EASD 2019 (European Association for the Study of Diabetes)

September 16-20, 2019; Barcelona, Spain; Day #2 Highlights – Draft

Executive Highlights

  • Hola from Barcelona, where Day #2 of EASD 2019 was highlighted by a slew of exciting new data. Full results from the SUSTAIN 8 trial of Novo Nordisk’s once-weekly Ozempic (injectable semaglutide) vs. J&J’s Invokana (canagliflozin) were presented, demonstrating superior A1c lowering and weight loss with semaglutide. Weight loss results here for semaglutide were truly impressive, and the overall data fit in nicely with PIONEER 2 results of oral semaglutide vs. Jardiance (empagliflozin). Speaking of oral semaglutide, we also have coverage of a new analysis of the PIONEER program, showing oral sema’s consistent benefits on A1c and weight across a wide spectrum of baseline A1c status. This is not surprising but nice to see the data racking up.

  • SGLTs in type 1 were back in the spotlight, with several presentations reviewing data old and new on this hot topic. A new post-hoc analysis of the inTandem program for sotagliflozin tackled the question of how the risk/benefit profile for the therapy might be shifted by applying the BMI >27 kg/m2 as mandated by EMA in its approval. Somewhat similarly to a parallel analysis done of the DEPICT program for dapagliflozin, this analysis of inTandem found greater A1c reductions, weight loss, and time in range improvements in the BMI >27 subgroup, coupled with potential DKA risk attenuation.

    • There’s always a lot to be said for the argument for greater education prompting better outcomes as well; Kelly’s pump failed the day she arrived in Barcelona and by following the STICH protocol, she measured her ketones, saw the score over 4.0, and had it down to 1.0 within three hours. Lots of vomiting in between, but an upside of more training for DKA stemming from SGLT-2s in type 1 “helping” outcomes in traditional DKA is a net positive. “The protocol for treating DKA isn’t intuitive, but it’s very learnable,” she shared with KOLs at The diaTribe Foundation’s Solvable Problems in Diabetes this evening.

  • In diabetes technology, Medtronic announced a CE Mark for its new, no-calibration, fully disposable, blinded professional CGM, Envision Pro, which will launch in the EU and Middle East in the fall. Launch will begin with pilots in the UK and Netherlands. It’s a big step for Medtronic to get to no fingerstick calibration (MARD: 11% - details below on the approach) and a fully disposable transmitter with Bluetooth-to-app.

  • Also in diabetes technology, we saw new six-month data on OpenAPS, where participants on the DIY hybrid closed loop achieved 80% time-in-range (70-180 mg/dl), 1.7% time spent <70 mg/dl, and a 1.2% A1c reduction (baseline: 7.9%) – all miles better than the control group on open-loop therapy.

  • The diaTribe Foundation’s 6th Annual Solvable Problems in Diabetes event just finished, and boy was it an incredible conversation with Profs. Chantal Mathieu, Stephanie Amiel, Juliana Chan, and Tina Vilsbøll joining our very own Kelly Close on stage. The speakers showed off their incredible range of diabetes expertise, touching on what they’re most looking forward to at EASD 2019, individualization of care and improving population level outcomes, new ESC guidelines, Time in Range impressions, diabetes advocacy, and much more. Check out some of our favorable quotes from the night below!

Salutacions des de Barcelona, where Day #2 of EASD 2019 is officially in the books. See our top highlights from the day below, and be sure to see our highlights report from Day #1 for can’t-miss-happenings from nearly 40 hours of industry symposia on Monday. Want to know what’s ahead? See our preview here for a detailed rundown of what’s to come over the final three days of the meeting.

Table of Contents 

Diabetes Therapy Highlights

1. SUSTAIN 8 Full Results: Head to Head Trial Shows Superior A1c Reductions and Weight Loss with Ozempic vs. Invokana

Dr. Ildiko Lingvay (UT Southwestern) presented full results from the SUSTAIN 8 trial of Ozempic (injectable semaglutide) vs. Invokana (canagliflozin), demonstrating superior A1c lowering and body weight reductions associated with semaglutide treatment. Full results were also published today in The Lancet Diabetes & Endocrinology. On A1c from a baseline of 8.3%, semaglutide 1.0 mg drove a 1.5% reduction vs. 1.0% for canagliflozin 300 mg (p=0.0001), while on weight from a base of 90 kg, semaglutide gave a 5.3 kg drop vs. 4.2 kg for canagliflozin (p=0.0029). Impressively, semaglutide helped a higher proportion of patients reach ADA’s A1c target of 7% (66% reaching goal vs. 45% for canagliflozin) and AACE’s more aggressive A1c target of 6.5% (53% vs. 24%). On this point, Dr. Lingvay noted that this 6.5% target is an especially challenging goal to achieve in clinic, and the fact that it was met by over half of patients on semaglutide is very noteworthy.

