Keystone 2014 – Practical Ways to Achieve Targets in Diabetes Care

July 17-20, 2014; Keystone, CO – Days #3-4 Highlights – Draft

Executive Summary

Greetings from back in San Francisco, where we just finished our coverage of the final two days of the Barbara Davis Center’s 2014 Keystone Conference. The vast majority of attendees stayed around through the conference’s finale, a testament to an outstanding closing agenda (and probably the fact that nobody wanted to leave the scenic surroundings). Highlights of the conference finale included an FDA keynote and full plenary session on the closed loop, lively discussions of reimbursement, and questions about the costs of cardiovascular outcomes trials. Below, you will find our synthesis of the top highlights, followed by coverage of the presentations, panel discussions, and corporate dinner events.

1. The FDA’s Dr. Courtney Lias’ (Director, Division of Chemistry and Toxicology Devices) keynote presentation on the artificial pancreas was perhaps the most optimistic and encouraging we’ve ever seen from the Agency – she noted that approval of closed-loop technologies will happen “sooner than you think.” She shared encouraging Agency perspective on device consolidation, mobile technology, and “component artificial pancreas systems” (in which different companies sell different components).

2. Dr. Fran Kaufman (Chief Medical Officer, Medtronic Diabetes, Los Angeles, CA) provided a broad overview of Medtronic’s pipeline, endorsing insulin-only closed-loop systems on the basis that the costs of a bihormonal system exceed the benefits of slightly enhanced glycemic control.

3. Dexcom VP of R&D Mr. Jake Leach shared a valuable overview of Dexcom’s very full pipeline, highlighting work on the company’s near- and long-term R&D priorities. A Q&A that included Dexcom CEO Terry Gregg shared that the company is still deciding whether to release the G4AP algorithm or to wait for Gen 5. This kind of detail? It’s why we go to Keystone!

4. Dr. Philip Home (Newcastle University, Newcastle upon Tyne, UK) suggested that current regulatory guidelines for biosimilar insulin approval might not be sufficient to fully confirm biosimilarity.

5. Medicare coverage of CGM was a major topic throughout the day. Dr. Satish Garg (University of Colorado, Denver, CO) said that when Medicare turns down CGM coverage, secondary insurance may pay for the CGM in those patients. Meanwhile, Dexcom CEO Terry Gregg said that the company is working with the FDA towards a non-adjunctive claim for CGM, a key CMS objection to covering the technology.

6. Cardiologist and FDA EMDAC panelist Dr. William Hiatt (University of Colorado, Denver, CO) expressed mixed views on the FDA’s 2008 CV Guidance – he believes it is important to take steps to exclude excess CV risk, but acknowledged that CVOTs can involve paying a lot and learning very little.  

7. In the same discussion on CVOTs, Dr. Robert Ratner (ADA, Alexandria, VA) suggested that the FDA might be ready to reconsider the appropriateness of its 2008 CV Guidance.

8. Earlier in the day, Dr. Ratner presented some tantalizing data on improved outcomes and cost savings with new evolving healthcare models such as patient-centered medical homes (PCMHs) and accountable care organizations (ACOs).

9. Drs. Irl Hirsch (University of Washington, Seattle, WA) and Philip Home (Newcastle University, Newcastle upon Tyne, UK) both stressed that the US is bearing a disproportionate burden in terms of funding diabetes drug development, and insisted that contributions from Europe are necessary to sustain innovation.

10. Dr. Howard Wolpert (Joslin Diabetes Institute, Boston, MA) expressed confidence that CGM would experience “accelerated adoption” in the near future. He called for more data and study of sensor adherence.

Table of Contents 

Top Ten Highlights

1. A packed room witnessed Dr. Courtney Lias’ (Director, Division of Chemistry and Toxicology Devices, FDA, Silver Spring, MD) testimony that the approval of a closed-loop system will happen “sooner than you think.” No kidding! Addressing FDA perspectives on the artificial pancreas, her keynote stressed that opportunities for industry growth will abound in the near future. She reported that the FDA is particularly interested in device consolidation and is “100%” behind the growth of mobile technology in facilitating this progress – this was huge news to hear, as the FDA has been criticized in the past for being overly cautious on the mobile technology front. Moreover, she said that the FDA is encouraging the growth of “component artificial pancreas systems,” in which different companies sell different components of a device. This flexibility is interesting to see from a patient perspective, since it would allow consumers to choose their algorithm, pump, and sensor of choice – we do think this will make things more complex for providers. Moreover, it should foster greater competition in the industry, since de-coupling closed-loop components gives small companies a chance to contribute (rather than full pump/CGM/algorithm systems, of which only Medtronic can produce at the moment, though we think others may be able to as well). In closing, Dr. Lias dismissed many “common misconceptions” associated with FDA approval, stressing that the FDA is not requiring that closed-loop systems be “perfect” solutions (see full details below). Cheers to that, as this is very helpful in setting patient expectations.

2. At Medtronic’s industry-sponsored dinner, Dr. Francine Kaufman (Medtronic Diabetes, Northridge, CA) gave a number of insights on automated insulin delivery (she is the master) and provided her personal opinion on bi-hormonal artificial pancreas systems, commenting that the approach adds too much complexity and cost. She highlighted that payers will “not be thrilled about the concept until we show benefit,” emphasizing that glucagon is extremely expensive in the current form and has little data on chronic use. Dr. Kaufman pointed out that “we can do a really good and safe job with insulin alone” and that insurers will not buy into the dual hormone system unless it’s shown to be significantly better than insulin on its own. She also spoke enthusiastically on Medtronic’s pipeline – there are certainly positive comments circling from the EU on the MiniMed 640G. For more on the company’s R&D pipeline, which includes the Enlite 3, hybrid closed-loop control, and the MiniMed Flex, see our coverage of its excellent 2014 Analyst Day.

3. At Dexcom’s dinner, Mr. Jake Leach (VP of R&D, Dexcom, San Diego, CA) discussed the future of Dexcom CGM innovation, highlighting the company’s near- and long-term R&D priorities: the cloud-based SweetSpot software (no timeline provided, but clearly in the works), ongoing feasibility trials on Gen 6, a mention of the new Dexcom/Insulet collaboration first announced on Day #1 of ADA 2014, smaller transmitters, new algorithms to support longer sensor life, accuracy/reliability in non-adjunctive use, and factory calibration (“Don’t know when, but it’s coming”). Notably, CEO Mr. Terry Gregg said in Q&A that no decision has been made on whether to make the G4 AP algorithm available soon or to wait for Gen 5. He emphasized that Dexcom remains committed to releasing a new product “every 18 months” – the G4 Platinum was originally released in October 2012 (22 months ago), though approvals have recently come for a G4 Platinum pediatric indication and a professional use indication. Mr. Leach said that Dexcom Share will be available “very soon” – as of Dexcom’s 1Q14 call in May, it was characterized as in “the final stages of review.” 

4. Dr. Philip Home (Newcastle University, Newcastle upon Tyne, UK) characterized the current regulatory guidelines for biosimilar approval as inadequate. He stressed that no single set of evidence – on PK/PD, A1c, pre-injection glucose, hypoglycemia, immunogenicity, etc. – will in itself be sufficient to prove biosimilarity, and that instead clinicians and regulatory authorities must consider the data package in its entirety. Dr. Home lamented the lack of access to data on manufacturing quality for biosimilars and predicted that the manufacturer’s reputation will serve as a stand-in for quality in many people’s minds – good news for companies like Lilly that are well-established in the insulin market.

5. Medicare coverage of CGM was a major topic of discussion throughout the day. Dr. Satish Garg (University of Colorado, Denver, CO) remarked that when Medicare turns down CGM coverage, patients’ secondary insurance might still pay for the technology. Meanwhile, Dexcom CEO Terry Gregg said that the company is working with the FDA towards a non-adjunctive claim for CGM, a key CMS objection to coverage. Still, he noted that coverage may be two years out at this point. The lack of Medicare coverage for CGM was a central focus of an inspiring panel discussion in which patients and providers shared their stories about life with type 1 diabetes and highlighted what they felt were the most pertinent issues in the field. Doctors and patients alike recounted their futile attempts at persuading Medicare officials of the necessity of CGM for many elderly patients – the only “success story” involved a three-part appeals process and multiple letters from highly respected physicians. This session and the one covered in our Day #2 report reinforced that Medicare coverage of CGM is one of the foremost advocacy issues in diabetes – we know that Dexcom, Medtronic, and JDRF have been working hard on this front, though most agree it is high time for more grassroots advocacy as well. More details from this panel are below.

6. FDA EMDAC panelist Dr. William Hiatt (University of Colorado, Denver, CO) provided his perspective on the value of cardiovascular outcomes trials – a perspective we saw as quite notable given that Dr. Hiatt sits on the FDA’s EMDAC panel. In his presentation and the subsequent panel discussion, Dr. Hiatt revealed a somewhat ambivalent perspective on the FDA’s 2008 CV Guidance – he acknowledged that the diabetes field paid an awful lot and learned relatively little with SAVOR and EXAMINE, and pointed out, as we continue to do, that one ruling that kicked off the CV controversy (Avandia and RECORD) ultimately ended up being reversed, suggesting that the entire premise behind the CVOT requirement may need revisiting.

