February 18-21, 2015; Paris, France; Day #2 Highlights - Draft

Executive Highlights

Greetings from our first full day at ATTD 2015 in Paris, France, where our team has been dodging the baskets of chocolate croissants (“pain au chocolat”) awaiting attendees within the CNIT conference center confines. Day #2 continued to feature much in the way of industry updates (Medtronic, BD, Dexcom, CeQur, Sanofi), though the agenda was also highlighted by presentations on glucagon development, pumps in type 2, immune interventions in type 1, and more. Please see below for our top highlights from Day #2, see our Day #1 coverage here, and see our preview for what’s in store the rest of the week. Also, a preview for tomorrow – hallway chatter has it that Dr. Moshe Philip, in collaboration with Tadej Battelino, will announce they got a CE Mark last week for their artificial pancreas algorithm, MD Logic; we can’t wait to hear more! Other notables working with him include Eran Atlas, and Thomas Danne.

1. Medtronic’s symposium on the MiniMed 640G shared new clinical data on the system and provided sneak peeks at MiniMed 670G camp data and a potential US regulatory pathway.

2. BD presented positive data from a pilot study of its new infusion set. The set has been submitted to the FDA, and we expect a clearance by this summer. It will be compatible with Medtronic Paradigm and Luer lock reservoirs.

3. CeQur’s corporate symposium shared never-been-seen data from the company’s pilot study (n=6) of the PaQ insulin delivery device (A1c dropped 1.8% vs. MDI in only 12 weeks!) and updated timing on near-term commercialization planned in the EU and US (both in 2016). We love seeing so much good data in type 2s on better insulin delivery; even small trials are telling.

4. A high-powered team of clinicians, researchers, and health behavior experts are putting together a working group on the psychosocial aspects of the artificial pancreas; an evening workshop on the topic was packed with some most valuable viewpoints and quotes. The late Richard Rubin would’ve been very proud to see so much focus on behavior.

5. In a talk on insulin pumps in type 2 diabetes, Dr. Richard Bergenstal (International Diabetes Center at Park Nicollet, Minneapolis, MN) questioned how serious the field is about reaching A1c targets and expressed interest in a trial comparing the effectiveness of insulin pumps (including patch pumps) vs. other type 2 diabetes therapies, particularly basal insulin/GLP-1 agonist combination therapy. We too would be excited to see this, and would throw in SGLT-2s, particularly as they are not part of GRADE (we’re wishing that had been designed as an adaptive trial …).

6. Dexcom’s Mr. Jake Leach headlined the company’s symposium with a wide-ranging R&D update. The major focus was on Gen 5 and the app ecosystem it should enable, including the Apple Watch.

7. Mr. Ole Henriksen’s (Last Mile P/S, Copenhagen, Denmark) health economic analysis of the cost-effectiveness of insulin pump therapy in type 2 patients suggested that CSII is economical at a daily cost <9.2 euros.

8. Phase 2a pump data on Xeris’ G-Pump glucagon found similarity though not exact therapeutic equivalence to Novo Nordisk’s GlucaGen; a presentation from management at the exhibit hall stage covered the status of the rest of Xeris’ glucagon pipeline.

9. Dr. Jay Skyler (University of Miami, Miami, FL) outlined a proposed combination approach to type 1 diabetes immunotherapy (ATG, GCSF, an anti-inflammatory agent, and a GLP-1 agonist) that he plans to investigate in an upcoming clinical trial.

10. Dr. Malgorzata Wilinska (University of Cambridge, Cambridge, UK) presented simulation study results demonstrating that CGM with a MARD of up to 10% can be just as effective and even reduce severe hypoglycemia risk compared to intermittent blood glucose measurements in the ICU.

Top Ten Highlights

1. Medtronic’s unopposed symposium on the MiniMed 640G filled the conference’s largest hall to the brim, shared new clinical data on the system, and provided sneak peeks at MiniMed 670G data and a potential US pivotal study. Dr. Tadej Battelino presented strong clinical data on the 640G from the 100-patient, two-week SportGuard RCT: a 41% reduction in hypoglycemia events <65 mg/dl (p=0.02) when SmartGuard was turned on vs. off. Reduction in time/AUC in hypoglycemia were similarly positive (see below). This is the first clinical data we’ve seen since the PILGRIM study was shown at ADA 2013. Notably, a six-month study (n=150) is ongoing in Australia (Dr. Tim Jones) and a six-week study (n=30) is starting in Germany (Dr. Thomas Danne). Meanwhile, the great Dr. Fran Kaufman took the podium and highlighted the progress of the MiniMed 640G in the US – she said that the regulatory trial is “poised to begin very shortly” in the “coming days/weeks.” As we understand it, this is the same in-clinic pivotal trial that started last October, though it was paused to allow time to tweak the algorithm. The expected completion date is April 2015, and as we mentioned in Medtronic’s earnings report this week, the company expects to submit the PMA to the FDA “later this calendar year” (launch by April 2015, per JPM 2015). Dr. Kaufman also shared early data from a camp study (n=21; 7 days) of the MiniMed 670G/Enlite 3 hybrid closed loop system. Mean time in range (70-180 mg/dl) during the day was similar: 69% in open-loop vs. 71% under hybrid closed loop, though mean percent time spent <70 mg/dl declined 75% overnight – we speculate on the former below, and will hear much more tomorrow from intrepid researcher Dr. Trang Ly. The MARD of the Enlite 3 sensor was 12.6% vs. fingersticks in the camp study and 10.8% vs. YSI in the inpatient setting. Dr. Kaufman concluded with a potential regulatory pathway for the MiniMed 670G/Enlite 3 in the US – a three-month study to evaluate the system’s safety might support approval. Medtronic is clearly pushing hard on the automated insulin delivery front, and we wonder how the rest of the field will respond. They are very strong on infrastructure, though they have lost a couple of strong minds in R&D – that said, that’s the name of the game, and with Dr. Bob Vigersky coming in, they’ll have another strong thinker (even though Dr. Vigersky is mainly going to be working on type 2, we imagine he’ll contribute to thinking across the board on the closed loop at Medtronic – a most valuable addition to the team!)

2. BD’s Dr. Larry Hirsch presented encouraging data from a pilot study comparing the company’s new infusion set to Medtronic’s Quick-set in 25 healthy participants. Each participant wore two of each set, one inserted manually and one with an inserter; diluent was used and pressure was tracked for basal and bolus infusion. The results were impressive and highly statistically significant using either an inserter or manual insertion – the BD set reduced flow interruptions by 77%-82% and occlusion events (which included occlusion alarms) by 77-79% vs. the Quick-set. Time spent with flow interruptions was extremely low with the BD set vs. the Quick-set (p=0.018 for insertions with the inserter, and p=0.006 for manual insertions). Leakage and kinking was infrequent and not significantly different, though we noticed the BD set did have a notably higher frequency of kinked cannulas upon removal (5 vs. 1 – NS) – Dr. Hirsch noted that all of these events occurred in the first 1-2 days of the study. Dr. Hirsch did not share any product design details, but revealed in Q&A that the set has been submitted to the FDA, and BD is hoping for clearance by this summer. He also said that it will be compatible with Medtronic Paradigm and Luer lock reservoirs (i.e., Animas, Roche, Tandem), a huge win for patients. Additional studies of the BD set with HCP insertion and patient self-insertion are underway/have been submitted for presentation. The company hopes to provide “a great deal more information” at ADA in June, AADE in August, and EASD in September.

