July 16-19, 2015; Keystone, CO Day #1 Highlights – Draft

Executive Highlights

Greetings from Rocky Mountain High, where the Barbara Davis Center’s Practical Ways to Achieve Targets in Diabetes Care (ATDC) Conference welcomed 600 participants – 25% more than last year and its largest attendance ever – to its 25^th annual conference. Researchers, educators, and members of industry hailed from 44 states and four countries – the US, UK, Germany, and China! – gathering in scenic Keystone, CO where we managed to avoid the anticipated drizzle and enjoy a sun-splashed opening day.

Company-sponsored technology workshops kicked off our day … and did not disappoint! Close Concerns Associates attending the Roche session were present for the US debut of the Accu-Chek Connect system (Bluetooth-enabled BGM + smartphone app with bolus calculator), while those attending the parallel event got a brief peak at another “Connect” ~ Medtronic’s MiniMed : >. Afternoon sessions shifted gears to look at outcomes, with plenary sessions that focused on ADA standards of medical care, the future of closed-loop therapy, lessons learned from the FLAT-SUGAR, and non-insulin treatment options in type 2 diabetes. Our report from day #1 features ten highlights along with an appendix of selected expanded talks.

1. Roche’s hands-on workshop debuted the Accu-Chek Connect BGM and paired smartphone app (with a bolus calculator!) in the US, sharing the planned August 7 launch date and reasonable $29.99 pricing.

2. A comprehensive overview of insulin injection from Dr. Irl Hirsch (University of Washington, Seattle, WA) raised questions about the lack of rigorous data on lipohypertrophy and featured a compelling analysis of its economic costs.

3. In a separate lecture, Dr. Hirsch defended the success of his FLAT-SUGAR study and raised a number of important follow-up questions.

4. Dr. Larry Hirsch confirmed that BD’s new FlowSmart infusion set will come to market next year in the US, Canada, and the EU

5. Medtronic’s Dr. Robert Vigersky presented recently published (Gehlaut et al., JDST 2015) data on the prevalence of hypoglycemia in type 2 diabetes as measured by blinded CGM. In a separate morning workshop, Medtronic’s Dr. Robert Mucic provided an introduction to MiniMed Connect, which generated a lengthy Q&A on the product details and cost.

6. Dr. Bruce Bode (Atlanta Diabetes Associates, Atlanta, GA) reviewed the pros and cons of non-insulin therapies for type 1 diabetes, expressing the greatest optimism for GLP-1 agonists.

7. An afternoon panel discussion sparked debate on the cost of diabetes therapies and a bit of international rivalry.

8. During a panel discussion, Dr. Robert Ratner (ADA, Alexandria, VA) argued that outcomes trials in diabetes are unrealistic due to low event rates and instead recommended the use of surrogate outcomes across the board.

9. Dr. Ratner also walked attendees through the ADA 2015 Standards of Medical Care, highlighting the growth and trends of the guidelines’ evidence as well as the field’s needs to move recommendations forward.

10. Dr. George Alberti’s (Imperial College, London, UK) overview of the last 50 years in diabetes treatment concluded with the striking assertion that the incorporation of self-management and patient empowerment into therapy has been the most significant development in the past 50 years.

Top Ten Highlights

1. Roche’s hands-on workshop debuted the Accu-Chek Connect BGM and paired smartphone app (with a bolus calculator!) in the US, sharing the planned August 7 launch date and reasonable $29.99 pricing. Area Business Manager Jeremy Mangelson ran us through the system, highlighting that the nationwide launch during AADE will come to market through an exclusive distribution agreement with Walgreens – a bold move indeed! The meter is priced reasonably at $29.99, roughly the same as the cost of LifeScan’s OneTouch Verio Sync (the only other Big Four smartphone-connected BGM), and the app is already available for free on Apple and Android devices [though you cannot do much with it without the meter]. The next-generation system is based on the Aviva Plus strip platform and features wireless Bluetooth compatibility with the paired apps. Most notably, the Connect app is the first FDA-approved mobile platform that offers insulin-dosing recommendations – a tremendous win for MDI users (who do not have access to bolus calculators on an insulin pump) and for Roche (first-to-market status with a bolus calculator mobile app). Roche confirmed that the insulin bolus feature will only be available to patients who purchase the Connect meter; those with other meters will be able to enter glucose values, but will be unable to use the Bolus Advisor function. We understand this from a competitive perspective, though it would be an amazing win for patients (and tremendous legacy for Roche) if the app was available to all meter users. Though the app is a free download, it is a “prescription” product (not the meter, just the app!), as patients must receive an HCP code to activate the Bolus Advisor software. The Connect also sends readings automatically to the cloud and can notify caregivers of values, a nice feature for parents. Our full report below features our first in-depth look at the meter, app, and web portal.

2. A comprehensive overview of insulin injection from Dr. Irl Hirsch (University of Washington, Seattle, WA) raised questions about the lack of rigorous data on lipohypertrophy and featured a compelling analysis of its economic costs. Dr. Hirsch opened by suggesting that lipohypertrophy studies to date have lacked rigor (e.g., often uncontrolled), resulting in many unanswered fundamental questions in the field. Fortunately, Dr. Hirsch believes that the quality of the research on lipohypertrophy is improving – indeed, he presented new data demonstrating the first, real scientific evidence both that insulin injection into regions with lipohypertrophy vs. normal adipose tissue worsens postprandial glucose control (17% higher in the first two hours, 58% higher from 2-5 hours, 39% higher > 5 hours) and leads to 4.5x more variable insulin absorption – “We’ve sort of known this for decades, but we’ve never had real data to show it until now.” The lack of answers, in his view, has contributed to tremendous waste that he estimated was on the order of billions (!) of dollars: ~$2 billion/year (as a conservative estimate) vs. ~$4.6 billion (as a “realistic” estimate). See the detailed report for Dr. Hirsch’s full assumptions along with commentary that lipohypertrophy does not affect CGM accuracy. On an unrelated note, Dr. Hirsch closed his presentation by advocating that physicians examine newly diagnosed patients for heart murmurs (in addition to injection sites, of course), since they can affect A1c readouts. He will present more on this soon; yet another reason we should move beyond this metric alone!

