Keystone 2018 (Practical Ways to Achieve Targets in Diabetes Care)

July 12-15, 2018; Keystone, CO; Days #3-4 Highlights – Draft

Executive Highlights

  • Greetings from San Francisco, where we’ve been hard at work preparing our second installment of Keystone 2018 Highlights. Below, you’ll find our top eight highlights in therapy, eight more in tech, plus five in additional topics – and don’t forget to check out the first two days of the meeting, too!

  • Diabetes therapy: Headlining the drug side of things, Dr. Ralph DeFronzo tackled combination therapy, revealing that if he resigned EDICT today, he would use triple therapy of GLP-1/SGLT-2/pioglitazone – kicking out metformin for the new SGLT-2 class. Following, Dr. Irl Hirsch called for a definitive study on the importance of glucose variability in type 1, reviewing tangential evidence from pregnancy supporting such a notion. More than anything, there was a strong type 1 focus in the last two days of Keystone. Dr. Stuart Weinzimer highlighted the potential for pramlintide or SGLT inhibitors in combination with closed loop, Dr. Antoinette Moran called for earlier intervention with combination therapy in type 1 prevention/modification, and Dr. Linda DiMeglio praised Afrezza + Fiasp despite the inherent PK/PD limitations of mealtime insulin, also introducing the DIY insulin movement. We also heard about the role of epigenetics and autoimmunity screening in type 1, as well as the state of the disease in under-resourced countries.

  • Diabetes technology: Day #3 of the meeting was headlined by another impressive keynote address from FDA’s Dr. Courtney Lias on the integrated CGM (iCGM) pathway, the possibility of iPump and iAlgorithm pathways, what’s happening with DIY systems like OpenAPS, and even work to better help companies work through cost/access issues. In a notable industry update, Dexcom CEO Mr. Kevin Sayer shared that direct G6 transmitter-to-Apple-Watch will happen “sometime next year,” as well as boldly noting that “over the next five years, there will be as much innovation on getting people to pay for technology as there will be in getting new devices out.” Yes! Senseonics’ Mr. Mike Gill announced that the first US Eversense insertions are slated for July 31 in “a big event,” and we learned the 180-day Eversense XL might be approved in late 2019/early 2020 (US trial to start enrolling in mid-fall). Tandem confirmed the early August launch of the t:slim X2 Basal-IQ with G6, and we appreciated a rundown of the prescribing and software update process. We also heard from Drs. Hirsch and Ratner on CMS CGM coverage issues, Dr. Satish Garg on installing micro-filters in insulin catheters (to filter out insulin excipients), Dr. Tim Bailey on patch pumps and expanding CGM, and Dr. Helen Murphy’s unpublished UK CONCEPTT cohort food log results (51% carb diet!).

Our team has descended back to sea level after a stellar Keystone 2018! In our Days #3-4 Highlights for the meeting, you’ll find a tremendous depth of KOL commentary on combination therapy, glucose variability, insulin and more, followed by another series of tech updates (from FDA, Dexcom, and Tandem, to start), plus a set of additional topics.

If you didn’t catch our Days #1-2 Highlights, you can find those here – every single day of this meeting packed a huge punch, so these are not-to-miss!


Table of Contents 

Diabetes Therapy Highlights

1. Dr. DeFronzo Continues to Push for Combination Therapy, Advocating for GLP-1/SGLT-2/Pioglitazone Over the GLP-1/Pio/Metformin Combo Used in EDICT; Maligns Continued Use of SUs + DPP-4s

Following the presentation of very positive six-year EDICT results at ADA 2018, Dr. Ralph DeFronzo shared that, if he had to redesign the study today, he would test triple therapy of GLP-1/SGLT-2/pioglitazone, rather than the GLP-1/metformin/pioglitazone combo used in the trial. While Dr. DeFronzo’s “Fab Four” of diabetes management includes GLP-1s, SGLT-2s, pioglitazone, and metformin, he has made it clear that metformin is the least favored of the four – and SGLT-2s weren’t on the market when EDICT began. Indeed, we also have to wonder that GLP-1 agonist Dr. DeFronzo would use today; exenatide was available first, but we feel confident that semaglutide especially, but also liraglutide or dulaglutide, would give even better results. Always taking a pathophysiological approach, Dr. DeFronzo outlined how these therapies all target the underlying defects of type 2 diabetes, referencing his now-classic ominous octet; in his words, “diabetes is not hopeless…combination therapy works very well if you’re picking the medications that target the basic pathophysiology of the disease.” Pushing for the aggressive early adoption of combination therapies, he explained that GLP-1 agonists correct six of eight defects: They suppress hepatic glucose production, improve neurological dysfunction and decrease appetite, improve peripheral glucose uptake through weight loss, improve beta cell insulin resistance/incretin effects, increase insulin secretion, and tamp down glucagon secretion from the alpha cell. TZDs complement action at the beta cell, liver, and peripheral muscle and also prevent lipolysis to reduce plasma free fatty acids, and metformin further suppresses HGP. But adding an SGLT-2 would target the final component of the ominous octet: increased renal glucose reabsorption. Dr. DeFronzo highlighted this mechanism as different from the others, as well as the weight loss and blood pressure benefits not offered by metformin. Thus, a GLP-1/SGLT-2/pioglitazone combination targets all components of the ominous octet, and would also offer cardioprotection through, to our understanding, three different mechanisms (not to mention likely renal-protection). We fully support this movement away from single-minded A1c lowering and toward improving outcomes by driving at the pathophysiological cause of disease, and even more so the idea of early, aggressive therapy to prevent progressive loss of beta cell function and give long-term reduction of CV risk factors. He implored the audience to move away from the algorithm of metformin, SUs, and insulin, as this treatment strategy does nothing to preserve beta cell function – or, we might add, lower CV risk. For reference, at three years, EDICT’s triple therapy demonstrated improvements in A1c vs. conventional therapy (5.8% vs. 6.7%, p <0.0001), supported by corresponding improvements in insulin sensitivity (p <0.0001), beta cell function (p <0.0001), weight loss (-1.5 kg vs. +3.7 kg, p <0.0001), and ~eight times fewer hypoglycemic events (p <0.0001). Positive results were maintained at six years.

  • Dr. DeFronzo lamented the continued use of SUs, which he called “as close to worthless as any drug you can get,” emphasizing that, as insulin secretagogues, these agents target no aspect of diabetes’ underlying pathophysiology. As such, on a long-term basis, SUs just won’t work; Dr. DeFronzo doesn’t love insulin either: “It hits everything over the head with a sledgehammer.” Recently, we’ve actually heard some fairly impassioned defenses of SUs, particularly those considered better within the class (e.g. gliclazide), and we’ve long known that there is some heterogeneity within the class – so they are not all as bad as the worst! But notably, we find it very difficult to support the widespread use of any therapy that causes weight gain and hypoglcyemia and doesn’t offer an ounce of cardioprotection or even long-term glucose lowering efficacy; indeed, the field knows this – but it also knows SUs are one of very few cheap diabetes therapies (more on this below). We’re glad for Dr. DeFronzo’s reminder that the basic science of SUs and diabetes is completely illogical, especially as far as long-term treatment and prevention is concerned.

  • Additionally, Dr. DeFronzo called DPP-4 inhibitors “pretty much worthless,” echoing comments from Drs. Jay Skyler and Steve Nissen, who have expressed a similar lack of enthusiasm for the class and deprioritized its use in their clinical practice. Dr. Nissen has been, perhaps, DPP-4s most outspoken critic, focusing on low glucose-lowering efficacy and CV neutrality. That said, the market remains relatively stable, if fluctuating, around $10 billion per year, reflecting how ingrained DPP-4s are in practice and treatment algorithms; providers continue to favor the class’s well-established safety profile (DPP-4s are remarkably tolerable and come with few, if any, side effects), especially in older patients and those with renal impairment. We do expect DPP-4s to be used perhaps slightly less over time, particularly as SGLT-2s go generic; on the other hand, not everyone can take SGLT-2s and it’s hard to say where GLP-1 pricing will go as a biosimilar – and the side effect profile for DPP-4s are the cleanest around. Indeed, we would so love to see increased utilization of DPP-4/SGLT-2 fixed-dose combinations (AZ’s Qtern, Lilly/BI’s Glyxambi, and Merck/Pfizer’s Steglujan), which have more impressive efficacy in an oral combination.

  • Although we could hardly agree more with Dr. DeFronzo’s push for combination therapy, we have to come back to cost, particularly if one is going to replace metformin with an SGLT-2 in this triple-therapy regiment. Some patients are going to be seriously burdened by out-of-pocket costs for two branded drugs unless more payers start to see the light – some, of course, have excellent coverage and we wish there were more. Indeed, this was an inescapable theme of AADE 2017, where Dr. Susan Cornell echoed many of the points Dr. DeFronzo made in this talk before admitting that she still prescribes sulfonylureas, entirely due to their low cost to patients. In the next decade, we’re incredibly excited for newer classes to go generic and become more widely available (though it’s hard to know what will happen on the pricing front for injectables); until then, we cannot stress enough that more work needs to be done to help patients gain access to therapy that is truly life-saving. Better reimbursement, patient assistance programs, and public insurance can all help, and we’re optimistic to see this conversation gaining momentum at a national level.  

