Medical Devices Advisory Committee




НазваниеMedical Devices Advisory Committee
страница2/16
Дата26.10.2012
Размер0.59 Mb.
ТипДокументы
1   2   3   4   5   6   7   8   9   ...   16
Agenda Item: Sponsor Presentation -- Introduction

MR. GREGG: Thank you very much.

We are getting our slides set up. So, if you will indulge me for a minute or so. We will get the first slide up and then we will proceed.

While they are doing that, I will go ahead and make my opening remarks.

My name is Terry Gregg, as indicated. I am the president and chief operating officer of MiniMed.

We appreciate your review of our premarket approval application. We are very pleased to be able to present you with information on the MiniMed Continuous Monitoring System.

Let me introduce the speakers, who will be discussing this new device that we believe will serve as an important tool in the management of people with diabetes. Our first speaker is Dr. John Mastrototaro. Dr. Mastrototaro has a Ph.D. in biomedical engineering from Duke University and is MiniMed senior director of sensor development and manufacturing.

John has extensive experience in research for subcutaneous glucose sensors and was responsible for the development of the system that you will be reviewing today. Dr. Mastrototaro will provide an overview of the Continuous Glucose Monitoring System and its principles of operation.

Our next speaker will be Dr. Jorge Mestman. Dr. Mestman is an endocrinologist and director of the University of Southern California, Center for Diabetes and Metabolic Diseases. Dr. Mestman has participated in both the initial feasibility studies and multi-center clinical trial of the continuous monitoring system.

Dr. Mestman will discuss his experience using the device in a total of 75 subjects at the University of Southern California Ambulatory Health Center.

Following Dr. Mestman, Dr. Todd Gross will discuss the overall results of the multi-center clinical trial and the statistical analysis of the data collected during the study. Dr. Gross has a Ph.D. in mathematical psychology and is MiniMed's manager of corporate statistics. Todd is also a faculty member in the UCLA Department of Humanities.

Our final speaker is Dr. Alan Marcus. Dr. Marcus is board certified in internal medicine and endocrinology and metabolism. Dr. Marcus is an associate clinical professor at the University of Southern California School of Medicine and has a large private practice in Southern California.

He also serves as a medical advisor to MiniMed. Dr. Marcus will discuss the use of the data provided by the Continuous Glucose Monitoring System in the management of people with diabetes.

In addition to our presenters, Mark Filase(?), MiniMed's director of clinical and regulatory affairs, Judy Spell, manager of clinical research and Jean Charleson, manager of clinical programs will be available to answer questions regarding our application.

Thank you again for considering our PMA for this exciting new device. We will begin our presentations with Dr. Mastrototaro.

Agenda Item: System Overview

DR. MASTROTOTARO: Thank you.

As Terry Gregg mentioned, in this portion of our presentation, I will give an overview of the MiniMed Continuous Glucose Monitoring System, which will be referred to often today as the CGMS. In the overview, I will discuss the indications for use for the product, what we believe is the importance of continuous monitoring and diabetes care, give a description of the CGMS, discuss some of the early feasibility studies, which led up to the multi-center clinical trial, which we will be presenting data on later, describe the important issue of sensor calibration when using this product and then I will describe how the CGMS will be used in practice by the health care team and provide you examples of patient data, which is presented to the health care professional and the team and finally discuss a simulated patient study we conducted at the end of last year at the FDA's request.

As you heard earlier, the indications for use for the CGMS, it is intended to continuously record interstitial glucose levels in persons with diabetes. This information is intended to supplement, not replace, blood glucose information obtained, using standard home glucose monitoring devices. The information collected by the CGMS may be downloaded and displayed on a computer and reviewed by health care professionals.

This information may allow identification of patterns of glucose level excursions above or below the desired range, facilitating therapy adjustments, which may minimize these excursions. Two important points I would like to make regarding the indications for use: One is that no real time glucose information is provided to the patient while wearing the product. Only the health care professional is able to view the data after the patient returns to the physician's office and the data is downloaded to a PC.

Second is that the glucose trend plots that are displayed for the health care professional are not to be used as the sole basis for changes in therapy, but rather should be used in combination with meter glucose values and other historical information to make recommendations for changes in monitoring practices, insulin administration, diet in patients and exercise, for example.