  • Semaglutide truly differentiated itself on weight loss outcomes, with 22% of patients achieving ≥10% weight loss (vs. 9% for canagliflozin, p<0.0001). Moreover, a post-hoc analysis showed that 7% of patients on semaglutide achieved ≥15% weight loss vs. 1% for canagliflozin (p<0.0001) – this is highly notable, as Dr. Lingvay explained that it’s the first time ever that a glucose lowering agent has shown such weight loss. This impressive weight loss isn’t surprising per se (after all, Novo Nordisk has initiated a dedicated obesity program for semaglutide (Phase 3a STEP program (four trials) initiated 2Q18; SELECT CVOT in obesity launched 3Q18; Positive phase 2 data in obesity in 2Q17) but is striking to see against another efficacious therapy that also has significant weight loss effects itself. Interestingly, the authors of the study, in the discussion section of the full paper in The Lancet Diabetes & Endocrinology, point to potential synergistic weight loss effects with SGLT-2s + GLP-1s given their putative distinct mechanisms of driving weight loss. Synergistic weight loss effects with these two drugs would be a major win for patients and we hope to see studies further exploring this possibility.

  • Regarding clinical implications, SUSTAIN 8 results heartily support the notion of using semaglutide over canagliflozin (and other SGLT-2s) as second-line therapy when A1c lowering and weight loss are the primary concerns. However, an individualized approach for each patient must still be applied, and we note that SGLT-2s may be favored in situations where patients have established heart failure, are at risk for heart failure, or have CKD. Patient preferences regarding using a once-weekly injectable therapy vs. a daily oral pill must also be considered. Regardless of head to head comparisons between the class, it’s clear that agents from both classes are highly efficacious and that the broader community should continue working to ensure as many patients as possible who can derive benefit are getting onto these drugs.

  • Full results from SUSTAIN 8 dovetail nicely with those from PIONEER 2 (oral semaglutide vs. empagliflozin) in generating a larger evidence base for semaglutide vs. the SGLT-2 class. As a reminder, full results from PIONEER 2 were presented at ADA 2019, showing that  from a baseline of 8.1%, mean A1c declined to 6.8% for those on the oral GLP-1 agonist (n=411) and to 7.2% for those on Lilly/BI’s SGLT-2 inhibitor Jardiance (n=410) over 26 weeks (p<0.05). In PIONEER 2, weight loss was similar in both groups: From a baseline 202 lbs, mean body weight fell by 8 lbs with both molecules (non-significant p-value) – a majority of this weight loss occurred in the first 26 weeks and was maintained through the full year. 

2. Sotagliflozin in Type 1: Post-Hoc inTandem Analysis Confirms Greater Efficacy in BMI >27 Group (Greater A1c Reductions, Time in Range Benefits, and Weight Loss) with Potential DKA Risk Attenuation

Dr. Thomas Danne presented a post-hoc analysis of sotagliflozin’s type 1 program sub-grouped by BMI (above and below 27 kg/m2, the cutoff point advised by EMA), demonstrating greater A1c reductions and time in range improvements with treatment in >27 BMI patients and slightly lower DKA rates between groups. We noted considerable enthusiasm for this presentation, which was part of a broader session on SGLTs in type 1 – in Europe, both dapagliflozin and sotagliflozin have been approved by EMA with the same guidelines, which understandably is driving increased interested here. Dr. Danne framed his talk by underscoring that any new type 1 therapy must balance its provided benefits with any risks it may introduce, and that the BMI cutoff proposed by EMA is attempting to better triangulate such a favorable risk/benefit profile. However, it’s been unclear so far how exactly that decision was made, and what data might support this guideline. We’ve seen pooled data from the DEPICT program for dapagliflozin in type 1 presented as a poster at ENDO 2019, which showed that patients with a BMI ≥27 kg/m2 experienced less DKA than the overall, pooled population, while efficacy and CGM-based outcomes were similar between the ≥27 kg/m2 subgroup and the entire cohort. Encouragingly, our understanding of the pooled data of inTandem presented by Dr. Danne for sotagliflozin in type 1 suggest a similar finding, with slightly lower DKA rates (low event rates surely complicate interpretation) accompanied by greater efficacy in the higher BMI group. See slides below for breakdown on DKA rates (slightly lower in higher BMI group), A1c differences (greater A1c drops in higher BMI group), time in range comparisons (large increase for higher BMI group taking highest dose), and body weight comparisons (greatest effect in higher BMI group taking highest dose). Of course, it’s paramount to keep in mind (especially on DKA endpoints) that these event rates are very, very low – a point that Dr. Danne emphasized when discussing the need for further investigation on this point.

