I'm delighted to talk to you in this summit on optimizing the use of sensor based glucose monitoring and persons across the diabetes risk continuum, I'm going to focus on applying C G M metrics and the use of C G M in optimizing outcomes metrics and those persons with type two diabetes on injectable regiments, including basil, insulin and basil insulin combination regimens. These are my disclosures that I submit for your consideration, and I would like to first begin by reviewing briefly the physiologic context of Type two diabetes. This is a multifactorial, complex dynamic disorder that cuts across multiple organ systems and metabolic pathways, and because of that it is associated with multiple cardio metabolic outcomes with which we are concerned now, as a consequence of this complexity, uh, the leading cause of death in diabetes remains that is, non covid related death remains heart attacks and strokes. But we should not underestimate or understate the implications of chronic kidney disease. In persons with diabetes, it simply speaks to the broad reach across the path of physiologic spectrum of the organ systems that are involved in this disease. Now, what has become abundantly clear is that this complex multifactorial disease requires multifactorial risk reduction in order to prevent or reduce its most guy outcomes and complications. Uh, this was vividly illustrated in the steno study, and this was a study small study that looked at a focus on intensive therapy targeting multiple risk factors. And as a result, after eight years, there were significant and clinically meaningful reductions across all of the microvascular as well as macro vascular complications. Now the findings from this study helped to liberate us in some ways, from the rigid focus on glucose centric approaches to the management of Type two diabetes. But it certainly did not diminish the importance of glucose, uh, management. It's very clear that while we have unequivocally established that there is a multifactorial, uh, risk reduction emphasis that's important in diabetes, an emphasis on glucose management remains central. It is, however, necessary, but not sufficient in order to prevent outcomes of diabetes complications. What we see here in this extensive literature review that was done by County and Jones is that the majority of patients who had a one C levels of greater than 7% who didn't miss the intensive therapy within one year of their diagnosis of diabetes shown in the dark line of the upper line. These were people who had five years significantly increased their risk of a variety of vascular complications. On the other hand, those persons who received intensive therapy before one year of diagnosis of diabetes shown by the Blue Line the lower line not only achieved their goals earlier, but they also avoided, uh, many of the vascular complications risk. So they avoided the disclosure systemic legacy, affirming the importance or centrality of glucose management in diabetes. Our talk is really going to be focused on elements of glucose, uh, management, because see GM is a tool that assists us in more appropriate and more precise management of glucose. So what we have in the management strategies for Type two diabetes is that a recognition of the need for glycemic control as a core urgency. And in fact, the attainment of glucose targets still remains. How we generally define control of diabetes. We see the generation of treatment algorithms from AIDS from a D A and E A S D would provide guidance. They provide direction. Uh, and, uh, they actually show the path towards intensification of treatments, which culminate ultimately in the use of injectable agents, including increasingly GLP one receptor agonist basil insulin is still commonly used and various insulin regimens, uh, using rapid acting or mealtime insulin therapy. Now, indeed, the pattern of treatment intensification in Type two diabetes involves historically, um, a dependence on basil insulin as the first injectable therapy. More recently, we're seeing, uh, more incorporation of GLP one receptor agonist therapy as the initial injectable therapy. But what happens is, if you don't achieve goal with that first injectable therapy, Uh, then, uh, there is a progression to combination injectable regiments. Uh, and that might be, uh, to GLP. One of basil. Insulin was the first store to basal insulin if the GLP was the initial injectable therapy. If goals are not reached, we have the options of adding more complex regiments, including short acting insulin's. We don't use a lot of premixed insulin any more, but they're still available as alternatives for increasing the complexity of the injectable regiments in order to meet patient needs depending on their responses. So with all of these different pathways towards intensification using injectable therapy, it is somewhat perplexing. Then why a one C goal attainment remain Type two diabetes virtually everywhere in the world. As we see summarized in these data, we are simply not getting the job done using the strategies that we have employed. So what have we learned? What are the key remaining challenges in for achieving glycemic goals and thus avoiding many of the burdens of increased vascular complications? Well, it's clear that, uh, the focus of treatment intensification remains on glucose without minimizing the importance of other risk factors. And we're trying to eliminate glucose variability. We can't do that without some dependence on glucose monitoring, to inform our treatment decisions and to see how well are we progressing towards our goals. Now see, GM has emerged as an essential tool for real time assessment of glucose fluctuations in response to the usual variables food or activity or even different medications. Now what we will show you, and what has been defined fairly clearly is that injectable therapy is associated with a broad range of variability. By far the worst, basil insulin has been very