Video Beyond CGM: Dual Sensing Technology as the Cornerstone for DKA Prevention Play Pause Volume Quality 1080P Fullscreen Captions Transcript Chapters Slides Beyond CGM: Dual Sensing Technology as the Cornerstone for DKA Prevention Overview CLICK HERE FOR CME CREDIT Back to Symposium Thank you very much, Professor Del Prato for this very kind introduction. Indeed, my name is Christoph De Bloch. I'm the head of the department of uh Endocrinology and Diabetology at the Antwerp University Hospital and also full professor at the same university. And I have the pleasure and honor to talk to you about dual sensing technology as a cornerstone for diabetic ketoacidosis prevention. So this is the outline. First of all, briefly discuss what we are available to do and achieve with the currently available hybrid closed loop systems. Then I will get to the indications of continuous ketone monitoring and talk finally about the current status and uh and with the case. So what can we achieve with currently available hybrid closed loop systems? All systems work fine and indeed when we look at the timing range that we can achieve, it varies between let's say 70 and 75% on average. So very, very good numbers that we can see. But despite the use of those automated. insulin delivery systems, still some people develop diabetic ketoacidosis. This is a study from Belgium where we used the, the, the, the tandem pump and indeed you see a hospital visit for diabetic ketoacidosis 12 months before um uh using an AID was 2.4 um events per 100 patient years. And it came down to 1.7 hospitalisations from 1.8 to 0.9. But still, there are patients who present with diabetic ketoacidosis, so it's not completely um um out of um out of the scope of what we see nowadays. So for whom might continuous ketone monitoring be very helpful? Well, first of all, it's important to realize that the worldwide incidence of DKA varies between 8 to 50 cases per 1000 patients, yes, in people with type 1 diabetes. It is a serious, potentially life threatening medical emergency, and it affects 4 to 8% of adults with type 1 each year. Mostly at onset of type 1, indeed in up to 80% of the patients, but even more important is that there's a risk of mortality certainly in children and adolescents, and this is on average 0.15 to 0.3% in developed countries, but this risk increases tremendously to 3.4% to 13.4% in developing countries, something we desperately need, definitely need to avoid. So indeed, what are the scenarios where addition of a continuous ketone monitoring sensor might help indeed when there's failure of insulin delivery by an artificial, an automated insulin delivery system already in the. Past when we did not have those AI days available, uh, AI and colleagues did a pump stop study and he saw that after 6 hours, most of the subjects developed, uh, beta hydroxybutyrate, the main ketone body, above 1 millimolar per liter. So pump stop is definitely, uh, something we, we, we need to take care of and, uh, alert the patient soon enough. These are data from Jennifer Scher, which she presented two years ago at ATTT and you notice in, in, um in um green, the ketone levels and in blue, the glucose levels, and the first bar, you, you see that this patient already had ketone levels of 1 millimolar at the glucose level of 160 mg per deciliter or 8.9 millimoles per liter. And at the glycemia of 250 mg per deciliter, ketone levels were already at 2.5. So the clinical thresholds that we used to uh use um up till one year ago, namely 250 mg per deciliter, are clearly way too late to uh detect ketones. Nowadays, the threshold has been moved to 200 mg per deciliter. But still, this may lead to a delay in the detection of ketones. And is there a temporal relationship between glucose values and ketone values? Well, these are data on file from uh uh about diabetes care, and uh on the top hand, uh, top left hand uh figure, you see that ketones and glucose rise concurrently at about the same rate. In the middle upper panel, you see that glucose levels rise faster than the ketone levels. On the right hand side, you see that the ketone levels rise faster than the glucose levels. And on the bottom side on the left, you see that the glucose rises, but ketones stay flat. And the final example is ketones rise and glucose stays flat. So, How to initiate ketone monitoring, uh, at which thresholds should we do this? There's no clear cutoff, so we definitely need continuous ketone monitoring to solve this problem. Another situation where there's a possibility of DKA is inadequate bolus dosing. This can be deliberately to lose weight, but also pre or post-operatively, certainly in non-diabetic wards where nurses sometimes are afraid to. To give a shot of insulin and usually under bolus or people with type 1 undergoing bariatric surgery and the surgeon