Memorandum

Positive 8-0 Vote for Senseonics Eversense Implantable CGM at FDA Advisory Committee; 90-day wear, adjunctive label – March 30, 2018

Executive Highlights

  • An FDA Advisory Committee voted unanimously in favor of approving Senseonics’s Eversense 90-day implantable CGM: 8-0 for overall benefit:risk, 8-0 for safety, and 8-0 for effectiveness. This is a great victory for expanding CGM options in the US, and we expect the FDA to rule favorably given the discussion and minimal debate.

  • Senseonics’s CEO Tim Goodnow carried the day, masterfully answering scientific and clinical questions and navigating the different levels of conversation brilliantly – this team was very well prepared and mixed confidence, data, and respect for the process (even when panelists asked surprisingly bizarre questions). Drs. Steve Russell (MGH) and Jeremy Pettus (UCSD) gave insightful, impassioned comments on unmet need for CGM.

  • Critical drivers of the positive vote, in our view, included: Senseonics outstanding accuracy (MARD of 8.5% and 87% of readings within 15 mg/dl or 15%) and conservative adjunctive labeling with two fingerstick calibrations per day. It was notable to see the comparison to the three other available CGMs in the US – Eversense stacks up well, and the data suggests (to us) that it meets the bar for non-adjunctive use. The application is for 18+ years and smartphone-only display (no receiver). There is no acetaminophen interference (a big win), though proposed contraindications/warnings for mannitol, sorbitol, and tetracyclines (which affect the fluorescent chemistry – we don’t think this is a big deal).

  • There was notably strong advocacy in the open public hearing, with 18 of the 19 people in the open public hearing speaking in favor of approval – this included Dr. Ed Damiano, Jeff Hitchcock, AADE’s Lisa Laird, Dr Mark Christianson, Dr. Dorothee Deiss, Dr. Tim Bailey, Dr. Katherine Barnard, JDRF’s Daniel Finan, AACE’s Sethu Reddy, our own Adam Brown (see his slides and talking points) and Maeve Serino (see her slides and talking points), and multiple trial participants and several current EU users – some were on their seventh sensors (each lasts 180 days)! Several trial participants said they cried upon returning the system.  

  • We heard plenty of enthusiasm for Eversense’s ability to expand CGM options, its strong quality of life benefits relative to current CGM, minimizing insertions and the number of “day #1’s” to once every 90 days (relative to one per 7-14 days with current CGM), slim removable transmitter (without pulling out a sensor), excellent adhesive, on-body vibration alerts, and the minimally invasive <5-minute insertion procedure. On the latter, Drs. Steve Russell and Tim Bailey commented that insertion is easy and fun – they actually look forward to doing the procedure! This bodes well for Senseonics, since scaling will rely on getting HCPs trained and bought in.

  • For now, the points of discussion will clearly center on labeling and the specifics of the post-approval study and/or US registry. The PMA has been under review for 17 months at this point (partly due to two pivotal trials), following submission in October 2016. We can’t imagine there are many more steps to get over the finish line, and the panel did not conclude more studies are needed. Assuming the remaining discussions go quickly, perhaps Eversense will launch at ADA. Kudos to the Senseonics team for driving this very positive Ad Comm vote for the first implantable, longer-term-wear CGM!

As we expected following the briefing documents, an FDA Advisory Committee vote unanimously in favor of approving Senseonics’ Eversense 90-day implantable CGM, on body transmitter, and mobile app, responding positively to overall benefit:risk (8 YES - 0 NO), safety (8 - 0), and effectiveness (8-0).

For those that would like them, we’ve scanned the slides at the following links:

The Senseonics team came out very strong and were impressively well-prepared with slides in response to nearly every conceivable panel question – CEO Tim Goodnow knows this technology inside and out. Eversense’s accuracy, of course, speaks for itself, and given the conservative adjunctive labeling, panelists were often trying to find anything to catch the company on – discussion often drifted for minutes at a time onto a variety of “What-If” scenarios, like MRIs or data on people with fragile skin and complications.

An enthusiastic Open Public Hearing provided tailwind to the discussion, with 18 out of 19 speakers strongly supporting approval – including several renowned clinicians, patient advocates, Eversense trial participants, and even a handful of Eversense users that flew in from Europe. We hadn’t appreciated – as summarized below – the quality of life benefits of Eversense over other CGMs. It will be fascinating to see how Eversense competes on the US market.

There is always a risk at FDA Advisory Committees that panels want to see “more data” before approval, but the dialogue was very tilted towards – “This shouldn’t hold up approval, but it might be nice to look into [pregnancy, ICU/hospital, people with complications, etc.].” We were also concerned the panel might focus too much on the four design changes implemented by Senseonics after the US pivotal study, but they instead applauded Senseonics (rightly) for continuously improving the product. As Dr. Goodnow noted: “This is a first-generation product. It’s been two years since we did this clinical study, and we’ve learned a lot.” As a first-gen product, it’s impressive to see how far Senseonics has pushed the accuracy and form factor – the first-gen Dexcom or Medtronic sensors had MARDs far exceeding 20% and were years away from being on the phone.

We fully expect the FDA to approve Eversense and look forward to seeing another accurate sensor with a new form factor in the US – likely this year assuming things go well in labeling discussions. We believe this will propel CGM as a “class” forward and will move forward CGM as standard of care.

