Video Dual CGM and Ketone Monitoring Across Diabetes Care Spectrum Play Pause Volume Quality 1080P 720P 576P Fullscreen Captions Transcript Chapters Slides Dual CGM and Ketone Monitoring Across Diabetes Care Spectrum Overview Continue to Test Back to Symposium Hello, I'm Rich Burgensta from the International Diabetes Center in Minneapolis, Minnesota, and today I'd like to discuss the profound impact of dual glucose ketone monitoring and maybe more appropriately titled today the potential profound impact because the dual glucose ketone system is currently in development. These are my disclosures. Back in 2018, I wrote that I thought continuous glucose monitoring was really transforming the management of diabetes step by step. I wonder if in 2028 I'll be able to write this statement that dual glucose ketone continuous monitoring is now transforming diabetes ketoacidosis prevention step by step. Let's look at those steps uh for the next few minutes, and I want to start with Looking at continuous glucose monitoring and how it came to be really the standard of care and see if we think that continuous ketone monitoring is going to do the same. We go back to DCCT. We all know the results from 83 to 93, glucose control matters. It's critically important to keep the glucose controlled. But along with preventing complications, we did get an increase. And hypoglycemia. So much so that for the next 20 years after DCCT ended, we knew glucose control was important. But the primary barrier to really getting down to 7, the reason these 1441 patients in the DCCT ended up about at 8, is because hypoglycemia was a limiting factor. And the more hypo you had, the more hypo you were prone to get. Now, if we expand out to after 20 years, we start to see a a little bit of a change in that A1C curve, and here's what I think happened hypoglycemia. Demanded or begets the need for CGM. We really needed a system. That could deal with the hypoglycemia and finger sticks and A1C didn't do it, so CGM came about. Yes, it was approved the years before, but it wasn't until 20 years after the DCCT in 2013 that we had a standardized report. We agreed on metrics. We started talking about time and range and time below range as our key metrics, and the dialogue really shifted from just A1C and finger sticks to this important data. The AGP continued to evolve, talking about evolution today. It went from this to this, and the metrics got codified and agreed upon. Then we added from the metrics we added the targets and the AGP visualization cleaned up a bit and was clearer to recognize lows and highs at a glance. And equally important to a good evolving visualization was the fact that we now can start to use the data. Patients were taught how to use CGM on their phones. clinicians and hopefully clinicians with patients were taught. How to look at an AGP report and make a wise and judgment call on adjusting lifestyle and medications. Here's an example. Patients look at their meals, they see the spikes, they adjust their meals. The spikes flatten out. This happens whether you have type 1 diabetes, type 2 diabetes, insulin, or no insulin. Clinicians look at an AGP and at first they're overwhelmed. Well, that's still a lot of data, even if you got it down to one page. But we teach them things like determine where to act. Look at the time and range and time below range, or is one of them out of target? If so, you need to go to the next panel and determine where you need to make an adjustment. Start with the red hypoglycemia first. Try to flatten and narrow the curve by making medication adjustments and lifestyle adjustments. And then number 3. Act on the data in a repeated fashion. Call people back for another review. So all of this was the evolution of CGM. And look what happened. You see that curve starting to bend, um, and finally the, the, the, the A1C is getting better and the glucose is having less hypoglycemia. But we still weren't quite at that goal. We still weren't hitting that 7%. We were still at just under 8, maybe 7.5. And then along comes CGM saying we need more, and we linked up with AID, and the CGM became part of that routine. Finally we could start to get not only avoid hypo, but we could Get the time and range uh improved and start moving that A1C. So look at the curve start to bend there. And, and AID as you know, took off for type one diabetes and now we're seeing it in type 2 diabetes again, driven by the CGM. So Hypo begets CGM. CGM begets AID. We get an improvement in time and range and time below range. Is there anything else we need to be dealing with? Well, yes. How about now a focus on another complication that doesn't get as much attention. A diabetic ketoacidosis, it's there. It's in the background. We don't discuss it as much unless you happen to be that patient