Technology Advances in the Management of Hospitalized Patients with Hyperglycemia and Diabetes

Dr. Guillermo Umpierrez discusses lessons learned from COVID-19, how technology supported the advancements many hospitals were forced to make, and how these changes could impact the future of inpatient glycemic management.

Dr. Umpierrez is a professor of medicine in the division of endocrinology and metabolism at Emory University School of Medicine, Chief of Diabetes and Endocrinology at Grady Memorial Hospital, and Director of the Grady Memorial Hospital Clinical Research Network, Atlanta-CTSA. He heads the Emory Diabetes Council and the nationally-accredited Emory Latino Diabetes Education Program.
Dr. Umpierrez is a national and international leader in the field of hospital management of diabetes and has published more than 300 scientific manuscripts and book chapters, including several landmark papers and guidelines. He has also presented more than 280 research abstracts at national and international scientific meetings.

Additionally, he is the recipient of numerous awards, including the American Association of Clinical Endocrinologists (AACE) Outstanding Service Award for the Promotion of Endocrine Health of an Underserved Population, the American Diabetes Association Outstanding Physician Clinician in Diabetes Award, and the Emory University Department of Medicine Best Grand Rounds Lecture Award.

He serves on the National Board of Directors of the American Diabetes Association and the American Association of Clinical Endocrinologists, and is a member of the Endocrine Society Clinical Guideline Committee and the American Association of Clinical Endocrinologists Diabetes Scientific Council.

Transcript

[00:00:00] Jordan Messler: Hi, everyone, uh, Jordan Messler, I'm the Chief Medical Officer here at Glytec. Honored to introduce our next speaker, Dr. Umpierrez. He's the Professor of Medicine in the Division of Endocrinology at Emory University School of Medicine, Director of the Grady Memorial Hospital Clinical Research Network and Director of the Diabetes and Endocrinology section at Grady Memorial Hospital.

He is a recipient of numerous teaching awards at Emory and national awards for his contribution to medicine and endocrinology, including from the American College of Physicians, American Association of Clinical Endocrinology. He's also the president-elect for the ADA in 2021.

His research interests include mechanisms of beta cell dysfunction in minority populations and the management of inpatient hyperglycemia. He is a national and international leader in the field of hospital management of diabetes and has published several landmark papers and guidelines in the field of inpatient diabetes. The way we manage insulin in the hospital today is largely guided by his research and contribution to national guidelines.

I have known Guillermo since my time as an Internal Medicine resident at Emory and subsequently as faculty at Grady, having an office down the hall from him. His work and support helps for my work, improving the care of patients in the hospital with diabetes and hyperglycemia, both locally at my health system and nationally with the Society of Hospital Medicine and now at Glytec.

He's on our scientific advisory board at Glytec and he's one of the reasons I've joined the company three years ago. It brings me joy to introduce Dr. Umpierrez, our keynote speaker, to discuss technology advances in the management of hospitalized patients with hyperglycemia and diabetes.

[00:01:42] Guillermo Umpierrez: Hello, I'm Guillermo Umpierrez, I'm a professor of medicine at Emory University School of Medicine in Atlanta, Georgia, and the President-elect of the American Diabetes Association. And it's my pleasure today to discuss with you Technology Advances in the Management of Hospitalized Patients with Hyperglycemia and Diabetes.

I should start with some potential conflict of interest. Emory University has received funding from AstraZeneca and Dexcom for investigator initiated research studies. And the agenda for today is to discuss in technology three areas: glycemic management systems, pumps, insulin pumps or continuous subcutaneous insulin infusion, continuous glucose monitoring, and we just started doing studies with closed loop systems.

We know that we're in the middle of this epidemic of diabetes. This slide to the left shows the number of people with diabetes from the 1980 to mid 2015, there was about 34 million people with diabetes according to the CDC. And we know that patients with diabetes are admitted more commonly more often than patient without diabetes. It's about 8 million discharges, one in four patients in the hospital have diabetes.