3. The clear highlight of CeQur’s corporate symposium was never-been-seen pilot study data from the company’s PaQ insulin delivery device, along with an updated EU and US timeline for commercialization. Preliminary results from the 12-week study indicate that PaQ treatment reduced A1c by 1.8% compared to MDI (baseline A1c = 8.9%) (n=6). These are the patients that need the most help, and it’s excellent to see such strong results in this pilot study. Looking at individual data, it was positive to see that every patient who transitioned to PaQ experienced an A1c reduction (i.e., 100% of patients saw some glycemic benefit). Considering too the short duration of the study, management emphasized that the “1.8% A1c reduction is pretty impressive by itself.” We would certainly agree with that.

• During Q&A, CeQur management shared new updates on the PaQ EU and US launch timelines. The company is now planning an EU launch in 2016, a slight delay relative to previous guidance that put a launch in 2015 (CE Mark approval came in November 2012). CeQur is planning a soft launch in Germany – and potentially Sweden – before expanding. In the US, CeQur plans to submit an FDA 510(k) application this year and expects a reasonable review period (potentially a 2016 US commercial launch). This is the first time we can recall the company providing US timing, a great signal of at least internal confidence. For more details on the PaQ launch timeline and pilot study, please see our below.

4. The concluding session of the day on the psychosocial aspects of the artificial pancreas revealed an initiative to form a working group on the topic. The dynamic workshop, led by Dr. Lutz Heinemann (Science & Co., Dusseldorf, Germany), Dr. Lori Laffel (Joslin Diabetes Center, Boston, MA), Dr. Katharine Barnard (University of Southampton, Southampton, UK), Dr. Korey Hood (Stanford University, Palo Alto, CA), and Dr. Jill Weissberg-Benchell (Northwestern University, Chicago, IL), was a can’t-miss for those interested in the practicalities and patient-centered elements of the closed loop. CGM was cited as an example of a technology whose uptake fell short of lofty initial expectations, and a warning to avoid an “if you build it, they will come” attitude towards the closed loop. Dr. Aaron Kowalski from the JDRF noted from the audience that there is no one definition of the artificial pancreas and that setting realistic expectations for patients will be key. Some suggested that the systems should be gradually introduced – as one attendee pointed out, few people today would immediately embrace self-driving cars without a period to build up trust. As usual, Dr. Bruce Buckingham (Stanford University, Palo Alto, CA) was right on target in pointing out how far the artificial pancreas has come for the conversation even to be turning towards human factors. We believe that a night-only system might be best to introduce first even if a 24/7 system is just as possible; that’s because there aren’t as many “interruptions” that come at night (meals, exercise, stress) and we’d love to see systems working well first without as many variables, to the point above about gradual introductions.

• The workshop was intended to lay the groundwork for a formal working group on the psychosocial aspects of diabetes technology, specifically the artificial pancreas. The starting group, led by the workshop’s five panelists and funded by the Helmsley Charitable Trust, is looking for broad involvement; anyone with interest should get in touch with them (if you do not know how to reach them, let us know). A key goal will be the development of validated self-administered measures for studying the psychosocial side of diabetes technology. Questionnaires are the bread and butter of this type of research, and Dr. Laffel’s five-minute “CGM self-efficacy” (aka confidence in abilities to use CGM) instrument may provide a good starting point. See the appendix below for a selection of the best quotes from the workshop as well as summaries of the panelists’ rapid-fire introductory presentations.

5. In a talk on insulin pumps in type 2 diabetes, Dr. Richard Bergenstal (International Diabetes Center at Park Nicollet, Minneapolis, MN) questioned how serious the field is about reaching A1c targets, emphasizing that earlier introduction of therapies might produce better results. Filling in for Dr. Bruce Bode (Emory University, Atlanta, GA), Dr. Bergenstal opened his presentation by asking whether using treatments like insulin pumps as “rescue therapies” is the best approach, noting that 80% of patients with type 2 diabetes on insulin are not at goal. He returned to this theme during Q&A, pointing out that despite stated A1c targets of 7-7.5%, most providers do not take action until a patient’s A1c reaches 8.5-9%. This is incredibly disheartening to hear from a patient (and system) perspective. With regard to insulin pumps specifically, Dr. Bergenstal believes the available evidence suggests that pump therapy (including patch pumps) can produce substantial glycemic improvements with lower insulin doses compared to MDI and potentially improve patients’ quality of life. He called for future randomized trials comparing the effectiveness of pumps (including patch pumps) vs. other type 2 diabetes therapies, particularly basal insulin/GLP-1 agonist combination therapy. We agree that an evaluation of the cost/benefit profile of insulin pumps compared to newer pharmaceutical options for type 2 diabetes (e.g., SGLT-2 inhibitors, once-weekly GLP-1 agonists, new combination drugs) would be a very interesting subject for future trials. Dr. Irl Hirsch (University of Washington, Seattle, WA) put forward during Q&A that the playing field has changed even since the OpT2mise trial began in 2012, and it remains to be seen whether implementing pump therapy for type 2 diabetes on a large scale is worth the cost and hassle factor, particularly for primary care providers. We continue to wonder the impact of “time in zone” on long-term complications and hope that moving forward, analysis of big data may help answer some of these questions.

6. Dexcom’s Mr. Jake Leach headlined the company’s symposium with a wide-ranging R&D update. He shared new pediatric data on the G4AP algorithm (MARD 10% vs. YSI, consistent with adults), news that the new Bluetooth-enabled Share receiver is available for AP research, at least 10 slides on the Gen 5 mobile app (submission expected “within the next couple of months,” which sounds on pace with the goal for March), multiple slides on the Apple watch integration and the potential ecosystem of apps (Tidepool, diasend, Databetes, etc.), the first screenshot of the Android Share app, encouraging Gen 6 data (MARD of 12% for 10-day wear and only two calibrations at startup), and more (see below). Dexcom has done an impressive job of bringing a regular cadence of products to market – the company had four FDA approvals in 2014 and the Share receiver approved here in 2015 – and we expect new offerings to come faster now that the FDA has down-classified secondary display of CGM data and retrospective software (Thanks Nightscout! Thanks Dexcom!). It’s remarkable to see how fast things are now moving – to think Dexcom could be on the Apple Watch at the same time the Apple Watch is released would have been unheard of a few years ago. Perhaps the biggest question facing the company is the cost of CGM, a barrier that was not really addressed during its symposium. This continues to be a significant issue, though we look to the ongoing DIaMonD study will address the technology’s cost-effectiveness. Of course, we really wouldn’t have expected Mr. Leach to address cost or reimbursement, given that often falls outside the scope of many R&D efforts. Still, cost is a major risk in today’s environment. Outgoing CEO Terry Gregg spoke extensively about medical device companies being paid for outcomes, and as Dexcom grows, the company is clearly thinking beyond R&D to reaching many more patients and payers with incredible, well-documented value. Another key issue is that so many providers lose money prescribing CGM, as Dr. Bob Vigerksy so elegantly put it in his provocative 2013 Diabetes Care paper – we wish there were far less fragmentation as some of this is merely knowing the right codes and some basics (not to understand the difficulties). Lots to do for both Dexcom and the entire field ...