with this dangerous condition. So let's look at that for a minute and see if it follows this same pathway. What was the effect of CGM on DKA? What was the effect of AI AID on DKA? First, we need to step back. DKA is a common emergency in people with diabetes. Type 1 and type 2, it turns out. It's a leading cause of death for children and adults under 58 with diabetes. So it's still a major problem, counts for a lot of hospitalizations, and if you get hospitalized with DKA with type 2 diabetes, you're even in worse shape. Than if you have type 1 diabetes, so it is a problem even though it doesn't seem to get as much attention. Did CGM solve the problem? When we got CGM, did DKA go away because we could control blood sugars better? Well, it made a big difference. We're really happy that it did. It cut. The rate of DKA in half for both type 1 and type 2 diabetes, the blue being before CGM and the yellow being after CGM, so you see the rates get cut in half both in type 1 and type 2 diabetes when CGM was introduced, but it's still the largest counter of, uh, annual events for acute hospitalizations. It's more than hypoglycemia, more than comas, more than hyperglycemia. DKA is still a major problem, even with CGM on the scene. So then it moves us to, well, OK, it's a problem, but who should I be looking at? Who's at risk for DKA? And here's a list that I put together just from clinical experience and looking at the literature Type 1, history of DKA, acute illnesses, use of a pump that goes back and forth a little bit. In the early days of pumps, a lot of DKA, and we got better with infusion sets, but it's still an issue. If somebody's had frequent high glucosis for a variety of reasons, they're at risk. Ketogenic diets, women, pregnancy, SGLT2 inhibitors is a whole separate category I think you all know about, but it's worth putting in red. So that was my list. To begin to think about, well, then along comes the DCCT study group, and they actually looked back at that curve that I showed you of those 1441 patients and said, you know, we've been following these people just as I showed for 30 to almost 40 years now. They had hundreds of DKA episodes. What was their inciting event or what was their correlation with the DKA? And it turns out that they verified the type 1 pumps, people who omitted their insulin or underserved populations, women had more episodes. So that was important to verify it in this population. They also added two new ones that hadn't crossed my mind or I hadn't seen in the literature. If you've had severe cardiovascular disease or if you had foot ulcers leading up to or on the edge of needing an amputation, those people had higher rates of DKA uh in this DCCT edict study group. And then they showed one group that did not have more DKA, which was a surprise to me, and I think that community, those with chronic kidney disease, that's important because those people would benefit from an SGLT2 inhibitor, but have been, we've been nervous to do it because of the risk of DKA, which I'll come back to. So lots of learnings of identifying who's at risk. Second, it's, well, OK, you're at risk, but what do I do with these levels of ketones that I'm getting one way or another, finger sticks are continuous. What level is important? And so there was a study done looking at those original people who were on SGLT2 inhibitors who had type 1 diabetes when they were trying to get the drug approved for use in type 1 and those 1,194 patients. Um, turns out when they checked their finger stick ketone levels and if they were over 0.8. On average for a month, the next month they had 3 times the rate of uh uh DKA risk. So we're starting to get some numbers that we can look at as a risk factor for going into ketoacidosis. We've been stuck with uh urine testing for a long time, not very popular, not very accurate, but something, and then along came ketone meters where you poke your finger and You get a number and you can kind of line it up to these, these numbers along the ketone strip, the urine strip, and people had to start saying, well, what are we going to call normal? It's around 0.6. What are we going to call elevated? And that was a little bit of a debate. Was it 1 or was it 1.5? And then it became high and then when everybody kind of agreed over 3 millimoles per liter was urgent high or needed acute attention. So we're starting to put the story together, who's at risk, what are the numbers, um. When do they indicate a risk? But when a survey was done just a couple of years ago of several endocrine clinics and they asked patients. It was kind of humbling. Patients were not familiar with the term DKA even. If you explain it a little more, maybe they understood what you were talking about. They didn't really weren't able to name right off the bat what