2. Does Glucose Variability Matter? Dr. Hirsch Thinks So But Calls for a Definitive Study in Type 1, Pointing to Evidence in Pregnancy, CVD

Tasked with answering the question of whether glucose variability matters, Dr. Irl Hirsch called for a study using closed loop technology vs. standard therapy (i.e. insulin vs. insulin) in type 1 to demonstrate a relationship between glucose variability and outcomes. Referencing Dr. Rich Bergenstal, who called this the most important question of the decade (we have to agree), Dr. Hirsch asserted that the type 2 population doesn’t lend itself well to this question. At a fundamental level, Dr. Hirsch argued, hypoglycemia and hyperglycemia are both bad, but there are also molecular and cellular principles that lead him to believe variability from frequent ups and downs could actually be worse than the effects of sustained hypoglycemia (arrhythmia, inflammation) or sustained hyperglycemia (microvascular complications) alone. To establish the validity of the question, Dr. Hirsch first pointed to evidence from Australia examining 1,604 adolescents stratified across four time periods. Rates of retinopathy fell over time, from ~50% in 1990-1994 to ~10% in 2005-2009; at the same time, the proportion of patients achieving an A1c <7.5% rose marginally but hovered around ~20%, but the proportion of patients on MDI or CSII (this was the era when mealtime insulin emerged) rose substantially, from ~15% to ≥90%. Retinopathy fell even though A1c stayed about the same, leading Dr. Hirsch to suspect a reduction in glucose variability, caused by the emergence of prandial insulin analogs, drove the reduction in complications. In his assessment, A1c is too simple of a biomarker, and it doesn’t solidly predict who will get complications and who won’t. So, what could be at play? Here, Dr. Hirsch turned to pregnancy and a biomarker called 1,5 anhydroglucitol (1,5 AG), which session chair Dr. Helen Murphy affirmed she uses routinely in pregnant patients with diabetes. He outlined how, in patients with A1c <8%, 1,5 AG is a powerful predictor of postprandial hyperglycemia – indeed, an important component of glycemic variability – and explained that low levels are bad. Typically, 1,5 AG is filtered at the glomerulus and reabsorbed at the renal tubule, but in diabetes the system is overrun, glycosuria occurs, and not all 1,5 AG can be reabsorbed. When a spike in glucose occurs, 1,5 AG is lost in the urine, giving lower levels in the blood; 1,5-AG below 6 ug/ml has been associated with high birth weight, despite no clear relationship between A1c and birth weight, as per a study from Dr. Hirsch’s group. Moreover, lower 1,5-AG is associated with neonatal hypoglycemia, pre-eclampsia, and C-section, and the ARIC (Atherosclerosis Risk in Communities) study even suggests a relationship between postprandial hyperglycemia and CVD, as measured by 1,5-AG. With 1,5-AG, Dr. Hirsch says, we now have biochemical metrics that can look specifically at variability (and this is not to mention CGM), so let’s answer the question once and for all. The implications of time in range for treatment goals, pharmaceutical development, and drug and CGM reimbursement are enormous – and we would certainly love to see a long-term outcomes trial conducted – but we also have to note that we’ve mainly heard skepticism that a true RCT to this end will ever be funded. As hybrid closed loop becomes more widely used, however, could real-world/observational evidence help solve this problem?

3. Dr. Weinzimer Highlights Potential of Pramlintide and SGLT Inhibitors in AID Systems, Emphasizing Benefit of Postprandial Blunting; “You Might Get Away with Not Needing User Input” with Amylin Analog

Discussing coming trends in type 1 diabetes management, Dr. Weinzimer underscored the co-emergence of AID systems and adjunct therapy for type 1, which presents a unique set of opportunities and challenges. In particular, he highlighted a renewed hope for the use of pramlintide, in hybrid and/or fully closed loop systems, referencing a compelling study just presented at ADA 2018 from Dr. Ahmad Haider’s McGill group. In this JDRF-funded, 24-hour, in-clinic study (n=27), simultaneous infusion of rapid-acting insulin + pramlintide resulted in a striking 86% time in 70-180 mg/dl, vs. 74% with rapid insulin alone, translating to an extra ~three hours/day in range. Mean glucose was 133 mg/dl with rapid acting + pramlintide, vs. 143 mg/dl with insulin alone, and peak postprandial glucose excursions were also significantly blunted in the pramlintide arm; there was no imbalance in hypoglycemia or other side effects. Dr. Weinzimer himself investigated this concept in 2012, and Dr. Jennifer Sherr spearheaded proof-of-concept work two years ago. As Dr. Weinzimer put it, the type 1 field is slowly learning from the type 2 field that insulin is not the only abnormality in diabetes, and amylin analogs – amylin is normally co-secreted with insulin from beta cells, but also lost in type 1 – represent a direct correction of one of those “other” abnormalities. Moreover, he says that amylin helps to suppress glucagon secretion, better enabling a fully closed loop: “You might get away with not needing user input” if you add an analog. So far, the only amylin analog on the market, AZ’s Symlin (pramlintide), has not seen significant uptake (selling $9 million in 1Q18). Mealtime bolus dosing of Symlin alongside insulin has proved tricky for many patients and providers, resulting in hypoglycemia, and the short half-life of the molecule means it has to be injected with every meal, so our sense is that most view Symlin as an overly-demanding adjunct therapy. Infusing pramlintide in a fixed ratio with insulin at the same and in the context of AID, however, could really make up for all of the inherent shortcomings of the therapy, and Dr. Weinzimer also noted recent interest in co-formulating pramlintide with rapid-acting insulin; Adocia just moved one of these candidates into phase 1 and Xeris has a preclinical candidate in development. We’re also intrigued by Novo Nordisk’s phase 2, once-weekly amylin analog – though it’s being developed for obesity, our sense is that management is very excited about the candidate and we can’t help but wonder about potential off-label use in type 1.

  • SGLT-2 inhibitors could be a very attractive addition to closed loop, according to Dr. Weinzimer. He cited a study presented at ATTD 2018 (P-024): On a fully-closed loop, addition of dapagliflozin conferred an additional 2.8 hours of time in range per day in the face of two unannounced meals. Mean glucose was 154 mg/dl with dapagliflozin vs. 187 mg/dl without (p=0.001), and total daily insulin dose was 11 units lower with dapagliflozin (p=0.001). We saw similarly positive data for use of adjunctive dapagliflozin with hybrid closed loop in a poster at IDF. However, the overall improvement in blood glucose variability offered by closed loop technology presents a tricky challenge relative to the DKA risk SGLT inhibitors carry in type 1: Patients lose the blood glucose signal to test for ketones, because blood glucose, theoretically, just doesn’t go up (particularly on fully closed loops). The risk of DKA is compounded when one takes into account the probability of infusion catheter occlusion or kinking. That said, we note that this so-called “euglycemic DKA” is a problem inherent to SGLT usage; while patients on closed loop may see a relatively higher proportion of DKA, the driver of DKA is insulin insufficiency and therefore patient education on this topic remains paramount either way. Further, the additional postprandial blunting offered by SGLTs could make for a truly remarkable closed loop system; while the combination of this therapy and technology will certainly need significant testing, we’re excited about the potential it holds. 

  • Dr. Weinzimer reinforced the value of predictive low glucose suspend technology (e.g., Tandem’s Basal-IQ) as an excellent option for some specific sets of patients: Those who want more aggressive daytime glucose targets than current hybrid closed loop (670G) systems allow, patients who don’t want excessive alarms, “diehard Dexcom users,” and pregnant women – this is a pretty big group! As a reminder, Basal-IQ was FDA approved the day before ADA with Dexcom’s G6 and will launch in August. In his assessment, while some would call PLGS an inferior strategy to hybrid closed loop, the reality is that the one approved hybrid closed loop option (670G) isn’t optimal for many. Some patients are willing to trade higher engagement for an A1c in the high 6% range, but that may come with a higher risk of hypoglycemia, and a PLGS system solves that problem while still allowing for a lower target blood glucose and, thus, a lower A1c.

4. Dr. DiMeglio Praises Fiasp and Afrezza Despite Cost + Inherent Limitations of Insulin, Introduces the DIY Insulin Movement; Dr. Weinzimer Gives an Optimistic Rundown of Glucose-Responsive Insulin

Offering her thoughts on prandial insulin, Dr. Linda DiMeglio praised the improved pharmacokinetics of Novo Nordisk’s Fiasp and MannKind’s Afrezza but stressed that insulin alone will not provide a “near cure” for diabetes. Given the narrow therapeutic objective of insulin – preventing hyperglycemia while avoiding hypoglycemia – Dr. DiMeglio asserted that insulin is inherently limited in its ability to treat diabetes; she referenced an old ADA campaign slogan: “Insulin is not a cure for diabetes, it just keeps people alive until we find one.” This brings to mind a concept we’ve been hearing more as adjunct therapy for type 1 becomes a bigger reality: Insulin is not the only defect in type 1, and our ability to manage glycemia in patients with type 1 will only get better with adjunct therapies targeting more pathways. SGLT inhibitors and GLP-1 agonists can help, but perhaps amylin and glucagon receptor antagonists will more directly target the beta and alpha cell defects present in type 1. That said, Dr. DiMeglio also seemed genuinely excited about how far insulin has come over the past century, even if progress has been slow and it’s been from a low base. She extolled the ultra-rapid-acting profiles of Fiasp and Afrezza, which she sees as the pinnacle of rapid-acting insulins, with the qualification that their advantages do not necessarily justify their prices for all patients. Indeed, we often hear that these insulins represent “marginal” improvements over previous prandial analogs (a sentiment we vehemently disagree with, particularly given our sense that patients love both options), but the point stands that these insulins do tend to be more expensive for patients, as reflected in the limited commercial uptake of both Fiasp and Afrezza to date. We remain encouraged that Lilly, Sanofi, and Adocia all have phase 3 or phase 3 ready ultra-rapid candidates in development; could a bigger “ultra-rapid-acting” class improve access and drive down costs for patients?

  • Dr. DiMeglio highlighted the work of the Open Insulin Project, part of a DIY movement to create an open-source insulin protocol – essentially, biohackers trying to make insulin more simply and cheaply. As Dr. Irl Hirsh outlined on Day 2, insulin prices are skyrocketing, a symptom of a convoluted healthcare and drug pricing system. In response, biohackers have initiated a DIY insulin movement, formalized as the Open Insulin Project, and are researching a novel way to produce newer, more affordable insulin. Open Insulin has partnered with Dr. Jim Wilkins, who believes an insulin-producing yeast strain could land a new generic on the market within three years. The DIY diabetes movement has gained a lot of momentum and success in tech, where OpenAPS, Loop, and other open-protocol initiaves have made it onto the JDRF agenda, into the hands of thousands of patients, and pushed available tech ahead of industry – prompting the first-ever DIY symposium at ADA this year. While we don’t doubt Open Insulin harbors a level of scientific skill and expertise, we do imagine most patients lack the lab skills and supplies to make their own insulin; our sense is that Loop isn’t necessarily the easiest system to set up, but the biochemistry know-how needed to make insulin is likely a barrier and carries a different set of health and liability risks than a self-constructed hybrid closed loop system. To this end, Open Insulin seems to have two goals: (i) developing an open protocol for insulin production and (ii) enabling commercial, “generic” production of insulin; it’s unclear if they might sell a kit or the necessary supplies to produce insulin, but they do seem very interested in pioneering research to do so more affordably.