7. Mr. Ole Henriksen (Last Mile P/S, Copenhagen, Denmark) presented an encouraging health economic analysis of the cost-effectiveness of insulin pump therapy in Germany for type 2 patients. The findings of the modeling study suggest that pump therapy could be “highly cost-effective” in Germany at a daily cost of 5.3 euros and “cost-effective” up to a daily cost of 9.2 euros. (Cost-effectiveness was defined according to thresholds establish by the World Health Organization.) The analysis was performed using the “UKPDS Outcome Model” to predict life-years saved and gained based on the cost of insulin therapy in Germany and patient outcomes data from Medtronic’s OpT2mise trial. While it was notable to hear the prediction that cost-utility can be achieved at < 10 euros per day, the findings do seem to raise as many questions as they provide answers: (i) How was the cost of MDI in Germany estimated? (ii) Will the average type 2 patients be motivated to use pump therapy as conscientiously as those in OpT2mise were, and will they obtain similar results? (iii) What does this mean for cost-utility in the “average” type 2 patient, since OpT2mise was conducted in patients in poor control? We applaud the effort to raise the level of conversation on this thorny issue as the ultimate goal is to reinforce to payers the value proposition. That of course requires showing economic utility – still challenging when the field does not have a good sense of the healtheconomic data payers want to see, or whether this is valuable at all. As we have heard, they want real-world studies to show similar economic utility, though the tradeoff for industry (longer studies, more expensive) does not seem feasible in the short-term. We’re glad to hear the cry for real-world studies and hope they will see the value. From our view, there are some major problems with the so-called “gold standard” of randomized controlled trials anyway – the “placebo” group isn’t really “placebo” since they have the care of HCPs they wouldn’t otherwise have (and they often want to do well and behave differently than they would “in real life”).

8. Xeris’ quick presentation on its non-aqueous stable glucagon in the exhibit hall’s international tech fair included a peek at new phase 2a data and a pipeline overview. Presenter Mr. Brett Newswanger (Director of Glucagon Products at Xeris) shared data from one of twelve adult type 1 diabetes patients in a phase 2a trial of the G-Pen Mini (mini-doses for moderate hypoglycemia) showing increases in blood glucose beginning after just a few minutes and lasting for enough time for a patient to eat a meal or receive medical help. Full results from this study will be presented at ADA, and subsequent phase 2 studies for the G-Pen Mini planned for 2015 include an outpatient study, an exercise study, and an alcohol study. Mr. Newswanger also touched upon a phase 2a pump study presented as a poster (abstract #0196) comparing G-Pump (the pump formulation) to Novo Nordisk’s GlucaGen. Although Mr. Newswanger stated that the PK/PD were comparable (the curves did look similar), the abstract states that the G-Pump was comparable in PK area under the curve (AUC) and that it achieved an 11% higher maximum concentration and took 10% longer to reach that point – that seems logical. The abstract states that G-Pump “was not different from GlucaGen but was not therapeutically equivalent.” We look forward to hearing how others perceive this nuanced distinction, but to us, the results certainly met the bar to continue development, especially given G-Pump’s clear efficacy and potential for the closed loop. However, this is an important point from a regulatory bioequivalence point of view, so we’ll have to see where it goes. Closed loop trials for the G-Pump formulation are still planned for this year.

• The rest of Xeris’ glucagon pipeline remains largely unchanged. The pivotal phase 3 trial for the G-Pen (for severe hypoglycemia) does not appear to have started yet, putting it behind previous guidance for initiation in 4Q14. See the appendix below for an overview of Xeris’ glucagon pipeline.

9. Dr. Jay Skyler (University of Miami, Miami, FL) outlined a proposed combination approach to type 1 diabetes immunotherapy that he plans to investigate in an upcoming clinical trial. After a rapid-fire overview of the many (mostly failed) studies of immune therapies to prevent or delay the progression of type 1 diabetes, Dr. Skyler argued that an effective immune intervention may require a combination of as many as four or five agents given the resilience of the human immune system. He outlined a general approach, consistent with his presentation on the subject at EASD, that would involve (i) an anti-inflammatory agent (anti-IL1B or anti-TNF); (ii) an immunomodulator (anti-CD3, anti-CD20, co-stimulation blockade, or ATG); (iii) stimulation of regulatory T-cell activity (IL2, GCSF, or a Treg infusion); (iv) a diabetes-related antigen (oral insulin or GAD vaccine); and (v) an agent that preserves beta cell health (GLP-1 agonist). His specific protocol includes ATG, GCSF (which have shown promise in combination in clinical trials), an anti-inflammatory agent, and a GLP-1 agonist. Acceptance of such an aggressive approach would clearly require a shift in mindset for many in the field – to illustrate this, Dr. Skyler contrasted the reaction to his proposed therapy from pediatric endocrinologists (“I would never do that to my patient”) with that of transplant surgeons, who argue that even more immunosuppression is likely necessary.

10. Dr. Malgorzata Wilinska (University of Cambridge, UK) presented simulation data suggesting that CGM with a MARD of up to 10% can be just as effective and even reduce severe hypoglycemia risk, compared to intermittent blood glucose measurements in the ICU. This study utilized a virtual population of 56 critically ill patients over a 48-hour ICU stay and tested three different ICU protocols (Yale, University of Washington, and NICE-SUGAR). The findings showed that mean glucose values and glycemic variability were similar for blood glucose measurements and for CGM with a measurement error of up to 15%. Notably, measurement errors of up to 10% for CGM use could lead to similar or favorable results in frequency and duration of hypoglycemia in this patient population. Interestingly, the type of protocol made a greater difference on glucose control than the different levels of CGM measurement error did – an indication of how essential it is to design an appropriate protocol. As Dr. Wilinska commented, CGM use in the ICU can significantly reduce workload compared to blood glucose methods as the devices can easily provide more frequent measures, trending information, and also improve safety with its alarms. CGM in the hospital has always struck us as a no-brainer from a clinical and cost perspective, which is why it’s been perplexing to see the challenges of Edwards, Echo Therapeutics, and very slow movement from others.