7. Both Dr. Hiatt and Dr. Robert Ratner (ADA, Alexandria, VA) suggested that it was possible (or in Dr. Ratner’s case, likely) that the FDA might be willing to review the appropriateness of the 2008 CV Guidance. As a basis for the theory, Dr. Hiatt referenced the 2013 FDA Advisory Committee meeting on the re-adjudication of the RECORD trial for GSK’s Avandia (rosiglitazone) that led to the exoneration of the drug, while Dr. Ratner cited a cryptic comment from the FDA’s Deputy Director for Clinical Science Dr. Robert Temple at the same AdComm suggesting that a review of the guidance might be on its way. These comments were highly speculative, but were still very exciting to hear, especially in light of the open hearing the FDA will be holding in August on a piece of the guidance.

8. Accountable care organizations and patient-centered medical homes are the future of healthcare delivery in the US, according to Dr. Ratner (in a keynote presentation that kicked off day #4). Early data on the implementation of these programs, which have largely been driven by the Affordable Care Act, suggest that they might be able to finally bend cost curves while preserving or even improving quality. Later, during Q&A, audience members nodded (some chuckling ruefully) when Dr. Ratner called Medicare “penny wise but pound foolish” with regards to its competitive bidding policy for test strips.

9. Continuing a discussion from day #2 on the high costs of insulin, Dr. Irl Hirsch (University of Washington, Seattle, WA) lamented the fact that the US “is being asked to fund all the innovation” in the diabetes pharmaceutical arena. Since other countries with government-run healthcare systems set strict limits on the amount they are willing to pay for products like insulin, companies compensate by raising prices in the US market. Pharmaceutical companies have been quite open about this issue; Dr. Philip Home (Newcastle University, Newcastle upon Tyne, UK) noted that “the CEO of Novo Nordisk has been hammering on the table” saying that Europe will need to pay more in the future in order for the current level of innovation to be sustainable. Dr. Hirsch said he can’t predict where this is going to end, but if the considerable focus on cost at this conference is any indication, patients and providers in the US will not accept such skyrocketing prices for diabetes drugs indefinitely. Others might say that prices went up so much in 2013 and 2014 in anticipation of insulin pricing dropping with biosimilar entry.

10. Renowned diabetes technology expert Dr. Howard Wolpert (Joslin Diabetes Center, Boston, MA) expressed confidence that “we are at an inflection point for accelerated adoption” of CGM. Citing recent technological advances in accuracy and reliability, Dr. Wolpert stressed that the “overwhelming” benefits of CGM will soon outweigh the hassles for the majority of type 1 patients. That said, he highlighted the need for more data examining sensor adherence in order to inform clinical decision-making, given that sensor usage is a key predictor of the benefits of CGM. In closing, Dr. Wolpert stressed that clinicians must be aware of the challenges that information overload and alarm fatigue can present to patients – he noted that information overload can be transformed into positive patient self-management through education, while alarm fatigue can be easily minimized through proper alarm optimization.

Detailed Discussion & Commentary

Keynote Presentations

FDA Perspective on Closed-Loop Studies

Courtney Lias, PhD (Director, Division of Chemistry and Toxicology Devices, FDA)

The audience was overflowing at Dr. Lias’ keynote talk addressing FDA perspectives on closed-loop devices. Emphasizing the “exciting” progress in this field, she also detailed the “many” challenges that will precede commercialization – her comments drew particular attention to the complex underlying pathophysiology of the disease as a larger hurdle than even device limitations. Looking ahead, Dr. Lias stressed that opportunities for industry growth will abound in the near future, as the FDA is particularly interested in device consolidation and “component artificial pancreas systems” with parts sold from multiple companies. In addition, the FDA anticipates that the growth of mobile technology will play a central role in this consolidation, a sentiment that reflected more openness and less caution than we can recall from past FDA talks. In closing, Dr. Lias dismissed many “common misconceptions” associated with FDA approval, stressing that the FDA is not anticipating or requiring that closed-loop systems be “perfect” solutions. We were very glad to hear this for the sake of patients and providers and society.

  • Dr. Lias opened her talk by stressing that “the FDA’s mission is not to squelch innovation”; rather, closed-loop systems fall under the umbrella of “promoting public health.” We were pleased to hear that the FDA recognizes the increased risks and diminished quality of life that are associated with diabetes. Notably, it appeared that the FDA considers the artificial pancreas a practical therapeutic option. This was the most direct response we can recall ever hearing to comments that the FDA is a roadblock to innovation. The device side of the Agency has been much more open, flexible, and responsive in the past 18 months, and it was good to hear Dr. Lias openly responding to these criticisms.
  • Noting that artificial pancreas approval will happen “sooner than you think,” Dr. Lias shared particular excitement regarding home and camp studies of closed-loop systems. She applauded clinical trials in camps for “stressing the system” by putting adolescents in novel environments and robustly testing the limits of devices. She also noted that the FDA is looking forward to the findings of upcoming home studies that will test the devices in even more practical, real-world settings.
  • The FDA is “100%” behind the growth of mobile technology as a method of facilitating device consolidation. Dr. Lias expressed apprehension that the number of devices (e.g., meter, CGM, pump receiver) in closed-loop systems will contribute to frustration and non-adherence in patients. Given that nearly everyone carries a cell phone, the FDA recognizes that mobile platforms provide a convenient potential solution for consolidation. Notably, Dr. Lias reported that the FDA is working closely with industry on requirements for mobile technology market entry and security, evidenced by their guidance on mobile medical apps issued in September 2013. We consider these comments very encouraging news not only for the future of closed-loop systems, but for the growth of mobile technology in diabetes self-management more broadly.
  • The FDA envisions the future of the closed-loop industry to consist of “component artificial pancreas systems,” in which different companies sell different components of a device. In such a system, a consumer would be able to purchase an algorithm on an app separately from the pump and sensor it communicates with. Such an arrangement would facilitate much greater patient choice and more competition within industry, though it also brings up many tough questions: How would the regulation process work for individual components? Would standards be developed to govern how components interface with each other? Who would develop such standards? How would companies collaborate to develop compatible platforms? Reservations aside, we enthusiastically applaud the FDA’s ambition and vision (as did multiple members of the audience) to foster this innovation.
  • Dr. Lias acknowledged that the diabetes population is at risk to begin with, so “it does not make sense” to try and develop a closed-loop device with zero risk. Instead, the approval process is made “in the context of the significant risk people with diabetes face every day.” We were happy to hear this perspective emphasized from the FDA, as the perception for many (incorrectly in our view) is that the Agency often seems to err on the side of “perfection or bust” for many devices and drugs in diabetes. The recent move to flexibility and openness on the device side is welcome for all stakeholders in diabetes, especially industry and patients.

Questions and Answers

Q: I have a question about the Dexcom G4 not being reliable in pediatric populations. Can you share the numbers regarding that?

A: It’s not that it’s unreliable. It’s approved and it’s reliable. However, if you look at the product insert, you’ll be able to see the data yourself. The tables are there with adult and children data, and you can compare it. What happens is that the sensor tends to miss alarms more often in the hypoglycemic range in children. People just need to be aware of that and make a decision that’s appropriate for them.

Dr. Peter Chase: I would add that some of those Dexcom studies have been repeated in children, and it is very similar to data in adults. So the guidelines on the pediatric part will likely be changed in the near future. You were talking about different stages of low glucose suspend. All of us thought that the next stage would be control-at-night, where people don’t have to worry about exercise, stress, eating, and whatnot. Is there anything you can add to the decision of skipping nighttime control and going straight to whole closed-loop?

A: It’s not so much a decision; it’s just not a requirement. FDA wouldn’t be hesitant to allow someone going to full closed loop without doing baby steps. Nighttime control is important. Any closed loop device has to look at performance at night. It’s a matter of looking at both things at once. In device development, you need to assess and mitigate the risks of that device for its intended purpose.

Q: You mentioned that the FDA would try to get companies to cooperate in developing products. How do you plan on doing that?

A: We can’t mandate they work together. We can encourage it, though. We begin by talking with them, often in a large forum, explaining the issues that need to be solved. Sometimes, these companies are starting to work together already and come to us for guidance in how they can interact. In that case, we can help them in planning for these things. However, when people don’t want to work together, it’s difficult to get cooperation. We do have other ways to put pressure on them though. We could even make it a public issue, for example, and make the information regarding cooperation public in order to put pressure on them that way. You have to remember, too, that you – the users – have a lot of leverage as well. Companies depend on consumers, so think about that as well.

Dr. Philip Home (Newcastle University, Newcastle upon Tyne, UK): I’m from outside the US. Tell us from within the Agency, what has been a problem in benefit/risk assessment? Sometimes, it seems as if they are maximizing safety issues and minimizing benefit evidence. It ends up feeling like they’ve been working against us. Once you begin to increase risk, it seems like you decrease benefit very quickly. Any comments about how different divisions put input into approvals?

A: Yes, there are different cultures in the FDA and different levels of risk tolerance. Diabetes has been nice in that we have been working with the diabetes community for a long time. A couple years ago, the center did change the approach to regulation of artificial pancreas devices. The Center decided to consolidate under one group so it was more streamlined. This helped manage the process better and helped make it go more smoothly. I do think it has helped. We’ve made a lot of progress in last few years in trying to get the balance better.

Q: Is there a way to link the EMA and FDA guidelines? Because current guidelines for closed-loop systems are very different.

A: You are right – they are very different. We occasionally talk with the EMA, but perhaps not as often as we should. We’re happy to talk with them though. One of the main things is that there are almost no requirements regarding CGMs in the EU – Very few. However, I think once we get to the artificial pancreas, the EU has more requirements. After all, ideally, we don’t want to have totally different requirements for approval in the US and EU. We want to allow companies to do one study and receive approval here and there. In doesn’t make sense to wait three years to bring a device to the US after it’s been approved in the EU.