  • Later, Dr. Stuart Weinzimer highlighted his bright spots in the glucose responsive (or “smart”) insulin landscape, pointing especially to the work of UNC’s Dr. Zhen Gu. While not identifying it by name, Dr. Weinzimer extolled Dr. Gu’s research on glucose-sensitive microneedle insulin patches, which has received ample funding from JDRF, MicroPort Scientific Corporation, and ADA, characterizing this as one of the key therapies leading the GRI field. Early last year, Dr. Gu published positive results for this therapy in mice. Indeed, Dr. Gu’s name has become all but synonymous with glucose-responsive insulin (GRI), and we saw new, very positive data highlighted at CODHy Tel Aviv in February. To be sure, it speaks to the youth of the field that a preclinical project can be called the leading GRI; to our knowledge, Merck discontinued the first-ever clinical GRI in phase 1 over efficacy concerns, and, according to Dr. J Hans DeVries at IDF, the company now has another GRI in phase 1 – the only current clinical candidate, to our knowledge. That said, Dr. Weinzimer did seem genuinely excited and optimistic about the future of the field and the promise of both molecular-level modifications to insulin and “smart cells” to result in a quality GRI. While overwhelmingly preclinical, the GRI landscape is host to a remarkable variety of modified insulin molecule and delivery strategies; for example, we most recently saw JDRF and Gubra partner on a candidate involving a glucose-sensitive linker at the lipidation site of human insulin, and that research is currently focused on defining the molecule’s glucose-sensitivity range. As Dr. Weinzimer explained, GRIs are also being developed in the form of “smart cells,” which involves embedding glucose and insulin transporters in the membranes of vesicles to effectively act as replacement beta cells. We’ve certainly heard less about this strategy; in many ways, it sounds like a more complex approach, but we’re nonetheless excited about the diversity of candidates in development. Finally, Dr. Weinzimer stressed that augmented devices should also be considered as researchers seek to improve insulin therapy. He touched on Insuline’s InsuPad, essentially a disposable heating pad, which enhances the efficacy of rapid acting insulins by warming the injection site. To our knowledge, this product hasn’t been cleared in the US, and we imagine the added hassle for an incremental benefit doesn’t appeal to most patients, but that’s not to say that other concepts can’t add value to insulin delivery.

  • During Q&A, Dr. Satish Garg sparked an interesting discussion on basal insulin regimens: “Are we at the end of basal innovation, or should we continue to investigate improvements?” Dr. DiMeglio noted that she would love a basal insulin that could be injected less often or taken orally, as well as cheaper options to address barriers of adherence and cost. Novo Nordisk’s phase 2 once-weekly basal is certainly promising to the first point, and we’ve heard that it could be a great “starter” insulin, though the dose essentially has to be multiplied times 7 – and some providers might balk at that. A provocative dialogue ensued, in which Dr. Larry Hirsh conjectured that taking insulin less often, for example three times per week, may result in patients forgetting when they took their last dose and lead to accidental overdosing or irregular regimens. Instead, he proposed that an insulin that is taken once per day with flexibility in timing would be most practical for patients (we note that Tresiba offers that – as a reminder it was originally positioned as a once “every other day” for some patients although this regimen wasn’t approved). In our assessment, a once-weekly insulin would be much easier than one taken every three days (i.e. it’ll always be dosed on the same day of the week), and we wonder what learning can be taken from once-weekly GLP-1s. Dr. Garg referenced the FLEX studies in which 8, 24, and 40 hour dosing schedules of insulin degludec showed comparable A1c reductions and less nocturnal hypoglycemia compared to glargine and fixed regimens of degludec in patients with type 1 diabetes; similar results were observed in patients with type 2, although there was no statistically significant difference in nocturnal hypoglycemia.

5. Immunotherapy for Type 1: Dr. Moran Calls for Combination Therapy Earlier in the Course of Disease + Strategy of “Induction and Maintenance”

In the final presentation of Keystone 2018, Dr. Antoinette Moran advanced the idea of “induction and maintenance” for type 1 immunotherapy, argued in favor of combination therapy and earlier intervention, and explained that designing these trials in children presents a variety of unique difficulties. A lesson she thinks the field should take from transplant research and literature, induction and maintenance refers to the short-term use of an intense agent(s) (e.g. immunosuppression) followed by lifelong maintenance with the lowest-possible drug doses (i.e. maintaining immunosuppression while minimizing complications). In Dr. Moran’s assessment, it’s possible the immunotherapies we’ve been testing are great at induction (indeed, many show very good short-term effects), but they need to be followed-up with something like an antigen or anti-inflammatory agent. Another factor contributing to the mixed success of trials to date, she says, is that they have almost exclusively focused on monotherapies given for short courses, when drug combinations could offer greater efficacy and durability. At ADA 2018, this sentiment was similarly advanced by TrialNet representatives in the exhibit hall, who shared a planned trial of rituximab + abatacept. As Dr. Moran outlined, rituximab has demonstrated the ability to slow C-peptide decline in new onset type 1 (four infusions over three weeks), but the effect only lasts to ~six months. Abatacept has a similar effect in new onset type 1 (monthly infusions over two years), but while the benefit over placebo is sustained, C-peptide still declines over 24 months. Notably, abatacept targets T cells, while rituximab targets beta cells, so we’ll be intrigued to see if efficacy is additive or even compounded in a combination trial. Dr. Moran also mentioned a potential benefit to starting earlier in the course of disease; most trials to date have been done in recent-onset type 1, and abatacept is now under investigation in stage 1 of disease. Going even earlier, Dr. Moran spoke to the role of oral insulin in prevention, offering that when the trial was started over a decade ago, the field knew a lot less than it knows now – some people enrolled had only one autoantibody, so their chances of progressing to type 1 were very low. She asserted that, if designed today, the trial would be very different; but also, at a basic level, the notion that one study will work for everyone with such a heterogeneous disease is probably false, and trials need to carefully consider subsets of populations to see success. One final challenge is targeting immunotherapy to children. Dr. Moran explained that FDA first requires demonstration of safety/efficacy in adults, even if kids are the target population and even though type 1 differs significantly among kids and adults in terms of disease process and response to immunotherapy. On the other hand, it’s important to consider that starting immunotherapy or immunosuppression in kids can then span decades, and these rules are in place because kids are especially susceptible.

  • During Q&A, an audience member asked about the role of the BCG vaccine in improving control for patients with type 1, and Dr. Moran was very skeptical of the clinical validity of this science. Recently, new phase 1 results from Dr. Denise Faustman (published in Nature) garnered a fair amount of public attention, and Dr. Moran pointed to the fact that, in Canada, BCG was on the national vaccination schedule for a couple of decades but had no impact on the prevalence of type 1. In her assessment, this is a bigger cohort than you’ll ever get in a study and, while not an RCT, represents solid evidence against the clinical significance of BCG for type 1 prevention. Similarly, an audience member noted that, until eight years ago, everyone in Finland received the BCG vaccine at birth, and no impact was observed (in addition to clinical trials showing no positive effect on C-peptide preservation, and even accelerated C-peptide loss in teenagers). That said, Dr. Faustman’s paper does assert that BCG offers improved glycemic control, rather than type 1 prevention, but we continue to point to small sample size (n=52), down to n=9 at five years and n=3 at eight years, as well as marginal statistical significance.

6. Dr. Brigitte Frohnert Underscores the Role of Epigenetics in Type 1 Development, Legacy Effects, and Complications; Potential Applications for Drug Development

Dr. Brigitte Frohnert stressed the importance of epigenetics in understanding type 1 diabetes pathogenesis, complication risk, and legacy effects, highlighting potential for epigenetics to illuminate potential therapeutic targets. Notably, risk of type 1 diabetes has increased more rapidly in the population than genetic changes can explain, and risk to offspring of a father with type 1 is higher than of a mother with type 1 (6% vs. 1%); further, monozygotic twins have a high discordance rate for type 1 (20% - 50%), especially if diagnosed at an age older than 15 years (our sense is that environmental exposure/other non-genetic factors are also widely thought to have a sizable impact). Focusing on a discordant type 1 diabetes twin study, Dr. Frohnert showed that there are significant DNA methylation differences between pairs of discordant twins. In the twin with type 1, methylation occurs at higher rates in certain genes that are involved in oxidative metabolism, mitochondrial function, cell metabolism regulation, and cell growth and proliferation, suggesting a role for DNA methylation and resultant gene transcription differences in type 1 pathogenesis, though our understanding is that a causal relationship hasn’t been established. Dr. Frohnert also detailed DCCT/EDIC evidence pointing to epigenetic effects in diabetes, as we heard earlier this year at Levine-Riggs 2018. Although the DCCT’s intensive treatment group saw improvements in A1c vs. conventional therapy during the 10 intervention years, both groups converged to an A1c of ~8.0% soon after follow-up began. Despite similar A1c between the two arms over the EDIC follow-up period, the intensive group continued to see significantly lower rates of microvascular complications than the conventional group, and Dr. Frohnert claimed that this lasting effect was due to “metabolic memory” and “legacy effects” driven by epigenetic changes. Analysis of these separate groups identified higher rates of methylation of the TXNIP gene in the conventional group; this gene has been associated with hyperglycemia and microvascular complications, as seen in the diagram below. Studies performed in vitro have provided added evidence for this working model, as cells grown in high glucose environments show persistent hypomethylation of the TXNIP gene, enabling increased TXNIP expression. That is, glucose levels can affect DNA methylation patterns, and those methylation patterns are persistent and have been implicated in long-term complication risk. Dr. Frohnert expressed hope that studies like these could be important in identifying future therapeutic targets, and it seems as if the approach is working: very recently, the drug verapamil (a blood pressure medication) was shown to have positive effects in type 1, apparently by lowering expression levels of TXNIP.

7. Dr. Rewers Claims Autoimmunity Screening Could Prevent 80% of DKA at Type 1 Onset and Diagnosis, Presents Preliminary Data from the Autoimmunity Screening for Kids (ASK) Program

Presenting preliminary data from the ASK program, Dr. Marian Rewers highlighted that 1.1% of children have pre-symptomatic type 1 diabetes and claimed that screening for autoantibodies could prevent 80% of DKA at onset. The Barbara Davis Center’s ASK (Autoimmunity Screening for Kids) program screened 9,511 children in Colorado (1-18 years old) for known islet autoantibodies (IA-2A, GADA, mIAA, and ZnT8A). Out of this initial population, 3.6% were positive for at least one islet autoantibody and 0.5% for multiple autoantibodies; we find it somewhat striking that >3.0% of those screened have at least one autoantibody, though we know that that the 20-year rate of progression to type 1 is only ~10% in this population – what’s different about these patients vs. those who progress through multiple autoantibodies and into type 1? Further, consistent with epidemiological data, ASK found that non-Hispanic whites had a higher prevalence of multiple autoantibodies (which is associated with a 5-year risk of onset of 44%) when compared to Hispanic and Other populations (0.8% vs. 0.4% and 0.2%, respectively). Dr. Rewers expressed hope that ASK can help clinicians catch type 1 onset earlier in the natural history of disease (i.e. stage 1 or 2, before typical clinical diagnosis), and we see opportunity for screening to be integrated into standard neonatal and even pediatric care. Indeed, GPPAD (the organization conducting POInT, a new oral insulin trial) is working to expand newborn screening in a number of European countries for the purposes of trial enrollment; universal autoantibody screening, we think, could go a long way in improving patient experiences and outcomes during and after diagnosis, and it could also offer a more robust and accessible sample for type 1 prevention and disease-modification trials, which can be stymied by enrollment challenges. Dr. Rewers reviewed data demonstrating that 80% of DKA at onset could be prevented by screening for islet autoantibodies and by proper anticipatory diabetes education, as well as that DKA at diagnosis predicts significantly worse long-term A1c outcomes. In a separate observational study of 3,364 Colorado children, patients with mild/moderate or severe DKA at diagnosis had significantly higher A1c levels than those without DKA at diagnosis over the course of 15 years (0.9% and 1.4% higher, respectively; p <0.0001 for both vs. no DKA), independent of ethnicity and health insurance. While there are very possibly other confounding factors, Dr. Rewers strongly advocated for screening in children plus structured education and monitoring programs to better inform patients about their type 1 risk and the how and why of avoiding DKA. On the other hand, as with all prevention programs, Dr. Rewers stressed that it will be important to track the cost and cost-effectiveness of such screening, as well as of pharmacotherapies used for prevention or delay; to this end, we note that organizations are working to improve immunological assays for type 1 autoantibodies as well as to develop technologies that require less blood – or even use saliva.

