American Diabetes Association 77th Scientific Sessions

June 9-13, 2017; San Diego, CA; Day #2 Highlights

Executive Highlights

Hello from San Diego, where the PACKED second day of ADA 2017 has just come to a close. This report details our top 22 (!) highlights from this data-heavy day of learning. On the therapy front, we bring you newly released results from trials such as DUAL VII, inTandem1 and 4, DURATION-7, Lira Pump, and the list goes on. Our tech coverage is headlined by the first data from the “limited learning launch” of Medtronic’s Sugar.IQ app with Watson, plus exciting updates from the exhibit hall.  Check out our day #1 highlights for a look at our top takeaways thus far, and take a look at what the next few days of ADA holds with our ADA 2017 resource hub, complete with category documents, our top picks for oral presentations, and a short list for the posters we’re most excited about. Onward!

Diabetes Therapy Highlights

1. To numerous rounds of thunderous applause, Dr. Liana Billings (North Shore University Health System, Skokie, IL) unveiled the impressive results of the DUAL VII trial demonstrating superiority for IDegLira (Novo Nordisk’s Xultophy) over basal-bolus therapy in terms of hypoglycemia, weight loss, insulin dose requirements, and injection burden. It recalled Dr. Buse’s words in 2015 about Xultophy – we hope clinicians and patients will be able to access this drug and not just the components.

2. Lilly presented new data on its novel alternative ultra-rapid insulin formulation LY900014 in two posters, demonstrating that the candidate shows faster absorption and improved postprandial glucose excursions vs. Lilly’s Humalog (insulin lispro) in both type 1 and type 2 diabetes.

3. In two back-to-back oral presentations, UNC’s Dr. John Buse and Lexicon’s Dr. Paul Strumph presented full data from the phase 3 inTandem1 and phase 2 inTandem4 trials, respectively, elucidating the benefit-risk profile of SGLT-1/2 dual inhibitor sotagliflozin in type 1 diabetes. While much of the data was previously reported in topline results, we notably learned that, in inTandem1, more than twice as many patients achieved net benefit (A1c<7% with no severe hypoglycemia or DKA) on sotagliflozin 400 mg compared to placebo: 44% of participants in the 400 mg arm achieved “net benefit” at week 24, compared to just 22% of those in the placebo arm (p<0.001). Those in the 200 mg arm also substantially more likely to achieve net benefit – 34% (p=0.002 vs. placebo). This reinforces the value for people with type 1 diabetes of therapies other than just insulin – truly valuable.

4. Drs. Richard Pratley and Jie Liu shared one-year findings from the VERTIS FACTORIAL and VERTIS SITA2 trials, respectively, underscoring the glucose-lowering and weight loss efficacy of Merck/Pfizer’s SGLT-2 inhibitor ertugliflozin in combination with sitagliptin (Merck’s DPP-4 inhibitor Januvia). This is particularly exciting as we await FDA decisions on ertugliflozin, ertugliflozin/metformin (fixed-dose combination), and ertugliflozin/sitagliptin (fixed-dose combination) expected by January 2018.

5. A team of researchers from Copenhagen University Hospital (Hvidovre, Denmark) presented the results of the highly-anticipated Lira Pump Trial demonstrating improvements in A1c, weight loss, and time-in-ranges with  GLP-1 agonist liraglutide as an addition to insulin pump treatment in people with type 1 diabetes.

6. In a packed oral session, Dr. Juan Frias (National Research Institute, Los Angeles, CA) presented the results from the DURATION-7 study, demonstrating clinical glycemic and weight benefits for AZ’s GLP-1 agonist Bydureon (exenatide once-weekly) as an add-on to insulin therapy in people with type 2 diabetes.

7. Zealand presented a poster detailing phase 2 results from a trial of its liquid-stable glucagon analog dasiglucagon, demonstrating a comparable PK/PD profile to native glucagon.

8. Findings from a preclinical study of Merck’s glucose-responsive insulin candidate (a lectin-bound insulin analog – it’s unclear if this is the same phase 1 MK-2640 GRI candidate) were presented in a poster.

9. Answering the question “What’s the solution to high insulin prices?,” former ADA Chief Scientific and Medical Officer Dr. Robert Ratner stated that there is only one real solution: “Someone has to make less money.” Dr. Ratner noted that rebates to pharmacy benefits managers (PBMs) rose substantially from $67 billion in 2013 to a whopping $106 billion in 2015 – he characterized this issue as “middlemen with no added value increasing expenditures.” Dr. Ratner discussed the potential of several free market trends, regulatory controls, and legislative proposals to impact insulin prices and was particularly optimistic about the forced transparency driven by the ongoing high-profile lawsuits alleging insulin price-fixing.

10. After delivering a rapid-fire overview of the past 25 years of CV data on diabetes drugs, Dr. Matthew Riddle highlighted two shortcomings in diabetes CVOTs to-date: (i) they don’t extend past ~five years and (ii) they mostly enroll participants at very high CV risk. We are very happy to hear the disappointment over a lack of “check-in’s” with patients after five years out articulated by the esteemed Dr. Riddle.

11. Diabetes therapy companies were out in full force at ADA 2017. Below we detail displays from Novo Nordisk, Lilly, Sanofi, AZ, Janssen, and more.

Diabetes Technology Highlights

1. Medtronic’s Dr. Huzefa Neemuchwala shared the first data from the “limited learning launch” of the Sugar.IQ app with Watson. Over an average of just two weeks, the 81 users experienced a solid 37-minute/day gain in time-in-range, an 11% reduction in prolonged hypoglycemia, and an 8% drop in prolonged hyperglycemia. We also saw promising initial engagement metrics. Once it launches, could this differentiate Medtronic’s upcoming standalone Guardian Connect mobile CGM (under FDA review) as a behavior change tool?

2. An Abbott corporate symposium featured tons of enthusiasm for both the FreeStyle Libre consumer and professional versions, as well as additional data from the >55,000 patient real-world cohort (increased time in range, less time in hyperglycemia from increased scanning), the SELFY Study (FreeStyle Libre benefits in 4-17 year olds), and retrospective FreeStyle Libre Pro data from India.

3. We enjoyed playing with several new devices in the exhibit hall (many for the first time!), including Medtronic’s MiniMed 670G, Insulet’s OmniPod Dash PDM, Dexcom’s G5 touchscreen receiver, and LifeScan’s One Touch Via. Lots of thoughts below on what caught our eye!

4. Dexcom’s Dr. David Price shared combined type 1 + type 2 data from the DIaMonD study testing CGM (n=179) vs. SMBG (n=128) in MDIs. From a pooled baseline A1c of 8.6%, CGM drove a 0.9% reduction in A1c at 24 weeks vs. 0.4% in the SMBG group (adjusted difference: 0.5%; p<0.001). Oh-so-notably, CGM users spent 72 more minutes per day in range (70-180 mg/dl) vs. 9 fewer minutes in the SMBG group (p<0.001).

5. A satisfaction analysis from the DIaMonD study reached a very important conclusion: contrary to common clinical belief, patients with type 2 on MDI are actually likely to find CGM at least as engaging and valuable as patients with type 1 diabetes. Combined with time-in-ranges data, this is a most valuable finding and we look forward to hearing, eventually, health outcomes result associated with this.

6. A poster from Dr. Guido Freckmann and colleagues compared the accuracy of the Dexcom G5 and real-time FreeStyle Libre head-to-head. Both devices showed results very consistent with the MARDs published by both companies. The outlier results, however, did show some aberrant sensors.

7. Swansea’s Dr. David Owens presented positive data from the SMBG study, an RCT of SMBG in non-insulin treated type 2 diabetes. Structured SMBG resulted in a 1%+ sustained A1c drop vs. -0.3% in the group that did not check blood glucose (baseline: ~8.6%). Yes! This was good to see for those that doubt the value of users knowing where their blood glucose is.

8. Fascinatingly, Stanford’s Dr. Molly Tanenbaum presented T1D Exchange data showing that surveyed type 1 diabetes clinicians (n=209) can be divided into three distinct profiles with respect to supporting patients with CGM: 20% are “Ready,” 41% are “Cautious,” and 39% fall in the “Not Yet” category. Each group requires different strategies. Our guess is that particularly with AGP, many more can go to the “Cautious” or “Ready” category and fewer need to be in “Not Yet” – the same for patients.

9. UVA’s Dr. Boris Kovatchev shared new preliminary data from the now “complete” training phase of the NIH-Funded International Diabetes Closed-Loop (iDCL) Trial: including a mean glucose of 145 mg/dl, 1% of the time <70 mg/dl, and 77% of the time in 70-180 mg/dl. We continue to hear incredible things about Boris’ work and were elated to see such a high time in range.

Other Highlights

1. As always, the ADA Pathway to Stop Diabetes symposium offered an inspiring glimpse at the future of diabetes management. This is a truly remarkable initiative.

2. Followed by a standing ovation, diabetes educator extraordinaire Ms. Davida Kruger (Henry Ford Medical Group, Detroit, MI) delivered the Outstanding Educator in Diabetes Award Lecture, as she shared valuable insights on her role in the DCCT, as a nurse practitioner in diabetes, and as an ADA volunteer throughout her career. We could not be more impressed with the incredible Davida – hard-won and incredibly deserved honors to her as one of the most impressive diabetes researchers on the planet in our book.

Table of Contents 

Diabetes Therapy Highlights

1. Impressive Dual VII Study Finds IDegLira Superior to Basal-Bolus Therapy in Terms of Hypoglycemia, Weight Loss, Insulin Dose Requirements, and Injection Burden

To numerous rounds of thunderous applause, Dr. Liana Billings (North Shore University Health System, Skokie, IL) presented the impressive results of the DUAL VII trial comparing the safety and efficacy of Novo Nordisk’s Xultophy (insulin degludec/liraglutide) versus traditional basal-bolus therapy in type 2 diabetes. People with type 2 diabetes on insulin glargine (n=506) were randomized in an open-label setting to once-daily Xultophy for 26 weeks or to basal-bolus regimen with insulin glargine and insulin aspart. The trial met its primary endpoint by demonstrating that treatment with Xultophy is non-inferior to basal-bolus therapy with respect to A1c lowering (-1.48% vs. -1.46%, baseline A1c 8.2%) and furthermore demonstrated superiority for a host of secondary outcomes. Compared to basal-bolus therapy, average total daily insulin dose among those taking Xultophy was less than half that of those on basal-bolus therapy (40 vs. 84 units; p<0.001). The end-of-trial body weight difference was also substantially in favor Xultophy (-0.93 kg [2.1 lbs] vs. +2.64 kg [5.8 lbs], p<0.001). Very notably, Xultophy additionally proved superior in terms of hypoglycemia: only 20% of participants in the Xultophy arm experienced blood glucose-confirmed symptomatic or severe hypoglycemia during the study period, versus a whopping 53% of participants in the basal-bolus therapy arm, an extremely impressive 89% risk reduction (HR= 0.11; 95% CI: 0.08-0.17; p<0.0001). Xultophy’s safety profile remained consistent with previous findings; unsurprisingly, nausea was the most frequently-reported adverse event (28 patients in the treatment arm vs. 4 in the basal-bolus arm), but Dr. Billings pointed out that only <3% of the participants in the IDegLira arm experienced nausea at any given time. Indeed, as we understand it, the basal insulin component of these combinations can mitigate the GI side effects typically associated with GLP-1 agonists. During Q&A, the esteemed Dr. Julio Rosenstock deemed this the "most robust, most impressive piece of data I've seen in years." We are equally impressed with these clinical advantages for Xultophy over traditional basal-bolus treatment regimen – especially in terms of the crucial outcome of hypoglycemia risk reduction – and hope these findings help to advance the uptake of the emerging class of GLP-1 agonist/basal insulin co-formulations. We expect to continue to see increasing preference for GLP-1 agonists over prandial insulin as a basal intensification option, and DUAL VII lends strong clinical evidence for this choice. We’ll be back with more on DUAL VII with our take on additional analyses, presented in two posters: one on the cost effectiveness of Xultophy (981-P) and another detailing patient-reported outcomes from DUAL VII (124-LB). While we are happy to see this research, it is our sense that it is extremely challenging for patients to actually gain access to this therapy, despite now years of robust praise.

2. Two Posters Describe a Faster Clinical Profile for Lilly’s Ultra-Rapid Insulin Formulation over Humalog

Lilly presented new data on its novel ultra-rapid insulin formulation LY900014 in two posters, demonstrating that the candidate shows faster absorption and improved postprandial glucose excursions vs. Lilly’s Humalog (insulin lispro) in both type 1 and type 2 diabetes. The posters presented are the first part of a 2-part, 6-period cross-over study of the candidate. In the first study, type 1 diabetes patients (n=30) were randomized to either LY900014 or Humalog. Results demonstrated that the LY900014 group had accelerated early insulin lispro absorption compared to the Humalog group: LY900014 reduced the time to early half-maximal drug concentration by 36.5% (p<0.0001) and the insulin lispro area under the concentration curve vs. time from 0 to 30 min post dose increased by 123% (p<0.0001) compared to Humalog. Additionally, the total glucose excursion over the 5-hour test-meal was significantly reduced by 44% for LY900014 vs. Humalog. Regarding type 2 diabetes, patients (n=29) were similarly randomized to either LY900014 or Humalog. LY900014 reduced the time to early half-maximal drug concentration by 23% (p<0.0001) and increased the insulin lispro area under the concentration curve vs. time from 0 to 30 min by 117% (p<0.0001) compared to Humalog. As for postprandial glucose excursions, the total excursion over the 5-hour test-meal was reduced by 105% (p<0.0001) for LY900014 vs. Humalog. We will have more information to follow on how long the insulin “lasts” etc. Regarding tolerability and adverse events, both patient populations demonstrated no differences in local tolerability and the number or severity of hypoglycemic events. As a reminder, LY900014 is Lilly’s internally-developed candidate that could begin phase 3 trials by the end of 2017. This ultra-rapid insulin has recently received more attention after the disappointing termination of Lilly’s partnership with Adocia on the phase 3-ready ultra-rapid insulin BioChaperone Lispro (see its data from ADA 2016). For more on Lilly’s discussion around the new candidate, please see our 4Q16 update.

