July 14-17, 2016; Keystone, Co; Day #2-4 Highlights – Draft

Executive Highlights

We made the 9,000 foot descent from Keystone, bidding farewell to ATDC 2016, and here we bring you our coverage of the balance of the meeting – we hope you saw our day #1 coverage on Friday! In the world of diabetes drugs, we saw a couple of controversies unfold: Dr. William Herman suggesting that the cardioprotective benefits of BI/Lilly’s SGLT-2 empagliflozin and Novo Nordisk’s GLP-1 liraglutide could be more cost-effectively achieved with the use of generic drugs such as sulfonylureas (!), and Dr. Phil Cryer proposing changes to the current definition of hypoglycemia. Additionally, Dr. Desmond Schatz continued his call for urgency in bringing diabetes to “212 degrees” and provided an update on five ongoing TrialNet type 1 diabetes studies. The conference also included the highly-anticipated Patient/Provider panel featuring illuminating testimonies from four of Dr. Satish Garg’s patients, and a talk by Dr. Ralph DeFronzo highlighting some “miraculous” past findings on GLP-1 agonists’ capacity to restore beta cell function that, years later, he asserts, most endocrinologists are not aware of. On the diabetes tech front, we received fascinating insights from Dr. Ramachandra Naik on the study design for J&J’s automated insulin delivery device with Dexcom’s G5 CGM, Dexcom VP of R&D Jake Leach on the company’s pipeline updates, and Dr. Satish Garg on upcoming results from the Medtronic 670G pivotal trial. Furthermore, we heard much discussion on the artificial pancreas, with the FDA’s Dr. Courtney Lias and the JDRF’s Dr. Aaron Kowalski offering respectively regulatory and scientific perspectives on the issue.  

 See below for our top highlights in diabetes drugs and diabetes technology from the absolutely information-packed days#2-#4. For more insight, check out our past highlights from day #1.

Top Diabetes Drug Highlights

1. Dr. William Herman (University of Michigan, Ann Arbor, MI) rather controversially suggested that the cardioprotective benefits of empagliflozin (Lilly/BI’s Jardiance) and liraglutide (Novo Nordisk’s Victoza) might be replicated more cost-effectively with the intensification of cardiovascular and other agents such as aspirin, statins, and even sulfonylureas. In our view, this assumes that patients adherence to SFUs will be just the same as other compounds that have fewer side effects – it also assumes an absence of beta cell burnout.

2. On the topic of the ongoing debate over how to define hypoglycemia, Dr. Phil Cryer (Washington University, St. Louis, MO) proposed the addition to the definition of severe hypoglycemia of a criterion of a plasma glucose <50 mg/dl.

3. Dr. Desmond Schatz (University of Florida, Gainesville, FL) picked up right where he left off in his address at ADA, calling for more urgency in bringing diabetes to “212 degrees” and suggesting that one reason for our inertia is the heterogeneity of the condition, focusing particularly on type 1 diabetes.

4. Dr. Desmond Schatz (University of Florida, Gainesville, FL) also provided an update on the enrollment progress of each of TrialNet’s five type 1 diabetes prevention and intervention trials – namely, the two oral insulin trials, the abatacept trial, the teplizumab anti-CD3 trial, and the ATG/GCSF combination trial. There was positive news on enrollment for most, outside TN18 abatacept, which has progressed more slowly than expected.

5. Four of Dr. Satish Garg’s (Barbara Davis Center, Aurora, CO) patients discussedunique stories of life with diabetes, provoking laughs, tears, and a standing ovation from a packed audience. We salute Dr. Garg for bringing patient voices into this meeting for so many years!

6. Dr. Ralph DeFronzo (UT Health Science Center, Dallas, TX) discussed what he characterized as the GLP-1 agonist class’ “miraculous” effect on the beta cell, highlighting two past GLP-1 studies he regards as “great breakthroughs in diabetes” that, he said, the majority of endocrinologists are not familiar with. Dr. DeFronzo also highlighted the promise of emerging combination approaches that merge GLP-1 agonism with action on other gut hormones. Among these are GLP-1/glucagon, GLP-1/PYY, GLP-1/GIP, and GLP-1/oxyntomodulin.

7. In his presentation of cardiovascular outcomes trials (CVOTs) for diabetes drugs, Dr. Jay Skyler (University of Miami, FL) expressed surprise that the SUSTAIN 6 CVOT topline results were able to claim a statistically significant reduction in cardiovascular risk for Novo Nordisk’s once-weekly GLP-1 agonist semaglutide with just 250 events. 

Top Diabetes Technology Highlights

1. Dr. Ramachandra Naik (J&J, Wayne, PA) shared the most detail ever on the planned pivotal study design for J&J’s automated insulin delivery device (a hypoglycemia-hyperglycemia minimizer) with Dexcom’s G5 CGM.

2. Dexcom VP of R&D Jake Leach shared several notable pipeline updates in his overview of Dexcom technology. Among his updates were that the FDA has approved the IDE for the G6 pivotal trial, which will begin shortly – this will be Dexcom’s largest study ever. Mr. Leach briefly referred to Thursday’s Advisory Committee meeting on whether G5 should be approved for insulin dosing – while he avoided speculation on potential outcomes, it was noted that this indication has been approved in Europe.

3. Dr. Satish Garg (Barbara Davis Center, Aurora, CO) announced, for the first time to our knowledge, his hope to publish results from the 670G pivotal trial extension phase by the end of 2016 (>1 year at-home use of the MiniMed 670G).

4. In a refreshingly titled talk, “Regulatory Promotion of Technology Innovation,” FDA’s Dr. Courtney Lias picked up where she left off at the NIH Artificial Pancreas Workshop, repeatedly asking for audience input as to how the FDA can make regulatory processes easier. In Q&A she touched on the OpenAPS movement, noting that as the movement expands beyond personal use, the risk increases and regulation is more likely to be enforced; on the plus side, she highlighted that “these movements apply pressure on the system to approve devices faster, so that we know that the technology that people are using is safe.”

5. In two talks on closing the loop, JDRF’s Dr. Aaron Kowalski highlighted the need to manage unrealistic expectations, to learn from the commercialization of CGM, his position as a “glucagon realist,” and that he is still a big believer in amylin.

Top Diabetes Drug Highlights

1. Dr. William Herman (University of Michigan, Ann Arbor, MI) rather controversially suggested that the cardioprotective benefits of empagliflozin (Lilly/BI’s Jardiance) and liraglutide (Novo Nordisk’s Victoza) might be replicated more cost-effectively with the intensification of cardiovascular and/or glucose-lowering agents such as aspirin, statins, and even sulfonylureas. He argued that if the cardiovascular benefits of empagliflozin in the EMPA-REG OUTCOME trial and liraglutide in the LEADER trial were due to a combination of their small improvements in A1c, weight, and cardiovascular risk factors rather than off-target effects, these risk factors can be addressed through cheaper, older medications. This would imply in our view that patients would be engaged with these compounds and take them! Dr. Herman noted that the absolute risk reduction of the composite three-point MACE outcome (cardiovascular death, non-fatal MI, or non-fatal stroke) was relatively small in these trials with a high number needed to treat and high cost to prevent one primary event. In the case of empagliflozin, with a number needed to treat of 143 and an average wholesale price of $470 per person per month, it would cost $700,000 to prevent one primary event. In the case of liraglutide, with a number needed to treat of 200 and an average wholesale price $928 per person per month, it would cost $2 million to prevent one primary event. Others would argue with this math (given the costs vary globally, etc.) – still, Dr. Herman argued that the participants in both trials were not optimally managed in terms of their glucose and various cardiovascular risk factors (we can’t imagine those who designed/executed the trial would not disagree). In the EMPA-REG OUTCOME trial, mean baseline systolic blood pressure was 135 mmHg while mean baseline LDL cholesterol was 86 mg/dl. In the LEADER trial, mean baseline systolic blood pressure was 136 mmHg. Dr. Herman pointed out that this indicates a substantial portion of participants had blood pressure and cholesterol levels that were significantly higher than the mean and that these averages are already rather high for patients with pre-existing cardiovascular disease. He also felt that cardiovascular therapies such as aspirin, beta blockers, anti-platelet agents, ACE inhibitors, and statins could be further intensified in these patients. For example, only about 71%-77% of participants in EMPA-REG OUTCOME and LEADER were on statins at baseline. He asserted that almost all of the participants eligible for these trials should have already been on high-dose statins at baseline. Furthermore, in an ensuing panel discussion, Dr. Herman went so far as to suggest that the 0.5% A1c difference at the end of EMPA-REG OUTCOME and LEADER could account for the cardioprotective effect and that the addition of sulfonylureas to achieve comparable glycemia could achieve a similar result. He also commented that sulfonylureas were shown to be safe and effective in the UKPDS.

