September 21-25, 2020; Virtual; Full Report – Draft

In this report, we provide our full coverage of diabetes/obesity therapy, diabetes technology, and big picture topics from EASD 2020. Dive into 16 chapters, ranging from SGLT-2s to digital health to obesity therapies. Titles highlighted in blue represent new or expanded coverage that wasn’t included in our daily highlights (day #1, day #2, day #3, day #4, and day #5).

Note that some talks may appear in multiple sections. See these links for therapy highlights, tech highlights, and big picture highlights! Navigate through all 347 pages of reporting using our table of contents below – happy reading!

Themes

Diabetes Therapy

A Bird’s Eye View on SGLT-2 Inhibitors: KOLS Step Back to Appreciate Successes in CKD and HFrEF, with or without Type 2 Diabetes from VERTIS-CV, DAPA-CKD, EMPEROR-Reduced, and more

This summer brought the results of three critically important SGLT-2 inhibitor CVOTs: first VERTIS CV at ADA 2020, then DAPA-CKD and EMPEROR-Reduced at ESC 2020. At EASD 2020, lead investigators took a step back and considered what these findings meant for the SGLT-2 inhibitor class as a whole. Through new analyses and expert discussion, we became certainty convinced of the tremendous potential of the drug class in chronic kidney disease (CKD) and heart failure with reduced ejection fraction (HFrEF), with or without diabetes.

• EASD SVP Prof. Chantal Mathieu (KU Leuven) praised the study design and results of DAPA-CKD, highlighting its (i) enrollment of patients with eGFR ≥25 and ≤75 mL/min/1.73 m² and UAC ≥200 and ≤5000 mg/g; (ii) excellent participant retention, collection of vital status, and diverse patient population (albeit with poor Black or African American representation); and (iii) incredible efficacy on the primary renal endpoint (39% relative risk reduction) and all-cause mortality. As a call to action, Prof. Mathieu encouraged usage of Farxiga (dapagliflozin) in patients with eGFR as low as 25, and for specialists to work more closely with PCPs to develop actionable guidelines for SGLT-2 inhibitors in treating CKD.

• We also got to hear from DAPA-CKD lead investigators Profs. Hido Heerspink and David Wheeler, who discussed how these results may influence risk stratification, primary prevention, CKD guidelines, and our understanding of the mechanism of SGLT-2 inhibitors, among other topics.

• In a pre-specified exploratory analysis of VERTIS-CV, SGLT-2 inhibitor Steglatro (ertugliflozin) delivered a 34% relative risk reduction on the composite renal endpoint of (i) sustained ≥40% decrease in eGFR; (ii) dialysis/kidney transplant; or (iii) renal death (HR: 0.66; CI: 0.50-0.88; p<0.01). During this session, Dr. Darren McGuire (UT Southwestern) presented meta-analyses with VERTIS-CV, EMPA-REG, CANVAS, CREDENCE, and DECLARE-TIMI to confirm a strong class effect for SGLT-2 inhibitors for renal and hospitalization for heart failure (hHF) risk reduction. This new analysis was received warmly by us and the audience after VERTIS-CV failed to show significance on a more rigorous renal endpoint at ADA 2020.

• In a presentation covering a recent meta-analysis of EMPEROR-Reduced and DAPA-HF published in The Lancet, Dr. Faiez Zannad (University of Lorraine) emphasized that the results of both trials were “nearly superimposable.” Indeed, on the primary outcome of first hHF or CV death, both dapagliflozin in DAPA-HF and empagliflozin in EMPEROR-Reduced were associated with a significant risk reduction in patients with heart failure with reduced ejection fraction (HFrEF), irrespective of diabetes. Significant benefits for CV and all-cause death were observed and driven by dapagliflozin, while significant benefits for renal outcomes were driven by empagliflozin. Despite these differences, Dr. Zannad emphasized that the evidence supports a protective, class effect of SGLT-2 inhibitors in renal disease and mortality. Dr. Milton Packer expanded on these findings by highlighting a potential cardiorenal protective mechanism of SGLT-2 inhibitors, which is independent of blood glucose lowering. With these findings in mind, he stated with conviction that “we are ready to add SGLT-2 inhibitors as one of the cornerstones of HF treatment.”

Many Paths Forward: GLP-1s Generate Promising Data in CKD, Stroke, Obesity and Type 1 Diabetes

While the results of large-scale SGLT-2 inhibitor trials certainly received greater attention during EASD 2020, we were excited to see new post-hoc analyses of completed GLP-1 receptor agonist trials and updates on those still ongoing. Novo Nordisk was involved in many of these presentations, which together brought more clarity and optimism for the drug class in reducing risk of CKD, stroke, and CVD in type 2, and in achieving indications for obesity and type 1.