  • LY900014’s formulation contains locally-acting excipients, citrate and trepostinil, to help accelerate insulin lispro absorption with the goal of providing ultra-rapid prandial insulin. Specifically, citrate increases vascular permeability at the injection site and treprostinil accelerates insulin lispro absorption by local vasodilation with no measurable systemic exposure. With this formulation, the goal is that LY900014 will more closely mimic physiologic prandial insulin secretion, improving postprandial glucose control and allowing greater flexibility in insulin administration timing.

3. Full Results from the Phase 3 inTandem1 and Phase 2 inTandem4 trials of Lexicon/Sanofi’s SGLT-1/2 Dual Inhibitor Sotagliflozin in Type 1 Diabetes

In two back-to-back oral presentations, UNC’s Dr. John Buse and Lexicon’s Dr. Paul Strumph presented full data from the phase 3 inTandem1 and phase 2 inTandem4 trials, respectively, elucidating the benefit-risk profile of SGLT-1/2 dual inhibitor sotagliflozin in type 1 diabetes. inTandem1 was a double-blinded trial that randomized 793 patients to either placebo, sotagliflozin 200 mg, or sotagliflozin 400 mg. The trial included a six-week insulin optimization period prior to randomization. As was reported in the topline results, even in the context of optimized insulin, the trial impressively met its primary endpoint by producing a mean placebo-adjusted A1c reduction of 0.35% in the 200 mg arm (baseline A1c post-optimization=7.6%; p<0.001) and 0.41% in the 400 mg arm (baseline A1c post-optimization=7.6%; p<0.001) at 24 weeks. inTandem1 employed a hierarchical statistical testing structure, in which secondary endpoints were assessed in the following order: (i) “net benefit” (defined as target A1c<7% with no severe hypoglycemia or DKA); (ii) weight; (iii) bolus insulin dose; (iv) fasting plasma glucose (FPG); and two measures of patient-reported outcomes – (v) the Diabetes Treatment Satisfaction Questionnaire status (DTSQ) score; and (vi) the two-item Diabetes Distress Screening Scale (DDS2) questionnaire score. We previously learned from a second release of topline data from inTandem1 that the 400 mg dose demonstrated superiority to placebo for all of these secondary endpoints, while the 200 mg dose only demonstrated superiority for A1c, net benefit, and weight. Dr. Buse very excitingly provided new detail on these secondary outcomes. More than twice as many patients achieved net benefit on sotagliflozin 400 mg compared to placebo: 44% of participants in the 400 mg arm achieved “net benefit” at week 24, compared to just 22% of those in the placebo arm (p<0.001). Those in the 200 mg arm also substantially more likely to achieve net benefit – 34% (p=0.002 vs. placebo). As was shared in the topline results, patients taking sotagliflozin experienced a 1.6 kg (3.5 lbs) and a 2.7 kg (6 lbs) weight reduction at 24 weeks with the 200 mg and 400 mg doses, respectively – this compares to a mean weight gain of 0.8 kg (1.8 lbs) in the placebo group (p<0.001). The results shared today revealed that the superiority of sotagliflozin 400 mg for the other secondary endpoints was similarly highly statistically significant (p<0.001). On the other hand, Dr. Buse noted that there was some heterogeneity between the impact of the two doses on bolus insulin dose – while sotagliflozin 400 mg was associated with a statistically significant decrease in bolus insulin (p<0.001; exact dosages not shared), there was no statistically significant difference in bolus insulin dose among those treated with sotagliflozin 200 mg and those in the placebo arm. On the other hand, the p-value for the difference between the 200 mg arm vs. placebo was below 0.05 for FPG (p=0.036), DTSQ score (p<0.001), and DDS2 score (p=0.002). That said, because of the hierarchical nature of the statistical testing, Dr. Buse emphasized that these findings may only be characterized as “descriptive” rather than significant. While some more conservative clinical trial purists may make a big deal about this, we’re very, very pleased to see consistency between the 400 mg and the 200 mg dose for these very critical endpoints. We’re especially excited about the quantification of the quality of life and patient satisfaction benefits of this drug – anecdotally, we’ve heard rave review from patients with type 1 diabetes taking selective SGLT-2 inhibitors and we’re glad to see this backed up with hard data. Safety data was as reported in the topline results, with fairly balanced adverse event rates across all three arms, somewhat higher rates of DKA with sotagliflozin, and somewhat lower rates of severe hypoglycemia. While there is some concern in some quarters about type 1 patients taking this class, we don’t think there is a way to stop it – it may be valuable to look at gaps in therapy for type 1s in order to better understand patients taking SGLT-2s off label. We applaud Lexicon for collecting so many outcomes and hope that these will be standardized soon.  

inTandem1 Primary and Secondary Efficacy Endpoint Results

Endpoints

Sotagliflozin 200 mg (p-value vs. placebo)

Sotagliflozin 400 mg (p-value vs. placebo)

Placebo

A1c

-0.43 (p<0.001)              

-0.49% (p<0.001)

-0.08%

Net Benefit (proportion of patients with A1c<7% and no severe hypoglycemia or DKA)

34% (p=0.002)

44% (p<0.001)

22%

Weight

1.6 kg (3.5 lbs, p<0.001)

2.7 kg (6 lbs, p<0.001)

0.8 kg (1.8 lbs)

Bolus insulin

Undisclosed (p=0.10)

Undisclosed

Undisclosed

Fasting plasma glucose (FPG)

Undisclosed (p=0.036)

Undisclosed (p<0.001)

Undisclosed

Patient-reported Diabetes Treatment Satisfaction Questionnaire status (DTSQ) score

Undisclosed (p<0.001)

Undisclosed (p<0.001)

Undisclosed

Patient-reported Diabetes Distress Screening Scale (DDS2) score

Undisclosed (p=0.002)

Undisclosed (p<0.001)

Undisclosed

inTandem1 Safety Results

 

Sotagliflozin 200 mg

Sotagliflozin 400 mg

Placebo

Proportion of patients with ≥1 treatment-emergent adverse events

67%

71%

68%

% patients with diarrhea

7%

10%

7%

% patients with genital mycotic infection

6%

10%

3%

Number of patients with DKA

3 (1.1%)

8 (3.1%)

0

Number with severe hypoglycemia

11 (4.2%)

12 (4.6%)

18 (6.7%)

  • Lexicon VP of Clinical Development Dr. Paul Strumph also took to the stage to discuss full results from the phase 2 dose-ranging inTandem4 study of sotagliflozin. The double-blinded, 12-week trial randomized 141 patients with type 1 diabetes to three doses of sotagliflozin (75 mg, 200 mg, or 400 mg), plus placebo. Unlike the phase 3 inTandem1 trial, inTandem4 only included a two-week placebo run-in period, rather than a six-week insulin optimization period.  The vast majority of the numeric data was already shared in the topline results, though Lexicon shared p-values for the first time, demonstrating that these results were highly statistically significant. The primary endpoint A1c results as well as the secondary urinary glucose excretion, postprandial glucose, body weight, fasting plasma glucose (non-inferior), and blood pressure results are summarized in the table below. Very notably, however, Lexicon shared data on beta-hydroxybutyrate (BHB) levels at baseline and at week 12 – as a measure of ketogenesis, BHB measurements are very helpful as we consider the potential DKA risk of sotagliflozin. Notably, the mean increase in blood BHB at week 12 was only 0.1 mmol/l, which is the lowest value detectable by a point of care BHB meter. Dr. Strumph also noted that this level of BHB increase is lower than what is typically seen with selective-SGLT-2 inhibitors. Overall incidence of DKA was very low this trial, with the only case of DKA in the entire trial occurred in the 400 mg arm – though, very notably, the patient who experienced DKA chose to continue the study after the event, demonstrating that the perceived benefits outweighed the DKA risk for this patient. Although just one case, this is unsurprising to us. The level of DKA observed in both inTandem4 and inTandem1 is somewhat lower in all arms than might be expected in a general type 1 diabetes population. Dr. Buse noted during Q&A following the inTandem1 presentation that all participants in the trial were very carefully and intensively educated on DKA risk management. Our sense is that this level of intensive education rarely happens in the “real-world,” and we hope that Lexicon will continue its leadership in pairing sotagliflozin with appropriate educational efforts when the product launches. Write Richard Wood of dQ&A for more on overall level of understanding by patients of DKA and DKA risk.

inTandem4 Primary and Secondary Efficacy Endpoint Results

Endpoints (placebo-adjusted for all efficacy endpoints)

Sotagliflozin 75 mg (p-value vs. placebo)

Sotagliflozin 200 mg (p-value vs. placebo)

Sotagliflozin 400 mg (p-value vs. placebo)

A1c

-0.25% (p=0.07)

-0.48% (p<0.001)

0.38% (p=0.006)

Urinary glucose excretion (g/day)

42 (p=0.006)

58 (p<0.001)

70 (p<0.001)

Change in 2-hour postprandial glucose reduction (mg/dl)

-20 (p=0.27)

-27 (p=0.15)

-49 (p=0.006)

Body weight

-1.3 kg (2.9 lbs; p=0.38)

-2.4 kg (5.3 lbs; p<0.001)

-2.8 kg (6.2 lbs; p<0.001)

Change in fasting plasma glucose (mg/dl)

-9 (p=0.51)

-9 (p=0.48)

-21 (p=0.10)

Change in systolic blood pressure (mmHg) in subgroup of patients with baseline ≥130 mmHg

-8.4 (p=0.26)

-6.8 (p=0.28)

-14.3 (p=0.013)

inTandem4 Safety Results

 

Sotagliflozin 75 mg

Sotagliflozin 200 mg

Sotagliflozin 400 mg

Placebo

Proportion of patients with ≥1 treatment-emergent adverse events

48.6%

28.6%

34.3%

50%

Number of patients with diarrhea

0

1

1

0

Number with genital mycotic infection

1

1

1

0

Number with DKA

0

0

1 (2.9%)

0

Number with severe hypoglycemia

1 (2.9%)

1 (2.9%)

1 (2.9%)

0

4. Two New VERTIS Readouts Underscore Efficacy of Ertugliflozin, Especially in Combination with DPP-4 Inhibitor Sitagliptin

Drs. Richard Pratley and Jie Liu shared one-year findings from the VERTIS FACTORIAL and VERTIS SITA2 trials, respectively, underscoring the glucose-lowering and weight loss efficacy of Merck/Pfizer’s SGLT-2 inhibitor ertugliflozin in combination with sitagliptin (Merck’s DPP-4 inhibitor Januvia). This is particularly exciting as we await FDA decisions on ertugliflozin, ertugliflozin/metformin (fixed-dose combination), and ertugliflozin/sitagliptin (fixed-dose combination) expected by January 2018 – we’ve been looking forward to this data for a long time! In VERTIS FACTORIAL, 1,233 patients with type 2 diabetes on stable metformin therapy were randomized to one of five arms: (i) a 5 mg dose of ertugliflozin, (ii) a 15 mg dose of ertugliflozin, (iii) a 100 mg dose of sitagliptin, (iv) a combination regimen of 5 mg ertugliflozin/100 mg sitagliptin, or (v) a combination regimen of 15 mg ertugliflozin/100 mg sitagliptin. Initial 26-week results were presented in a poster at EASD 2016, and Dr. Pratley today discussed data from the trial extension out to 52 weeks, which demonstrated sustained superior reductions in A1c and fasting plasma glucose for the SGLT-2/DPP-4 combinations vs. either agent alone. After one year of treatment, A1c dropped by 1.4% from a baseline 8.6% for both the combo therapy groups vs. an A1c decline of 1%, 0.9%, and 0.8% for the 5 mg ertugliflozin, 15 mg ertugliflozin, and 100 mg sitagliptin groups, respectively (and that’s in an RCT – unlikely that would be seen in “real life” from our admittedly speculative view). The treatment difference for A1c decline was statistically significant in favor of the combination vs. either monotherapy in all comparisons (p<0.05). In both combo arms, 40% of participants reached an A1c goal of <7% after one year vs. 26% of patients on 5 mg ertugliflozin, 23% of patients on 15 mg ertugliflozin, and 27% of patients on sitagliptin. A “global” word on findings such as these - we’d love to see what the A1c goals would look like with other interventions combined after the initial findings (diet, exercise, behavior modification, peer to peer counseling, etc.)