• This statement doesn’t, of course, adequately reflect sulonylureas’ well-characterized negative impact on increasing hypoglycemia risk and weight gain (to say nothing of beta cell burnout), which have a negative impact on cardiovascular health. In the LEADER trial, participants in the placebo group were significantly more likely to initiate sulfonylurea therapy during the trial (p<0.001) (and were also more likely to initiate other antihyperglycemic agents), yet the liraglutide-treated group was still able to achieve a lower mean A1c and experienced a significant reduction in risk for the primary MACE outcome and cardiovascular death. In fact, Dr. Anne Peters (USC, Los Angeles, CA) previously interpreted the LEADER results as further evidence against the use of sulfonylureas. Furthermore, we imagine the intensification of therapy through the addition of numerous cardiovascular and antihyperglycemic agents could have negative consequences on patient medication burden and adherence. There are a variety of factors to consider regarding empagliflozin and liraglutide use; for example, patients are able to achieve the combination of improvements in A1c, weight, and cardiovascular risk factors without the addition of multiple new agents they would otherwise need. Some of these new agents could be low-cost, although then the problems associated with low adherence may prompt other problems that are more costly (i.e., patients not taking medicine so going onto dialysis, having preventable heart disease etc).
• In his discussion of CVOTs the day before, Dr. Jay Skyler (University of Miami, FL) bluntly stated his distaste for the cost argument for the use of sulfonylureas and even metformin. As he put it, sulfonylureas are known to increase the risk of hypoglycemia, cause weight gain, and increase cardiovascular risk, so “why the hell are we using any sulfonylureas?” Dr. Skyler considered the argument for their use on the basis of cost and accessibility to be a “cop-out” and characterized it as “irresponsible” to prescribe sulfonylureas. More controversially, Dr. Skyler also suggested that the evidence to justify use of metformin is “slim to nonexistent” and its A1c-lowering efficacy is fairly low at -0.4%, so he personally doesn’t use metformin in his own practice. Dr. Skyler felt similarly about DPP-4 inhibitors, noting that the class is much weaker than GLP-1 agonists and three CVOTs thus far has demonstrated a decidedly neutral effect of the class on cardiovascular outcomes. On the other hand, Dr. Skyler supported the use of the TZD pioglitazone, the SGLT-2 empagliflozin, and GLP-1 agonists based on their antihyperglycemic and cardioprotective potential. Overall, Dr. Skyler advocated for a shift toward a value-based reimbursement system that will provide coverage for drugs with proven benefits.
• Notably, ADA Chief Medical Officer Dr. Robert Ratner was similarly concerned about the high direct and indirect economic costs of diabetes, and advocated for the prevention of diabetes and diabetes-related hospitalizations through better outpatient management. He shared that the ADA’s 2012 economic analysis of diabetes costs found that $76 billion of the total $176 billion direct medical costs of diabetes was attributable to hospital inpatient care, by far the largest category of direct diabetes costs. For comparison, the cost of diabetes medications and supplies at $21 billion was less than one-third of the cost of diabetes-related hospitalizations – some argue this figure should be higher given the impact on long-term outcomes, though this is very hard to prove without decade-long trials etc. Physician office visits tied with nursing/residential facility stays for the lowest spending category, accounting for only $15 billion of costs each. He persuasively argued that reducing hospitalizations would have a remarkable beneficial health economic impact. For instance, he noted that the cost savings of keeping a single patient out of the ER for a single day could pay for a year of CGM for that patient. Keeping one patient out of the ICU would pay for all of the insulins and other diabetes-related medications for 10 people for a year. In an afternoon session on diabetes in the elderly, Dr. Ratner noted that the prevention of one sulfonylurea-induced hypoglycemia-related ER visit would pay for five years of DPP-4 inhibitor therapy for that patient. Ultimately, Dr. Ratner supported the reduction of diabetes-related hospitalizations through (i) improving outpatient management through diabetes medications and prescriptions for diabetes complications; (ii) increasing access to and use of outpatient diabetes diagnosis and care through Medicaid expansion under the ACA; and (iii) preventing diabetes through widespread reimbursement of the Diabetes Prevention Program (DPP). To reduce diabetes medication spending, rather than advocate use of sulfonylureas, Dr. Ratner called for Medicare Part D to have the ability to directly negotiate drug prices with manufacturers, in place of the current system in which the for-profit pharmacy benefits managers (PBMs) for individual commercial sponsors for Part D plans negotiate with the pharmaceutical companies. We appreciated Dr. Ratner’s candid take on rising diabetes costs and his proposal of solutions that don’t involve limiting patient access to medications that significant benefits in terms of safety and quality of life as well as A1c efficacy.

2. On the topic of the ongoing debate over how to define hypoglycemia, Dr. Phil Cryer (Washington University, St. Louis, MO) proposed the addition to the definition of severe hypoglycemia of a criterion of a plasma glucose <50 mg/dl. Currently, the ADA defines severe hypoglycemia as an event requiring the assistance of another person and defines documented symptomatic and asymptomatic hypoglycemia as a plasma glucose <70 mg/dl with or without symptoms. He suggested that the definition of severe hypoglycemia be expanded to include hypoglycemic events with a self-monitored plasma glucose, CGM, or lab measurement of plasma glucose <50 mg/dl, along with the previously-defined hypoglycemic events requiring assistance from another person. Dr. Cryer prefers <50 mg/dl as the plasma glucose threshold for hypoglycemia because it is unambiguously dangerous, should be scrupulously avoided, and is a clear marker of severe hypoglycemia even if assistance from another person is not required. He did not propose changes to the current consensus on non-severe documented hypoglycemic events, which we appreciate as plasma glucose values <70 mg/dl can still often negatively impact a patient’s quality of life. Dr. Cryer did note that only the plasma glucose criterion of <50 mg/dl should be used to diagnose severe hypoglycemia in children, since pediatric patients typically require assistance no matter the severity of the event. Furthermore, he pointed out that there is tremendous variation in how different individuals experience hypoglycemia and that plasma glucose <70 mg/dl is not necessarily dangerous for all patients. As Dr. Cryer put it, “it’s not possible to cite a specific plasma glucose concentration that ‘defines’ hypoglycemia in diabetes, because the critical glucose concentration varies within and among individuals based on frequency of hypoglycemia, A1c, and other factors.” We imagine it’s possible that, using a stricter <70 mg/dl definition, the FDA has underestimated the benefit of drugs like Novo Nordisk’s Tresiba (insulin degludec) in reducing the risk of more unambiguously dangerous severe hypoglycemia (Novo Nordisk used a confirmed hypoglycemia plasma glucose threshold of <56 mg/dl in its trials for Tresiba). Hard-set or inappropriate hypoglycemia thresholds may oversimplify patient diversity and restrict our full understanding of new medications. We’d love to hear more consensus on different levels of hypoglycemia from key opinion leaders in the field, and for the FDA to consider this heterogeneity when evaluating new drugs and proposed label updates. We do have a growing sense that what payers care the most about is patients avoiding hospital visits and ambulance calls, not particularly avoiding hypoglycemia (however defined).

3. Dr. Desmond Schatz (University of Florida, Gainesville, FL) picked up right where he left off in his address at ADA, calling for more urgency in bringing diabetes to “212 degrees” and suggesting that one reason for our inertia is the heterogeneity of the condition, focusing particularly on type 1 diabetes. Not only does type 1 diabetes manifest distinctly in adults vs. children and in unique individuals, but the disease itself presents differently today than it did 20 years ago – “the changing face of type 1 diabetes” as Dr. Schatz put it. Only 67% of organ donors with type 1 diabetes aged 1-14 and 29% aged 15-36 present with insulitis today, compared to what used to be 100%. Patients positive for insulitis also have greater beta cell mass. Interpreting these findings, Dr. Schatz posed a contentious question: “Is it possible that insulitis is not destructive, but in fact, may be protective?” He shared type 1 diabetes incidence data from the TEDDY cohort to demonstrate a divergence in the first autoantibody that appears (41% GADA only, 39% insulin only, 13% both GADA and insulin autoantibodies) and asked, does a different “first” autoantibody suggest a different disease? Dr. Schatz urged the audience to confront questions like this as the heterogeneity of type 1 diabetes becomes too powerful to ignore. He made the case that acknowledging this heterogeneity would improve the way we understand, prevent, and treat type 1 diabetes, pushing us toward personalized medicine. Dr. Schatz warned that, while big data may lead us to the “coalescence of factors” that cause type 1 diabetes, healthcare providers must continue to approach each patient as a unique case in order to achieve the best possible outcomes in diabetes care. Dr. Schatz emphasis on the heterogeneity of the type 1 diabetes echoed Dr. Ratner’s presentation on the heterogeneity of type 2 diabetes on the first day of the conference – we hope that greater attention to this topic can lead to practical approaches for truly individualized and personalized medicine in both type 1 and type 2 diabetes.

4. Dr. Desmond Schatz (University of Florida, Gainesville, FL) also provided an update on the enrollment progress of each of TrialNet’s five type 1 diabetes prevention and intervention trials: the two oral insulin trials, the abatacept trial, the teplizumab anti-CD3 trial, and the ATG/GCSF combination trial. He repeated TrialNet chair Dr. Carla Greenbaum’s update from the first day of the conference that the TN07 trial of oral insulin is fully enrolled and the results will be presented at ADA 2017 in San Diego. Furthermore, he also noted that the final enrollment of the trial was 562 participants, significantly exceeding the estimated enrollment of 400 on ClinicalTrials.gov. In addition, he shared that a second oral insulin trial, TN20, will initiate soon. This trial will be a mechanistic study similarly aimed at patients with stage 1 type 1 diabetes (with the presence of multiple autoantibodies and normal glycemia). The study hopes to determine how different doses and dosing intervals of oral insulin might impact immune response. In particular, the trial is evaluating 75 mg of oral insulin daily and 500 mg oral insulin every other week. TrialNet hopes to complete enrollment by the end of December 2016 and enroll up to 80 participants from the US and 20 from Canada, Europe, or Australia. Enrollment began in March and has already exceeded expectations thus far with 56 participants enrolled by June. The trial is expected to complete in December 2017 (primary completion expected September 2017) according to ClinicalTrials.gov. On the other hand, enrollment for the TN18 abatacept trial similarly aimed at participants with stage 1 type 1 diabetes has progressed slower than expected. That said, Dr. Shatz shared that as of May 2016, the trial had enrolled 114 of the total 206 expected participants (compared to a target of 130 participants enrolled by May 2016). The trial is expected to complete in February 2018 according to ClinicalTrials.gov. Enrollment for the TN10 anti-CD3 teplizumab trial is slightly ahead of schedule and had enrolled 58 of the total 71 expected participants by May 2016 (compared to a target of 54 participants enrolled by May 2016). This study is aimed at participants with stage 2 type 1 diabetes (with the presence of multiple autoantibodies and abnormal glucose tolerance) and is expected to complete in January 2022 (primary completion expected January 2021) according to ClinicalTrials.gov. Finally, the TN19 trial of ATG/GCSF in patients with new-onset type 1 diabetes just reached its target enrollment goal of 84 participants, significantly exceeding enrollment expectations. ClinicalTrials.gov lists the estimated primary completion date of this trial as October 2016 and Dr. Schatz expressed his optimism that results would be available next year. It’s encouraging to hear that that most of these trials, except TN18 abatacept, have reached or are nearing complete enrollment. Dr. Schatz noted that it’s important to begin thinking about the next generation of TrialNet trials and potential immune and therapeutic agents that will build on the insights from these trials. In the meantime, see our coverage of Dr. Greenbaum’s TrialNet update from November 2015 to learn more about the aims behind each trial.

5. Four of Dr. Satish Garg’s (Barbara Davis Center, Aurora, CO) patients – Courtney, Kimberly, Patricia, and Troy – discussed their unique stories of life with diabetes, provoking laughs, tears, and a standing ovation from a packed audience. This discussion spanned a vast array of topics, but two predominant themes emerged: mental health and weight management. Courtney recounted her experience grappling with the combination of type 1 diabetes and bipolar disorder, describing the difficulty of simultaneously managing the ups and downs in both blood sugar and mood. The topic of psychological health remained prominent as Kimberly, Patricia, and Troy followed with emotional descriptions of their journeys of weight loss and gain with diabetes. Kimberly, who has had type 1 diabetes for over 40 years (since the age of 6), explained the difficulty of diabetes management during her two pregnancies and how it led her to gain 40 pounds. A severe cardiac event at age 38 served as an impetus for her to begin a lifestyle modification program that helped her lose the weight with diet and exercise, although in subsequent years she has vacillated between gaining and losing the final 20 pounds. “People with diabetes are not an equation,” she remarked, explaining how the complexity and heterogeneity of diabetes makes it difficult to understand which measures will be effective. On this note, Troy (Kimberly’s brother) and Patricia closed the session with descriptions of their two very different experiences with bariatric surgery. Troy has lost hundreds of pounds (down from 450 lbs) after a Roux-en-Y gastric bypass and proclaimed passionately that “bariatric surgery saved my life.” For Patricia, however, bariatric surgery was a significant disappointment. Diagnosed at age 26, after beginning insulin therapy Patricia saw her weight increase to over 250 pounds. The Lap Band procedure (presumably years ago) helped her lose about 30 pounds, but the weight slowly came back despite very strict adherence to the recommended diet and exercise modifications. A former high school athlete, she expressed her desire to try exercising again, but lamented that no exercise program or diet to date has worked. This Patient/Provider Q&A panel is a beloved and highly-anticipated tradition at Keystone, and we encourage you to read the transcript from this illuminating discussion below.

