~1

Present Global Diabetes Education

~2

Hello, my name is William Tamborlane. I am a Professor of Pediatrics at Yale School of Medicine and the title of my talk today is Advances in the Treatment of Childhood Diabetes: Better Care Through Technology.

~3

Before we address how far we have come in the treatment of childhood diabetes, I think it is always good to refresh our memory about the, what I had to call the bad old days of diabetes namely how we treat the children and adolescents with diabetes before 1980. At that time, many of the children were on only one or two insulin injections a day usually in NPH and perhaps getting rapid acting insulin, regular insulin as a rapid acting insulin component. We only had urine test to check metabolic control of diabetes and aggressive therapy to try to normalize blood glucose who thought to be not only unsafe but also of unknown benefits. Therefore, once hemoglobin A1c assays were introduced in the late 1970s or early 1980s, it was clear that many of our patients had hemoglobin A1C levels in excess of 10% and it is also not surprising that when children mature into adulthood with mean plasma glucose values between 350 and 400 mg/dL that they would face the devastating microvascular complications involving the eye and kidneys.

~4

In the early 1980s, intensive insulin therapy was made possible by the introduction of: glucose meters for self-monitoring of blood glucose by patients at home, by the introduction of hemoglobin A1c assays that gave us a simple way of estimating average blood glucose levels in individual patients for the past three months and new strategies for more physiologic insulin delivery including MDI or multiple daily injection therapy and CSII or continuous subcutaneous insulin infusion pump therapy.

~5

These techniques were used in the Diabetes Control and Complications Trial or DCCT to show that improved diabetic control with intensive versus conventional treatment decreased the risk of development and progression of early retinopathy by 50-75%, early nephropathy by 35-55%, and early neuropathy by 60%.

~6

The DCCT included a total of 1441 patients, of those only approximately 200 were actually adolescents between the ages of 13 and 17 years on enrolment. Nevertheless, despite this small sample size, a subset analysis was carried out in the teenagers in the DCCT and that analysis demonstrated that intensively treated adolescents in the DCCT show the same lowering of risk of retinopathy in comparison to conventionally treated adolescents as did adults who were intensively and conventionally treated. Therefore in the Journal of Pediatrics Paper, published in 1994, the DCCT group recommended that most children and adolescents should be treated with intensive therapy to prevent or markedly delay diabetic complications.

~7

Nevertheless, DCCT data itself indicated that there will be special challenges in trying to apply these recommendations to adolescents with type 1 diabetes in the general practice setting. Specifically, compared with adults, intensively treated adolescents in the DCCT had more hyperglycemia with higher hemoglobin A1C levels of 8.1% versus approximately 7.1% in intensively treated adults. Now usually a higher A1C value would be associated with a lower risk of hypoglycemia, however, being a teenager in the DCCT actually was an independent risk factor for severe hypoglycemia and in fact, it was twofold the risk of severe hypoglycemia in adolescents compared to adults during the first twelve months of therapy, and finally there was a similar risk of excessive weight gain. Intensively treated adolescents and adults became obese in about twice the rate in the intensively treated patients compared to conventionally treated subjects.

~8

One of the management of children and adolescents with type 1 diabetes is particularly challenging. Of the past 14 years or so since the end of the DCCT, there have been a number of technological advances that hopefully will allow us to achieve the DCCT recommendations in increasing large numbers of our patients. For example, there has been the introduction of both rapid and long acting insulin analogs. For many years, insulin pump therapy would have very little use in children but over the past 6 or 7 years, it has been a rebirth of interest in using the new and improved insulin pumps in youth with type 1 diabetes and most recently, there have been the introduction of real time continuous glucose monitoring systems.

~9

So let us quickly review the current events that we have now available to us in the treatment of children and adolescents with type 1 diabetes. As you can see, if you recall up until the 1990s, all that was available were regular insulin which as you can see in yellow at time action curve that could extend peak to 2 hours but could have a duration of action of up to 7 or 8 hours and NPH insulin or Lente insulin which were intermediate acting insulin which are shown here in white peaking at around 5 hours with duration of insulin of 16 hours. More recently, we had the introduction of rapid-acting insulin analogs shown in purple aspart, glulisine, lispro peaking very rapidly with duration of 4 hours and relatively flat more optimal basal insulin shown in green, glargine and detimer insulin with the duration of action that approach 24 hours.