are the symptoms. And if you ask them if they tested their ketones, you know, almost 40% never really ever tested their ketones, despite most people being told, well, you should have it handy in case. Well, what are the levels really mean? One way to look at these levels and determine is it OK just to follow your glucose? Will it tell you what the ketones are going to be? So why do ketones? Well, here's some really interesting studies, uh, led often by Jen Sheer at Yale, who's done some amazing studies where you have people on an AID. pump system and you stop the Insulin infusion and you watch them carefully in the clinic and you see their glucose go up when you stop infusing insulin and you see the ketones go up. So here's a whole bunch of patients glucose on the top, ketone levels on the bottom. But you notice they start at different levels, they go up at different rates, um. And you begin to wonder, well, what is this message telling me? I see them both going up and so here's, here's 4 or 5 examples I pulled out that I think is an instructive lesson from that really beautiful work. Ketones go up, but glucoses are pretty steady. Glucoses go up, ketone steady. Both of them, uh, uh, glucoses are steady, ketones go up. So there's all different patterns. You can't rely on glucoses to be your guide for ketones. So my takeaway from this really elegant work is if you want to know if you're at risk for ketones, you've got to measure them, and you've got to measure them continually to really see the pattern. Don't trust your glucoses to be your guide to ketones. So back to the figure of these decades. CGM was introduced and it cut the rate of DKA in half. Great, we still have a high risk. AID, a little plus minus. It keeps the glucoses down so you don't get the hyperglycemia that could lead to DKA, but infusion sets sometimes come undone unknowingly and that puts you at risk. So that would be a good population still in my mind to benefit. What else do we need to do to really move this field of continuous ketone monitoring forward? Well, the ketone monitoring system has to be easy, accurate, and actionable. Those are my three terms. It's going to be easy if it looks like this, and this is what's in development. That's all I can say. This is what's being tested is one sensor that has both a filament for glucose. The glu glucose and ketones are on one measurement filament. One system measures them both. It looks accurate, but that'll be in the testing for the company to submit to the FDA and get approval. But here's just one example where someone took a ketone solution and the ketone levels went up and the lab on the continuous sensor in the lab. Reference results were right on top of it, so it looks encouraging, but we have to wait for the final approval of a big data set. And then is it actionable? Well, a lot of the AID companies are signing up to say if you create this dual sensor, we're on board to put it on our AID patients because we think that many of them would say they could benefit from that. But we need to know the terminology, and that's important, and we're starting to get there. Remember I showed you before, this is a consensus opinion, an expert panel opinion saying maybe 0.6 is a good one to call our normal range, elevated above that. When you get to 15, you really need an alert. When you get to 3, you need an alarm. Others say, well, maybe it shouldn't be 15, maybe 1.0 is where we should have that alert. So I'm hoping there'll be an option for different patients. The one maybe is the alert or maybe 1.5, but everybody gets an alarm at 30 because that's when you're at risk for ketoacidosis by everybody's opinion and past history. So We also need an action plan. Patients, when they're seeing these day to day, need to know. When to look at their values, how to respond to the alerts, to take some fluids, to keep up with your insulin, to be checking your infusion sites. So there's information for patients. They should have a wallet card if they're on an SGLT2 inhibitor because their glucose may be perfectly normal and their ketones are out of whack, and that's going to be hard for, you know, an emergency room physician or team to know what's going on, but this wallet card would explain, I'm on this agent. And then clinicians need to know, well, what if I get a call from a patient. We can modify the stop DKA protocol that gave people advice of how to deal with high blood sugars and ketones, but now we can revise it with continuous ketone data and put in there such things as stop your SGLT2 if your ketones get above a certain level. When your ketones are at a certain level, give even more insulin than you thought from before. So I'm optimistic we're going to have good training materials. For an action plan. And then I personally think we need to update the AGP report to have a glucose panel, to have a ketone metrics panel. Why not have a ketone time and ketone range bar and tell people how their clinicians, how they've been doing for the past 90 days. And what their average and highest levels were. And then finally I want to close. With a case we were asked to pull a case out and I went, I pondered and pondered, what am I going to do about a case of continuous glucose monitoring and it turns out just last month was the first published case in history of uh Of a continuous ketone and glucose monitoring individual. So I took that case and from my friends in Australia, Dr. O'Neill and his team, Dr. Kong, and here's the case. I'll do it in 2 minutes. 