3. In a separate lecture, Dr. Hirsch defended the success of his FLAT-SUGAR study and raised a number of important follow-up questions. As a reminder, results from the study (n=102 patients with type 2 diabetes) presented at ADA demonstrated that it is possible to achieve a modest but significant difference in glycemic variability between two groups (exenatide + glargine + metformin vs. rapid-acting insulin + glargine + metformin) while maintaining a similar A1c. During this talk, Dr. Hirsch reviewed the results and emphasized that the trial successfully met its primary endpoint, even if it could not achieve a “whopping difference” in variability. He was also “delighted” that all secondary variability endpoints and hypoglycemia trended in the right direction, though they were not statistically significant – he attributed this to lack of power since fewer patients than expected completed the trial. Dr. Hirsch closed by raising several key questions for his group moving forward, including whether effects of exenatide itself (independent of reduced variability) could have confounded the results, why there was so little hypoglycemia in the rapid-acting insulin group, and ultimately whether such a modest difference in variability is sufficient to support a definitive trial examining the effect of variability on long-term outcomes – we believe the latter is the $200 million question! These questions illustrate some of the tension inherent in clinical trials between the desire for clearly interpretable data vs. applicable “real-world” results. For example, Dr. Hirsch rhetorically asked whether it ultimately would matter from a clinical perspective whether outcomes were improved due to reduced glycemic variability or another effect of a GLP-1 agonist. On hypoglycemia as well, the results could very well have been different with another comparator group. While it is clear that designing a sufficiently simple, well-powered trial to examine the effect of glycemic variability on outcomes will be challenging, we are certainly rooting for Dr. Hirsch’s team!

• On a related note, we heard Dr. Guido Freckmann (University of Ulm, Germany) highlight CSII therapy as a tool to reduce glycemic variability. He pointed to Swedish registry data showing that pump therapy reduced variability and mortality compared to MDI – this was an observational study though certainly directionally interesting.

4. Dr. Larry Hirsch confirmed that BD’s new FlowSmart infusion set will come to market next year in the US, Canada, and the EU – in line with Medtronic’s previous timeline to launch in 2016 (as per the partnership announcement signed at ADA). He also presented new results from a follow-up trial to BD’s pilot study first presented in oral form at ATTD 2015 (an expanded 60 patients vs. 25 shown at ATTD). Consistent with the previous results, the BD set significantly reduced the amount of time with flow interruptions (p=002) and reduced the risk of pressure events by ~75%. There were no observed statistical differences in set site leakage upon removal compared to the Quick-set (1 vs. 0 events, respectively). We believe this set is an important advance and look forward to longer, real-world studies and wearing it ourselves; as closed-loop trials have shown, bringing things out into the real world from the clinic is a whole new endeavor. More details on this talk are below.

5. Medtronic’s Dr. Robert Vigersky presented recently published (Gehlaut et al., JDST 2015) data on the prevalence of hypoglycemia in type 2 diabetes as measured by blinded CGM. The trial enrolled 108 patients with type 2 diabetes that wore Medtronic’s iPro blinded CGM for five days. Hypoglycemia was frequently documented and unrecognized by the patients in this study – 49% had at least one hypoglycemic episode and 75% of those patients experienced at least one asymptomatic hypoglycemic episode. As expected, hypoglycemia was seen frequently in individuals on insulin (66%) and those on oral hypoglycemic agents (81%; sulfonylureas). However, hypoglycemia was also observed in patients on non-hypoglycemia causing agents like DPP-4 inhibitors and metformin – we wonder if this stemmed from CGM inaccuracy at lower glucose values (e.g., 70-90 mg/dl) that read as hypoglycemia, but actually weren’t. Dr. Vigerksy stressed that the mechanism behind this hypoglycemia remains unknown, but is an active area of interest for Medtronic. Ultimately, he noted that the findings resulted in treatment modification in 64% of patients, leading Dr. Vigersky to conclude that retrospective CGM data can provide valuable feedback in this population. We agree that both blinded CGM and real-time CGM have lots of potential in type 2 diabetes, though there are many keys to get right: (i) making it cost-effective enough; (ii) reducing provider burden; (iii) building meaningful data output to drive changes in type 2 diabetes medications. Certainly, Medtronic needs to think about this in a big way as Abbott comes to the US with FreeStyle Libre and presumably the blinded version, FreeStyle Libre Pro – both offer meaningful improvements on the three aforementioned criteria. Dr. Vigersky did provide a brief overview – but did not provide any updates – on the status of Medtronic’s closed-loop efforts (the MiniMed 670G pivotal study began during ADA), the new insulin infusion partnership with BD (commercialization in early 2016), and the coming Fall 2015 launch of MiniMed Connect.

• In a morning workshop, Medtronic’s Dr. Robert Mucic provided an introduction to MiniMed Connect, which generated a lengthy Q&A on the product details and cost. This remote monitoring keychain device – slated to launch this fall for $199 – was unveiled at ADA 2015 and subsequently shown last week in the FFL exhibit hall. Notably, we heard attendees raise concerns regarding cost, specifically of patients who may not afford an iPhone or computer, to which the Medtronic team responded by noting their efforts looking at various financial aid and payment plan options. We thought this was a valid but somewhat unfair criticism, given the device’s reasonable price, no ongoing cost, and target at existing pump/CGM users. Medtronic’s Ms. Kim Larson importantly stressed MiniMed Connect’s advantage in its opportunity for providers to be reimbursed on a monthly basis for reviewing CGM data (the device uploads data automatically to CareLink). We believe cloud-based technology like MiniMed Connect and Dexcom’s Share/Gen 5 can help drive better outcomes, especially as reimbursement structures shift towards remote population management.