Q: If you’re testing a child at age one for antibodies, and it comes back negative, shouldn’t you also test again at a later age as well – such as at eight, for example – to conclusively see whether these antibodies are present?

Dr. Rewers: If you look at the study in terms of who was being tested, there’s a peak early and then there’s a later peak in early adolescence. If you only test them once and don’t test them again, yes, you could miss the onset of islet autoantibodies. The model should not be that you just test once and then you’re done. Instead, you should ideally check multiple times like you would for things like lead or anemia. In the ASK study, we do have the option where families can come back in a year to check again if they would so like. But this is definitely something we need to look into regarding what the cost/benefit analysis is for the number of times you screen someone.

8. Dr. Moran Sheds Light on the State of Type 1 in Under-Resourced Countries: Lack of Diagnosis is the Leading Cause of Death in Children with Diabetes

What’s the most common cause of death in a child with diabetes? Dr. Antoinette Moran explained that, while the answer used to be a lack of insulin, today it’s a lack of diagnosis. In an eye-opening talk on underserved youth with type 1 diabetes, Dr. Antoinette Moran shared her own experiences working in Africa, outlining the unique problems patients with type 1 diabetes face on much of the continent. She praised organizations including Life for a Child and Novo Nordisk’s Changing Diabetes in Children for improving insulin availability around the world, as well as ESPE and ISPAD’s PETCA initiative in Nairobi, which trains HCPs in pediatric endocrinology in Africa. However, public understanding of the disease, while improving, remains low, and structural challenges hinder diabetes care across the continent. Low diagnosis rates mean that patients go undiagnosed far longer than in other parts of the world. Until about 10 years ago, she said, there was almost 100% mortality among kids who developed type 1 from ages 1-6, and the vast majority still present in DKA. In some areas, a lack of clean water means that, even when children have very good hygiene, they experience many of issues with skin and other infections. Among patients who do survive into adulthood, complications are the rule – but we were struck that the set of complications Dr. Moran listed was quite different from those that first come to mind in the US. Severe growth stunting and pubertal delay is common, resulting in adults with childlike stature, and infection is also a big concern. Hypoglycemia remains a significant issue, as does resulting severe cognitive dysfunction, and Dr. Moran also noted very limited microvascular treatment options. As she explained, regional and national healthcare in Africa is set up to test for and treat HIV/AIDS and malaria – and has sophisticated tech to that end – but there’s a lack of infrastructure for chronic illness. Especially regionally, health care facilities are sparse and poorly equipped, and the sheer cost of treatment is prohibitive. While studies have actually demonstrated that insulin can be kept cool enough in a clay pot in a hut, in many places there isn’t a way to get the insulin there in the first place. It can be difficult to look beyond our own borders when patients with type 1 and type 2 diabetes, on average, aren’t even doing well in the US, but we’re thankful for the efforts providers, professional organizations, and industry are making to equalize care on a global scale.

  • The United States also has its own problem with care inequality in type 1.  As we saw at ADA 2018, T1D Exchange data indicate that meaningful differences exist in both technology utilization and A1c between white, black, and Hispanic patients under the age of 18. White patients on injections and pump therapy achieve an average A1c of 8.6% and 8.3%, respectively, while black patients average 10.3% and 9.1% and Hispanic patients 9.5% and 8.6%. Moreover, pump utilization remains unequal even when controlling for socioeconomic status. At low (<$50,000), middle ($50,000-<$75,000), and high (≥$75,000) income levels, white patients <18 years old use pumps at a rate of 53%, 70%, and 78%, respectively, while black patients are at 27%, 32%, and 60% and Hispanic patients at 41%, 59%, and 72%. At 53%, the lowest income group of white patients uses pumps almost as much as the highest income group of black patients (60%). In the lowest income group, white patients are twice as likely (53%) to be prescribed a pump as black patients (27%). In Dr. Moran’s assessment, this discrepancy comes primarily from providers and an inherent bias in how HCPs are prescribing devices and treating patients; she called for greater vigilance and awareness of biases in prescribing patterns, and we think this aspect of health disparity certainly warrants more research and attention.

Diabetes Technology Highlights

1. Dr. Lias on FDA’s Work on Interoperability, Access, Reimbursement, Business Motivators; OpenAPS “Coming into Compliance”

FDA’s Dr. Courtney Lias delivered a well-received keynote similar to her talk at ADA a couple weeks ago, reviewing the exciting integrated CGM (iCGM) pathway, part of the Agency’s plan to “try to make your diabetes life like the rest of your life.” After giving background on iCGM and AID integration again – with equal passion and enthusiasm, we might add, and including the Tandem example on iCGM compatibility (FDA doesn’t need to get a new submission from Tandem when they want to integrate a new iCGM) – Dr. Lias delved into other irons the Agency has in the fire to push diabetes life further toward resembling the rest of life. As we’ve heard from the Agency, JDRF, and other stakeholders, “integrated pump” (iPump) and “integrated algorithm” (iAlgorithm) categories are under serious consideration – enabling a more interoperable AID ecosystem, where component combinations are not tied to each other and won’t require separate regulatory submissions or complicated contracts. Dr. Lias explained that “if we can create secure and reliable protocols, this type of paradigm might work well” for pumps. SOOIL will reportedly submit a de novo request to be registered as an integrated pump (and therefore pave the way for a new category), with ambitious hopes for a launch by June 2019. On the algorithm side, Dr. Lias laid out a series of possible benefits, as well as some unknowns. On the plus side, an “iAlgorithm” or “iController” category could allow more players to enter the field (don’t have to develop CGM and pump as well), may allow for fewer/smaller clinical trials as performance and specifications become better understood, and integration into smartphones will enable a more “streamlined device experience” for patients. On the other hand, the Agency still needs to understand whether special controls can be made for algorithms, and how business models would work (e.g., would the algorithm reside in a pump or “separate” – Dr. Lias thinks either could work). We were also immensely excited to hear that FDA is making it a priority to continue to understand how it can influence and incentivize innovation. That includes listening to patient and HCP needs, “optimizing the regulatory touch,” and…drum roll please…learning more about other external barriers that keep patients from getting devices (e.g., coverage/reimbursement, business motivators). That means not only that the Agency will work to avoid doing something to inadvertently hurt access, but also to work actively with CMS and other payers to help them understand what FDA has done and the value of integrated diabetes device systems. For example, Dr. Lias said her division will talk with CMS about therapeutic CGM to allow patients to only use the app, rather than requiring the app and the DME receiver to be used, because “that’s the way it has to go” (see last month’s coverage of the Medicare update to finally allow smartphone CGM use). Helmsley Trustee Mr. David Panzirer has also offered to get FDA in the room with payers to ask questions – though they won’t necessarily be able to directly influence payer policies, Dr. Lias intends to learn more about what’s important to them, and then be able to pass that information on to manufacturers or even modify regulatory policy accordingly. Lastly, FDA isn’t authorized to consider cost in its decision-making, but it can indirectly drive down cost by a number of mechanisms, for example increasing competition (as we’re seeing in CGM) or “preventing manufacturers from having to pay millions of dollars to lawyers.” This busy federal division couldn’t be further from the FDA CDRH of 5-10 years ago. 

  • Dr. Lias On OpenAPS: “We don’t endorse it, though we know patients are doing it. We’re talking with them, and trying to encourage working toward coming into compliance, and they are all doing so – both Loop and OpenAPS. We do want to apply a level playing field – if we are going to create feasible pathways, we want people to use them.” We’re assume for DIY systems to “come into compliance,” they will need to do some adverse event collecting and reporting, at minimum. We doubt they will be asked to perform clinical studies or have FDA assess incremental algorithm updates. We assume the community is more than happy to do this kind of reporting if it’s feasible and it allows the algorithms to be more “accepted.” Reflecting on ADA, we wrote, “Brave moves from SOOIL Korea and Diabeloop to embrace and incorporate software from the OpenAPS community show how industry and the DIY movement don’t have to stand on opposite sides of the street” – ADA was all about industry crossing over to the DIY side, but the opposite is equally interesting where the DIY moves toward an accepted industry path too.

  • Regarding DreaMed’s recently-cleared Advisor Pro clinical decision support software for optimizing pump settings based on CGM data, Dr. Lias said “the level of data we asked for wasn’t very high, it was really just to show that it was similar to or no worse than available.” Referring to other decision support products – particularly those that leverage CGM data to guide therapy – in the pipeline, Dr. Lias added, “depending what it is, it either needs no data, or just a little.” The response came in response to a question from a very complimentary Dr. Irl Hirsch, who “opened a can of worms” by harping on the need for smart pens and standardization across bolus calculators (“they all give different results!”). Of course, much of this relates to patents, which in part led to the demise of Cozmo following a lawsuit with Medtronic. Companion Medical’s InPen app does have a bolus calculator, and recently achieved a major milestone in Android clearance (launch in 4Q18).

  • Where are we in diabetes now, in Dr. Lias’ view? “Who remembers when we had CDs, but cars didn’t have CD players, so we needed to use that tape deck adapter. I think that’s where we are with diabetes. And sadly, I think in some ways, we are where we were when I was a kid, trying to record a song on the radio, with a little tape and a microphone, timing it perfectly to avoid all of the commercials. I want to do diabetes like we’ve done with everything else: Connect everything. But we can’t necessarily connect pumps, or glucose meters, and that’s something that has to change. Diabetes devices need to work together.”