thinks that the patient can be left without insulin or transition from the ICU to the regular ward, all dangerous situations. And then finally, of course, uh, or, or, uh, sequentially, the recurrent DKA, the frequent flyers, and those with the high A1C and some ethnic racial minority status, psychological factors and underinsurance, they're more likely to, uh, uh, develop recurrent DKA. So, uh, also sick days and stress can, uh, evoke DKA. Reduced carb intake, this can also lead to uh uh elevated ketone levels. And when we do a low-carb diet, uh, up to point, uh, up to 50. Sorry, Milo, can I start over? I, I will start over from the beginning. Sorry for that. Mm Is it OK if I start over? Thank you, uh, Professor Del Prato for the nice and kind introduction. My name is indeed Christoph de Bloc. I'm the head of the department of Endocrinology and Diabetology at the Antwerp University Hospital and also full professor at the same university, and I have the honor and pleasure of talking to you about dual sensing technology which might be, uh, the trick to prevent diabetic ketoacidosis in the near future. So we will talk about hybrid closed loop systems and uh what results do they give nowadays, but also for whom continuous ketone monitoring might be very helpful and might save lives and what is the current status and finally we end up with a case uh presentation. So indeed what can we achieve with currently available hybrid closed loop systems nowadays? Well, those systems are really performing very well and you can see timing ranges that vary between let's say 70-75%. Every system works. But nevertheless, and these are Belgian data, uh, in a large group of patients who are transitioned to automated insulin delivery, we see that 12 months before start and 12 months after start, there's a reduction in hospital visits for. There's a reduction in hospitalization due to DKA, but this was not significant, and still there are people experiencing DKA, so clearly there is a need for diabetic continuous ketone monitoring. So for whom might this be very interesting? Well, first of all, it's important to realize that the worldwide incidence of DKA varies between 8 and 50 cases per 1000 patient years in people with type 1 diabetes. It is a serious, potentially life threatening medical emergency affecting approximately 4 to 8% of our adults with type 1 each year. Mainly at onset of type 1 diabetes in up to 80% of the patients, but even more important is it's not a trivial situation. There's a risk of mortality certainly in children and adolescents. It ranges between 0.15 to 0.3% in developed countries, but increases to 3.4% to 13.4% in developing countries. So the need for continuous ketone monitoring is really clear here. And indeed, uh, we should be able when using continuous ketone monitoring to detect failure of an insulin delivery by an artificial, uh, by an automated insulin delivery system soon enough. Um, and indeed, when we do a pump substudy, and this was already done, uh, almost 30 years ago, we see that, um, the main ketone bodies, beta hydroxy butyrate are increasing, uh, above 1 millimo per liter in the majority of subjects after 6 hours of pump stop. Um, so this is very important, but Jennifer Scher also showed, uh, during the ATTD meeting in 24 that, um, Already at glucose levels of 160 in the first bar, you see that the ketone levels were already at 1 millimoles per liter and at the glucose level of 250 mg per deciliter, the ketone levels had risen up to 2.5 millimoles. So the clinical thresholds that we used to use, namely 250 and that we use nowadays, namely 200 mg per deciliter, might be too late and it may lead to a delay in the detection of ketones, and this is something we clearly need to avoid. And also looking at data on file from Abbott Diabetes Care, we see that there's no clear relationship or correlation between glucose values depicted in light blue and ketone values depicted in dark blue. On the left upper side you see. that glucose and ketone levels uh rise concurrently. Uh, in the middle upper panel, you see that glucose levels rise faster than the ketone levels. On the upper left, on the upper right-hand panel, you see that the ketone levels rise faster than the glucose levels. And on the bottom figures you see a case where glucose values rise and ketone values stay low, and the other way around to where ketone levels rise and glucose values stay quite stable. So it's difficult which threshold should we use of glucose to tell the patients to monitor or check for ketone levels. So clearly we need continuous ketone monitoring. Inadequate bolus dosing is also something we often see sometimes in young people or in adolescents who want to lose weight, but also pre and postoperatively in non-diabetic wards sometimes insulin is interrupted and then patients develop