Delivery of Quality Care in the Evolving Payer Landscape

Robert Ratner, MD (American Diabetes Association, Alexandria, VA)

Dr. Robert Ratner presented on how a critical trend in the overall healthcare system – the rise of patient centered medical homes (PCMH) and accountable care organizations (ACOs) – can be seen as an opportunity to improve diabetes treatment. He reviewed several studies on the effects of PCMH implementation on diabetes outcomes and costs: by emphasizing integrated, team-based care and quality-based reimbursement, organizations (in many cases) were able to significantly reduce ER visits and inpatient hospitalizations, improve achievement of treatment goals, and save substantial amounts of money ($161 million annually, in the case of Community Care of NC). Dr. Ratner also discussed several ACO pilot projects, including the Medicare Shared Savings Program, which will eventually involve 360 Medicare ACOs in 47 states and Washington, DC. Out of the 114 organizations that have participated thus far, 54 spent less than their budget benchmarks and 29 qualified for shared savings (an average $126 million savings to the healthcare systems and $128 million for the Medicare Trust Fund). Dr. Ratner believes that ACOs will grow rapidly in prevalence, and he encouraged the diabetes community to take the lead as a “chronic disease model” for providing high quality care at a reasonable cost.

  • Citing the fact that preventing one hospitalization pays for a pump for a year (at least!), Dr. Ratner stressed the importance of keeping patients out of the hospital. The Affordable Care Act is one of many factors driving a transition from a fee-for-service reimbursement model to one that better aligns provider incentives with improved outcomes. PCMHs and ACOs are two of the proposed strategies to fundamentally change the way care delivery is structured, and Dr. Ratner suggested that the two new care delivery models could be an opportunity to greatly improve diabetes care.
  • Dr. Ratner listed the seven key components of the PCMH:
    • Coordination and integration of care
    • Decision supports for quality and safety
    • Whole-person orientation
    • Each patient is associated with a “personal physician”
    • The personal physician is the director of patients’ care
    • Enhanced access to care
    • Quality-based reimbursement
  • Dr. Ratner defined ACOs in a easy-to-understand manner, characterizing them as an updated kind of capitated care. A group of 32 “pioneer” ACOs covering nearly 700,000 Medicare beneficiaries are early implementers of the ACO model; 25 of them have reduced risk-adjusted hospital readmission rates below their benchmarks, and 30 of them have made money, suggesting that this model could be an effective means of bending the healthcare cost curve.

Questions and Answers

Q: Medicare restricts the thing that costs the least: test strips. If a cop pulls you over and asks how fast you’re going and you say “I looked yesterday,” you get a ticket.

Dr. Ratner: That’s traditional Medicare. In an ACO, you give them as many strips as it takes.

Comment: I hope that Medicare will provide smarter reimbursement in the future, because it is not right now.

Dr. Ratner: This is where we are in an evolution. What does it take to keep people out of the ED? That’s what’s ultimately going to win. Penny wise and pound foolish, that’s what Medicare is right now.

Q: This was a great overview, but I didn’t see any emphasis on personal responsibility. No health care system will survive unless patients are personally responsible

Dr. Ratner: You’re right but we need to be very careful. We provide education and support to do all of that. To throw a person with diabetes under the bus if they can’t do everything because they don’t understand or they don’t have the finances is unfair. We should not use pay for performance on selected individuals or a percentage of individuals; that’s not a patient centered approach. Shared decision-making and decision support is how you get buy-in.

Comment: With car insurance, having a bad record means you pay more money. Why can’t we put that concept into health insurance? You’re overweight, you don’t take care of yourself, you pay more.

Dr. Ratner: We have to differentiate car insurance from health insurance. You have 100% of the responsibility for driving; that’s not true for your health.

Q: I work for a PCMH and ACO. The biggest problem is that patients aren’t interested. We give them behavioral health and diabetes education, but they don’t want to do it, so my data’s not as good. The administration understands for now, but in the future, if I don’t meet their guidelines, I may have to drop patients.

Dr. Ratner: The idea of cherry picking is what killed capitated care. The issue is that given the Affordable Care Act, you can’t deny someone coverage based on a previous diagnosis or previous expenditures. People won’t be able to cherry pick. We’ll have to learn to accommodate people. Not everyone will have an A1c under 7% and stay out of the hospital. The question is, do you shift the distribution curve? Everything is a normal distribution: if you can shift it to the left so the vast majority are doing better, you’ll have outliers who don’t get there, but you’re saving money for the system.

Q: Can you comment on the very high reimbursement for procedures relative to the type of care we endocrinologists do? It is a real challenge if the salaries of other disciplines keep going up compared to us endos.

Dr. Ratner: That’s the past. Radiologists are in trouble. Increased procedures (radiology, scoping folks who don’t need it) will cut into an ACO’s ability to provide care, so they’ll get penalized. There will be increased reimbursement for keeping people out of the hospital and keeping costs down, not high cost procedures that defeat the purpose.

Dr. Satish Garg (Barbara Davis Center, Aurora, CO): Costs are so overinflated. How do we account for that? In many countries, you can get a CT scan for $40.

Dr. Ratner: That’s the past. What we’ve done is pay for individual procedures. That’s changing. With bundled care, if you get charged $3,000 for MRIs, you’ll stop doing $3,000 MRIs. It will be $200 or it won’t be done.

Q: There was recently an editorial in NEJM on what’s preventing us from preventing type 2 diabetes. What are we doing to expand on environmental factors, personal responsibility, etc. to prevent costs?

Dr. Ratner: Do I have another hour and a half? The Diabetes Prevention Program is very active and expanding. Social determinants of diabetes and obesity were reviewed by the prevention committee at ADA and a piece was published in 2013 about what the environmental influences are. The difficulty is that we don’t have good data on changing determinants. We need those. We need the public will to consider diabetes a public health problem. Mayor Bloomberg tried to do that and he got shot down. It’s an evolving situation. We managed to come up with the public will to get rid of lead in paint and gasoline, and that’s had a major impact on kids and lead poisoning. Do we have public will to change issues around sugar-sweetened beverages and high-calorie, low-density food? I don’t know that we do yet but that’s where we have to go.

Plenary: Cardiovascular Disease and Diabetes

Is CVD Outcome for Diabetes Drugs Appropriate?

William Hiatt, MD (University of Colorado, Denver, CO)

Cardiologist Dr. William Hiatt offered a thorough and somewhat ambivalent perspective on the much-debated topic of diabetes drugs and cardiovascular risk – we were listening closely, as Dr. Hiatt is a member of the FDA Endocrinologic and Metabolic Drugs Advisory Committee (EMDAC). He began by suggesting that A1c does not represent a useful biomarker for cardiovascular risk, based on results from ACCORD and ADVANCE that indicate that the most intensive control may increase the risk of mortality. Dr. Hiatt said little about the value of ongoing outcomes trials except to express his disappointment that trials on newer diabetes drug classes have not yet demonstrated any cardiovascular benefit, though he is hopeful about the potential of SGLT-2 inhibitors due to their beneficial effects on blood pressure and weight (there is also great hope in the field that GLP-1 agonists may prove to be cardioprotective). Much of Dr. Hiatt’s talk was devoted to recounting the controversy over rosiglitazone that led to the FDA’s 2008 CV Guidance. It was hard to gauge his overall viewpoint on the saga: he disparaged some elements of the RECORD trial, but also acknowledged that the Duke Clinical Research Institute (which ran the re-adjudication that ultimately exonerated the drug) performs excellent work. He concluded his presentation by suggesting that the TZDs and the DPP-4 inhibitors that have had CVOTs completed likely do not posed increased CV risk.

  • During the panel discussion, Dr. Hiatt more forthrightly expressed his concern and ambivalence over the appropriateness of the FDA’s CV Guidance (see below). He acknowledged that SAVOR and EXAMINE (the CVOTs for AZ’s Onglyza and Takeda’s Nesina) cost a great deal and provided relatively little notable data. Ultimately, should the FDA choose to review the appropriateness of the 2008 CV Guidance, EMDAC voices such as Dr. Hiatt that are neither bluntly for or against the policy might be key tiebreakers in the discussion.
  • As a reminder, the FDA is holding an open hearing on August 11 regarding the reporting of interim CVOT results - it should provide some insight into the FDA’s current views on the current CVOT paradigm.

Recent and Ongoing Cardiovascular Outcomes Trials with Diabetes Drugs

Jay Skyler, MD (University of Miami, Miami, FL)

Dr. Jay Skyler reviewed four of the biggest recent cardiovascular outcomes trials (CVOTs) with diabetes drugs and highlighted ongoing trials, expressing frustration with the way the trials are being designed. He emphasized that current CVOTs are too short to prove much about long-term CV outcomes – this has been disappointing. He also argued that trials should not rely on enrolling patients with pre-existing CVD, given that the primary goal of diabetes therapy is to prevent CVD in the first place. Dr. Skyler thoroughly reviewed results of SAVOR, EXAMINE, AleCardio, and ORIGIN, whose results have all been previously presented in detail at other meetings. Looking ahead, he presented a slide with the large number of ongoing CVOTs, emphasizing that these studies will require large pools of patients as well as money. He also pointed out that MannKind’s inhalable insulin Afrezza will not be required to do a CVOT in the classic sense, but that the FDA-mandated pulmonary safety study will provide good long-term data on CV safety – a CVD trial in disguise! To conclude, Dr. Skyler emphasized that he is concerned with the design of these CVOTs, stating that longer studies enrolling people without cardiovascular disease are in order to assess risk reduction, not just neutrality.