6. Dr. Ralph DeFronzo (UT Health Science Center, San Antonio, TX) discussed what he characterizes as the GLP-1 agonist class’ “miraculous” effect on the beta cell, highlighting two past GLP-1 studies he regards as “great breakthroughs in diabetes” that, he asserted, the “majority of endocrinologists are not familiar with.” Dr. DeFronzo suggested that GLP-1 agonists “saves the beta cell,” promoting weight loss and reducing A1c. It is well-understood that GLP-1 agonists increase insulin gene transcription (Pdx-1) and inhibit beta cell apoptosis, but two lesser-known studies demonstrate that GLP-1 agonists do so almost instantaneously, and for a long period of time. A 2003 study in Diabetes – which DeFronzo hailed as “the most important study in diabetes literature in the last 15 years” – demonstrates that a single dose of liraglutide (7.5 ug/kg) restores type 2 diabetes patients’ beta cell insulin response to normal in just eight hours. To complement this, Dr. DeFronzo pointed to data from a three-year study that demonstrated that patients treated with exenatide present normal insulin secretion in response to a 15mM clamp – a 319% improvement over insulin glargine (p<0.0001). Dr. DeFronzo concluded his talk by passionately underscoring the critical importance of the GLP-1 agonist drug class. “If you understand what’s in the literature, there is no excuse why this wouldn’t be first-line therapy for type 2 diabetes.” 

• Dr. DeFronzo also highlighted the promise of emerging combination approaches that merge GLP-1 agonism with action on other gut hormones. Among these are GLP-1/glucagon, GLP-1/PYY, GLP-1/GIP, and GLP-1/oxyntomodulin. The most advanced GLP-1 combination approach is the GLP-1/glucagon dual agonist class, which has been shown to reduce body weight, increase energy expenditure, and improve glucose tolerance better than GLP-1 alone. According to Dr. DeFronzo, such combinations have the greatest potential to “outthink the gut hormones” to treat both diabetes and obesity. We’ve seen significant and increasing interest in GLP-1/glucagon dual agonists over the last year, including plenty of buzz surrounding gut hormone dual agonism at this year’s ADA.

7. In his presentation on cardiovascular outcomes trials (CVOTs) in diabetes, Dr. Jay Skyler (University of Miami, FL) expressed surprise that the SUSTAIN 6 CVOT topline results were able to claim a statistically significant reduction in cardiovascular risk for Novo Nordisk’s once-weekly GLP-1 agonist semaglutide with just 250 events. He noted that it is difficult to determine statistical superiority with a low number of events because the confidence intervals are usually too narrow – he pointed out that intensive therapy just missed statistical significance for cardiovascular risk reduction in the UKPDS trial and the beneficial effect was only evident after many years of follow-up during which more events accumulated. The SUSTAIN 6 trial accumulated 250 MACE events (cardiovascular death, non-fatal MI, and non-fatal stroke) in 3,300 subjects over a little over three years and two years of treatment with semaglutide. Dr. Skyler suggested that he would prefer a p-value of 0.025 or less, as a result with the traditional p<0.05 threshold for significance only has a 50% probability of being replicated in a second study. We do think conventional acceptance is the traditional p<0.05 though we are always eager to hear various opinions on statistical results. Dr. Skyler noted that to demonstrate a hazard ratio with an upper bound of the 95% confidence interval of 1.3 or less with 80% power and achieve a p<0.025, 456 primary cardiovascular events are needed. For 90% power, 611 events would be needed to demonstrate non-inferiority with a hazard ratio with an upper bound of the 95% confidence interval of 1.3. To demonstrate superiority, even more events would likely be needed. As we understand it, a number of others would agree that SUSTAIN 6 trial results would not support a label indication for cardioprotection and we expect Novo Nordisk will have to conduct a larger CVOT powered for superiority – indeed, of course, they are developing an oral once-weekly version of the drug. We’re certainly looking forward to EASD in Munich in September hear more about the full results for this trial.  Dr. Skyler indicated that we will need to wait until the full presentation and the paper to understand what was meant by the press release.

• Dr. Jay Skyler suggested that it’s “probably an overreach” for the FDA to continue to require CVOTs for all diabetes drugs. He suggested that, instead, CVOTs should be selectively conducted for certain drugs and focus on specific questions that need to be addressed for the specific drug rather than uniformly required for all new diabetes compounds. Furthermore, Dr. Skyler argued that CVOT design should be more focused, aimed at demonstrating a preventive benefit in people who don’t have acute coronary syndrome or pre-existing cardiac disease, as this is the population that would benefit most from cardioprotection. Dr. Skyler suggested the neutral results of ELIXA study for Sanofi’s GLP-1 agonist Lyxumia (lixisenatide) were the results of a less than optimal study design that enrolled patients with acute coronary syndrome. He then suggested the study organizers may have been seeking a short trial with a fast event accumulation rate, rather than an opportunity to demonstrate a potential meaningful cardioprotective benefit for the drug – this was effectively pointing out in our view that this was a trial to prove safety rather than cardioprotection, which seemed reasonable. 

Top Diabetes Technology Highlights

1. Dr. Ramachandra Naik (J&J, Wayne, PA) shared the most detail ever on the planned pivotal study design for J&J’s automated insulin delivery device (a hypoglycemia-hyperglycemia minimizer) with Dexcom’s G5 CGM. Dr. Naik confirmed that the trial will begin in the fourth quarter of this year, will test the device in 200 adults, adolescents, and pediatrics down to two years-old (though children will only be enrolled once it has proven to be safe). The study will include a several-day in-clinic assessment, followed by a three-month home portion, and will take place at 25-30 centers in the US. A post-approval commitment will be necessary, though no details were shared. Animas confirmed with us that, like Medtronic’s MiniMed 670G pivotal study, this will be a single-arm trial. The inclusion of pediatrics in Animas’ pivotal is particularly notable, as the 670G pivotal only included children down to 14 years of age; Medtronic has already started pediatric studies, but it’s unclear if these will be included in the FDA label out of the gate. These study design details confirm that J&J is IN this field, and could potentially be second to market with a commercial hybrid closed loop system. The timing also confirms what we last heard from Chief Medical Officer Brian Levy at ADA and J&J’s May Medical Device Business Review, which both slated the pivotal to begin later this year. It’s good to see there is now FDA clarity on the path forward. J&J’s Medical Device Business Review called for a US launch of this AID system by November 2017, though J&J updated the expected launch timing at ADA to November 2017-May 2018. The Via was not discussed, though we think that has enormous potential to do well.

2. Dexcom VP of R&D Jake Leach shared several notable pipeline updates in his overview of Dexcom technology: (i) an enhanced G5 app will launch “as soon as possible,” adding backfill of up to three hours of missing data when the phone and transmitter are separated (e.g., when showering) and a mute switch override (e.g., to enable hearing the low alarm when the phone is on silent mode); (ii) the touchscreen receiver was submitted to FDA last month, on par with 1Q16 timing expectations; (iii) an FDA submission of the Android G5 app is expected in the next month or two (making the previous “2016” launch expectation tight but doable); (iv) the Android G5 app will be compatible with Android Wear smartwatches; (v) the next-gen insertion system and 50% smaller transmitter will be filed with the FDA in the “next couple of months” (a 2016 launch was the previous goal, making this timing possible but also tight); (vi) the FDA has approved the IDE for the G6 pivotal trial, which will begin shortly (Dexcom’s largest study ever); and (vii) the Verily partnership is on pace to launch a first-gen product in 2018 (“within two years”). Mr. Leach briefly referred to Thursday’s Advisory Committee meeting on whether G5 should be approved for insulin dosing – though he avoided speculation on potential outcomes, he did comment that the device is extremely accurate and approved for this indication in Europe.

3. Dr. Satish Garg (Barbara Davis Center, Aurora, CO) announced, for the first time to our knowledge, his hope to publish results from the 670G pivotal trial extension phase by the end of 2016 (>1 year at-home use of the MiniMed 670G). This will be very scrutinized as a follow-up to the three-month results presented at ADA, and it will be interesting to see if the efficacy is sustained in the 80%+ of participants who continued on the device. To our knowledge, this is also the longest and largest group of patients that has continuously worn automated insulin delivery, representing ~99 years of patient exposure by our math (assuming 12 months per user and ~99 users in the extension phase). Many patients have told Dr. Garg that when they see the shield (indicating that the pump is in closed-loop mode), they feel safe – a cool reminder that user interface and icons can really drive psychological satisfaction with a device. A 64-year-old patient told him after the pivotal study, “There’s no way I’m giving this back; I need to keep it. My husband travels to Phoenix all the time, I really need this.” She was the first of nearly 100 patients in the pivotal trial to write the FDA and ask for continued access. Dr. Garg expressed gratitude to the FDA and Medtronic (who is funding the extension arm) for allowing the study to continue, both for the sake of collecting data (during the FDA review – a big win!) and for the patients. Medtronic’s Dr. Fran Kaufman expressed high confidence in the potential speed of approval, guaranteeing that Medtronic will do everything it can to allow its hybrid closed loop to be approved by next summer. As a reminder, the MiniMed 670G was submitted to FDA at the end of June, and Medtronic’s pre-ADA Analyst Meeting maintained the expectation to launch in the US by April 2017, timing they had originally given a year earlier.

