Jared Watkin - Freestyle Libre (Vienna, EASD 2014)

You'll see where we believe flash glucose monitoring fits into the world of diabetes care. As a company, we've had technology for both strip and meter testing and CGM for a number of years now, and we've seen the benefits, as shown by the previous speakers, of comprehensive reports that can be generated by the monitoring technologies. We've seen that ambulatory glucose profiling can create clear and actionable reports that can be understood quickly and acted upon by both physicians and their patients in an understandable manner. As we looked at the available technologies, we realized and recognized as a player in that field that the current methods for generating those comprehensive data sets required for things like AGP had limitations. Finger stick testing in reality in day-to-day life limited the number of data points that could be obtained. Overnight data obviously as well is very impractical; it's the rare patient who is willing to wake themselves up and test overnight. CGM, continuous glucose monitoring, clearly addressed that data density issue, but as we know, the penetration of CGM on a worldwide basis is relatively low. Overall latest figures would suggest less than 5% of the market is actually CGM. We talked about some of the strides that are going on there in the US; we don't see that trend elsewhere in the world. There are a number of reasons for that, but a large part of it is due to cost, lack of reimbursement in many markets, and another thing is the data interpretation challenges. This is where we introduced and developed the new category of flash glucose monitoring. The product received the CE mark for European launch in August this year, and it's called FreeStyle Libre. A little bit about the system: essentially a patient will receive a sensor kit which contains the sensor and the applicator. On the left hand side there is the reader kit, which is a reader, a wall plug, and a USB mini USB cable, used for both charging and for downloading the reader to the software management packages on either a PC or a Mac. So we designed FreeStyle Libre to try and address some of the unmet needs, to allow some liberation from the hassles of glucose monitoring. Some of the key features: normal routine finger pricks required? We're essentially saying that on a general day-to-day basis, finger stick testing is no longer required. You do not need to calibrate the system with finger sticks, unlike conventional CGM systems. We do recommend finger stick testing under a couple of conditions: one, when glucose is changing rapidly, where we know ISF and blood glucose can be somewhat different, and also when you're confirming or testing for hypoglycemia. So if you intended to make therapy changes under those conditions, we would recommend a finger stick test. The way a patient obtains the data from FreeStyle Libre, and we'll have a demo at the end of my talk, is a painless one-second scan. No calibration during that time period, and it's really designed to replace BGM use for day-to-day therapy decisions. So a little bit about what information the patient gets when they conduct a scan: they get three key pieces of information in that one-second scan. They get the current glucose reading at any one time, and the system generates glucose information once every minute, so you can scan every minute and get a fresh glucose reading. Also during that scan, you get up to eight hours of history that's held on the sensor itself, so you get the previous eight hours of glucose data collected every 15 minutes for the previous eight hours. And then finally, as also similar to CGM, you get the directional arrow, the trend arrow, which shows the direction glucose has been moving. Also, to connect it to a Mac or a PC for downloading of the information, the reader itself will hold up to 90 days of 24-hour glucose information. It has a test strip port built into it, so when on the rare occasion you need to do finger stick testing, you can do it off the same reader, and that uses the FreeStyle Precision or in other markets known as the FreeStyle Optium test strip. Built into the system itself, it also does ketone testing as well. The home button is the single button on the reader that turns the reader on to allow scanning or to go further into the submenus; everything else is activated through the backlit color touch screen through an icon-driven menu. Something about the sensor itself: obviously sensor size has been one of the key barriers to penetration of such systems. The sensor itself is also water resistant, so for daily activities including swimming, showering, and exercise, water should not present a problem to the use of the device. I'll show you some more data a little later on, but most people agree it's painless to apply and wear. As I mentioned, it automatically measures, captures, and stores glucose data day and night, so even between scans, the system is collecting and holding that data on the on-body sensor for up to eight hours. And at the end of the way, unlike conventional CGMs, this sensor is fully disposable; there is no reusable piece. At the end of the 14 days, you throw the whole piece away and you apply a fresh sensor. Some detail on the reports: whereas there are comprehensive data management reports that we have, with patterns, timing, and