Below, you’ll find thematic highlights from the panel meeting, including a detailed summary of the compelling open public hearing arguments.

Voting Summary

The panel voted unanimously in favor of approving Senseonics’ Eversense 90-day implantable CGM, on body rechargeable transmitter, and mobile app – a positive 8-0 vote for overall benefit:risk, 8-0 for safety, and 8-0 for effectiveness. As noted in our detailed description of the Advisory Committee Briefing Documents, this was the sort of vote we expected, given  stronger-than usual panelist experience in diabetes, strong accuracy data, and few areas of major FDA uncertainty in the submission. Oddly, panelists were not afforded the opportunity to provide commentary on their voting decision as in past advisory committee meetings, so the rest of this report integrates their insights from throughout the day – we were disappointed by this. Overall, panelists found very little to criticize in the device, though there were still plenty of moments of head scratching where the panel seemed to be asleep – hours into the day, one asked whether it was for an adjunctive or non-adjunctive indication (!). In another example, panelist Dr. Avery Tung asked how CGM changes diabetes management – a reminder that the panel did include members with expertise far outside of diabetes. Dr. Steve Russell did an excellent job of explaining the value of CGM, as well as what Senseonics adds to the equation. Overall, we were relieved the panel made the right call here, as that is never a certainty with diverse FDA panels.

  • The table below summarizes the votes (all positive), some of the topics/questions/concerns that panelists brought up repeatedly, and how they voted at the Dexcom G5 Ad Comm (if present).

Name

Eversense Vote

G5 Non-Adjunctive Vote (If Present)

(Safety, Effectiveness, Risk:Benefit)

Specific Concerns, Questions, Insights, and Our Observations

Andrew Bremer, MD, PhD

Program Direction, Division of Diabetes, Endocrinology, and Metabolic Diseases, National Institutes of Health

Eversense: YES, YES, YES

G5: YES, YES, YES

- Outstanding panel chair, kept discussion on point, on time, and summarized concerns in a balanced, productive way

- Significant diabetes technology expertise, given work at NIH

- As Chair, was mostly quiet on opinions and questions for FDA or Senseonics

- Definite positive for any future diabetes tech panels

George Grunberger, MD

Medical Director, Grunberger Diabetes Institute

YES, YES, YES

YES, YES, YES

- Terrific champion for CGM and more options, as a practicing endocrinologist

- Positive, real-world clinical perspective on the four major design changes

- Focus on post-market surveillance, including “Clinician Satisfaction scale” with Eversense

- Rightly pointed out BGMs can be inaccurate, which also pose risks to non-CGM users

- Definite positive for any future diabetes tech panels

Marc Rendell, MD

Creighton Diabetes Center, Creighton University School of Medicine

YES, YES, YES

 

YES, YES, YES

- Questions focused on risks/safety of device and insertion procedure

- Asked how Eversense compares to other implanted devices in other therapeutic areas (Quite favorably, with low infection rates)

Katheen Wyne, MD

Division of Endocrinology, Diabetes, The Ohio State University Wexner Medical Center

YES, YES, YES

 

YES, YES, YES

- Practicing endocrinologist, definitely motivated to increase CGM penetration – but also cognizant that barriers have as much to do with clinicians as patients

- Brought up concerns with Senseonics’ Eversense for slender type 1s, particularly with repeated insertions, scar tissue, and insulin injections

- Positive, real-world clinical perspective on the four major design changes

- Likely positive for any future diabetes tech panels

Avery Tung, MD

Professor of Anesthesia & Critical Care, Quality Chief for Anesthesia, University of Chicago

YES, YES, YES

 

N/A

- Critical care experience, so most questions focused on esoteric hospital issues (heparin, systemic inflammation, bypass)

- Little knowledge of diabetes (e.g., “How Does CGM change diabetes management?”)

Robert Burr, MD

Endocrine Center of Cape Cod, Falmout Hospital

YES, YES, YES

 

YES, YES, YES

- Practicing endocrinologist, focused on older population

- Acknowledges real-world issues – e.g., patients will use this non-adjunctively anyways

- Wondered about other skin types and whether they are people for whom product is not suited

- Big fan of post-marketing data, very persuaded by EU experience on MRIs

Walter Kraft, MD

Director, Clinical Research Unit and Division of Clinical Pharmacology, Thomas Jefferson University

YES, YES, YES

 

N/A

- Excellent comments on dexamethasone data – low risk, does not believe it needs to be focus of post-market

Edward Gregg, PhD

Chief of the Epidemiology and Statistics, CDC

YES, YES, YES

 

N/A

- Works in diabetes epidemiology, so brings a big-picture/population perspective – definite fan of registries

- Focused a lot on low alert accuracy and hypoglycemia; didn’t reference the briefing documents much

Stephen Clement, MD

Inova Fairfax Hospital

Not present

 

N/A

This was a real loss thatDr. Clement could not make it – he’s a very respected voice

Barbara Goldsmith, PhD

Chair and Professor, Medical Laboratory Sciences & Biotechnology, Thomas Jefferson University

YES, YES, YES

 

N/A

- Barely talked all day, minimal diabetes tech experience

Gabriella Lakos, MD

Senior Associate Medical Director, Abbott Laboratories

N/A – Industry Rep

 

N/A

- Supported use of modeling with new software algorithm

- Nicely redirected off-the-rails discussion back to intended use, rather than “what if”

Carolyn Petersen, MS, MBI

Senior Editor of mayoclinic.org

N/A – Consumer Rep

 

N/A

- Wondered if decision aids might help find patients who have the appropriate motivation for Eversense

Anna McColister-Slipp

Chief Marketing and Communications Officer, Galileo Analytics

N/A - Patient representative

 

N/A - Patient representative

- As a T1D and CGM wearer, raised multiple questions on how CGM accuracy is measured

- Very vocal panelist

- Focuses discussion and questions on data in people with complications

 

Courtney Lias, PhD

FDA, Division Director

N/A – FDA

N/A

- Mostly quiet throughout the day, but as always, insightful and balanced questions

- Frequently redirected discussion back to intended use and current submission when it drifted towards 180-day wear, etc.