  • Baseline characteristics: patient characteristics were generally well-balanced between the two different BMI groups (above and below 27), with similar baseline age, male/female ratios, A1c, and time in range. Notably, and understandably, baseline total insulin doses were significantly higher in the BMI >27 group (~76 IU/day v s. ~46 IU/day). On this point, Dr. Danne speculated that this higher baseline insulin dose associated with a higher BMI may be “part of the protective effect of being overweight” in regard to DKA risk

3. A Miss in Type 1 Prevention: GABA/GAD Therapy Induces Minimal Changes to C-Peptide and A1c But Shows Potential to Lower Fasting Glucagon

Dr. Kenneth McCormick (University of Alabama at Birmingham) presented intriguing data from the first-ever human study of GABA (neurotransmitter beta-aminobutyric acid) in type 1 diabetes, both alone and in combination with self-titled “diabetes vaccine” Diamyd (GAD-alum). Past pre-clinical research, primarily a 2011 paper published in PNAS studying GABA’s effect on beta cells in mice, has suggested that GABA may be able to hinder the progression of type 1 diabetes through suppression of glucagon, increase in endogenous insulin secretion, or modulation of the autoimmune response (the jury is still out on mechanisms of action). Notably, this year’s EASD community did not seem particularly excited about GABA as a potential treatment, with minimal attendance despite a large assigned auditorium.

In the study, patients aged 14-18 years (n=97) with newly diagnosed type 1 diabetes (within 5 weeks of diagnosis) were randomized to (i) GABA and placebo-GAD; (ii) GABA and GAD; or (iii) placebo-GABA and placebo-GAD. GABA was administered twice-daily, orally (1 g/M2/day, maximum 1.5 g/day), while GAD was administered subcutaneously (20 ug/dose), once at baseline and again at month 1. Preliminary results from the phase 1 trial did not demonstrate a significant difference in the primary endpoint of C-peptide response to MMTT, in either the GABA or GABA/GAD group compared to placebo; however, fasting glucagon was reduced in the GABA/GAD group at 12 months (p<0.013). In addition, no statistical difference was seen in A1c, with nearly superimposable curves between the three groups. Dr. McCormick did note a trend towards greater efficacy from the GABA/GAD combination vs. placebo on fasting C-Peptide/glucagon ratio and daily insulin dose; however we are a bit wary of this supposition without statistical significance. While we look forward to further investigation into GABA/GAD’s effect on glucagon, as well as full results from the study expected in early 2020, the outcomes again show how incredibly challenging it has been to translate type 1 prevention research from the lab bench to the clinic. To date, only teplizumab has shown any kind of convincing efficacy to prevent the onset of type 1 diabetes, demonstrating marked reductions on cumulative diabetes onset (72% vs. 43%, HR=0.41, p=0.006) and time to diagnosis (48 months vs. 24 months) compared to placebo. Novo Nordisk’s combo therapy of anti-IL-21 and liraglutide also recently demonstrated efficacy in a phase 2 trial of newly diagnosed type 1s; full results from this trial are anticipated soon.

  • According to Dr. McCormick, several significant study limitations prevented more robust results. Most importantly, the FDA-mandated dose of GABA was quite low compared to past studies in animal models, and Dr. McCormick mentioned anecdotally that he would increase the dose at least 5-fold if possible, during Q&A (we don’t understand how it is possible that trial design winds up getting done so differently than it otherwise would?). In addition, the young age of trial participants made compliance with a twice daily oral medication fairly difficult, though Dr. McCormick suggested that studies taking the drug thrice daily were warranted following results. 

  • We last heard of Diamyd (GAD) during J&J’s 2Q11 earnings call, after the compound was returned to Diamyd Medical following disappointing phase 3 results. As a reminder, the European trial enrolled individuals recently diagnosed with type 1 diabetes but failed to meet its primary efficacy endpoint (preservation of stimulated C-peptide) at 15 months.

4. ONSET 9: Fiasp Demonstrates Superior 1-Hour PPG vs. Novolog in Patients with Long-Term Type 2 on Bolus Insulin

Mountain Diabetes and Endocrine Center’s Dr. Wendy Lane unveiled new positive phase 3 data on Fiasp (fast-acting insulin aspart) in adults with long-term type 2 diabetes (≥10 years from diagnosis), who have been on a basal-bolus regimen ± OAD for ≥1 year. Fast-acting insulin aspart was found to have superior 1-hour post-meal glucose compared to insulin aspart. Baseline A1c for inclusion ranged from 7.0-10.0%, with a final randomization criterion of A1c ≤9.0%. Compared to Onset 2 and Onset 3, which looked at bolus naïve patients (fast-acting insulin aspart vs. insulin aspart, both with insulin glargine U100; and fast-acting insulin aspart vs. basal insulin, respectively), Onset 9 was specifically targeted to patients with longstanding type 2 who require bolus insulin. As Dr. Lane described in Q&A, “We know this is a progressive disease. These are people with 20 years of type 2 diabetes, and you’re going to have them on multiple agents, and you still have to use bolus insulin … There’s a population that will come to full intensification.”