Honorable Mentions

1. While the headline results from the Medtronic-sponsored OpT2mise trial were a net positive for pumps, the trial provided evidence of pumps’ costs as well. The presentation of secondary results regarding treatment satisfaction and resource utilization by co-investigator and co-author Dr. Ronnie Aronson (LMC Diabetes & Endocrinology, Calgary, Canada) provided evidence for both sides of the equation. Although Dr. Aronson highlighted the across-the-board 10-20% improvements in a treatment satisfaction seen with pumps in OpT2mise, he also shared that there were nearly triple the number of unscheduled physician’s office visits in the pump group (0.76 visits/patient) than in the MDI group (0.28 visits/patient). The difference persisted when only study-related unscheduled visits were counted (0.66 visits/patient with pumps vs. 0.20 visits/patient with MDI), and there was a non-significant trend towards more hospitalizations in the pump group (15 vs. 9) although based on relatively few events. Dr. Aronson attributed the imbalance in office visits to the learning curve with pumps. Indeed, providing a peek at numbers from the second six months of the trial, he showed a roughly threefold drop in the incidence of unscheduled visits compared to the first six months in patients in the pump arm. Even so, we can see from data like this why providers perceive a possible hassle factor (and cost!) in getting patients started on pumps, despite the clinical benefits and improvement in overall patient satisfaction.

2. In a discussion on customized support for youth with type 1 diabetes, Dr. Thomas Danne (Diabetes Center for Children and Adolescents, Hannover, Germany) noted that digital technologies “might be the next big step in diabetes education.” He emphasized that the field is hugely worthwhile to explore right now considering:  (i) the number of youth familiar with smartphones, apps, and websites; (ii) that over 70% of type 1 youth have not met glycemic targets; and (iii) that more than 1,000 digital support tools are publically available. As we have seen over the past 12 months, digital health has a lot of momentum, though it is rare to hear a clinician sharing such positive sentiments. He also pointed out the positive role that social integration and peer engagement can have in translating these digital interventions into relevant outcomes. Indeed, we believe that the potential of diabetes online community is often understated, and it was great to hear him call attention to the benefits of sharing the burden of the disease with others.

Detailed Discussion and Commentary

Oral Presentations

Continuous Subcutaneous Insulin Infusion (CSII) Sets - Reduced Flow Interruptions With A Novel Investigational Catheter Infusion Set

Larry Hirsch, MD (BD Diabetes Care, Franklin Lakes, NJ)

BD’s Dr. Larry Hirsch presented encouraging data from a pilot study comparing the company’s new infusion set to Medtronic’s Quick-set in 25 healthy participants. Each participant wore two of each set, one inserted manually and one with an inserter; diluent was used and pressure was tracked for basal and bolus infusion. The results were impressive and highly statistically significant using either an inserter or manual insertion – the BD set reduced flow interruptions by 77%-82% and occlusion events (which included occlusion alarms) by 77-79% vs. the Quick-set. Time spent with flow interruptions was extremely low with the BD set vs. the Quick-set (p=0.018 for insertions with the inserter, and p=0.006 for manual insertions). Leakage and kinking was infrequent and not significantly different, though we noticed the BD set did have a notably higher frequency of kinked cannulas upon removal (5 vs. 1 – NS) – Dr. Hirsch noted that all of these events occurred in the first 1-2 days of the study. Dr. Hirsch did not share any product design details, but revealed in Q&A that the set has been submitted to the FDA, and BD is hoping for clearance by this summer. He also said that it will be compatible with Medtronic Paradigm and Luer lock reservoirs (i.e., Animas, Roche, Tandem), a huge win for patients. Additional studies of the BD set with HCP insertion and patient self-insertion are underway/have been submitted for presentation. The company hopes to provide “a great deal more information” at ADA in June, AADE in August, and EASD in September.

• There are numerous reports of CSII insulin delivery irregularities. Patients find elevated blood glucose that may/may not respond to a correction bolus – so called “Unexplained hyperglycemia.” According to Van Bon et al., DT&T 2011, ~60% of subjects reported at least one of these “unexplained hyperglycemia” events over a 13 week period, and ~30% of subjects had at least one occlusion set alarm. Added together, that means ~2/3 of patients had at least one of these events. When unexplained hyperglycemia occurs without an occlusion alarm, it is often termed “Silent occlusion.” Patel et al. (DT&T 2014) found a ~15% failure rate of initial insertions (hyperglycemia that was not reduced with a correction bolus, within about eight hours) with a polymer set.
• BD has developed an investigational set with a different design, intended to improve flow reliability – no details were shared on how the set improves flow, but we do know it is a plastic cannula. In preclinical swine studies, the new BD set reduced pressure rises and time with flow interruption by ~75% vs. sets from Medtronic, Roche, and Animas (Poster 221 at this ATTD).
• This trial was a single site, open label comparison between the BD set and the Medtronic Quick-set. It enrolled 25 participants without diabetes. Each participant had four basal/bolus diluent infusions in the abdomen through two of each set – one was manually inserted and one was inserted with Medtronic’s Quick-serter. All insertions were done by study HCPs. After set removal, the infusion site was assessed for leakage, swabbed, and measured; no leakage meant <0.5 units. In-line pressure data was collected throughout the study through use of a transducer.
• Four insulin pumps were connected by line extension tubing to the transducer. On the other side, tubing went to the infusion sets of interest. Device placement took five minutes, followed by one minute of catheter priming. Basal infusion then took place at a rate of one unit per hour, followed by bolus infusion of 10 units for 10 minutes. Then, basal infusion of one unit per hour resumed for 30-60 minutes. A forced occlusion tubing clamp ended the study and served as a positive control.

Introducing the New Frontiers in Artificial Pancreas System (Sponsored by Medtronic)

Introducing The MiniMed 640G

Tadej Battelino, MD (University Children's Hospital, Ljubljana, Slovenia)

A completely packed auditorium listened intently as Dr. Tadej Battelino presented clinical data on the MiniMed 640G from the 100-patient, two-week SportGuard randomized trial (SmartGuard ON vs. OFF). In the primary outcome, use of SmartGuard led to a 41% reduction in the median number of hypoglycemia events/week <65 mg/dl (p=0.02) and a 56% reduction in events <60 mg/dl (p=0.007). Median hypoglycemia AUC and time spent in hypoglycemia <65 mg/dl were also reduced by 56% and 41%, respectively, when SmartGuard was turned on. Time spent in hyperglycemia >140 mg/dl was significantly greater, though the significant difference disappeared for time spent >180 mg/dl. This is perhaps to be expected, given the suspension of basal insulin and depending on where patients are setting their low threshold. Notably, Dr. Tim Jones has an ongoing six-month RCT evaluating the MiniMed 640G with SmartGuard, which should give valuable longer-term data (including A1c) on the device – we even wonder if a significant finding could emerge on severe hypoglycemia. Medtronic is putting admirable effort on the clinical data front with this system, and we look forward to what comes out of these studies.