- Noted that CGM submissions stand on their own, but comparisons to other systems can be directionally useful and provide context

- Wondered what kind of post-approval information should be collected, especially related to insertion/removal

- Her comments and questions suggested (to us) an approval is quite likely

- Dr. Lias is an incredible asset for the diabetes ecosystem – she is an incredible leader and has very high standards and helps move things through FDA very expertly

Patricio Garcia, MPH

FDA, Designated Federal Officer

N/A – FDA

N/A

-

Open Public Hearing (OPH)

We were elated to see fantastic turnout for the FDA open public hearing – of the 19 speakers, 18 of presented compelling, passionate arguments in favor of an Eversense approval. Those in support of Eversense approval included representatives from organizations such as AACE, AADE, JDRF, CWD, Close Concerns and the diaTribe Foundation -the latter two were well-represented, with our own Adam Brown who spoke eloquently on behalf of the diaTribe Foundation (view his remarks/slides here) and Maeve Serino representing Close Concerns (her remarks/slides are posted here) very impressively in the view of her team! The one dissenting opinion came from Dr. Danielle Shapiro, representing the National Center for Health Research. While Dr. Shapiro acknowledged the unique features Eversense provides, she was not convinced that the benefits outweighed the risks – she felt the data on the four minor design changes was insufficient for demonstrating the safety and efficacy. She expressed particular concern regarding the updated sensor algorithm, which showed poor fluency with CGM and the precedent for doing such a post-hoc analysis to improve the product. She also bizarrely argued that there is “no urgent need for the device,” given other systems on the market. Thankfully, the remaining 18 speakers expressed valuable opinions that were quite the opposite – getting more patients on CGM is perhaps one of the most urgent priorities in diabetes to date and we strongly believe in Eversense’s potential to help expand penetration. As Adam said in his remarks, “CGM is the most important tool to manage diabetes … As someone who’s benefited enormously from CGM, I’m on a mission to make sure people have access.” See the below table summarizing the positive arguments.

OPH Arguments in Favor of Approval

  • More CGM options are needed to expand penetration: CGM is the most important advance in diabetes in recent years, and yet T1D Exchange data shows that ~3/4 of the US type 1 diabetes population in the US are not using it.

  • Benefits of CGM: real-time information instead of educated guesses with limited fingerstick data, improves A1c and time-in-range, reduces hypoglycemia.

  • Minimizes number of day #1’s and sensor insertions – four times per year vs. 36-52 times per year with current CGMs.

  • Eversense’s milder adhesive seems to reduce skin irritation, a source of frustration for some current CGM users and a barrier to uptake. (As a side note, Senseonics did not play this up too much. Dr. Russell also reminded us that adhesive is not even required to keep the transmitter in place over the sensor – a wearable sleeve would suffice, a definite win for those irritated by adhesive.)

  • Eversense transmitter can be removed without taking out the sensor – beneficial for special occasions and allows for increased lifestyle flexibility. Patients also don’t have to worry about accidentally tearing out their sensor/transmitter and having to reinsert.

  • The 90-day sensor wear “pre-commits” the user to wearing CGM – a great behavioral economics point! With traditional CGM, each 7-10-day sensor removal serves as an opportunity to not insert another sensor. Could it reduce the number of sensor vacations?

  • The haptic on-transmitter vibration alarms can notify users of glycemic excursions during situations in which they may not notice visual or auditory alerts on their phone. These alerts are particularly valuable to those who are visually impaired or have difficulty hearing.

  • 90-day implant is one less thing to worry about, potentially improving psychosocial outcomes and reducing diabetes distress.

  • Visual- or dexterity-impaired people may especially benefit from the longer sensor wear, as they do not need to worry about repeated weekly insertions – only two calibrations/day.

  • Safety and peace of mind: the highly-respected Dr. Katharine Barnard cited a study showing 73% of the trial participants reported feeling more secure.

  • Discretion of viewing glucose readings from a mobile device or smartwatch, without having to carry an additional handheld device.

  • Ability to see historical CGM data and patterns right within the Eversense app.

  • Insertion/removal procedure is quick and painless, leaving only a small cut. Providers find they can integrate the procedure easily into their daily routine. Dr. Dorothee Deiss anecdotally noted that she has not seen any serious complications with Eversense, including during “several emergency MRI.”

  • Senseonics might help reduce cost relative to SMBG although more cost-effectiveness studies are needed.