Following a 12-week run-in period, patients (n=1,091) were randomized to either (i) fast-acting insulin aspart + insulin degludec ± metformin; or (ii) Novolog (insulin aspart) + insulin degludec ± metformin. Any OADs other than metformin were discontinued during the trial. After 16 weeks, change from baseline A1c, the primary endpoint, as well as change from baseline 1-hour PPG increment, a secondary endpoint, were assessed. Fast-acting insulin aspart demonstrated non-inferior A1c (ETD -0.04%; 95% CI: -0.11 to 0.03), as well as superior 1-hour PPG increment (ETD -0.40 mmol/L; 95% CI: -0.66 to -0.14). In addition, fewer overall or BG-confirmed hypoglycemic events were found for insulin aspart – overall, daytime, and nocturnal (we are looking for more specific data on this).

  • Overall, we’re excited to see another possible area for Fiasp to fit into across quite a competitive market. Despite its positive profile, results aren’t over-the-top “wow,” and our sense is that thought leaders seem somewhat split over Fiasp’s added value vs. current mealtime options. Some feel that the advantages it brings over NovoLog are essentially marginal, while others have identified potential for its use in closed loop systems due to the molecule’s faster PK/PD profile – we agree here, and can’t help but wonder whether Fiasp will shine brightest not as a standalone diabetes drug, but as a component of closed loop systems and maybe even a dual hormone artificial pancreas. It’s certainly costly to try to develop faster insulins, though it’s incredibly clear the value – EASD is buzzing about how very, very much hyperglycemia there is and the answer to this is not an algorithm – it’s faster insulin or better food chocies, in our view.

5. PIONEER Analysis Shows Consistent Efficacy with Oral Semaglutide Across Baseline A1c Levels; Greatest Effects with Highest Doses + Higher Baseline A1c

Dr. Juris Meier presented a new analysis of the PIONEER trials for oral semaglutide demonstrating the therapy’s consistent effect on glucose and weight across a wide spectrum of baseline A1c levels. In the analysis, all patients from PIONEER 1-5, 7, and 8 were included (n=5,657) and binned into baseline A1c groups of ≤8%, between 8% and 9%, and >9%. A1c reductions associated with oral semaglutide showed consistent benefit across each baseline A1c group, with larger A1c drops understandably seen in the higher baseline A1c groups. Results reinforce oral semaglutide’s potential in a wide patient population, allowing HCPs to confidently prescribe the therapy across a broad spectrum of patients, ranging from those achieving relatively decent glucose control (A1c <8%) to those who might be experiencing greater challenges with glucose control (A1c >9%). Notably, this analysis further confirmed oral semaglutide’s robust superiority on glycemic effects when compared to the active comparators it was tested against in the PIONEER program: oral semaglutide at a higher dose (7 mg and 14 mg) achieved greater A1c reductions across baseline A1c levels when compared to active comparators of empagliflozin in PIONEER 2, sitagliptin in PIONEER 3, and liraglutide in PIONEER 4. See summary table below:

6. More Hope for Type 1? Positive Results from Teplizumab Trial Spur Hope For New Biomarkers and Earlier Type 1 Identification

According to principal investigator Dr. Kevan Harold from Yale (he’s a big deal), positive results from TrialNet’s teplizumab type 1 diabetes prevention trial will light the way for the discovery of more biomarkers to further delay type 1 onset. As a refresher, stunning results first shared at ADA 2019 revealed marked reductions on cumulative diabetes onset (72% vs. 43%, HR=0.41, p=0.006) and time to diagnosis (48 months vs. 24 months) with the anti-CD3  teplizumab treatment when compared to placebo. These results have raised hope for the existence of a treatment window to prevent type 1 diabetes – to further open this window, Dr. Harold believes that future studies should focus on identifying new biomarkers for subjects most likely to respond positively to treatment, investigating the effects of repeated dosing, and determining the potential of combinations therapies.

  • Although lauded for its positive effects, Dr. Howard cautioned that teplizumab has been associated with decreased immunity (reduction in circulating lymphocytes) and rashes, making it similar to any other drug where cost-benefit analysis must be performed before use. With this, response to teplizumab also differs among patients, conferring more delay in patients with C-peptide responses in the lower half of the study’s overall sample.

  • Dr. Anette Ziegler lauded the prevention trial as a clinical breakthrough and suggested that there is a biomarker of teplizumab response that can be used to find other agonistic therapies. Exhausted CD8+ T cells are a promising biomarker, and analysis of the differences in C-peptide responses within the original sample can potentially elucidate the underlying mechanism at work. In order to confirm or deny these suspicions, Dr. Ziegler suggests revisiting pre-existing data, like that found in the TEDDY trial (which showed similar progression of type 1 diabetes in the general population and children with a diagnosed first degree relative). For example, the recently launched German FR1DA trial is taking the right steps to learn more about type 1 disease progression by screening antibodies in children nationwide. So far, the study has determined that ~20% of all children have dysglycemia or hyperglycemia, increasing the need for new studies to understand type 1 disease progression at a large scale.