The advent of finger stick blood glucose monitoring represented a major advance in the treatment of diabetes. Today, we will discuss why we believe continuous monitoring provides even more valuable information necessary in the management of diabetes. On this plot, glucose concentration on the vertical axis in milligrams per deciliter versus time of day for a two day period is shown.

The pink line shows the output from a person who does not have diabetes; whereas, the yellow line shows output from a sensor inserted in a person with diabetes. The blue triangle shown superimposed on the yellow line are the referenced YSI values which confirm the output from the sensor used.

There are two important points I would like to make on this slide. The first is that glucose concentration in the person without diabetes is relatively stable throughout the day, varying from approximately 70 to 120 milligram per deciliter; whereas, in the person with diabetes, glucose concentrations can vary from below 70 milligram per deciliter to up near 300 milligram per deciliter and this is in an intensively managed patient.

Second, because of the wide fluctuations in glucose in the person with diabetes, monitoring blood sugars only four times a day, for example, as illustrated with the white circles, which are obtained each day at the time of meals and then at bedtime is not sufficient to capture trends in glucose levels and, therefore, makes diabetes management unrealistic.

In summary, we believe that the ability to provide the health care professional with these trend plots will allow them to better understand the glycemic control in a patient and help physicians and the entire diabetes care team provide appropriate recommendations for modifications to therapy.

More on this subject will be discussed later by Drs. Mestman and Marcus.

The first generation CGMS product, which we are here to discuss today, is essentially like a holter style monitoring product, where the system would be set up at the physician's office and the sensor would typically be inserted there. The patient would then wear the system at home for up to three days for one sensor and longer if more than one sensor was provided to the patient.

The patient would then return to the physician's office where the data would be downloaded for review and analysis and by looking at the trend plots generated from the system, an assessment of glycemic control and the identification of patterns could be made by the diabetes care team.

Again, I want to emphasize that while wearing the product at home, no real time glucose information is displayed or available to the patient.

The CGMS consists of five components: the sensor assembly, glucose monitor, cable, communications station and test plug. The test plug is used only as a diagnostic tool to troubleshoot suspected system problems and will not be discussed further.

The PC provides means for retrospectively viewing the data from the CGMS following its use by a patient.

I will now spend some more time going through the four primary components of the system, which are the sensor assembly, the monitor, cable and communications station.

The glucose monitor, the cable attached to it and the sensor itself are the three components of the system, which are worn by the patient during use. When connected together, the system is waterproofed to prevent potential problems if the system is exposed to moisture while being worn.

The glucose sensor is inserted in the subcutaneous tissue using a rigid introducer needle, which is removed following insertion, leaving behind the sensor in a small flexible cannula. The system requires a one hour stabilization period after which time meter values can be entered into the monitor used for calibration of the sensor. The range of the sensor is 40 to 400 milligram per deciliter and each sensor can be used for up to three days. Although the implementation of the sensing method into the product is novel, the assay method we use to measure glucose is not very different than that used in a YSI glucose analyzer, for example.

The working portion of the sensor has an outer biocompatible membrane, which regulates glucose and oxygen diffusion through it to an inner glucose oxidased layer, which rests on top of a platinum electrode.

There are two reactions that take place in the measurement of glucose. Glucose and oxygen, which diffuse through the outer membrane, react at the glucose oxidase to produce gluconic acid and hydrogen peroxide. Hydrogen peroxide generated at the glucose oxidased layer can diffuse down to the platinum electrode on which we apply a .5 to .6 volts that oxidizes the peroxide and creates electrons, which we measure as a current.

The important thing to note here is that the more glucose that there is in the environment, the more peroxide is generated and the more peroxide that is generated, the more current we measure. Later, when I describe the calibration approach, basically, that is going to involve a description of how we convert the currents measured by the sensor into glucose concentration.

The glucose monitor is the size of a pager and has the ability to be connected to a belt with a clip or worn without a clip in a pocket. The monitor collects and stores the sensor current output values in memory every five minutes, with 14 days total memory capacity. The monitor memory may be cleared after use.