  • OPH patient advocates included: Adam Brown (Close Concerns, the diaTribe Foundation, T1D), Jeff Hitchcock (CWD, father of T1D), John Pettingale (T1D), Lisa Powell (T1D, read by Dr. Steve Edelman), Tobias Schultz (T1D), and Julia Vider (T1D). Patients were very impressed with the transmitter’s on-body haptic vibration alarms, ease of use, longer sensor wear, and discretion. Several commented on how prior issues with adhesives and skin irritation prevented them from using CGM. Many cited the easy and painless insertion/removal procedure, as well their relief not to carry an additional reader. We were also impressed by some of the extreme athletes who spoke – one patient has run 90 marathons, including two on Eversense! This particular patient noted that the alarm functions helped him feel safer while training and that being able to monitor his glucose levels via a quick glance at his watch is particularly convenient. See below for some of our favorite quotes:

    • “I cried all over the engineers the day I had to give [Eversense] back. I’m grateful for Eversense for giving people with type 1 diabetes the power to make informed decisions. I would not hesitate to use it.” – Lisa Powell (read by Dr. Steve Edelman) 

    • “With traditional CGM I have to set aside time each week to insert the sensor. It’s a pretty big hurdle and this means I take weeks off and it has a negative impact on my numbers. With the Eversense, I pre-committed to having CGM and it’s always waiting. Eversense completely removes need to insert.” – John Pettingale [Editor’s note – this was a particularly strong eompetitive advantage historically – other manufacturers at this stage have made a lot of progress on insertions]

    • “The best thing my family and friends can say is, “Toby, with that [Eversense] system no one realizes anymore that you have diabetes. Eversense has given me the highest level of security, comfort, and freedom both in real life and in competitive sport.” – Toby Schultz

    • “There are 42 factors that affect blood sugar – no one can manage this on 2-4 glucose data points every day. CGM is the most important tool to manage diabetes. This approval is such a no brainer. It’s a conservative application for adjunctive labeling; the device is good enough for non-adjunctive labeling based on what the FDA has previously approved. Let’s think about expanding the toolbox because we need so many more CGM options on the market.” – Adam Brown

    • “I have so much higher quality of life – the Eversense has given me my life back. I can do what I want to do. I can perform when I do sports, or if I’m at work.” – Josefine, EU user on her seventh sensor (equivalent to about 3.5 years).

    • “I have tested and used many devices in my 18 years with diabetes. This is by far the best device I’ve ever experienced. It was removed last Thanksgiving, and I still miss it.” – Randy, Eversense Trial participant (read by JDRF’s Bennie Johnson)

    • “For some people with type 1, today’s CGM devices are a struggle. Many would benefit from a CGM that is implanted. All deserve to hear their body’s symphony.” – Jeff Hitchcock

    • “Eversense provides more CGM options. It reduces the frequency of sensor insertions, eliminates the receiver as a required device to be purchased, and reduces common frustrations around adhesives and site irritation.” – Maeve Serino

    •  “We have more options in home goods and tech, but drastically fewer options when it comes to our constant daily companion of diabetes. Please give diabetics like me options.” – John Pettingale

  • Researchers and clinicians in the OPH included: Beta Bionics’ Dr. Ed Damiano, Dr. Mark Christiansen, Dr. Dorothee Deiss, Dr. Tim Bailey, Dr. Katherine Barnard, JDRF’s Dr. Daniel Finan, AACE’s Dr. Sethu Reddy, and AADE’s Lisa Laird. They provided compelling arguments demonstrating the benefits of CGM and elaborated on how providing more options could meaningfully expand penetration. Dr. Finan reminded the panel that less than one-third of adults with type 1 diabetes in the US meet recommended A1c targets. With rates of severe hypoglycemia and DKA “unacceptably high,” CGM adoption is more important than ever – and not moving as fast as it could. Regarding Eversense, providers emphasized that the insertion/removal procedures were easy to learn and implement. We were also pleased to hear a focus on the psychosocial benefits of Eversense, as many commented on how Eversense helps patients feel safer and improve quality of life. See below for some of our favorite quotes:

    • “New CGM technologies like Dexcom, Abbott, and Senseonics are raising the bar higher. Senseonics is a bold alternative CGM therapy. This is a categorically different choice, one that has been long overdue. People are individuals; it is essential that diversity [in CGM] be encouraged and nurtured.” – Dr. Ed Damiano

    • “It’s always difficult as a researcher to take back a device that participants want to keep. There’s an urgent need for reliable and discrete CGM devices in terms of glycemic control and psychosocial burden, and I can’t stress the quality of life improvements with this device enough.” – Dr. Katherine Barnard

    • I’ve been lucky to provide Eversense to my patients, lucky because a lot have had serious skin reactions or allergies to adhesives which forced them to stop wearing CGM. No one had any skin reactions with Eversense and some are using their 7th sensor. I performed 120 insertions and 60 removals. It’s a painless, quick procedure which I can integrate into my daily routine. I have not seen any serious complications even during several emergency MRI.” – Dr. Dorothee Deiss

    • “As an endocrinologist, I quickly learned how to insert Eversense. It takes approximately five minutes, including time of anesthesia. I have seen and used the new insertion device, and the removal takes less than 10 minutes in almost all cases. The small incision is quite acceptable to all the patients I’ve worked. My patient experiences have been positive, and many are veterans of CGM. The accuracy is comparable to devices that are already approved.” – Dr. Mark Christiansen

    • “I’m an endocrinologist of 33 years and I’ve guessed at diabetes management for most of that time period. CGM allows evidence-based, proactive care. Imagine driving at night with the headlights on every couple of hours. That’s the reality for many. Having the headlights turned on is more judicious and safe always. More options for technology in a patient-centric world will improve management.” – Dr. Sethu Reddy