4. In a refreshingly titled talk, “Regulatory Promotion of Technology Innovation,” FDA’s Dr. Courtney Lias picked up where she left off at the NIH Artificial Pancreas Workshop, repeatedly asking for audience input as to how the FDA can make regulatory processes easier. She affirmed that the Agency is prioritizing: (i) device consolidation and convenience via mobile apps; (ii) the optimization of pivotal trial design (namely pre- vs. post-market considerations); and (iii) device interoperability (the FDA announced an internal initiative to encourage interoperability just last week). Touching on the topic of pivotal studies, Dr. Lias reminded the audience that the FDA is flexible and encouraged early communication to ensure that studies have an appropriate balance of pre- and post-market data collection – not just for approval, but also to improve odds of early reimbursement. We were very glad to hear reference to regulatory studies having even a potential impact on reimbursement and FDA wanting to influence that positively, since historically, there was often a much clearer line between the two. Medtronic, of course, went with a super-fast single arm, three month trial to demonstrate pre-market safety, to be followed by a planned 1,000-patient post-market outcomes study. While some are concerned every company will have to follow this path, the FDA is clearly flexible on this topic. It was valuable to hear her acknowledge in Q&A that study design can really drive patient access and cost, and the FDA can work with companies to encourage proper design (hopefully with CMS and/or other payer input!). Following her enthusiasm last week, Dr. Lias also continued to emphasize that device interoperability will actually improve the safety profile of devices and speed up the regulatory process without sacrificing cybersecurity. We will be interested to learn more about the internal initiative, what form it will take, and how that might impact the diabetes device landscape. In Q&A she touched on the OpenAPS movement, noting that as the movement expands beyond personal use, the risk increases and regulation is more likely to be enforced; on the plus side, she highlighted that "these movements apply pressure on the system to approv devices faster, so that we know that the technology that people are using is safe." We completely agree and recall that the Dexcom G4 Share receiver was approved quickly as the Nightscout community gained steam. We hope the same is true of automated insulin delivery. More highlights from this talk are below!

5. In two talks on closing the loop, JDRF’s Dr. Aaron Kowalski highlighted the need to manage unrealistic expectations, to learn from the commercialization of CGM, his position as a “glucagon realist,” and that he is still a big believer in amylin (AstraZeneca’s Symlin). Dr. Kowalski’s main message was that, even though hybrid closed loop is on the horizon, the community has a great deal of work to do to ensure access, uptake, and efficacy. First, he said (as has been often emphasized over the last 6-12 months), we must prevent unrealistic expectations – patients have to know that these systems will not be a cure, and will require plenty of interaction. He told healthcare providers in the audience to prepare for complaints about early systems, particularly in often demanding early adopters. Dr. Kowalski also recommended that we learn from the evolution of CGM – what can be learned from early CGM device use that automated insulin delivery might incorporate? Many of the same issues will almost certainly come up with automated insulin delivery (“It doesn’t work like I expected”), and healthcare providers should be prepared to address them. Above all, he said making devices smaller and more convenient will be critical for adoption – areas where JDRF is focusing resources. Dr. Kowalski also touched on reimbursement (characterizing this as the diabetes community's “bugaboo”), expressing slight apprehension because of the lack of large efficacy trials and appreciation for outcomes beyond A1c. Fortunately, a number of groups – Medtronic, UVA/TypeZero, Cambridge – are all launching longer efficacy trials soon. Moving forward, Dr. Kowalski highlighted his “Diabetes Scorecard” (Diabetes Care 2015) to judge the value of artificial pancreas systems beyond A1c – scorekeepers are the people with diabetes (and loved ones), clinicians, and payers, and the categories are glycemic control, “burden”, and value.

• Regarding glucagon, Dr. Kowalski quipped, “I’m not a glucagon skeptic, but a glucagon realist. It’s not as simple as you may think.” The frequently-used analogy, he elaborated, that an insulin-only artificial pancreas is like driving a car without brakes, is not appropriate – “when you drive a car, hopefully the brake works every time you press it.” Dr. Kowalski argued that glucagon, if included, should only be used as a rescue – relying on it for control could be dangerous if the glucagon set becomes occluded, he emphasized. He was careful to point out that he isn’t totally opposed to the use of glucagon in closed loop, but that the health economics aren’t favorable at the moment – he asserted the marginal cost is currently greater than the marginal benefit over insulin-only. We aren’t sure what assumptions Dr. Kowalski is using (no long-term large trials yet) or what may be appropriate as the field progresses and we are eager to see the differences in the systems. With Lilly Diabetes behind the Damiano group, of course, there may be flexibility in how the glucagon element is priced, though we have no information on this.   
• “I’m still a big believer in amylin,” shared Dr. Kowalski, and JDRF is currently working on use of the hormone with “a number of companies.” Dr. Kowalski acknowledged that pramlintide has not been wild success commercially – adding burden without a dramatic increase in glycemic control or quality of life. However, JDRF has been looking at single cartridge, co-secreted insulin+amylin, and so far, it seems promising for restoring normal physiology. This echoed Dr. Steve Russell’s comments at the NIH Workshop two weeks ago and what we’ve heard anecdotally from patients pumping amylin for years.  Dr. Kowalski remains optimistic for the potential, particularly if insulin and amylin can be co-formulated. He did not discuss the cost considerations, which presumably would be similar to the costs constraints he discussed related to glucagon (see above). As usual, Dr. Kowalski raised extremely valuable questions, such as how much glycemic/user experience benefit would another hormone have to add relative to additional cost or complexity it brings to the system? How would payers reimburse for a multi-hormone closed-loop and weigh the glycemic and quality of life outcomes?

Detailed Discussion and Commentary

Challenges in Diabetes Management

Challenges (and Opportunities) in Inroducing Closed-Loop Systems

Aaron Kowalski, PhD (JDRF, New York, NY)

In his final talk of the conference, the in-demand Dr. Aaron Kowalski spoke about key upcoming areas of focus in automated insulin delivery: expectation management, learning from CGM’s commercialization, reimbursement, and DIY innovation. To begin, he highlighted the short video shared by Dr. Jeremy Pettus and Dr. Steve Edelman at this year’s ADA – the humorous skit depicts the year 2026, in which closed loop technology has completely solved diabetes, forcing healthcare providers like Dr. Edelman to take up jobs at Starbucks and shutting down organizations like JDRF. This is not how Dr. Kowalski imagines it will work, as the challenges of artificial pancreas implementation will be vast and variable. First, as the technology nears the market, patients must understand that the artificial pancreas is not a cure – it will not restore euglycemia (because of the lag with subcutaneous sensing and insulin infusion). In addition, the systems will require two insertion sites, mealtime bolusing, infusion set and sensor changing, fingerstick calibrations (for now), and a working knowledge of how to manage diabetes in case the system fails. Will patients appreciate the early hassles? Dr. Kowalski wondered if CGM’s commercialization can provide some lessons. The discontinuation rate with early sensors was alarming (no pun intended) – for many patients, the added benefit of using the inaccurate early devices didn’t outweigh their cost and hassles. Dr. Kowalski asserted that healthcare providers should be prepared to hear similar grievances from early artificial pancreas users. [We assume, however, that first-gen AP systems will add more value than first-gen CGM devices, even in early adopters who only use them overnight.] He also touched on reimbursement (the diabetes community's "bugaboo”), expressing slight apprehension because of the lack of a large efficacy trial and appreciation for outcomes beyond A1c. Fortunately, a number of groups – Medtronic, UVA, Cambridge – will all be launching longer efficacy trials soon. Moving forward, Dr. Kowalski highlighted his “Diabetes Scorecard” (Diabetes Care 2015) to judge the value of artificial pancreas systems beyond A1c – scorekeepers are the people with diabetes (& loved ones), clinicians, and payers, and the categories are glycemic control, “burden”, and value.

• “I’m still a big believer in amylin” and JDRF is currently working on use of the hormone with “a number of companies.” Dr. Kowalski noted that pramlintide has not been wild success commercially – adding burden without a dramatic increase in glycemic control or quality of life. However, JDRF has been looking at single cartridge, co-secreted insulin+amylin, and so far, it seems promising for restoring physiology. This echoed Dr. Steve Russell’s comments at the NIH Workshop two weeks ago. Dr. Kowalski remains optimistic for the potential, particularly if insulin and amylin can be co-formulated. He did not discuss cost at length related to the combination.
• Dr. Kowalski asked “What if devices were less ‘medical’?” He pointed to DIY hacking solutions, such as bolusing from an Apple Watch (as we saw from Mark Wilson at the D-Data Exchange), as a great way to minimize the psychosocial burden of diabetes – patients may feel more comfortable managing glycemia from a non-medicalized device. The big unknown we’ve detected is whether industry will harness the innovation in the DIY community and incorporate it into commercial products.
• Dr. Kowalski’s comments on education and expectation management remind us of a poster presented at ADA last month, in which 42% of individuals at a workshop thought that the artificial pancreas was an organ grafting procedure. Worries about provider education were also a clear theme at the NIH AP Workshop two weeks ago. The field must clearly prepare and do a better job of educating those new to technology so they know exactly what to expect from automated insulin delivery. The term “artificial pancreas” is a bit confusing, and we assume many/all commercial products will steer away from it (e.g., MiniMed 670G “hybrid closed loop,” Animas “Hypoglycemia-Hyperglycemia Minimizer,” etc.).
• After the session, a diabetes nurse from Missouri approached Dr. Kowalski and asked him “Why can’t patients talk to their pumps?” What a great idea, particularly as voice recognition becomes more popular and widely used in consumer technology (Amazon Alexa, Apple’s Siri, etc.) A voice-responsive pump would certainly improve the quality of life of visually-impaired patients, and also eliminate pump interface hassle for those who have normal vision. Dr. Kowalski appreciated the suggestion, but had not heard of any companies working on this technology. As diabetes devices move to the smartphone, this seems like a logical foray for the user interface to expand to: “Hey Siri, how much insulin on board do I have?” What may be a challenge is the loftier, “Hey Siri, please give me a three-unit bolus.”

Panel Discussion

Dr. Satish Garg (Barbara Davis Center, Aurora, CO): Do we have any justification for a rise in the cost of insulin?

Dr. William Herman (University of Michigan, Ann Arbor, MI): Not that I’ve been able to figure out.

Dr. Garg: I don’t understand why our government isn’t able to negotiate the way Europe is.

Dr. Herman: Under Medicare Part D, the government cannot negotiate prices. There is a real lack of transparency surrounding insulin pricing. It’s not well understood and it’s unclear where the money is going, whether it’s to manufacturers, rebates, or something else.

Dr. Aaron Kowalski (JDRF, New York, NY): I’m going to play devil’s advocate a little bit. I remember the first time I went to my college pharmacy, the pharmacist charged me $15. I thought that was egregious! I threw my arms up and stormed out because of my outrage. But the counter-side to this debate, which I think is important, is the incredible cost of drug development. I’m impressed by my experience on Afrezza – there was $2 billion spent on that drug. The company is in dire trenches right now. We say we want better insulin (i.e. glucose-responsive insulin that lowers the risk of hypoglycemia), and that’s going to cost $2.5-$3 billion to bring to market. Drug development is hugely expensive and it’s a high-risk endeavor. I absolutely agree that we need more transparency, and I don’t understand why insulin costs so much, but I empathize with these companies because drug development costs are much too much. Is it the FDA causing drug development costs to be so high?