Meter values used for calibration or periodic monitoring can be entered into the monitor, as well as markers for meals, insulin, exercise or other events. As mentioned earlier, the monitor has the capability to download its data to a personal computer through the communications station.

Once downloaded, the PC software calculates the calibration constants, which are applied to the sensor's current output to convert these values into glucose values. The MiniMed communications station is shown in this slide with the monitor aside of it. The communications station allows the monitor data to be downloaded to the PC through infrared communication ports in the cradle of the communications station.

I would now like to briefly discuss the feasibility studies that have been conducted previous to the multi-center clinical trial and these studies collectively resulted in the study of 70 patients. These studies were conducted in controlled settings and lasted four days.

In the four day period, where we studied Type 1 patients primarily, we inserted glucose sensors into various sites of the participants and many patients had two sensors inserted simultaneously. While wearing the product, they underwent their typical meal and insulin schedules and also while wearing the product, simultaneously we collected blood glucose information obtained, using a YSI, HemoCue and various meters.

This slide shows an example of the output from one of the feasibility studies in a person with Type 1 diabetes on multiple daily injections. On this slide, we are showing glucose concentration, a milligram per deciliter, versus time of day for a three day period. The red tracing is the output from the glucose sensor and the dots on the plot represent the output from the YSI, Accu-Chek and HemoCue used in this study.

As you can see, the sensor does a good job of tracking the trends in glycemic excursion. The other interesting thing to notice in this person who was on a very regimented schedule of injections and meals is that each evening at 12:00 a.m., they had a high glucose excursion of approximately 350 milligram per deciliter and each morning prior to breakfast, they had a low excursion.

So, we believed that the use of this kind of information to identify these trends will be useful in improving diabetes management.

This slide shows representative data from a patient with two sensors operated simultaneously for a two day period to demonstrate that sensors in multiple sites behave similarly. One sensor in this patient was inserted in the abdomen, the other in the upper arm. The two sensor tracings are shown by the red and black lines and, again, as you can see, they track the glycemic excursions in the patient.

I would now like to take a moment to discuss the calibration approaches used in the multi-center clinical trial, which followed after the feasibility studies. Each of the panel members previously received a binder of all the daily trend plots obtained in the multi-center trial. These plots were generated using a 1-point calibration method to convert the sensor signal to blood glucose.

Subsequently, we evaluated a second approach, which uses a daily calibration based upon regressing the sensor currents to the meter values entered by the patient. The second binder of trend plots you received shows the daily sensor trend plots with both methods of calibration applied to illustrate the differences between the two calibration approaches.

Later, Todd Gross will present the results from the multi-center clinical trial we conducted. Before he does that, I would like to discuss the two methods of sensor calibration, which we have evaluated. In the 1-point sensor calibration scheme, the patient entered all the finger stick glucose values into the monitor while wearing the device. And in our instructions for use, it is recommended that the patient enter at least four values per day minimum.

Values that were entered as calibration values provide a real time sensor sensitivity check while the system is being worn. When calibration values are obtained to enter into the monitor, we recommend that two meter values be obtained at the same time and the average of those two values be entered into the monitor.

Calibration of the system is performed by the download software after the system has been worn by the patient and the daily calibration curves resulting from the downloaded data are based only on values entered as calibration values.

As I mentioned earlier, the glucose monitor collects and stores the five minute sensor current outputs and the meter values entered by patients as shown in this slide. So, here we see on this access we are showing the current output from the sensor and that is the red tracing. So, this is a value in nanoamperes.

On the right vertical access we are showing the glucose concentration or the meter values entered by the patient. They are represented by the black circles. The first black circle represents a meter value that was entered by the patient as a calibration value. Again, these calibration values calculate a real time sensitivity factor to verify proper sensor function while the system is being worn.

After downloading the data to the PC, the graphing software converts these sensor currents into glucose values using the calibration approach. With the 1-point calibration method, the meter value, which was entered as a calibration value by the patient, which is this first value, was used to determine the multiplier to convert sensor signals to blood glucose and the other meter values were not used for calibration at all, but were still plotted on the daily trend plots.

There are two potential risks with this approach. The first is related to potential dynamic differences between blood and interstitial glucose. If blood glucose is rising or falling rapidly at the time of calibration, the glucose concentrations in the interstitial and blood may be different. This can lead to an inaccurate calibration, which is enforced until the next calibration is performed.