    • “I’ve been wearing the Eversense for more than one year and I can confirm the experiences of my patients. It’s highly beneficial, effective, and safe to use. I’d even consider using it in my pediatric patients.” – Dr. Dorothee Deiss

    • “AADE believes that Eversense has proven safe and effective, and it will greatly assist in adoption of CGM. Self-insertion and skin sensitivity are eliminated with Eversense. Statistics show that only a small minority of people with diabetes choose to wear CGM. We urge a positive review for the Eversense CGM.” – Lisa Laird

Labeling, Accuracy, Comparison to Other CGMs, and Interference

  • Senseonics has proposed an adjunctive indication for use (as a complement to BGM) in adults >18 years. Similar to other adjunctive CGMs, Eversense would be approved for tracking and trending glucose values and providing alerts for high/low glucose events. Users will still need to take fingersticks for treatment decisions; panelists like Dr. Robert Burr pointed out that non-adjunctive use will likely happen in real-world use (similar to Dexcom’s G5, prior to its approval), and it will be interesting to see how Senseonics and the FDA navigate this (if at all). Presumably Senseonics could resubmit for non-adjunctive following approval, or perhaps go for the iCGM pathway (510(k)) to follow Dexcom’s G6. The age indication matches Some other CGMs are 2+ years, can be used for treatment decisions. Two calibrations per day, no receiver.

  • Senseonics’ primary US clinical studies demonstrated very strong accuracy over 90 days (2 fingersticks/day), with 85%-87% of sensor readings within 15 mg/dl or 15% of YSI reference. MARD was a very strong 8.5% in the US pivotal (PRECISE II, n=90) using the improved new algorithm, and it was 8.8% with the original algorithm. Interestingly, MARD was not reported in the slides for the follow-on 90-day PRECISION study (n=35) to gather more early sensor wear data – this clearly reflects FDA’s focus on the 15/15 threshold for CGM accuracy. (PRECISION had a 9.6% MARD< as summarized in Sponsor Executive Summary – likely reflecting more in-clinic days in the early-wear period.)

  • Eversense has accuracy in line with Dexcom’s G5, Abbott’s FreeStyle Libre, and Medtronic’s Guardian Sensor 3 on the 15 mg/dl / 15% threshold (per labels). Of course, these studies were not head-to-head, so data might look different in a head-to-head study; however, directionally, the data suggest to us that a non-adjunctive label (like G5 and FreeStyle Libre) is possible.

  • Consistent with other CGMs, Eversense does have lower accuracy on day #1 – 77%-79% of points within 15 mg/dl or 15%. Still, given that there is only a single day #1 every 90 days with Eversense, this is not a clinical concern in our view. In fact, it’s actually a huge advantage for Senseonics in our view though less of a big deal as all sensor accuracy improves. This topic of day #1 inaccuracy did dominate panel discussion, which we felt was absurd (!) given Eversense’s comparable day #1 performance, minimization of such days, and adjunctive labeling. As noted in the product feature learning section below, Eversense has a 24-hour warmup after insertion (no data), after which four calibrations are required on day #1; on days 2-90, users must enter two fingersticks per day with a slight grace period between them. The calibrations will be seen as a relative weakness now that two systems (Abbott Libre and Dexcom G6) have no calibration required – the G6 will require a calibration code).

  • It was widely agreed that the accuracy data collected in the early wear period was sufficient to demonstrate accuracy. This was an FDA concern in the briefing documents, since Eversense’s 90-day wear means fewer in-clinic visits in proportion to overall wear time. However, given the alignment with home SMBG data and consistent sensor performance over time, this did not receive much actual discussion.

  • Will Senseonics pursue the FDA’s new integrated CGM (iCGM) 510(k)-clearance pathway, as established via Dexcom’s G6 clearance? It was clear from our discussion with Dr. Courtney Lias that pivoting to this path would require a new FDA submission from Senseonics, so we imagine this would come following PMA approval (if at all). We believe Eversense’s accuracy data will meet many of the performance standards established in FDA’s Special Controls, but we’re not positive if it will meet the hypoglycemia standard. We’ll have more analysis on this front soon comparing all the CGM systems.

  • Eversense also looks strong on alerts for hypoglycemia (<70 mg/dl) and hyperglycemia (>180 mg/dl): 95%+ detection rates with ~7%-16% false alert rates. The false alert rate for hypoglycemia was arguably high at 16% in PRECISE II, but dropped to 8% in PRECISION. These data are strong and complement the accuracy shown above.

Alert Setting

PRECISE II study

PRECISION study

Detection Rate

False Alert Rate

Detection Rate

False Alert Rate

Low glucose at 70 mg/dl*

96%

16%

95%

8%

High glucose at 180 mg/dl*

98%

7%

99%

7%

*Using updated algorithm; Includes threshold and 10-minute predictive alerts

  • Sensor longevity looks good too – 91% of sensors in PRECISE lasted to 90 days, rising to 100% of sensors in PRECISION. Given this data and the sensor’s ability to actually go to 180 days, this was not a point of concern or discussion.