And if you look at the distribution of patients with hyperglycemia in the hospital, this is data from the group of Curtiss Cook and Mayo Clinic in Arizona, so in the top panel, in the ICU and the bottom panel, the Non-ICU, and you're looking at the percentage of patient who had a blood glucose greater than 180 and 200.

And you see that one out of three patients have glucose more than 180, and one out of four patients have a glucose greater than 200. And it doesn't matter if you are in the ICU or Non-ICU, the number, the prevalence of hyperglycemia is similar in both groups. And what is important is that hyperglycemia is associated with an increased rate of complications.

Here you have the levels of blood glucose going from 100, 200, over 300 and this is the composite of complications that including pneumonia, accute kidney injury, respiratory failure, acute MI, bacteremia, and death. The higher the blood glucose, the higher the rate of complications.

So this is quite important. Hyperglycemia increased the inflammatory response, increase the oxidative stress endothelial dysfunction. And it doesn't matter if you're admitted to the ICU, Non-ICU. This is data, all data for Mercedes Falciglia from Cincinnati that shows that in patient with hyperglycemia in the ICU, this is data in over 200,000 patients admitted to different VA hospitals.

So based on where the glucose, the higher the blood glucose in no history of diabetes, or with history with diabetes is associated with an increased rate of mortality. There's particularly no history of diabetes with hyperglycemia, mortality was much higher than in patient with diabetes and similarly, Non-ICU, this same story repeat.

So here to the left is patient with diabetes in the red bars are patient with glucose greater than 180, and in the yellow, less, less than 180. Then you see here to the left, in diabetes, hyperglycemia is associated with an increased rate of death reoperation and infections, but look to the right, patients with no diabetes.

This is in 11,000 patients that underwent colorectal and bariatric surgery. Hyperglycemia, glucose greater than 180 is associated with increased rate of death, need for reoperation and infection.

So hyperglycemia is important, both in patient with, and without diabetes. And this has led to current guidelines.

That's right. We used to say target between 80 to 140, now we want 140 to 180 in most patients admitted to the hospital. In our unit, we just try to get somewhere between 80 to 180 or 100 to 180. It doesn't matter if you go from 100 to 140, 140 to 180, we just want to avoid hypoglycemia.

Glycemic target, we modify according to the clinical status. That is important. That's right. If you have a patient with repeated episodes of hypoglycemia who have already kidney failure or heart failure or it's terminal illness, just keep the glucose less than 180, 200 should be okay. The most important thing that I should avoid hypoglycemia.

So you remember the Van den Berghe studies in the 1990s, we recommend the glucose less than 110, not anymore. We don't recommend glucose greater than 180. We want blood glucose between 100 to 140 for most people in the hospital setting.

How we manage patients. In the ICU, we use IV insulin for people who are critically ill. Why? Because you can titrate rapidly, you can adjust according to the patient meals. There's no really contraindication for the use. So very critical ill patient, the IV insulin has been shown to be effective.

And there are several protocols that have been published. That's right, the Leuven, the Yale, the Glucommander, NICE-SUGAR. They all achieve glycemic control, but the thing is that you have to measure glucose every hour to every two hours, but you can manage the patient very well.

So, if you look at the different series, this is the glycemic values that are achieved with IV insulin protocol versus control, in the intensive treatment group, the blood glucose that they achieved was excellent. That's right. Most of these trials was to achieve a glucose less than 120, and the control to glucose less than 180, you see that they get about 150. So, reaching glycemic targets can be done with multiple, multiple protocols.

The problem has been hypoglycemic events. Hypoglycemia with intensive treated group, we've seen in, uh, in two or three fold increase the rate of hypoglycemia. We reported between 5 to 20%.

So, this is why the American Diabetes Association has relaxed the recommendation for target. And of course, in the NICE-SUGAR studies, those patients who has moderate or severe hypoglycemia, was associated with 30 to 100% increase in mortality.