6. Dr. Bruce Bode (Atlanta Diabetes Associates, Atlanta, GA) reviewed the pros and cons of non-insulin therapies for type 1 diabetes, expressing the greatest optimism for GLP-1 agonists. He took a fairly dim view of pramlintide, the only currently approved adjunctive therapy for type 1 diabetes, as very few people remain on the drug long-term due to side effects and the need for frequent injections and challenging dosing. Perhaps more surprisingly, he also offered a fairly skeptical take on metformin, stating that the risk of GI side effects overwhelms the modest benefits on weight and insulin dose in most cases. That said, he acknowledged that if the University of Glasgow/JDRF-funded REMOVAL study, scheduled to complete in 2016, is able to demonstrate cardiovascular benefit with metformin in type 1 diabetes, it could absolutely become standard of care. Dr. Bode expressed greater optimism about the potential of GLP-1 agonists, highlighting promising results from a study presented at AACE 2014 demonstrating improvements in a number of parameters with Novo Nordisk’s Victoza (liraglutide) in type 1 diabetes. Excitingly, he shared that Novo Nordisk’s phase 3 ADJUNCT ONE and ADJUNCT TWO studies on Victoza in type 1 diabetes have completed data collection and should hopefully report results later this year – as of the company’s 1Q15 update, results should be available by the end of October. On the “hot topic” of SGLT-2 inhibitors in type 1 diabetes, Dr. Bode reviewed the promising data from phase 2 trials but acknowledged that the risk of euglycemic DKA needs to be further investigated. On the whole, while acknowledging that these drugs can “do wonders,” he ultimately recommended a fairly conservative approach, stating that he would not prescribe these drugs off-label until their efficacy and safety are better understood. However, he also presented T1D Exchange data showing that some patients and providers are clearly taking matters into their own hands: off-label use of non-insulin therapies has risen across the board from 2010-2012 to 2014-2015.

7. An afternoon panel discussion sparked debate on the cost of diabetes therapies and a bit of international rivalry. Dr. George Alberti (Imperial College, London, UK) suggested that his preferred approach to treating type 2 diabetes would be a stronger initial emphasis on education, followed by metformin and a sulfonylurea if needed, with newer, more expensive drugs used only with caution or in situations where the benefits are very clear (such as GLP-1 agonists for very overweight patients). Dr. Robert Ratner (ADA, Alexandria, VA) countered by citing recent ADA cost analyses demonstrating that prescription drugs account for a very small share of total spending on diabetes compared to hospitalizations for complications. The discussion got a bit heated as attendees from the US and the UK debated the methods and consequences of the UK’s National Institute of Health and Care Excellence’s (NICE) cost-effectiveness analyses for diabetes therapies. The US and many European countries have clearly taken divergent paths in recent years in terms of cost and access to newer therapies, with (as a generalization) US patients often gaining quicker access to a wider range of new treatment options but at substantially higher cost (to the healthcare system and to individual patients who lack coverage). Ultimately, we do not see this gap as sustainable over the long term, particularly given the increasing focus on reducing costs within the American healthcare system.

8. During a panel discussion, Dr. Robert Ratner (ADA, Alexandria, VA) argued that outcomes trials in diabetes are unrealistic due to low event rates and instead recommended the use of surrogate outcomes across the board. After Dr. Satish Garg (University of Colorado, Aurora, CO) stressed that most trials are too short or too solely focused on safety parameters to identify hard cardiovascular outcomes, Dr. Ratner pointed to the low event rate of major cardiovascular events in diabetes. Citing the low event rate of 3% to 3.5% in the ELIXA trial’s high-risk population, he questioned how outcomes trials can provide clinically relevant information and specifically suggested that Dr. Irl Hirsch (University of Washington, Seattle, WA) pursue surrogate outcomes in his follow-up study to FLAT-SUGAR. We are encouraged to hear continuing discussion around the limitations of the current CVOT paradigm and believe this is an important perspective to consider, though we realize there are plenty of data purists at the FDA and elsewhere who are unlikely to agree that outcomes trials should be abandoned entirely.

9. Dr. Ratner also walked attendees through the ADA 2015 Standards of Medical Care, highlighting the growth and trends of the guidelines’ evidence as well as the field’s needs to move recommendations forward. Dr. Ratner opened his presentation by showing that the Standards of Medical Care have had a “remarkable increase in the quality of evidence” as over half of the current guidelines’ evidence are now graded at either the “A” or “B” levels. Specifically, he touched on the broad influence of the guidelines’ recommendations for testing diabetes, as both the US and Community Preventive Services Task Forces have recently followed these guidelines. However, he noted that understanding the rapid development of type 2 diabetes in children remains a significant unmet need, as the minimal evidence in this area leaves its guidelines at only an “E” grade level. In addition, he stressed the importance of diabetes self-management education, commenting that it has been challenging to conduct randomized controlled trials in this area but that expert opinion agrees that such education should be paid for by insurance companies. Regarding pharmacotherapy treatment, Dr. Ratner highlighted that we are “truly moving toward a patient-centered approach” in the context of glycemic goals, but noted that the data behind the individualization of anti-hyperglycemic therapies are “woefully inadequate.” On the technology front, Dr. Ratner pointed out that evidence surrounding glucose monitoring is strong but that data on hypoglycemia unawareness continues to be sparse and that CGM’s biggest limitation today is its adjunctive requirement of SMBG. Additionally, he touched on the guidelines’ discussion of dyslipidemia and lipid management, pointing to the recent changes on statin recommendations – he believes that these will evolve as they are “highly problematic” for patients who lie on the guidelines’ boundaries. Ultimately, we applaud the ADA’s leadership in providing such a comprehensive set of guidelines that address diabetes’ many multifactorial factors from comorbidities to racial differences – for more on how these recommendations have most recently evolved, please see our coverage of this year’s changes.

