aggressive, like most things, when it's uncertain, I'll start out not with conservative targets for whatever device I'm using, and then work with it. But you definitely have to do adjustments based on the ancillary drugs or lifestyle issues that are going on. So thanks, I know we're past time, but I'm happy to hang around. My flight isn't for a bit, so thanks. Thank you.

closed-loop insulin delivery devices

practical management

advancements

new devices

device settings

parameters

target range

algorithm

postprandial hyperglycemia