So, you have to, when we manage patient now, we take in consideration the importance of the reducing hyperglycemia because you decrease complications. But in the other hand, you have to prevent hypoglycemia which can be associated with increased rate of neurological, cardiovascular, and overall mortality in these people.

So how can we manage patients in the low range and the upper range? And I think that this Glucommander or these computer guided insulin infusion algorithm, has been important in achieving those goals.

We reported the results of a study that we call GLUCO-CABG trials, and we, we divided 300 patients in two groups, 100 to 140 and 141 to 180, and both were using the Glucommander guided insulin infusion. These are patients who underwent a coronary artery bypass surgery.

And here you have the intensive is the orange and the conservative is the green. So in the intensive, we got an average blood glucose at 130. So we wanted 180 to 140, you got 130. In the conventional, we want 140 to 180 and you got 160. And this is in the ICU and you will see very good control. And the most important, is the rate of hypoglycemia was minimal.

So, how can you achieve good control, maybe with a glucose monitoring with this ancillary support of these infusion protocols and with daily or hourly testing of blood glucose what you get is no patient with glucose less than 140 very few patients with glucose in intensive group with hypoglycemia. And if you go to 140 to 180, really no hypoglycemia at all.

The improvement of glycemic control was associated with reduced rate of complications in patient with no history of diabetes. In the upper panel is the perioperative complications in patient with diabetes, there was no significant difference in MACE, major adverse cardiovascular event, respiratory, acute kidney injury.

But look in patient with stress hyperglycemia. These are patient with no diabetes, no hyperglycemia, that underwent bypass surgery with blood glucose elevated, you have a significant reduction. So keeping the blood glucose less than 180 was associated with an reduced rate of complications in patient with no diabetes.

And how, one of the things that we look in these studies, what was the acceptance of the nursing staff? And most people agreed that they like the, reminders every hour to blood glucose and more importantly, the guidance on how to adjust the insulin therapy. So I think that now they have several of these software-driven insulin dosing, but I think they have now good, good data on the Glucommander both in the ICU, Non-ICU, patient with DKA and others.

And the tool also has been shown to, to improve diabetes control. There are a couple of publications with Glucostabilzer, and there are some others done in Europe.

So what is the future? I think that they help improving glycemic control, but more importantly, they reduce hypoglycemia or allowing you to achieve target with less hypoglycemia.

Let's move now to insulin pumps, continuous subcutaneous insulin infusions. And here you have the number of people using pumps in this country, and you have 400, about half a million people. Everybody with type one diabetes should have an insulin pump, if possible. Of course this is not, not real, but in the number have increased.

 So now, we have patients using more pumps. They're admitted to the hospital, should they be allowed to use pumps in the hospital? And there is no prospective randomized studies. We only have data and retospective observational data, but it shows that it works very well. With a good number, it better equal similar to patients with SubQ insulin. The hypoglycemia is even less, less than 70 was seen in very few people. And you have a retrospective data.

And this is from Curitss Cook in Arizona. The Six-year experience, 250 patients. And when you look in is all insulin pump ,people were treated with pumps versus people who were taken off pumps to SubQ insulin. And here you have the rate of hypoglycemia and it's equal or better.

So these are, these type of study have allow that we make the recommendations that maybe you can use, eh, these insulin pumps in the hospital. Unfortunately, there is no randomized control studies. This shows hypoglycemia when the pump off or pump on, there is absolutely no difference in this group of people.

So I think that in our unit, we allow for most patients use an insulin pump. They can be safely managed in the hospital setting if the patient is oriented, is conscious, can help managing the pump. Many nurses don't know how to deal with pump. Many doctors don't know how to deal with pump, so you need to have a good intra, infrastructure and support system to allow people to use the insulin pumps in the hospital.