10. Dr. George Alberti’s (Imperial College, London, UK) overview of the last 50 years in diabetes treatment concluded with the striking assertion that the incorporation of self-management and patient empowerment into therapy has been the most significant development in the past 50 years. Dr. Alberti introduced his lecture as a counterpoint to coming talks that would delve deep into cutting-edge treatment, drugs, and devices. Instead, he provided a comprehensive overview of how the diabetes landscape has changed, highlighting in particular the sheer difference in number of patients: fewer than 100 million in 1980 vs. an estimated 385 million in 2013. He then reviewed progress in glucose monitoring – from urine glucose in the 1600s to the introduction of A1c in 1977 to sensors and CGMs today. Next up were developments in treatment, from diet and oral agents to insulin and insulin delivery. Dr. Alberti did mention that he is not completely convinced by the viability of stem cells as a type 1 diabetes treatment or cure. He concluded by suggesting that the most tremendous change in the past half-decade has been an increasingly patient-centered approach to care. We certainly appreciated his patient-centered spirit, especially amidst all the biomedical innovations he reviewed. We couldn’t agree more!

Detailed Discussion and Commentary

Practical (Hands-on) Experience with CGM and/or CSII Treatment

Introduction of the Accu-Chek Connect System

Jeremy Mangelson (Area Business Manager, Roche Diabetes Care, Indianapolis, US)

Roche’s hands-on workshop debuted the Accu-Chek Connect BGM and paired smartphone app (with a bolus calculator!) in the US, sharing the planned August 7 launch date and reasonable $29.99 pricing. Regional Business Manager Jeremy Mangelson ran us through the system, highlighting that the nationwide launch during AADE will come to market through an exclusive distribution agreement with Walgreens – a bold move indeed! The meter is priced reasonably at $29.99, roughly the same as the cost of LifeScan’s OneTouch Verio Sync (the only other Big Four smartphone-connected BGM), and the app is already available for free on Apple and Android devices [though you cannot do much with it without the meter]. The next-generation system is based on the Aviva Plus strip platform and features wireless Bluetooth compatibility with the paired apps. Most notably, the Connect app is the first FDA-approved mobile platform that offers insulin-dosing recommendations – a tremendous win for MDI users (who do not have access to bolus calculators on an insulin pump) and for Roche (first-to-market status with a bolus calculator mobile app). Roche confirmed that the insulin bolus feature will only be available to patients who purchase the Connect meter; those with other meters will be able to enter glucose values, but will be unable to use the Bolus Advisor function. We understand this from a competitive perspective, though it would be an amazing win for patients (and tremendous legacy for Roche) if the app was available to all meter users. Though the app is a free download, it is a “prescription” product (not the meter, just the app!), as patients must receive an HCP code to activate the Bolus Advisor software. The Connect also sends readings automatically to the cloud and can notify caregivers of values, a nice feature for parents. Our full report below features our first in-depth look at the meter, app, and web portal including pictures and an overview of the smartphone/cloud-connected market.

• The Accu-Chek Connect is an important move toward more seamless, automatic, and hassle-free transfer of blood glucose data. Roche’s laudable aim is to reduce the patient and provider hassle in the downloading process by removing the need for manual data entry. The Connect builds on a trend of smartphone/cloud-connected meters – see below for a competitive landscape of the industry. Notably, the launch of the Connect marks the second smartphone-connected BGM commercialized by a Big Four company (LifeScan’s OneTouch VerioSync). We believe much of the diabetes device industry is moving in this direction, and it will only be a matter of time before most devices (BGM, CGM, pumps) are cloud/smartphone app connected. (Google’s 2Q15 earnings call today reminded us of the power of mobile: More Google searches now take place on mobile devices than on computers in 10 countries, including the US and Japan.)
• The biggest highlight of the Connect system is undoubtedly the app – see pictures below – which provides the convenience of insulin bolus advice on a smartphone based on blood glucose results and entered carbohydrates. The FDA clearance marks the software as the first mobile platform cleared for insulin-dosing recommendations. The app is a “prescription” product, as patients must consult with a provider to receive an HCP code that activates the Bolus Advisor software. It’s interesting that this is required, since there are many free bolus calculator apps on the app store right now that are unapproved by the FDA. There’s certainly a danger in the wrong settings, but there’s also a danger in not using a bolus calculator app at all. We wonder if Roche proposed the HCP code, or if the Agency required this. HCP involvement is of course critical for proper bolus calculator setup, so the requirement is understandable. 
• The app stands as big win for US insulin users on MDI (the vast majority of type 1s and 2s) who do not have access to bolus calculators on an insulin pump, and a big win for Roche (first-to-market status with a mobile bolus calculator). We have yet to see rigorous data that such an advisor can improve glycemic control, though anecdotally this seems like a no-brainer (and Roche is reportedly working to get these data published). Dosing insulin – even with the best technology today – is still a guessing game at best, and the math is hard even with ratios that end in “0” and “5.” Down the road, perhaps Roche and others will build algorithms to better optimize bolus calculator settings – certainly, Medtronic is thinking about this with the IBM Watson partnership.
  