  • Dr. Liu presented 52-week results from VERTIS SITA2 investigating ertugliflozin as an add-on to a metformin/sitagliptin regimen – a follow-up to the 26-week data shared in an oral presentation at EASD 2016. Participants (n=464) with type 2 diabetes experienced a mean 0.8% A1c decline from a baseline of 8.1% with 5 mg ertugliflozin after one year, while patients in the placebo arm showed essentially flat A1c from baseline 8% (p<0.05). Similarly, people randomized to 15 mg ertugliflozin experienced a mean 0.8% A1c reduction from a baseline of 8% (p<0.05 vs. placebo). Dr. Liu displayed graphs to clearly depict this sustained glucose-lowering effect out to one year, which is certainly compelling, in our view. In VERTIS SITA2, 33% of patients on ertugliflozin reached an A1c target <7% at week 52, regardless of dose, compared to only 14% of the placebo group. We’re glad to see such strong data from the VERTIS program, and we’re excited by the prospect of a fourth SGLT-2 inhibitor coming soon to the commercial market by a group that could more successfully get this class to the patients that need it. Expanded options within this class will be a noteworthy win, and we look forward to seeing how ertugliflozin may grow whole class sales – while patients and clinicians in the US are rarely “choosing” compounds anymore, as much as the formularies are making choices for them, we do believe that a powerhouse combo of Merck/Pfizer marketing in this class would be positive for all.
  • Ertugliflozin also demonstrated a profound weight loss benefit in VERTIS FACTORIAL and VERTIS SITA2. Dr. Pratley showed that ertugliflozin was associated with 2.4 kg (5.3 lbs) weight loss at a 5 mg dose, 3.2 kg (7.1 lbs) weight loss at a 15 mg dose, 2.4 kg (5.3 lbs) weight loss at a 5 mg dose in combination with sitagliptin, and 2.8 kg (6.2 lbs) weight loss at a 15 mg dose in combination with sitagliptin. Meanwhile, patients on sitagliptin alone in VERTIS FACTORIAL experienced 0.1 kg (0.22 lbs) weight loss on average over the course of a full year. Baseline body weight was 87 kg-89 kg (192 lbs-196 lbs) across all study arms. Weight loss effects were consistent between week 26 and week 52 in VERTIS SITA2. According to Dr. Liu, body weight decreased by a mean 3.5 kg (7.7 lbs) from a baseline 88 kg (194 lbs) for the 5 mg ertugliflozin group, by a mean 2.8 kg (6.2 lbs) from a baseline 87 kg (192 lbs) the 15 mg ertugliflozin group, and by 1 kg (2.2 lbs) from a baseline 87 kg (192 lbs) for the placebo group. Weight loss is one of the key aspects to SGLT-2 inhibitors that makes these products so appealing for type 2 diabetes patients (not to mention the superior A1c-lowering and possible CV benefits). We’re not at all surprised to see marked improvements in body weight with ertugliflozin treatment vs. sitagliptin or placebo, but we’re always happy to see corroborating evidence on this front. We’re also pleased to see that combination with sitagliptin does not appear to substantially attenuate the weight loss effect of ertugliflozin.
  • Both presentations highlighted SGLT-2/DPP-4 fixed-dose combination tablets as an important advancement in type 2 diabetes care. Both Drs. Pratley and Liu began their presentations with the statement that SGLT-2 inhibitors and DPP-4 inhibitors have complementary mechanisms of action, which suggests that a combination of the two could be a more efficient way to get patients to goal – especially those with high baseline A1c. VERTIS FACTORIAL and VERTIS SITA2 provide hard evidence for the benefits to this particular combination vs. either monotherapy, with more patients getting to A1c goal while also losing weight and facing a lower pill burden. As Dr. Liu put it, DPP-4 inhibitors are already commonly considered as a second-line therapy option after metformin, and this data now supports earlier intervention with an SGLT-2 inhibitor alongside. Merck’s Januvia (sitagliptin) leads the DPP-4 inhibitor class in sales and has surely cultivated familiarity among diabetes care providers, so we’re hopeful that a fixed-dose combination of ertugliflozin/sitagliptin – if approved – also becomes an early consideration for type 2 diabetes treatment. That said, frustratingly in the US, clinicians and payers seem to be less comfortable with fixed dose combinations than do their counterparts in other geographic regions. We’d love to get more patient sentiments, particularly that relate to adherence, to the payers.

5. Lira Pump Trial Demonstrates Improvements in A1c, Weight Loss, and Time-in-Range with Liraglutide Added to Insulin Pump Treatment in Patients with Overweight and Type 1 Diabetes

A team of researchers from Copenhagen University Hospital (Hvidovre, Denmark) presented the results of the highly-anticipated Lira Pump Trial demonstrating the clinical potential of the GLP-1 agonist liraglutide as an addition to insulin pump treatment in people with type 1 diabetes. People with type 1 diabetes using CSII and with overweight (n=44) were randomized to receive liraglutide 1.8 mg or placebo in addition to regular insulin pump therapy for 26 weeks. Dr. Thomas Dejgaard outlined the positive primary outcomes: compared to placebo, liraglutide as an add-on to insulin pump treatment produced significant improvements in A1c (-0.6% vs. +0.2%; p<0.001) and striking weight loss (-7.3 kg [16.1 lbs] vs. -0.6 kg [1.3 lbs]; p<0.001) results without increasing the time spent in hypoglycemia (3.2 vs. 3.5 events; p=0.77). While the A1c reduction may be perceived as fairly modest by some for a GLP-1 agonist, we’re especially impressed given the low A1c baseline, by the substantial weight loss, and by what increased time in ranges were. Dr. Kirsten Nørgaard followed with a deeper dive into the impact of add-on liraglutide to insulin dosing and time-in-range. Excitingly, liraglutide as an add-on to insulin pump therapy gave participants significantly more time spent in the normoglycemic range (57% vs. 45%) – a crucial outcome for patient quality of life and perceived therapeutic and personal success (as Close Concerns, The diaTribe Foundation, and dQ&A will elucidate tomorrow in our very own poster!). That said, there was no statistically significant difference in time spent in hypoglycemia (defined as <3.9 mmol/l [70 mg/dl]). We were extremely surprised to hear that liraglutide also produced no change in total daily dose of insulin – this seems extremely surprising to us and we are checking again with the authors, particularly given that the highest dose of lira was given to these patients. There was also no change in distribution between basal and bolus insulin. Furthermore, as expected for a GLP-1 agonist, nausea was the most commonly reported adverse event for participants randomized to the liraglutide treatment arm; 64% of participants experienced this, though it was transient and disappeared over the first few weeks of the study (we’d love to know more about how tolerable the nausea was). Currently approved for type 2 diabetes (under the trade name Victoza) and obesity (under the trade name Saxenda), these results underscore the potential of liraglutide for type 1 diabetes as well. For more detail on the Lira Pump results, plus relevant Q&A, refer to our Detailed Discussion and Commentary section below.

6. Exenatide Once-Weekly offers Advantages in A1c and Weight Loss as an Add-On to Insulin Therapy in the DURATION-7 Trial

In a packed oral session, Dr. Juan Frias (National Research Institute, Los Angeles, CA) presented the results from the DURATION-7 study, demonstrating clinical glycemic and weight benefits for AZ’s GLP-1 agonist Bydureon (exenatide once-weekly) as an add-on to insulin therapy in people with type 2 diabetes. Participants already on insulin therapy (n=461) were randomized to once-weekly Bydureon or placebo in combination with their existing insulin regimen. With Bydureon, participants experienced a significantly greater mean reduction in A1c (-1.0% vs. -0.3%, baseline A1c=8.5%, p<0.01) after 28 weeks, and 25% more patients in the Bydureon treatment arm achieved an A1c of <7% by the end of the study. Therapeutic intensification with Bydureon furthermore reduced fasting plasma glucose and 2-hour postprandial glucose by an average of 9 mg/dl (p=0.028) and 27 mg/dl (p<0.001) versus placebo, respectively. Notably, participants in the Bydureon arm additionally lost an average of 1.5 kg (3.3 lbs) relative to placebo (p<0.001), and 20% more participants using Bydureon achieved the composite endpoint of an A1c <7% with no weight gain or major hypoglycemia. Participants taking add-on Bydureon also experienced a 2 unit/day decrease in daily insulin requirements, but this did not achieve statistical significance (p=0.07). Adverse events were not significantly different between the treatment and placebo arms (125 vs. 133 events); gastrointestinal complaints were numerically higher in the exenatide group (35 vs. 25 events), as expected with a GLP-1 agonist, and no severe hypoglycemia events occurred in either treatment arm. Together these results underscore the superiority of exenatide once-weekly in combination with basal insulin over basal insulin therapy alone.

7. Zealand’s Liquid-Stable Dasiglucagon Shows Comparable PK/PD Profile to Native Glucagon in Phase 2 Study

Zealand presented a poster detailing phase 2 results from a PK/PD study of its liquid-stable glucagon analog dasiglucagon. The trial compared four doses of dasiglucagon (0.1 mg, 0.3 mg, 0.6 mg, and 1.0 mg) to native glucagon (Novo Nordisk’s GlucaGen; 0.5 mg and 1.0 mg) in participants with type 1 diabetes (n=58), demonstrating that dasiglucagon, like native glucagon, was able to rapidly and effectively increase blood glucose concentration following a controlled, insulin-induced hypoglycemia event in which blood glucose fell to 55 mg/dl. On the pharmacodynamic front, the three highest dasiglucagon doses (0.3 mg, 0.6 mg, and 1.0 mg) each took a median of six minutes to raise plasma glucose levels to above 70 mg/dl, versus 6 and 7 minutes for the 0.5 mg and 1.0 mg doses of native glucagon, respectively (and 10 minutes for the 0.1 mg dasiglucagon dose). Dasiglucagon furthermore showed comparable performance to native glucagon in terms of the average time required to increase plasma glucose levels by 20 mg/dl; the four dasiglucagon doses (0.1 mg, 0.3 mg, 0.6 mg, and 1.0 mg) respectively took a median of 14, 10, 9, and 9 minutes to achieve this level of plasma glucose reduction, versus 10 minutes for both the 0.5 mg and 1.0 mg native glucagon doses. Notably, the increase in plasma glucose was longer-lasting and more pronounced with dasiglucagon: total plasma glucose excursions (as measured by area under the effect curve [AUE]) were significantly higher with 0.3 mg dasiglucagon versus 0.5 mg native glucagon (p<0.0001) and for both 0.6 mg (p=0.0043) and 1 mg (p<0.0001) dasiglucagon versus 1.0 mg native glucagon. This larger glucose excursion profile promisingly suggests that dasiglucagon may be superior to native glucagon at preventing recurrent hypoglycemia. On the pharmacokinetic front, dasiglucagon reached maximum exposure significantly later than native glucagon (~35 minutes; p<0.01), but demonstrated a significantly higher total exposure for all dose comparisons (p<0.001). In a separate conversation with Zealand, the company indicated that this could be suggestive of higher viability for dasiglucagon over native glucagon, meaning that dasiglucagon could achieve the same the same glucose increasing efficacy at lower doses – but of course this would need to be verified in future studies. Importantly, dasiglucagon was well-tolerated with a similar safety profile as native glucagon. For both dasiglucagon and native glucagon, the most commonly reported adverse events were nausea and vomiting (40 events with dasiglucagon, 35 events with native glucagon) and mild injection site reactions which disappeared within 30 minutes post-dosing (7 events with dasiglucagon, 5 events with native glucagon).

  • Following the release of these topline results in 3Q16, Zealand commented in its 4Q16 earnings update that the FDA is expected to provide guidance of dasiglucagon’s development in 1Q17. The candidate is currently in phase 2 development both as a single-dose rescue pen (phase 3 to begin as soon as 3Q17) and as a multi-dose component of a dual hormone artificial pancreas system. We are very pleased to see this progress toward a more patient-friendly and feasible alternative to the current very complex, error-prone native glucagon kits – a longstanding area for improvement in diabetes care. For more on the soluble glucagon arena more broadly, check out our soluble glucagon competitive landscape.

8. Merck Shares Preclinical Data on Glucose-Responsive Insulin

Findings from a preclinical study of Merck’s glucose-responsive insulin candidate (a lectin-bound insulin analog – it’s unclear if this is the same phase 1 MK-2640 GRI candidate) were presented in a poster. The study investigated the insulin in a dog model (n=8) vs. canine or human insulin, with each animal receiving both treatments in random order. Infusion of the insulin was adjusted in a low-glycemic condition to match canine/human insulin levels to bioequivalence. In a condition of hyperglycemia, the liver then reduced extraction of the insulin from the blood, allowing more of the glucose-responsive insulin to remain in active circulation. Greater plasma levels of the drug were correlated with increased hepatic glucose uptake (p<0.05 vs. control), non-hepatic glucose uptake (p<0.05 vs. control), and overall glucose disappearance (p<0.001 vs. control). Tracer-determined glucose levels declined in the presence of the insulin analog (p<0.05 vs. control). These results – which are extremely early-stage – support the potential of a lectin-bound molecule for smart insulin delivery, and the poster concludes that this is an approach to glucose-responsive insulin worth further consideration. That said, this very early-stage study was limited in that there’s no indication (yet) of how sensitive this lectin-bound analog will be in humans or whether it will prove clinical utility. Glucose-responsive insulin is one of the most challenging areas of diabetes research, given insulin’s narrow therapeutic range, the complexity of insulin dosing/titration, and the safety concerns that stem from releasing insulin into the bloodstream but expecting it to turn “on” and “off” automatically. We’re not giving up hope, and we’re so glad that Merck’s continued research interest in this area. Merck also has a glucose-responsive insulin candidate – MK-2640 – that finally completed its phase 1 trial in August 2016 (per its ClinicalTrials.gov page) after being delayed several times.   Whether or not the company decides to move forward with this preclinical candidate or with a phase 2 program for MK-2640 (and at this point, we’re not sure, at least in the near-term), these results will be incredibly informative to the community of researchers working on smart insulin. This includes Dr. Zhen Gu’s lab at the University of North Carolina (which recently published positive preclinical results on a smart insulin patch and was highlighted today during the ADA Pathways symposium) and many other preclinical candidates outlined in our insulin competitive landscape. Merck’s MK-2640 remains the most advanced pipeline candidate in this area, and we would of course love to see ongoing clinical development, but we’ll have to wait-and-see. Dr. Gu continues to offer big inspiration to us!