2. Dexcom CEO Kevin Sayer: Direct-to-Watch Integration Coming “Sometime Next Year;” Next five years: innovation on access; Decision Support – Sleep Button; Inpatient and Obesity Markets

For the first time, Dexcom shared specific timing for the highly-anticipated direct CGM transmitter-to-Watch integration (primary display) – CEO Mr. Kevin Sayer expects the connection will “happen sometime next year.” Last we heard at ATTD, direct-to-Watch (not needing a nearby phone) was still “in development,” but no specific timing was attached to it. (And contrary to what the NYT reported in December, we confirmed direct-to-watch is not under FDA review.) Mr. Sayer only referenced Apple, characterizing the connection efforts as “the next big Apple project they’re working on.” We assume the Dexcom/Fitbit partnership could go direct-to-Watch over time, though the initial Dexcom app for Fitbit smartwatches in 2H18 almost certainly will be secondary display. We’re delighted to hear Dexcom is still working on this integration – we continue to hear from parents that this will be very welcome, as use with the Series 3 Apple Watch would mean their kids can avoid carrying around a smartphone and have data sent direct from transmitter to watch to the cloud. Plus, direct-to-watch avoids being tethered to the phone all the time to get CGM data – something more and more would love. We assume the G6 iCGM designation is a key contributor to this new timeline; whereas before the Watch app would have required a PMA, now it can undergo a Class II pathway. We had hoped that would accelerate things, so hearing this is a 2019 event is slightly surprising. Of course, the team probably has their hands full with the G6 launch and now maintaining both Apple and Android apps for G5, G6, Share, and Clarity. Whew!

  • Mr. Sayer refreshingly claimed that “over the next five years, there will be as much innovation on getting people to pay for technology as there will be in getting new devices out.” Yes! He rightly cited cost as one of Dexcom’s three main focuses, along with performance and convenience. To this end, Mr. Sayer reiterated that the goal for Dexcom/Verily’s second-gen sensor will be to reduce the cost to less than that of the G6 inserter. Whoa! However, he cautioned that “it’s hard to get payers to move.” Despite offering rebates if reductions in hypoglycemia and hospitalizations are not met, Mr. Sayer explained that payers are “afraid to make this available to everyone for fear of what the cost will be.” As he put it, a child whose parents wake up eight times a night to manage his blood glucose is not likely to be hospitalized – from a payer’s perspective, CGM is only adding cost for this patient. Mr. Sayer asserted: “It’s our responsibility to teach those who pay for this device that that’s not good enough.” Hear, hear!

  • In line with previous comments, Mr. Sayer said that Dexcom plans to “work very hard to make decision support a reality.” Referencing an earlier presentation from the evening, during which paper copies of arrow/trend dose adjustment guidelines were distributed (e.g., Endocrine Society’s guidelines for G5), Mr. Sayer exclaimed: “all those worksheets – I cringed! That should all be automated on an app.” We agree and have long felt CGM-based bolus calculators are a “when,” not an “if.” Mr. Sayer mentioned efforts to create a “sleep button,” which could give insulin dosing/food decision support around bedtime. Such a feature is already in development and can predict hypoglycemia ~70% of the time, but Mr. Sayer described it as “not good enough,” as it lacks integration of insulin and carb data. We love the idea of a “sleep button,” especially for parents of children with diabetes who feel the need to put their kids to bed with high blood glucose values to avoid nocturnal hypoglycemia. During Q&A, Mr. Sayer remarked that Dexcom is “hopefully” going to integrate with Siri for decision support – wow, imagine being able to ask Siri how to treat a low or high! Nice!

  • Mr. Sayer cited substantial interest in using the no-calibration G6 in inpatient settings. He assured attendees that Dexcom will “actively” pursue this market, especially now that G6 has no acetaminophen interference. We think zero-cal CGM could massively reduce nurses’ burden, not to mention drive positive outcomes. Dexcom hopefully learned a lot from the Edwards intravascular CGM partnership, which did not push forward. G6 is a completely different product and well-suited for the hospital in our view.

  • Mr. Sayer also expressed enthusiasm in entering prediabetes, referencing Dexcom’s collaboration with a “large payer on predicting prediabetes.” We’re not sure if he was referring to the ongoing 10,000-person pilot with UnitedHealthcare in type 2s or another one. Intriguingly, he also noted potential to enter the obesity space: apparently, Dexcom CGM has been used on the weight loss TV show The Biggest Loser “forever.” News to us!

  • Also in the pipeline are efforts to make the receiver optional. As of 1Q18, Dexcom was “putting plans in place” to file this with FDA, but specific timing was not shared. This would allow G6 to match Medtronic’s Guardian Connect and Senseonics’ Eversense, both of which have smartphone-only display in the US. Mr. Sayer noted that “less than half” of Dexcom users purchase the receiver in Europe, where it is not required. Although he kept timing vague, announcing plans to “eventually” file with the FDA, there is clearly FDA precedent here (with Medtronic and Senseonics) and making the receiver optional would cut down patient startup costs and improve margins. Mr. Sayer also referenced Dexcom’s pump partnerships with Insulet, Tandem, and Lilly, adding: “There will come a day when pens will talk directly to our app.” He didn’t provide further detail beyond hinting that his presentation in five years will be “vastly different” than what it is today. Agreed!

  • Mr. Sayer also highlight Dexcom’s public API interface whereby apps can access Dexcom data directly (launched last September); for now the API is retrospective data, but the interface could ultimately move to real-time data streaming without the current three-hour delay. He noted that “literally 20,000” companies have inquired about writing software through the interface. Still, Mr. Sayer acknowledged that “some very complicated business decisions” have made, making real-time API launch details uncertain. He noted that the FDA will have to clear any app to which Dexcom provides data access, raising potentially complex issues surrounding responsibility and security. Despite these underlying intricacies, Mr. Sayer was very optimistic about the potential for the platform to drive innovation, noting: “There will come a time where people can do this as good if not better than we do” (i.e., referring to CGM apps/software).

3. Tandem to Ship t:slim X2 w/ G6 and PLGS in First 2 Weeks of August, Overviews Flow for Upgrade; Control-IQ Pivotal Recruiting; Peds Control-IQ Study in France (6+)?; Dr. Forlenza’s PROLOG and Algorithm Commentary

At the very end of a Tandem dinner symposium, a rep explained in detail the process for getting a patient onto t:slim X2 with Basal-IQ (G6-integrated PLGS), which is expected to begin shipping in the first couple weeks of August. FDA approved t:slim X2 with G6 integration the day before ADA began, far earlier than we expected and earlier than Tandem guided for. In the next couple of weeks, t:slim X2 users will receive an email from Tandem, asking them to input their personal information, along with their provider’s information, if they are interested in the free PLGS software upgrade. Tandem will then draft a statement of medical necessity to obtain a new prescription for the system, and the provider simply needs to sign it and send it back to the Tandem team. The patient will then receive an email telling them a 45-minute, online training module is ready. Once they pass the quiz following the training, they will receive an update code, and the software update process will take ~7-10 minutes to complete online. Amazing! Tandem has clearly done a lot of work to make the process as seamless as possible, in line with the t:slim X2 software update to add G5 that launched last year. BDC’s Dr. Forlenza said that he wants to do a “fair day” at the Barbara Davis Center, where they pick a Saturday in August for everyone to come in, do training, take the quiz, update the system, and then work with clinical and research staff on the practicalities of starting with a PLGS algorithm. We love it! In Q&A, Dr. Forlenza reiterated that the system only works with G6, even though the pivotal study (PROLOG, presented at ATTD) was performed with G5 – this is a testament to the forward-thinking FDA, of course. For patients who are struggling to get their hands on G6 – in some cases because Dexcom’s secondary distributors are still shipping G5s despite orders for G6 – Dr. Forlenza advised sticking with the G5 integration on the device until a G6 system is obtained. (G6 is also seeing tremendous demand through Dexcom, as we understand it from social media.) In terms of access, the Tandem rep noted that the upgrade pathway from older t:slims to the t:slim X2 has expired, though the Touch Simplicity Today pathway for in-warranty Animas patients ($999 up front, which then goes toward out-of-pocket costs once the Animas warranty is up) remains open until September. For the Medicare population, CGM-integrated pumps are apparently not covered, despite the fact that pumps and therapeutic CGMs are individually covered. Absurd! The rep hopes this will change soon, and stated that FDA is helping Tandem move the dialogue with CMS along – nice! Dr. Forlenza pointed out what should be obvious to CMS – PLGS should ultimately save Medicare money in the form of fewer hypoglycemia-related hospitalizations.

  • As noted in our ADA full report, recruitment has begun for the pivotal study of the t:slim X2/Dexcom G6/TypeZero Control-IQ hybrid closed loop system with automatic correction boluses – see the post here, which indicates the study is already recruiting at six centers. As of late April, a US launch was expected in 1H19, with a rolling FDA PMA submission in 2H18; the study page indicates expected completion in April 2019, which leaves pretty tight timing to get this approved and launched in 1H19.

    • The IDCL study only includes ages 14+, though we heard murmurs of a study with Prof. Eric Renard’s group in France that would investigate the system in kids down to age 6. Presumably this could also pave the way for EU approval. If such a study were to come to fruition, we’re not sure if FDA would accept it for pediatric review or expect a US-based study.

  • Dr. Forlenza explained the “frighteningly simple” Tandem PLGS algorithm, and in particular how it might be able to maintain a lower average glucose level than Medtronic’s MiniMed 670G. Every five minutes, the controller takes a linear regression (slope) of the past four glucose points (i.e., looking 20 minutes into the past). It then projects that slope into the future; if the line projection indicates that glucose will go below 80 mg/dl or if current glucose is below 70 mg/dl, insulin delivery is paused. Insulin delivery then resumes on the first glucose rise, more aggressive resumption criteria than that in Medtronic’s 640G – allowing the Tandem system to better prevent rebound hyperglycemia and thereby keep average glucose lower. Of course, the 670G suspend/resume feature is part of a broader PID hybrid closed loop controller, so “it’s really not comparable in terms of how it works.” With all automated insulin delivery systems that minimize hypoglycemia, Dr. Forlenza explained that by the time a patient is actually low or near-low, he/she has probably not received insulin for one to two hours, and therefore has virtually no active insulin on board. To treat a low while on one of these systems, many people find that they need to take about half as much carbs as they’re used to in order to avoid rebound highs. This is among the greatest lessons that users need to learn when going on one of these systems!

  • The PROLOG results were accepted by Diabetes Care last week and should be published in the near future. Dr. Forlenza was extremely positive on the system all around: “The main takeaway is it’s just beneficial. It doesn’t add burden. It takes away hypoglycemia and doesn’t add user interaction or increase blood glucose – it just takes away a bad thing. Relieving a stress and providing a benefit.” In reviewing the PROLOG findings (presented at ATTD), Dr. Forlenza made a number of insightful observations: (i) The study wasn’t limited to previous pump users (MDIs welcome!) and there was no A1c restriction – these provisions tried to mimic real-world use. “As long as you can use it safely, I shouldn’t be a gatekeeper for how you take care of yourself.”; (ii) Researchers were very surprised to see a greater decrease in hypoglycemia during the day vs. overnight – Dr. Forlenza posited that it may have something to do with the fact that the study wasn’t blinded, and people may have wanted to try to stress test the PLGS (“Many went to the gym every day. They ate as much as they wanted to, just to see how it behaved. You see that a lot in kids.”); (iii) Hyperglycemia borderline improved in the PLGS group (-1%, though not significant). Again, Dr. Forlenza believes the lack of blinding played a role here, leading patients in PLGS to believe that they had a safety net on the low end so they could be more aggressive treating highs; and (iv) He was “very impressed” that non-CGM users saw such a good improvement with PLGS – a testament to the system’s usability!