In the few years ago, there was a, here you have continuation of insulin pump. This is a different series, a small series, just 92 patients. Most people with type one diabetes during surgery, they did very well. And there you have during surgery if you continue the insulin, continuous insulin pump therapy, the blood glucose was okay. If you stop the pump and you start treating with IV insulin, with SubQ insulin, a little higher blood glucose levels. So more than 180 in 80% if you started giving SubQ insulin but again, for surgery patient, we only allow for a small surgery just less than two hours. For longer, we don't recommend.

So this has led to several position statement of the use of insulin pump in the hospital setting. So this is what we do in our hospital. This is what we reported with David Klonoff in diabetes care in patient critically ill transitioned to IV insulin.

The patient is not critically ill or surgery is less than two hours, you can continue if you have a trained personnel, if you have a protocol in the hospital. If the patient is able to operate, self management is important because not everybody knows how to deal with insulin pump.

Look at severity of illness. It can not be used with MRIs. The CT is controversial, some hospital allow CT some do not. In our hospital, we use a lead apron, and we don't take it off for a CT scan, but you must take it out if, if you have, patient undergo an MRI. And that's the decision, if you continuous insulin pump or you transition to multiple daily injections.

So something to take in mind is the level of educations. There are many patients that we see in clinic, the same in the hospital. They have very poor, very poor knowledge and is somewhere about 30 to 40% of the series in the Cleveland clinic shows that they have very little education.

So, if you don't know, if the patient don't know what kind of rate, basal rate they have, how to adjust, how to supplement, how to do correction doses, better take them off insulin. If you don't have the appropriate team, take them off.

In addition, patient with DKA and HHS do not be, cannot be treated with insulin pump. As we know, many times if a pump fails and patient develop DKA. So you cannot treat them with insulin pump because you are going to start adjusting insulin every hour in these patients. Critical illness or patient who have loss, loss of consciousness, or long-term anesthesia, they should not be on pumps.

Eh, as I mentioned, people cannot appropriate demonstrate the use of the pump. Psych, refusers or not willing to participate, if they don't have pump supplies, does your hospital has, pum, pump supplies or not. Then of course the most important thing is the lack of trained healthcare providers. And you have to have the support of the institution.

So what is the future? I think that we should continue, but if you don't feel comfortable, you can switch the patient what we call pump holiday. The pump holiday protocol, you are going to take away, replace the 24 hour basal dose and you see 20, 24, 30 units, whatever it is. And then you'd give this as a basal insulin.

Remember the half-life of of insulin is short, so you have to give the basal insulin one or two hours before you stop the insulin pump. And you have to combine this with mealtime SubQ rapid-acting insulin.

Most people who use pumps are type one, so they have to be on a basal/bolus plus correction protocol and you use the same insulin-to-carb ratio that they were using before. The correction-dose of algorithm should be ordered, according to the usual insulin sensitivity factor, this seldom happen in most hospital.

So what we do at Grady is that we just take the number of insulin, basal insulin, the patient is receiving and we give it SubQ. We gave the same one to one. So, some doctors order 80% of the basal rate to, in SubQ insulin eh, especially if you use data mirror that has a peak. Glargine has less than a peak. Glargine U300, Degludec, maybe one-to-one should be okay. Eh, and you do basal/bolus approach given according to the prandial, according to the insulin-carb ratio that the patient was using before.

So insulin pump. In summary, we can say that we don't have randomized controlled trial, but there are multiple publications, retrospective studies and consensus report that we should continue insulin pump if the patient is able to manage their insulin pump in the hospital. But no data that is better than to switch them to SubQ or not, or MDI.

Eh, definitive patient like it. So there is improved patient satisfaction, but you must have, you must have a support team. If this is not available, you should take away the pump and start them in the holiday protocol. There are several studies suggesting retrospective data, again, that if you switch them to basal/bolus approach you do fairly well in this group of people.

And finally, policies and procedures on pump use need to be in place at each institution. They have to be aware that not everybody should go continuous insulin pump, especially if you don't have the support system.