  • When reviewing the results from SAVOR, Dr. Skyler paused to consider the slight (5-to-12) asymmetry in pancreatic cancer rates that trended towards favoring Onglyza. He argued that if the imbalance had gone the other way, even though the difference would not have been statistically significant, the news headlines would have been completely alarmist. Both here and at other moments during the conference, Dr. Skyler suggested that the mainstream media has done a poor job of framing the results of major trials, causing far-reaching shockwaves.
  • During the panel discussion, Dr. Skyler noted that many of the individuals behind the CV guidance (which direct large amounts of research funding towards academic institutions) were from the institutions that now run many CVOTs. He characterized this as a major conflict of interest, and questioned why industry-linked conflicts of interest were weeded out whereas academic conflicts of interest were not.

Panel Discussion

Q: Following this session, how many of you feel better prepared to deal with cardiovascular disease?

[Only a few people raised their hands]

Q: How many of you feel more confused?

[Sheepish laughter]

Q: What do we do about these unnecessary CVOTs? Do we complain to senators?

Dr. Robert Ratner (ADA, Alexandria, VA): The FDA is clearly starting to look at this issue again. At the last rosiglitazone meeting, the one that took away the black box with the restrictions, one panel member said, ‘There’s the elephant in the room that we’re not talking about. With the removal of restrictions on rosiglitazone, does that mean we’re going to stop requiring CV outcomes trials?’ And Bob Temple’s response to that was, ‘That’s not why we’re meeting today.” On the other hand, in August, they are going to talk about opening up interim data from the CVOTs for review for the approval of drugs. So I think the discussion is coming back. Do you see that happening, Will?      

Dr. Will Hiatt (University of Colorado School of Medicine, Denver, CO): I do. I must say. I’ve really wrestled hard with the diabetes guidance. It extended to drugs such as sibutramine that clearly caused harm. And I think that thinking about it is okay, but the problem is that you need enough events to rule out a certain amount of risk, so these tend to be low-risk populations, particularly with obesity drugs, and you don’t necessarily require a lot of events. And so you have these really high upper bounds on the confidence intervals driven by the lack of events. So what triggered the 2008 guidance and what was in the public discussion was the rosiglitazone meta-analysis. That’s what triggered it – it was that big. Looking back now and thinking about the saxagliptin and sitagliptin publications, I say to myself that we spent a lot of money to learn not a whole lot. So I worry about that exact same issue. Going forward, if you have a drug with a mechanism that you think actually might reduce cardiovascular disease, by all means go for it [the CVOT]. But the fact that we require every drug in the world … is it overdone? I don’t want to sit and do too much judging, but I am concerned about the same issue.

Dr. Jay Skyler (University of Miami, Miami, FL): Regarding the 2008 Advisory Committee meeting, if you go back and look at the minutes of the meeting, there are people with conflicts of interest who were not allowed to participate because they’re involved in industry. And then the people who are allowed to participate and who are giving recommendations happen to be from the Cleveland Clinic, Oxford, Duke, and other institutions. They suggest that trials should be done and conducted by independent academic groups, the same groups that they represent! How is that not a conflict of interest? They’ve created their own machines and they’re driving them themselves.

Q: The cost of all these outcomes studies and getting the medications in front of the FDA are huge. These medications aren’t even supposed to affect cardiovascular disease. They’re for diabetes. These drugs lead to relatively small reductions in A1c. Is it really worth it? I propose a brand new medication – its name is one less slice of bread a day. We could probably get more significant reductions at much lower cost.

Dr. Jay Skyler: I agree with you. I think it would have been nice for the people to have planned the meals for this conference to have taken that into consideration.

Dr. Satish Garg (Barbara Davis Center, Aurora, CO): Now I have to step in here. Every year, we take these comments into account. We got rid of the pizza from last year! [Laughter]

Q: I didn’t hear any mention of Cycloset? Can you comment on cardiovascular safety trials? 

Dr. Skyler: Trials were actually not initiated prior to the time that the FDA put the new guidelines into effect and it wasn’t powered in the same kind of way. The results were actually quite surprising and showed a dramatic decrease in cardiovascular events with Cycloset. And the FDA let them put it in the label under side effects. But, there weren’t any. They were concerned it was going to be an increased risk, but actually, there was a 50% reduction, which is really dramatic. The explanations for this were not entirely clear.

Dr. Ratner: Cycloset is a perfect example of the danger of small numbers. When you have a small number of events, chance alone is going to have a positive p value of one in 20. That’s essentially what happened. There is another example of cardiovascular benefit from low numbers, in which acarbose had a dramatic effect in reducing myocardial infarctions. But no one really believes it because it wasn’t really validated in other studies. I think that’s the situation with Cycloset. Small numbers become difficult to replicate.

Dr. Skyler: But with a 50% reduction in cardiovascular events, I think the data still says something [about Cycloset]. If it were 10-20%, I might dismiss it. I think what it says is that we need another bigger trial to understand it better.

Q: I’m on the hospital side and in a lot of my consults, I see people with chest pain post-myocardial-infarction, a new diabetes diagnosis, and no health insurance. I would love to use a DPP-4 inhibitor but I have no choice but to use cheap things like metformin and sulfonylureas. We all have that love/hate relationship with sulfonylureas but they are inexpensive. I’ve heard a lot about the ischemic preconditioning problems with glyburide. Can you elaborate more on sulfonylurea therapy in people who have just had a coronary event? Once, I was told that I’m endangering their lives by prescribing sulfonylureas and that they need to come up with the money to get more expensive therapies.

Dr. Hiatt: We have gotten a lot better at treating heart disease and diabetes outcomes because background therapy has gotten a lot better. I’m not a complete nihilist around the A1c. I do think any drug you can use that can adequately manage their metabolic problems to some appropriate level is absolutely appropriate. We’re not advocating bad therapy. We have drugs that do that and you should use them. I too am concerned that we’re adding lots of new expensive drugs with small A1c benefits, where I’m not convinced that it helps the patient that much. In your case, you’re not just worried about sulfonylurea isolation. You can use good blood pressure medications that are generic as well as good statins that are generic. You can do all the things that Dr. Eckel referred to that have a huge impact and feel good about that.

Dr. Philip Home (Newcastle University, Newcastle upon Tyne, UK): I’ve been looking at sulfonylureas for years and have published reassuring papers. More importantly, there is a supportive paper from the Diabetes Care editor’s forum, in press. These are serious people who are respected who have concluded that there are no issues of safety. From outside the US, I just wonder why you cannot get generic sulfonylureas. We pay $5 a month for gliclazide, which is a perfectly acceptable alternative.  

Q: We can use gliclazide and glimepiride for fairly inexpensive prices. And I usually do use glimepiride because you can titrate it much better. But again, those are all the choices we have. If patients get that, we’re also getting prescriptions for statins and all these other things, etc. so we have more co-pays. Sometimes we really have no choice but to use sulfonylureas. I’m just really concerned that I’m increasing their cardiovascular risk.

Dr. Home: If I can bring up another point that arises here, Jay showed us all those twenty studies on CVOTs. These studies cost five or ten billion US dollars – who’s going to pay for these? It’s going to be the person with diabetes and health insurance. I don’t offer thanks to FDA’s inappropriate attitude in driving costs of future therapies up. Patients and insurers are ultimately going to have to pay.

Q: For those of us who practice exclusively in the pediatric world, what is your take-home message for us?

Dr. Skyler: I don’t have any data on this. As Bob pointed out, there’s not adequate cardiovascular data, so you make individual recommendations to try to deal with it.

Dr. Home: The simplest answer, I think, is that in the absence of clinical trial data coming from the pediatric population. We do have clear data that you can manage CV and arterial risk. Because we do know that it is associated. The message is that within the limits of your individual track, good blood glucose control with our current therapies and good education and support groups should minimize cardiovascular risk.

Dr. Robert Eckel (University of Colorado, Aurora, CO): We are making a lot of assumption about atherosclerosis in type 1. Many people think that it’s just like type 2, but happens earlier. We’ll be having a workshop on basic science on this in the fall. I’m entering with a bias that there is more calcification and more fibrous processes. But this has never been specifically studied in type 1. We don’t have enough science in type 1 to know what to do. In adults with type 1, we really need science. Type 1 has been ignored, and many assumptions have been made.

Dr. Ratner: The ADA just recently came out with type 1 diabetes specific guidelines. And we do have some data on adolescents, but we have very little on children. With adolescents, we have the subgroup from DCCT with 25 years of follow-up. So the glycemic guidelines have been brought down to 7.5%, even for small children. We know the blood pressure benefits, even in those adolescents. But we have no data on lipids.

Dr. Marian Rewers (University of Colorado School of Medicine, Denver, CO): Regarding the clinical trial data, there’s a lot of observation from DCCT. Let me add one more bias in terms of cardiovascular disease in type 1. A major uncontrolled risk factor is peripheral hyperlipidemia – in an effort to control A1c in these patients, we’re giving way too much insulin in the peripheral system. Cardiac data show that there is insulin resistance in all of the target organs. So I think the several sessions here about the artificial pancreas should look at minimizing the dose of insulin in type 1 patients by optimization of timing. Part of the delivery is going to help tremendously down the line in terms of cardiovascular risk in this population.