• SportGuard was an investigator initiated, 100-patient RCT of the MiniMed 640G. Patients were randomized to use the 640G with or without SmartGuard for two weeks with a three-day run-in. The primary outcome was the number of hypoglycemia events, defined as sensor glucose <65 mg/dl (<3.6 mmol/l). Of the 100 subjects enrolled, 98 completed the study at two sites (Slovenia and Israel). The mean age was 13 years, the mean A1c was 7.7%, and the mean duration of type 1 diabetes was 6.5 years. The study wrapped up on February 4, 2015, and the data analysis is ongoing. Dr. Battelino showed preliminary results.
• In the primary outcome, use of SmartGuard led to a 41% reduction in the median number of hypoglycemia events/week <65 mg/dl and a 56% reduction in events <60 mg/dl (2.9 vs. 1.7 events, p=0.02; 1.6 vs. 0.7 events, p=0.007). There were a median of zero events <50 mg/dl in both groups. We felt this was strong data on avoiding hypoglycemia for such a short RCT.
• Median hypoglycemia AUC and time spent in hypoglycemia <65 mg/dl were reduced by 56% and 41%, respectively, when SmartGuard was turned on (1721 vs. 749 mg/dl*min, p=0.009; 206 vs. 121 minutes, p=0.01). As would be expected, patients also had a much shorter magnitude and duration in hypoglycemia with the PLGM algorithm.
• Time spent in hyperglycemia >140 mg/dl was significantly greater with SmartGuard turned on: 6,002 minutes vs. 6,576 minutes (p=0.01). However, the difference largely disappeared for time spent >180 mg/dl: 3,846 minutes vs. 3,981 minutes (NS).
• Dr. Tim Jones has an ongoing six-month RCT evaluating the MiniMed 640G with SmartGuard. The objective is to compare the % time in hypoglycemia (<63 mg/dl) during six months of therapy with SmartGuard turned ON vs. OFF. The five-center, prospective study will randomize participants to SmartGuard ON or OFF.  Over 50 participants have been enrolled to date, and the goal is to enroll a total of 176 patients. This study should offer better perspective, certainly for A1c and potentially for severe hypoglycemia.
• Dr. Thomas Danne has a six-week, prospective, single-arm (before/after comparison) of the MiniMed 640G starting up in Europe. The primary outcome is area under the curve in hypoglycemia with and without SmartGuard. Twenty-four participants will be enrolled at three sites in Germany.

The Evolution of the Closed Loop System

Fran Kaufman, MD (Chief Medical Officer, Medtronic Diabetes, Northridge, CA)

Ending the session, Dr. Fran Kaufman took the podium and shared updates on the MiniMed 640G in the US and provided new details on the MiniMed 670G hybrid closed loop system.

• Dr. Kaufman said that the US regulatory trial of the MiniMed 640G is “poised to begin very shortly” in the “coming days/weeks.” As we understand it, this is the same in-clinic pivotal trial that started last October, though it was paused to allow time to tweak the algorithm. The expected completion date is April 2015, and as we mentioned in Medtronic’s earnings report this week, the company expects to submit the PMA to the FDA “later this calendar year” (launch by April 2016, per JPM 2015).
• She also shared early data from in-clinic (n=8; 3 days) and camp (n=21; 7 days) studies of the MiniMed 670G/Enlite 3 hybrid closed loop system. Stanford’s Dr. Trang Ly will present more detailed results on the camp study during ATTD Day #3, though the topline results were positive:
  
  • Mean time in range (70-180 mg/dl) during the day was similar: 69% in open-loop vs. 71% under hybrid closed loop. This was not statistically significantly different (p=0.6), which was a bit surprising. That said, kids at camp get excellent care, probably don’t miss many boluses, and tend to be in good control already, so the study may not be representative of the real world. It may also be that the algorithm is a bit conservative on dosing insulin and doesn’t kick in until very high values. We assume that is per safety and would change over time (presumably it’s a change in algorithm that the lawyers don’t want at this stage).
  • Mean percent time spent <70 mg/dl declined 75% overnight (5.2% in open loop vs. 1.3% in closed-loop; p=0.003). This part is of course the most exciting piece of hybrid closed loop systems, which is effectively fully closed-loop overnight.
  • The MARD of the Enlite 3 sensor was 12.6% vs. fingersticks in the camp study and 10.8% vs. YSI in the inpatient setting. That represents an improvement over Enlite, though to what extent is unclear, as the calibration scheme was not shared.
• Dr. Kaufman shared several slides with never-before-seen details on the MiniMed 670G/Enlite 3 hybrid closed loop systems.
  
  • US IDE – “Potential regulatory pathway.” The study will be entitled “Safety evaluation of the hybrid closed loop (HCL) System in type 1 diabetes”. The objective is to show that hybrid closed loop algorithm is safe and performs as indicated. The study will be three months in length with “intensive evaluation” in either month 1, 2, or 3. A multi-day hotel study is included in the protocol of visits. Patients will have a nighttime companion for the duration of the study. It was not clear what the comparator is in this study, but we assume it is open-loop sensor-augmented pump therapy. Devices will include the Enlite 3 sensor, MiniMed 670G insulin pump, Bayer Contour Next Link BGM, and CareLink (though CareLink will be set up quite differently for closed loop). As soon as Medtronic starts the study (patients ages 14-75 years), a pediatric study in children will also move forward. Kudos to Medtronic for including over 65 – we hope they recruit aggressively for this. Same goes for starting the study in children simultaneously.
  • Hybrid closed-loop algorithm and settings: Low and high alerts will be set at 70 and 300 mg/dl, respectively. The target glucose for the hybrid closed loop algorithm will be set at 120 mg/dl. An exercise feature may be used that raises the target to 150 mg/dl during activity – pretty smart! Several alarms are fixed into the system: sensor glucose <50 mg/dl, sensor glucose >300 mg/dl for one hour; and sensor glucose >250 mg/dl for three hours. The slide did not say, but it sounds like patients will bolus for meals as normal, and the system will take care basal insulin in the background.
• Medtronic now describes automating insulin delivery in four steps: low glucose suspend (MiniMed 530G/Veo), predictive low glucose suspend (MiniMed 640G), “overnight and hybrid closed loop” (we assume this is the MiniMed 670G, though the slide did not say), and “closed loop.” We assume the last step would be fully automated and not require any manual bolus dosing for meals.  

Workshop: Psychosocial Aspects Relating to the Use of the Artificial Pancreas

Introductory Presentations

Lori Laffel, MD, MPH (Joslin Diabetes Center, Boston, MA), Korey Hood, PhD (Stanford University, Stanford, California), Jill Weissberg-Benchell, PhD, CDE (Northwestern University, Chicago, IL), Lutz Heinemann, PhD (Science & Co., Düsseldorf, Germany), and Katharine Barnard, CPsychol (University of Southampton, Southampton, UK)

The packed panel began the workshop by introducing their shared goal: the creation of a dedicated working group focused on the psychosocial aspects of diabetes technology, with a focus on the artificial pancreas. The group is looking to pick up steam, and already has a funder in the Helmsley Charitable Trust – many thanks as always to HCT for doing so much good work in diabetes. The second half of the allotted time was dedicated to a lengthy panel discussion and audience comment session, but first the panelists rose to give brief presentations, summaries of which can be found below.