In this example, notice the rapidly increasing sensor currents at the time that the calibration point was entered. The second problem is that the meter value may be inaccurate. In the multi-center clinical trial, we tried to reduce this potential problem by requiring that the patient make two sequential meter measurements and use the average as the calibration value.

But as you can see in this slide, for meter values, which were obtained within five minutes of each other in the multi-center trial, and typically they were obtained within one minute of each other, there is often a lack of reproducibility between the two measures.

Therefore, because of the two potential issues with the timing of the calibration and then the accuracy of the calibration method, use of the 1-point calibration is not ideal. For the data that I just showed you, if we apply the 1-point calibration to the sensor signals that you saw previously, this is the resulting sensor output in milligrams per deciliter shown. So, here is the calibration point, the signal from the sensor naturally lines up exactly with the calibration value and then the subsequent outputs from the sensor are determined using that calibration factor.

As you can see, the ill-timed calibration resulted in the sensor overestimating the blood glucose values. Because of that, we have looked at a regression calibration approach, which we have now incorporated into the product. The patient uses the CGMS identically to how they would and how it was described earlier with the 1-point calibration method. They still enter all of the meter values that they obtain into the monitor.

Values that are entered as calibration values continue to provide a real time sensor sensitivity check and calibration is again performed by the download software at the time of download from the patient. The difference with this new approach is that now we take advantage of all the entered meter values to generate sensor glucose values based upon a daily regression between the sensor currents and the meter glucose values.

This approach reduces the effects of an ill-timed meter entry and helps average out the error of individual meter readings. Here we show the previous 1-point calibration output as denoted with the red tracing again and the regressed output as denoted with the blue line. As you can see, the regression approach minimizes the effect of the initial calibration value and greatly improves the accuracy of the trend plot.

Now, I would like to discuss how the system will be used in practice and describe the format of the data as it is presented to the health care team. In terms of using the product, first, the health care professional would insert the sensor and secure the area with tape. Using the user interface of the monitor, they would then initialize the sensor and this starts a one hour clock.

At that point, they could elect to send the patient home or they could ask the patient to remain in the office until they perform the initial meter entry, which is used for the sensitivity check and that would occur one hour after the sensor was inserted in the body. While wearing the product at home, the patient would enter other meter values and any events as recommended by the physician.

After wearing the sensor or sensors for as long as the physician asked, the patient would then return to the physician's office to remove the sensor and download the data. At the time of downloading the data, the regression calibration approach would be applied to the signals that are stored in the monitor's memory.

The physician would then review the data plots and summary information with the patient and make recommendations for any changes in their diabetes management.

This graph shows an example of how the daily trend plot on the PC is presented to the health care professional. The glucose concentration in milligram per deciliter is shown on the vertical axis versus time and in the product we are showing daily trend plots, calendar day plots, which are from 12:00 a.m. to 12:00 a.m.

In addition, meter glucose values entered by the patient are displayed on the trend plot as would be any event, such as meals and insulin administration. And you will see some plots with those markers identified later.

Dr. Marcus will discuss how these daily trend plots will be used by a physician and the entire diabetes care team to make recommendations for changes in the patient's diabetes management. In addition to the daily trend plots, the MiniMed CGMS also provides a summary of daily statistics from the glucose sensor values, which are shown in these columns here, and any of the meter values entered by the patient into the monitor while being worn.

For each day, the number of readings, their average standard deviation, minimum and maximum values are shown. The overall statistics for all the days of use are combined and shown at the bottom of the table. This format of the data allows the health care professional to easily see minimum and maximum daily excursions, as well as the average glucose values.

In addition, the correlation coefficient between the sensor and meter values is shown, as is the mean absolute error associated between the meter values and the sensor values. This information is provided to aid the physician in assessing the reliability of the sensor's output. The sample graph I just previously showed you was from the date October 7th, shown here, and its correlation was 0.89, which is slightly lower than the median correlation we obtained in the multi-center study.

Finally, our version of a modal day plot is provided, which superimposes the sensor outputs from each day of operation onto a single daily trend plot. We believe this presentation of the data will aid in the identification of glycemic excursions at certain times of the day. For example, in this patient's chart, all of the daily profiles showed an elevated glucose level late in the day.