~10

So as I will try to illustrate in the next couple slides, one of the issues that lead to more hyper and hypoglycemia in the adolescents in the DCCT was the fact that the rapid-acting insulin component that we have available at the time was only regular insulin. Now it turns out that regular insulin when used in adolescents causes a lot of problems because the pharmacokinetics and pharmacodynamics of regular insulin actually adversely interacts with the insulin resistance of puberty, which actually contributes to both hyper and hypoglycemia. So let us look at this, this slide just summarizes studies a lot of data on the insulin resistance of normal puberty. These are studies that we carried out with biostasticians in the 1980s which showed that adolescents with or without diabetes have a physiologic state of insulin resistance. This insulin resistance we believe is primarily related to the increases in growth hormonal level that occurred during puberty and it is very specific. The insulin resistance of puberty appears to affect primarily the ability of insulin to stimulate peripheral glucose uptake; however, the liver which is normally very sensitive to insulin does not appear to be adversely affected by the hormonal changes of adolescents.

~11

Now in order to overcome the peripheral insulin resistance of puberty, the teenager require large premeal bolus doses of rapid-acting insulin and when those bolus doses are increased, it has a major effect on the pharmacodynamics of regular insulin. Mainly you see two things happening. First, there is a delay peak which may occurred not in 2 hours but 4-5 hours which contribute to the large glucose excursions that frequently occur after meals in an adolescent with type 1 diabetes and the increase post meal glucose values.

~12

Furthermore, increasing the bolus doses of regular insulin markedly prolongs the duration of its actions. Therefore, you get a prolonged elevation in plasma insulin levels above basal. The liver which retains normal sensitivity to insulin is shut down and hepatic glucose production is suppressed, these contribute to late post-meal hypoglycemia.

~13

In contrast, the advantages of rapid-acting insulin analogs are sharper peaks, less overshoot hyperinsulinemia, better control of postprandial hyperglycemia, and reduced risk of pre-prandial and nocturnal hypoglycemia.

~14

There are currently two long-acting insulin analogs, insulin glargine and insulin detimer.

~15

The advantages of long-acting insulin analogs are that they are the first soluble long-acting insulins and that they are prolonged and relatively peak-less pharmacodynamic action makes them better basal insulins.

~16

It is important to point out that there are also significant disadvantages of long-acting insulin analogs particularly as they relate to the management of children and adolescents. First of all, it requires separate injections and cannot be mixed with rapid-acting insulin analogs. And furthermore, their peak-less pharmacodynamic action profile puts a premium on compliance with injections of rapid-acting insulin before each main meal and large snack. We know for a fact that many teenagers are very non-compliant with taking every bolus injections before each meal

~17

So let us turn our attention to insulin pump therapy.

~18

As shown in this slide, this is scenario of considerable interest on our part for now almost 30 years. What we see here is a picture of some of our initial children who were using insulin pump therapy and the title of the paper that we published in New England Journal of Medicine in 1979 discussing the reduction to normal of plasma glucose by using subcutaneous administration of insulin with a portable infusion pump

~19

We and others were very enthusiastic about the potential advantages of pumped over injected insulin. First of all, the pumped use only rapid-acting insulin. Bolus doses were given before each main meal and snack. The pumped treatment afforded the patients increased flexibility and improved lifestyle. And furthermore, patient is required only one injection every 2-3 days rather than 2, 3 or more injections per day.

~20

However, it is actually quite disappointing that there is limited use of insulin pump therapy in children and adolescents prior to the reported DCCT results in 1993. There are a lot of good reasons for that. First of all, the large size and technical limitations of early pumps made it difficult for children to use. There is psychological issue with respect to body image and wearing a device would expose the fact that you had diabetes with some people felt might be difficult for patients. But most of all, by far and away, the largest would have to go with the lack of commitment to the goals of intensive therapy part of that being able to demonstrate that it was beneficial and this act of commitment was in a part of patients, parents but most of all us as physicians.