• Novo Nordisk sponsored a popular symposium on the first day of the conference, supporting the potential for Rybelsus (oral semaglutide) across a diversity of type 2 diabetes patients. Post-hoc analysis of PIONEER 1 through 8 demonstrated that regardless of baseline A1c, diabetes duration, age, background mediation, and race/ethnicity, all participants in the GLP-1 arms exhibited clinically meaningful A1c reductions. Dr. Ofri Mosenzon (Hebrew University Medical School) singled out PIONEER 5 – which investigated Rybelsus in patients with moderate renal impairment – to emphasize that the agent is safe in eGFRs as low as >15 mL/min/1.73 m². The efficacy of Rybelsus in CKD is being explored in FLOW, which will conclude in August 2024.

• Post-hoc analysis of SUSTAIN 6, PIONEER 6, and LEADER also revealed a potential class effect of GLP-1 receptor agonists on stroke. In patients with type 2 diabetes and high cardiovascular risk, combined results of semaglutide and liraglutide showed a significant 18% relative risk reduction (RRR) on time to first occurrence of all strokes (HR: 0.82; 95% CI: 0.68-0.98; p=0.03). Importantly, these benefits may manifest earlier in the treatment course than those seen with PCSK9 inhibitors.

• A separate post-hoc analysis of SUSTAIN 6 and PIONEER 6 also revealed that semaglutide addition to standard of care was associated with a 6% mean absolute risk reduction (ARR) in 10-year CVD risk (95% CI: 1.9-10%). Using the DIAL model based on data from the Swedish National Diabetes Registry, the investigators determined that semaglutide added an additional 1.7 CVD-free life years to people with type 2 diabetes, with the greatest benefits achieved with earlier treatment.

• Novo Nordisk also updated us on the STEP 1-4 program evaluating semaglutide in obesity. In all four trials, 2.4 mg semaglutide has been associated statistically significant weight reductions. These trials investigated the GLP-1 receptor agonist in a diversity of patient populations and study designs: STEP 2 enrolled participants with established type 2 diabetes and STEP 3 offered intensive behavioral therapy (IBT) in addition to pharmacotherapy, for example. We are looking forward to phase 3b of the STEP program, which includes the SELECT CVOT (n=17,500). These results should give more clarity on the association between weight loss and CV benefits.

• We were excited to again see the positive results of Novo Nordisk’s phase 2 trial evaluating anti-Il-21/liraglutide combination therapy in type 1. These results, which were originally presented at ADA 2020, showed that the combination regimen precipitated a 48% improvement in post-meal C-peptide secretion relative to placebo. Although non-significant, beneficial trends for A1c, total insulin dose, and reduced hypoglycemia were also observed. This treatment regimen has received Orphan drug designation by FDA in type 1, and phase 3 trial design is in the works.

Closing the Divide: Epidemiological Data Show Widespread Underuse of Effective Therapies Despite New Data; Whose Guidelines to Follow?

Despite the very promising advances for SGLT-2 inhibitors, GLP-1s agonists, and many more up-and-coming classes, global data continues to show that the majority of patients are not yet receiving these potentially life-saving therapies, and comorbidities like cardiovascular and renal disease are still major concerns worldwide. Data from the Novo Nordisk-sponsored CAPTURE study showed that in 9,823 adults with type 2 diabetes across thirteen countries (Mexico, Argentina, Brazil, Israel, Italy, France, Czech Republic, Hungary, Turkey, Saudi Arabia, China, Japan, and Australia), overall CVD prevalence was estimated at 34.8%. The use of SGLT-2s and GLP-1s, medications with documented CV (and other) benefit, was very low for those with and without CV at ~16% for SGLT-2s and ~10% for GLP-1s. More promisingly, society guidelines seem increasingly enthusiastic about earlier usage of these therapies. Recently published guidelines from Primary Care Diabetes Europe encouraged early combination therapy (i.e. metformin + GLP-1 or SGLT-2 inhibitor) over the traditional, step-wise approach. In addition, interest in differences between the ESC vs. EASD/ADA guidelines continues to persist.

Diabetes Technology

Signs of a Maturing AID Field: More and More Real-World Data in Well-Studied Groups (e.g., Adults) and Building Evidence in Other Populations (e.g., Pediatrics, Technology-Naïve)

Coming into the year, we wrote that 2020 was sizing up to be a big year in AID. With groundwork laid in 2019 and the several years prior, we were looking at up to four (!) new AID product launches in the year: Tandem’s Control-IQ, Insulet’s Omnipod Horizon, Medtronic’s MiniMed’s 780G, and Tidepool’s Loop. Of course, many of these timelines were affected by COVID-19, which quickly reached pandemic status ~2-3 months into the year. Still, Tandem successfully launched its Control-IQ AID system in January 2020 and has quickly been able to roll the system out to pediatrics and internationally. Less than a month after EASD, Medtronic also officially announced that its second-gen MiniMed 780G AID system has launched in Europe. In addition to the two new product launches, this year’s EASD meeting featured lots of data on AID – clinical data in newer populations (e.g., pediatrics) and real-world data in more well-studied populations (e.g., adult type 1s).