Dr. Eckel: However if you think about insulin resistance, which we’ve clearly documented in type 1, ultimately, cholesterol is high and triglycerides are lower in type 1. Why is that? In general, the lipid protein abnormalities that were deemed appropriate for atherosclerosis just don’t appear to be there in type 1. The hypothesis is that this may be a different process in terms of the arterial wall.

Dr. Ratner: I would also look at the ORIGIN data to suggest peripheral hyperinsulinemia probably can’t be blamed.

Dr. Home: The associations are relatively poor between peripheral hyperinsulinemia and atherosclerosis. In terms of giving insulin, you’re improving insulin insensitivity and if you measure insulin levels, they fall in type 2 diabetes. We have to be very careful in interpreting that data. I personally don’t believe the peripheral hyperinsulinemia and atherosclerosis story.

Comment: [Dr. Rewers] We’re mixing type 1 and type 2. What we call it in type 2, insulin levels are in ranges of 20-40 units per milligram. In type 1s, it’s insulin levels after injections of short-acting or long-acting insulins at three or four times that level. I suggest that we have a separate session next year on this issue.

Dr. Skyler: The hyperinsulinemia question goes back to three papers in the 1970s that studied non-diabetic populations. A bunch of studies found that circulating levels of insulin were correlated with increased coronary risk. But when we looked at it later, it wasn’t insulin. It was insulin resistance at those tissues that’s really the culprit. Hyperinsulinemia reflects insulin resistance in those organs.

Q; We’ve talked a lot about cardiac events this morning. At the 2013 meeting of EASD, Dr. John McMurray stated that we need to make heart failure more important and that it is the most important complication of diabetes. What is your response?

Dr. Home: I know Dr. John McMurray well and he is a heart failure guy. But honestly, who is saying that heart failure is more important than cardiovascular death or myocardial infarction? In most studies these days, we do adjudicate data regarding heart failure. So that data is there already. But I would go against saying that cardiovascular death, stroke, and heart failure should be consider equally important complications of diabetes. That’s not a very intelligent way to approach that.

Dr. Hiatt: I think that the number of deaths that occur from ischemic heart disease overwhelms the number from heart failure. We should still prevent heart failure – After all, there is evidence that people with diabetes have stiff left ventricles. Other than that, you want to keep things in perspective.

Plenary: Special Issues in Diabetes

Biosimilars in Diabetes Care

Philip Home, DM, DPhil (Newcastle University, Newcastle upon Tyne, UK)

 Dr. Philip Home shared his thoughts on what should be required for clinical confidence in a biosimilar insulin, stressing that current regulatory guidelines may not be sufficient to ensure the safety and efficacy of these products. “Biosimilars are coming our way” due to the patent expiry of Sanofi’s Lantus (insulin glargine), Dr. Home noted, and there is currently a lack of clear regulatory guidelines surrounding their approval. The EU has the most detailed guidelines specific to biosimilar insulins, but Dr. Home critiqued several aspects of that guidance, saying that the confidence intervals of 20-25% for pharmacokinetic (PK) and pharmacodynamics (PD) studies are too wide and expressing strong disagreement with the EU’s statement that randomized controlled trials will likely not be necessary. He believes that it is essential to study biosimilar insulins in the most sensitive clinical population (in this case, C-peptide negative patients) and to review a totality of clinical outcomes rather than focusing on a single measure (PK, PD, A1c, pre-injection glucose, hypoglycemia, immunogenicity). He expressed concern about the lack of access to manufacturing quality data (differences in manufacturing techniques could significantly affect biosimilar performance), saying that a company’s reputation will likely be the best indication of quality for many clinicians. Dr. Home closed by reviewing data presented at ADA on Lilly’s biosimilar insulin glargine LY2963016, pointing out that while the studies appeared to be well run, the erratic PK data in type 1 patients and the wide confidence intervals for many of the efficacy measures might not individually confirm biosimilarity conclusively.

  • A review of the current regulatory guidelines on biosimilar insulins illustrated the high degree of variability in this area. The EU has been somewhat of a pioneer, with a set of guidelines for general biosimilars since 2005 and guidelines specific to biosimilar insulins more recently. However, Dr. Home sharply critiqued some of those guidelines, saying that the required confidence intervals of 20-25% for PK/PD studies were far too wide and that he fundamentally disagreed with the statement that there was “no anticipated need” for randomized controlled clinical trials. Elsewhere in the world, the guidance is even less clear – the US and Canada have only general guidelines for biosimilar approval, and, concerningly, in many other countries like Mexico, the requirements for approval are extremely lax.
  • Dr. Home believes that reviewing a “package of clinical outcomes” is the best way to ensure the clinical safety and efficacy of a biosimilar insulin. He outlined several outcomes that should be part of such a package, including PK and PD studies, A1c, pre-injection glucose, hypoglycemia, and immunogenicity, cautioning that focusing on any of those outcomes alone will not be sufficiently powered to demonstrate biosimilarity. He emphasized the need for clinical studies in the most sensitive population possible, such as C-peptide-negative type 1 diabetes patients, as any endogenous insulin secretion could buffer out any difference between a biosimilar and its reference product.
  • It will be difficult to monitor and evaluate the quality of the manufacturing process for biosimilar insulins. Slight differences in the manufacturing process compared to the reference product are inevitable due to the complexity of insulin production, but it is essential to ensure that those alterations do not change the product’s clinical profile or lead to adverse effects like antibody production. Dr. Home lamented the lack of access to data on the quality and consistency of the manufacturing process and predicted that clinicians will most likely rely on the manufacturer’s reputation to assess the quality of the product – this would work to the advantage of companies like Lilly that are well-established in the insulin market, whereas Asia-based companies such as Biocon may have a larger perception barrier to overcome.
  • Dr. Home offered a mixed assessment of LY2963016, Lilly’s biosimilar insulin glargine, emphasizing that the current data package is fairly strong but that there are issues about the strength of individual parts of the data. He praised some aspects of the comparative data presented at last month’s ADA against Sanofi's Lantus – the PK profiles of two different doses of LY2963016 in healthy subjects gave him “some confidence” that the two insulin glargines have similar absorption profiles, and he was satisfied with the data showing no significant difference in clinical outcomes between the two glargines. However, he characterized the PK data in patients with type 1 diabetes as “all over the place” and criticized the wide confidence intervals in the efficacy trials.

The Role of CGM in Achieving Glycemic Targets

Howard Wolpert (Joslin Diabetes Center, Boston, MA)

In an impressively wide-ranging talk, Dr. Wolpert spoke about the utility of CGM and strategies for maximizing the technology’s potential (an update from his presentation at ENDO 2014). Dr. Wolpert acknowledged that the decision to use CGM involves a tradeoff between significantly improving hypoglycemia awareness (“the real benefit of CGM”) and the hassles associated with sensor inaccuracies and nuisance alarms. He expressed confidence that recent technological advances have put the industry at an “inflection point” at which the overwhelming benefits will soon be realized by society. Dr. Wolpert also shared his views on interpreting CGM downloads, emphasizing the utility of “Modal Day Statistics” in identifying frequent trouble spots. In closing, he shared his Joslin group’s approach to alarm optimization, noting that their procedure of trial-and-error replicates many features of setting basal rates on insulin pumps.

  • Dr. Wolpert noted that the adoption of CGM entails a trade-off that must be assessed on a per person basis. Indeed, he suggested that the only patients for whom the quality-of-life benefits of CGM tend to significantly outweigh the issues with sensor inaccuracy, burnout, and alarm fatigue are those suffering from significant hypoglycemia unawareness and those with the greatest risk for episodes of severe hypoglycemia. However, he believes that recent technological advances are going to shift this trade-off toward favoring CGM use in the coming years – “We are at an inflection point for accelerated adoption,” he said. We imagine the upcoming move to more automated insulin delivery (e.g., Medtronic’s MiniMed 640G) and sending sensor data to mobile devices (Dexcom’s Share, Gen 5) will play a key role in this shift.
  • Dr. Wolpert highlighted the need for data examining sensor adherence, as it is a key predictor of patients that benefit from CGM. Dr. Wolpert noted that the Dexcom G4 Platinum has already made impressive strides on the accuracy and reliability fronts – especially in comparison to the “less accurate” Medtronic Enlite – but that the lack of data on sensor adherence may be limiting clinical adoption of the technology. Dr. Wolpert emphasized that randomized control trials are of little value in predicting real-world adherence, given the issues with subject selection bias and protocol-enforced adherence. We imagine data from the T1D Exchange could help in this regard, though it still has some selection bias.
  • Dr. Wolpert stressed that real-world adherence requires a proper conceptualization of the challenges. Patients are often overwhelmed by alarm fatigue and information overload, though these can both be addressed – information overload can be transformed into positive patient self-management through education, while alarm fatigue can be easily minimized. In Dr. Wolpert’s experience, measurements are often seen as a reflection of self-worth, leading patients to feel as though they are “failing.”
    • Dr. Wolpert shared Joslin’s approach to CGM alarm optimization and glycemic thresholds. He noted that the clinical challenge in setting these thresholds is to derive benefit from the alarms while minimizing fatigue. Echoing what we have heard from other experts, Dr. Wolpert called for alarms to be set quite widely at the beginning, with tightening over time as patients gain comfort with the system. He stressed the utility of setting high alarms appropriately, a feature that is of particular value during the night when it can prevent patients from trending excessively hyperglycemic. Dr. Wolpert closed by highlighting snooze alarms as an underutilized feature of CGMs, encouraging providers to consider increasing the duration between alarms to minimize nuisance.
  • Dr. Wolpert highlighted “Modal Day Statistics” as a particularly helpful way to analyze CGM downloads and identify glycemic trouble spots. Modal day statistics show data from every hour of the day, indicating the typical range of a patient’s blood glucose levels at that time. Dr. Wolpert highlighted this “bird’s-eye” point-of-view as a way of identifying when a patient is having frequent dysglycemia and the variability associated with these excursions. He stressed the clinical utility and simplicity of this information and the ability to actionably treat off this download.