~21

Surprisingly, there continued to be limited use of insulin pump therapy in children and adolescents even after the DCCT results were published. One of the problems was that many pediatric endocrinologists remained reluctant to use pump treatment in their patients and a lack of data in children was often cited.

~22

To fill this gap in data, we undertook an outcome study to examine the changes in clinical outcomes in patients in our routine diabetes clinic when they switch from injection to pump therapy. The study included the first 161 patients started on CSII on our clinic from 1997 onward. They had to have at least 12 months of data both pre and post switching from injection to CSII. And the major outcomes were changes in the hemoglobin A1C and changes in the rate of severe hypoglycemia. The dataset was closed around 2000 and the paper was published in Pediatric Diabetes in 2002.

~23

It is important to point out that these and all of the other similar clinical outcome study that followed were not randomized and did not have a control group with which to compare the clinical outcomes. However, we had historic benchmarks based on the outcomes of intensive treatment in DCCT adolescents. In the intensive treatment DCCT adolescents were only able to lower him with over A1c value to 8.1 percent. In addition they have high rate of severe hypoglycemia especially when it’s defined as seizure and coma events. There were 39 event for 100 patient years in the first 12 months of the treatment, and 27 events for 100 patient years in the study in its entirely Finally, intensively treated adolescents had a two fold increase in risk of excessive weight gain.

~24

Improvements in metabolic control of diabetes that were achieved when our patient went in injections to CSII are shown in this slide. The y-axis shows the mean hemoglobin A1C levels in this population; population was divided into 3 age groups, 12 to 18, 7 to 11, and less than 7. The dotted line across the top of the graph gives us our benchmark of comparison from the DCCT adolescents about 8.1 to 8.2 percent. And the green bars shown here depict the mean hemoglobin A1c values in each age group when the patients were on injections therapy prior to switching to pump therapy. I think you can see they were pretty well regulated while they were on injection therapy. In contrast, after 12 months of pump therapy, there was a 0.6 to 0.7 percent decline in hemoglobin A1c value across all three age groups. Now it turns out that if one is able to sustain that kind of reduction in hemoglobin A1C, the DCCT data would suggest that you would be reducing the risk of retinopathy on the onset and progression by almost 40%. That is the question, could this be sustained? The final bar, the heads bars actually show the mean hemoglobin A1C levels in each of the three age groups at their last visit approximately 26 months after starting pump therapy, and as you can hemoglobin A1C levels and metabolic control was improved for the entire 26 months.

~25

Normally, a lowering of hemoglobin A1C would increase the frequency of severe hypoglycemic events. So let’s look at what happened in our subjects to the frequency of severe hypoglycemia. In this graph, the y-axis shows the percent of patients with a seizure or in coma event during as shown on the left, the 12 months prior to pump therapy and shown on the right will be the 12 months after the pump therapy. Once again, the dotted line shows the bench mark for comparisons of the rate of severe hypoglycemia in the DCCT adolescents during the first 12 months of that study. As shown by the green bar on the left during the first 12 months, but the 12 months just prior to starting pump therapy, the hypoglycemic rate was very similar to that in DCCT adolescents about 37% of patients had a seizure and coma event during the 12 months prior to switching to pump therapy and in fact this was significantly reduced during the 12 months on pump treatment to 24% of patients with a seizure and coma event during those 12 months.

~26

From the publication of our paper, the Ahern paper shown on top of this table. A number of other investigators have published the results of similar clinical outcomes that is in the sampling as shown in this slide. It has shown here the number of subjects and their age range.

~27

As shown in this slide, the outcomes of these other non-randomized pediatric pump studies were very similar to those observed in our patients namely there was a consistent reduction in hemoglobin A1C values that average to 0.5%, hypoglycemia rates were reduced despite the lowering of hemoglobin A1C values and interestingly there was not, in this studies it did not appear that improved control with insulin pump therapy led to a substantial increase in weight or adiposity as reflected by changes in BMI-z scores. Again, the mean A1C in all these studies was 7.6%