• Building on a set of ~2,000-3,000 person data shared at ADA, Tandem shared very positive real-world data from ~6,000 Control-IQ users – representing “~10%” of the total Control-IQ user base. Across the ~6,000 users, Time in Range increased by nearly 2.5 hours/day within the first two weeks of Control-IQ use and was sustained through at least six months. At baseline, mean Time in Range in the sample was 65%, quickly rising to 75% after just two weeks and remaining there through the 30-day, 90-day, and 180-day time points.

  • On the pediatric front, we also saw results from the Free-Life Kid AP Study, which used Control-IQ in kids ages 6-12 with type 1. The outcomes were positive, with those using Control-IQ 24/7 seeing a 15% percentage point increase in Time in Range (+3.5 hours/day) from 53% to 68%. Encouragingly, the time in closed loop (94%) demonstrated that the Dexcom G6/Tandem t:slim X2/Control-IQ system is easy to use in this pediatric population. During a Tandem symposium, UVA’s Dr. Marc Breton showed a slide with four studies evaluating Control-IQ in four age groups: toddlers (ages 2-6), pediatrics (ages 6-13), adults (ages 14+), and seniors (ages 65+). All four studies showed strong results, with Time in Ranges near or above 70% and high time in closed loop.

• In addition to some new details on Medtronic’s MiniMed 780G pivotal, we saw encouraging presentations on implementation of the first-gen 670G system, which is already in the hands of more than a quarter million people. One poster on from Sidra Medicine in Qatar showed successful initiation of the MiniMed 670G AID system in youth (ages 7-18) transitioning to AID straight from MDI. Through six months, participants’ mean A1c fell from 8.2% to 6.9% and their mean Time in Range increased from 46% to 73% (+6.5 hours/day!). The results show that, with a successful training protocol, technology-naïve patients can move directly to AID from MDI – as easier to use systems come to market, we imagine this process might become even more seamless. The data from Qatar also bolster an important implication we drew from the FLAIR study, presented at ADA: technology-naïve patients can benefit just as much, if not more, than patients already using CGM and CSII, and should not be excluded from AID by clinicians and researchers.

• At its industry symposium, Insulet announced that the Omnipod 5 hybrid closed loop pivotal trial has completed. The three-month trial initially began at the end of December 2019, temporarily paused in March due to a “software anomaly,” and resumed again in June. At this point, a “majority” of participants have already moved onto the extension phase. The pivotal trial included patients as young as 6 years old and, pre-pivotal data suggested the system not only has strong glycemic outcomes, but very strong system usability results. System usability is particularly important as AID looks to enter into newer populations, such as pediatrics, elderly patients, developing countries, and more.

Time in Range Validation Continues: Studies Link TIR with MACE, Severe Hypo, Retinopathy, Quality of Life; COVID-19 Further Boosts Value of TIR vs. A1c

At EASD 2020, conversations around Time in Range continued to accelerate. After consensus guidelines for Time in Range (and other CGM metrics) were developed at ATTD 2019, many have asked how the field can further validate Time in Range as a meaningful endpoint for drugs, devices, regulations, payers, physicians, and patients – though preferably without having to invest in expensive, time-consuming trials that will be challenging to conduct. Encouragingly, we saw several presentations at EASD that answered the call, linking Time in Range with MACE, severe hypo, retinopathy, and quality of life. Additionally, with the COVID-19 pandemic and its resulting shelter-in-place orders, we heard many clinician presenters further highlight the value of CGM-derived Time in Range and GMI (eA1c) metrics, as lab A1c measurements were often unavailable.

• Dr. Rich Bergenstal was seemingly everywhere at EASD, but perhaps his most noteworthy presentation was a post-hoc analysis of the DEVOTE trial linking “Time in Range” with major adverse cardiovascular events (MACE) and severe hypoglycemia events. Every 10% increase in Time in Range was associated with a 6% relative risk reduction for three-point MACE (non-fatal MI, non-fatal stroke, CV death). Every 10% increase in Time in Range was associated with a 10% reduction in the risk of severe hypoglycemia events. On this, Dr. Bergenstal commented that the data clearly dispel the notion that “tighter control” (i.e., higher Time in Range) would lead to more hypoglycemic events.

• Another study, done in Belgium, showed a strong correlation between Time in Range and several microvascular complications, particularly retinopathy. In the study, the average retinopathy rate in the Time in Range > 70% group was 22% lower than the rate of retinopathy of Time in Range <40% group. In addition, higher Time in Range was also linked to fewer diabetes-related hospitalizations. In the poster hall, we caught an analysis showing that higher Time in Range (but notably, not A1c) was independently correlated with improved quality of life.