ketosis, but also a transition from the ICU to the regular ward. And then the frequent flyers, those with recurrent DKA, they, they definitely need uh continuous ketone monitoring. It presents an opportunity, but also perhaps a challenge for these people who are, who are at risk, those with a high A1C, with, with uh some uh. Ethnic racial minority status, psycho psychosocial factors play a role, and underinsurance is, is clearly also playing a role in the recurrence rate of those people with frequent recurrent DKA during sick days and stress. We also need to monitor ketone, uh, levels, and a reduced carb intake, um, can also lead to, um. Increased ketone levels when you consume, let's say 50 g of carbs, there is a risk of slight increasing ketone levels, and this was nicely shown with the cybiotics, cybionics continuous ketone monitoring, and patients followed a very low carb diet and an intermittent fasting diet. And look at the ketone levels of 0.6 millimoles or above, the rates of events per week, 5.2 on a very low carb diet. Ketone levels of 1 millimole or above were luckily very rare, 0.9 events per week, and none of the events was clinically important. So a very low carb diet does increase the ketone levels but not up to levels where it can be life threatening. That's also important to realize. High-intensity exercise, as you can see on the right-hand side on the figure, can also cause ketone levels to rise due to the emergence of counter-regulatory hormones, uh, the, the, the glucagon insulin ratio, which, which, uh, which varies. So also high-intensity exercise can, um, um, um, illustrate when there's a need for continuous ketone monitoring. SGLT2 inhibitors changed the way we manage our people with type 2 diabetes nowadays, but in people with type 1 diabetes, they also run the risk of nephropathy and carddiopathy, and there were some studies that depict one, the EE study, the tandem study with dapa-hiliflozin, EMPA and sotahiliflozin, and there we noticed that the prevalence of DKA uh varied between 1.4, 1.5 to 4% of the people. And important to realize here is that DKA may not necessarily be associated with hyperglycemia above 200 mg per deciliter in these cases. So, and here you see uh a case being described where after the introduction of an SGRT2 inhibitor, uh, the black bars, the autobasal came down, the autocorrection boluss came down, the meal boluss came down, the carb intake was reduced, and this, uh, although the patient had a very good time in range, the ketone levels, uh, started increasing and this patient ended up with a diabetic ketoacidosis. So clearly, the algorithm of the currently available um auto automated insulin delivery systems need to be fine-tuned. Besides the SGLT-2 inhibitors, there's also more frequent use of GLP-1 and GLP-1 JIP receptor agonists in our overweight and obese subjects with type 1. It's off-label, but it's used, and also there people can uh experience nausea, vomiting, and ketoacidosis. And also look at immune checkpoint inhibitor therapy. Some patients develop ketosis. Pregnancy, heavy alcohol intake, prolonged sedentary state, and social isolation are all situations where the addition of CKM might enhance the value of using uh um Uh, an, uh, an automated insulin delivery system and can evo avoid the occurrence of severe diabetic ketoacidosis. So, what about real-world ketone testing? This was nicely shown by Rich Bergensthal. So here they evaluated 135,000 patients using um readers with ketone data analyzed, 2.5 million ketone tests, and You'll be not surprised 92.5% of them had no ketone testing performed whatsoever and even when glucose levels were above 250 mg per deciliter, in the majority of cases there was delayed ketone testing. It took more than 5 to up to 8 hours before patients started. Testing their ketone levels and they were also rarely rechecked within 24 to 70 hours. So, uh, and also what uh uh Richbergensthal and colleagues noticed is that elevated ketones were present even at lower doses so. Um, when you, uh, uh, when you need to check your ketones using a capillary, uh, finger stick, yeah, there's always the initiation that comes from the patient. So the elimination of testing barriers and early identification of elevated ketone levels using continuous ketone monitoring. could really prompt the patient to identify the reasons for the rise in risk in ketones, such as an issue with a, with a pump, uh, a catheter dislodgement or whatever, uh, and to seek early treatment. And CKM indeed does not require patient prompted ketone measurement in contrast to the finger sticks. Of course, When we, when we develop continuous ketone monitoring sensors, it can act as an airbag. You, you might rarely need it, but it could save your life, um, but it needs to be, uh, analytically very accurate. Uh, we need still need to prove