55-year-old with a 53 year history of type 1, so congratulations to that person for a long life with type 1 diabetes, doing really well, 83% in range. A1C of 6.9. And he was in a study where they were taking an SGLT2 inhibitor or or not, a placebo, and just using CKM. To monitor them, they had a continuous ketone monitor and they had an AID system with its own glucose sensors. So in this case it wasn't a dual ketone monitor, but it was the same ketone monitoring continually. And they were given advice of what to do. So how did that patient do? For months the patient did well, and then there was one episode that I'll show you. So here's the case report. This individual doing nicely, getting along fine. And here's the report of the glucose values in the lower panel and the ketone values in the upper panel and the insulin delivered with the AID system down here, OK. Patience was working out in the out in his garden in the afternoon 4 to 6 p.m. Came in He was going to have dinner, took a bolus of insulin through his pump ahead of the meal. Nice, good work. Took a little extra bolus after the meal, which this person often did. And was feeling, was feeling fine. The glucose went up a little bit and then look what happened. It turned out that his glucose sensor, glucose sensor now with the AID system timed out. It was, it was expired on that moment. So no big deal. He took it off and said, I'll just recharge, I'll just uh I'll just recharge my sensor. This sensor needed recharging and I'll I'll start up a new one. That was actually a 2.5 hour process to recharge a sensor, but He had no glucosis, but his ketones started to go up, and his ketone sensor was working independently, and it showed he had reached this level of 1.0. And he was puzzled and said, Well, gosh, I wonder what that's about, and just thought about it. Then it went up to 1.5, and he got another alert, so he said, I better take some insulin. So he took insulin through his pump, insulin through his pump, insulin through his pump. And enough and, and, and the ketone finally went to 3.0 and he got an alarm and said, whoa, something is not right. And he looked at his sensor more I mean sorry, his infusion set more carefully and it was unattached. It was just hanging there. And so he said, well, gosh, all that insulin I gave never got in. He changed the, the infusion set. And at that time his glucose sensor came back on so he could see his sugars were around 300. He had a new infusion set. He gave insulin through the infusion set and he started to come down, but you could see the ketones were such a good warning for him. Now, did he do it all the way we would instruct somebody? No, you know, he was supposed to change his sight if he ever hit a 1.5. He was supposed to take insulin with an injection, not through his pump if you start seeing ketones. But he at least paid enough attention to them to generate action, to get it going and the ketones were in this case, you know, possibly life-saving or emergency room visit saving for him to get action. So in conclusion, looking at glucose and ketone dual ketone sensor, I think this has the potential to transform prevention if the following happen. There's a bunch of boxes to check, but I hope I gave you the data to say this is a clinical problem. We need to measure it accurately. It needs to get approved by the FDA. We need to continue to work to who's going to benefit most. I think we have a pretty good list building. It has to be easy. Patients have to be able to take action. Clinicians have to be able to take action. We need to give them the educational materials, and then it needs to get into the standards of care. And get covered by insurance. When those boxes get checked, I hope we're going to see the rate of DKA improve just like we've seen glycemic control improve with continuous glucose monitoring. Thanks very much for your attention. Published Created by Related Presenters Richard Bergenstal, MD Executive DirectorPark Nicollet International Diabetes CenterAdjunct Professor, Department of MedicineUniversity of MinnesotaSt. Louis Park, MN