useful for controlling fasting glucose levels, but it's often over used in an attempt to try and control plan your glucose levels as well This is something that is best addressed by using rapid acting instruments or, more recently, the GLP one receptor agonist. What the adequacy is of any given combination has to be assessed by doing appropriate glucose monitoring so we can avoid extremes and see GM provides the most assured pathway to the attainment of gold glucose levels using a patient centered approach. Now there's a fair amount of, uh, monitoring that is being done, and yet we're still not getting greater, uh, numbers of people to go. What is it that we're actually getting from our monitoring activities? And do we need more Well, it's clear that the A one c has some limitations, even though it's been considered the standard of care. This is a single laboratory value, and it gives us the 90 day representation of the 90 day average of glucose. But it doesn't detect the daily highs or lows that are so important. An overall control situations night fluctuations during the day are unknown. Similarly, self monitoring of glucose only helps you to the extent that you do the measurement. If you don't test, you miss that particular time point. Uh, even frequent testing can often Miss Hypo and hyper glycemic swings, and it's impractical to be getting a lot of data overnight, and identifying patterns from logbooks is very, very time consuming and cumbersome. See, GM has brought us a lot of advantages. But there's still a lot of data which can be difficult to interpret and gain insight into what's happening with the patient. And this kind of information load can be time consuming to assess. And this is why it's been important that we've been developing different metrics and different tools that are now a lot more helpful and a lot more convenient in their youth. We still suffer from low penetration because of access issues for some patients. Now, ultimately, if we take a step back to assess what is our greatest need and our greatest challenge in the quest to achieve more normal glucose profile, well, the greatest urgency is actually in glycemic variability. What we see here are the two dimensions of glycemic variability on the Y axis. It's a function of amplitude. How high are the excursions? How low on the other end or the excursions on the on the X axis? We see the time dimension. This essentially tells us what's the dwell time? Uh, at those different levels of excursion on the Y axis, we have metrics that include coefficient of variation, standard deviations mean at amplitude of glycemic excursions and so forth. We have various measures of what that volatility our liability is. We can capture it quantitatively on the X axis, the time component. We actually can see how much dwell time there is in each of those different excursions, whether they are in the desirable ranges or whether they are outside of the desirable ranges. And we have quantitative metrics like the mean of daily differences or the continuous overall net glycemic action, which is a quantitative parameter for assessment of interstitial glucose changes. So we now have ways of capturing data that is helpful in assessing variability. Now, the adequacy of self monitoring to share meaningful light on glycemic variability, uh, has problems. Uh, this is highlighted by many of the shortcomings that we see listed here. There are few patients who consistently test their glucose, uh, in sufficient numbers of times, uh, to to yield the most helpful information. The timing may be off, the accuracy may be compromised. Uh, there are some patients that don't actively download the data. And there are few clinicians who have enough time nowadays to spend to detect patterns, especially when they have to be captured from handwritten log books. Sometimes that can look a lot like gibberish in the short and office visit. The patterns of variability are not easily discerned from these kinds of data collections. So what, then, are our most pressing unmet needs, especially during treatment intensification in Type two diabetes? And how might see GM, with its array of metrics and new tools, really help us? Well, a one c clearly has limitations, especially in its inability to account for glucose variability. Um, we don't always, um, accurately reflect related conditions. Avoidance of hypoglycemia is absolutely crucial. It's one of the first things that we try to do because of the safety concerns. And we always want to know whether there is undetected or whether there is uncontrolled exposure to serious levels of hyperglycemia so that we can make treatment decisions, Uh, and those dozing adjustments in an appropriate way. C g m in some ways comes to the rescue because we can get real time actual glucose measurements. Uh, the excursions, whether they are high or low can be detected and the duration, the extent the timing of them can be assessed so that we can pick out patterns and warnings to avoid the risk of hypoglycemia. And this gives us much clearer guidance on the scope and timing of any dose adjustments that need to be made in therapy. This is a very strong template for shared decision making. Our quest for glycemic control is, in truth, the pursuit of reduced or eliminated glycemic variability. Nature disdains wide fluctuations and a person who doesn't have diabetes in the run of the day. Your low glucose will rarely, ever four below 70. Your high glucose, the independent of your diet will rarely, ever go above 138 135 mg per deciliter. Nature. This disdains wide fluctuations. So should we. We must develop strategies to avoid these excursions, and its variability occurs. We must be equipped to promptly detect them and resolve them. Now, it is an important study that reflects the design approach that is used to, uh, evaluate coefficient of variation and glucose changes, uh, and to evaluate the clinical utility of C G. M uh the investigators in this