that we can reduce the frequency of DKA episodes, that we can shorten the in-hospital stay, uh, when patients have a DKA, and there should be an increment of benefit of CKM added to the CGM, but I think that's already obvious from what I told you so far. And then there should be an economic analysis and also patient reported outcomes should be. Analyzed and as I already alluded to probably novel algorithms uh for the AID device need to be constructed to integrate the data of the CKM and then we need to uh choose uh specific cutoffs and thresholds for alarms and also trend information will be very useful um uh to be visualized. Alvine colleagues developed the first continuous ketone monitoring system. It was tested in 12 volunteers over 14 days. They used wired enzyme technology and they noticed a very nice linear relationship over the 0 to 8 millimolar range with a good stability and accuracy for 14 days. Indeed, this was done with retrospective calibration. But, uh, up to, uh, uh, ketone levels below 1.5 millimolars, the, the, the mean, uh, median, um, absolute difference was 0.13 millimolars and above, uh, 1.5 millimolars, the, the mart, uh, the, the median absolute relative difference was 14.4%. But the sensor was not tested in a setting with dynamically changing uh glucose or ketone levels. We also um tested an implantable um multi-biomarker sensors, uh, able to sense glucose, ketones, and lactate levels, um, using near infrared spectrometry. And here you see a case where this patient was on a low carb diet. The pump was stopped to check the rate of rise in glucose and the accuracy of the sensor, and the patient was asked to drink ketone esters, and then we checked against capillary ketone. levels and you see that there's a very good accuracy and indeed when we look at the ketone accuracy on the right hand side, the mean absolute difference was 0.12 millimolars and glucose readings on the left hand side you can see the Parkes error grid was very, very nice. We need, I think, continuous ketone monitoring, but do we need continuous display of those values? That's another question. Patients also need, need not to be burdened the whole time, uh, by a relatively low ketone levels, so, um, uh, that. The arrows also are very important. One arrow up means a rate of change between 0.25 and 0.5 millimoles per liter per hour. And two trends, um, uh, two arrows up or down mean a change of more than 0.5 millimoles per liter. So it's not only the value that's important but also the direction it's going to. And what about the recommendations? When you have elevated ketones between 0.6 and 1.5, well, you need to monitor your glucose levels. You, you check your insulin pump and the tubings, uh, or your insulin pen, take extra fluids, uh, if you're on a low carb diet, take extra carbs and an extra shot of insulin. And of course when ketone values become higher and higher, you should really reiterate this process, monitor more frequently, drink enough, and indeed add insulin as recommended by your healthcare professional, and above 3, you should seek immediately help and go to the ER. Um, and what about reporting of those continuous ketone monitoring? Well, this was the first attempt by, uh, Haider and colleagues at the latest ATTT conference, and level one ketosis up to, uh, um, 1 millimoles per liter. It should be below 3% and level two ketosis above. Uh, 1.5 millimoles should be below 1% and I think we indeed should report the time above 0.6, 1.0, 1.5, and 3 millimoles per liter. Um, voila, but there are lots of unknowns. What is the time lag between interstitial and capillary ketones? Is this time lag variable or stable in different situations? Is there a certain, uh, ketone value, um, that, uh, calls for action, um, at, uh, to do the same action in different conditions? For instance, a ketone level of 1 millimolar, does it mean the same when you do sports, when you're following a low carb diet, when you consume. Um, and, or when you take an SGLT2 inhibitor or when there's a, uh, a catheter, uh, dislodgement. So different situations probably require different cutoffs and we should also take into account the trend, uh, arrows as I mentioned. And uh we already know that there's no uh clear temporal relationship between the glucose and the beta hydroxy bitterrate levels as I showed you before, and indeed we need uh integration in new algorithms. So here I end with uh two cases. Uh, this is a 55 year old man. He, he's been living with uh type one diabetes since age 2. His A1C is really well controlled, 6.9% and a timing range of 83%. And his last episode of DKA was during childhood. So very well controlled patient. He was gardening, um. And his ketone sensor gave alarms and as you can see, uh, in, in the, in the figure on the left-hand side, his continuous glucose monitor uh uh stopped, uh, so