Dr. Herman: Why was insulin glargine introduced at $40/ vial, which presumably covered R&D, and then made more expensive (it’s now $300/ vial)? I’d also like to point out that the DCCT was done with animal insulin and with human insulin. There is a benefit to newer insulins, but at what cost?

Dr. Garg: The University of Colorado created a foundation that allows patients with type 1 or type 2 diabetes to get any sensor, pump, insulin, or drug they need for free. There is zero co-pay. For the past four years they’ve been coding this experiment, and on average they’re not spending more than $10 million/year. Next year will be the fifth, and we’re planning to analyze this data to answer one critical question: If you take away cost barriers, does it improve patient outcomes? The reason I bring this up is because in Britain as well as Europe, none of these things cost anything. Some of our patients who visit Euorpe, they walk in to a clinic there because they didn’t have enough insulin, and clinics just give them insulin and they’re shocked that it’s free. So I don’t understand.

Dr. Brian Frier (University of Edinburgh, UK): It’s not quite that simple. Patients don’t pay for any drugs in NHS, but the health authorities very carefully watch the budget. If they think drugs are too expensive, they move to strict restrictions on their use. Insulin degludec launched in the UK some time ago but uptake was low because of cost. Now in desperation, Novo Nordisk has dropped the cost by 30%.

Dr. Garg: Artificial pancreas will be available next year, but it is a big deal to start anybody on it. Even in our latest study, there was a lot of handholding. It’s not just something you just slap on. Many patients are confused about the lack of basal rates – there is so much education to do. I could go on and on, we have two pages of comments for Medtronic and the FDA to consider. It’s extremely important, and there needs to be some sort of artificial intelligence so it learns from the last six days how much patients have bolused, and then adjusts the basal/micro-boluses accordingly.

Dr. Kowalski: A number of the algorithms do adapt over time. Medtronic is a PID controller, which is more calculus-based. Other groups’ algorithms (Cambridge, UVA, bionic) do adapt and can get into activity, meal, and menstrual patterns, and get more sophisticated on a person-to-person basis.

Q: I was thinking about how my CGM and how little reaction it has to the tissue in my body, compared to the pump where you get inflammation, etc. when you leave it in. The main barrier seems to be the subcutaneous administration of insulin. Are any efforts being made to subvert the natural defenses so insulin can be infused in a more physiologic way?

Dr. Kowalski: This is a great question, and a high priority for us at JDRF. The challenge with the implantable is most people don’t want to do it. I know team from Roche is here, but most people don’t want a permanent implantation. So the question is, how small does it need to be for people to adopt it? On the counter side, we know that’s going to take some time, so we’re looking at what we can do right now to improve insulin infusion? Looking at ways to make catheters better, insulin that’s more reliable. I don’t know you’ll get same physiologic benefit as peritoneal delivery, but if we had better ways to infuse insulin right now we could maybe get there.

Dr. Frier: Why is the absorption of short-acting insulin delayed as pregnancy progresses?

Dr. Sarit Polsky (Barbara Davis Center, Aurora, CO): They weren’t able to sort it out in that study. There’s only one I’ve seen that actually looked at that, with closed loop. There’s a question about slight edema in peripheral tissue as pregnancy progresses, and of course increased risk for pre-eclampsia. t’s a great question. Thankfully, at this point, we have data to support that bolusing earlier can help.

Dr. Herman: Are there any issues related to altered GI motility during pregnancy? Earlier bolusing might cause hypoglycemia?

Dr. Polsky: I haven’t seen any data to support that. Early gestation women are hyper motile, later on as pregnancy progresses usually we don’t’ see that. Occasionally if someone has had more advanced diabetes complications going into pregnancy, maybe we’ll see that, but thankfully women tend to be on the lower end and that hasn’t happened.

Dr. Garg: I wonder if some of that has to do with hormones. As pregnancy advances, perhaps the counter-regulatory hormones go up? I agree with Dr. Polsky that some of it must have to do with changes to subcutaneous flow.

Q: We see quite a few women in our clinic with gestational and pre-existing diabetes. We treat our gestationals to targets that you mentioned, but for pre-existing diabetes we treat pre-prandial to <100 as opposed to <90. I’m curious about the difference between pre-prandial targets for gestational and pre-existing diabetes? Why the difference?

Dr. Polsky: There’s actually a lot of controversy in the pregnancy world about what those fasting and pre-prandial levels should be. It’s caused a lot of whiplash, to be honest. A HAPO study looked at different levels of glucose below hyperglycemia ranges and adverse events. What they could see is that 75 mg/dl is a reference range, any level of hyperglycemia above that you see an increased risk in adverse events (c-section, macrosomia, things like that). So the question was, what’s the threshold level, and there was none. There’s lots of groups working on this, international diabetes/pregnancy study group, WHO have their own ideas. We can argue about 90 vs. 100 mg/dl targets but they’re pretty similar for most people and pretty difficult to achieve for every meal regardless. So the question is how low can you go without significant hypoglycemia, and that’s what you should aim for.

Dr. Garg: On the EMPA-REG and LEADER studies – some people are overwhelmingly supporting the use of these drugs based on this data. But contrary to all, Dr. Herman presented that some of these advantages might actually be related to other risk factors. If you can just imagine that Dr. Ratner and Dr. Skyler were here – how would you refute this?

Dr. Herman: With great difficulty – they’re quite persuasive. But again, I don’t think it’s unlikely that a 0.5% reduction in A1c and other benefits could result in a 12-14% reduction in CV outcomes. The question is how easily could this be achieved by titrating the dose of ACE inhibitor or adding sulfonylureas to the existing regimen.

Dr. Garg: Are these CVOTs really needed? We’re adding billions of dollars and enrolling hundreds of thousands of patients in these trials. Some patients said at this conference last year that CVOTs are the best thing that ever happened, but others vehemently disagree.

Dr. Herman: A CVOT would be a useful trial if they signaled increased risk. It’s difficult to interpret a benefit without having equal control of multiple CV risk factors. I think there is going to be an advantage to medications that cause weight loss, improvement in glycemia, blood pressure, and other cardiovascular risk factors.

Dr. Frier: If I understand you correctly Bill, you’re saying that people in these trials were selected to have advanced CV disease, so while there was a tremendous benefit in these groups, if you try and apply the findings to the whole type 2 diabetes population the benefit is not going to be as profound. So we should be cautious about changing algorithms.

Dr. Herman: Absolutely. We have to apply these findings to patient populations included in these trials. But these results should not imply that for those failing metformin three years into the natural history of type 2 diabetes, adding one of these medications is going to have same result.

Dr. Frier: And when you add the economic assessment, it becomes even less compelling.

Comment: I used to work on designing CVOTs. You have to realize that companies are trying to make lemons into lemonade. In order for an NDA to be accepted based on 2008 FDA guidance, companies have to demonstrate the drug is no more than 1.3x more dangerous than placebo in terms of coronary adverse evens. They only have the time from the start of program to finishing efficacy studies, and they have people with a very high rate of CV outcomes who you’re least likely to demonstrate benefit in. CVOTs need 122 MACE events to demonstrate the necessary 1.8 hazard, then the post-approval commitment to demonstrate no more than 1.3 needs about 300 MACE events for that. To demonstrate superiority, or a hazard ratio <1, they need 600 MACE events. For the sponsors to demonstrate safety, they have to go to high-risk people, and then it’s hard to show efficacy. The fact that efficacy has been shown is very, very powerful.

Q: I believe there are biosimilar insulins (glargine) that have been or soon-to-be approved in other parts of the world. What has that done to the brand of glargine?

Dr. Herman: Biosimilars have been slow coming to market. Lutz Heinemann gave a very nice presentation on the limitations and lingering questions as to how similar they really are. The drop in price has been much less in price than one would hope or predict, especially after the experience with generic drugs.

Dr. Frier: It’s about 20-30% reduction in price.

Q: Certainly we need research, but how much actual scientific research is being done by Eli Lilly and Sanofi? How does their research budget actually compare to their marketing budget? Say Dexcom cut the price for CGM in half, but got more insurance companies to agree to cover it. Would that improve things?

Dr. Kowalski: I’m sure marketing budgets are massive, but I know for a fact that research budgets are just huge. Full disclosure: JDRF has partnered with Sanofi and Lilly on glucose-responsive insulin and they’re spending a tremendous amount of dollars on next-generation insulin. You get this tunneling effect into phase 3 costs. My experience with the big three insulin manufacturers is that they’re spending a ton on research. Pricing of sensors is an economics question that I can’t profess to understand. Our JDRF philosophy is that competition can help drive some of this. You could say competition hasn’t worked for insulin, but I hope at some point that it can help.

Dr. Herman: I don’t have a real answer to your question, although I must say I’ve been quite amazed that the cost of U500 human insulin is up to almost $2000/vial.

Patient/Provider Question & Answer Panels

In this patient-centric panel, a beloved and highly-anticipated tradition at Keystone, four of Dr. Satish Garg’s (University of Colorado, Aurora, CO) patients took the stage to discuss their unique stories of life with diabetes. Their testimonies provoked laughs, tears, and a standing ovation from a packed audience, and this emotional session concluded with a thoughtful and illuminating Q&A session between Courtney, Kimberly, Patricia, and Troy and the audience of providers.

Patient Testimony: Courtney

I had just turned 14 when I was diagnosed with type 1 diabetes – this was in the early 80’s. I started to notice a difference in my moods in late teens and early 20s. I had no idea what it was – in those days, mental illness was never talked about or suggested. By the time I reached college, I was extremely depressed and was sleeping 12-14 hours/day. Everything was centered around that. I also experienced extreme highs where I hardly ever slept, and was always awake with rapid and racing thoughts. My mood went along with my diabetes, and my A1c was very affected by my bipolar illness. My numbers went high especially in periods of depression, because depression slows you down. You don’t test and you miss shots because you’re in your own world.

I couldn’t finish college because of the lack of motivation that came with my depression, so I moved back home. I was in and out of jobs, and bipolar kept me in those patterns. I was lucky to have very supportive parents, but they didn’t know what to do or how to help me. Eventually I was hospitalized, and they misdiagnosed me with major depressive disorder. The antidepressants made things worse, and I was finally diagnosed with bipolar in my early 30’s. Diabetes alone, much less a mood disorder, can be very difficult on marriage. It has been difficult at points, but my husband is my advocate and my provider, and he’s very aware of my symptoms that take place.

Two and a half years ago I became completely nonfunctional and was hospitalized again. I started electroconvulsive therapy (ECT) and it saved my life. Urge your patients to do this if it fits for their particular situation. I hadn’t felt that good, that normal, since I was a little girl, and it was amazing. Having that kind of stability in moods – oh my gosh! It helps control your diabetes a lot better. My diabetes made ECT treatment complicated since it requires anesthesia, so I started transcranial magnetic stimulation (TMS), which is an amazing thing, because treatment for mental illness used to be only talk therapy and meds. The fact that other options exist, gives me hope in and of itself, and that’s a wonderful thing.