• Dr. Lutz Heinemann warned the audience not to assume that all patients will be excited to use the artificial pancreas when it arrives on the market. We think this is incredibly important and we were so glad to hear this emphasized. We believe this is especially true in the daytime. We think improvements are pretty likely for most people at night, that they will cause better days, and that there is likely to be less – perhaps much less - perceived hassle likely at night. He pointed to CGM as an example of a technology that was initially heralded as a must-have but which is still not used by close to a majority of diabetes patients – some of this is perceived or real patient or provider hassle, some is cost, etc. He lamented the dearth of data on the psychosocial aspects of diabetes technology but noted that alarm fatigue, feedback from patients’ families, and other patient-reported outcomes have a clear impact on a technology’s success. He expressed disappointment that the German IQWiG agency refused to consider patient-reported outcomes data in its cost-effectiveness assessment for CGM – we wonder how advocacy could help impact this.
• Dr. Korey Hood brought a psychologist’s perspective to the panel, noting that patient attitudes, wishes, emotions, and beliefs are often understudied (or never studied, we might add). Describing the working group, he noted that the Helmsley Charitable Trust has agreed to fund the project – how wonderful it was to hear this! The goal is to develop and validate a set of measures to assess the psychosocial and psychological impact of the artificial pancreas. These measures will be self-administered, which should allow data to be collected efficiently from a broader range of patients. Some of the variables to be evaluated include ease of use, human factors, impact on daily function, trust in a device, and perceived impact on diabetes control.
• Dr. Jill Weissberg-Benchell shared feedback from patients in recent and ongoing studies for the Bionic Pancreas. The results, a few of which we summarize below, were largely in line with what one would expect and laid an excellent foundation for the subsequent discussion. She cited trust in the device as a factor that impacted both adults and children. We were so happy to hear her impressions; Dr. Weissberg-Benchell and other leaders of this session are truly brilliant and incredibly insightful.

Table: Selected feedback from adult and pediatric type 1 diabetes patients in Bionic Pancreas studies

• Dr. Lori Laffel discussed an instrument she helped develop to assess CGM self-efficacy (CGM-SE), or confidence. Validation studies, which will soon be published in Diabetes Technology & Therapeutics, found that pediatric type 1 diabetes patients with higher CGM-SE scores used CGM significantly more three and six months after initiation and also had significantly lower A1c levels than those with lower CGM-SE scores (delta A1c of ~0.8% at three months and ~0.5% at six months between the higher and lower scoring subgroups).
• Dr. Katharine Barnard spoke about psychosocial factors in diabetes management earlier in the day during a Sanofi-sponsored symposium on Challenges in Diabetes. She strongly advocated for the inclusion of psychosocial measures as endpoints in clinical trials, and ideally as the primary endpoint in some studies – in her view, the fact that funders will only accept trials with a primary outcome related to A1c is a fundamental problem. She believes closed-loop devices represent a completely new paradigm because they would be the first to completely take diabetes management out of the patient’s hands – a welcome prospect for some but potentially anxiety-inducing for others.

Panel Discussion – Selected Highlights

Dr. Jordan Pinsker (Sansum Diabetes Center, Santa Barbara, CA): I run the artificial pancreas clinical trials at Sansum and I believe I share a common experience when patients come in for evaluation who say they wear CGM, but when I ask to see it they don’t have it with them. It turns out they don’t actually use it often, not because of cost or a lack of interest in glycemic control, but because they don’t see the benefits as outweighing the costs. I have grave concerns that when we tell patients that they will still need to bolus with an artificial pancreas or will need to tell the system when they eat, patients will perceive that the benefits may not outweigh the costs. We recently published an article on this with Dr. Barnard. What will artificial pancreas users expect of the system?

Dr. Aaron Kowalski (JDRF, New York, NY): I find it interesting and – well I wouldn’t want to say disappointing – to hear presenters throw around the term “artificial pancreas” without a proper definition of what an artificial pancreas is. I would argue that there isn’t just an artificial pancreas. There are a series of systems that will become more automated over time. We have seen here the 640G system that will reduce some hypoglycemia. There will be treat-to-range systems. Jill mentioned the anecdote from the boy in the Bionic Pancreas study who said that the tool stopped him from thinking about a plate of food as a math problem. That comment made me sad, because no artificial pancreas will fix that problem now or in the near future. Insulin does not work fast enough yet to have a fully automated system. If we set the wrong expectations, we’re doomed to failure. If you set the expectation that a closed loop system will reduce some hypoglycemia and will provide some degree of help, you’ll set the balance properly. Starting with defining what the artificial pancreas is will be very important. There is not going to be a fully automated system until we have faster-acting insulin.

Audience Comment: We should be designing different systems at different levels, to follow up on Aaron’s point. Even if we can get to a fully automated system, it may not work for everyone. There should be different levels of automation. Perhaps these systems should be designed for specific age groups. What works for an adult may not work for a teenager. I love Ed Damiano’s Bionic Pancreas, but while it may work for his son, it won’t work in everyone in my practice from age three to age 40. Also, I would like to see the gradual introduction of these types of systems. If you gave me a fully self-driving car today, I wouldn’t use it. It would take me many months before I could build the confidence. We can’t just give patients the artificial pancreas and tell them to go out and start using it right away. I would want to see some process for introducing it – maybe having monthly visits where different features were activated step-wise.

Audience Comment: I have had type 1 diabetes for 46 years and I have no idea how I’m going to be able to completely switch off from managing my diabetes. I would like to build on Aaron’s point: yes we have been talking about the artificial pancreas as a finished product, but it is a system that has multiple components. We’ve been hearing about overnight trials for years. I’ll have that now, please. As a patient, I’ve been working hard for 46 years and I’d really fancy a day off. It would be nice to be on call rather than on duty. I have an insulin pump and I’m always using either CGM or Flash Glucose Monitoring. When I calibrate and see that the sensor is tracking closely, I gladly bolus and correct based on its readings. I trust the technology I’m using. I’ll take whatever you can give me – better is better.

Audience Comment: Many of the current psychosocial studies are conducted as part of pilot studies, meaning that they happen in very self-selected subjects who are motivated and love technology. That group may not be representative of the general population.

Audience Comment: As a community we would benefit if we did not use terms like adherence and compliance. [Audience applause] If we make a lousy product then it is our problem and not the patient’s problem. When we use those terms, we assume that we are right and the patient is wrong. This applies even for CGM. Yes, CGM is helpful. But from another vantage point, it’s like a GPS: it will tell you where you are, but not what to do. Or put another way, it is like sitting in a room full of mirrors when you don’t like the way you look.

Dr. Bruce Buckingham (Stanford University, Palo Alto, CA): At the first ATTD, everything was in silico. Now, we’re talking about human factors. This is so critical, thinking about family members and significant others and notifications. We’re now looking not at what the artificial pancreas can do over a couple of days, but over years. We’ve made tremendous progress to get to the point where we can now really deal with human factors. We’ve moved from the engineering to the human relationships.