after, after 7 days, it stopped working, he needed to replace it, but the ketone, um, The sensor gave alarms, so the, the patient gave fake carbs to give extra insulin, but he, he, he realized that nothing happened and he was feeling a bit nauseous and became more and more nauseous and then he checked his uh his glucose levels and he noticed indeed high. Glucose levels and dislodgement, uh, and, and he fixed it and prevented hospitalization for DKA. So if he would not have wear, wore, um, worn, uh, a CKM, a continuous ketone monitoring, perhaps he would not have, uh, be in time to prevent hospitalization for DKA. Finally, I end up with a 30 year old uh uh woman. Who's had diabetes for 20 years. She's using also um this uh hybrid growth loop system, the Minimet 780, with a total daily insulin dose of 0.6 units per kilogram per day. Uh she's also really well controlled, an A1C of 7%, time and range of 73%. And uh she's been prescribed an SGLT2 inhibitor and she took uh part in the cybionics continuous ketone monitoring study, and this was a well-educated patient. She knew what to do, uh, and when, when, uh, she needed to test her ketones using the capillary finger sticks, um. In case of symptoms, in case the, the CGM readings were above 200 mg per deciliter, um, so she, she was really well educated. And on the top, uh, you see that, uh, this is a, a, a, a regular day in her life. Uh, at that time, she was not yet taking an SGRT2 inhibitor and she was consuming, uh, um, 200 g of carbs, uh, on a daily basis, and you can see there are sometimes ketone levels, uh, marginally increased but not nothing to be worried about. On day 106, she had a high activity day, and she was already using 10 mg of MA glyflozine. She was still consuming enough carbs, 186 g of carbs, but you can see that there are some occasions where the ketone levels rose above 1.5 millimolars. Now, what do you think the lady should be advised when she's, when she saw ketone levels above 0.6 millimovers uh um and rising ketone levels? Should she check the tubing of her AID system in case of symptoms and change it only if her glycemia is above 250, or should she do this uh earlier? Or for ketone levels above 0.6, already start drinking extra water, ingest extra carbs, and take an extra insulin bolus of 10%, for instance, or for, is it only for ketone levels above 1.5 that she should ingest because she She, she did extra, extra exercise, ingest extra carbs, drink a large glass of water every hour, and administer a bolus based on the carb intake hourly and monitor ketones until they drop down to 0.6 millimolars. What do you think? Well, this is what I would advise. Indeed, in the first scenario, we should not wait until 250. Certainly when this patient is using an SGRT-2 inhibitor, we should already start thinking about it. Certainly, when glucose levels go up to 200 mg and uh the patient experiences symptoms, she should uh check her ketones. Um, and, uh, when the ketone levels are above 0.6, but no symptoms, Yeah, you should not overreact. You can drink an extra glass of water, but whether an extra insulin shot is necessary, yes or no, I think it depends on the situation. What is clear if ketone levels are above 1.5 and you're on a On a, on a low carb diet, uh, ingest extra carbs, take a glass of water, administer an extra bolus, and monitor your ketones. So the follow up, the monitoring is also critically important for our people with type 1 diabetes now. Um, later on, this lady indeed started a low carb diet. She wanted to lose extra weight, and she's very active, so she went for a 60 minute run. And from her previous runs that she did, her previous experience, she sometimes saw that she was going to hypo, so, so she deliberately interrupted her insulin delivery. It's only a 1 hour run. It's not a 2 hour run or a 4 hour run. It's a 1 hour run, and she, she. Uh, consumed 12 g of carbs, and after the run, and administered the bolus of 2 units after the run was finished. So before the run, just uh interrupted uh um uh her insulin delivery. So what do you think might happen? Um, do you think that her ketone values stay flat because it's aerobic exercise, it's running, uh, she did not uh do uh um um um resistance training or uh intermittent high intensity training, it's aerobic exercise, or do you think. That ketone levels rise to 0.6 and then go down again, or they rise to 2 and then go down again, or they rise to above 3.0 millimeters per liter necessitating a visit to the emergency room for to start the DKA protocol. What do you think might happen here? Remember. Type 1 SGRT2 inhibitor, temp stop, uh, temporary stop uh for uh the insulin uh uh delivery for 60 minutes. Well, um, It will not stay flat if there is if there is instant interruption. Indeed, if you, if you continue using the AID and using the temporary