  • Regarding sensor interferents, Eversense is not impacted by acetaminophen – a clear positive. However, the fluorescence chemistry is impacted by mannitol and sorbitol (IV, dialysis) and the class of antibiotics known as tetracyclines. Senseonics has proposed a contraindication for mannitol and sorbitol, plus a warning for tetracyclines. The latter dominated all of the interferent discussion, and panelists recommended clear instructions for those taking tetracyclines – when can sensor values be trusted again? We assume these specifics will be carefully addressed during FDA discussions to come. Overall, however, these interfering substances should have minimal-to-no impact on uptake of the system – blocking acetaminophen is obviously the big win.

  • Concerns over MRI scans with the implanted Eversense dominated 15+ minutes of discussion. The issue was ultimately put to rest by Senseonics’ CMO Dr. Lynne Kelley – EU experience has been limited, but thus far, “it’s benign.” Senseonics has taken a conservative approach on MRI scans, recommending that patients have the sensor removed before a scan occurs. Patient rep Anna McCollister-Slipp pointed out the inconvenience of this recommendation, wondering if it is really needed and what might happen if an emergency removal wasn’t possible. Dr. Robert Burr pressed too, asking about post-marketing MRI experience in Europe. Dr. Kelley revealed that ~5-10 Eversense patients in the EU have successfully undergone MRIs with the sensor still implanted, with no safety issues. The sensors worked well post-MRI, a very good sign. Presumably this will still be discouraged in the initial label, though we imagine Senseonics will submit data to this effect – especially once the indication expands to 180 days, where removal/reinsertion will happen only twice per year.

Design Changes and New Learning on Eversense Features

  • The four incremental design changes since the pivotal study were discussed, but across the board, these improved the product experience and were not a source of concern – the much smaller gen 2 transmitter (big improvement), an improved software algorithm for hypoglycemia and early-post-insertion accuracy, an improved insertion/removal tool (to prevent accidental deep insertions), and a flatter sensor end cap (to improve removal safety). Endocrinologists Drs. Wyne and Grunberger shared enthusiasm for these updates, emphasizing that these changes are important for safety and improve the product. The panel was also reassured by the new algorithm, which seems incrementally better but was not formally tested in a clinical study – only via post-hoc modeling with the raw sensor data. However, this does have precedent with Dexcom’s Software 505 (G4AP) – we see low risk here and expect the ultimate label will indeed use accuracy data from the improved algorithm.

    • “This is a first-generation product. It’s been two years since we did this clinical study. We’ve learned a lot.” – CEO Tim Goodnow. Indeed, these four changes reflect Senseonics’ continuous product improvement, something the panel appreciated.

  • We did learn a few new details about Eversense’s fingerstick calibration scheme and requirements. We’d guess factory calibration is still more than a year away at this stage, though one calibration per day seems nearer term. (Neither was discussed at the panel, since the PMA submission is for two calibrations/day.)

    • Eversense has a 24-hour silent period after insertion, during which no data is displayed. Day #1, therefore, is actually from 24-48 hours post-procedure.

    • Four fingerstick calibrations are performed on day #1, after which two per day are required. Given that day #1 happens only four times per year (with a 90-day sensor), this is not a big barrier in our view.

    • The two daily fingerstick calibrations must be entered within a window of 10-16 hours of each other. If the device is not calibrated in 16 hours, real-time glucose results will stop appearing. “We are pretty onerous,” said CEO Tim Goodnow. We are glad to see this conservatism, and would note that Dexcom’s G4/G5 will continue to show real-time data even if a calibration is missed, which makes it easier for a patient to skip calibration.

    • The 90-day sensor actually has a hard shutoff at 97 days, allowing for a seven-day grace period following day 90. After day 97, no data will be displayed – again, great conservatism for safety and a positive for Senseonics’s business model (no sensor restarts). The Eversense system will send reminders to schedule a removal/reinsertion appointment along the way.

  • Eversense does have a few cool safety sensor features we learned about:

    • If the sensor signal drops below a certain threshold, it turns off and will not display data. “We would prefer to give no result than to give a wrong result,” said CEO Tim Goodnow. This seems similar in theory to Medtronic’s Guardian Sensor 3, which does include sensor diagnostics in the transmitter to monitor sensor health.

    • Eversense’s fluorescence chemistry is sensitive to extreme temperatures, and therefore, it includes temperature sensors. It operates within a range 26-40 degrees Celsius (79-104 degrees Fahrenheit) – given that the transmitter is on the skin and the implant is in the body, it is rare that temperatures exceed this range. CEO Tim Goodnow said that the reliability of the system is ~96%, and more than half of that (3.5%) is “typically user initiated” – e.g., out of compliance with the calibration scheme, charging the transmitter while showering, etc. About 0.5% - or ~1-2 readings per day, on average – are not shown because of temperature changes. For instance, going outside without a coat in a very cold place might drop the local skin temperature enough where the system would not show data. We think it is very cool it can monitor this! This system also has high temperature alerts, which have not been an issue so far – even in EU users running marathons in hot weather. During the OPH (see above), several athletes emphasized how much they love the system for exercise.

  • Eversense is the first CGM sensor with a drug component, dexamethasone (DXA) – a corticosteroid that reduces local inflammation. This was not a source of controversy given the reassuring data collected so far. Systemic levels of dexamethasone acetate approaching 1 ng/ml can be deleterious; Senseonics evaluated exposure to dexamethasone acetate in PRECISION and found levels to be below 0.05 ng/ml. When the sensor implant is removed, Senseonics found that ~85% of DXA was still left in device – confirming the slow release works as expected. Panelists were very reassured by the data, especially compared to much higher doses of DXA in other therapeutic areas.