    • We’d also and that users in the pivotal were doing extremely well at baseline (64% time-in-range) and the control condition was already t:slim X2 with Dexcom G5 CGM. Mean time <70 mg/dl declined from 4.5% during sensor-augmented pump (t:slim X2 with G5) to 3.1% with PLGS, a 31% reduction (-19 minutes/day; p<0.001). Though the absolute reduction was small, these rates of hypoglycemia were low and users were mostly already experienced with pump and CGM. In real-world use, we imagine outcomes would be even better.

4. First US Eversense Insertions Scheduled for July 31; Three New CGM CPT Category III Codes for Insertion/Removal Procedures; 180-Day in US By Late 2019/Early 2020?

At a Senseonics-sponsored corporate symposium, VP of the US Region Mr. Mike Gill shared that the first US patient insertions of Eversense will be performed in a “big event” on July 31. A Senseonics representative later elaborated that Dr. Bruce Bode’s and Dr. Steven Russell’s clinics will be among those that participate in the July 31 insertions. This aligns with the investor call following approval in June. Of course, it is likely to be a gradual ramp, since this will be quite an HCP-gated launch with the in-office insertion procedure and need for HCP training. (On the May call, management guided for “modest” revenue contributions from the US in 2H18.) During his presentation Mr. Gill provided additional color on the Eversense reimbursement landscape, announcing three new CGM CPT Category III codes to account for the additional provider time associated with the Eversense insertion and removal procedures: (i) Code 0446T – creation of subcutaneous pocket with insertion of implantable interstitial glucose sensor, including system activation; (ii) Code 0447T – removal of implantable interstitial glucose sensor from subcutaneous pocket via incision; and (iii) Code 0448T – removal of implantable interstitial glucose sensor with creation of subcutaneous pocket at different anatomic site and insertion of new implantable sensor, including system activation. Unfortunately, it seems that Code 0446T will provide the same reimbursement as the current code for initiating CGM (Code 95250), which may not add extra physician motivation here. A Senseonics representative also mentioned the possibility of a bundle, which would include a CPT III code to reimburse both the physician’s time and 90 days-worth of supplies; we also heard this in June. While Mr. Gill characterized these new Category III codes as “entry-level,” he explained that the ultimate goal is for the codes to evolve to a Category I designation, which is accepted by all payers including Medicaid and Medicare. Senseonics has already approached both national and local payers, receiving “very, very strong” responses. Payers have been especially impressed with the Eversense 90-day wear-time and robust adherence rate. In a separate conversation with a Senseonics representative, we learned that the first payers will hopefully come on board in the Fall, with at least one payer already engaged in a 90-day review process. For the initial users this summer without payer contracts, we assume it will be cash pay. Additionally, the representative shared that Senseonics is also looking at pharmacy channels – an exciting strategy, which would reduce patient and provider burden and match Abbott’s FreeStyle Libre in the US. Given the novelty of the device and increased time required by physicians to both learn and perform the procedures, attractive reimbursement will prove critical in uptake.

  • The 180-day Eversense XL is expected to be approved in late 2019/early 2020. The US clinical trial for the XL is expected to start enrolling in mid-fall, a slight delay from previous timing shared during the company’s 4Q17 call for an XL US clinical trial with “reduced calibration” to begin enrollment in summer. Additionally, the representative confirmed that an EU trial for a 360-day version of the Eversense sensor is ongoing, with an EU launch also expected by the end of 2019. She explained that the sensor itself will not require modifications; rather, the algorithm needs to be adjusted for extended wear. While 90 days certainly makes a dent in reducing patient burden over 7-14 day wear, a 360-day wear sensor would be a step-function change and could radically expand CGM – especially if Senseonics can deliver on the June expectation to reduce calibration near-term.

  • Mr. Gill detailed the three-step, “very straightforward” process flow for obtaining Eversense. Physicians will identify six patients interested in using the Eversense (three for the certification process and three as follow-ups). Patients are entered into a portal via Salesforce, which streams the information to Foundation Care to facilitate a prior authorization request with the patient’s insurance company. Once the prior authorization request is accepted, Senseonics will ship supplies directly to the clinic, and a clinical manager team will provide training and credentials with three of the previously identified patients. Supplies will be purchased on a cash-pay basis until payer coverage is obtained; initially this will make Eversense very expensive at the expected ~$10/day pricing (implying ~$900 for each 90-day sensor). Senseonics notes that once “Experimental” and “Investigational” are removed from payer coverage policies, claims could be paid in the “next couple months” with the same out-of-pocket costs as competitors.

  • We learned and performed the sensor insertion/removal process ourselves – on a dummy arm, of course! We were impressed with the procedure’s simplicity: after watching just one insertion and removal, we performed both procedures in a matter of minutes. While several endocrinologists have expressed enthusiasm at the prospect of performing procedures – Dr. Tim Bailey coined the term “interventional endocrinologist” – Senseonics recognizes that time constraints may prove to be a limiting factor. To this end, a company representative shared that Senseonics is working to receive authorization for nurse practitioners and other mid-level professionals to perform the insertions/removals.

  • Eversense user Tobias Schulte again shared his own positive experiences with the system via video. (He presented live in the Eversense product theater at ADA.) He described wearing the Eversense during an Iron Man competition, crediting the system’s unique features, specifically the on-body vibration alerts, with his ability to finish the race. As he put it, “feeling the transmitter vibrating is a big benefit and allows me to join this wonderful event like I did before my diagnosis.”

5. Dr. Tim Bailey on Flash Glucose Monitoring and Implantable CGMs as “Important Additions to Our Toolbox”

Dr. Tim Bailey strongly claimed that “everyone with diabetes” can benefit from CGM use, pointing to Abbott’s FreeStyle Libre and Senseonics’ Eversense as two additional options that might serve to increase uptake. Unfortunately, Dr. Bailey explained, the majority of patients with diabetes still don’t use or have access to CGM. He referenced an internal claims data analysis revealing that of 2.9 million mealtime insulin using patients in the US, just 0.3 million (10%) use CGM. Of those who do use CGM, he said an estimated 27% discontinue use within one year, citing poor accuracy, problems with adhesive/insertion, and cost as the top three reasons. (This data is at least a couple years old, and we’d guess discontinuation has improved a bit since that publication.) Dr. Bailey noted that features on his patients’ “CGM wish list” include: (i) longer sensor wear and less frequent sensor insertions; (ii) no calibrations; (iii) ease of use; (iv) discretion; (v) less alarms and more alarms; and (vi) reduced cost. That’s where the FreeStyle Libre and Eversense come in. Dr. Bailey remarked that with the FreeStyle Libre, it’s “the first time I haven’t had to push people into CGM, they’ve come in asking.” He credits much of FreeStyle Libre’s success to its lack of required calibrations and reduced cost. In fact, he mentioned that from viewing his patients’ copays, he’s noticed that the FreeStyle Libre is down to ~$40/month in many users. He also claimed that “41% of the ~100,000 US patients on the FreeStyle Libre have type 1 diabetes” – if this statement is true, it suggests FreeStyle Libre use has roughly doubled in the US from “>50,000 users” as of Abbott’s May call. It also means there’s been strong type 2 uptake of Libre in the US, as ~2/3 of global FreeStyle Libre users had type 1 as of the May call. We’re delighted to see FreeStyle Libre gaining traction in type 2, as traditional CGM has not had enough studies to drive significant reimbursement or adoption in type 2. (With some of that changing for MDI/intensive insulin users.) As for Eversense, Dr. Bailey lauded the 90-day sensor’s extended wear-time, as well as the on-body vibration alerts and lack of a receiver. He emphasized that because Eversense is only changed every three months, patients experience fewer “day ones” associated with lower sensor accuracy. The obvious downside is the need to visit a provider four times a year for 90-day sensor insertion/replacement – no small task for many people with diabetes. Dr. Bailey compared Eversense to Intarcia on the utilization front, commenting: “You just stick it in and they can’t really take it out.” Ultimately, Dr. Bailey characterized both flash glucose monitoring and implantables as “important additions to our toolbox” that will hopefully expand the CGM segment.

6. Dr. Tim Bailey: Will Patch Pumps Eventually be Preferred for AID? Comments on Calibra (OneTouch Via), U500 Omnipod, Valeritas V-Go

UCSD’s Dr. Tim Bailey overviewed the patch pump landscape, providing commentary on the 12 “semi-patches” (short tubing) and patches (no tubing), even though “CGM is kind of [his] favorite thing.” While he gave specific commentary on type 2-oriented patches (Calibra/OneTouch Via, U500 Omnipod, and Valeritas V-Go), at a high level, he proposed that the patch form factor may eventually be preferred for automated insulin delivery. If pipeline supply and JDRF funding are any indication, a patch or semi-patch system (i.e., a tubed pump in the pocket with no screen) could become the norm over time. Insulet (~2020 launch), Lilly (launch late 2019-late 2020), Cellnovo (2018 EU launch with Diabeloop), Kaleido (2018 EU launch with Diabeloop?), Tandem (t:sport with integrated Control-IQ; 2020-2021 launch), EOFlow (~2021 launch with TypeZero algorithm), SFC Fluidics (single pod AID; no timeline), and Cam Med (no timeline) all have plans to lend their pumps to AID systems or build their own. The latter three (EOFlow, SFC Fluidics, and Cam Med) all received JDRF dollars to drive development. AID systems from Lilly and Bigfoot will have tubed pumps, but will leverage designs without on-pump screens and instead use an external device for the complete user interface – an app in the case of Bigfoot and an app and/or handheld for Lilly. Also in the pipeline are Roche’s tubeless Solo patch pump, expected to launch this year in Europe (no automation initially), as well as BD’s type 2 patch pump (presumably still expected to launch by September 2019).

  • Product-specific comments:

    • J&J’s OneTouch Via: “I’m intrigued by the Via device. New data presented at ADA. For type 2 diabetes, this is a wonderful option. It’s bolus-only, two units per click. It’s great, you have a whole supply of mealtime insulin when you leave the house in the morning – you don’t have to carry anything else around.”

    • Insulet/Lilly U500 Omnipod: “In VIVID (RCT of U500 Omnipod vs. U500 MDI), these patients had really high doses - 200-600 units per day. You increasingly see patients like this. U500 is an ancient insulin, and tough to use. You see a significant A1c drop in VIVID. Also, that’s achieved with a drop in insulin dose. It is at the expense of slightly more nocturnal hypoglycemia. U500 clearly is a different insulin. The duration we usually use is like five hours or so. I think the duration is six hours in this VIVID study. But I’m not sure that’s quite long enough. Because patients eat big dinners. The other thing about concentrated insulin is you could potentially fit more into the patch.”