  • Dr. Ziegler believes we can better define presymptomatic stages of type 1 diabetes by looking at low points of C-peptide reactivity in early study stages. She believes that using rapid progression markers (age, genetic risk score, antibodies and epitopes, and antibody score) is integral to this process. It’s also important to increase the therapeutic options available to young children, as so many children develop type 1 diabetes annually. She also emphasized that it’s crucial to have a high enough drug supply to support these new proposed studies, and that subsequent novel treatments need to be affordable for patients.

7. DynaHEALTH Consortium Brings a Bio-Psychosocial Approach to Diabetes Epidemiology

Finland’s Dr. Sylvain Sebert provided an update on the exciting DynaHEALTH consortium, an ambitious project to study healthy vs. unhealthy aging in terms of glycemic health. DynaHEALTH is a research consortium funded by the European commission through the Horizon 2020 initiative. The consortium integrates together data from 20 separate longitudinal studies from eight European countries (Belgium, Denmark, Finland, Germany, Italy, the Netherlands, and Spain). This encompasses a remarkable 1.3 million participants, with the oldest born in 1934 and the youngest born in 2018. With data this long-ranging, Dr. Sebert pointed out that DynaHEALTH is extremely well-positioned to provide insight into the challenges of an aging population, particularly obesity and type 2 diabetes. Another unique aspect of the consortium is its bio-psychosocial approach to studying Europe’s aging population, factoring in stress and social adversity as well as traditional clinical measures.  The consortium’s main areas of focus are the relationships between: (i) adiposity and the risk of deteriorating glycemic health/type 2; and (ii) type 2 and stroke, coronary heart diseases, dementia, and Alzheimer’s.

  • Dr. Sebert highlighted a 2018 study as an example of the kinds of insights we can expect from the DynaHEALTH consortium going forward. The study examined longitudinal data spanning 46 years from the Northern Finland Birth Cohort (n=5,078), searching for factors that explain the participants’ variability in fasting glucose (i.e. diabetes, prediabetes, or diabetes-free) over the lifespan. Very interestingly, fasting glucose levels at age 46 could be predicted using a mathematical model combining socioeconomic factors (education level, occupation, household income), metabolic factors (insulin, waist circumference, BMI, HDL cholesterol, triglycerides), blood pressure measurements, and psychosocial factors (marital status, employment status, home ownership, depression, sleep quality, life satisfaction) from when a person was only 31 years old. We are especially intrigued about the psychosocial dimension of diabetes management and prevention and look very forward to more output from DynaHEALTH.

8. Dr. Rayaz Malik Argues for a New Biomarker for Diabetic Neuropathy in Camillo Golgi Lecture

In the highly-prestigious Camillo Golgi lecture, Weill Cornell Medical College-Qatar’s Dr. Rayaz Malik argued for a new biomarker (! seriously …) for diabetic neuropathy and diabetes complications more broadly: corneal confocal microscopy (CCM). He explained that traditional tests of diabetic neuropathy (e.g. 10 g monofilament, vibration perception threshold, electromyography) are not sensitive enough to detect early-stage neuropathy, in part because they measure only large nerve fibers as opposed to the small, unmyelinated fibers that are the first to be damaged in the natural history of the disease. CCM measures the small neural fibers of the cornea with a two-minute eye-imaging test – standard practice in many ophthalmology clinics. According to Dr. Malik, this measure correlates almost perfectly with intra-epidermal fiber density, the current “gold standard” measurement for diabetic neuropathy. Beyond this diagnostic validity, CCM also has prognostic validity – that is, it is predictive of future development of neuropathy as well. Diabetic neuropathy is a huge area of unmet need in diabetes care. According to the ADA’s 2017 position statement, available treatments yield only about 50% pain relief, and in only 1/3 of patients. Dr. Malik suspects that earlier detection of neuropathy – using a biomarker like CCM – would vastly improve outcomes, and from a therapy perspective enable the development of drugs that act earlier in the progression of the disease. To this end, Dr. Malik coined neuropathy the “Cinderella complication.” He elaborated, “there are fantastic ways of assessing other complications like retinopathy and nephropathy. But in contrast neuropathy relies on clinical examination of the foot.” Of course, visible foot ulcers are a late sign of neuropathy, leaving clinicians unable to detect the disease earlier, when intervention would be more effective.