study assessed a variety of therapeutic interventions, including highlighted in red insulin's GLP one receptor agonist injectable treatment. The patients wore two sensors to secure redundant data sets, and they provided logs of their food and any episodes of Hypo uh Listen, Mia. They were subjected both the participants and the investigators, too extensive questionnaires before and after evaluation of the glucose profiles that were achieved. And when the data from such studies has been assessed for patterns of glucose variability, it is clear that the highest variability you see the coefficient of variation on the X axis. The highest variability shown in red was observed in patients taking insulin, including basil insulin. And it made little difference if they had Type one or Type two diabetes across age ranges and across baseline a one CS. Now, some of the key takeaways from these variability studies was that the coefficient of variation was the most sensitive measure of glucose variability for each therapy, and it accounted for almost 40% of the variation. There was increased variability over a broad range across all type to therapy groups, compared to the reports that were generated in healthy subjects. Not surprisingly, the most elevated variability where people with Type one diabetes. But there are many gaps, many unmet needs that persist as barriers to optimal outcomes in patients using injectable therapies that include basil instrument and combination measurement. Now, a major advance in the quest to reduce glycemic variability through improved clinical decision making has been the adoption of a variety of useful metrics from C. G M that provide insights on glycemic stability in much greater detail and with much greater clarity, then can be provided from the A one C We see on the left here, for example, that we can get an assessment by using these metrics of how much time is spent, either in range in the most desirable physiological glucose ranges that we've targeted or how much time is spent below them below that range and how much time is spent above that range. And and this kind of clarity that has emerged has helped us better understand dimensions of average glucose. There was a great deal of ambiguity sometimes between the laboratory measured a one C and the estimated A one C that was derived from C G M For that reason, a new metric called the glucose management indicator has been adopted, and it is based on current C g M. It has replaced the estimated A one C. You're actually looking at actual glucose measures that are used to derive this G M. I. And it gives you a reflection of more recent trends in glucose measurements because the weird times that a user would generate these data from, uh, provides the best alignment of glucose control metrics. And we can couple not only more recent insights into what the glucose trends have been. But we've also announcing the evolution of graphics that have facilitated, uh, efficient interpretation and has facilitated problem solving and shared decision making. Now this G M E is a very useful tool because what has been determined is the following. If the G m I, uh, differs from the laboratory defined a one C by greater than 0.5% because the GMI reflects what the more recent glucose trends have been. If that difference is greater than 0.5, you have to really be careful of an increased risk for hypoglycemia. If you intensify therapy in such a patient, so the G m. I gives us important insights that can be useful in guiding therapy. Now what we have available to us with these core metrics, uh, from C g. M. Uh, this is a rich variety of of insights into the dynamics of, like, systemic changes in the pursuit of reduced variability. We see how much time the person has warned the, uh the sensors, the percentage of time data is actually being collected, and then a whole host of additional metrics that are very useful not only for the termination of being glucose, but for generating things like the glucose management indicator. And how much variability is there, Uh, and where our patients spending their times with respect to these different ranges. And we can get a pictorial profile in terms of the ambulatory glucose profile. So one can ask how compelling is the evidence that use of these metrics can make an important difference clinically. Consider that the most urgent priority in the pursuit of reducing variability in type two diabetes is the prevention or reduction of hypoglycemia. This is by far the most dangerous circumstance that we can encounter. Well, what you see here is real world data from almost 15,000 users with their first sensors. What you see is that there is a clinically meaningful and statistically significant reduction in, UH, low blood glucose readings, whether it's at the level of 70 mg per desolate, all the way down to the more dangerous level, 5 g per deciliter. And most importantly, these important clinical reductions were seen very quickly after the start of sensor wearing observed in the first two days. So if we look at the other extreme, the data from 6 to 7 weeks of where indicate clinically meaningful reductions in hyperglycemia in high frequency scanners, so we can avoid the excursions that characterized variability in glucose. Uh, and this can be seen from data that's been generated in sensor wearing wearers. Well, how do we best use the C G M derived tools in clinical decision making to improve outcomes in those patients who appear to be at the highest risk of volatility and instability of glucose patterns, which are the harbingers of poor outcomes? How do we, uh, bring these tools to better advantage in our clinical, uh, decision making? Now, if we recall from the variability studies across different therapy types. Uh, the insulin taking patients are those shown in red who are especially likely to have the highest range of variability, or simply put the poorest overall control and the highest potential for poor