target in the center, you can set your targets in conjunction with your healthcare professional, and then the system will detail the time above target, the time within target, and the time below target to give you an assessment of your overall level of control. Similarly, on the right hand side, it will also record low glucose events based upon your set target for hypoglycemia, and it will show you how many episodes of hypoglycemia you've had and at what times during the day they were occurring. Those are an example of the reports that are available on the reader itself. As I mentioned, obviously you can download the reader itself to a PC or a Mac, where we have a number of reports. I'm showing you one here; this is very similar to what Ramsey showed you from the science study, but it's essentially a combination of the Joslin Center to try and provide some extra very simple to understand traffic light information, helping to interpret what was being seen in the AGP profile. In terms of a technical achievement that we believe we've met, we believe the system is clinically proven to be accurate and stable and consistent over 14 days without the need for the finger stick calibration. That entails factory calibration, and that's really enabled by our unique Wired Enzyme technology, which I'll talk you through in a minute, and I will show you how that's manifested itself in terms of the accuracy trial. So what is calibration and why is it required? When we've talked about this over the past year, there's been some questions about how we've achieved the factory calibration and avoided the need for finger stick calibrations during the sensor wear. We determine that calibration factor by doing a large amount of testing where we correlate blood results to the electrical output of lots of sensors that we produce. Typically in the past, the reason why sensors have had to be finger stick calibrated during the wear period is that their calibration factor has changed during the wear period. So when you do a finger stick, you essentially reset the calibration to make sure that the system provides the right answer and relates that to the electrical current. We do not need to do that with the chemistry we have. We enable it through our unique Wired Enzyme chemistry, and this allows the sensor to be very stable during the wear period. The other thing that has to be done in order to make it feasible from a factory calibration perspective is that the chemistry allows a 14-day no-cal sensor. The Wired Enzyme is actually the core technology of the original FreeStyle Navigator product. A couple of key features: it doesn't require oxygen to provide the results, even though it uses glucose oxidase. The normal natural electron acceptor for the glucose oxidase reaction is oxygen; in this case, we have an osmium mediator that becomes the electron acceptor, and that's a key part of the overall chemistry. The sensor obviously as well operates at a very low potential, which helps to minimize things like interference and keeps the signal-to-noise ratio at an appropriate level. So there's a little animation here to try and explain what we mean by Wired Enzyme technology. Really, in the end, it means that the enzyme plus the mediator are actually wired via that polymer network. What this really does is it means during the wear period, you do not get any loss of activity, so you do not get a reduction in signal response through the 14 days of wear. You can see the glucose binding into the glucose oxidase there, and whereas normally the electrons would be used to pass to oxygen to produce hydrogen peroxide that would then be oxidized, in our case, you see the electrons are actually passed to the osmium mediator through the polymer network chain, transferred to the electrode surface where we measure it as an electrical output that stays extremely stable through 14 days. I mentioned as well the need to produce the sensors very precisely, very accurately, with very minimal variability within a lot. What you can see here on the left hand side is the equipment we use to calibrate each lot, that's used to apply a lot-specific calibration factor that obviously stays consistent through the shelf life and through the 14-day wear period when a patient applies it. What that really means is signal stability over a 14-day period. This is a paper published by my colleague; we see no trend, no systemic change in response once in the body over the 14 days, and that's what's required in conjunction with very precise production to allow a factory calibrated sensor. So how does that manifest itself in terms of performance? Obviously as part of the CE mark process, we have to generate an accuracy data set to support the approval of the product. This is the trial that we did: there were 75 subjects, comparing against a blood glucose system. That blood glucose system has been used to calibrate the sensor, but there's no self-correction with this; the blood glucose response has got nothing to do with the sensor. The sensors are prospectively calibrated, so the comparison is through 14 days of sensor wear. These studies generate a lot of data; this ends up with over 13,000 paired data points for comparison with performance to blood glucose monitoring. This is the information that we obtained; this is what will appear in the labeling