My diabetes and my bipolar are the two worst things that I deal with, because you don’t know if the ups and downs in blood sugar are causing ups and downs in mood or the other way around. You don’t know what came first. When I am in depressive or manic episodes, it is difficult to keep diabetes right there in front of you, since diabetes changes every minute of every day. Diabetes is manageable but it is affected by so many things, it’s not in a box. I’ve had many surgeries and hospitalizations, and all of these things contribute to messing with my diabetes. I am very blessed to have help with the medical teams I do. Dr. Garg and his staff are amazing, and I’m thankful for my family and friends to get me through it day by day.

Patient Testimony: Kimberly

I’m 46 years old and I’ve had type 1 diabetes for 40 years. In the early days, when I was diagnosed, we did urine testing and my insulin was derived from beef cattle. My A1c was 10%, and that was considered good! Over my teen years, and in college, human insulin was developed. This opened the door to control blood sugar better. My A1c went down to between 8% and 9%, and that was considered pretty good.

Dr. Garg became my doctor when I was 16, so we’ve worked together for a long time! In college I didn’t take care of myself as well; I wasn’t good about blood sugar testing. I got married at 21, and pregnant at 22. I went in for a checkup and my A1c was 13% – that was quite the wakeup call. We worked very strictly to get my blood sugar down to 7% in just a few months, which was amazing. But at 20 weeks of the pregnancy I became pre-eclampsic and gained about 40 pounds. Two years later, I was pregnant again. I actually found out because I had to take a pregnancy test before starting pump therapy! During that pregnancy my A1c was in the 5s, and I was told it was almost too low. Despite trying traditional methods of diet and exercise I gained another 20 pounds.

I had a cardiac event at age 38 – a 95% blockage in my right coronary artery. I went to cardiac rehab for several weeks, and learned about diet and exercise. I lost about 40 pounds with the diet and exercise I learned on the cardiac rehabilitation program. But I’ve gained 20 pounds since then, so I’ve vacillated back and forth.

Many of the concerns people with diabetes have is about weight and how to manage it best with diet and exercise. The other thing we deal with is nighttime lows. Midnight comes around, your blood sugar is super low, and you eat everything in the refrigerator – I’m tired of eating those glucose tablets! The other thing that’s concerning is getting support from people who know what you’re going through. People with diabetes are not an equation, and all of our bodies are different.

Patient Testimony: Patricia

I’m a physical therapist and I was diagnosed with diabetes in 1996 at age 26 (you can do the math to see how old I am!) In high school I was heavy but not obese. Gradually, when I decided not to go to college, I gained a lot of weight because I was eating a lot of fast food. I decided I needed to go to college after I began charging groceries to my credit card. And in college I actually lost weight…and was thirsty as all get-out. One of my roommates was a nursing student and she had just studied the chapter on diabetes and the warning signs. I heeded her advice, got tested, and my A1c was 11.6%. I was sent immediately to diabetes education classes. My doctors told the diabetes educators – but not me – that I might be type 1 and not type 2, but to still do the classes anyway. I was spilling ketones in my urine, so I began doing injections. Insulin made me gain back 20 pounds – which I was not happy about – but after the dehydration and the fatigue, you get that first insulin injection and you feel like you got your life back.

That was amazing, but the weight gain wasn’t. I gained more weight and my insulin doses increased to 200 units of per day between long-acting and short-acting. I weighed 263 pounds and made the decision to have the lap band procedure. I quickly lost 33lbs, but then plateaued. I kept very strictly with the program, but nothing got me out of that plateau, and I slowly gained all of that that weight back. A proper lady would never tell you what she weighs, but luckily for you, I’m not a proper lady! I weigh 270 pounds today. Because of my weight I had to get a knee replacement, and I know that my other knee will need it at some point too. You think that would be enough motivation: ‘just eat and exercise and the weight will come right off!’ But I failed at every diet out there. I decided that I’m never going to be thin. This is how I am and I have to come to terms with that.

Today I take 145-160 units a day through my insulin pump. With that much insulin I find it very difficult to lose weight. It’s a vicious cycle: insulin makes you gain weight and that makes diabetes more difficult, and then you need more insulin again. Hopefully with a bit more time and two good knees I can start exercising again. And enjoying it like I used to.

Patient Testimony: Troy

Before I begin, I just want to say that before I came up here I was taking care of my low blood sugar. We deal with this 24 hours a day, 7 days a week. Diabetes is a constant struggle.

I didn’t start off overweight. When I was younger I was very athletic; I played football at Texas Tech. I’m just an eater! My problem is that I like to eat. Looking back, I was 35 when I was diagnosed and met Dr. Garg, who I call “The Hammer!” (Laughs). Every time I went in for a visit it was a screaming match! (Laughs.) He’s my hero but what he was trying to tell me was that I had to lose weight. After about 2 years of him banging on me, I decided to get bariatric surgery. It was the best decision. There’s two real key things to know if you have a patient get bariatric surgery. For one thing, at first Dr. Garg wasn’t allowed to come in and help me get my blood sugar down post-surgery. The second thing is to make the patient get up and walk immediately following the surgery. Get active immediately!

I was always in shape, and even went into body building for a time, but as soon as I was diagnosed with diabetes the weight came on and wouldn’t stop. For everybody that’s overweight, bariatric surgery saved my life. I used to weigh 450 pounds. I’ve lost over 100 pounds with the surgery and I’ve kept it all off.

Dr. Satish Garg (Barbara Davis Center, Aurora, CO): I don’t often say this, but I was so scared about how to manage his diabetes after bariatric surgery. The surgeon and I worked together to figure it out, and the outcome has been amazing.

Questions and Answers

Q: Courtney, can you give us some insight into how the ECT and TMS impacted your blood sugars? During the days, acutely.

Courtney: My A1c runs from 6.8%-7.2%, but right before ECT I was at 8.6% because I was so depressed. That was 2.5 years ago, and it was done under general anesthetic, which is a mess with type 1 diabetes. I typically ran high going into the procedure, but it was mostly out of my hands.

Dr. Garg: Being in a psychiatric hospital, you’re taken away from your insulin so it’s very difficult. Troy, how did you manage your diabetes when you were in the hospital for your bariatric surgery?

Troy: I monitored myself very closely with my Dexcom. It was a constant hassle.

Dr. Garg: He managed it because the surgeon and I talked and decided to let Troy manage himself.

Troy: The hospital folks were very resistant to that! They tried to take away my Lantus. Without Dr. Garg intervening it would have been terrible.

Q: Troy and Patricia, I find it interesting that both of you tried the lap band first. How did you make the decision about which procedure to undergo? And if you were talking to someone with type 1 diabetes and considering surgery, what would you say?

Patricia: I chose the lap band because I work in the hospital and have seen a lot of bad outcomes – infection, feeding tubes – from RYGB. At the time that I did it, all the research said that the weight loss is the same between the two procedures, but just takes longer from the lap band. Going into the surgery, I didn’t have a handle on what an emotional eater I was. The guy giving me the psychological exam was clearly reading the questions the insurance company wanted to have asked. You’d have to be an idiot not to know what answers he wanted! So I told him what he wanted to hear. I wish there had been a better pre-surgery psychological exam. And more than one of them.

Troy: I would go straight to the Roux-en-Y. I don’t know why I wasted my time with the lap band for 6 years!

Q: Did anyone try the weight loss-aiding diabetes medications, like GLP-1 agonists? And what do each of you think is the most important diabetes technology you have – what would you bring with you to a desert island?

Patricia: I had tried metformin and phentermine and I saw only minimal success with those. Atkins got me to lose about 20-30 pounds but I couldn’t go much beyond that. For technology, I would bring my insulin pump and CGM combo, it is amazing. Dr. Garg has promised me the closed loop system once it’s ready!

Kimberly: I’m on metformin and have been for 15 years. I lost a lot of weight initially, but that was probably from all of the nausea the drug gave me! It’s good for my blood sugar but I haven’t experienced much weight loss. In terms of technology, I’m old school so I would bring my syringes and my insulin. And my glucometer. I don’t like tapes – they pull out, they itch, they give me scar tissue, and the beeping alarms drive me crazy.

Troy: I would take every electronic deal I could possibly have! And in terms of drugs, I tried all those things – all the diets, all the drugs. I’ve lost 100lbs several times and always gained it back. The only thing that stuck was the Roux-en-Y.

Q: I’m wondering if you have any sort of observations about which kind of insulin therapy made you gain or lose weight.

Troy: The pump has helped me regulate things better. I used to be always over-injecting and under-injecting, it was a mess. With the pump it has been easier to keep the weight off.

Kimberly: For myself, when I was on the pump I experienced my lowest lows, which are really scary. For a while I was scared just to bolus because crashing is such a scary thing. I’m on Tresiba now and that’s pretty good; my blood sugar has been very consistent. But certainly with better control you will gain weight.

Q: We have a glucose management service at our hospital and we keep statistics to see how well we’re doing. In other hospitals glycemic control is very poor for patients coming in for other procedures, and we need to fix this.

Dr. Garg: Hospitals tend to put type 1 and type 2 diabetes in the same boat; type 1’s tend to get lost. At the Barbara Davis Center we tell the hospital the diabetes treatment guidelines we want the surgeons to do.

Patricia: I work in a hospital as well. We have on-site diabetes educators. If you’ve had diabetes for a long time – 16 years for me – they should leave you alone and let you take care of it yourself. I can do this better than you can, so just go away with your stuff! (Laughs, applause). But I do think they’re an excellent resource for the newly-diagnosed and the very poorly-controlled.

Comment: I work at a hospital and we work really hard to let patients stay on their pumps throughout their stay. We understand that you want to stay on your pump and we’re trying to work on that!

Corporate Symposium: Partnering to Improve Patient Care: A Technological Evolution in Insulin Delivery (sponsored by J&J)

Predictive Insulin Delivery Devices: Roadmap and Algorithms

Ramachandra G. Naik, MD (J&J, Wayne, PA)

Dr. Ramachandra Naik shared the most detail ever on the planned pivotal study design for J&J’s automated insulin delivery device (a hypoglycemia-hyperglycemia minimizer) with Dexcom’s G5 CGM. Dr. Naik confirmed that the trial will begin in the fourth quarter of this year, will test the device in 200 adults, adolescents, and pediatrics down to two years-old (though children will only be enrolled once it has proven to be safe). The study will include a several-day in-clinic assessment, followed by a three-month home portion, and will take place at 25-30 centers in the US. A post-approval commitment will be necessary, though no details were shared. Animas confirmed with us that, like Medtronic’s MiniMed 670G pivotal study, this will be a single-arm trial. The inclusion of pediatrics in Animas’ pivotal is particularly notable, as the 670G pivotal only went down to age 14 years; Medtronic has already started pediatric studies, but it’s unclear if these will be included in the FDA label out of the gate. These study design details confirm that J&J is IN this field, and could potentially be second to market with a commercial hybrid closed loop system. The timing also confirms what we last heard from Chief Medical Officer Brian Levy at ADA and J&J’s May Medical Device Business Review, which both slated the pivotal to begin later this year. It’s good to see there is now FDA clarity on the path forward. J&J’s Medical Device Business Review called for a US launch of this AID system by November 2017, though J&J updated the expected launch timing at ADA to November 2017-May 2018.