9. Insulin Pricing: Complicated Issues with Complicated Solutions

Answering the question “What’s the solution to high insulin prices?,” former ADA Chief Scientific and Medical Officer Dr. Robert Ratner stated that there is only one real solution: “Someone has to make less money.” He emphasized that the system of insulin pricing/coverage/rebating is extremely complicated and that a lack of transparency makes it difficult to understand where the bulk of the money is going. That said, Dr. Ratner noted that rebates to pharmacy benefits managers (PBMs) rose substantially from $67 billion in 2013 to a whopping $106 billion in 2015. While acknowledging that an appropriate level of profit from insulin supports R&D and prices in the US subsidize insulin in low- and middle-income countries, Dr. Ratner stated to great applause, “The issue of middlemen with no added value increasing expenditures is something we can get rid of without having any downward pressure on insulin pricing elsewhere” – hear hear! Dr. Ratner views free market competition mechanisms as the most likely to be successful in driving insulin pricing reform. He acknowledged that the very first biosimilar insulin glargine – Lilly/BI’s Basaglar – has not been discounted quite as much as some had hoped (in the US, the follow-on biologic is discounted about 15% relative to originator Lantus). On the other hand, he underscored that we typically must wait for two or more generics to see price driven down, and Dr. Ratner was fairly optimistic on the future launch of additional biosimilars given that new FDA commissioner Dr. Scott Gottlieb has publicly highlighted simplification of the biosimilar regulatory process as one of his goals. Even with the current, rather complicated process, Merck has already submitted a biosimilar insulin glargine formulation and Mylan/Biocon are planning to submit their own formulation shortly. Dr. Ratner also sees greater proliferation of direct-to-consumer programs as promising – he highlighted Walmart’s $26/vial human insulin as an excellent example of how the complex insurance/PBM/rebate system can be bypassed entirely to bring insulin directly to patients at a lower cost. We think that the Blink Health and InsideRx direct discount programs are a good example of this principle as well, though the prices for medications through these programs are admittedly still much higher than that of Walmart insulin. Most promising in Dr. Ratner’s view, however, are the recent proliferation of lawsuits accusing insulin manufacturers, payers, and PBMs of price-fixing in insulin. While he reserved judgement on whether any of these lawsuits had merit, he pointed out that they will force transparency in this opaque field by forcing various stakeholders to disclose their contractual agreements as part of discovery. We think it’s highly unlikely that any actual price-fixing is occurring and, while we wholeheartedly agree with the need for transparency, we wish it wouldn’t take several high-profile lawsuits to encourage companies to disclose their financial agreements. We certainly believe voluntary disclosures would generate more goodwill, while these lawsuits that only further exacerbating the public furor over drug pricing. The “story” has spun out of control in many ways and we continue to hope that manufacturers and other stakeholders will take the initiative to address these concerns head-on – without being pushed by external forces. Some progress has been made in recent months, but there is still much more room to help relieve the financial burden of diabetes care for patients.

  • On the other hand, Dr. Ratner was less optimistic about the potential for legislative or regulatory controls to offer near-term relief on insulin pricing. He noted that there has been significant discussion among politicians and in legislature (both at the state and the national level) on potential policy solutions to the issue of high drug costs. He highlighted proposals to allow Medicare Part D to negotiate with manufacturers, government-imposed price controls, and challenges to “pay for delay” tactics, but ultimately emphasized that the legislative process can be incredibly slow. On the regulatory front, he suggested that some of the proposals floating around could have unintended, negative consequences from a global perspective. For instance, he noted that calls to allow the importation of drugs from other counties have become popular in some circles and that this would actually be possible within the current purview of the FDA, without need for legislative change. That said, Dr. Ratner pointed out that the last four FDA commissioners (including Dr. Gottlieb) has come out “categorically against” this idea. Further, Dr. Ratner suggested that cross-border importation of insulin would actually incentivize companies to only sell insulin in companies that can command a higher price, like the US, and potentially create insulin shortages in other countries.

10. CVOTs: Past, Present, and Future

After delivering a rapid-fire overview of the past 25 years of CV data on diabetes drugs, Dr. Matthew Riddle highlighted two shortcomings in diabetes CVOTs to-date: (i) they don’t extend past ~five years and (ii) they mostly enroll participants at very high CV risk. He presented both as important future directions for CVOTs – from a patient perspective, we are very disappointed that these important trial didn’t include in the design even simple check in points for patients, particularly given the investment required to get the trials going. Research shows that interventions can have legacy effects – as was the case of intensive glucose-lowering in the DCCT – and Dr. Riddle explained that follow-up longer than five years is necessary in order to fully capture the risk/benefit profile of therapies. He summarized baseline characteristics of patients in Lilly/BI’s EMPA-REG OUTCOME (for SGLT-2 inhibitor Jardiance), Novo Nordisk’s LEADER (for GLP-1 agonist Victoza), and Novo Nordisk’s SUSTAIN 6 (for GLP-1 agonist semaglutide) – all positive CVOTs, and all featuring people with high baseline CV risk. A key justification for this element of trial design is that conducting a CVOT is massively expensive and time-consuming. In order to accumulate sufficient CV events for statistical analysis, these studies must continue for many years and they benefit from enrolling people more likely to experience a CV event. Nonetheless, this remains a limitation of CVOTs published so far. Indeed, the new Jardiance indication incorporating EMPA-REG OUTCOME data is for the reduction of CV death only in type 2 diabetes patients with a prior history of CV disease – CVOTs in a broader population might support a broader indication, which would be an incredible victory for the patient community. Evolving thinking of CVOTs is already shaping up to be a major theme of this meeting and it’s only day #2 – we also heard Dr. Lawrence Leiter discuss this topic in detail on Day #1 of ADA. We’ve heard from other leading thought leaders as well, including Dr. Silvio Inzucchi, that investigating primary CV prevention should be the next frontier for diabetes CVOTs, especially SGLT-2 inhibitors. Lucky for us, J&J’s CANVAS trial investigating the CV effects of SGLT-2 inhibitor Invokana (canagliflozin) reports full results on Monday, and includes a subset of participants without existing CV disease at baseline. AZ’s DECLARE for SGLT-2 inhibitor Farxiga (dapagliflozin) features an even larger pool of low CV risk participants, but this CVOT isn’t expected to complete until April 2019.

11. Diabetes Therapy Exhibit Hall Highlights

Diabetes therapy companies were out in full force at ADA 2017. Novo Nordisk’s commanding booth emphasized Tresiba (insulin degludec) and Victoza (liraglutide), with the combination product Xultophy taking a back seat. Both the standalone Lilly and Lilly/BI booths – which each commanded space on opposite ends of the exhibit hall – highlighted Jardiance’s unique CV indication (the words “Only Jardiance” repeatedly flashed within a heart shape). AZ’s booth was dominated by Farxiga (dapagliflozin), consistent with the company’s positioning of the drug as a cornerstone of its new Cardiovascular and Metabolic Disease unit. Janssen continued to highlight Invokana (canagliflozin) as the #1 prescribed SGLT-2 inhibitor in the US. Sanofi’s booth heavily featured the company’s newest offerings Toujeo (insulin glargine U300) and Soliqua (lixisenatide/insulin glargine) along with an exciting virtual reality demonstration of severe hypoglycemia. Both AZ and Sanofi displayed their diabetes and cardiovascular products side by side, a testament to the blurring distinctions between those two fields.

Diabetes Technology Highlights

1. New Data on Medtronic’s Sugar.IQ with Watson: 37 Minutes/Day Improvement in Time-in-Range; Encouraging Engagement Over ~2 Weeks of Use in 81 Users

Medtronic’s very smart Head of Innovation Dr. Huzefa Neemuchwala shared the first data from the “limited learning launch” phase of the Sugar.IQ app with Watson (“Intelligent Diabetes Assistant App”). Results came from de-identified CareLink data in 81 users of the Sugar.IQ app using MiniMed 530G/Enlite + MiniMed Connect to send CGM data to the app. Relative to baseline metrics (one month prior), this small group of Sugar.IQ users has experienced a solid 37-minute/day improvement in time-in-range (p=0.04; baseline not shared), an 11% reduction in sustained hypoglycemia (>120 minutes; p<0.001), and an 8% drop in sustained hyperglycemia (>120 minutes; p<0.001). Notably, within three days of the app delivering a pattern “insight,” 65% of users have experienced fewer lows and 55% experienced fewer highs. In total, Sugar.IQ has now been used by 97 people for an average of two weeks each, and engagement has been encouraging in this very limited launch: an average of 1.5 unique app sessions per day, 78% of users logging food, and 4.8 logged food items per day – persistence over time, particularly with food and opening the app up, will be THE key question ahead. The very cool Glycemic Assist feature, allowing users to “follow” a particular food item over time (we love this!), has been popular too: 1,886 views in these 97 users so far. Users have “followed” their glycemic response to Dunkin Donuts, Panera Bread, corn flakes, ice cream, buttermilk biscuits, etc. – pretty squarely in the junk food (“Diabetes Landmines,” as Adam would put it) category, but hopefully the app will gradually nudge people away from eating them! We include examples below of the insights Sugar.IQ delivers – so far, this group of users has received 1,119 different contextual, personalized insights ranging from glycemic control and behavior to hyper- and hypoglycemia to rapid rate-of-change to boluses. Users have “liked” a notable 89% of the insights, indicating they are finding useful patterns. Dr. Neemuchwala also shared two Sugar.IQ case studies from patients with longstanding diabetes – the app identified trends (over-correcting highs, eating a high-carb lunch) and nudged them to change their behavior (the latter is a phrase Dr. Neemuchwala emphasized). It’s great to see the significant investment in this project continue, and if engagement remains high, Sugar.IQ could prove to be an excellent behavior change/decision support differentiator from other competitive standalone CGM offerings. (To say nothing of the service model upside, ability to apply learning to closed-loop algorithms, etc.)

  • It’s unclear when Sugar.IQ will launch fully, but we assume the biggest gating factor is approval of the standalone Guardian Connect mobile CGM (under FDA review and currently in human factors testing). Guardian Connect will stream CGM data to Sugar.IQ directly via Bluetooth, and should help Medtronic add significant value to its mobile CGM. Per Medtronic’s JPM presentation (the last update), a full launch of Sugar.IQ and Guardian Connect were expected in May-October, though today’s Medtronic Diabetes Analyst Day update said that human factors work is ongoing and FDA discussions continue. Medtronic also told us it is working with IBM to finalize the algorithm for the Sugar.IQ commercial launch.
  • Read the detailed write-up below for much more, including several screenshots from this fascinating talk, Medtronic’s strategy, and an update on the improved hypoglycemia prediction feature.

2. New Abbott FreeStyle Libre Data: Real-World Country-by-Country, in 4-17 year Olds, & Pro in India

  • Dr. Olga Kordonouri (Children’s Hospital, Hannover, Germany) presented data from the single-arm SELFY Study of FreeStyle Libre in pediatrics (ages 4-17), showing strong 1+ hr improvements in time-in-range per day vs. baseline. At the end of the study period (two weeks blinded compared to eight weeks unblinded), participants (n=76; 58% pump users) saw time in range improve by a whopping 1 hour/day (p=0.0056), A1c drop by 0.4% (p<0.0001; baseline 7.9%), and time >180 mg/dl fall by 1.2 hours/day (p=0.0038). To our surprise, time spent <70 mg/dl did not change significantly vs. baseline, even though decrease in hypoglycemia was the primary achievement in the IMPACT study of FreeStyle Libre in type 1 adults – perhaps this pediatric group had parents much more mindful of hypoglycemia. Pediatric patients scanned 12.7 times per day, on average, nearly as frequently as the 15.1 times per day seen in their adult counterparts in IMPACT, whereas SMBG frequency fell from a median of 8.0 to 1.0 per day during the eight week portion – talk about confidence in the system! Median sensor duration was 13.4 days (a good sign for adhesive in an active population), there were only three reported mild device-related events, and both parents and teens indicated increased treatment satisfaction. There was no formal control group in this study, but it is evident that pediatric patients have improved glycemic outcomes on FreeStyle Libre and will use the sensor, and the increased discretion and reduced need to perform fingersticks are equally compelling, particularly for this age group. This data will also be presented Sunday in the poster hall (110-LB).
  • Dr. Ramzi Ajjan supplemented real-world FreeStyle Libre data presented at ATTD with new cuts showing increased time in range and decreased time in hyperglycemia from increased scanning, and even a country-by-country breakdown! The ATTD data showed that when individuals from the >55,000-user-strong cohort scanned at higher frequencies, they presented with lower A1cs and spent less time in hypoglycemia. The additional results today also suggested that scanning with a higher frequency increased time in range and reduced time in hyperglycemia – we didn’t get clear pictures to parse out the magnitude of the effect, but we’ll be back in short order as soon as we do. On a more encouraging note, we were able to snag the very cool photo below – it depicts the regional differences in baseline hyperglycemia (hours per day above 180 mg/dl) and hypoglycemia (hours per day below 70 mg/dl), as well as the effects of sustained use (clockwise from top left: Germany, Spain, France, UK, Italy, and other). It is fascinating to see how the healthcare (and diabetes care) systems, diets, and other customs contribute to slightly different population-wide glycemic patterns. That said, more FreeStyle Libre scans, no matter which country, consistently resulted in lower hyperglycemia and hypoglycemia – very confidence inspiring. With two major outcomes trials (IMPACT, REPLACE), 300,000+ patients using the device (many non-adjunctively) overseas, and real world in this >55,000-user group, it will be interesting to see how the FDA approaches this product review (especially for non-adjunctive use).

Real-world Data: Regional Differences

  • Acclaimed diabetologist Dr. V. Mohan presented retrospective, real-world data demonstrating significant A1c reductions from the use of FreeStyle Libre Pro across seven diabetes clinics in India. Dr. Mohan et al. mined the EMRs of these diabetes clinics for data from patients from March 2015 to October 2016 – one group had used FreeStyle Libre Pro (n=2,536), while the other consisted of matched controls (n=2,536). Overall, patients in the control group saw an average A1c reduction of 0.7% over the time period (baseline 9.3%), while the FreeStyle Libre Pro group declined 1.0% (from 9.3%). The type 1 cohort who used Pro saw a 0.7% reduction (baseline 9.6%) vs. a 0.2% reduction in the matched controls (baseline 9.6%), while the type 2 Libre Pro cohort experienced a 0.9% decline (baseline 9.2%) vs. a 0.7% reduction in the matched controls. There were no significant differences between male and females in the study, and impressively, Libre Pro use yielded A1c reductions across the board, irrespective of age. We did find it odd that the control group improved so much, since this wasn’t a clinical trial. The incremental gain with Libre Pro was a 0.5% A1c advantage in type 1 and a 0.2% advantage in type 2, both from very high baselines. We might have expected larger declines, but this wasn’t a formal study. A1c reductions were seen at all seven sites, though there was a very wide range of observed declines from site to site – the smallest A1c drop was 0.4%, while the largest was 2.1%! Why are some clinics doing so much better than others, despite starting from similar baselines? Is it an inherent property of the demographic a clinic serves, or is it a modifiable procedural component of care, or a mixture of both? Dr. Mohan and co. did perform multiple regressions to determine the factors that made patients more likely to be Libre Pro non-responders (aka to not see a marked A1c drop after 14 days of wear): They found that longer duration of diabetes, time to follow-up A1c test (the beneficial A1c effect disappears), and insulin user were all associated with a lack of A1c reduction.
    • Dr. Mohan offered some grim commentary on diabetes care in India: He reminded us that patients pay for everything out of pocket – for this reason, there are a miniscule number of patients on pumps, and “you’re lucky if you can get them on MDI.” He later added in Q&A that a sensor in India costs 2,000 rupees (~$30), but physicians tack on ~1,000 rupees for application, data download, data review, etc., putting the total closer to ~$45. A Libre Pro sensor in the US costs $60, but is covered by most insurance plans for up to four bouts per year. FreeStyle Libre is definitely a technology that can scale to meet a market like India, so we hope with volume the cost of Pro can come down even further – or perhaps other relationships could be forged with government to cover some of the costs.