Before I summarize, I would like to take a moment to describe a simulated patient study we conducted at the end of last year at the FDA's request. This study evaluated the CGMS in a well-controlled in vitro environment to simulate that of a patient. Basically, sensors were placed in beakers of various glucose solution, where the sensor would reside for a period of time before being moved into a beaker of a different concentration.

The study lasted three days, which is the maximum use that the sensor would be used in practice. An Accu-Chek Advantage meter was used to calibrate the sensor as it would be used by a patient. Subsequent Accu-Chek measurements were obtained from each beaker of glucose that the sensor was moved into and these values were entered into the monitor as were markers for meals, insulin and exercise events as the product would be used by a patient.

This slide shows a typical daily trend plot from that study. As you can see, the sensor output in vitro in the beakers of constant glucose solution is exceptionally stable and its measured glucose values accurately reflect the Accu-Chek values, which were also more stable in this controlled study.

Also notice that approximately 3:30 p.m. in this daily trend plot, we purposely entered an Accu-Chek value of approximately 380 milligram per deciliter as a calibration value, even though the sensor had been moved into a solution of 100 milligram per deciliter glucose at the time. The monitor performed the real time sensitivity check, alarmed the user of a calibration error and did not provide glucose measures until the sensor was recalibrated with an acceptable glucose value, which was done at approximately 4:00 p.m., where a value of around a hundred milligram per deciliter was entered.

The results for the final three calendar days of use following the initial set up of the experiment showed an excellent agreement between the Accu-Chek and the sensor system with a bias of minus 1.7 milligram per deciliter, a mean absolute error of 4.7 percent and a correlation of .99. This is for 60 paired Accu-Chek and sensor values.

We believe that this study confirms that the sensor provides an excellent measure of glucose and that inaccuracies observed in vivo are partially due to timing differences between the interstitial and blood glucose compartments, partially due to inaccuracies in the meter and partially due to the interstitial environment where the sensor resides.

Yet, despite these inaccuracies, the sensor still provides an accurate profile of daily glycemic excursions.

In summary, we have designed the CGMS to provide the health care professional and the entire diabetes care team, including the patient with continuous glucose monitoring data, which can be used in combination with meter glucose values, meal and insulin dosing data and other information to enhance their ability to effectively manage diabetes.

It is important to note again that the CGMS is intended to supplement, not replace, finger stick measurements, that the CGMS provides retrospective graphs of glycemic excursions, facilitating therapy adjustments, that the regression calibration method, which we now have incorporated into the product is superior to the 1-point calibration approach and improves the overall reliability of the system and the feasibility studies and simulated patient studies we have performed demonstrate the utility of the CGMS to track glycemic excursions.

I would now like to introduce Dr. Jorge Mestman, who will discuss his clinical experience with the MiniMed CGMS, having been the principal investigator of several of the feasibility studies and one of the multi-center clinical trial investigators.

1   2   3   4   5   6   7   8   9   ...   16

Похожие:

Medical Devices Advisory Committee iconVeterinary medicine advisory committee

Medical Devices Advisory Committee iconRanch hand advisory committee

Medical Devices Advisory Committee iconAdvisory committee on immunization practices

Medical Devices Advisory Committee iconNational Vaccine Advisory Committee (nvac)

Medical Devices Advisory Committee iconExternal Advisory Committee on Cities and Communities

Medical Devices Advisory Committee iconWildlife Diversity Policy Advisory Committee

Medical Devices Advisory Committee iconSchedule 5 Appendix c other Medical Devices

Medical Devices Advisory Committee iconPeer reviewed by the Arizona Department of Commerce Economic Research Advisory Committee

Medical Devices Advisory Committee iconFood and drug administration national institutes of health advisory Committee on: transmissible spongiform

Medical Devices Advisory Committee iconAdvisory Committee, Cuyahoga Valley School-to-Career Consortium, Broadview Heights, Ohio 1996-2002

Разместите кнопку на своём сайте:
Библиотека


База данных защищена авторским правом ©lib.znate.ru 2014
обратиться к администрации
Библиотека
Главная страница