  • CCM’s detection of early-stage diabetic neuropathy suggests that this complication is more widespread than previously believed. Dr. Malik’s research has found CCM indications of very early stage neuropathy (that is, decreased corneal nerve fiber length and decreased corneal nerve fiber density) in people with type 1 diabetes who had not yet developed retinopathy (traditionally viewed as the earliest-onset diabetes complication), and even in people with prediabetes. Even children with type 1 diabetes (mean age 14 years, mean A1c 9%) showed signs of slight CCM impairments, suggesting the very earliest signs of neuropathy setting in.

    • CCM may be an indicator of neurodegeneration more broadly, also correlating with the severity of diseases of the central nervous system like Parkinson’s, MS, and Alzheimer’s disease (itself another complication of diabetes).

    • Interestingly, several interventions have shown improvements in diabetic neuropathy, as measured by long-lasting changes in CCM. These include bariatric surgery, pancreas/kidney transplantation, omega 3 supplementation, and even vitamin D supplementation.

  • Dr. Malik expressed optimism that one day CCM will be recognized as a valid clinical trial endpoint with the FDA. Although this may be an uphill battle, Dr. Malik pointed out parallels between the current CCM movement and one decades ago for A1c to be accepted as a valid clinical endpoint. Dr. Tony Cerami, the discoverer of A1c, once personally told him “Don’t get discouraged. It took years for A1c to be accepted by the FDA as a measure of diabetic control.” To this end, Dr. Malik pointed out that 562 research centers across 45 countries now use CCM in clinical practice. He closed with a powerful sentiment for the future of CCM use: “The eye is the window to the soul, and beyond.” This is something diabetes advocacy should be watching and helping  - they now have marching orders!

9. Skin Autofluorescence (Measure of AGEs) Predicts Incident CVD and Mortality in Type 2, On Top of Age, Baseline CVD, Smoking Status, Statin Use, etc.

According to a new prospective study of 2,327 people with existing type 2 diabetes detailed by University of Hroningen’s Prof. Bruce Wolffenbuttel, baseline CVD and age were the strongest predictors of CVD and mortality after a median follow-up time of four years. In newer news, baseline skin autofluorescence – a noninvasive measure of vasculature-stiffening and atherosclerosis-inducing advanced glycation end-products – was also significantly associated with CVD and mortality, independent of traditional risk factors, as were current smoking status, waist circumference, systolic blood pressure, and statin use. In those with no baseline CVD (n=2,050), age remained the strongest predictor of the composite endpoint, followed by skin autofluorescence, statin use, current smoking status, waist circumference, and systolic blood pressure. See the tables below for odds ratios, 95% CIs, and p-values. Prof. Wolffenbuttel pointed out that these data were limited in that the exact time of CVD events were not collected, cause of death is not (yet) available, the study was performed only in people of Western-European descent, and there was no follow-up questionnaire or AGE quantification. An audience member added during Q&A that chronic caffeine intake and smoking status elevates AGEs in the skin, as do other nutrients (likely in more chronic fashion). Still, AGE as a one-time, noninvasive measure on top of age and current CVD disease status is clearly an easier way to further risk stratify populations in the clinic. To that point, Prof. Wolffenbuttel referenced a 2009 paper showing that skin autofluorescence provides additional predictive power to the UKPDS risk engine. Interestingly, skin autofluorescence also predicts incidence of type 2 diabetes, CVD, and mortality, in those without baseline CVD (van Waateringe et al., Diabetologia 2019).

10. Phase 3a GLP-1 data Shows Favorable A1c-lowering Profile for Hansoh Pharma’s Novel Once-Weekly GLP-1 Agonist PEX 168

A poster showed favorable phase 3a data for Hansoh Pharma’s once-weekly GLP-1 agonist PEX168 (polyethylene glycol loxenatide) in people with type 2 diabetes. The trial randomized people with type 2 diabetes (n=406) to either 100 or 200 µg PEX168 or placebo for 24 weeks. At the end of the study, participants on PEX168 experienced significant improvements in A1c: -0.99% from baseline for the 100 µg dose and -1.34% for the 200 µg dose(vs. a non-significant -0.15% with placebo). Furthermore, the proportion of patients who achieved an A1c <7% by the end of the 24 weeks stood at 35% and 47% in the 100 µg and 200 µg dose groups, respectively, both significantly higher than in the placebo group (16%). No serious adverse events or severe hypoglycemia events occurred in the trial, and most adverse reactions were mild to moderate GI reactions, as is expected for the GLP-1 agonist class. According to the study’s ClinicalTrials.gov page, a 28-week extension period is planned, where participants randomized to placebo will cross over to one of the active comparator arms. We look forward to following up on the full readout and will be particularly eyeing the weight loss data, which is listed as a secondary outcome. While the phase 3a data presented here look encouraging, we note that the GLP-1 agonist space is highly competitive, and PEX168 will have a high bar to clear if it is to compete against heavyweights like Trulicity, Victoza, and Ozempic – market leaders with impressive profiles for combination glucose-lowering, weight-loss, and CV protection. Notably, PEX168 is already approved in China, launched earlier this year under the brand name Fu Laimei.