Industry Workshop: CeQur – Simple Insulin Infusion for People with Type 2 Diabetes

Demonstration of Glycemic Control in PaQ Feasibility Study

Julia Mader, MD (CeQur, Marlborough, MA)

Dr. Julia Mader presented encouraging preliminary data from CeQur’s pilot study of its PaQ insulin delivery device in patients with type 2 diabetes (n=8). The trial consisted of three study periods: baseline MDI (one week), transition from MDI to PaQ (1-2 weeks – to ensure that patients were on an appropriate basal rate), and PaQ treatment (12 weeks). Preliminary results indicated that PaQ treatment reduced A1c 1.8% from a baseline of 8.9% (n=6; one patient has yet to report and one patient dropped out of the study). Looking at individual data, it was great to see that every patient who transitioned to PaQ experienced an A1c reduction (i.e., 100% of patients saw some glycemic benefit). Considering the short duration of the study, Dr. Mader emphasized that the “1.8% A1c reduction is pretty impressive by itself.” We would agree! Based on these positive results, the company is hoping to transition into a larger clinical trial (n=25) later this year. This study will pave the way toward a hoped-for 2016 launch of PaQ in both the EU and US (details on this new timeline below).

• During Q&A, CeQur management shared new updates on the PaQ launch timeline. The company is now planning an EU launch in 2016, a slight delay relative to previous guidance that put a launch in 2015 (CE Mark approval came in November 2012). CeQur is planning a soft launch in Germany – and potentially Sweden – before expanding.
  
  • In the US, CeQur plans to submit an FDA 510(k) application this year and is cautiously optimistic for a 2016 US commercial launch. This is the first time we can recall the company providing US timing, a great signal of internal confidence. As a reminder, the company appointed Mr. Douglas Lawrence as Chief Executive Officer in April 2014, and we are hopeful that his background at BD will help CeQur navigate the complex US regulatory landscape.
  • CeQur will continue to do clinical studies throughout 2015. Dr. Mader described the next trial in the pipeline, a larger version (n=25) of the pilot study that will use blinded CGM. This is great to hear, and we will be eager to see data on time in zone and hypoglycemia.
• Returning to the pilot study, we would point out that secondary outcomes were positive as well. Total daily dose of insulin and body weight were not statistically different from baseline. On the insulin dose front, 75% of patients successfully transitioned to PaQ with the first basal dose, which is hopefully a good sign that the device should be easy for HCPs to transition patients onto. No severe hypoglycemic events, serious adverse events, or catheter site infections were reported.
  
  • Hopefully, future studies will include data on the bolus adherence front (e.g., average number of daily boluses on MDI vs. PaQ). We see the ease of bolusing on PaQ vs. MDI as a major advantage to this device and suspect that a larger study would likely bear out the expected improvement in adherence, particularly in patients sub-optimally controlled and/or forgetful.

Simple Insulin Infusion is Cost-Effective

Ole Henriksen (Last Mile P/S, Copenhagen, Denmark)

Mr. Ole Henriksen presented a health economic analysis of the cost-effectiveness of insulin pump therapy in type 2 patients in Germany. The modeling study was designed to answer one question in particular: Given that pumps were shown to be superior to MDI in poorly controlled patients (per Medtronic’s OpT2mise study), at what cost level is CSII a good value for money? The researchers utilized the “UKPDS Outcome Model” to predict life-years saved, running the analysis based on the cost of insulin therapy in Germany and patient outcomes collected from the OpT2mise trial. Results indicated that pump therapy would be “highly cost-effective” at a daily cost of 5.3 euros and would be “cost-effective” up to a daily cost of 9.2 euros. (Cost-effectiveness was defined according to thresholds establish by the World Health Organization). During Q&A, Dr. Henriksen noted that the cost-utility does not derive entirely from projected reduction in complications associated with pump use (~19,000 euros over a lifetime) as one might expect. Rather, he emphasized that the addition of life years projected by the model for patients on pumps plays a major role as well.

• It was notable to hear the prediction that cost-effectiveness with pumps in type 2 patients can be achieved at under 10 euros per day – however, the findings seem to raise as many questions as they provide answers: (i) How was the cost of MDI in Germany estimated? (ii) Will the average type 2 patients be motivated to use pump therapy as conscientiously as those in OpT2mise were? (iii) What does this mean for cost-effectiveness for the “average” type 2 patient since OpT2mise was conducted in patients in poor control?
• It is great to see the effort at modeling what has become a big topic of controversy for type 2 patients. However, we would note that the goal is to get payers to appreciate this value. We wonder how they would respond to the modeling data and whether they would demand more “real-world” findings to support these claims. It’s hard for us to know exactly what data payers want to see to cover simple CSII.

Panel Discussion

Q: When talking about the improvement in glucose control with continuous infusion, what contribute more to the improvement we see? Better adherence or the flatter profile of the resulting insulin levels?

Dr. Juan Pablo Frias: Probably both. But, better adherence is probably the bigger factor. I think certainly having the device attached to you and not missing the multiple doses is the factor that helps patients. The better bioavailability certainly plays a role as well in terms of more physiological delivery.

Q: How does the device stick on the body?

Dr. Peter Gravensen: The device is built with adhesive tape, so it is taped onto the body.

Dr. Leslie Lilly: We have done a number of adhesive studies to assess for dermal irritation and for comfort. So we have identified the type of dermal tape that meets those criteria.

Q: I would worry that the PaQ insulin-containing balloon would suddenly burst and dump insulin faster than desired into my patient? Can you explain why this can’t happen?

Dr. Gravensen: When we have pressurized insulin, we have to ensure that the insulin cannot get into the user any more quickly than we intend. Any pathway from the product into the user consists of checks and balances. There is redundancy built in, just like it is for aircrafts or space shuttles.

Q: Can I buy the pump today?

Dr. Lilly: We do have a CE Mark for the product. We are planning a limited launch in Germany by the end of this year. And then, we’ll expand. [Correction: Management shared after the symposium that the EU launch is planned for 2016.]

Q: Dr. Henriksen, in your model, where is the cost-utility coming from? I assume it is coming from the fewer complications?

Dr. Ole Henriksen: Actually, the cost of insulin plays a big role. The reduction in complications is not the biggest factor, but the addition of life years in the model is a major factor as well. That’s the two parts of it.

Q: Can you comment on clinical trial results and the significance of the A1c drop?

Dr. Julia Mader: Our first two-week study showed an improvement in A1c even though you would not expect it in such a short period of time. We just completed our first three-month study. Now, we’re looking forward to another three-month study. However, the 1.8% A1c reduction we saw here is pretty impressive by itself. That is a substantial improvement in and of itself. That too, anecdotally, we’ve heard that some of the patients in our trials that have gone back on MDI have seen their A1c increase again. They want to come back to our trial.