  • Roche confirmed that the insulin bolus feature will only be available to patients who purchase the Connect meter. Those with other meters will be able to enter glucose values, but will be unable to use the Bolus Advisor function. We were disappointed to hear this, though not surprised as it makes a lot of sense from a competitive advantage point of view. We wonder if patients on other meters would be interested in using the app, and whether Roche would reconsider this stance – certainly, it would be an amazing legacy for the company.  Bolus calculators are tremendous tools and MDIs only have access to unapproved – and frankly not very good – free versions on the app store.
• Glucose readings can automatically be sent (via text) to up to ten additional mobile devices – a huge patient and caregiver convenience that enables remote monitoring and peace of mind. Roche is coming out with strong marketing towards the pediatric population, positioning the monitoring as a way to reduce the burden of pediatric diabetes management and to enable kids to live a more normal life (e.g., attending sleepovers).
  
  • Additional tabs within the mobile app offer insight into time-in-range, average number of tests per day as well as pages for viewing results. Among these is a convenient page to summarize paired points (e.g., before vs. after breakfast) that offers a great way for patients to connect the dots between behavior change and clinical results (e.g., a 15-minute walk actually reduces blood sugar!). Roche has done some outstanding work on the BGM education front, and we salute the company for going beyond simply display of glucose values.
• Upon testing, the Connect app also prompts users to enter the bolus advisor page. This tab offers the chance to log carbs and take a picture of the meal. We are particular fans of this latter feature, which can create “teachable moments” – after all, carb counting is hard and this enables providers/caregivers to compare the carbs logged with a picture of the actual meal. The following tab provides the patient with the bolus advice, showing graphically the amount of insulin on board and amount of insulin to be injected. Data is automatically sent to the online portal.
• We were also provided a short demo of the online web portal, which comes in both provider- and patient-facing versions. The provider-facing side enables population management in a notable way, showing all of a provider’s patients on one screen including a symbol to indicate hypoglycemia risk (we like this being front and center!). Following pages in the portal enable providers to see the number of times patients are overriding bolus advice, to change a patient’s insulin factors and ratios, to view a color-coded logbook, or to view a graphical display of a patients’ average blood glucose over time.
• It’s not clear how Roche is thinking about interoperability (e.g., Apple’s Health Kit, Glooko, Tidepool, diasend). Ideally, the glucose results should flow freely between apps, enabling the ecosystem that many in diabetes technology hope for. Our test drive of Meal Memory and the Dexcom Share receiver has given us a lot of optimism over targeted apps that leverage passively collected data to drive meaningful therapeutic change – let us know if you are interested in a sneak preview of this article. We hope Roche is strongly considering data interoperability with these growing platforms.
• We were also intrigued to learn more about the timeline over which various functionalities of the Accu-Chek Connect System received clearance: (i) the Connect BGM was cleared in March 2015; (ii) the Connect app for Android platforms was cleared in late March; and (iii) the Connect app for iOS systems was cleared in early June. Notably though, the Android version of the app was recalled in April (before being launched in the US) after clearance due to problems with inaccurate bolus advice associated with changes of the screen orientation (e.g., from portrait to landscape or vice versa) during calculations. The app had been downloaded a total of 644 times (Italy: 219, South Africa: 24, Germany: 401). As we understand it, the software bug was fixed “within a few days.” Both the Apple and Android apps are currently available on the Apple and Android stores.

The Accu-Chek Connect App – Bolus Advisor

The Past, Present, and Future of Diabetes Injection/Infusion Therapy – According to Hirsch²

Thinking About Exogenous Insulin and the Skin: Lipohypertrophy

Irl Hirsch, MD (University of Washington, Seattle, WA)

Dr. Irl Hirsch provided a comprehensive overview of lipohypertrophy and insulin injection. He opened with a persuasive case that lipohypertrophy is dramatically underappreciated. Unsurprisingly, Dr. Hirsch’s informal poll of the audience showed that 100% of providers working with patients had challenges with lipohypertrophy – “Everyone sees this!” Unfortunately, Dr. Hirsch stressed that the problem with this area of research is that lipohypertrophy has historically not been rigorously examined – studies are often uncontrolled or anecdotal, resulting in many unanswered fundamental questions: Can lipohypertrophy tissue regress? Does needle change frequency matter? What is the ideal site change frequency? What sort of intervention can shift behaviors toward more optimal injection techniques? Fortunately, Dr. Hirsch believes that the quality of the research on lipohypertrophy is improving. Indeed, he closed his presentation by sharing new data from two recent BD studies from ADA 2015 to illustrate that the body of evidence is growing – we summarize those studies below. On a very interesting note, Dr. Hirsch closed his presentation by advocating for a change in the way physicians examine newly diagnosed patients – listening for heart murmurs, since they affect A1c readouts.

• Hovelmann et al. (ADA 2015): This BD study compared the effects of subcutaneous insulin injection into lipohypertrophy and normal adipose tissue regions on postprandial glucose control and insulin absorption in 13 patients with type 1 diabetes and lipohypertrophy. Patients received two identical mixed meals (75 g carbohydrates) separated by at least 6 hours, and each covered with 0.15 U/kg insulin lispro. Insulin injection into lipohypertrophy regions had a pronounced negative effect on insulin absorption and postprandial blood glucose control compared with injection into normal tissue – mean postprandial glucose levels were significantly increased (17% higher in the first two hours, 58% higher from 2-5 hours, 39% higher > 5 hours, all p<0.05) and postprandial hypoglycemia (glucose ≤ 50 mg/dl) occurred slightly less frequently with lipohypertrophy injection (2 vs. 6 patients, p=0.20).
  