target of 150, then it might stay flat, but it will not stay flat. Uh, if you interrupt your insulin, transiently rise to 0.6 millimoles per liter, while the combination here of insulin interruption, and this would increase your ratio of glucagon over insulin. The counter-regulatory hormones start kicking in, the low carb diet, the exercise, and the SGLT2 inhibition use altogether, the ketone levels will rise to higher levels, and indeed, they transiently rose to 2 millimoles per liter. Um. Would it rise to more than 3? I do not suspect it. It was only 1 hour of instant suspension, and she took, as I already mentioned, after finishing her her run immediately extra cops and a bolus administration was administered, so. Um, uh, I think, uh, this is very important, uh, to realize it. So, uh, on another occasion, uh, later on, uh, the, she also manually suspended. Uh, her insulin during biking, uh, having a fitness class, and again biking on the same day, and insulin was suspended during biking in response to a downward trend. I'll show it immediately. So here is the one hour run, uh, on the same day. So this is a very Active lady, yeah, so on the same day biking and uh she interrupted her uh insulin uh delivery because there's a downward trend in her glucose, um, here she followed the fitness class, um, and here again biking, um, for the second time. Um And then um she consumed her dinner later on and uh went to bed then. So, um, indeed this is um a scenario where we learn a great deal of this lady on a high activity day taking 10 mg of carbs uh of, of agliflozin and only 42 g of carbs, so. As I already mentioned, she took her dinner then and at 23 hours she went to bed, but she reported fatigue and she thought, oh this is due to the high amounts of physical activity that I've done a 5 kilometer run, and an 11 kilometer bike ride, then an exercise fitness class, and again 11 kilometer bike ride home, and she had hunger, so she consumed. 50 g of carbs and because of the amount of exercise she did, she said, OK, I will not need a bolus. Her glucose level at that time indeed was only 130 mg per deciliter, but then half an hour later, ketone levels started rising and the following morning, um, she woke up, um, um, she had to uh. Pee several times during the night. Unfortunately, she did not really check her ketone levels overnight, um, and she was very thirsty and very fatigued, uh, the, the next morning. And indeed, as you can see, uh, during the night, her, her ketone values peaked just to 2.9 millimoles per liter. She woke up then, gave an extra shot of insulin, drank enough water, and, uh, the, the ketone levels came down, but still it was not easy to control because she, uh, uh, the, the, the, the, the next day, uh, there was also a tendency, and this is sometimes what we see following, um, uh, intensive exercise. So Indeed, uh, what, what, what should she do, uh, when she wakes up with 2 2.9, go to the ER, um, we should also perhaps, uh, instruct her to interrupt her low carb diet, ingest extra carbs, and take an extra bolus, or interrupt the low carb and HRT. Inhibitor ingest extra cups and additional bolus. I think here the answer is very clear. Indeed she needs to do the combination of the three things, but it's very important here that the continuous ketone monitoring as soon when she woke up enabled her to visualize the recovery process. And this really provided also reassurance for her uh of the treatment effectiveness as the ketone levels uh started to decline. So, um, to conclude, continuous ketone monitoring, I hope that I've convinced you that it's really necessary in the near future. Uh, it can act as an airbag. It can save your lives, as David O'Neill, uh, uh, um, uh, mentioned, uh, in his talks about continuous ketone monitoring. Uh, continuous display might not be necessary, certainly at the low levels in order not to burden the patient with information overload, but it can be very helpful for those who have recurrent DKA or frequent flyers, those on SGLT-2 inhibitors or sometimes also perhaps incretin-based agents, those who want to follow a low carb diet, uh, on the and, and be safe, uh, with that. And for intra and uh perioperative care, it can also be. Uh, be very helpful. So of course we still need to learn a great deal on how ketone levels behave in different scenarios. We need to learn whether there's interference, what about lag time, and so on, but, uh, experience is building up and I hope that we will, um, uh, get CKM or dual glucose ketone monitoring soon available, uh, for our patients and with that I'd like to thank you for your attention. Published September 23, 2026 Created by Related Presenters Prof Dr Christophe De Block, MD Head of the Dept of Endocrinology-DiabetologyAntwerp University HospitalFull Professor, University of AntwerpAntwerp, Belgium