3. Exhibit Hall Highlights: Insulet’s OmniPod Dash PDM, MiniMed 670G, Dexcom’s G5 Touchscreen receiver, LifeScan’s One Touch Via

We enjoyed playing with several new devices in the exhibit hall (many for the first time!), including Medtronic’s MiniMed 670G hybrid closed loop (interesting takeaways on user interface nuances), Insulet’s OmniPod Dash PDM (major, major improvement!), Dexcom’s G5 touchscreen receiver (very durable – built like a tank), and LifeScan’s One Touch Via (manufacturing process cleared, and great to FINALLY see this in a hall). We also enjoyed visiting Abbott, Ascensia, BD, and Tandem for a refresher on their latest offerings. Read the exhibit hall section below for a company-by-company snapshot!

4. T1 + T2 Pooled Data From Dexcom’s DiaMonD Study: -0.5% A1c Advantage with CGM, 72 More Minutes in Range vs. SMBG

Dexcom’s Dr. David Price shared combined type 1 + type 2 data from the DIaMonD study testing CGM (n=179) vs. SMBG (n=128) in MDIs, the first pooled analysis following type 1 results at ADA 2016 (later published in JAMA) and type 2 data at ATTD 2017. The combined outcomes were very consistent with the by-group data: from a pooled baseline A1c of 8.6%, CGM drove a 0.9% reduction in A1c at 24 weeks vs. 0.4% in the SMBG group (adjusted mean difference: 0.5%; p<0.001). Consistent with the by-group results, age, numeracy, and education had no impact on the A1c benefit of CGM – a very important finding for expanding CGM to broader populations, especially seniors. Time-in-range metrics strongly favored CGM – at 12/24 weeks (pooled), CGM users were spending 72 more minutes per day in range (70-180 mg/dl) vs. 9 fewer minutes in the SMBG group (p<0.001). This improvement came almost entirely from less time spent over 180 mg/dl: -50 minutes per day in the CGM group vs. -13 minutes per day in the SMBG group (p=0.003). Time in hypoglycemia (<70 mg/dl) also favored the CGM group, but just barely missed statistical significance: -8 minutes per day vs. +3 minutes per day (p=0.05) from baseline. Notably, 93% of study participants were using CGM >6 days per week at six months, a sign of (i) strong patient acceptance of the technology; and (ii) excellent screening by investigators to ensure those in the study would actually wear CGM. Dr. Price said the type 2 results are pending publication, and we wonder if they could drive more reimbursement of CGM in the insulin-using type 2 population. Taken with the results of Abbott’s IMPACT and REPLACE, DIaMonD definitely shows the value of CGM in the type 1 AND type 2 populations, though there is definitely benefit to be gained once decision support is paired with glucose data.

  • As expected, the impact of CGM rose with baseline A1c – those starting DIaMonD at a baseline of >7.5% saw a 0.9% improvement in A1c, while those starting at >9.0% saw a 1.4% improvement. Again, this is a good sign that those not in good control stand to have greater A1c reduction benefits from CGM.
  • Interestingly, only 16% of the CGM group achieved an A1c <7.0% at 24 weeks, lower than we would have guessed but still much higher than 9% in usual care (p=0.05). Patients were obviously starting at quite a high baseline A1c in this study, so perhaps a greater percentage was unrealistic with CGM alone. We expect further behavioral feedback and insulin dosing decision support will help close the gap further.
  • Patients in DIaMonD were not a young, tech savvy, “early adopter” population: those in the CGM group had a mean A1c of 8.6%, a mean age of 52 years, a median diabetes duration of 17 years, a mean SMBG frequency of just 3.6/day, a mean BMI of 31 kg/m2, 5% had >1 severe hypoglycemia episode in the last year, and 52% had less than bachelor’s degree.

5. DIaMonD Study Analysis: Type 2s Find CGM as Satisfying, Engaging, and Valuable as Type 1s

A satisfaction analysis from the DIaMonD study reached a very important conclusion: contrary to common clinical belief, patients with type 2 on MDI are likely to find CGM at least as engaging and valuable as patients with type 1 diabetes. The study gathered results from the 44-item CGM Satisfaction Scale that was administered at the completion of DIaMonD (24 weeks) in 102 adults with type 1 and 76 adults with type 2. Notably, CGM satisfaction was high in both groups, with no significant differences between the type 2 (4.31) and type 1 groups (4.26) – we emphasize continued design improvement over time as contributing to this result. Deeper analysis of the satisfaction subscales demonstrated that mean perceived benefits were significantly higher (p <0.05) among participants with type 2 vs. type 1 (4.39 vs. 4.24), though no significant differences in mean perceived hassles were observed (1.76 vs. 1.71). In the individual questionnaire, type 2 patients reported greater satisfaction on those items focused on acquiring new knowledge or skills (e.g., “CGM taught me new things about diabetes that I didn’t know before”). Ultimately, we love this focus on the type 2 patient experience of wearing CGM, especially because it runs so counter to common beliefs: “Type 2s just won’t wear CGM ... type 2s are unengaged, type 2s don’t need real-time CGM, etc.” Three cheers for providing everyone with diabetes the opportunity to wear real-time sensors and learn things about their own physiology, food choices, and behavior! Here’s hoping these results spawn more studies that counter the conventional views.

DIAMOND STUDY ANALYSIS: Mean Scale Score

6. FreeStyle Libre Vs. G5 in Head-to-Head Accuracy Comparison

A poster from Dr. Guido Freckmann and colleagues examined the accuracy of the Dexcom G5 CGM and real-time FreeStyle Libre in a head-to-head comparison of the two EU sensors approved for non-adjunctive use. The small study assigned 20 patients to wear two of each sensor in parallel for 14 days (G5 sensors were replaced after seven days). Patients returned to the center three times (48 hours per visit), during which up to 115 comparison measurements with a BGM were performed. Results indicated that combined MAD/MARD was 9.6 for the G5 and 11.0 for Libre, while individual sensor MAD/MARD values ranged from 4.8 to 21.6 for the G5 and from 6.9 to 37.2 for Libre. The distribution of individual sensor results is shown below – showing good clustering around the mean, but definitely some outliers. Dexcom had more sensors with MAD/MARD under 10, while Abbott’s concentrated around 8-12. While the study is small and the accuracy comparator was BGM, it’s highly notable to see head-to-head results, especially because they are in line with what the companies themselves have reported. Significant outlier values (MAD/MARD values in >20) occurred with both sensors, though not too often. We hope more studies like this are done to illustrate real world and comparative performance of different systems. Ultimately, we think both systems are safe for insulin dosing on aggregate, though obviously some sensors still see outlier values. The key point, however, is to remember the paucity of data most non-CGM (SMBG) users are using every day to titrate insulin – just a few sporadic measurements.

FREESTYLE LIBRE VS. G5: MAD/MARD Distribution

7. SMBG Study: Structured Testing Leads to 1%+ Decline in A1c in Type 2s Not on Insulin Therapy

Swansea’s Dr. David Owens presented rather positive data from the SMBG study, an RCT of SMBG in non-insulin treated type 2 diabetes. Random assignment to a group that performed structured SMBG resulted in an early and sustained A1c drop vs. assignment to a group that did not test. Following screening and basic diabetes education, ~450 patients were assigned to either (i) No SMBG; (ii) SMBG alone; or (iii) SMBG + telecare. Both SMBG groups received structured SMBG education in an additional visit. Every three months, A1c, cholesterol, weight, patient-reported outcomes, and (when applicable) 7-point SMBG profiles were collected – A1c at 12 months was the primary outcome. At one year, the no SMBG group experienced a 0.3% decline in A1c (baseline 8.6%), while the SMBG and SMBG+telecare groups had whopping declines of 1.1% (baseline 8.5%) and 1.3% (baseline 8.6%), respectively. We suspect (but have no way to know) that time in ranges would have been better for the SMBG and SMBG+telecare groups. Impressively, these curves began to diverge as early as three months into the study, when the no SMBG group had only declined 0.1% and the SMBG groups had both dropped ~0.6%-0.7%. At the end of the 12-month period, there were nearly three times as many patients at goal (A1c=7.0%) in both of the SMBG groups vs. the no SMBG group (p<0.001). Notably, there was no difference in A1c reduction between the SMBG alone and SMBG+telecare groups, suggesting that remote provider contact in this study was non-inferior to clinic visits. These findings bode positively for the debate over SMBG in non-insulin users, and fit well into the context of existing literature – the STeP study, the oft-cited Diabetes Care study from the T1D Exchange (Miller et al. 2013), etc. Even though the patients in this study didn’t take insulin, checking BG clearly caused A1c to drop meaningfully, presumably from behavioral modifications resulting from the structured testing approach. We wonder what would happen if this huge trial was run with CGM – SMBG without CGM, blinded CGM, real-time CGM? 

8. Clinician Prescribing of CGM: 20% are “Ready,” 41% are “Cautious,” and 39% are “Not Yet”

Stanford’s Dr. Molly Tanenbaum presented T1D Exchange data showing that surveyed type 1 diabetes clinicians (n=209) can be divided into three distinct profiles with respect to supporting patients with CGM: 20% are “Ready,” 41% are “Cautious,” and 39% fall in the “Not Yet” category. The profile determinations were based on self-reported attitudes toward technology, time in clinic to review data, ability to keep up with technology, perceived patient barriers to using CGM, as well as provider and practice characteristics. Clinicians who fall in the “Not Yet” bin have the hardest time keeping up with new technology, the lowest proportion of patients with type 1 diabetes, and generally have inadequate clinic time to review CGM data. “Ready” clinicians – which make up a greater proportion of surveyed clinicians than CGM penetration metrics might predict – have the easiest time keeping up with new technology, the highest proportion of patients with type 1 diabetes, and have more time to review CGM data in clinic. Finally, the “Cautious” group generally had positive attitudes toward CGM, but falls in the middle with regards to ability to keep up with new technology, number of type 1 patients, etc. Interestingly, this group was the most likely to endorse barriers to patient uptake – cost, data overload, on-body burden – they consider CGM to be net beneficial, but they are hesitant to change current therapy. Segmenting providers in this way is necessary, Dr. Tanenbaum said, if we wish to support and make them feel comfortable maintaining or shifting to positive attitudes toward CGM. Dr. Tanenbaum suggested that “Ready” clinicians should be encouraged to think about what’s working well, and what will change if uptake increases; the “Cautious” group should assess whether patient and provider barriers align and feel comfortable with education and coaching techniques; and for “Not Yet” clinicians, Dr. Tanenbaum suggested systemic change – if possible, healthcare system barriers should be addressed to enable increased acceptance. She reminded the audience that “physicians aren’t the reason for low uptake of CGM, but they can be part of the solution” – the fact is that the current model of prescribing and paying for CGM still demands additional effort from clinicians (see below image). We absolutely love this work – it’s easy to fall into the trap of assuming that the market will respond as soon as a product becomes available or the tech improves, but changing clinical practice takes a long time, and providers must see a clear cost-benefit win to making the prescribing leap.

Clinician Roles in CGM Use

  • Dr. Tanenbaum bookended her presentation with a theoretical case study to illustrate how crucial it is to get patients and providers on the same page. At first, “Megan” likes the accuracy of CGM A, but her doctor wants her to wear CGM B. Because the two aren’t aligned, Megan doesn’t wear a CGM. At the end of the presentation, the doctor has agreed to learn how to access and review data for Megan’s preferred CGM, and Megan is wearing her CGM regularly, is less concerned about hypoglycemia, and has seen a reduction in A1c. This is obviously a simplified case, but it shows how provider attitude and initiative can have a direct impact on the patient’s wellbeing.
  • In November, Dr. Tanenbaum and colleagues published an important paper in Diabetes Care, an online survey of T1D Exchange adults (n=1,503) investigating barriers to pump/CGM uptake and profiling device users versus nonusers. The Stanford team is taking a multi-pronged attack to address adoption of diabetes devices, focusing on reducing barriers for patients and providers alike.