Finally, let me talk to you about the CGM in the hospital setting. So the way that we have adjusting insulin therapy, of assessing, glucose monitoring in the hospital during the last 50 years, is with fingersticks. And it works well. They tell you what the blood glucose is 2, 3, 4 times a day. It can tell you if the patient is low or high, it depends on how well you going to use it.

In the ICU, if the patient is receiving insulin infusion, we measure every one or two hours, in a Non-ICU, most hospital measure before meals and at bedtime. So everybody in all hospital has a procedure how to transmit the glucose from fingerstick to the electronic medical records, so it's good. But, the problem with fingersticks, they are isolated points. For example, you do before meals and at bedtime, but you don't know what happened in between meals or especially at night. So now we have several studies that have shown that if you use continuous insulin infusion, a continuous glucose monitoring, you're going to produce a much better assessment of glycemic control.

The other thing is important is that in the hospital, asymptomatic hypoglycemia is pretty common. This is study in 250 patient, prospective study that we conducted at Grady Hospital. In the brown area is asymptomatic, in the blue is symptomatic. 45% of episode of glucose less than 70 were asymptomatic.

So we went to the bedside. We were called, beep by the laboratory as soon as the blood glucose was less than 70, we went to the patient's room and we asked for symptoms and scale of hypoglycemia and they were asymptomatic in about half, especially with increased age. Those patients over the age of 50, 65, look at the ratio of having asymptomatic hypoglycemia.

For some reason, don't ask me why, males have less recognition of hypoglycemia than females. Or those patients who has kidney failures are also less likely to have symptomatic hypoglycemia. So the fact that we don't test, the fact that many people have asymptomatic hypoglycemia, especially during the pandemic, we have now recognized the value of using continuous glucose monitoring in the hospital setting.

CGM comes in different flavors. Invasive: intravascular, venous, arterials. They're not very common now, but there are several studies from many companies who are trying to use these intravascular venous and arterials. We use the subcutaneous, that's right, and there's some trends there, also and the research, but we're using mostly factory calibrated, eh, CGM. Sampling frequencies take 1 to 15 minutes.

There are several studies in the ICU that have shown very good correlation, more than 20 studies have been published. To tell you the truth, 35 studies in a recent review, that list they use a CGM in the hospital setting in the ICU.

And here you have the Clarke Error A and B. That's what you want to be, a good correlation with fingersticks we're seeing in somewhere around 98, 99% of patients. So it works well, but you measure glucose every hour in the ICU by fingerstick. So you don't really improve glycemic control, you facilitate the care, but therefore it has not been approved by the FDA for use in the ICU.

Eh, a recent systematic review that I mentioned before, 37 studies combining randomized control study, observational study, shows that they're very good. They're, the efficacy was good in the 98, 95, 99%, but you don't improve eh, the outcome with these patients.

So we're still doing some ICU, especially during the pandemic. Because what we wanted to do is to decrease the rate of fingersticks

and expose the patients, and this has a few of our publications using continuous glucose monitoring in the OR, in the ICU and in a Non-ICU during the COVID pandemic. And what we reported is that in the ICU, if you put the CGM together with the Glucommander, you can place the CGM up to 20 feet away, from the CGM, and you can place this at the patient's wall and they together next to the insulin infusion.

So this is the studies that in New York by Dr. Agarwal, who reported in diabetes care earlier this year, that you can have continuous, or you can fit your insulin infusion according to the glucose values.

Eh, this is the data from our unit. The, we placed the CGM next to the Glucommander and we have an iPad at the nurses stations and the same like the cardiology have the Holter monitor, we have 24/7 glucose monitoring. So if the glucose drop below 80, you go on to the bedside and you confirm the hypoglycemia or you treat. The, but more importantly, you can adjust the insulin infusion protocol. You can adjust the insulin infusion by just looking at the meter instead of going to the room and prick fingers of the patient.