Panel Discussion

Q: This links more to the conversation on cost from yesterday, but how much does a bottle of Lantus cost?

Dr. Philip Home (Newcastle University, Newcastle upon Tyne, UK): I can’t answer in your terms. In the UK we think about monthly costs, which is about $80 per month.

Comment: Dr. Hirsch, the cost of a bottle of Lantus is around $200. My concern is that pharmaceutical companies are really abusing us. But we’re in bed with them, too. We’re being held hostage. On one hand, they are sponsoring this conference, while some of our patients must give up insulin to eat. How do you justify that?

Dr. Hirsch: The fundamental problem is that this is an international problem, but the US is being asked to fund all the innovation. Other governments around the world are drawing a line in terms of how much they are paying. We don’t have that system where the government sets a price. What happens here is that if we want to see better insulins, less hypoglycemia, more rapid-acting analogs, we’re the ones who are going to pay for it. We need the international community to understand that everyone has to pay for it. I don’t know where this is going to end.

Dr. Home: You have a market and they set the market price. In other countries like the UK, there’s no market, there’s an agreement between a monopoly supplier and a monopoly purchaser. The CEO of Novo Nordisk has been hammering the table saying Europe has to pay more because the US can’t go on funding them alone.

Comment: This happened with test strips, they tried to lower the cost and it still doesn’t even matter if you have insurance, the copays keep going up and up.

Dr. Satish Garg (Barbara Davis Center, Aurora, CO): How will pay for performance affect compensation in telehealth? If you only do one visit a year and locals do three visits, what happens?

Dr. Paul Wadwa (Barbara Davis Center, Aurora, CO): That’s a good question, and we don’t know the answer yet. We provide that care even though we’re not seeing the patient in person, we are the clinicians in the pediatric clinic seeing patients. Even if they come to the remote site, our team sees them; we’re still reaching out and making contact. We might only see them in person once a year, but we’re seeing them virtually ideally four times a year.

Q: With the use of telemedicine, did you notice an increase in device usage with your patients? Maybe they were more inclined or had better access?

Dr. Wadwa: We’re analyzing that data now. In the data I showed, we were working with diabetes educators and matching them to patients. We think they were getting better access to technology, more easily. But we’re still analyzing that data.

Q: Dr. Wolpert, your slides on the accuracies and inaccuracies of CGM were older: from 2002-2007. Is there more up to date evidence? And how about SMBG, is there anything that shows the accuracy of that?

Dr. Wolpert: One slide I showed with comparative data of Medtronic and Dexcom was more recent, from 2012-13. I don’t want to overemphasize the issue, I just think it’s important for patients to appreciate that you’re using plasma glucose as a reference for CGM and you’re measuring interstitial glucose, so there’s a physiological challenge in establishing accuracy. People need to appreciate that the discrepancy enters into the accuracy assessment, that there’s a practical limit to what degree you can establish accuracy. We happen to use BGM in terms of self-management because that’s what we have. With improvements in CGM, we will be able – and many patients do – to use interstitial data. There needs to be a shift in the regulatory standpoint; people need to appreciate the nuances of how interstitial glucose changes relative to blood glucose and incorporate that. I’m not focusing on inaccuracy, just on appreciating the difference.

Comment: Well when I started, I was checking my urine and looking at colors on the side of a bottle, so I think CGM is amazing.

Dr. Mohan: Abbott has an average blood glucose profile coming out now. Given the disconnect between blood and interstitial plasma glucose, can we predict or do we have algorithms that can address this difference?

Dr. Wolpert: As you know, there have been recent studies looking at lag. From my studies, in which we’ve had two devices on a patient at the same time, it seems that there’s no consistent lag time, such that it’s going to be hard to come up with consistent recommendations. I think that there are two ways to use CGM. One is that you use it as a synthesis of major trouble spots. I think that what gets lost are the benefits that results from patient insight into self-care habits. That entails getting them to look at details.

Q: I have a question for Dr. Wadwa. When you’re selecting new sites for telemedicine, do you really screen the people or the infrastructure? Do you look for a place that’s already routinely doing A1cs or is it more about finding motivated providers?

Dr. Wadwa: It’s a combination. We get feedback that there’s a need somewhere. We’re focused on Colorado and Wyoming right now; in Durango we use an outpatient pediatric office because they’re eager to work with us, in other places we work with a diabetes center because they have access to tools.

Plenary: Obesity and Type 2 Diabetes

Pharmacological Approaches to Weight loss

Daniel Bessesen, MD (University of Colorado Denver, Denver, CO)

Dr. Daniel Bessesen gave an engaging presentation on weight loss medications and the role of weight management in diabetes. He emphasized that although weight loss medications are often not discussed with patients, pharmacotherapy of obesity is “mainstream,” and many professional societies (NHLBI, ACP, AACE, Endocrine Society) have now included weight loss drugs as part of their guidelines. Dr. Bessesen referred to NHANES data showing that weight loss medications do help people lose weight, but noted that the use of the old generation of weight loss medications actually fell from 2008 to 2010. Regarding specific pharmacotherapies, he provided an overview of the new generation of agents, including Vivus’ Qsymia (phentermine/topiramate) and Arena/Eisai’s Belviq (lorcaserin) noting that Qsymia comes out ahead of Belviq with regards to weight loss. He brought up the idea of prescribing phentermine along with Belviq, but noted that the safety and efficacy of this approach is unclear. Dr. Bessesen also mentioned Novo Nordisk's liraglutide 3.0 mg (which will be discussed during its FDA AdComm on September 11) and Orexigen’s Contrave (naltrexone/bupropion) as potential new medications. Dr. Bessesen emphasized the importance of providing weight loss medication information to people with diabetes honestly and accurately, concluding that he hopes he has made attendees “feel more comfortable discussing the risks and benefits” of weight loss medications with their patients.

Novel Approaches to Weight Loss

Robert Eckel, MD (University of Colorado, Denver, CO)

Dr. Robert Eckel highlighted several new approaches to weight loss beyond the standard set of lifestyle, surgical, and pharmacological options. He began by discussing Aspire Bariatrics’ AspireAssist technology, which allows patients to “aspirate” (drain) a portion of their stomach contents through an endoscopically-implanted tube. Results from a small proof-of-concept study (n=18) demonstrated that the AspireAssist removed 25-30% of the calories from each meal and led to ~20% average weight loss over 52 weeks (“that beats all pharmacotherapy”) as well as higher self-reported quality of life. Though Dr. Eckel said that he personally finds the idea of gastric aspiration “kind of unappealing myself,” survey results suggest that a large majority of patients find the procedure acceptable. An ongoing 10-center pivotal trial involving 175 patients is expected to complete in April 2015 (ClinicalTrials.gov Identifier: NCT01766037).

  • Dr. Eckel also expressed cautious interest in the endoluminal barrier approach (which GI Dynamics is developing as the EndoBarrier), which attempts to mimic the physiology of gastric bypass surgery in a less invasive manner by inserting a sleeve into the small intestine to block nutrient absorption. This method has not yet been studied in humans, but it has been shown to lead to modest weight loss and substantial improvements in glucose tolerance in rodents.  He also briefly evaluated the gastric balloon approach (he is skeptical of its ability to truly alter body weight regulation) and transoral gastric suturing, which limits stomach volume by tying together regions of the mucosal lining (preliminary data suggests it may be effective).

Role of Bariatric Surgery in Obesity & Diabetes

Phillip Schauer, MD (Cleveland Clinic, Cleveland, OH)

Bariatric surgery proponent Dr. Phillip Schauer presented data showing that surgery results in superior clinical outcomes compared to medical treatment – read our coverage of an Ethicon product theater at AACE for another recent presentation on this topic from Dr. Schauer. He first reminded us of the Look AHEAD study results, in which intensive lifestyle intervention led to weight loss initially but most of the weight was regained within ten years and no reduction in cardiovascular events was observed. From this, Dr. Schauer stressed that greater, longer sustained weight loss is necessary to produce clinical benefits, claiming that bariatric surgery is the principal way to achieve this. He presented background data showing that bariatric surgery results in improvements or resolution in hypertension, diabetes, and hyperlipidemia – drawing from the STAMPEDE trial (which we covered in-depth at ACC 2014). Dr. Schauer concluded that the evidence behind bariatric surgery is making it more mainstream, promoting it as a worthy option for people with uncontrolled diabetes and obesity.

Patient/Provider Panel Sessions

Adult

In one of the most moving and inspiring sessions of the conference, several patients with type 1 diabetes joined Drs. Satish Garg, Peter Gottlieb, Irl Hirsch, Aaron Michels, Jay Skyler, and Howard Wolpert onstage to share their stories and discuss pertinent issues in diabetes management. The audience of HCPs was blown away by the patients’ experiences – one patient said he had recently celebrated his 80th birthday after living through 66 years of type 1 diabetes and over 500 seizures due to hypoglycemia; another described how he fell in love with exercise and lost almost 200 pounds after struggling with his weight for years. The conversation eventually turned to the urgent need for Medicare coverage of CGM – both patients and providers expressed immense frustration with the futility of their efforts to obtain coverage, though one patient had succeeded after going through a lengthy appeals process and writing multiple letters explaining that he could quite literally die without the device. One audience member expressed hope that change could be possible if enough doctors and patients “were to start screaming and hollering at these people to get a move on,” but Dr. Hirsch predicted that Medicare would want to see a specific clinical trial showing a difference in outcomes between CGM and non-CGM users in Medicare patients before changing the policy. Dr. Satish Garg said that when Medicare turns down CGM coverage, patients’ secondary insurance may pay for the technology – while not a perfect solution, this could be a useful workaround for some patients.