  • Though unsurprising, Dr. Hirsch stressed that this is the FIRST scientific evidence we’ve had that insulin injection into regions with lipohypertrophy worsens postprandial glucose control vs. injections into normal adipose tissue. “We’ve sort of known this for decades, but we’ve never had real data to show it until now. This to me is really important.”
• Famulla et al. (ADA 2015): This follow-up study to Hovelmann et al. investigated the pharmacodynamics of insulin lispro injected into abdominal areas with lipohypertrophy or normal adipose tissue in the same 13 patients. Dr. Hirsch stressed that this was the first-ever euglycemic clamp study to look at glycemic control in lipohypertrophy patients. Patients received single doses of 0.15 U/kg lispro every ~6 hours, twice into a region with lipohypertrophy and twice into normal tissue region in random order, under clamp conditions (blood glucose target: 100 mg/dl). Ultimately, findings confirmed the clinical experience that both insulin absorption and action are substantially blunted and considerably more variable when insulin is injected into areas with lipohypertrophy – results indicated that the variability in insulin pharmacodynamics was 4.5x higher with lipohypertrophy vs. normal adipose issue.
• We were intrigued to hear Dr. Hirsch’s commentary that lipohypertrophy does not affect CGM accuracy. Dr. Bruce Buckingham reportedly presented this data in an abstract at ADA 2015 though we – and Dr. Hirsch, himself – have been unable to track down the hard data. In Dr. Hirsch’s words, the poster showed “that lipohypertrophy had zero impact on CGM accuracy and lag time … I don’t have the data but I know what I saw!” WE are following with Dr. Buckingham on this front.
• Dr. Hirsch presented a compelling analysis of the economic costs of lipohypertrophy: ~$2 billion/year (conservative estimate) vs. ~$4.6 billion (a “realistic” estimate, in his words). Dr. Hirsch’s model drew from statistics from the literature. Assumptions came from both the CDC National Diabetes Statistics Report (2014) and multiple published studies.

Table 1: Economic Burden of Lipohypertrophy

• On a very interesting note, Dr. Hirsch closed his presentation by advocating for an adjunct to the way physicians examine newly diagnosed diabetes patients – of course, examining injection sites for proper rotation but also listening for heart murmurs. He suggested that patients with heart murmurs will have A1c values that read low! Indeed, his claim drew many murmurs (“oohs” and “aahs” … not heart murmurs) from the audience. He shared that heart murmurs indicate the sheering of red blood cells as they pass through valves. As a result, these cells do not live the presumed 90 days and lead A1c values to underreport glycemia. He will be presenting additional data to support his claims on Sunday (see our Keystone Preview here), and we will looked forward to hearing more about this notable phenomenon.

The Past, Present, and Future OF Insulin Infection / Infusion Therapy

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

The highlight of this comprehensive review of insulin delivery was commentary on BD’s new set, expected to launch in 2016 in the US, Canada, and the EU – as stated in the BD-Medtronic collaboration agreement announced at ADA in Boston. Dr. Hirsch did not share any other major new details on the set, though we appreciated expanded results from BD’s pilot study first presented in oral form at ATTD 2015.

• Dr. Larry Hirsch presented encouraging data from an expanded study comparing BD’s new infusion set to Medtronic’s Quick-set in 60 healthy participants (ADA 2015 Poster: 1071-P) – the previous pilot study shown at ATTD was only in 25 participants. As a reminder, the study had each participant wear two of each set; diluent was used and pressure was tracked for basal and bolus infusion; it was powered to show at least a 50% reduction in pressure rise events. Consistent with results from BD’s pilot study, the findings were impressive – the BD set significantly reduce the amount of time with flow interruptions (p=0.002) and reduced the risk of pressure events by ~75%. There were no observed statistical differences in set site leakage upon removal compared to the Quick-set (1 vs. 0 events, respectively).
• Preclinical in vivo tests against the Medtronic Quickset, the Animas Inset, and the Roche UltraFlex also showed large, significant reductions (~75-80%) in pressure events. BD conducted a first-of-its-kind MRI imaging study (a poster at ADA) showing increased subcutaneous diffusion of diluent vs. other marketed infusion sets. 
• Dr. Hirsch reminded us of the set’s design and features, including the side-ported 6 mm, 28 gauge polymer catheter (allowing insulin to flow out of both the bottom and side) and the 30 gauge introducer needle (the smallest available). For a close-up experience with this set at BD’s headquarters, read Closer Look coverage here.
• Similar to his ATTD talk, Dr. Hirsch also summarized previous studies showing approximately 2/3 of CSII patients experienced at least one episode of unexplained hyperglycemia over 13 weeks, with most of such events not triggering the pump occlusion alarm (“Silent Occlusions”).
• Dr. Hirsch began with an inspiring photo-illustrated story of his journey with diabetes, starting nearly 58 years ago with U-40 glass syringes, a 25-gauge/16 mm needle, a sharpening stone, and a boiling case for sterilization – “good control was an oxymoron for the first 30 years.” When Dr. Hirsch graduated high school, his “severe disability” (type 1 diabetes) earned him a full merit scholarship from the state of New York for college and medical school. In contrast, he noted, “We would laugh now if someone suggested a scholarship for diabetes.” Indeed, how far we have come!
• Dr. Hirsch presented unpublished data from BD – an international survey of > 13,000 patients from 41 countries at > 400 centers – to make the highly persuasive case that “there has been a true ‘shift to short’ [needles].” He shared new data indicating that a substantial portion of current insulin-injectors on pens use 4 mm and 5 mm needles, emphasizing that there has been a drastic shift from five years ago when 50% of patients were on 8mm needles and none on 4mm. The preliminary data are still in processing and cannot yet be published in detail. This is an impressive testament to BD’s work to use hard science to develop and support its products, which have ultimately benefitted patients through less pain and a less intimidating injection experience.
• “Needles that we call ‘short’ are actually the right length.” Dr. Hirsch reviewed several studies of skin and subcutaneous fat thickness at injection sites and of studies comparing different length pen needles and glycemic control. The latter have all demonstrated consistent findings: (i) 4 or 5 mm pen needles deliver insulin reliably to subcutaneous tissue, including obese patients; (ii) short pen needles provide equal glycemic control as do long needles, even in obese patients; (iii) patients prefer shorter, finer gauge needles; (iv) injection site anatomy matters, with almost no one having skin thickness over 3.5 mm; and (v) the risk of intramuscular injection is higher with longer needles and 2-4 times higher in the thigh vs. the abdomen (the two most commonly-used injection sites), in males vs. females (the latter have more SC adipose thickness), and in patients with lower BMI. As a reminder, IM injection results in more rapid and potentially unpredictable insulin absorption, especially with mild exercise of the muscle site. 