~28

As you may recall, in the beginning of this talk, we mentioned in the bad old days of diabetes hemoglobin A1C values were generally in excess of 10%. However, this slide will illustrate that we can do much better and it provides the data in our Pediatric Diabetes Clinic and the mean hemoglobin levels in our patients as of January 2006. We have a total of over a thousand patients and the mean hemoglobin A1C of all those patients is approximately in January 2006 was some 7.4%. In contrast, the insulin pump treated patients, which will comprise almost 2/3 of our population, the mean hemoglobin A1C in that group was 7.2% and injection therapy patients, the mean A1C values was a 7.8%. These are improvements in control that honestly, I never expected to be able to see in children with diabetes using the techniques that we have available so it is very encouraging that we can really improve in the management of children using the new tools that we’ve had available to us.

~29

However all of the non-randomized clinical outcome studies have not answered the question with respect to how pump therapy compares to multiple daily injection therapy in the prospective or randomized trial in children. Back there until relatively recently, there had been no randomized clinical trials comparing pump therapy with multiple daily injection therapy in youth with type 1 diabetes

~30

So, to examine this question, we developed and implemented our randomized clinical trial that compared CSII versus glargine-based MDI therapies. The study included 32 subjects who were between 8 and 20 years of age. They had a duration of type 1 diabetes more than 1 year so that we could avoid the confounding effects of the honeymoon phase, A1C had to be between 6.5-11% and they had to be naïve to both glargine and pump therapy. Finally aspart insulin was used in both groups as their rapid-acting insulin component.

~31

In this slide, it will show the changes in hemoglobin A1C levels during the 16 weeks of studies in the two treatment groups. Hemoglobin A1C values were shown on the y-axis and the duration in weeks into the study were shown on the x-axis. As can be seen at baseline, glargine group is shown in blue, the mean A1C at onset of the study was 8.2% in the glargine group. There was a slight improvement during the first few weeks of the study but at the end of the 16 weeks, A1C levels had returned to baseline values at about 8.1%. In the pump group, hemoglobin A1C values at baseline were not significantly different than the glargine MDI patients. However with pump therapy, there was a significant fall in hemoglobin A1C levels. Those levels were significantly lower than the baseline values and significantly lower than what was observed in the pump group.

~32

This slide shows the percent of patients in each of the two groups who are able to achieve very strict control of diabetes namely hemoglobin A1C value less than 7.0%. In the CSII of pump group, half of the subjects or 8 of the 16 patients were able to lower hemoglobin A1C values to less than 7% whereas only 2 of the 16 in the glargine group or 12.5% rate were able to achieve this level of diabetes control.

~33

We also had very good data including diary records and meter downloads regarding average pre-meal blood glucose levels throughout the 16 weeks of the study in both treatment groups and these data are very interesting in that as shown here, the fasting pre breakfast blood glucose in the glargine group was approximately 149 mg/dL and in fact with CSII or pump therapy, the mean pre-breakfast glucose was 148 mg/dL. We can interpret this observation suggesting that we were as aggressive and successful in regulating overnight glucose control with glargine as we were with pump therapy. If that is true, where was the problem? Why were the glargine MDI patients are higher at the end of the 16 weeks with respect to A1C. Well, let’s look at what happened during the day in the glargine group. As you can see here, there was a progressive increase in average pre-lunch and pre-dinner values, so at the pre-dinner the mean glucose was almost 210 mg/dL and it is still quite high at bedtime. In contrast, the pre-meal average blood glucose of lunch, dinner and bedtime in the pump group was identical to the pre-breakfast value. So these data suggest that this difference was statistically significant. All these data suggest to us that it is not the problem with the glargine-based MDI versus CSII is not in regulation of basal glucose during the overnight period, it is actually difficulties that many teenagers have with complying with the frequent pre-meal doses of rapid-acting insulin analogs not only every time they eat their main meal but also when they have a large snack.

~34

It is important to point out there have been several other randomized clinical trials comparing pump with multiple injection therapy which reported no difference in A1C. One of those is a Weintrobe study shown here on the table at the top that involve 23 children, 9-14 years of age, while they did not show a difference between CSII and MDI, this was a cross-over study and the large majority of patients at the end of the study chose pump therapy. The other 3 are as a RCTs shown here involved young children 6 years of age or younger and although there was no difference in A1C between the two groups, in fact because of younger age, A1C values in both groups tended to be somewhat high.