GLP-1 Agonist and DPP-4 Inhibitor Highlights

Post-Hoc Analysis of LEADER, SUSTAIN 6, and PIONEER 6 Shows 18% Reduced Risk of Stroke with GLP-1 Agonist Therapy

A post-hoc analysis of the LEADER, SUSTAIN 6, and PIONEER 6 trials demonstrated that GLP-1 agonist therapy was associated with an 18% reduced risk of stroke (95% CI 0.68-0.98, p=0.03) in people with type 2 diabetes and high CV risk. As a reminder, all three of these past GLP-1 CVOTs have shown compelling risk reduction for 3-point MACE, and though the individual stroke outcome has shown a favorable point estimate in all three trials, only semaglutide in SUSTAIN 6 was associated with a significant risk reduction for stroke (HR 0.61, 95% CI 0.38-0.99, p=0.04). (The LEADER trial of liraglutide showed a HR of 0.89 [95% CI 0.71-1.11] and the PIONEER 6 trial of oral semaglutide showed a HR of 0.74 [95% CI 0.35-1.57]). With this in mind, it is quite encouraging to see (with the benefit of a larger study population and more stroke events) that GLP-1 therapy as a whole appears to be beneficial for stroke prevention. To this end, lead author Dr. David Strain (University of Exeter) highlighted the almost-immediate separation of the Kaplan-Meier curves between the GLP-1 and placebo groups, underscoring that GLP-1 therapy appears to have a much more rapid effect on stroke prevention than current standard-of-care therapies such as statins or anti-inflammatory agents. Interestingly, the effect of semaglutide (both injectable and oral) was stronger than that of liraglutide (HR 0.68 vs. HR 0.86).

• The study further analyzed the effect of GLP-1 agonists on the various stroke subtypes: ischemic, hemorrhagic, and undetermined. Within ischemic strokes, the investigators also analyzed further subtypes such as large artery atherosclerosis, cardio-embolism, small vessel occlusion, other determined cause, and undetermined cause. Given low event numbers for these different subtypes, none showed a statistically significant risk reduction with GLP-1 therapy. However, encouragingly, the point estimates for each stroke subtype favored GLP-1 agonists for all but ischemic strokes of undetermined cause (HR 1.25, 95% CI 0.34-4.65, p=0.74). Overall, GLP-1 agonists seemed to show the strongest effect for small vessel occlusion strokes (HR 0.78, 95% CI 0.59-1.02, p=0.07).

Post-Hoc Analysis of SUSTAIN 6 and LEADER Identifies the Extent to Which A1c, Blood Pressure, and Body Weight Lowering Mediates GLP-1 Renoprotection

SUSTAIN 6 and LEADER demonstrated that GLP-1 receptor agonists semaglutide and liraglutide, respectively, reduce the presence and severity of kidney disease in type 2 diabetes patients. In a new counterfactual mediation analysis of these studies, Dr. Johannes Mann (Friedrich Alexander University of Erlangen, Germany) identified the degree to which these renoprotective effects were indirect consequences of A1c, body weight, and systolic blood pressure lowering. Out of these three variables, the researchers identified A1c lowering as having the greatest effect on renal outcomes, mediating an estimated 26% and 25% of the renoprotective benefits of semaglutide and liraglutide, respectively. Notably, systolic blood pressure also had a meaningful effect, mediating an estimated 22% of the benefits in the semaglutide group although just an estimated 9% of the benefits in the liraglutide group. Lastly, body weight reduction was estimated to mediate 8% of the renal effects of semaglutide and 9% of those of liraglutide. Given these variable results, the researchers concluded that GLP-1 receptor agonist renoprotection can be attributed more so to direct mechanisms, including inhibition of the renin-angiotensin-aldosterone system (RAAS) and blunting of local inflammation, among others.

Pooled Analysis of SUSTAIN 6 and PIONEER 6 Shows Addition of CVD-Free Life Years After Treatment with Semaglutide

Dr. Jan Westerink  (University Medical Center Utrecht) presented findings of a pooled analysis of data from SUSTAIN 6 and PIONEER 6, showing that with the lifetime DIAL prediction model, adding semaglutide to standard of care can add CVD-free life years in people with type 2 diabetes. Enrolled patients were at high risk for CVD, and this study aimed to quantify the absolute benefit of incorporating semaglutide to standard therapy in people with and without established CVD at baseline. The free, open-access DIAL model, based on data from over 300,000 people with type 2 in the Swedish National Diabetes Registry, found a mean increase of 1.7 CVD-free life years (95% CI: 0.5 – 2.9). At baseline, the pooled cohort (n=6,480; 73% eCVD, average A1c 8.4%, mean diabetes duration 14.4 years) showed a HR of 0.76 (95% CI: 0.62 – 0.92) for 3-point MACE. There was a range of CVD-free life years gained, with the most benefit observed in patients with eCVD at baseline and ages 50-54 years old. An inverse relationship between age and CVD-free life years gained was observed, highlighting that semaglutide provides the most benefit if implemented early in the life course. With this, adding semaglutide to standard therapy was associated with a 6% mean absolute risk reduction in 10-year CVD risk (95% CI: 1.9-10%).