9. New Preliminary Data from the Recently “Completed” IDCL Training Phase: Mean 145 mg/dl, 77% in 70-180 mg/dl, 1% <70 mg/dl

UVA’s Dr. Boris Kovatchev shared new preliminary data from the now “complete” training phase (n=20) of the NIH-Funded International Diabetes Closed-Loop (iDCL) Trial that continues to show promising time-in-range and hypoglycemia numbers in a small subset of patients. Dr. Kovatchev shared data from 11 individuals in which the training protocol achieved impressive glucose control consistent with past results shared at ATTD: overall mean glucose of 145 mg/dl (150 mg/dl in the “first half of the night”; 125 mg/dl in the “second half of the night”), just 1% of the time <70 mg/dl, and 77% of the time in 70-180 mg/dl (75% in the “first half of the night”; 91% in the “second half of the night”). We have not seen this cut of the overnight iDCL data before, though the improvement as patients get further into sleep is consistent with this algorithm and awesome for the user experience. Notably, Dr. Kovatchev also highlighted the very notable win that patients saw no readings <60 mg/dl or >300 mg/dl, evidence that the system is mitigating the most dangerous of highs and lows. We hope this data will also be borne out by the main phase of the trial (CT.gov posting here), which Dr. Kovatchev confirmed will begin “soon” (no specific timing update, though Tandem said the same yesterday). As a reminder, the main study will serve as Tandem/TypeZero’s hybrid closed loop pivotal (still not yet open for recruitment, according to ClinicalTrials.gov) and will randomize 240 patients in a 2:1 ratio of closed-loop control vs. sensor-augmented pump therapy over six months.

iDCL Training Study

 

Other Highlights

1. ADA Pathway to Stop Diabetes Symposium Showcases Promising new Research

As always, the ADA Pathway to Stop Diabetes symposium offered an inspiring glimpse at the future of diabetes management. We were particularly excited to hear from Dr. Zhen Gu (University of North Carolina, Chapel Hill, NC) on the variety of smart insulin delivery approaches his lab is researching. Echoing his talk at the JDRF Mission Summit in January, Dr. Gu discussed the merits of hypoxia-sensitive vs. pH-sensitive delivery systems and his early forays into the possibility of using red blood cells for smart insulin delivery. In his view, the main challenges facing the smart insulin field today are (i) achieving a fast response comparable to normal beta cell activity; (ii) avoiding hypoglycemia; (iii) achieving ease of administration; and (iv) ensuring biocompatibility. Other highlights from the symposium included:

  • Dr. Daniel Ceradini (NYU Langone Medical Center, New York, NY) presented data showing that that siRNA-mediated Keap1 knockdown resulted in significantly accelerated diabetic wound closure in a humanized mouse model. The approach reduced wound burden by 60% and could potentially allow for improvements in quality of life and significant reductions in healthcare costs. We were very excited to see this research given this is an enormously expensive complication and one that is addressable.
  • Dr. Joshua Thaler (University of Washington, Seattle, WA) discussed his research on how astrocyte and microglial inflammatory signaling promotes obesity susceptibility in high fat diet-fed mice.
  • Dr. Praveen Sethupathy (Cornell University, Ithaca, NY) discussed recent findings demonstrating that a high-fat diet alters small intestinal morphology and physiology. His lab found that 20 weeks on this diet led to a significant increase in epithelial proliferation and a disproportionate allocation of stem cells and progenitors. In a promising proof of concept in fruit flies, he found that overexpression of a particular mRNA in stem cell compartments can almost completely suppress this proliferative response.
  • Dr. Kathleen Page (USC, Los Angeles, CA) summarized recent findings implicating central adiposity in childhood as an early indicator of intrauterine exposure to gestational diabetes and maternal adiposity. We find the subject of epigenetics extremely interesting and hope to see more on this as an approach to reducing obesity longer-term.
  • Dr. Phillip White (Duke Molecular Physiology Institute, Durham, NC) presented findings demonstrating that inhibition of the enzyme BDK improves glucose tolerance and lowers hepatic triglyceride content. BDK is an enzyme that regulates branched-chain amino acid metabolism along with PPM1K; its beneficial effects appear to be exerted through regulation of ATP-citrate lyase.

2. Ms. Davida Kruger Named Outstanding Educator in Diabetes

Followed by a standing ovation, diabetes educator extraordinaire Ms. Davida Kruger (Henry Ford Medical Group, Detroit, MI) delivered the Outstanding Educator in Diabetes Award Lecture, as she shared valuable insights on her role in the DCCT, as a nurse practitioner in diabetes, and as an ADA volunteer throughout her career. Our team joins the diabetes world in our deep gratitude and admiration for Ms. Kruger’s leadership and service and example to the field – please see our Detailed Discussion and Commentary section below for Ms. Kruger’s wisdom on the evolution of clinical research, the need for nurse practitioners in diabetes, and patient-centered care.

Detailed Discussion and Commentary

Oral Presentations: GLP-1s and SGLT2s—To Do or Not to Do in Type 1 Diabetes Mellitus?

Efficacy and Safety of Liraglutide in Insulin-Pump-Treated People with Type 1 Diabetes: The Lira Pump Trial

Thomas Dejgaard, MD, PhD, Kirsten Nørgaard, MD, and Christian Frandsen, MD, PhD (Copenhagen University Hospital, Hvidovre, Denmark)

A team of researchers from Copenhagen University Hospital (Hvidovre, Denmark) presented the results of the highly-anticipated Lira Pump Trial demonstrating the clinical potential of the GLP-1 agonist liraglutide as an addition to insulin pump treatment in people with type 1 diabetes. People with type 1 diabetes using CSII and with overweight (n=44) were randomized to receive liraglutide 1.8 mg or placebo in addition to regular insulin pump therapy for 26 weeks. Dr. Thomas Dejgaard outlined the positive primary outcomes: compared to placebo, liraglutide as an add-on to insulin pump treatment produced significant improvements in A1c (-0.6% vs. +0.2%; p<0.001) and weight loss (-7.3 kg [16.1 lbs] vs. -0.6 kg [1.3 lbs]; p<0.001) without increasing the time spent in hypoglycemia (3.2 vs. 3.5 events; p=0.77). While the A1c reduction could be characterized as modest for a GLP-1 agonist, we’re especially impressed by the reduction from a presumably low baseline A1c combined with the substantial weight loss. As expected for a GLP-1 agonist, nausea was the most commonly reported adverse event for participants randomized to the liraglutide treatment arm; 64% of participants experienced this, though it was transient and disappeared over the first few weeks of the study (we’d love to better understand the experience and intensity of nausea since this percentage isn’t that helpful overall). Currently approved for type 2 diabetes (under the trade name Victoza) and obesity (under the trade name Saxenda), these results underscore the potential of liraglutide for type 1 diabetes as well.

  • Dr. Kirsten Nørgaard followed with a deeper dive into the impact of add-on liraglutide to insulin dosing and time-in-range. Excitingly, liraglutide as an add-on to insulin pump therapy gave participants significantly more time spent in the normoglycemic range (57% vs. 45%) – a crucial outcome for patient quality of life and perceived therapeutic and personal success. That said, there was no statistically significant difference in time spent in hypoglycemia (defined as <3.9 mmol/l [70 mg/dl]). Very surprisingly (shockingly), liraglutide also produced no change in total daily dose of insulin; it is less surprising that the distribution between basal and bolus insulin also wasn’t changed.
  • Dr. Christian Frandsen closed with a closer look at the effect of liraglutide on weight loss. In addition to producing significant reductions in the primary outcome of body weight, liraglutide also produced reductions in a number of more specific measures of body composition, including total fat, android fat, gynoid fat, and total lean mass. The mechanism of this effect is unclear but likely is not related to food preferences. As revealed by food preference surveys given throughout the study, liraglutide did not appreciably impact food preferences.

Questions and Answers

Q: Your group has already presented nice studies in obese and non-obese people with type 1 diabetes on MDI. What is different about this CSII population that the outcomes are so much better?

A: The A1c reduction in liraglutide treated patients were exactly the same (0.6%) between these trials. The major difference is that here we didn’t see an appreciable A1c change in the patients in the placebo arm, whereas we saw a -0.4% A1c reduction in the placebo arm in the other studies, leaving no significant difference between the treatment and control groups.

Q: Why the choice of a 1.8 mg dose of liraglutide? You showed very impressive weight loss – is this because people with type 1 are more sensitive to the drug? Did any participants need a reduction in dose?

A: I don’t think it’s true that people with type 1 diabetes are more sensitive to liraglutide. A few patients had to reduce to a 1.2 mg dose to minimize nausea side effects.

Q: Do you think these improvements you’re seeing are due to weight loss?

A: A large part of it could be the weight loss, that's possible.

Posters

Effect of Ertugliflozin on Glycemic Control, Body Weight, Blood Pressure, and Bone Mineral Density in Type 2 Diabetes Mellitus Inadequately Controlled with Metformin Monotherapy: VERTIS MET Trial

J Rosenstock, J Frias, D Páil, B Charbonnel, R Pascu, D Saur, A Darekar, S Huyck, H Shi, B Lauring, S Terra

Results from the VERTIS MET trial demonstrated that combination therapy of ertugliflozin and sitagliptin significantly improved glycemic control and reduced body weight and systolic blood pressure over a 26-week treatment period in people with type 2 diabetes inadequately controlled on metformin monotherapy. The double-blind study randomized 291 adults with type 2 diabetes to ertugliflozin 5 mg/sitagliptin 100 mg (E5/S100), ertugliflozin 15 mg/sitagliptin 100 mg (E15/S100), or placebo. Results showed that both E5/S100 and E15/S100 provided significant A1c reductions compared to placebo, with reductions of 1.6% (baseline of 8.9%) and 1.7% (baseline of 9%), respectively, vs. 0.44% with placebo (baseline of 9%). In addition, the odds of achieving an A1c <7% at week 26 were significantly higher in the ertugliflozin groups vs. placebo: The proportion of patients with A1c <7% was 36%, 31%, and 8.3% in the E5/S100, E15/S100, and placebo groups, respectively. Regarding other analyses, the ertugliflozin groups also demonstrated greater reductions in body weight and blood pressures compared to placebo. E5/S100 and E15/S100 achieved weight losses of 2.9 kg (baseline of 91 kg) and 3 kg (baseline of 91 kg), respectively, compared to placebo’s weight loss of 0.9 kg (baseline of 95 kg). Additionally, E5/S100 and E15/S100 demonstrated systolic blood pressure reductions of 2.0 mmHg (baseline of 130.7 mmHg) and 4.0 mmHg (baseline of 129.2 mmHg), respectively, vs. placebo’s systolic blood pressure increase of 2.4 mmHg (baseline of 127.4 mmHg). Regarding adverse events, ertugliflozin was generally well-tolerated, though the drug was associated with a higher incidence of genital mycotic infections compared with placebo – an unsurprising finding for an agent in the SGLT-2 inhibitor class.

  • The E5/S100 and E15/S100 groups demonstrated temporary, modest reductions in mean eGFR from baseline. eGFR values in both study arms demonstrated drops up to ~4 and ~2.5 ml/min/1.73 m2, though mean eGFR returned to or near baseline by week 26. This slight decrease in eGFR immediately following new initiation of SGLT-2 inhibitor therapy has also been noted for other agents in the class, and again, isn’t particularly surprising and doesn’t seem to give researchers much cause for concern.

Oral Presentations: Where Is Glucose Monitoring Taking Us?

Sugar.IQ Insights: An Innovative Personalized Machine-Learning Model For Diabetes Management

Huzefa Neemuchwala, PhD, MBA (Head of Innovation, Medtronic Diabetes, Northridge, CA)

Medtronic’s very smart Head of Innovation Dr. Huzefa Neemuchwala shared the first data from the “limited learning launch” phase of the Sugar.IQ app with Watson – now tag-lined, “Intelligent Diabetes Assistant App.” Results came from de-identified CareLink data in 81 users of the Sugar.IQ app using MiniMed 530G/Enlite + MiniMed Connect to send CGM data to the app. Relative to baseline metrics (one month prior), this small group of Sugar.IQ users has experienced a solid 37-minute/day improvement in time-in-range (p=0.04; baseline not shared), an 11% reduction in sustained hypoglycemia (>120 minutes; p<0.001), and an 8% drop in sustained hyperglycemia (>120 minutes; p<0.001). Within three days of the app delivering a pattern “insight,” 65% of users have experienced fewer lows and 55% experienced fewer highs. In total, Sugar.IQ has now been used by 97 people for an average of two weeks each, and engagement has been encouraging in this limited launch: an average of 1.5 unique app sessions per day, 78% of users logging food, and 4.8 logged food items per day – persistence over time, particularly with food and opening the app up, will be THE key question ahead. The very cool Glycemic Assist feature, allowing users to “follow” a particular food item over time (we love this!), has been popular: 1,886 views in these 97 users so far. Users have “followed” their glycemic response to Dunkin Donuts, Panera Bread, corn flakes, ice cream, buttermilk biscuits, etc. – pretty squarely in the junk food (“Diabetes Landmines”) category, but hopefully the app will gradually nudge people away from eating them! We include examples below of the insights Sugar.IQ delivers – so far, this group of users has received 1,119 different contextual, personalized insights ranging from glycemic control and behavior to hyper- and hypoglycemia to rapid rate-of-change to boluses. Users have “liked” a notable 89% of the insights, indicating they are finding useful patterns. Dr. Neemuchwala also shared two Sugar.IQ case studies from patients with longstanding diabetes – the app identified trends (over-correcting highs, eating a high-carb lunch) and nudged them to change their behavior (the latter is a phrase Dr. Neemuchwala emphasized). More details and screenshots below!