~35

One of the factors that is helping to drive the increased use of pump therapy is the technologic advances in the new advanced pump models.

~36

Among a practical advantage of the new pump models are better basal replacement. Basal infusion rate can be varied every 30 minutes in 0.1-0.5 or actually 0.025 U/hr increments. This is especially useful in young children. You can also suspend or reduce the basal rate during the sports and exercise. You can actually program the pump for different profiles. For example, teenagers may have a different profile for weekend when they get up at 10 or 11 o’clock in the morning versus weekdays when they wake up at 6 o’clock in the morning and night following exercise compared to 7:30 days.

~37

The newer pump is actually allowed for better and smarter bolus dosing. Once again, the bolus doses can be varied in very small increments, 0.1, 0.05, .025 unit increments. You can actually give unlimited number of bolus doses, this is especially useful for young children or you’ll never know how much they are going to eat so you can actually divide the meal bolus into three separate small boluses. There is a bolus history function and this bolus history is particularly useful in adolescents, we can actually track to see how compliant our children are with taking their pre-meal boluses. And in fact non-compliant of bolus dosing in pump therapy just like with MDI therapy is the most common reason why you see hemoglobin A1C values rising in the teenager. The new pumps have dose calculators so that the patient enters the glucose values and the amount of carbohydrate for that meal and an insulin dose can be calculated, in fact some of the pumps do not even require entry of the glucose value, they actually have wireless link between the pump and the meter.

~38

There are number of benefits to pump therapy that extend beyond the changes in hemoglobin A1C values and risk of hypoglycemia. Their decreased glucose variability has been reported in a couple of studies. Many have reported patients indicating that it is easier for them to cope with diabetes while on pump therapy. They have greater treatment satisfaction of pump versus injection therapy. For example, 2/3 prefers CSII in Weintrobe cross-over study. And in fact, in all of the studies randomized, non-randomized have been very low discontinuation rates reported when patients start on pump therapy.

~39

In April of 2006, a Consensus Conference was convened in Berlin that was attended by pediatric diabetes experts in the area of insulin pump therapy. These experts concluded that almost all pediatric patients with type 1 diabetes are candidates for CSII. Furthermore, they indicated that CSII was strongly recommended for children with the following clinical features: they have recurrent severe hypoglycemic events, hemoglobin A1C values above the target for age, unacceptable fluctuations in blood glucose, and/or they already had microvascular complications, insulin regimen was compromising their lifestyle was another feature that they thought would strongly recommend the use of CSII. Furthermore, these investigators pointed out that CSII may be beneficial in very young children, adolescents with eating disorders, patients who suffer from the dawn phenomenon, and competitive athletes.

~40

So let’s turn our attention to the most recent technological advance to help us with treating type 1 diabetes in youth mainly the introduction of the continuous glucose monitoring systems.

~41

This slide shows the first continuous glucose monitoring systems that were approved by the Food and Drug Administration, the Medtronic MiniMed CGMS which was in continuous to be a very satisfactory, retrospective analytical system with data collected three days and then subsequently downloaded like a Holter monitor to examine glucose profiles. At the bottom is a first real time continuous glucose monitoring system namely Glucose Watch 2 Biographer. Unfortunately, technical problems made this device very unsuccessful in using children with type 1 diabetes.

~42

This slide shows the second generation CGM devices. On the left is the Medtronic MiniMed Guardian Real Time System and on the right is the Diabetes Freestyle Navigator System which is not yet approved by the Food and Drug Administration.

~43

While none of the currently available real-time continuous glucose monitoring systems are perfect and all need further improvement. I did truly believe that this advance will really revolutionize our ability to manage type 1 diabetes. First of all, use of real-time CGM will allow us to have improved bolus dosing and that there will be trend arrows showing the rate of change in glucose for real-time adjustments prior to each main meal. Secondly, we have retrospective data from sensor downloads to optimize carbohydrates to insulin ratios and correction doses. The real-time continuous glucose monitoring also lead to improved overnight control by having hypoglycemia alarms to alert families that their child’s blood glucose is getting low. This is one of the most feared complications of diabetes. Those families can also again download their overnight data for retrospective analysis to optimize overnight basal rates.