• For patients, these results mean that 10-year risk of being free from recurrent CVD events may to 21% from 27% if semaglutide becomes standard therapy. Dr. Westerink emphasized that these findings should be using in shared decision making when discussing lifestyle interventions and beginning pharmacotherapy. We would like to see more studies performed with the DIAL prediction model and wonder about its potential for clinical uptake. If future studies with different patient populations show greater benefits in risk reduction of CV events, and maybe even adverse renal outcomes down the line, we also wonder if these positive results will encourage further investment from health systems into making therapies like semaglutide more accessible. 

Select Q&A

Q: The greatest gain in life years is present in patients under 55 years old, but what proportion of patients was this in the study? Likely few, can you comment?

A: You’re correct, but for the current presentation, it isn’t necessary to have a large group of patients in this group because data from SUSTAIN informs the model. If there is a patient above 55, then the results still apply. More interestingly, would patients under age 50 be willing to take a drug for the rest of their life? Still, benefits in this group would be the largest.

Q: Do you have plans for cost effective analyses, and how will this translate to money saving for health authorities?

A: This is an important and provocative question and throws us to thinking about what cost effectiveness is. Is it money or life years? From an earlier study with statins in patients with CVD, there’s a large difference in what patients and GPs think is a worthy absolute benefit. In patients, there’s a wide distribution of what they think is a good benefit. We need to think about that before we go to guideline committees and economic evaluations, but I’m perfectly fine with those analyses in the future.

Results of LIRAFLAME Show No Change in Vascular Inflammation with Treatment of Liraglutide

Dr. Rasmus Ripa (Rigshospitalet Copenhagen) presented findings from the LIRAFLAME study, which was inspired by liraglutide’s lowering of CVD risk in the LEADER trial. The double blind, randomized trial (n=102; mean A1c ≥ 6.5%) examined the primary endpoint of atrial inflammation by FDG-PET/CT and the secondary endpoints of carotid inflammation (by FDG-PET/CT), coronary artery calcium score, and carotid thickness. There was no significant difference in the primary outcome of FDG uptake in arteries between the liraglutide and placebo-treated groups (p=0.53). Despite being exploratory given missed significance on the primary outcome, secondary outcomes also all missed significance. Dr. Ripa concluded that despite no difference in inflammation between treatment and placebo in the population with type 2 diabetes and no known CVD, exploratory analyses may support possible differences in patients with established CVD. He suggests future studies focus on this patient population to learn more about liraglutide’s systemic effects, which will hopefully yield clarity on the therapy’s mechanism of action.

EXSCEL Post-Hoc Shows Early Exenatide (AZ’s GLP-1 Bydureon) Therapy May Be Renal-Protective

Dr. Annemarie Van der Aart (University of Groningen, Netherlands) presented a post-hoc of the EXSCEL trial, indicating renal protective effects of AZ’s GLP-1 agonist Bydureon (exenatide once-weekly) for patients with T2D and higher risk of kidney disease. The analysis looked at eGFR slope and percentage change in UACR across subgroups with albuminuria or other risk markers of CKD progression. In patients with baseline UACR >100 mg/g, eGFR slope improved significantly with weekly exenatide vs. placebo. No significant difference was observed between the GLP-1 and placebo among participants with baseline UACR <100 mg/g. Notably, this renal benefit of Bydureon was consistent regardless of baseline eGFR, history of cardiovascular disease, RAAS inhibitor use, or systolic blood pressure. While these findings are promising for patients with type 2 diabetes at risk for advancing CKD (and we should note that all type 2s face substantial renal risks), Dr. Van der Aart emphasized that this analysis was hypothesis-generating and requires further confirmation in clinical trials. Additionally, the UACR data was only available for subjects with albuminuria, which may have introduced bias.

Real-World Study of ~9,000 Type 2s Suggests SGLT-2s Give Greater Risk Reduction for Three-Point MACE and Heart Failure vs. GLP-1s

Dr. Gian Paolo Fadini presented a real-world study (n=8,596) to illuminate how HCPs should choose between GLP-1 agonists and SGLT-2 inhibitors for patients with diabetes/high CV risk. He started with data on biomarkers: GLP-1s showed greater A1c-lowering efficacy vs. SGLT-2s, whereas SGLT-2s were significantly superior on systolic blood pressure reduction. Turning to hard outcomes, Dr. Fadini highlighted a directional signal that SGLT-2 inhibitors give greater protection against three-point MACE (non-fatal MI, non-fatal stroke, CV death) as well as heart failure, though neither finding reached statistical significance. We were surprised by this result, given the growing consensus among thought leaders that GLP-1s are more effective in reducing risk for atherosclerotic events (e.g. MI and stroke) while SGLT-2s are more effective in heart failure. Some guidelines explicitly recommend GLP-1s for patients with atherosclerosis, SGLT-2s for patients facing higher heart failure risk. Ultimately, the choice between a GLP-1 and an SGLT-2 should be personalized; studies like this one provide important insight, though it should be interpreted in the context of other real-world and RCT data as well. We should note that getting on either of these therapies will be meaningful for CV prevention, and we hope to see both drug classes expand substantially in the coming years.