Post-Approval Study

  • Although the panel suggested several additional areas of investigation to be explored post-approval, they agreed that Senseonics’ proposed post-approval study (n=175) will be sufficient in answering the core questions. If Eversense is approved, Senseonics intends to collect safety and efficacy data on repeat sensor insertions over two years in 175 patients (up to 20 clinical sites), comprising 350 patient-years of device exposure. The primary safety endpoint will be the rate of device- and insertion/removal-related serious adverse events (SAEs) over the first 12 months, with an easy-to-meet threshold of <7%. For context, there were zero SAEs in PRECISE and PRECISION. The study’s primary effectiveness endpoint will be time-in-range (70-180 mg/dl) at 12 months compared to the first month post-first sensor insertion. Importantly, the ongoing European Patient Registry (n=1,686 patients; see the briefing documents) is not collecting effectiveness data to our knowledge, so we’re looking forward to seeing how Eversense drives outcomes in a real-world setting. Panelists were interested in creating a similar US registry to allow for patient tracking with the ultimate goal of determining who might benefit the most from Eversense. As Ms. Carolyn Peterson (Consumer Rep) pointed out, such a registry might lend towards the creation of a decision aid to help clinicians identify patients best-suited for Eversense. By collecting longitudinal data over longer use periods, findings from a greater diversity of patients without having to cherry-pick participants might be obtained. Panelists also suggested including patient-reported outcomes and noted that taking an app-based approach might facilitate such data collection. It’s not clear if Senseonics will create a similar US registry, or simply run the post-approval study alone.

  • Panelist Dr. Walter Kraft found the dexamethasone (DXA) exposure data collected in PRECISION to be sufficient and advised against recording additional DXA metrics. As he put it, given limited resources, it’s critical to focus energy on areas that provide additional information, and the data from PRECISION “put the nail in the coffin about any concerns for DXA.” Dr. Kathleen Wyne mostly agreed, but recommended that the amount of DXA still remaining on the device upon removal be measured as in PRECISION.

  • Patient representative Ms. Anna McColister-Slipp raised repeated concerns surrounding a lack of data on the efficacy and safety of Eversense in those with microvascular complications and hydration fluctuations. While she emphasized that these limitations should not preclude Eversense from receiving approval, she’d like to see patient-reported outcomes included in post-market surveillance as they can be incredibly useful for patients when choosing between devices. We were relieved this commentary did not drive the committee off track to request additional studies; other CGMs also have limited data in these populations. We strongly agree that more patient-reported outcomes would be useful.

  • Panelist Dr. Robert Burr noted that a US registry could be helpful in identifying a population of patients for whom the insertion procedure proves challenging. He expressed concern regarding a lack of data on those undergoing chronic glucocorticoid therapy, which tends to impact the skin. Dr. Lias commented that patients on glucocorticoid therapy were excluded from Senseonics’ studies, and while Dr. Goodnow acknowledged Senseonics’ lack of experience with these particular patients, he believes age has been the “best indication,” with ~10% of the study population at 65 or above. When pressed by Dr. Lias to provide specific recommendations on what metrics could be helpful to collect in a registry, Dr. Burr called for reliable, sequential data on complication rates, unexpected injury to the skin or subcutaneous tissue, insertion failure, or infection. We’d note that the proposed post-approval study will collect all serious adverse events and AEs, which would undoubtedly include the variables described by Dr. Burr.

  • Dr. Grunberger noted that a clinician satisfaction scale is currently missing from the proposed post-market study. He humorously acknowledged that “Drs. Bailey and Russel are happy interventionists now,” though it will be important to gather input from clinicians on Eversense – particularly regarding the use of the slightly updated (and safer) blunt dissector tool. As he pointed out, clinicians who have previously used the original tool might even serve as their own controls and could be contacted to gather feedback.

Sensor Insertion and Removal Discussion

  • Both Drs. Russell and Bailey characterized the Eversense insertion and removal procedures as “fun” and easy. As Dr. Russell noted: “It’s kind of fun to do procedures again. I actually wouldn’t be surprised if physicians would find themselves enjoying doing these. I volunteer to do any in my practice.” Senseonics presented strong data in support of an easy insertion/removal procedure, showing pooled clinical trial data from 670 insertion/removal procedures in which less than 1% of removal procedures required a secondary approach to remove the sensor. In fact, in PRECISE II and PRECISION, 100% of insertions and 99% of removals were successful on the first attempt, with 91% of insertions and 80% of removals completed under five minutes.

    • We also wonder if there are business implications here – presumably,  Senseonics will have a reimbursement code for the procedure. To what degree will this make Eversense more financially attractive for clinicians vs. other CGMs?  

  • Senseonics presented an impressively comprehensive certification process for HCPs. All endocrinologists/diabetologists performing the procedure are required to undergo training and must complete a rigorous training checklist. Outside of the US, 461 clinicians have been trained on insertions, 94% of whom received their certification after just one session. 258 clinicians have been trained on removals, and 86% were certified to perform removals following one session. Dr. Russell noted that closure is performed with a steri-strip, so no sutures are needed – Senseonics management did not emphasize this point during their presentation, but it appeared to make a positive impression on the panel. Despite these robust statistics, Dr. Burr expressed some concern, noting that clinicians performing the procedure to date “know they’re being watched,” insinuating that results might be different as the population broadens. He cautioned that upon expanding Eversense to the general population, not only will the patient base be substantially more diverse, but the providers will also vary significantly in capability. We look forward to watching this – especially how excited endocrinologists are to do procedures.