• We haven’t heard many specifics on the Animas hypoglycemia-hyperglycemia minimizer algorithm, but like the MiniMed 670G, it will only modulate basal insulin delivery (not issue automatic correction boluses). Animas’ system will obviously monitor manual boluses and on-board insulin as it makes dosing decisions, and we assume it will work well overnight. Still, there are many questions: How will it initialize? Will the algorithm learn and adapt over time? How will it compare to the MiniMed 670G algorithm in terms of robustness to errors, glycemic target, level of overnight and daytime control, etc.? One comment in Q&A suggested users “must” take a fingerstick for mealtime boluses, as the label for CGM mandates, though we assume many patients will simply enter the CGM value into the bolus calculator. Animas will obviously be closely watching Thursday’s Advisory Committee to see if Dexcom’s G5 will be approved for dosing insulin. Presumably an update there could have implications for this product’s label.

Partnering for a Patient-Centric Approach to Innovation

Damon Tanton, MD (Florida Hospital, Orlando, FL)

Dr. Damon Tanton led a highly interactive session that encouraged attendees to consider diabetes technology from the perspective of patient, provider, family member, educator, and manufacturer. He began by telling the story of his chance encounter with boxer Evander Holyfield near a vending machine filled with cuff links – the important detail of the story, shockingly, has nothing to do with meeting the former boxing champion. Dr. Tanton saw the cuff links and wondered, “What would it be like to have insulin pumps in vending machines?” From a consumer perspective, the idea makes sense, but he pointed out that so much more goes into pump use than simply putting the device on and pressing start. The same goes for the artificial pancreas – even in five years when there are (hopefully) multiple systems on the market, patients and providers will still have to consider sensor changes, communication errors, battery changes, insulin refills, infusion set changes, and other issues. Dr. Tanton remarked that “the technology is here; we just don’t have the cohesion” – presumably, he meant the whole ecosystem surrounding the core devices, including psychosocial factors, usability, device interoperability, prescribing and training hassle, etc. We like this term “cohesion” and believe it could drive adoption more than any clinical or medical factor – Will the average PCP be able to prescribe automated insulin delivery? When will the technology be such a no-brainer than MDI+SMBG users start switching in droves? The progress in algorithms, sensors, and real-world clinical research has been tremendous in the past ten years, but human and patient factors must now be the focus of the field. We really enjoyed the audience responses to this session’s interactive questions – see below. 

• In the interactive portion of the session, Dr. Tanton posed four discussion questions to the audience, yielding fascinating responses:

1.How would you want a device to interact with friends and family? Audience responses: We want patients, as well as family and friends, to be able to determine what data they get – they might not want all that data.

2.What is the role of a diabetes educator? Audience responses: (i) Assess patients’ knowledge and skillset; (ii) Teach families to use the product and empower them to manage child’s diabetes by making adjustments on their own; (iii) “I don’t want anything I prescribe to put more burden on the patient.”

3.What is the role of a healthcare provider? Audience responses: (i) Reading downloads and finding out patient’s compliance; (ii) Should be able to provide options to patients; (iii) Ensure patient is capable of learning, testing their own blood sugars, and complying with a regimen.

4. What is the role of the manufacturer? Audience responses: (i) Create a safe and reliable product; (ii) “Number one, access. Number two, safety.” (iii) Provide enough manpower in territories: “If you’re going to launch a great product, support us clinicians so we can use it effectively” [applause from audience]; (iv) Work with payers to ensure that products are covered so they can actually be prescribed.

Questions and Answers

Q: Can you give us an update on insulin patch product [OneTouch Via]?

Dr. Naik: So the product is already cleared by the FDA. We are currently doing a clinical outcomes trial. People on MDI tend to miss doses for a number of reasons; this lack of adherence can lead to poorer glycemic control. If they have insulin on demand, they can discreetly dose. We are testing whether or not adherence is better with it, and we hope to have results by early next year on our outcomes trial. We hope to have the product available in a limited manner by the end of this year, and then we will roll it out from there. That’s all I can say about the OneTouch Via. [Per J&J’s update at ADA, the OneTouch Via will launch in limited markets OUS in late 2016, with US launch to follow shortly thereafter.]

Q: How do closed loop insulin systems handle acute illness?

Dr. Naik: We don’t have any data to that effect, but the expectation is that it should work the way it works otherwise. We’re planning to do a CRC assessment. We will induce over- and under- insulinization (giving an extra dose, for example) and see whether the mechanism kicks in for all these extraneous circumstances. Once we’ve tested hypoglycemia, hyperglycemia, and these other extraneous circumstances, then there should be no concern for acute illnesses.

Q: I have no experience with pediatric diabetes patients, but do they require different algorithms for any reason?

Dr. Naik: We’ll test the pediatric population as well, but the algorithm should work the same way. After we show safety in adults, then we’ll move on to efficacy in lower age ranges.

Dr. Argento: All these algorithms learn. Illness is something people deal with all the time. Women on birth control are influencing their hormones every month, and the system should be able to compensate to a new normal. All of these algorithms are built to learn from recent experience.

Dr. Tanton: The same thing is true for Dexcom algorithms. That’s the measure of any good algorithm – it’ll be better on days three, four, and five than it was on days one and two.

Dr. Argento: The FDA requires a dedicated device right now, so forget about your iPhone. It’s a security reason, and some people don’t want their blood glucose controlled through their phone because their phone doesn’t always do what they want it to. [Editor’s Note: This is Dr. Argento’s person opinion, not Animas’ view. As we’ve followed this topic, the FDA has said very clearly that it is open to and excited about moving artificial pancreas devices to phones (see our NIH AP Workshop coverage); however, appropriate safety factors must be designed into the system. We assume the path to market for a dedicated medical device is easier, but it is not a requirement based on what we’ve heard from the Agency at recent meetings.]

Q: In these early devices, how is the pump going to be sensitive enough, or is it going to be too sensitive? What about varied activities, like rowing a boat for an hour?

Dr. Naik: The infusion set issue is real. Studies essentially use saline in healthy volunteers for short durations, so they’re probably not reflecting real world scenarios. We’re looking into innovations for infusion sets alongside our partners. It’s high-time – innovation has to come not only on a systems level, but at the infusion set level.

Dr. Argento: The effect of infusion sets is highly variable. For some people it just won’t work after day three. But the problem isn’t insurmountable. Most of my pump patients do pretty well. What we really need is a sensor that tells us when the set is out, not only when it’s occluded. At least then you’ll have the feedback about blood glucose as an alarm. It’s never going to be “plug it in and forget about it,” at least not for the foreseeable future.

Q: Will there be any more calibrations besides the two per day, as far as the Dexcom sensor goes, with these closed loop systems?

Dr. Naik: For our study, we’re using the Dexcom G5, for which one of the inclusion criteria is two fingersticks per day to calibrate. For closed loop to function well, the sensor must obviously function well. The fingerstick must also still be taken before meals to get mealtime bolus. In the future, we might be able to completely close the loop.

Dr. Argento: Longer term, Dexcom is working on factory calibration too. Also of note, there is an FDA hearing on Thursday on whether Dexcom can get direct dosing, which is what I and half of my patients do, because it works.

Dr. Tanton: One of the hardest talks to have with patients is that their meters are not the gold standard. The Dexcom G4 is probably more accurate than most meters, so if you’re using other CGMs that have higher MARDs, you’re probably actually introducing errors into system.

Dr. Argento: If fingersticks are done properly, they give you a midpoint of a range of possibilities (+/- 15%). The first thing you have to do with a new CGM user is go over fingerstick technique – I have dozens of examples of miscalibrations that led to pretty bad results. We put a lot of weight in fingersticks, but they can be all over the place, even when done well. There should be a system to reevaluate strips that come off the factory shelf – not only what the FDA approved. It’s not like they came down from Mt. Zion – they may be introducing highly deleterious variability into the system. [Editor’s note: Animas Chief Medical Officer Brian Levy emphasized to us that J&J has a rigorous post-market surveillance program to ensure strip accuracy and precision.]

Corporate Symposium: Practical Review of Real-Time Continuous Glucose Monitoring in Clinical Practice at the University of Colorado, Barbara Davis Center and the Future of Dexcom Technology (Sponsored by Dexcom)

Dexcom Technology 2016 & Beyond

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

Dexcom VP of R&D Jake Leach shared several notable pipeline updates in his overview of Dexcom technology: (i) an enhanced G5 app will launch “as soon as possible,” adding backfill of up to three hours of missing data when the phone and transmitter are separated (e.g., when showering) and a mute switch override (e.g., to enable hearing the low alarm when the phone is on silent mode); (ii) the touchscreen receiver was submitted to FDA last month, on par with 1Q16 timing; (iii) an FDA submission of the Android G5 app is expected in the next month or two (making the previous “2016” launch expectation tight but doable); (iv) the Android G5 app will be compatible with Android Wear smartwatches; (v) the next-gen insertion system and 50% smaller transmitter will be filed with the FDA in the “next couple of months” (a 2016 launch was the previous goal, making this timing possible but also tight); (vi) the FDA has approved the IDE for the G6 pivotal trial, which will begin shortly (Dexcom’s largest study ever); and (vii) the Verily partnership is on pace to launch a first-gen product in 2018 (“within two years”). Mr. Leach briefly referred to Thursday’s Advisory Committee meeting on whether G5 should be approved for insulin dosing – though he avoided speculation on potential outcomes, he did comment that the device is extremely accurate and approved for this indication in Europe.

• Dexcom has never had this much going on, and many of these projects will be very meaningful for driving further adoption of CGM – lower cost (Verily), smaller on the body (new transmitter), less painful and intimidating to insert (new applicator), fewer calibrations and better accuracy and alerts (G6), more convenient to view data (Android), etc. Dexcom has blazed a strong trail of product improvements, and we’ll be psyched to see what form factor changes drive the biggest increase in CGM uptake.