11. EASD President David Matthews Discusses Aims, Sessions to Watch, and Organizational Updates in EASD 2019 Presidential Address

President Matthews’ talk centered on the similarities between the study of diabetes and one of Barcelona’s most iconic landmarks: La Sagrada Familia. Both are complex and remain unfinished, but a crowd is waiting to learn more about its intricacies. Dr. Matthews emphasized EASD’s mission of promoting excellent diabetes care through research and education, advertising their commitment to post-graduate education with a new e-learning platform with virtual courses available to anyone in addition to formal programs and fellowships. Excitingly, the organization is also working on setting up a European-wide diabetes forum under the EUDF consortium (comprised of the EASD, EFSD, FEND, and JDRF). The EUDF will focus their research on healthcare systems, diabetes self-management, and technology to make life better for people with diabetes. Dr. Matthews also announced the new EFSD/Lilly EXPAND program, which will be dedicated to translational diabetes research.

  • Notably, Dr. Matthews also listed a few talks to look out for at EASD – we believe this is unprecedented. These include (i) Dr. Rayaz Malik’s session on diabetic neuropathy; (ii) Timo Otonkoski’s session on the right amount of insulin; (iii) Dr. Daniel Drucker’s session on metabolic solutions; and (iv) Filip Knop’s session on glucagon. Be sure to reference our EASD 2019 preview for our team’s own suggestions!

Diabetes Technology Highlights

1. Medtronic’s No Calibration, Bluetooth, Fully Disposable Envision Pro Blinded Professional CGM CE Marked; EU & Middle East Launch This Fall, Starting With UK & Netherlands; Professional CGM Comparisons

Medtronic announced this morning that it has received CE Mark approval for the fully-disposable, no-fingerstick-calibration, Bluetooth, blinded, seven-day-wear, 11.0% MARD (vs. YSI) Envision Pro blinded professional CGM system for use in type 1 and type 2 diabetes. Launches are expected to commence in Europe and the Middle East this fall, beginning with pilots in the UK and Netherlands. This international launch comes on the early side of prior expectations for a launch before April 2020. The system features two major CGM firsts for Medtronic: zero fingerstick calibrations and a fully-disposable transmitter design – both critical for the professional CGM market (simplicity!) and helping close the gap with Abbott’s FreeStyle Libre Pro and Dexcom’s under-FDA-review G6 Pro. We confirmed with Medtronic that Envision Pro will not be launched in the US, though Medtronic hopes to build on this for its US product cadence. In a big step up from Medtronic’s current iPro2 professional CGM, data from the Envision Pro transmitter is also sent directly to the Envision Pro app via Bluetooth, which then relays the data to CareLink. That means patients don’t have to return the transmitter to the clinic. During the wear session, users can capture events in the Envision Pro app, similar to the myLog app that launched for iPro2 in 2016. The on-body form factor appears to be identical to the current iPro 2 (blinded professional CGM), Guardian Connect, and Guardian Sensor 3, with the purple sensor and white clamshell transmitter (picture below). The paired app looks identical to the iPro2 myLog app (minus fingerstick logging since calibration is no longer needed) and the diagnostic reports for providers are the same that have been used for iPro 2 (improvements in decision support are in the works as Medtronic looks to target more PCPs).

  • Medtronic’s Envision Pro CGM system is not factory calibrated; instead, it is powered by a new retrospective algorithm that leverages multiple sensor signals and applies machine learning to optimize the sensor’s accuracy. There are two major factors that allow the algorithm to perform with zero fingerstick calibrations: (i) an improved sensor used to collect data – the Envision sensor is designed and manufactured to collect more accurate sensor data, with less noise and more signal consistency (details not provided); and (ii) Utilization of multiple sensor data inputs – the algorithm in Envision Pro leverages Electrochemical Impedance Spectroscopy (EIS) collected through the Envision recorder’s diagnostic chip (see an explanation from DTM 2018); this allows the algorithm to adjust the sensitivity factor and calculate sensor glucose values without the need for calibration of blood glucose inputs. At DTM 2018, Medtronic termed this “self-calibrating,” as opposed to “factory calibrated.”

  • On the real-time CGM front, as of last month, the pivotal trial to obtain real-time iCGM labeling with day one-only calibration remained underway; FDA submission was previously forecasted for April-October 2020.


  • Accuracy data from the  user manual suggests that the new Envision Pro algorithm’s accuracy appears to be in line with that of the iPro2, with the big upgrade to no calibration (vs. four fingersticks required per day for iPro2). Reps didn’t specify how many people were included in the accuracy studies, but they all had either type 1 or type 2 diabetes.