So there are some technology limitations that we're still working on. There's some substance interference, they have improved significantly. There is about 5 to 10 minutes measuring lag. Eh, some groups say more than 10 minutes, an hour, look like it's somewhere between 5 to 10.

Eh, we don't have accuracy, let's say for arterial hypotension, hypothermia, or hypoxia.

Intravascular CGM, don't use it because many risks. The other thing is cost, that's right. And the other is that limited data in favor of tight glycemic control in the ICU. So I think that in the ICU the use of CGM is okay if, during the pandemic to reduce the exposure, but the real use of CGM is going to be in the Non-ICU setting.

This table just group a few of the studies that have been reported with the use of CGM in the Non-ICU setting. So the study that we use with, with Medtronic CGM, the average daily glucose here, CGM in red, point-of-care testing in green, this is days in the medical ward, and this is the accuracy. So very good accuracy with Medtronic and in another studies, you will see that you recognize hypoglycemia significantly better with CGM. And this is between midnight to six o'clock in the morning.

This where the more cases of hypoglycemia recognize that of course we don't do fingersticks after bedtime. That's right. And that's the beauty of the CGM. You also recognize some hyperglycemia, but what is, I think the most important thing is the recognition of hypoglycemia.

In a paper published eh, for our group, eh, Rodolfo Galindo last year shows, the mean hospital daily blood glucose, and this was comparing the Freestyle Libre Pro Flash, so the Abbott CGM, and you see, there's a very nice correlation between point-of-care testing and CGM. This is from day 1 to day 10. CGM reads a little low, about 10 milligrams lower than point-of-care testing. But again, the beauty of using CGM is recognition of hypoglycemia, especially nocturnal hypoglycemia.

So if you order fingerstick before meals and at bedtime, there is nothing in between 10 o'clock or midnight to six o'clock in the morning. That's where a lot of episodes of hypoglycemia, occurs, less than 70, less than 54. That's something that we call prolonged hypoglycemia, greater than two hours. So if we believe that hypoglycemia is serious or may have serious consequences, recognizing this hypoglycemia should be of great value.

Very good correlation between Point-of-care testing and the Libre with 98, 98.5% in Clarke Error Grid A and B. Overall, the MARD was about 14 and not very good in hypoglycemia, although, the rate of hypoglycemia was small to make a definitive conclusion.

Those were using the Libre Pro, what about Dexcom? And here you have in a paper published by Dr. Georgia Davis that 98.5% has very good correlation with the first 24 hours in the overall, do you see that there is about 98.5% in the A and B Clarke Error Grid with MARD of about 14.

Things to take in consideration is that there are two situations in which the MARD is not very good, the mean absolute relative difference between point-of-care testing by fingersticks and CGM. One, is in the hypoglycemia array. In the hypoglycemia, the MARD is much more, less than 70 with higher blood glucose concentration, and the other is the patient is very anemic with a hemoglobin A1C, less than 7 in grams, they hit, the MARD was not as good. So if you have a patient with a glucose less than 80 you have to order a fingersticks.

One of the deficiencies or limitations of the CGM, is that you get the reader in the cell phone or in the apps, but it doesn't be transmitted to the nurses stations. So, we have been working what we call the Glucose Telemetry System. Can you send the information all the way to the nurses stations, and more importantly, alert the nurses when they blood glucose less than 80 or 85 to prick fingers and treat accordingly?

So here you have, we placed this CGM, it can be in the arm or in the abdomen, this is a Dexcom, by the way, it's only approved for the abdomen for Dexcom, but it can be any way. Both are very accurate. And here you don't see very well, but you have the Glucose Telemetry System.

So here you have patients in Day 1, on admission. Day 2, how the blood glucose changes, and Day 3, blood glucose concentration. And then you have what is called the Hospital Glucose Profile. You see the lows, the in-target range, or the high blood glucose. You can also look at glycemic variability and the percent time. So if you use this, can you use this method, the Glucose Telemetry System, can you reduce the rate of hypoglycemia?