of the device. So for 14 days with no finger stick calibrations, we achieved a MARD, mean absolute relative difference, of 11.4%. I'll put that in context of other sensor performance in the next slide. The distribution in terms of color: the brighter the colors, the more data points are in that section. So what does 11.4% MARD mean? This is the plot of performance of MARD over the 14 days. You can see that other than day one, which is slightly worse, which is a fairly typical phenomenon when you're using implanted sensors, for the rest of that time period up to 14 days, there is no change in accuracy during that period of wear, again without any finger stick calibrations during that time period. So you can see that the stable chemistry, the precise manufacturing processes, results in leading edge performance in terms of the accuracy of the system. Let's compare it; there's no head-to-heads available yet. Perhaps the only fair way of comparison is to look at other products' own pivotal trials. Our own FreeStyle Navigator 2 in its own pivotal trial achieved 12.3% versus blood glucose; that product requires five calibrations over 5 days, so during that 14-day wear, it would have required 15 finger sticks to achieve that performance. For Dexcom G4 Platinum product, 14% versus BG is what's in the labeling, with a need for two finger sticks per day, so over that period, 28 finger sticks would have had to be carried out to calibrate the product. And Medtronic Enlite, 14.1%; in this case it's against YSI, there's no published data against BG, with a four per day calibration scheme that would have resulted in 56 calibration finger sticks done over the 14-day wear period. Now it's worth pointing out that FreeStyle Libre really gives leading edge performance when compared to other sensors in the field. So a little bit about our user experience: we've executed some user experience studies, and as we said, we were trying to design the product to reduce the hassles of glucose monitoring for patients with diabetes. We carried out a user experience study: 93% of people that tried the device agreed that FreeStyle Libre was comfortable to wear. 83% agreed that it was actually painless to apply the sensor, and 100% of people said it was either painless or almost painless to apply the sensor. Obviously the ease of application and the discomfort felt is a key barrier to the use of other sensors. 96% agreed that the FreeStyle Libre system would reduce the hassles of glucose monitoring for them. So a little bit about our next series of plans. Clearly in these days of constrained healthcare costs, we have to provide evidence that a product such as Libre delivers a benefit in terms of health to patients at a cost position that is affordable by an increasingly constrained healthcare system. We're running two randomized control trials to generate outcome data to support reimbursement positions in Europe and the rest of the world. The first is entitled REPLACE; it's a randomized control trial evaluating the impact of this Libre sensing technology on standard of care with SMBG, and there is a six-month extension for the device intervention group as well. It's more than 210 subjects with type 2 diabetes, all of which are on multiple daily injection therapy, and the subjects have an HbA1c of 7.5% or above going into the study. The study details are on clinicaltrials.gov, and recruitment for this study is in progress as we speak. Obviously a very important study to demonstrate the benefit of the product when using type 2 diabetes. The second study is called IMPACT, and this again is a randomized control trial, and this time we're evaluating the impact of the Libre technology on hypoglycemia in type 1 diabetes. You can probably see the parallels to the studies here to what some of the other companies have done. The Libre product again compared to a control group using SMBG for six months. In this case, there's a total of 26 sites across the Netherlands, Germany, Spain, Austria, and Sweden, and there will be more than 225 subjects in this study, all of which have type 1 diabetes and are either on MDI or pump therapy, and these subjects will come into the study with A1Cs of less than 7.5%. Again, this trial is on clinicaltrials.gov and recruitment is in progress. So two, certainly for this space, large randomized control trials to show the impact of use of FreeStyle Libre on key outcomes in patients with both type 1 and type 2 diabetes. So in summary, we believe the current methods that are widely used for glucose monitoring have limitations. FreeStyle Libre addresses the need to collect the glucose data. It's a 14-day sensor wear, a fully disposable low profile sensor that is factory calibrated, so no finger stick calibrations are required during the 14-day use. Accurate, stable, and consistent performance over 14 days without finger stick calibrations: 11.4% MARD, and 99.7% of all data within Zone A and Zone B of the consensus error grid. It will produce, amongst other things, AGP-based reports to allow the visualization of the comprehensive data that the product collects. Ongoing randomized control trials to focus on reimbursement through improved outcomes when people with type 1 or type 2 diabetes use the product. And obviously that's a very important step. I'll hand the product over to you, Jenny, and that's it.