outcomes. What C. G M metrics do for us is to provide a consolidated summary of all of the dynamic parameters of changes in glucose patterns so we can spot if there is a problem. If so, how severe and what is the pattern over what time does it occur and thus provide us with prompts for appropriate treatment directions? You can see here that you have not only the days of where and how much time it was active, but the glucose targets, the ranges that we desire that have been adopted for best outcomes. And then you see what the average glucose looks like, what the G. M e, the derived average glucose measure looks like and how much variability almost 50% variability in this particular representation. And then you see the stacked bar, which essentially tells you how much deviation does the patient have from what is desired? Are they spending, um, more time in the undesirable ranges and less time in the desirable ranges. And then if you look at the pictorial graphic the ambulatory glucose profile at a glance, you can see where the problem areas. All right, So what these represent are different data points that can now be used to assess what the time and range targets are and where the patient is now. These are collapsed, uh, summaries. These stacked representations, these stacked bars, uh, and and they show here in simplistic passion, what percentage of time is the patient spending in each of these ranges? And obviously we'll talk a little bit more about why the more green you see, the better. And when you want to look at real time data in terms of where the patient is as if you're looking at one day's worth of data, you can see that these data represent different periods of where for 14 days each, and they show you the different averages that have been generated over those two different periods of time so you can see what the variability looks like. You see the percentiles 90th percentile all the way down to the lower percentile. Now the question has been raised. Which of these is better. The profile, uh, that you see with real time flow or the time in ranged stacked bar. Well, uh, I think that both of these offer considerable benefits because you can see that you can't always tell by simply looking at one or another representation, what the prediction is going to be of what the average glucose is, and you need a better sense of where is the problem area? Where do you need to spend more time in terms of range, and where do you need to spend less now? What we have here is the ambulatory glucose profile that provides much greater detail providing problem spots and, in addition, uh, to the scope and amplitude of the variability. It lets you know whether it's high or whether it's low in terms of the excursions and the readers I can become quite accustomed to seeing at a glance how near, or how far the pattern that is observed is from the ideal and the ideal. Here is what we see represented, uh, graphically as flat, narrow and in range. It's the f n i ara, if you notice that all of the more desirable characteristics of the glucose profile are highlighted in green, and what we really want is this progression to green. Now Green is extremely important. It's a focal point because Green embodies the core features of our desired outcomes. We want people in range in that green range time in target. Uh, we want them to be there for longer periods of time because that leaves less time for people to be out of range in those excursions that are either on the high end or the low end of the ranges. Now the compliment two more green is to have less red. It is especially urgent to avoid hypoglycemia for reasons that are beyond the scope of this presentation today, but they are well known to us. This is one of the most dreadful, uh, complications in glucose management that we encounter now the clinical utility. Uh, the ambulatory glucose profile. The A G P is best illustrated by the ease with which the provider can assess. Is there a problem? And the targets tell the tale. We know what the targets are in terms of the ranges of glucose and the amount of time that we would like to have patients in those ranges. But we can also see by looking at the graphic rep representations or at the stacked bar, whether or not the patients are spending sufficient time or insufficient time or excess time in the appropriate or desired ranges. And here we see for comparison that while you would like to have, uh, a person in the target range of 70 to 1 80 mg per deciliter to greater than 70% of the time, this person was their only 47% of the time. At the other extreme, you don't want people at the very low range of less than 54 mg per deciliter. You want them? They're less than 1% of the time here. They were in a very low range for 6% of the time, so we can see very quickly that you need more green, less red in a patient of this type. And we can see what the deviations are from the desirable range. So if you take the glucose profile, this is very effective in focusing on where or if you will win. Is the problem occurring? For example, if we look here at the low and very low ranges we see when the patient is dipping into those dangerous zones. So at a glance, you see what your first priority has to be in order to avoid the danger of clinically meaningful hypoglycemia. This is especially important in people who are on intensification regimens involving injectable therapies like y. As you can see what where the peaks are, What times of day are you having inadequate periods of coverage such that you are now experiencing these high excursions and you can see for how long people are dwelling in those zones. So if we look briefly at a case, uh, let's use this as a case in point of how these metrics can help us. This is a 60 year old man with Type two diabetes, no history of CBD or CKD, and his A one c is not a goal. Now he weighs 95 kg. He was tried on a GLP one didn't tolerate it Well, he's now on a regimen of metformin, sulfa, nearly