  • The session concluded with an informal poll of the panelists’ interest in an artificial pancreas device – five of the eight panelists said yes, they would take it in a heartbeat; two said they would potentially be interested but would need more information, and Dr. Robert Eckel said that he would turn it down for the same reason he turned down the Medtronic MiniMed 530G (“if it’s working, don’t break it”).

Industry Sponsored Dinner: Dexcom

Continuous Glucose Monitoring Technology

Jake Leach (VP, Research and Development, Dexcom, San Diego, CA)

Mr. Jake Leach provided a wide-ranging summary of Dexcom’s pipeline, updating us on the company’s cloud-based SweetSpot software (no timeline provided, but clearly in the works), ongoing feasibility trials on Gen 6, and a mention of the new Dexcom/Insulet collaboration first announced on Day #1 of ADA 2014. Factory calibration of devices was highlighted as an upcoming milestone (“Don’t know when, but it’s coming”) that will facilitate patient convenience and adherence, as will smaller transmitters, new algorithms to support longer sensor life, and accuracy/reliability in non-adjunctive use – great to hear Dexcom is working on all these fronts, and we wonder how the company will prioritize each piece. Mr. Leach also reviewed the company’s current and upcoming products, including the G4 Platinum (adult, pediatric, and professional use all approved); Animas and Tandem pump integrations (no timeline provided; the Animas Vibe has been at the FDA for over a year; the integrated Tandem t:slim was supposed to be submitted in June); Dexcom Share (“available very soon”); the new G4 AP algorithm (data shared at ADA showing nearly comparable accuracy to fingersticks), and Gen 5 mobile platform. Notably, CEO Mr. Terry Gregg said in Q&A that no decision has been made on whether to make the G4 AP algorithm available soon or to wait for Gen 5. He emphasized that Dexcom remains committed to releasing a new product “every 18 months” – the G4 Platinum was originally released in October 2012 (22 months ago), though approvals have recently come for a G4 Platinum pediatric indication and a professional use indication, and Share should be available any time now – as of Dexcom’s 1Q14 call in May, it was characterized as in “the final stages of review.”

  • CEO Mr. Terry Gregg said in Q&A that no decision has been made on whether to make the G4 AP algorithm available soon or to wait for the Gen 5. The first study of the G4 AP algorithm was shown in a late-breaking poster #75 at ADA 2014, demonstrating a sub-10% MARD and nearly comparable accuracy to fingerstick meters. There is the possibility that the G4 AP algorithm could be done via a remote web update, though that could only occur after it is approved by the FDA. Given the tentative timeline to launch Gen 5 mobile platform sometime next year, it may make sense to roll the innovations into one FDA submission.
  • Dexcom’s cloud-based SweetSpot software will be Mac compatible, solving a gripe of many patients/providers, who are unable to download the G4 Platinum using the Windows-only Studio software. As a reminder, the SweetSpot system will integrate data on blood glucose, CGM, and insulin delivery in an online platform. The last update on this platform came in 3Q12 (!), where Dexcom’s goal was an FDA submission by the end of 2012 – there has been no timing update since. This is an area where Medtronic is definitely ahead of Dexcom with its cloud-based CareLink software, particularly in the minds of providers we’ve heard at recent conferences. We’ve seen previews of some of the sleek SweetSpot reports in conference exhibit halls, and cannot wait to see the advanced features (e.g., pattern recognition, glycemic variability stats). Mr. Leach emphasized that the newest update solves the Mac compatibility issues that has plagued the Windows-only Studio software, a statement that received a rousing round of applause from the audience.
  • Ongoing Gen 6 feasibility trials are aggressively testing Dexcom’s next generation sensor membrane system (no timeline provided), and the company is focused on moving toward factory calibration (“Don’t know when, but it’s coming”). The Gen 6 platform will build upon the improvements of Gen 5, including a redesigned phone-compatible transmitter. The combination of a new algorithm and sensor should support longer sensor life, enhanced accuracy and reliability in non-adjunctive use, and entirely mitigate signal responses acetaminophen.
  • Dexcom is also hoping to integrate CGM data with the new “HealthKit” feature of Apple’s iOS 8 (release scheduled for Fall 2014). This Apple-developed tool will consolidate data from various health and fitness apps to offer a one-stop-shop for all the data. We love the idea, particularly because it would allow easy pairing of exercise apps with glycemic data. To date, this has only been possible through manual tagging and logging.
  • Dexcom is creating an employee-funded foundation that seeks to assist patients who cannot afford its products. Mr. Gregg emphasized that product development at Dexcom is a balance of considering the financial burden on patients (“98% of our patients in US have some type of third-party payers”) and outmaneuvering competitors.

Real Time CGM and Clincial Practice

Jay Skyler, MD (University of Miami, Miami, FL; Board of Directors, Dexcom)

In line with his presentation from Keystone 2013 and AACE 2014, Dr. Jay Skyler reviewed the therapeutic benefits of CGM and commented on the present competitive landscape of the industry. As a reminder, he is on the Board of Directors of Dexcom. He compared the performance of Dexcom’s G4 Platinum to Medtronic’s Enlite and Abbott’s Navigator using multiple clinical studies, including the recently published head-to-head-to-head study from Damiano et al. (Journal of Diabetes Science Technology, 2014). As we’ve come to expect, Dr. Skyler was quite critical of the Medtronic Enlite, provocatively referencing data that its previous Sof-Sensor was actually more accurate than the newer device (Calhoun et al., Diabetes Technology and Therapeutics 2013). That study compared the accuracy of the Enlite (Veo algorithm) to the Sof-Sensor (Guardian Real-Time algorithm), finding that the overall median sensor vs. reference difference was -15 mg/dl for the Enlite and -1 mg/dl for the Sof-Sensor (p<0.001), while the median absolute relative difference was 15% for the Enlite vs. 12% for the Sof-Sensor (p=0.06). Dr. Skyler critically joked, “I’m not sure why Medtronic chooses to go from bad to worse when they allegedly make improvements.” We appreciate Dr. Skyler’s data-driven discussion, though would note that study data is conflicting on the accuracy of the Enlite. The same work from Drs. Ed Damiano and Steven Russell actually suggests the opposite – based on two related studies, Medtronic’s Enlite appeared more accurate than the Sof-Sensor. The group’s 2014 study found an overall MARD of 17.9% for the Enlite (Damiano et al., JDST 2014), while the 2013 study found an MARD of 20.3% for the Sof-Sensor (Damiano et al., Diabetes Care 2013). The later paper notes that these studies are comparable, since the same Navigator sensor was used in both studies and yielded accuracy results that were not significantly different.

  • Dr. Skyler was far less critical of Abbott’s technologies relative to Medtronic, which was no surprise considering Abbott’s CGM is no longer available in the US. However, we assume speakers’ talks will increasingly begin mentioning the upcoming launch of FreeStyle Libre in the EU later this year, as the factory calibrated, 14-day wear product certainly merits a role in any competitive landscape discussion of CGM.
  • Dr. Skyler noted that Dexcom is working towards broader pharmacy coverage of CGM in an effort to reduce the hassle and paperwork of CGM adoption and “make life easier” on providers. As noted in our AACE 2014 coverage, Dexcom signed a contract with CVS/Caremark in May to have CGM covered as a pharmacy benefit for an impressive ~25 million covered lives. This built on Dexcom’s addition to Express Scripts’ Preferred Drug List in January (pharmacy benefit coverage for ~5 million members). Pharmacy benefit coverage means that patients would only need to pay a copay to get CGM, rather than the average ~20% co-insurance on sensors, transmitters, and receivers. The additional benefit of moving to a pharmacy benefit is it helps patients avoid the high deductibles typically associated with DME.
  • As a reminder to the audience, Dr. Skyler summarized the findings from landmark CGM studies: the JDRF CGM trial (JDRF Study Group, Diabetes Care 2009) and SWITCH (Conget et al., Diabetes Technology and Therapeutics 2011). Both studies demonstrated clear relationships between CGM use and reductions in A1c and risk of severe hypoglycemia.  

How Patients with Type 1 Diabetes Translate Continuous Glucose Monitoring Data into Diabetes Management Decisions

Jeremy Pettus, MD (University of California San Diego, San Diego, CA)

Dr. Jeremy Pettus presented intriguing data on how people with type 1 diabetes use continuous glucose monitoring (CGM) trend arrows to make major modifications to their insulin therapy, a follow-up to Dr. Steve Edelman’s initial presentation at ATTD 2014 and two posters at ADA 2014 (836-P and 858-P). Dr. Pettus first highlighted that the current recommendations for adjusting insulin based on real-time CGM data only call for dose changes on the order of 10-20%. His group surveyed CGM users with both type 1 and type 2 diabetes, and this talk discussed the data from the 222 participants with type 1 (very well controlled, with a mean A1c of 6.9%). Respondents filled out an online survey of 70 multiple choice and scenario-based questions that explored their use of real-time CGM data. Results revealed much larger changes in insulin doses than current recommendations suggest – 41% and 48% dose reductions when responding to falling trend arrows (one and two arrows, respectively) and 111% and 140% increases in doses when responding to rising trend arrows. Dr. Pettus’ take home message was that “the rate of change information provided by CGM devices is just as important, if not more important, than the actual glucose value.” He concluded by pushing for more research on the development of dose recommendations based on trend arrows in different clinical situations as well as for this information to be better distributed to patients.