Usage Patterns

  • It’s unclear when Sugar.IQ will launch fully, but we assume the biggest gating factor is approval of the standalone Guardian Connect mobile CGM (under FDA review and currently in human factors testing). Guardian Connect will stream CGM data to Sugar.IQ directly via Bluetooth, and should help Medtronic differentiate its standalone mobile CGM offering from rising competition (Abbott, Dexcom). Per Medtronic’s JPM presentation (the last update), a full launch of Sugar.IQ and Guardian Connect were expected in May-October, though today’s Medtronic Diabetes Analyst Day update said that human factors work is ongoing. Medtronic also told us it is working with IBM to finalize algorithm for the Sugar.IQ commercial launch.
    • As a reminder, this app has been fairly delayed. The plan as of last year was to launch Sugar.IQ by the end of 2016, timing that was updated at JPM. Sugar.IQ was demoed and “beta launched” in September at Health 2.0 – presumably “the beta launch” was a previous group, and this data is from the wider release that was alluded to at ATTD.
  • Sugar.IQ insight examples: In a word, wow! “Planning your day? I see you tend to go low on Saturday between 12 PM and 3 PM.” “I notice that you tend to go low after meals with >20g of protein.” “Great! I noticed that you had only 1 nighttime low(s) in the last month. Whatever you’re doing, seems to be working very well.” “I see that between 6AM and 9AM, your glucose often goes high (300+ mg/dl) after taking an insulin injection.” “After you glucose is high for more than 120 minutes, you then tend to go low.”
  • Looking ahead, Medtronic has also expanded the research on the hypoglycemia prediction feature, which now has >90% accuracy at predicting hypoglycemia within a 2-4 hour window (80%+sensitivity, 67% positive alert rate). This feature is now using 100+ behavioral models based on unsupervised clustering techniques. Last we heard, this will be included in a future version of the app, but not the one at launch.
  • Dr. Neemuchwala provided two case studies of Sugar.IQ noticing a specific glycemic/insulin/food pattern, giving the user an objective insight, and a resulting human behavior change. “Simple judgment-free nudges can lead to sustained behavior improvement.” Both were in people with long-standing diabetes (one with type 2 for 20 years on insulin, another with . One case concerned over-correcting highs, while another concerned eating a high-carb vs. slightly lower-carb lunch. See the slides below!

Insight Can Enable Discovery &amp; Behavior Change

Insights Can Improve Meal Behavior and Time in Range

  • Sugar.IQ uses machine learning to find patterns in diabetes data, and ultimately, hopes to combine many data sources: CGM and insulin, biometrics, meals/logbook, CRM, medical and claims, mood, sleep, location. The focus is on putting all this data in one place, then driving insights and predictions.

The Right Insight at the Right Time

Therapy, Routines, Lifestyle

Glycemic Assist

Special Lectures and Addresses: President, Health Care & Education Address and Outstanding Educator in Diabetes Award Lecture

A Toast to Our History

Davida Kruger (Henry Ford Medical Group, Detroit, MI)

Upon receiving the Outstanding Educator in Diabetes Award, a humbled and passionate Ms. Davida Kruger shared her work in diabetes over the past 35 years, as she provided insights on her role in the DCCT, as a nurse practitioner in diabetes, and as an ADA volunteer. In an emotional opening, Ms. Kruger shared her personal reasons for working in diabetes, as she touched on the diabetes diagnoses and complications she watched her mother and other family members experience. Moving into the beginnings of her career, she discussed her role in the DCCT, emphasizing the development and evolution of the role of trial coordinators during this time and the emergence of different team member roles throughout the process. Notably, Ms. Kruger continued by stressing the increasingly important role of the nurse practitioner, as diabetes prevalence increases and the clinical endocrinologist workforce declines. Along these lines, she presented data showing that nurse practitioners make equally accurate diagnoses as physicians and are also more likely than physicians to provide health education and promotion, adapt medical regimens to the patient’s preferences, and listen more to the patient. Lastly, Ms. Kruger touched on her work as an ADA volunteer, specifically pointing to the importance of fundraising for the Association’s research foundation. In conclusion, Ms. Kruger expressed appreciation for her family, mentors, and colleagues, as she closed by stating that the field “must remain open to all team members” and keep the patient at the “head of the team.” Ms. Kruger left the podium with a standing ovation and our team joins the diabetes world in our deepest gratitude and admiration for Ms. Kruger’s incredible leadership and service to the field.

Exhibit Hall

Abbott

Abbott’s trademark yellow booth highlighted its blinded, retrospective FreeStyle Libre Pro. The product launched in the US last fall, meaning that this was the first ADA where Libre Pro could be demoed … and, boy, were attendees taking advantage! This was one of Abbott’s bigger booths in recent memory, and still, it was PACKED. Clearly, the high enthusiasm we saw earlier this year at ENDO has yet to abate. As expected, no update was shared on the US launch of FreeStyle Libre consumer, which remains under FDA review for both an adjunctive claim and a non-adjunctive (replacement) claim following submission in 3Q16. (The last timing update in January suggested a 2H17 US launch, though FDA is hard to predict.) On the global front, reps highlighted the recent acquisition of national reimbursement for FreeStyle Libre in France, sharing that the company ended up fielding “thousands of calls” on the very first day of the announcement. Abbott also confirmed it currently markets FreeStyle Libre consumer in 35 countries (previously “30+” as of a few weeks ago), with more than 300,000 global users.

Amgen

Amgen in a big booth is new for ADA. Amgen’s booth was relatively small and placed in the corner of the exhibit hall, with all of its attention dedicated to its PCSK9 inhibitor Repatha (evolocumab). In its red/white/blue color scheme, Amgen promoted Repatha’s unique dosing option, as wording included “maximize LDL-C reduction in one move” and “the first and only PCSK9 inhibitor with a single monthly injection.” Other promotional material focused on the drug’s “two dosing options” and how it doesn’t require titration (“no titration necessary”).  Additionally, the company seemed to focus in on the drug’s earlier challenges with determining coverage, as touch screens highlighted resources including in-office insurance support and emphasized that “91% of adults are covered” (Repatha is on the formulary of 91% of adults with prescription drug coverage). The booth also featured videos on Repatha’s clinical data and safety information, a medical info section, and cookies for attendees (we wish they had offered something a bit healthier!). Please see our Amgen 1Q17 update for more on the company’s latest.

Ascensia

Ascensia boasted a sizable booth near the rear of the exhibit hall. The company has come to ADA on the heels of a couple big partnership announcements (Insulet, Voluntis), which were both featured on handouts within the booth. Reps wouldn’t divulge any new details on either announcement, but were more than happy to chat at length about the Bluetooth-enabled Contour Next One meter and Contour app. These two products were far and away the focus of the booth, with signage prominently highlighting the ability of the Next One to light up to indicate in-range (green), low (red), and high (yellow) results. Just like at ATTD, the motif of the booth played on this association with “illumination” using slogans that read: “Seeing how everyday activities affect blood glucose. That’s illuminating”; “Their diabetes. Illuminated”. Notably, the company also exhibited the Contour Next Link 2.4 meter (though to a lesser extent vs. the Next One), which is available with Medtronic’s MiniMed 640G internationally and in the US with the 630G and upcoming 670G.

AstraZeneca

AstraZeneca’s booth was dominated by SGLT-2 inhibitor Farxiga (dapagliflozin), with only a small corner dedicated to the GLP-1 agonist Bydureon (exenatide once-weekly) pen. This emphasis is consistent with the company’s positioning of Farxiga as a cornerstone of its new Cardiovascular and Metabolic Disease unit in its 1Q17 update. The new company structure was also evident in the significant real estate devoted to cardiovascular drug Brilinta (ticagrelor); it was clear from this booth and others that the line between diabetes and cardiovascular drugs is becoming increasing blurred. AZ’s booth also included a fairly substantial research section with handouts on ongoing clinical trials, a video demonstration of the Bydureon autoinjector (currently under FDA review and a poster on the autoinjector will be presented at ADA on Sunday), and a very realistic dummy patient that apparently gave one of the security guards quite a scare this morning!

BD

We were hoping to see the MiniMed Pro infusion set with BD FlowSmart technology back on display in the exhibit hall after a hiatus relating to a halted launch, but no dice. A rep told us that BD is still “figuring it out with Medtronic … it’ll be another slow launch, we want it to go well,” confirming what we heard at AACE a little over a month ago – we agree that proceeding with caution in order to optimize patient experience is absolutely the right move. See our test drive of the set – we still believe that it has tremendous potential once the training process is optimized and manufacturing works out the kinks we experienced. Elsewhere in the booth, BD was, as one rep put it, “preaching what we’ve always preached: short needle length.” An adjacent (unbranded) BD booth called “Type 2.0 Lab” invited participants to select their top three most important features for a device to have – at the top of the leaderboard when we walked through were: flexible dosing (adjustable basal/bolus dosing), dose capture, and reduced number of injections through pump therapy. We’d note that BD is already developing products in all three of these areas!

Dexcom

Dexcom’s welcoming booth proudly showed off the just-approved and quickly launched Android version of G5 – already available on the Google Play store (16 reviews, 4 stars) following the approval announcement earlier this week. We also got to hold and play with the very rugged new touchscreen G5 receiver for the first time following FDA approval in March. We would describe the new receiver as “built like a tank” – it’s bigger than the current option, and the touchscreen is more resistive than one found on a consumer-grade phone. This thing looks like it could be thrown at the wall and not break, fixing the reliability issues that have challenged Dexcom’s current receiver. On the other hand, this receiver definitely feels like much more of a medical device than the current (kind of cool) iPod-nano like receiver, a surprising move for Dexcom in our view. The receiver’s user experience is now in line with the G5 app, though using the device next to the iPhone or Android apps makes the user experience choice pretty obvious – the smartphone version of G5 is miles better than the new receiver. The phone app obviously benefits Dexcom and users too, as Android/iPhone bring continuous data upload to Clarity/Dexcom’s cloud, continuous over-the-air app updates, integration with other apps, etc. Plus, with direct Apple Watch-to-G5 transmitter communication coming (no phone needed; see our coverage from this week), there will be another phone-agnostic option for viewing Dexcom data

Dexcom

  • The booth also showed the major new additions to Dexcom Clarity, most notably adding an AGP report and the (new to us) massively improved Dexcom Clarity mobile app. See the screenshots below from the new Clarity mobile app, a spectacular improvement over the previous basic PDF-download version. It now feels like a real app! We love that time-in-range is so prominent and all reports are accessible, including AGP – indeed, the standard of care continues to evolve for patients and we saw this as a major win. A rep told us that Dexcom has removed the “estimated A1c” metric from Clarity at the request of the FDA – apparently the discrepancy between CGM-derived estimated A1c and lab values was confusing, and all companies have reportedly been asked to examine this metric. Clarity also recently added a more detailed day-by-day and hour-by-hour statistics recap, which we deem incredibly valuable for patients.

Dexcom data Dexcom data

  • Otherwise, Dexcom marketed several handouts on non-adjunctive use of CGM, Medicare coverage of CGM (“now available” – though the ongoing talks with CMS about local coverage decisions were not mentioned), and shared printouts of the DIaMonD study testing CGM in MDI.

Insulet

Insulet’s busy booth showed off the new Bluetooth-enabled OmniPod Dash PDM for the first time. Reps gladly took us through the PDM’s brand new user interface on the locked down, consumer-grade Android phone (see picture below) – and boy is it a major improvement in form factor and user experience over the current PDM, and a definite upgrade over the “prototype” from ADA two years ago. We’re glad to see IOB prominently displayed on the home screen, clear display of last blood glucose and last bolus, easy access to take a bolus (center button on the bottom), a simple swipe to display the current basal profile graphically and key pod status, a nicely laid out bolus calculator (click to see all the math in detail), and great use of a side-bar menu to put less-used functions and settings. The off-the-shelf Android phone is very slim and retains the touchscreen quality and consumer-grade hardware that we’ve all come to expect out of modern devices. This does not feel like a medical device – that is for sure! After the demo, booth-goers filled out a market research survey soliciting opinions and feedback. We noticed several questions on expectations for battery life, and it will be interesting to see what longevity Insulet comes to market with – as a locked down Android phone not running cellular (micro USB charge), we assume it would last multiple days on a single charge (this is our speculation). As a reminder, an FDA submission for this product is expected in 2H17, enabling a possible late 2017/early 2018 launch. Along the perimeter of the booth, attendees could also demo the pod and get a free copy of our own Adam Brown’s bestselling diabetes guide, Bright Spots & Landmines.

  • Ascensia’s Bluetooth-enabled Contour Next BGM was situated next to each Dash PDM – we didn’t get to demo the integration, but it was great to see this on display so soon after the integration announcement just two days ago! Users can also enter a BG manually using an on-screen scroll wheel (like the One Drop app) or via typing the number.

Insulet app

  • Other UI/UX features in Dash: It’s easier to view and change basal segments, a feature placed quite prominently in the app’s user flow (one swipe from the main home screen). (Arguably this does not warrant being so upfront, given how infrequently patients and HCPs are changing basal profiles. We agree that Dash has made it much easier to do so!) The food library has also been revamped and expanded, and Dash has a nice “Shopping Cart”-like feature to add multiple foods into one meal. Users can also see a progress bar as a bolus is delivered and cancel it while in progress. We did notice the phone has a camera, which in a future version of the UI/UX could possibly enable pictures of meals overlaid with CGM and insulin data.

Intarcia

Intarcia had a smaller presence at the exhibit hall this year, with only a small medical information booth, presumably due to the fact that it’s ITCA 650 device has been submitted to the FDA and the company must be very careful about pre-marketing. Notably, Intarcia sponsored its very first corporate symposium at this year’s ADA, and the company is clearly focused on reaching and educating providers about ITCA 650 through new and alternative channels.

J&J – Janssen

The Janssen section occupied approximately two-thirds of J&J’s booth and devoted about equal attention to standalone Invokana (canagliflozin) and Invokamet (canagliflozin/metformin). The booth had attracted a decent amount of foot traffic when we stopped by, with people lining up to sample sugar-free iced vanilla lattes and “donate a photo” (you get your picture taken, J&J donates $1 to the ADA). As at recent conferences, the Invokana materials featured data showing superiority vs. Merck’s Januvia (sitagliptin) and banners trumpeting the product’s status as the #1 prescribed SGLT-2 inhibitor in the US. While Invokana continues to lead the class for now, the trend lines do not look great for J&J: US Invokana sales dropped 17% YOY in 1Q17. We expect that future within-class competition among SGLT-2 inhibitors will depend heavily on the CANVAS results that will be reported on Monday. If the results suggest an SGLT-2 inhibitor class effect on cardiovascular outcomes, Invokana’s first-in-class status should serve it well in the US. On the other hand, if the results are substantially less impressive (or feature a more mixed safety message) than those from EMPA-REG OUTCOME, we would expect Lilly/BI’s Jardiance (empagliflozin) to make significant market share gains.