Corporate Symposium: Dexcom Continuous Glucose Monitoring – Superior Technology Leading to Quality of Life Improvement and Innovation with the Artificial Pancreas (Sponsored by Dexcom)

Dexcom Continuous Glucose Monitoring in Technology – Leading the Way in Accuracy, Performance and Innovation

Jake Leach (Senior VP of R&D, Dexcom, San Diego, CA)

Mr. Jake Leach headlined the Dexcom symposium with new pediatric data on the G4AP algorithm (MARD 10% vs. YSI, consistent with adults), news that the new Bluetooth-enabled Share receiver is available for AP research, at least 10 slides on Gen 5 (submission expected “within the next couple of months,” on pace with the goal for March), multiple slides on the Apple watch integration, the first screenshot of the Android Share app, and more.

• The new Dexcom G4AP algorithm (Software 505) demonstrated a MARD of 10% vs. YSI in a 79-patient pediatric study (2-17 years old). This was impressively similar to the 9% MARD originally collected in adults. Notably, 91% of points were within 20%/ 20 mg/dl, consistent with the 93% observed in adults. Sensors were inserted in the upper abdomen or upper buttocks and calibrated every 12 hours with SMBG. The seven day study included one in clinic session vs. YSI in 6-17 year-olds (they spared the 2-5 year olds!) on day one, four, or seven. Dexcom collected 2,262 paired CGM-YSI points in a wide data range of 40-40 mg/dl. The study was presented as a poster at this ATTD (Laffel et al.).
  
  • This data should allow Dexcom to secure pediatric approval for the new algorithm, which was initially approved and launched in November for patients 18 and over. Mr. Leach noted that patients should especially see more responsiveness with the new algorithm when glucose is changing.
• The new Bluetooth-enabled Share receiver (approved in January) “really enables remote monitoring 24 hours per day.” With the original Share system + docking cradle, remote monitoring was really limited to nighttime or when an outlet was nearby. The portability benefits are significant. We also couldn’t help but think of this as an indirect reference to Nightscout. As we noted in our Share receiver approval report, Nightscout is not going away, though the convenience of the Share receiver seems quite high, and once an Android version is out, the reasons to choose Nightscout may be less compelling.
  
  • In line with the new Share receiver, Dexcom is making available an orange Share AP receiver for investigational use (“a new tool for artificial pancreas research”). The receiver includes the new G4AP (Software 505) algorithm. This device will allow investigators to wirelessly send results from the Dexcom receiver to an artificial pancreas controller via Bluetooth – an especially big benefit for home studies. In the past, researchers have had to hardwire the artificial pancreas controller to the Dexcom receiver (e.g., the Bionic Pancreas brick), or use other solutions like the Mazlish box. Dexcom has an FDA master file and an investigator brochure; Mr. Leach asked attendees to contact Dexcom through its website.
• The Gen 5 mobile platform is “on track” for an FDA filing “within the next couple of months.” As a reminder, the goal as of Dexcom 3Q14 and JPM 2015 was a filing in calendar 1Q15 (by March), which means things seem to be going according to plan. Mr. Leach showed several screens of the interface, which we previously saw at JPM 2015. However, he did point out that the app enables “in-app help screens” to answer common questions and help troubleshoot problems. As we’ve reported in the past, G5 will also build in the Share capabilities and offer the Share receiver to communicate directly with the Bluetooth enabled transmitter (backup for the phone or if someone doesn’t want to use a phone) – the transmitter has the flexibility to communicate with two devices. See a picture on Twitter here.
• Mr. Leach had multiple dedicated slides on the Apple Watch integration with Dexcom apps, starting with the Follow app – as we reported last week, the plan is for this app to be ready once Apple releases the watch in April. See a picture on twitter here. Said Mr. Leach, “We intend to link up most of our apps with the Apple Watch once it’s released.”
• We got a first glimpse at the Android Follow app, which looked identical to the Apple version – see a picture on Twitter here. Though the Android app is trailing the Apple Follow app quite a bit (the Apple version was launched in October), Dexcom hopes to develop these in parallel going forward. This is great news for patients, since “Where is the Android version?” is one of the most common questions we hear. Dexcom is also working on an Android version of Gen 5.
• Mr. Leach devoted a full slide to the renewed Dexcom and Insulet collaboration – see a picture on Twitter here. Insulet was the only pump company to have a dedicated slide. Indeed, Asante was not listed on any slide, though Mr. Leach mentioned them verbally. Tandem and Animas were not mentioned at all. The gating factor on this partnership will be Insulet’s new PDM, which is slated to be shown at ADA 2015, though not submitted to the FDA until early 2016.
• In line with Dexcom’s JPM 2015 presentation, Mr. Leach showed the new sensor applicator, which permits sensor application with one hand. The single push button device hides the needle and deploys the sensor in 50 milliseconds. See a picture on Twitter here.
• Also from JPM, Dexcom showed its new 30% smaller transmitter vs. the current G4 Platinum slim transmitter. Notably, the new transmitter is also lower cost to manufacture!
• Mr. Leach showed encouraging topline data on the Gen 6 sensor from a 40-patient study (more than we recall seeing at JPM) – a single calibration at startup (two BGM values) resulted in an 11.7% MARD vs. fingersticks (days 1-7) and similar 12.1% MARD vs. fingersticks (days 1-10). The pilot study didn’t use YSI and values were blinded to patients, something Dexcom plans to change in studies going forward. Mr. Leach said the team is “very excited about the performance.” As we noted at JPM, this is slightly worse accuracy than G4AP (MARD: 9.0% with two calibrations per day), but the tradeoff of only calibrating at startup and a longer approved wear time is likely worth it. Gen 6 will block interfering substances with a new membrane, as well as employ new algorithms to detect more outliers.

International Technology Fair

Xeris – Clinical Development Overview of XeriSol Glucagon

Brett Newswanger (Xeris Pharmaceuticals, Austin, TX)

Mr. Brett Newswanger, Xeris’ Director of Glucagon Products, provided an overview of the company’s glucagon portfolio. It was valuable to get an up-to-date look at the status of Xeris’ pipeline and a window into what to expect later this year (see table below). In addition to a peek at data from one patient in a phase 2a mini-dose trial (covered above), we also saw phase 2a results from a pump study (G-Pump) vs. Novo Nordisk’s GlucaGen delivered via OmniPod (abstract #0196). Though the pharmacokinetic profile was comparable in terms of area under the curve in the first 150 minutes, the G-Pump reached an 11% higher maximum concentration and took 10% longer to reach that point than GlucaGen. To us it makes sense that it would take longer to reach a higher peak concentration, and we will be taking a closer look at the full poster presentation. Despite the slight nuanced differences, it appears that the results are enough to support G-Pump’s continued development, especially given its great potential for the closed loop. That said, we’ve heard that bioequivalence is a major factor on the regulatory side of glucagon, so we’ll have to see how the data plays out.

Table: Xeris’ glucagon pipeline

-- by Melissa An, Adam Brown, Varun Iyengar, Emily Regier, Manu Venkat, and Kelly Close