    • Valeritas V-Go. “This has been on the market for quite some time. Meant for use in type 2s. I initially thought I wouldn’t use it in any patients, but now I’m more comfortable. V-Go may be an excellent option for people with very high A1cs, but may not bring A1c down to “the 7s or the 8s.”

7. Drs. Ratner and Hirsh Identify Issues with Medicare and Medicaid CGM Coverage; Dr. Ratner Argues that Notion of Over-Treatment in Elderly is a Function of Hypoglycemia Due to Old Guidelines, Calls for Individualization at All Ages

As he has a knack to do, Dr. Irl Hirsch sounded the alarm about a “disaster” in diabetes care, this time pertaining to Medicare coverage of therapeutic CGM: Coverage hinges on MDI/pump therapy, documented blood glucose testing 4x/day, and an insulin treatment regimen requiring frequent adjustments, and Dr. Hirsch has found that type 1s with high C-peptide levels struggle to meet these criteria. He didn’t elaborate, but we suspect that people with high residual beta cell function are at lower risk of hypoglycemia, so they don’t need to check their blood glucose quite as often; it could also be that their insulin regimens are less precise/demand fewer adjustments. Dr. Bob Ratner conveyed his own concerns regarding CGM reimbursement, pivoting the spotlight to variability in Medicaid coverage for CGM, since the decision is left up to each state. Currently, 17 states do not cover CGM through Medicaid, though California could be close. Concluding on a brighter note, Dr. Hirsch ended the discussion on CMS coverage with what he sees as the next goal: coverage of sensor-integrated pumps and hybrid closed-loop systems.

  • Discussing diabetes management in the elderly, Dr. Ratner emphasized the importance of individualization, praising ADA’s 2018 Standards of Care and explaining why it’s erroneous to claim all older patients are over-treated. Specifically, Dr. Ratner pointed to the relaxation of glycemic control goals in elderly patients with poor health as key to preventing hypoglycemia. This guideline, which was first developed in 2012, was not reflected in studies published in 2015 and 2016 that argued elderly patients were being over-treated, a claim authors said was evidenced by high rates of hypoglycemia. As such, Dr. Ratner dismissed those results as reflective of the faults of previous guidelines and suggested that the new, individualized standards of care will likely give lower rates of hypoglycemia in future studies. The issue of individualizing A1c targets has gained new life in the wake of very controversial ACP guidelines; while we do think relaxed targets can be appropriate for a specific type of patient, we’re also wary of the slippery slope target relaxation represents and we think more clear consensus is needed as to how clinicians should assess the appropriateness of relaxation. We appreciate Dr. Ratner’s dispelling of the dangerous notion that all older patients are being dangerously over-treated and his call to consider all patients as individuals.

8. Dr. Garg Discusses In-Development Catheter Micro-Filter to Reduce Occlusions, Proposes “Stud Finder” for Lipohypertrophy

When the discussion turned to infusion set fallibility and insulin delivery variability, gracious host Dr. Satish Garg spoke about ongoing work (company/lab not specified) to put a “micro-filter” in catheters and asked for a “stud-finder” for fibrosis and lipohypertrophy. We had never heard of the micro-filter concept, which aims to reduce catheter occlusions by removing insulin excipients – which commonly polymerize and cause clogs –  before the actual insulin molecule goes in to the subcutaneous space. He said this work is “very exciting, very early, but it looks very promising.” We’d love to learn more about this approach – particularly, who is working on it, if it makes the catheter larger, and if it changes the properties of insulin. Dr. Tim Bailey believes the lower-hanging fruit is to prevent kinking with solutions like stiffer catheters, multiple ports/slits, or mesh around the catheter. As for the stud-finder idea, Dr. Garg remarked on how difficult it is to find the right site to inject insulin or to put a sensor. “If we can find studs all around the wall, why can’t we find hypertrophic sites or scar tissue on our bodies? I think it would be simple, but I’m not an engineer.” Dr. Bailey and Dr. Stu Weinzimer replied that it’s an interesting idea, though they’re not sure how it’d be accomplished. We agree!

Additional Topics

1. Dr. Stuart Weinzimer Provides “Wake-Up Call to Advocate More for Diabetes Funding”; Room for Both AP System and Islet Cell Transplant Development

Yale’s Dr. Stuart Weinzimer asserted that in considering the development of artificial pancreas systems vs. islet cell transplant research, it’s not necessary to choose one over the other. There’s room for both, and there may even be novel means by which to combine the two, although Dr. Weinzimer didn’t elaborate much further on this interesting and politically difficult topic. Dr. Weinzimer shared the latest JDRF funding statistics below, underscoring that the vast majority of JDRF’s ~$1 billion budget over the past ten years has been allocated towards biological interventions – nearly a five-fold difference relative to artificial pancreas research. Wow has the latter provided terrific return-on-investment! Still, Dr. Weinzimer cautioned against thinking that the two areas are really in competition for the same dollars, which is not really true – JDRF has only so much money and has to decide where to put it. However, other disease states are the “real competition,” not to mention other avenues of public funding, like defense. To illustrate his point, Dr. Weinzimer provided data showing that while diabetes prevalence in the US dwarfs that of cancer and HIV/AIDS (30.3 million vs. 15.5 million and 1.1 million, respectively), NIH funding both in raw dollars and per capita scarcely reflect it (see our 2015 piece on this in diaTribe from based on a JAMA article). In fact, cancer receives $361/person, HIV/AIDS receives $2,727/person, while diabetes is allocated just $37/person! Although cancer and HIV/AIDS undoubtedly warrant adequate funding, as Dr. Weinzimer put it: “This should be a wake-up call for all of us to advocate more for diabetes funding.” 

2. Dr. Bob Ratner on Bariatric Surgery in Type 2 Diabetes: “No Question” Regarding Efficacy, but Cost and Complications Concerns Remain

Georgetown’s Dr. Bob Ratner took a somewhat conservative stance on bariatric surgery for the treatment of type 2 diabetes, acknowledging that while there’s “no question” regarding the efficacy of GI operations in those with overweight and type 2 diabetes, cost, complications, and varied outcomes should be duly considered. Dr. Ratner was particularly encouraged by five-year data from the STAMPEDE study, one of the few randomized controlled trials investigating bariatric surgery outcomes. 150 obese patients with type 2 diabetes were randomized to receive either intensive medical therapy alone or medical therapy plus a Roux-en-Y gastric bypass or sleeve gastrectomy. At five years, those who underwent surgical procedures had greater mean A1c reductions than those who received medical therapy alone (2.1% vs. o.3%; baseline: 9.2%). Body weight, triglyceride level, high-density lipoprotein cholesterol level, use of insulin, and quality-of-life measures were also superior in the surgical groups. Interestingly, a meta-analysis revealed a very rapid diabetes remission rate, typically within the first two-to-three months post-surgery, followed by a leveling off; however, as Dr. Ratner pointed out, results indicate that the reverse is also true (i.e., progressive deterioration in glucose) mainly due to weight regain. Still, while seven-year data show diabetes remission to be maintained fairly well, other comorbidities such as hypertension recur fairly consistently with weight regain. Moreover, Dr. Ratner pointed to cost and safety concerns, noting that one study found the hospitalization expenses associated with bariatric surgery are never recaptured over 20 years, even with the benefit of decreased medication! We’re glad to see a more balanced discussion of risks must be weighed with benefits.

  • Dr. Ratner acknowledged that rates of surgical complications are “pretty low,” referencing a study (n=4776) that found just 4.3% of Roux-en-Y gastric bypass or laparoscopic adjustable gastric banding procedures to be accompanied by at least one major adverse outcome. However, he cautioned that this rate may be misrepresentative as “only the best surgeons are reporting.” As for metabolic complications, one study found “virtually all” procedures resulted in anemia. Nutritional deficiencies were by far and away the most common complication amongst bariatric surgery patients, and while vitamin supplementation guidelines exist, as Dr. Ratner noted, it’s “a lot of meds to take, especially for a population with adherence issues.” Other severe complications include dumping syndrome and postprandial hypoglycemia linked with expansion of beta cell mass. In one study of 20 self-identified “centers of excellence,” metabolic complications were shown to be directly associated with surgical skill – as Dr. Ratner warned: “Pick your surgeon carefully.” Other potential risks include an increase in depression, substance abuse, and other psychological issues. To this end, Dr. Ratner characterized psychiatric evaluations prior to surgery as “imperative.”

Norwich Medical School’s Dr. Helen Murphy shared unpublished nutritional data from the UK CONCEPTT cohort. 100 pregnant women with type 1 diabetes enrolled in CONCEPTT kept detailed records of everything they ate at baseline and at 34-weeks gestation. Dr. Murphy found that on average, women consumed 200 grams of carbs, 60 grams of protein, and 60 grams of fat per day. Yikes – this is a 51% carbohydrate diet in pregnancy with type 1 diabetes, which is quite a difficult challenge for keeping blood sugar in range! For main meals, the participants’ food choices “reasonably aligned” with ADA recommendations, and carbs in particular were found to be “broadly speaking” in line with ADA recommendations, comprised mainly of “healthy whole grains.” Notably, women’s snacks were shown to be composed “entirely of ADA non-recommended sources” (i.e., junk food) so that of the 200 grams of daily carbs, 92 grams were actually from non-recommended sources. Ultimately, 50% of participants’ average daily food intake were derived from non-recommended sources. As Dr. Murphy reminded the audience: “we can have all the smart insulins we like, but no smart insulin is able to be absorbed quickly enough to safely match this intake.” Given that these women were highly motivated participants in a clinical trial, we can only imagine that a more diverse population might select even less-nutritious items. It’s clear that more attention on diet and lifestyle are necessary – Dr. Murphy mentioned that such factors are responsible for 70% of the variability within closed loop and automated insulin delivery trials. As noted in our days #1-2 coverage, we hope there is more work to test low-carb in pregnancy with type 1 diabetes, given all the new learning that’s happening around this eating pattern as more wear CGM. 