  • Senseonics proposed two changes related to the insertion/removal procedure: (i) the updated Blunt Dissector Tool; and (ii) a re-designed end cap to reduce instances of breakage upon sensor removal. Although the panelists carefully contemplated the implications of these changes, they rightly agreed that Senseonics sufficiently demonstrated safety and efficacy. As Dr. George Grunberger astutely noted: “As a basic scientist it makes me shudder because you’re making changes and asking for approval in what’s essentially an educated guess. But as a clinician I was pleased to see that these changes: A) they make sense; and B) they improve the equation. As long as there’s enough post-market surveillance, I don’t think it should affect the safety or effectiveness of the system.” Ms. McCollister-Slipp echoed these sentiments, adding: “I think this is encouraging that we’re not requiring companies to stick to the old design, we’re allowing them to make improvements to the design as they move through the process.” Indeed, although there was only one severe adverse event reported, Senseonics believes it likely occurred due to inappropriate placement of the sensor. To mitigate the risk of insertion/removal adverse events, the new blunt dissector tool includes two guides plus an extended handle that prevents too-deep insertions. Dr. Steve Russell felt strongly that the new tool makes the insertion process safer and less vulnerable to mistakes. Similarly, after two reported instances of missing end caps, which are composed of biocompatible materials and thus are unlikely to pose any serious threat to the patient, Senseonics re-designed the end cap so that it is flush with the sensor, reducing the likelihood of it breaking off upon removal.

  • As requested by the FDA, panelists discussed the lack of clinical trial data on repeat sensor insertions, but it did not prove to be a major sticking point. Dr. Kraft expressed substantial confidence, noting he would feel “comfortable moving forward with multiple insertions with the data provided.” He finds no evidence of a decrement in accuracy and remarked that the side effective profile seems “mostly procedural and not exponential.” He added that the risk profile at three months or six months is probably pretty similar, except that you’re 3-6 months older. Dr. Wyne noted that the current package insert recommends alternating arms for insertion, expressing concern related to insulin injections in her slender type 1 patients – will this cause scar tissue buildup and reduce skin real estate on the arms for insulin injections? Eversense does recommend that insulin injections occur at least four inches away from the sensor, a precautionary perspective” to account for transient local insulin/glucose changes that could affect the sensor.  Dr. Burr was curious to hear whether direction was provided on how far apart repeated incision sites should be placed and was surprised to learn that after 90 days, the small scar at the insertion site is often not visible; if clinicians can locate the prior incision, they are trained to use the same scar.

Unmet Need and Expanding CGM Adoption

  • UCSD’s Dr. Jeremy Pettus and MGH’s Dr. Steven Russell provided a strong case for the unmet need of CGM, underscoring the importance of addressing barriers to adoption. To this end, Dr. Pettus presented T1D Exchange data from over 20,000 patients in the US, showing that CGM penetration stands at just 24%, and an estimated 27% of patients discontinue CGM use within one year. Of those quitters, a 61% reported problems with adhesives or the insertion site (patients could select multiple reasons; see below). This is a potential feature win for Senseonics, given anecdotal feedback on the sensor’s gentler adhesive patch, which is recommended for daily change. Providing a clinician’s perspective, Dr. Russell said he would encourage his patients to use Eversense, noting “tremendous potential” in the system, particularly in encouraging increased CGM adoption. As he put it: “the goal is to increase the use of CGM. I have patients that are waiting for this technology, including patients who have chosen not to use what is currently available. As a physician that sees many patients with diabetes daily, the benefits are numerous and certainly outweigh the risks.” We could not agree more! See below for some of the related slides from Drs. Pettus’ and Russell’s presentations, which highlight how Senseonics is likely to pitch the device in a competitive CGM market.

Close Concerns’ Questions

How quickly will Eversense be approved? What the remaining hurdles or tough discussion points for labeling? Will Senseonics need to submit any more data?

How quickly will Senseonics launch in the US? How controlled will the launch be? How many trainers are there? How long will it take to train US endocrinologists/diabetologists?

How will reimbursement look for Eversense? Will the procedure be reimbursed? Will pricing take the longer 90-day life into account?

How much could Eversense expand the market for CGM? How will patients view its pros and cons vs. other US systems: Dexcom G6 (launching this quarter, 2Q18), Abbott’s FreeStyle Libre, and Medtronic’s Guardian Connect?

Will Eversense steal share from other player, as it has mostly done to date in the EU?

How will the 180-day indication change Eversense’s value proposition? Could this tip even more people in favor of CGM?

How will clinicians view Eversense and the procedure? What kinds of patients will they recommend Eversense to? Will they reserve it for those with sensitive skin and/or adhesion issues who have found other CGMs not to work? Or will they use it to attract new CGM users?

Might Senseonics’ training program eventually expand to include other healthcare professionals such as nurses, physician assistants, and/or general practitioners? Many patients, especially those with type 2 diabetes, receiving their care from GPs – if Eversense is intended to fill an unmet need, might restricting it to those who regularly see endocrinologists limit access?

How will Senseonics be priced compared to other systems?

 

--by Maeve Serino, Adam Brown, and Kelly Close