Corporate Symposium: Two Ends of the Diabetes Spectrum (Sponsored by Medtronic)

from Artificial Pancreas Technology to Insulin Insecurity

Satish Garg (Barbara Davis Center, Aurora, CO)

Dr. Satish Garg announced, for the first time to our knowledge, his hope to publish results from the 670G pivotal trial extension phase by the end of 2016 (>1 year at-home use of the MiniMed 670G). This will be very scrutinized as a follow-up to the three-month results presented at ADA, and it will be interesting to see if the efficacy is sustained in the 80%+ of participants who continued on the device. To our knowledge, this is also the longest and largest group of patients that has continuously worn automated insulin delivery, representing ~99 years of patient exposure by our math (assuming 12 months per user and ~99 users in the extension phase). Many patients have told Dr. Garg that when they see the shield (indicating that the pump is in closed-loop mode), they feel safe – a cool reminder that user interface and icons can really drive psychological satisfaction with a device. A 64-year-old woman told him after the pivotal study “There’s no way I’m giving this back; I need to keep it. My husband travels to Phoenix all the time, I really need this.” She was the first of nearly 100 patients in the pivotal trial to write the FDA and ask for continued access. Dr. Garg expressed gratitude to the FDA and Medtronic (who is funding the extension arm) for allowing the study to continue, both for the sake of collecting data (during the FDA review – a big win!) and for the patients. Medtronic’s Dr. Fran Kaufman expressed high confidence yesterday in the speed of approval, noting, “I can probably guarantee that it will be before the next ADA meeting” – strong optimism from the usually cautious Dr. Kaufman. As a reminder, the MiniMed 670G was submitted to FDA at the end of June, and Medtronic’s pre-ADA Analyst Meeting maintained the expectation to launch in the US by April 2017.

Keynote

FDA Perspective: Regulatory Promotion of Technology Innovation

Courtney Lias, PhD (FDA, Silver Spring, ND)

In a refreshingly titled talk, “Regulatory Promotion of Technology Innovation,” FDA’s Dr. Courtney Lias picked up where she left off at the NIH Artificial Pancreas Workshop, repeatedly asking for audience input as to how the FDA can make regulatory processes easier. She affirmed that the Agency is prioritizing: (i) device consolidation and convenience via mobile apps; (ii) the optimization of pivotal trial design (namely pre- vs. post-market considerations); and (iii) device interoperability (the FDA announced an internal initiative to encourage interoperability just last week). Touching on the topic of pivotal studies, Dr. Lias reminded the audience that the FDA is flexible and encouraged early communication to ensure that studies have an appropriate balance of pre- and post-market data collection – not just for approval, but also to improve odds of early reimbursement. Medtronic, of course, went with a super-fast single arm, three month trial to demonstrate pre-market safety, to be followed by a planned 1,000-patient post-market outcomes study. While some are worried every company will have to follow this path, the FDA is clearly flexible on this topic. It was great to hear her acknowledge in Q&A that study design can really drive patient access and cost, and the FDA can work with companies to encourage proper design (hopefully with CMS input!). Following her enthusiasm last week, Dr. Lias also continued to preach that device interoperability will actually improve the safety profile of devices and speed up the regulatory process without sacrificing cybersecurity. We will be interested to learn more about the internal initiative, what form it will take, and how that might impact the diabetes device landscape. In Q&A she touched on the OpenAPS movement, noting that as the movement expands beyond personal use, the risk increases and regulation is more likely to be enforced; on the plus side, she highlighted that “these movements apply pressure on the system to approve devices faster, so that we know that the technology that people are using is safe.” We completely agree and recall that the Dexcom G4 Share receiver was approved quickly as the Nightscout community gained steam. We hope the same is true of automated insulin delivery. More highlights from this talk are below!

• Dr. Lias was actually asked to speak about “regulatory challenges in technology approvals”, but she changed the tone and title to be more positive: “Regulatory Promotion of Technology Innovation.” We love it! The FDA has made amazing progress on diabetes devices in the past few years, though there is sometimes still a negative tone associated with the Agency’s work. We applaud Dr. Lias and colleagues for improving FDA’s openness to innovation, and would like to see the FDA more frequently acknowledged for the great and extremely difficult work that they do – particularly given the very under-resourced Agency.
• In her discussion of artificial pancreas pivotal trials, Dr. Lias alluded to Dr. Steve Russell and Dr. Roy Beck’s recently published Diabetes Care paper on the topic – see our ADA coverage here for the highlights. The paper proposed that a “recommended” pivotal trial would be a six-month, parallel group RCT with a representative population (MDIs, high A1c’s, spectrum of ages), a 2:1 randomization (artificial pancreas vs. usual care), and primary outcomes of A1c and time <60 mg/dl.
• Dr. Lias elaborated on some of the mobile app challenges, saying that developers need to consider how the app interacts with the phone’s operating system – How does the app handle updates? What happens to the app’s functioning when the operator receives a phone call? What happens to alarms if the phone is muted? The FDA is “trying to help where they can,” but app developers really have to work with mobile phone companies to address these concerns. This appears to be a big challenge – e.g., Will Apple work with a diabetes device company to help an app override the phone’s mute button when a hypoglycemia alarm comes up?
• Dr. Lias highlighted the FDA’s very rational, tiered approach to regulating mobile apps. At the bottom of the pyramid are wellness apps that don’t require FDA regulation: tracking food consumption, exercise, etc. These are not considered medical devices. The next tier consists of lower risk mobile applications that meet the “device” definition, but are not considered mobile medical applications – patient self management apps and organization/tracking tools fall in this category. At the top of the pyramid are “mobile medical applications,” which are most heavily regulated by the FDA and would do things like communciate with BGMs, CGMs, etc.

• Regarding the pre- vs. post-market balance of data, we recommend reading our take on last week’s JAMA article from FDA’s Drs. Jeffrey Shuren and Robert Califf. The senior FDA officials argued that pre-market trial requirements might be reduced with better post-market data collection. We see high potential for this to be used with automated insulin delivery, particularly for devices that are cloud connected.

Questions and Answers

G. Todd Alonso (Barbara Davis Center, Aurora, Colorado): In September, it will be three years since Medtronic’s threshold suspend was approved. Yet some payers still refuse to pay. We have this peer-to-peer system, but I often find that I’m talking to someone who’s never even thought about treating a patient with diabetes. So we have patients under 16 who do great on the device, but they can’t afford it.  What can the FDA do to help us get coverage for them?

A: I agree – the stories I hear of patients fighting for some devices and test strips are frankly terrible. The main thing the FDA can do is talk to manufacturers, basically urging them to consider how they can design their studies so they have the highest chance of achieving reimbursement. We do what we can to try to encourage this kind of thinking. I’d also love to hear your ideas to help with this!

Satish Garg (Barbara Davis Center, Aurora, Colorado): There are two big groups making artificial pancreas devices in their homes – the DIYPS movement and the #WeAreNotWaiting movement. What is the FDA perspective on that?

A: Regardless, they are making a medical device. If they make it for personal use only, the FDA may not get involved. But as the movement grows and the proprietary technologies and algorithms are distributed, the public health risk is increased, the device is not really for personal use, and it is something we would be more likely to engage on. They know that. The other thing to consider is that these movements apply pressure on the system to approve devices faster, so that we know that the technology that people are using is safe.

Q: When I’m outdoors I can’t read my devices, and I feel that’s unsafe. I’ve gotten alarms before and not been able to respond appropriately because I couldn’t see my screen. That should maybe be an FDA issue? Also, if someone loses or misplaces a device, shouldn’t there be a way for whoever finds it to call the manufacturer and have them locate and send it to the owner? Just a thought.

A: See, this is why I love coming to these conferences – those are two suggestions that I’d never heard before. I’ve definitely considered the screen readability issue with my phone before, but never with pumps. Those are human factors issues that I will certainly pass along.

Comment: The FDA used to be a deterrent for getting things done – I’d like to thank Courtney and the FDA for making things move as smoothly and rapidly as they are for the artificial pancreas.

Recent Advances in Diabetes Treatment

Overview of JDRF Role of Closed-Loop in T1D

Aaron Kowalski, PhD (JDRF, New York, NY)

Dr. Aaron Kowalski highlighted JDRF’s instrumental role in the development of the artificial pancreas field, shared some pragmatic thoughts on glucagon, and expressed hope in Q&A that devices will increasingly be used in type 2. Dr. Kowalski discussed the remarkable progress the artificial pancreas field has made in the past ten years, from concept stage to the cusp of commercialization. He noted that JDRF has now contributed nearly $150 million to closing the loop, established the development roadmap, created the artificial pancreas consortium, drove the FDA to release guidance, and is now working on the T1D Outcomes Program (to define better value propositions for all stakeholders beyond A1c). Regarding glucagon, Dr. Kowalski remarked, “I’m not a glucagon skeptic, but a glucagon realist. It’s not as simple as you may think.” The frequently-used analogy, he elaborated, that an insulin-only artificial pancreas is like driving a car without brakes, is not appropriate – “when you drive a car, hopefully the brake works every time you press it.” An occluded glucagon infusion set could be deadly, and Dr. Kowalski argued that glucagon, if included, should only be used as a last-resort rescue. He was careful to point out that he isn’t opposed to the use of glucagon in closed loop devices, but that the health economics aren’t favorable at the moment (the marginal cost is currently greater than the marginal benefit over insulin-only). We look forward to more discussion and data on this topic, not just for glucagon but for other hormones like amylin. How much glycemic/user experience benefit does another hormone have to add relative to additional cost or complexity it brings to the system? How will payers reimburse for a multi-hormone closed-loop and weigh the glycemic and quality of life outcomes?

• During Q&A, Dr. Kowalski commented that type 2s should also use technology (especially CGM) whenever possible. Currently, providers are “dosing all kinds of drugs with no information,” and CGM and closed loop algorithms could drive much more informed treatment (particularly insulin titration). It’s an excellent point, and we see the entire field moving to give type 2s and PCPs more actionable glucose data through better product form factors/cost, clinical decision support, and even behavioral advice (Abbott’s FreeStyle Libre, Dexcom/Verily’s bandage-like CGM, Medtronic/ Qualcomm’s disposable professional CGM, Glooko’s basal insulin dosing system, WellDoc/LifeScan’s partnership, etc.).
• As a person with diabetes himself, Dr. Kowalski expressed right off the bat that it “really gets under [his] skin” to hear people say that people with diabetes aren’t “compliant.” This is a common occurrence, and we agree with Dr. Kowalski – it’s not as if some patients decide to be compliant and others blow diabetes off. Rather, we observe lots of reasons why patients may not be as engaged as others with diabetes: cost of therapy, side effects, real life getting in the way of diabetes, the invisibility of diabetes, the long-term nature of benefits vs. short-term hassles, feeling like a failure, data overwhelm, etc. The most game changing products will help address these issues, since these are likely to be patients running very high A1c’s.

--by Abigail Dove, Helen Gao, Brian Levine, Payal Marathe, Adam Brown, and Kelly Close