4. BDC’s Ms. Spiegel Addresses Low-Carb and Keto Diets, Recommending Against Use in Children Due to Safety Concerns but Calling for HCP Support of Patient Choices

Ms. Gail Spiegel addressed the rising popularity of very low carb, ketogenic, and paleo diets, cautioning against their use in children for fear of underdevelopment and psychosocial burden; however, she underscored the ultimate need to work productively with families, giving them agency to make decisions and resources to use any diet most safely and effectively. Illustrating the burgeoning popularity of these various low-carb diets, Ms. Spiegel asked the audience how many have had patients ask about these options – and nearly everyone raised their hands, a phenomenon she attributed to the multiple sources of information that patients can now draw from, certain books, magazines, websites, and social media support groups that promote such diets. In reviewing one survey of patients on a very low carbohydrate diet, Ms. Spiegel discussed how a significant portion (27%) of these patients did not discuss this new diet with their diabetes providers, and nearly half of them felt that their providers were not supportive of their diet choice – both of which are problematic. Specifically, she voiced concerns over ketogenic diets in children, commenting on higher risk of DKA, nutritional deficiencies that have been associated with poor growth in children, risk of bone disease due to decreased bone mineral density, and the increased psychological burden of the diet in children. Ms. Spiegel further explained how restrictive eating behaviors can contribute to social isolation in children, increase anxiety, and provide an additional source for conflict among child and parents. Despite the opportunity for a difference in views between patient and HCP, Ms. Spiegel advocated for patient agency in making dietary decisions, and emphasized the need for better education and support for patients. This is certainly an interesting issue; our understanding is that, for years, children diagnosed with type 1 were often written a “blank check” to eat whatever they want and take insulin for it (many likely still are), setting them up for poor dietary habits and significant swings in blood glucose. We imagine HCPs struggle to walk a tightrope in recommending diets to kids with type 1, not wanting to be too prescriptive or too lenient, and we applaud Ms. Spiegel’s call to allow and enable patients to make their own informed decisions. We ourselves have anecdotally observed or experienced the benefits of ketogenic diets in managing blood glucose, but we respect that evidence of safety in children in very limited.

  • Ms. Spiegel described BDC’s recently-implemented protocol for families trying low carb or ketogenic diets, expressing some reservation about the role of these diets in children. Whether the child/family is/are already on one of these diets or is just curious about these options, they’re assessed in a meeting with a dietitian and given proper educational resources. While the choice remains with the family, a low-carb/ketogenic diet isn’t directly recommended; the HCP encourages a well-balanced, low glycemic diet. If the patient continues to follow a low carb/keto diet, then the HCP advise frequent ketone monitoring, ongoing management including various lab testing of metabolites, and the addition of supplements (multivitamin, Vitamin D, calcium) to the diet.

Q: You talked about the psychological burden of diabetes and also the burden of restrictive eating. Is there anywhere in the protocol you discussed that involves referral to psychological healthcare?

Ms. Spiegel: We are fortunate enough to a have a social work team that’s awesome. We need to get involved sooner, probably, than what’s on the protocol right now. We need to think about how we will integrate that more into the process.

Q: Would you give insulin for fat and protein?

A: Two years ago, I gave a talk on this. There’s more research showing that those nutrients have effects on blood glucose as well. With kids, you don’t see the protein effects as much as the fat effects. To give a general recommendation, I’ve seen some people saying to increase insulin by 30% for fats. We aren’t as aggressive as that, but maybe we should be. I might say to increase insulin by 1 or 2 units.

Q: I am a type 1 diabetic on a keto diet. It has changed my life. After many rollercoaster years, I now finally have a normal A1c and I feel better. Please consider giving patients a choice. I appreciate what you are doing and that you are having these conversations with patients. It would be fantastic if at any meeting you guys could also let us [those on keto diets] speak, because we have experience with this type of lifestyle. After 18 years of living with diabetes, I didn’t know I could have an A1c in the low 5% without an insulin pump and with no low blood sugar. We on keto diets are not so bad and not so crazy.

A: We hear you. The issues with kids are different than with adults, though. That’s where we are coming from.

5. Patient/Provider Panels Illuminate the Good and Bad of Childhood & Adult Experiences with HCPs, Friends, and Family

Each year, Keystone includes two parallel panels of patients with type 1 diabetes, adult and pediatric, and these are always a valuable opportunity to gain insight on the day-to-day of living with diabetes. This year’s sessions certainly did not disappoint; below, we’ve highlighted our favorite and most impactful quotes from the discussion.

Quotable Quotes – Adults

Dean, Debby, Heidi, Leslie, Mike, and Dr. Paul Strumph

On How Far the Field Has Come

  •  “The DCCT that showed tight control mattered didn’t get published until 1993. So, Mike, 20 years after you were diagnosed, people were still arguing about whether glucose control mattered! When I took the boards in 1990, we had to answer a question about whether intensive management improved eye, kidney, or nerve disease or prevented fetal malformation. The only correct answer at that time was fetal malformation. My first A1c in 1980 was over 13%. They said “great!” There was no scale. They were just happy they had another sample.” – Dr. Paul Strumph

  • “I thought urine testing was totally cool – I was like a chemist. The original glucometer was huge and you did the wiping with a wet cotton ball, and when my one friend thought she was high she would squirt more water and rub it down more.” – Leslie

  • “I got diabetes in 1973, and they told my parents I wouldn’t see 16. At 18, they told me I wasn’t supposed to live this long. With that in mind, I chose that I didn’t want to be different, so I did all the stuff you’re not supposed to. I didn’t want to do what I was supposed to do, didn’t care what my blood sugar was, and was consistently in the 300 range. About a year ago, I lost my left foot and then all my right toes. I’m currently on the 670G, and it has been a godsend for me. My A1c a couple weeks ago was down to 6.7%.” – Mike

On Pain Points and Language

  • “My only suggestion to the field is to know your patients as something other than a number. I know it gets so busy, but if we sit down with someone who knows a little about us – just what your job is, something like that – it really helps. This disease affects individuals so differently, and a lot of factors play into that.” – Dean

  • “Outside of the BDC, there have been several bad words. A lot of doctors think they know everything, so open your minds and let the patient teach you. My bad words come more from colleagues and others uneducated about diabetes. They’ll say, ‘You’re eating a piece of cake, but you have sugar diabetes!’ It’s kind of about educating others.” – Heidi

  • “When people find out I just got diabetes, they say, “No, you’re not type 1.’ And I’ve had other doctors tell me that, and I say, ‘Well, I am.’ It kind of bugs me sometimes, and they ask me how I got it, or what did I do. Well, I don’t think that I did anything, but I feel like I have to explain myself.” – Debby

  • “My biggest one is, ‘It’s an easy disease, all you have to do is eat.’ My friends and ex-wife have all said it’s not that hard. I tried, with a couple coworkers, to figure out how many times I’ve had a needle in my body, and we came up with 52,000 pricks. Don’t tell me it’s easy.” – Mike

  • “I hear professional colleagues say, ‘Well, if they only use insulin right, they’d achieve targets.’ Targets are targets because they’re usually unachievable. Not many people can get there, and many HCPs and physicians continue to think it’s just about using insulin appropriately...I’m an endocrinologist and I can’t get my A1c below 7%.” – Dr. Paul Strumph

On Access & Reimbursement

  • “I’ve been fortunate on the city’s health plan, and the cost through that is relatively low. But how am I going to leave that job and that insurance? I have kids and a wife to worry about too. My biggest problem with coverage is that if I’m not employed, I don’t know what my insurance is going to be. Will they keep me on or kick me off? – Dean

  • “The company I used to work for had excellent insurance. Due to my amputation, I was forced to go to other insurance, then I lost my job and I’m on Medicaid. So far, it’s been okay, but I was just informed by Nicole that it doesn’t sound like they’ll pay for the CGM part of the 670G I’m now on, so I have possibly a lot of out of pockets in the near future.” – Mike

  • “My husband has very good insurance, but it’s still $300-400/month out of pocket. I don’t do well with Humalog but that’s the one insurance likes, so Dr. Garg and his team have been huge advocates in getting NovoLog for me. It’s frustrating when the insurance company decides that they know what’s best.” – Leslie

Quotable Quotes – Pediatrics

Sevara (age 6), Christian (age 9), Sage (age 15), Alex (age 16), Melissa (Sevara’s mother), and Tara (Christian’s mother)

On Talking to Parents About Diabetes

  • “There have been a couple of incidences where I have not communicated well with my parents. I think I’ve gotten better at telling them when there are issues. I try to work at that even though I want to be independent. I need to talk to them. It depends on the situation. Sometimes I wish they would lay off a bit but usually it’s good to know they care and want to know how I’m doing.” – Alex

  • Independence is a big thing. The only time I talk to them is if my sensor malfunctions and I tell them that I’ve changed it. Or, if I have a very severe low in the 30s or below. Other than that, everything is pretty independent.” – Sage

  • “He [Christian] actually does really well with communicating with us. I just stressed with him from the beginning that we have to work together. But, I want him to take as much responsibility as possible. We’re constantly keeping an eye on things. He’ll let me know if he’s feeling low and we’ll double check with his Dexcom. If he’s feeling high he’ll let me know. We have phones and we communicate through texting at school.” – Tara

On the Parental Experience and Coping

  • It’s very challenging at times. I’m very grateful that I have a husband to be there and be supportive. When Christian was first diagnosed I tried to take control of everything. It was exhausting. I was getting up for everything in the middle of the night. I talked to my husband after that, and said that we have to share this, we have to do what we can together for him, working together as a team. It is emotional. The key is to talk and to communicate.” – Tara

  • “It’s exhausting. It gets more routine. But the nighttime situation, my husband and I take turns. Sleep concerns are real. Technology is definitely helping us out in regards to that, which is something we are excited about.” – Melissa

On Experiences with Technology

  •  “I don’t hear my alarms. I will wake up when I feel low, and I’ll go downstairs and treat myself. If it needs a calibration, I don’t hear the alarm. My parents do hear them sometimes though…Getting alarms on my phone would be great since it’s always right next to me on my bedside. Part of why I don’t hear my pump alarm is because sometimes I’m on top or under my blankets.” – Alex

  • “In terms of the 670, having access to phone or Bluetooth would be great.” – Melissa

  • “Let’s just say that I don’t hear my alarms. I wake up in the morning with 6,000 alarms. I’m a deep sleeper; I don’t hear them...If I could change one thing, it would be about when my sensor says >200 mg/dl and checking that reading before getting kicked out [of closed loop]” – Sage

On Talking to Others About Diabetes

  • “Classmates at school make assumptions about how you get diabetes and make jokes. I would want people to know the difference between the different types of diabetes. There is a lot of discrimination at school about this, so that would help.” – Sage

  • Education for the general population would make me feel a lot less isolated and different. If they actually knew and understood what was going on, I wouldn’t get as many weird questions on a daily basis.” – Alex

  • “We need to let people know what these kids are going through. I tell Christian—[diabetes] makes you different, yes. But you have to let them know how it doesn’t make you different. He has a lot of kids at school that are very understanding, even at the young age they are at. He still has days where he comes home from school and says that someone made fun of him, and, but then he’ll come back the next day and say that they apologized. Everybody is different and no two people dealing with this are the same.” – Tara


-- Ann Carracher, Martin Kurian, Brian Levine, Peter Rentzepis, Maeve Serino, Adam Brown, and Kelly Close