J&J – OneTouch

Crowds flocked to the OneTouch portion of the J&J booth – fresh green makeover and all – to see the OneTouch Via bolus-only insulin delivery patch device in the flesh for the first time in an exhibit hall. Reps disclosed that the product’s updated manufacturing process (submitted in November) received FDA 510(k) clearance on Wednesday, but did not give any specifics on the launch timing front. We truly appreciated the logic behind the lack of timing updates, despite guidance from ADA 2016 calling for an “early 2017” launch and a December email exchange indicating a 1H17 launch: “We are focused on ensuring a perfect launch for patients. That means getting reimbursement and training down. Our goal is that no patient will start on Via and not continue.” Contrary to what we were told in the December email, the rep indicated that the launch, when it does happen, will not be focused, but “filled to capacity.” Presumably J&J will start with a limited release to learn from early adopters, make sure manufacturing and access is in place, build inventory, etc. We expect strong patient demand for this product. It was great to finally get to hold the device, and it was clear that booth-goers were intrigued – we overheard one who was amazed by the patch’s small form factor, and another who was complaining about the limited basal options of other patch pumps (of course, OneTouch Via doesn’t have basal capabilities – this requires a separate prescription for insulin and pen/syringe delivery device). We believe the supreme discretion offered by the Via will make it a very viable asset for a J&J diabetes business that is currently exploring strategic options, including sale (depressing!).

J&amp;J One Touch

  • The inspired OneTouch reps were also detailing a new and improved version of the OneTouch Reveal app (Apple Store, Google Play) for its Bluetooth-enabled BGMs, complete with a timeline feature, a logbook, and sharing capabilities. The timeline tracks important blood glucose events, calling attention to patterns – it also features weekly content with general diabetes information, though the goal is to make the content more personalized. The logbook makes it easy to visualize blood glucose readings with contextual information. The “share” feature allows users to share their last reading or a report a email or text, or even over the cloud, with a clinician. The upgrade also includes the introduction of physician- and patient-facing web apps.
  • There was no launch timing update on the OneTouch Vibe Plus integration with Dexcom G5 (approved for ages two and up in December), but the company is “moving toward commercialization.” This would be a nice stepping-stone on the way to a hypoglycemia-hyperglycemia minimizer closed loop system.
  • We were glad to see that WellDoc once again represented at the OneTouch booth, marketing the BlueStar/OneTouch Verio Flex BGM integration allowing passive blood glucose data transmission into the type 2 diabetes management software. A WellDoc rep told us that the integration hasn’t quite rolled out to patients yet, but she was over the moon about this partnership as well as others (namely Samsung).

Lilly

The Lilly-only booth at ADA 2017 was essentially an enlarged version of its recent booths at AACE 2017 and ENDO 2017. The front and center of the booth cleared a wide path for caffeine-deprived attendees to make a beeline for the espresso stand – Lilly has quickly achieved a reputation for best espresso in exhibit halls, in our humble (but very expert) opinion. The first line of product information, flanking either side of the espresso bar, featured SGLT-2 inhibitor Jardiance (empagliflozin), biosimilar Basaglar (insulin glargine), and GLP-1 agonist Trulicity. The signage for Jardiance was particularly eye-catching – in addition to touting the new CV indication for Jardiance, Lilly really drove the point home with animations of the words “Only Jardiance” coming together to form a heart shape. Directly, behind the Jardiance signage, Lilly strongly highlighted the rest of the empagliflozin franchise (Synjardy and Glyxambi), debuting the tagline “Meet more of your patients’ needs with the rest of the Jardiance family” for the first time. Once-weekly Synjardy XR (empagliflozin/metformin) won the label “partnered for convenience,” while Glyxambi (empagliflozin/linagliptin) was headlined with “partnered for power.” Additional square footage in the booth was also devoted to the Humulin U500 KwikPen, Humalog U200 KwikPen, and the DPP-4 inhibitor Tradjenta (linagliptin) franchise. We were also pleased to see the continued presence of the Patient Access and Affordability section – which is now newly promoting the new InsideRx partnership, in addition to Lilly’s Blink Health collaboration.

Lilly/BI

Capturing the attendees at the very other end of the exhibit hall, the partnered Lilly/BI booth featured a huge, outer swath of the booth devoted to Jardiance, with the product’s signature teal coloring (a departure from the minimalist white and gray tones of the rest of the booth). Most notably, a giant, teal plastic heart sculpture commanded the attention of passerby. Upon placing a hand on the display in front of the heart, upbeat music and a steady heartbeat sound began to play, while facts about the CV indication for Jardiance flashed on the heart (“Only Jardiance,” and statements highlighting the ADA Standards of Care inclusion, the size of the EMPA-REG OUTCOME trial, and more). According to a Lilly representative, the displayed messages change with each 38th person that interacts with the display – representing the 38% risk reduction for CV death in EMPA-REG OUTCOME that supported the indication.

MannKind

MannKind’s small booth for inhaled mealtime insulin Afrezza was centered on the same playful message we saw at AACE 2017 last month: INsulin, 12-15 minutes; OUTsulin, three hours. The company is clearly focused on the faster onset/offset of Afrezza, which leads to lower hypoglycemia risk and improved patient quality of life, though an “ultra-rapid-acting” label claim for the product is still under FDA review (a decision is expected by September). We do hope Afrezza receives this designation, at which point we imagine MannKind’s marketing around this theme will ramp up. We truly believe in the value of faster onset/offset for prandial insulins, a therapy area that is currently lagging behind patient need, since available options aren’t good enough and still cause far too much hypoglycemia. A blown-up model of the inhalation device was mounted on the booth’s back wall, and conference attendees could read additional information about Afrezza on a video monitor or on printed handouts.

Merck

Merck presented with a relatively large booth near the center of the exhibit hall, with the majority of the marketing focused on its Januvia (sitagliptin) franchise. Most of the booth’s boards promoted Januvia, while a couple promoted its combination product Janumet (sitagliptin/metformin); and approximately one-fifth of the booth’s real estate was dedicated to the company’s vaccine Pneumovax 23 (pneumococcal vaccine polyvalent). As usual, Merck brought a patient-centric atmosphere, with large table touch screens, in which attendees could choose to follow different patients and review their medical histories and labs. Additionally, the boards featured named patients, alongside clinical trial data showing strong FPG and PPG reductions and artistic explanations of the DPP-4 inhibitors’ mechanism of action. In addition to a medical affairs section, the booth had its fair share of interactive activities including the opportunity to create a customized wall chart, a virtual reality station, and a puzzle challenge – not to mention the company’s popular frozen yogurt parfait offerings. To learn about Merck’s latest updates, please see our 1Q17 update.

Medtronic

Medtronic’s bustling, expansive booth was filled along the perimeter with the company’s latest products, including the MiniMed 670G hybrid closed loop (finally – in the US not under glass!) and the Bluetooth-enabled standalone Guardian Connect Mobile CGM available outside the US (under FDA review and currently in human factors testing, per today’s Analyst Day). We got to demo the MiniMed 670G pump for the first time and found several things interesting about the user interface. First, automated basal dosing actually appears as pink dots on the 670G’s trend graph screen and is denoted as a “micro bolus” (e.g., 0.15 units) – we wonder if this terminology discrepancy between a “micro bolus” and “automated basal delivery” could be confusing for some patients. On the other hand, it is semantics and they are the same thing in practice – a basal rate is simply a series of micro-boluses. We were also reminded of the tough balance between transparency (what is the algorithm doing?) and simplicity (keep the under-the-hood nuances to a minimum) – from what we saw, the 670G errs on the side of less clarity about what the system is doing at any given time to keep glucose in range. For instance, if glucose is trending high, the 670G would increase basal insulin delivery, but it’s hard to know as a user how much additional basal insulin the system has delivered to keep blood glucose in range (unless one scrolls to every pink dots and tallies them up). The “active insulin” metric on the 670G home screen also does not include increased basal insulin delivery from hybrid closed loop (it’s only from manual boluses), and there is no CGM “projection” – in our view, this could make it hard to know if additional bolus insulin is needed. We’ll be interested to see how patients feel about this, since some might like any insulin that is delivered over the pre-programmed basal to be considered “active insulin.” (This could be useful for decision making, but could also be confusing and unnecessary for most users.) We’re also not clear on how aggressive the 670G’s auto-basal can be – can it deliver 2x or 3x or 4x the pre-programmed basal? Reps reminded us that the Guardian Sensor 3 recommends 3-4 calibrations per day and has shown an MARD under 10% with that scheme – the only sensor FDA approved to power a hybrid closed loop system. Attendees could examine the sensor through a popular virtual reality experience, though we didn’t have a chance to try it ourselves.  

Novo Nordisk

Novo Nordisk’s bright, modern booth was hard to miss as one of the largest displays in the exhibit hall – with people lining up to have their A1c measured, as is ADA tradition. A bustling social space filled with sleek low-slung white couches and light wood tables, the booth was focused largely on Tresiba (insulin degludec) and Victoza (liraglutide), with a lesser amount of signage promoting the newly-launched combination product Xultophy (insulin degludec/liraglutide). Tresiba materials featured a skydiving/parachute motif with the slogan “a proven A1c descent” – complete with a corresponding virtual reality headset experience (these seem to be popping up everywhere in exhibit halls!) in which visitors could simulate skydiving from an airplane.  Data featured in the booth focused especially on the flexible dosing indication on Tresiba’s label and the fact that it is the only basal insulin pen with a max injection dose up to 160 units – this latter point is clearly part of Novo Nordisk’s efforts to position Tresiba for patients with type 2 diabetes and higher insulin requirements. Victoza materials highlighted the drug’s “3-for-1 benefits” – A1c efficacy, weight loss, and low rate of minor hypoglycemia – and emphasized that it is the #1 prescribed GLP-1 agonist globally (perhaps a nod to the increasing competition in the GLP-1 agonist space). Demo stations were interspersed within the booth where company representatives guided visitors through the use of each product’s corresponding pen. Aside from products, we were thrilled to see a sizable portion of the booth devoted to encouraging visitors to join the company’s annual ADA 5K run, and a sizeable presence from members of Team Novo Nordisk – bikes and all.

A separate, smaller booth was devoted entirely to obesity medication Saxenda (liraglutide 3.0 mg), beckoning providers to “help your patients with obesity get the reductions they need.” The booth featured a photo of a women holding a larger pair of jeans – the classic “before and after” weight loss pose – with the phrases “excess weight,” “high cholesterol,” “large waistline,” and “high blood pressure,” emblazoned on the pants. Notably, the promotional materials referred to Saxenda as “the first and only GLP-1 agonist for chronic weight management” – we got the sense that Novo Nordisk is very much trying to drive home the message that obesity, like diabetes, requires ongoing therapeutic management, unlike the “diet pills” of the past. The presence of an exclusively obesity-focused booth further underscores Novo Nordisk’s fairly recent strategic commitment to increasing R&D in this challenging space.

Sanofi

Sanofi occupied a commanding booth at the center of the bustling exhibit hall, featuring both its diabetes and cardiovascular franchises. Newly-launched Soliqua (insulin glargine/lixisenatide) was the clear focal point (dominating the booth’s main overhead banner) with signage also dedicated to Toujeo (insulin glargine U300) and the PCSK9 inhibitor Praluent (alirocumab). Demarcated by the product’s signature dark blue, purple, and gold color scheme, Soliqua materials emphasized the combination product’s ability to “help patients put high A1c behind them.” Amidst a giant statue of a percentage sign, presumably representing the notion of an A1c goal, Sanofi representatives ran through touchscreen displays of clinical data on Soliqua and how to initiate this in therapy in patients. This messaging was largely centered on how to initiate for people hoping to intensify existing Lantus therapy (“continue current therapy, or start something different?”). These displays were quite popular, indicating a growing interest among physicians in this long-awaited new drug. Toujeo materials focused on a patient vignette of Ray, “a musician and father living with type 2 diabetes,” and applauded Toujeo’s ability to “play to stay in range” and “get patients in tune with their goals.” Amidst touchscreens displaying the clinical data, further signage reminded the booth’s visitors that Toujeo has a “more stable and prolonged action profile than Lantus,” perhaps subtly encouraging physicians to switch their patients from the dominant (and off-patent) Lantus to a regimen involving the similar but more advanced Toujeo. In the Praluent corner stood two giant statues of blue and green arrows, crashing down toward the ground to parallel the drug’s LDL cholesterol-lowering action. Signage here emphasized Praluent’s efficacy and ease of use, posting large statistics boasting the drug’s “more than 60% LDL reduction” and the fact that “up to 80% of patients achieved their goal” when taking Praluent. As at recent conferences (and in stark contrast to previous years), there was hardly a mention of former star product Lantus (insulin glargine). Giant SoloStar pens for Toujeo, Lantus, and Soliqua were interspersed within the booth, and an engrossing virtual reality experience depicting “hypoglycemia at 30,000 feet” that drew an eager crowd. In a separate location, Sanofi’s unbranded Glycemic Explorer booth made a reappearance – this is becoming somewhat of a fixture, present at the exhibit halls of AACE, ENDO, and last year’s ADA.

Tandem

Following a successful Media Day yesterday, Tandem didn’t have much more to divulge in its booth: One employee lamented that the t:slim X2-Dexcom G5 approval hasn’t yet come through – we hear it could happen any day now; it would’ve been nice for the company to be able to detail it at ADA, especially right across the aisle from Medtronic’s booth. We look forward to seeing the feasibility data (n=10 between Barbara Davis and Stanford) on Monday for the predictive low glucose suspend device with G5 – the rep hinted that the algorithm simply “does its job.” The pivotal study, which is preparing for enrollment, will have 90 patients and could “hopefully be approved by FDA in January, February 2018.”  As usual, Tandem reps were extremely complimentary of all interactions that they have with the awe-inspiring FDA CDRH division.

-- by Melissa An, Adam Brown, Abigail Dove, Helen Gao, Varun Iyengar, Brian Levine, Payal Marathe, Emily Regier, Maeve Serino, and Kelly Close