  • While no calibration, Bluetooth, and fully disposable are big steps for Medtronic, Envision Pro still lags behind Abbott’s FreeStyle Libre Pro and the Dexcom G6 Pro (currently under FDA review) in terms of wear time. See a comparison of the three major professional CGM players in the following table. Note that Medtronic previously aimed to launch a real-time professional CGM, which could very well allow for toggling between real-time and blinded, after April 2020.


Medtronic Envision Pro

Abbott FreeStyle Libre Pro

Dexcom G6 Pro (under FDA review; 2020 launch expected)

Wear time

7 days

14 days

10 days

No calibration?




Fully disposable?







(Assuming it’s the same sensor and algorithm as G6)

Bluetooth to app?


All data uploaded automatically to CareLink


Patient must come to clinic to have sensor downloaded with reader


Data Uploaded to Dexcom Clarity (via G6 app) or stored on receiver for in-clinic download

Blinded vs. Real-Time?

Blinded Only
App allows manual entry of contextual info – meals, medication, activity

Blinded Only

Blinded and Real Time Modes


2. OpenAPS Outcomes Study (n=51): 80% TIR, ~2% Time <70 mg/dl vs. 72% TIR, 7% Time <70 mg/dl in Pump-Only Group

Dr. Anastasios Koutsovasilis (General Hospital of Nikaia) presented strong, new data from 23 patients on OpenAPS vs. 28 on open-loop, with participants initiated on OpenAPS achieving a notable 80% time-in-range (70-180 mg/dl) and 1.7% time spent <70 mg/dl after six months. The study was done in participants who had been on pump therapy for > three months with baseline A1c between 7.5% and 10%. Both groups used the DANA S pump with a Dexcom G5 CGM for the study. Due to the small sample sizes, the time-in-range achieved by OpenAPS was not statistically different than that of the control group (72%), though time in hypoglycemia was very significant: 1.7% with OpenAPS vs. 6.9% on open loop (p=0.04). The improved numbers in OpenAPS users reflects an extra two hours/day spent in range and ~1.2 hours less time <70 mg/dl. It’s hard to believe that Time in Range was not officially statistically different with these kinds of results - again, we stress the “n” was the difference. You could ask absolutely any patient on CGM if they would prefer 80% TIR or 72% TIR and you’d receive a strong argument for 80%. A1c also improved by 1.2% with OpenAPS (baseline: 7.9%) vs. no statistically significant reduction with open loop (7.8% to 7.4%; p=0.1). We view this as another great validation of automated insulin delivery, even in an engaged population doing technically quite well at baseline if the “below 70” and “below 53” and “above 250” are ignored – these represented a total of two hours in hypoglcyemia per day and two hours above 250 mg/dL, compared to 26 minutes and 17 minutes, respectively.


OpenAPS (n=23)

Open-loop pump (n=28)


Time-in-range (70-180 mg/dl)




Time-in-range (90-180 mg/dl)




Time <70 mg/dl




Time <53 mg/dl




Time >180 mg/dl




Time >250 mg/dl




  • Over six months, participants on OpenAPS saw A1c reduction from 7.9% to 6.7% (p<0.001), while the control group showed Fasting plasma glucose, post-prandial glucose, and bedtime glucose values were significantly reduced for OpenAPS patients. Lastly, no patients in the OpenAPS group showed nocturnal hypoglycemia or hypoglycemia “during the day before meals,” despite every participant in the study having experienced “at least one hypoglycemic episode” in the month before the study. Notably, the control group did not see significant change in hypoglycemia after the study began.




3 Months

6 Months

A1c (%)









Fasting plasma glucose (mg/dl)









Glucose before bedtime (mg/dl)









The diaTribe Foundation’s Sixth Annual Solvable Problems in Diabetes Forum

1. The diaTribe Foundation’s “Solvable Problems in Diabetes” Brings Drs. Mathieu, Amiel, and Chan, and Vilsbøll Together to Discuss EASD 2019, New ESC Guidelines, Diabetes Advocacy, and More

Tonight, the diaTribe Foundation hosted the 6th Annual Solvable Problems in diabetes, putting together a fascinating panel featuring Professors Stephanie Amiel, Juliana Chan, Chantal Mathieu, and Tina Vilsbøll. In time, we’ll provide a number of quotable quotes to capture the many incredible insights shared this evening on a range of topics:

  • On what to most look forward to at EASD 2019

  • On individualization of care and improving population level outcomes

  • On the new ESC/EASD Guidelines (see ESC Days #1-3 Highlights)

  • On Diabetes Advocacy

  • On Tele-medicine

What an incredible group of leaders! We can’t wait to report back.

--by Albert Cai, Ursula Biba, Anirudh Gururaj, Rhea Teng, Abigail Dove, Brian Levine, Martin Kurian, and Kelly Close