Tirzepatide Modulation of Branched-Chain Amino and Keto Acids Links to Superior Effect on Insulin Sensitivity than Traditional GLP-1s

A study of tirzepatide on the metabolic profiles of people with type 2 diabetes showed that the GLP-1/GIP dual agonist, currently being developed by Lilly, is linked to significant changes in metabolites related to insulin sensitivity but not weight loss. Using data from tirzepatide’s phase 2b study in adults with poorly controlled type 2 diabetes (baseline mean A1c 8.2%, BMI 32.4 kg/m²), a metabolomic analysis was performed in n=259 of the original 315 participants to determine tirzepatide’s effects on metabolic profile and markers of insulin sensitivity and A1c. Mass spectrometry metabolomics was used to measure plasma samples for branched-chain amino acids and keto acids (BCAAs, BCKAs, respectively), as well as other catabolic products that may be associated with insulin sensitivity and weight loss. BCAAs and 2-hydroxybutryric acid were strongly associated with improvement in A1c. Tirzepatide treatment was associated with a dose-dependent effect on circulating BCAAs. With this, changes in levels of BCAAs and their associated catabolites correlated with HOMA2-IR and A1c, linking tirzepatide’s effects to insulin regulation, but not weight loss. Dr. Valentina Pirro (Duke University Medical Center) highlighted that tirzepatide significantly reduced levels of 2-hydroxybutyrate – a marker of insulin resistance and glucose intolerance. She concluded that the metabolic changes observed after treatment with tirzepatide were greater than those observed with dulaglutide. As tirzepatide is also being considered in obesity, we are curious to see further mechanistic studies on how tirzepatide is exerting its unprecedented weight loss effects. Only time will tell with upcoming phase 3 results for the drug.

Select Q&A

Q: Do you think these metabolites studies are indicators of hepatic sensitivity, or is it more peripheral sensitivity?

A: We want to have further studies to try to answer that question. All of the data in this analysis was retrospective and collected from plasma samples, so we’re currently studying to get more mechanistic understanding.

Q: Did you observe an increase in glycine, and did you measure GCSG index?

A: We did notice an increase in glycine, and levels seemed to dose dependently increase at the last time point (26 weeks) and was significantly greater compared to baseline, placebo, and dulaglutide modification. Dose dependency levels were higher in the intermediate time point (12 weeks), then plateaued at the end. I don’t have information available on the GCSG index.

Japanese RWE Study Shows 32% Reduced Risk of Death with GLP-1 Agonists, But No CV Protection

A new real-world analysis from the large scale receipt database of Japan indicated that people with diabetes treated with GLP-1 agonists had significantly lower mortality, but no reduced risk of a variety of cardiovascular outcomes. The study used insurance claims and health check-up data from 2011 to 2018, focusing on 2,401 people with diabetes who were prescribed a GLP-1 agonist and 53,946 who were on a different diabetes therapy. According to propensity score matching analysis, GLP-1 treated participants had a 22% lower risk of death than people with diabetes on a different therapy (HR 0.78, 95% CI 0.61-0.99, p=0.039). However, for all other outcomes examined, there was no significant benefit for GLP-1 therapy: 3-point MACE (HR 0.92, 95% CI 0.74-1.14, p=0.450), 3-point MACE and HF (HR 0.90, 95% CI 0.757-1.07, p=0.244), nonfatal MI (HR 0.70, 95% CI 0.45-1.09, p=0.111), nonfatal MI and HF (HR 0.91, 95% CI 0.73-1.13, p=0.382), and cerebrovascular events (HR 0.96, 95% CI 0.73-1.27, p=0.773).

• These real-world findings are inconsistent with past GLP-1 CVOTs, which on the whole demonstrated compelling cardioprotective benefits. The lead author, Dr. Masaya Koshizaka provided several potential explanations for this. First, the administered dose of liraglutide and dulaglutide, one of the more popular GLP-1 agonists in Japan, is half of the dose used in US and Europe. This could indicate that Japanese patients are not receiving the full therapeutic benefit of GLP-1 therapy. Second, the non-GLP-1 control group in this study included several people on SGLT-2 inhibitors, another cardioprotective therapy, thereby minimizing the differences between the treatment and control groups.  Other explanations could include a low number of cardiovascular events across the study, and a low percentage of study participants with baseline CVD. The latter point is particularly salient since CVOT data, particularly from EXSCEL, points to a greater effect of GLP-1 agonists for secondary as opposed to primary prevention.