Trend Arrow

Mean Patient Adjustment to Bolus Dose*

Example Bolus Dose*

-->

---

3.0 units

^
(single up arrow)

+111%

6.3 units

^^
(double up arrow)

+140%

7.2 units

v
(single down arrow)

-41%

1.8 units

vv
(double down arrow)

-48%

1.5 units

*“It has been four hours since your last dose of insulin and meal and your CGM receiver shows a value of 220 mg/dl (matching your fingerstick blood glucose of 220 mg/dl) with an arrow and trend graph that is flat (straight across). If you are not planning on eating or exercising, what dose of insulin would you give yourself to bring your glucose to 120 mg/dl?”

Panel Discussion

Mr. Terry Gregg (CEO, Dexcom, San Diego, CA): I’ve been chasing CGM since I joined MiniMed back in 1994. I’m going to hand over the reigns now to Kevin Sayer. What I’ve realized is that translational information is important. We’ve got to take information and make it actionable. For example, we recently entered into partnership with Joslin to create a database to educate non-diabetes specialists. As we look at the number of diabetes patients that are seen by specialists, it’s only 38% for those over the age of 18. So, the majority of patients are being treated by non-specialists, who we know are only getting something like eight hours of diabetes-related training. So, we’re trying to create a database that educates other doctors so that they can provide more adequate care.

Kevin Sayer and I also hear, “You guys need to do something with big data.” I don’t even know what that term means. When I asked big data specialists, they say, “We don’t have any idea what this is going to do. We just want to collect it and analyze it later on.” Stayed tuned on this idea of translational data. I’m spending a lot of my time on how we can use and share information.

Our CGM is a class III medical device. It’s highly regulated by the FDA. Even though you heard the FDA official earlier today [Dr. Courtney Lias] say “Go for the endgame,” I will say that going for the endgame is very difficult. Stay tuned, a lot more will come of that. We asked the FDA how fast can we transition into this? They said that if it’s real-time actionable, then it’s a class III medical device, and you have to submit it as such. So hurdles remain. Although we want to do all these things and are capable, we are highly regulated and the regulatory process does have challenges at times. Take Gen 5, for instance. It’s been developed for two years. All companies in this industry have these restrictions. We can only go as fast as the FDA wants us to, but we work in a very close, collaborative relationship with the FDA.

Q: This is a question related to the cloud presentation for CGM. As a parent with multiple children, will there be capabilities to screen child to child? If so, how soon will this be out?

Mr. Jake Leach (Vice President, Research and Development, Dexcom, San Diego, CA): The Dexcom Share app can actually follow up to five people. You can even put pictures next to each person, so you can easily tell them apart. The product was submitted to the FDA last summer, so we do expect it to be available very soon.

Q: So if we’re part of a provider team and we have more than five patients, it doesn’t really leave room for that?

Mr. Leach: It wasn’t really designed for healthcare providers. It was mainly designed for parents or family members. But that’s a future idea.

Dr. Jay Skyler (University of Miami, Miami, FL): As a physician, I wouldn’t want to follow all my patients’ data all the time. [Laughter]

Q: I’m relieved to hear that the Gen 5 receiver will be on the phone. I often lose the receiver. If there was an app on the phone that could make a noise on the receiver so I could find it, I would love that.

Dr. Skyler: That’s awesome feedback. Thank you. But you could lose your phone, too.

Q: Are you guys working towards Medicare approval? I’ve been type 1 for a long time and I will not leave my CGM at home. I’m going to be 65 in three years. It’s a scary thought.

Mr. Gregg: We have an individual who’s working with CMS and she’s working on getting approval. We’re also partnered with Medtronic and with professional societies because we all have a common goal. And that’s greater access to that population. My greatest agony as CEO is to see someone with coverage at 64 and then lose that coverage. I think it’s criminal and we’re doing the best job we can possibly do. I don’t want to candy coat it because I think we’re still looking at two years. There’s a lot of skin in the game but a lot of players.

Dr. Robert Ratner (American Diabetes Association, Alexandria, VA): Non-adjunctive use. That’s the term CMS is focusing on. They categorically turned us down with that saying that therefore, CGM is not necessary. Until you get non-adjunctive use dealt with, and I’d love to hear about that, CMS isn’t going to cover. As soon as you get rid of that, we’ve got all the ammunition in the world to get CMS coverage.

Mr. Gregg: I’ll give you update on that actually. FDA is well aware that a significant portion of G4 users dose off their CGMs. So we asked them what we have to do to achieve this. We have a close, collaborative relationship with the Agency.

Q: The cloud idea is great from the user’s point-of-view, but from provider’s point-of-view, it would be great if there were more efficient integration with EHR. Have you reached out to major providers regarding this?

Mr. Leach:  Excellent question. Our folks have had a lot of experience with integrating into EMRs. A lot of that integration was unique for each healthcare setting it was developed for. Additionally, one of the things we’ve been looking at is a new feature in Apple’s iOS 8 called the “HealthKit.” It’s a way to share health data on an iPhone. We’re hoping that plugs into EPIC and makes this more feasible. We need to get that data without the effort.

Q: Dr. Pettus, I was interested in the data you collected on hypothetical systems. Were patients they making these changes based on hypothetic data without fingersticks? Did you ask them about different situations specifically?

Dr. Jeremy Pettus (University of California San Diego, San Diego, CA): We did ask them what they would do in particular situations. At night, most said they would dose without confirming a CGM reading with a fingerstick. In fact, the majority of people would dose CGM at night without confirming it. Or treat a low without confirming it.

Q: It’s impressive how much more accurate the G4AP algorithm is. You also mentioned how it handles SMBG. Maybe it can detect faulty SMBGs?

Mr. Leach: G4AP is quite a bit smarter than the G4 Platinum algorithm. It looks at the sensor signal and looks at the SMBG reading. It doesn’t believe exactly what the SMBG reading says because of potential error. We’re trusting our CGM sensor more than the SMBG reading. The sensors have gotten so stable that we can do that now.

Q: I have a quick note before my question. Accuracy data gives us confidence as clinicians. I think we get too bogged down in data without thinking about what it means to patients. And improved accuracy leads to more adherence and better outcomes. Ultimately that improves quality of life, which unfortunately is something, as a clinician, that I don’t necessarily get to feel like I’m doing on a day-to-day basis when I’m reminding patients to take insulin, etc. My question is about G4 AP. Will it be available soon? Or will we have to wait till the Gen 5 is available?

Mr. Gregg: No decision has been made. We have approval for downloading software through the Internet. The challenge that Dexcom faces is that we like to iterate quickly. One of the regulatory challenges we have is that when you have the Gen 5 on the horizon, it complicated our decision-making. Which version do you put into which product? It all depends on what’s going to be approved. I can’t answer your question yet. But as soon as we can, we’ll make a statement. It’ll be the Gen 4 Platinum or Gen 5. It’s really based on the timing of approval. We do things concurrently. We don’t wait for things to happen. After all, because of the regulatory challenges at times, we have to be working on things in parallel. Our goal is to get a new product into your hands every 18 months, which requires parallel processing.

Q: I have an application question. You talked about parents watching their kids play sports with their G4s. What kind of sports? Can kids play football with them?

Mr. Leach: I know of two cases of them being used in soccer and basketball. I also know kids play football with them. We even have NASCAR drivers wearing them. They do find places to wear them.

Comment: I have a patient who’s a national BMX racer, and he sports his diabetes paraphernalia with pride.

Q: Dr. Pettus, how has CGM made a difference in your life?

Dr. Pettus: I was just thinking about that. The biggest difference is the feeling of safety. Really feeling like I’ve finally got a handle on my diabetes. I believe that the Gen 4 is the real deal to the point that I really do trust it. In fact, I would rather go back on regular NPH than give up my CGM. My control has always been pretty good, but my ease of achieving that control has changed dramatically. It just means so much to know what your blood glucose is doing all the time. It’s always on your mind to see where you are at. It seems simple, but it’s a big deal.

Q: When you say you’re not happy unless you’re putting out a new product every 18 months, I think it’s important to note that the patient isn’t happy with that. There are financial constraints. Rather than putting out new products that we need to buy, I think it’s more ideal to just update them.

Mr. Gregg: I agree with you completely. I’m probably the most patient-centric CEO on this planet and I’m very sensitive to that subject. I’ll announce here the employees of Dexcom have created a foundation, but not as a company. We’re cash flow positive and we can’t take shareholder money for this. But, we wanted to do something for patients who can’t afford our products. So we are in process of creating an employee-funded foundation that we, including the executives, fund so that we can take care of patients who are needy. That’s kind of the culture of the company. Now there are some things we can’t do when we transition to certain technologies. Ninety-eight percent of our patients in the US have some type of third-party payers. We try to time things appropriately as we can. But at same time, we’re racing. Jake gets told every day that we need to make a replacement for fingersticks. We have to do this in a stepwise fashion and we’re making great strides toward that. I will remind the audience that we spend 30% of our budget on R&D while other companies spend less than 10%. We’ve never taken the position that we do what we do for financial reasons. We do it for patients. I hear your message and it very much resonates to the management team here.

 

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