~44

We are very fortunate to be a participant in what is called the Diabetes Research in Children Network with DirecNet. DirecNet is consisting of 5 clinical centers, Colorado, Iowa, Nemours, Stanford and Yale. We have a coordinating center in Tampa, Florida and central laboratory in University of Minnesota which performs the DCCT hemoglobin A1c.

~45

The DirecNet study group has reasonably carried out a pilot study using the Abbott navigator system. This study was in 30 insulin pump treated patients ranging an age from 4-17 years. The sensor we used were encouraged to use a sensor close to 24 hours a day, 7 days a week if possible. It was a 3-month study with optional 3-month extension. And patients and parents use a Navigator to adjust insulin doses in real-time and use sensor download for retrospective adjustments.

~46

This slide shows some of the results of the Pilot Study. The Navigator proved to be quite accurate with accuracy within 10-15% of reference glucose values for up to 5 days. It was used extensively by the patients averaging 130 hours/week. Parents and patients were very satisfied with the device and the data provided by the device. And in fact, even though these patients were very well controlled at entering study with an A1C of 7.1%, A1C actually fell to 6.8% after 13 weeks and this was due to an increase in the percent of values within the target range. There was also decrease glucose variability as expressed by MAGE values. However, it was a little disappointing that we weren’t able to reduce the percent of values that were in hypoglycemic range.

~47

I would like to use this slide, which many people have seen that it is a slide that is taken from the Diabetes Control and Complications dataset which relates the risk of developing microvascular complications to chronic A1C values. A1C values in this graph are shown on the x-axis from 5-12% and then the rate of developing early retinopathy expressed as the rate per hundred- patient years is shown on the y-axis and that rate varies from 0-16%. Now if we look where we started back in the 1980s before 1980 where our mean A1C values were over 10 or 11%, we would expect our patients to convert on an annual basis approximately 12% of patients who had no retinopathy would develop retinopathy per year and in fact when we designed the DCCT, the biostastician said that there would be little problem if there was an effect because by 7 or 8 years of type 1 diabetes back in those days, nearly every patient has some mild background diabetic retinopathy. Let’s see where we are now, in 2006 what we can achieve with pump therapy here is a hemoglobin A1C down to 7.1% - 7.2%, and when you achieve that level of control in fact you reduce the risk of developing retinopathy down to 1-1.5% per year which is an incredibly important difference.

~48

It is important to note however, that the ultimate reality is that no treatment of diabetes will ever be perfect until there is a feed back control of insulin delivery that is regulated by fluctuations in plasma glucose levels.

~49

One way to accomplish this is to use an artificial, mechanical beta cell, the component design elements of artificial pancreas would be insulin pump to accurately and precisely deliver a variable rate of amounts of insulin, and those pumps are already available. We also need continuous sensor to accurately determine ambient glucose levels. And finally effective algorithms to vary insulin delivery rates based on real-time sensor outputs.

~50

We have actually been collaborating with investigators at Medtronic Diabetes to test the feasibility of an external closed loop system that uses an external sensors shown here and a external pump and a laptop computer in this case to use a computer algorithm to vary the rate of insulin delivery. With this system, the sensor actually transmits the glucose data to the laptop computer which then converts the glucose signal into an insulin deliver rate and that rate is send into the insulin infusion pump.

~51

This is actually a picture of one of our study subjects teenage girl wearing two sensors and her insulin infusion pump.

~52

So we have recently completed the study involving 17 teenagers who are in the closed loop control in our general clinical researcher center for approximately 36 hours. Our preliminary observations are that a short term closed-loop control is feasible in children with type 1 diabetes. The night time control is outstanding. And further more that meal-related glucose excursions are as good as or better than traditional open-loop therapy and in fact can be improved with a high bridge system where the patient is given a small priming dose prior to eating main meal.

~53

So finally I think we’re very encouraged by the results of this preliminary study and feel that they are practically applicable. Artificial pancreas may be just over the horizon and really a new hope for a child with diabetes and finally I thank you for your attention.