With the University of Maryland, we've reported last year that you reduce the rate of hypoglycemia. So if the blood glucose is less than 85, the nurse that went to the bedside, prick the finger, confirm a blood glucose low and gave the patient some orange juice, crackers before the patient dropped to 70, 50 or 40, or who knows.

We also were able to demonstrate the reduction in the percent time below range. Using the telemetry. These were done with a Dexcom, eh, CGM.

So what are the challenges? Well, it's new technology, relatively. Not all the hospitalists know how to deal with that. Our primary care don't use much CGM so, we are right now educating them.

Some of this CGM needs calibration. Fortunately, we have two that do not require calibration. That is, I've been informed that a third one would not require, a C, calibration is the Medtronic that is sometime during the summer or fall this year.

There is interfere with some of these medications. This has been getting better. They could be information overload. I mean, what happened? Did you get all of this glucose every day, multiple times a day. Can the doctor get, you know, freeze, frozen because too much information. Eh, we don't have data on hypotension, hypothermia, hypoxia. Eh, as I mentioned, the real time or telemetry is not everywhere so this is something that we're working, how you can do this. Cost, I think that there is some data suggesting that maybe if you decrease hypoglycemia, is cost savings.

So what's in the pipeline? Well, closed-loop. The group from, from England, from the UK has reported three studies on Non-ICU and patient receiving enteral feedings, and that you see that the closed loop, so connecting an insulin infusion to a CGM, reduces the blood glucose, improved blood glucose concentration, and more importantly, decreases the standard deviation.

So it is a quite important, quite impressive. We are right now conducting studies in, at Grady with the use of CGM, CGM connected to an insulin infusion. You know, is this going to be for everybody? I don't think so but of course, if somebody comes on, an, on a closed loop system that now we know there are a couple in the market, you may continuous using it if the patient is alert and is able to manage their glucose levels.

What can we said about CGM? I think that data from clinical studies is emerging and I think that this data is very promising. You can use CGM in the hospital setting. We still need some randomized control study if you can really manage patient with CGM alone. I mean, the preliminary data is very, very exciting, but we still need more data.

Eh, and I think that the improve accuracy and they reduce needle calibration and especially during COVID has make our use of CGM much more appealing in the hospital setting.

So in summary and future directions, we need to conduct appropriate studies for FDA approval on the use of CGM in the hospital setting. It is my understanding that Dexcom is working with the FDA in a, in a prospective studies, randomized studies to see the accuracy, but not compared to point-of-care testing because fingersticks are inaccurate. So it has to be compared with the laboratory or the YSI, the Yellow Springs Instrument, that is very reliable blood glucose concentration.

We are working very hard in educating and training the hospital personnel, and more importantly, primary care with the use of CGM. We need to work on the simplified system to data transmission, to the bedside, to the nurses station, and more importantly to the electronic medical records.

So we are there, there are a couple of publications that a couple of hospitals who are doing that this is going to be very important. We need pharmaco-economic analysis and of course the accuracy of the CGM together with the insulin infusion protocol have been shown to be quite effective.

Close loop system in the hospital setting. Is this going to be for everybody? I doubt very much. It takes several hours to connect new patients on a close loop. Maybe to have a patient with type one, that could be there, but if you, somebody comes with an insulin infusion, they may continue. If somebody comes with a closed loop system, they may continue using the device in the hospital setting.

 I thank you very much for the opportunity to review with you what is new advances in technology in the hospital setting. Thank you.

[00:36:31] Jordan Messler: Thanks Guillermo for that terrific presentation. We're honored to have you here today. Thank all of you as well for attending these two days at our inaugural conference, Time to Target. We appreciate your attendance as well as appreciation to all of our amazing speakers. And we look forward to seeing you again next year.

MAR-0000458 Rev 1.0