Plenary One: Ways to Improve Diabetes Outcomes

Panel Discussion

Dr. Irl Hirsch (University of Washington, Seattle, WA): Here’s my problem: I’m seeing these patients and as we’re developing these new tools, our patients can’t afford our old tools. George, I think you come from a better perspective than any of us. In the UK, you’ve really focused on the cost efficacy better than we have in the United States. I think GLP-1 agonists are fabulous drugs, but we don’t see as many using them as I think could benefit from them and I think the main reason is not the nausea but the cost. I’m involved in these clinical trials and I’m amazed at what I see with the research, but how are we going to afford these new tools to treat type 1 and type 2 diabetes when right now we can’t even afford what we’re using even with insulin? George, I think your perspective is maybe a little more mature than the rest of us because you’ve been dealing with this longer than we have.

Dr. George Alberti (Imperial College, London, UK): I think by “more mature” you just mean I’m older. [Laughter] It is insane and I have reflected on the cost and whether in fact we shouldn’t be spending money we do have, certainly in the UK, on more patient education and nurse educators and saying, “Ok that comes first.” Then metformin, which is cheap, comes next. And my European colleagues would probably not agree with this but sulfonylureas are still there and still can give a good effect. Be cautious about the further drugs you need. In the very overweight, once you’ve done your education program, then I think there is a very strong case for the GLP-1-type compounds because of the beneficial effects on weight.

Dr. Robert Ratner (ADA, Alexandria, VA): We do a cost of diabetes analysis every five years. This is exclusive to the United States. It’s a very in-depth look at what society spends on diabetes, both direct and indirect costs. We published in 2013 the survey that was done in 2012. The previous one had been done in 2007. We found that the cost of diabetes in United States had gone up 42%. The cost per person went up less than medical inflation. So it was all driven by the increase in the number of people developing diabetes. When you break down the cost of diabetes in this country, over 40% comes from hospitalizations. Physician office visits are about 10%. Glucose lowering agents and glucose monitoring supplies? Not even 11%. So in fact, we’re not spending the money on all these expensive drugs; we’re spending the money on the people coming into the hospital with heart attacks and amputations and strokes. Remember the old commercial: pay me now or pay me later? If we really look at the data in both type 1 and type 2 diabetes on controlling complications, if we are aggressive in outpatient management and utilize the drugs we have available in a cost-effective manner, we will cut down on hospitalizations. How many months of GLP-1 therapy do you think a 10-day stay in the CCU will cover?

Dr. Alberti: I still think you need to be sensible. I don’t think you need very expensive insulin analogs at all.

Dr. William Tamborlane (Yale University, New Haven, CT): It is a slippery slope, this cost-effectiveness analysis. First of all, it is the most unscientific way of trying to assess value. I could call it voodoo medicine. In the UK, they have a committee called NICE – which, by the way, always means the opposite of what it sounds like – it is not nice. It tries not to approve new drugs and devices because they cost more money. Every government program wants to reduce costs. Cost-effectiveness analysis ultimately revolves around quality-adjusted life-years gained. You have to say that with this intervention, the diabetic patient is not only going to live a year longer, but a year of good quality. The NICE committee, to assess quality of life changes with a drug, uses what’s called the EQ-5D. It’s five questions, three of which are: I can dress myself, I’m not bedridden, and I’m not anxious or depressed. At least two of those are not going to change over six months of using a GLP-1 agonist or a new rapid-acting analog.

Dr. Alberti: I think you’re showing profound ignorance of what NICE does. We have the chair of the NICE type 1 diabetes committee here. Perhaps she can put things right.

Dr. Stephanie Amiel (King’s College London, UK): It is one of the measures, but I’m sorry, you are wrong about how it works. If I went into great detail about it, it would take us quite a long time. Like you, I approach my task with deep skepticism, but they do the best they can with the available evidence and they look at a range of things including hypoglycemia, HbA1c, and quality of life. They go to extreme effort to get patient input as well as healthcare professional input.

Dr. Tamborlane: I talked to someone on the NICE committee and she said it was safety that gets some of the type 2 drugs approved. We know that certainly in type 2, safety related to hypoglycemia is a very soft outcome. As we pointed out in our two talks, Guys Hospital was one of the leading CSII centers late 70s and early 80s – they were our rivals at Yale. In 2000, there were more children on pumps in our clinic at Yale than there were pumps in entire United Kingdom. And that was because it was felt that there was not enough evidence to support the use of pumps.

Dr. Amiel: The NICE guideline on pumps in the UK, were it to be implemented, would have 20% of adult patients and most of our children on pumps. The problem’s not NICE, it’s the people who are trying to implement their recommendations.