• Dr. Koshizaka also pointed out several limitations in study design that could have impacted the results. Most notably, given the nature of the data reporting in the receipt database of Japan, it was not possible to factor in participants’ level of glycemic control or other important covariates, such as dyslipidemia, hypertension, smoking status, obesity, and lifestyle factors. The propensity score matching technique, which essentially matches patients with similar characteristics between the treatment and control groups, partially circumvents this problem, but indeed the results must be taken with a grain of salt in the absence of statistical adjustment for possible confounding factors.

• Methodological issues aside, people of Asian descent have been extremely under-represented in major GLP-1 CVOTs, and it may be the case that these agents are not as effective in this population. This will be critical to assess in future, larger scale trials. We hope this Japan-specific RWE inspires a larger-scale, multi-country analysis to more comprehensively assess the effects of GLP-1 agonists on cardiovascular events in the real world, similar to what CVD-REAL and CVD-REAL 2 have done for SGLT-2 inhibitors.

Semaglutide Tsunami: Novo Nordisk Symposium Spotlights Strength of Ozempic and Rybelsus Data Across Diverse Patient Populations, Points to Future Expansion Outside of Diabetes

In Day #1’s most-watched session, Drs. Tina Vilsbøll, Juris Meier, Richard Pratley, Ofri Mosenzon, Stephen Bain, Alice Cheng, and David Strain unleashed a deluge of clinical trial data in support of GLP-1 semaglutide as a cornerstone for the treatment of type 2 diabetes. While the KOLs presented data from both the phase 3 SUSTAIN and PIONEER programs for Ozempic and Rybelsus, respectively, much of the session centered on the safety and efficacy of the oral option. Considering Rybelsus’s recent launch in the UK, we were not surprised by this focus and look forward to following the drug’s upcoming rollout across Europe.

Despite similar levels of prioritization for GLP-1s and SGLT-2 inhibitors in current guidelines, a live audience poll revealed that the vast majority of EASD attendees still prefer prescribing an SGLT-2 inhibitor over a GLP-1. In response to these results, Dr. Richard Pratley stated that he was “not really surprised” by the numbers, given the greater ease of an oral medication. However, he expects Rybelsus will help close this divide in the near future, especially as the drug is further utilized in primary care.

• Using data from PIONEER 1 through 8, Dr. Stephen Bain (Swansea University) emphasized Rybelsus’s efficacy across diverse patient populations – divided by baseline A1c, diabetes duration, age, background medication, and race/ethnicity. While Rybelsus, as with most glucose-lowering therapies, did demonstrate greater A1c reductions for those with higher values at baseline, the oral GLP-1 showed clinically relevant reductions regardless of any of the aforementioned factors. In an interesting point, Dr. Bain noted that while the number of adverse events was similar irrespective of age, older patients generally had higher discontinuation rates than their younger peers. In our purview, these findings suggest that improving titration guidelines for older patients may improve uptake and adherence in this population.

• Dr. Ofri Mosenzon (Hebrew University Medical School) used her rapid-fire talk to highlight semaglutide’s efficacy in two vulnerable patient populations: those with renal impairment and those with late-stage diabetes. Citing data from PIONEER 5, Dr. Mosenzon commented on the “beautiful” declines in A1c and body weight seen in patients with moderate renal impairment on oral semaglutide. While some audience members expressed hesitation with prescribing GLP-1s at lower eGFR levels (see poll below), Dr. Mosenzon assured them that the drug was safe to use in eGFRs as low as >15 ml/min/1.73m². To note, SGLT-2 inhibitor canagliflozin can only be prescribed to patients with an eGFR of >30, despite being indicated to treat diabetic kidney disease, making GLP-1s an attractive treatment option for patients with renal impairment. For the strongest evidence on GLP-1s and potential renal protection, however, we’re waiting on results from Novo Nordisk’s FLOW trial of semaglutide in CKD – set to complete in August 2024.

• Rounding out the session, Dr. David Strain (University of Exeter Medical School) unveiled highly informative data on the effects of GLP-1s on stroke. Combined data from SUSTAIN 6, PIONEER 6, and LEADER showed that GLP-1s significantly reduce the risk of stroke in patients with type 2 at high CV risk. Impressively, the Kaplan-Meier curves diverged almost instantaneously for patients on GLP-1s vs. placebo, suggesting an early treatment benefit. In comparison, while PCSK9 inhibitor Repatha (evolocumab) also demonstrated a 21% relative risk reduction on stroke in the FOURIER trial, this benefit took almost four years to become apparent. Looking ahead, Dr. Strain alluded to possible expansions outside of diabetes from the SELECT CVOT in people with overweight or obesity. In a similar talk at ESC 2020, Dr. Kim Connelly suggested that findings from SELECT would also yield key insights on GLP-1 benefits in prediabetes. Pending positive results, we imagine that a distinct CV benefit outside of diabetes would significantly increase uptake across a number of specialties.