area and clergy at 70 units at night. What you would immediately note is that he's on 0.74 units per kilo of basil insulin, well above the general recommendation of 0.5 units per kilo and his A one c is not at goal. This is a classic picture of an over bazelon ized patient, and you got to start looking for solutions. What do we do as we look at the fact that he is not spending sufficient time in range and he has way too much red? And you look at his ambulatory glucose profile, you see that it is not flat and it's not narrow, it's not in range. And so you you are now beginning to see a pathway towards solutions. Uh, one solution might certainly will, will involve reducing his basil insulin and starting him on some kind of coverage that would provide mealtime coverage. And it may be now when you start mealtime coverage, time to stop is thoughtfully Yuria, so the metrics can effectively target whether action is needed for a patient. Based on the observations that you make from their profiles, you want to generate more green, you want to generate less red, and so you need to do something that will keep the patient from having, uh those periods of of exposure to low blood glucose is in the wee hours of the night of the morning in the late hours at night. And the actions that are needed are going to be actions that not only avoid those lower areas, but that avoid those high peaks those high excursions in the mid to late afternoon periods. So this is what we can now leverage among the metrics to give us a better sense of how do we get to the ideal of flat, narrow and in range? The clear definitions and standardization of the C G M metrics now make it possible to provide guidance for clinicians in how to optimize the use of these data for clinical decision making. Providers can actually use notes on these graphics to capture nuances of the patient that may prove helpful. For example, what the patients wake up time is when they go to bed. When do they have periods of increased activity, what their food history looks like? And if they've missed doses of key medications and the visual feedback that is afforded lends itself to easier assessment of progress? Is it flat, narrow or not? Is there more green, less green, more red or not, there are some additional tools that can add even greater value to the clinical decision making process. For example, we now have directional and rate or pace UH, indicators pace of change indicators in the form of trend arrows. These can provide important teachable moments to patients, and and that's very important for patient empowerment. Um, what we see here is that you can have glucose levels that are rising quickly. Uh, and that can be quantified, uh, at the other extreme. You can see evidence that glucose levels are falling quickly. Likewise, it can be quantified, and you have assessments in between. Now what does this do for patients? Well, this can greatly improve the adequacy of corrective dozing, and I might add that for some systems, these trend arrows can actually provide optional alarms, depending on patient needs. So in addition to the insulin requirements to cover the carbohydrates that are being consumed, uh, in addition to the insulin requirements for correcting the deviation from the measure glucose compared to what the target glucose is, we now can add the additional value of the trend arrow. In a sense, this is helping us to skate to where the puck is going to be, uh, to use an old hockey term, because now, in addition to those usual, uh, insulin doses to cover the basic needs if we anticipate where our needs will be based on the trend arrow if, for example, the trend, Arrow says it's rapidly rising glucose. Now we can add a Bullis that anticipates the influence of that rapid rise in glucose. Uh, and that gives us a total insulin dose that would be quite different from what it might have been without the trend arrow and at the other extreme, If it's rapidly falling, we can drop the insulin, Uh, that is being given by a pre specified amount and again anticipate what the long term need to be. So with these various tools and metrics that are afforded through C G M, uh, it's become clear that optimal clinical utility demands that we take advantage of the cognitive paradigms that have been made famous by Dan Conner, man in the work that he famously did with Amos Tversky, where he essentially, he pointed out that we need to be able to think fast in terms of making real time decisions, things like those corrective actions based on trend arrows the corrections that need to be made based on changes that take place in our dietary intake or in physical activity that we undertake. But but at the same time, in addition to real time decisions thinking fans, we need to be able to look at the data retrospectively to look at what the patterns have been to leverage the power of the ambulatory glucose profile so that we can see what patterns have been in place where the problems have been. What decision making needs to take place in order for those patterns to be moved towards the ideal that flat, narrow and in range pattern. So what we have shown here is that standardized CGM metrics and things like time and range targets can really help patients and providers work together to target the what and the when there are glycemic control problems in order to help craft strategies that will achieve agreed upon goals. Using both the C G M data and the story that is graphically told in a G P reports can help personalize glucose management decisions, C g M. Thus has the potential transform diabetes care with fast. That is real time and slow that look back, that retrospective type of thinking to individualize care and nowhere is that more important than in trying to control and mitigate the volatility that is commonly seen in the insulin requiring Type two patients who are on intensification treatment regimen. So let me thank you very much for your kind time and attention.