Question: I’d like to ask a more philosophical question that has to do with where we’re putting our resources as diabetes experts and professionals. I see anybody who comes in and I do best to get that person to the best control safely. But the patients I most enjoy seeing are those that are doing worst, who I can make a small intervention and make an enormous difference in terms of outcome. The frustration I have is that as an endocrinologist, I spend a disproportionate amount of time on people whose A1c’s are in the 7% range who want to do better and I try to help them to do better. I see relatively few people who are in the 9, 10, 11% and above, where I think I could have greatest impact with the smallest number of drugs at the least cost. How do you try to get to this extremely vulnerable group where evidence extremely strong that with an intervention you could make an enormous difference?

Dr. Bruce Bode (Atlanta Diabetes Associates, Atlanta, GA): In US, as you see patients, you keep them as a chronic condition and you don’t discharge. In the UK, I think it’s probably different. Their primary care takes care of everybody and if they aren’t in control they get referred. Theoretically, the only way you could do that is to see a lot more new patients and be open to say, “Give me your people above 8 to 9 and anybody below that go back to the primary care.”

Dr. Tamborlane: For pediatrics, we have weekly clinic meeting of all providers. We spend 90% of time in that meeting talking about patients with A1c’s over 9%. We never move anywhere with those patients, it’s often the same relatively small group. I think that the hypothesis that you can have a bigger impact on the really poorly controlled patients is not always the case. The problem is that we have challenges in pediatrics in improving parenting and getting kids to take medicines that they don’t want to take. It is really not so easy. And poverty. Many studies point out that low socioeconomic status, with all things that contribute to poverty, is an important factor in overwhelming control.

Dr. Alberti: I think one of the advantages we do have in the UK is a total care system. One program we’re been involved with is a team who goes for the hard to reach groups, the homeless etc. They’re having a really dramatic impact on the 20s, 30s, and 40s of patients – not hundreds of patients – that they can get at and field their cardiovascular risk factors and improve their diabetes quite dramatically.

Dr. Hirsch: It’s a generalization, it’s not always the case, but at least where I live people are willing to drive up to 3 hours to come to our clinic if they have type 1 diabetes. The type 2 population won’t drive across our lake. It’s a problem, absolutely. What we’re doing – and I haven’t seen data yet, I hope to in the next few weeks – is sending a junior endocrinologist to work in a primary care clinic. It’s not such a novel idea, but they don’t have expertise or more importantly the infrastructure in primary care. If we bring that to them, we will improve the care. Because you’re right, there are too many people with double digit A1c’s in those clinics.

Dr. Ratner: The smartest providers I’ve ever known are the ones who know what they don’t know. Essentially what that means if they hit their limit of capability and then they refer. Any of you who heard David Marrero’s presidential lecture at ADA will recognize that one area where we are least capable of dealing with our patients is the behavioral and mental health component. Depression is so common in diabetes. The inability to cope with everything we’re asking them to do and yet how often do we actually make the referral. When you look at all the clinical trials, they all have an upper limit of A1c. You don’t want people with A1cs in 11% and 12% because that’s not a medical problem, that’s a behavioral and mental health problem. I think we need to shift our attention to mental health issues and then we can get to those folks with A1c’s of 9 and above.

Dr. Hirsch: There are a lot of exceptions to that though, especially in type 2 patients in rural communities. They have little access to medical care. There are still a lot of places like that. The big thing that I see is the misdiagnosis – these patients come in almost catatonic because they have type 1 diabetes.

Dr. Ratner: That’s the knowing what you don’t know and when to refer.

Dr. Hirsch: That’s right, and we see that a lot.

Dr. Satish Garg (University of Colorado, Denver, CO): So Bruce, you talked about DKA with SGLT-2, what is your hypothesis for this combo product? With SGLT-1 and SGLT-2, do you anticipate a different response to glucagon and thus an outcome for DKA?

Dr. Bode:  In the dual SGLT-1 and SGLT-2 inhibitor study, we did measure GLP-1 levels and they were numerically higher; I am not aware if we measured glucagon levels.  We did have two cases of DKA in the sotagliflozin group, both of which were related to dislodged or crimped infusion sets.  So I do not know if  DKA events will be less with this new dual SGLT-1 and 2 inhibitor. Did we measure glucagon levels – were they down?

Dr. Garg: If we did not measure them, we should.

Dr. Tamborlane: I think the obvious gorilla in the room with SGLT-2 is with type 1, with our pump patients. As opposed to type 2, it’s really the type 1 patient on a pump who is regularly going to have infusion site problems.

Dr. Tamborlane: Dr. Ratner has just told us you can’t depend on randomized controlled trials either because you don’t accept patients with A1c over 10 or renal failure or on 10 other drugs. I love randomized trials, we do randomized trials all the time, but it’s not rocket science.

Dr. Ratner: Guidelines are suggestions.

Q: I’m from Hawaii and I work with a lot of Medicaid patients. I feel like I’m in my own little world because we have managed Medicaid. We have five plans and every plan – except Kaiser – covers GLP-1 agonists with some rules. Someone got the message that people in poverty tend to be more obese and will benefit from these drugs. I’m curious if anyone else has generous similar Medicaid coverage?

Dr. Hirsch: Let me ask, raise your hands if you are able to get GLP-1 agonists for your patients on Medicaid. [Decent number of raised hands] Wow that’s impressive.

Comment: I’ve had more problems with Medicare patients with poor Part D coverage and exorbitant copays. Now that everyone knows that insulin prices are ridiculous, I have drug representatives supplying patients with pens. These patients are a few thousand dollars over the cutoff Medicaid. They’re too rich for assistance, but not rich.

Dr. Hirsch: All of us deal differently. There are a lot of local and regional differences – where you live it’s easier to get GLP-1, where I live it’s easier to get marijuana.

-- by Melissa An, Adam Brown, Helen Gao, Varun Iyengar, Emily Regier, and Kelly Close