Debating the Most Effective First- and Second-Line Therapies for Type 2 Diabetes: Insights from Nephrology, Cardiology, Endocrinology, and Primary Care

As part of Lilly/BI’s ACROSS T2D program – an interactive educational resource platform for HCPs who care for type 2 diabetes patients – this industry-sponsored symposium addressed three critically important medical questions: (i) should metformin remain the foundational therapy for patients with type 2? (ii) are SGLT-2 inhibitors preferred over GLP-1 receptor agonists for second-line therapy, or vice versa? And (iii) which organ is deserving of greater attention in the context of type 2, the heart or the kidney? Framed as a series of debates, this session gave endocrinologists Dr. Silvio Inzucchi (Yale School of Medicine) and Dr. Francisco Giorgino (University of Bari Lado Moro, Italy); cardiologist Dr. Nikolaus Marx (University Hospital Aachen, Germany); and endocrinologist and nephrologist Dr. Merlin Thomas (Monash University, Australia) the (virtual) stage to present their expert opinions on each topic. Dr. Sarah Jarvis (Patient Platform Limited, United Kingdom) wrapped up the symposium by providing her perspective on each of these questions as a primary care physician, paying particular attention to tremendously relevant topic of shared decision making between patient and provider. See below for the most significant take-aways from each debate.

• Metformin as foundational therapy. Dr. Giorgino cited data from VERIFY, UKPDS 34, and two smaller-scale CVOTs (by Kooy et al. and Hong et al.) to demonstrate metformin’s: (i) robust and sustained effects on glucose-lowering; (ii) potential CV and mortality benefit; and (iii) excellent safety profile alone or in combination with other therapies (e.g. with DPP4 inhibitor vildagliptin in VERIFY). He also pointed out various confounding factors within the large-scale GLP-1 receptor agonist and SGLT-2 inhibitor trials. Not only do they fail to enroll sufficient numbers of low-risk patients, but the majority of patients in both placebo and treatment arms of the trials are on background metformin therapy to begin with. According to Dr. Marx, however, this latter point is of little interest: across this library of trials, patients receive similar CV benefits irrespective of metformin background therapy. In CANVAS, the presence of metformin background actually blunted canagliflozin’s relative risk reduction for hospitalization for heart failure (hHF) and fatal and non-fatal stroke. Primary care physician Dr. Jarvis appeared to side with Dr. Giorgino in her commentary, believing there was “no question” that metformin gives rise to meaningful CV risk reduction and improved glycemic control. If used early on, metformin’s effect on glucose lowering can significantly reduce incidence of microvascular, macrovascular, and mortality events.

• SGLT-2 inhibitors versus GLP-1 receptor agonists as second-line therapy. While both agents have been shown to have significant benefits for glycemic control, CV risk, and cardiorenal health, among other characteristics, Dr. Inzucchi and Dr. Giorgino took opposing sides on which class should be preferred as second-line therapy. Dr. Inzucchi advocated for SGLT-2 inhibitors, paying particular attention to the drug class’ consistent benefits in hHF. Heart failure, as Dr. Inzucchi reminded the audience, is the complication associated with the single greatest excess risk in type 2 diabetes and is associated with a significantly worse clinical course (relative to that of diabetes without heart failure). SGLT-2 adherence was also shown to be superior to that of GLP-1 receptor agonists in a small BMJ study, presumably due to route of administration (oral versus injected, respectively). On the opposition, Dr. Giorgino pointed out than GLP-1 receptor agonists lower A1c to a greater extent than SGLT-2 inhibitors; as shown in a meta-analysis of GLP-1 CVOTs, the extent of A1c reduction has been directly and significantly related to the magnitude of 3P-MACE and stroke risk reductions. In his argument, Dr. Giorgino also advocated for the utility of GLP-1 receptor agonists in providing cellular protection (for beta cells and neurons) and in active COVID-19 infections. An opinion summary published in the The Lancet Diabetes & Endocrinology suggested that clinicians discontinue SGLT-2 inhibitors upon hospitalization for COVID-19, conversely. The adverse GI side effects of GLP-1s were brought up in both presentations, although Dr. Giorgino that these events can be minimized by dose titration and smaller meals.

  • On this question, Dr. Jarvis took the opportunity to highlight the importance of shared decision making in contemporary medicine. As both SGLT-2 inhibitors and GLP-1 receptor agonists are reasonable therapy options for type 2 diabetes patients, patients and clinicians should work together to reach a medical decision harmonious with the clinician’s knowledge and the patient’s experience with the condition.

Focusing on cardiac or renal risk reduction in type 2 diabetes. In advocating for a cardiac focus for type 2 diabetes care, Dr. Marx echoed many of the facts that Dr. Inzucchi also presented concerning the relationship between heart failure and diabetes. He showed that irrespective of age, heart failure is one of the most significant, frequent, and earliest CV complications associated with type 2 diabetes. Heart failure events reduce quality of life and are associated with future hospitalizations and continuous declines in cardiac function. On the opposition, Dr. Thomas advocated for a renal focus in type 2 diabetes. Not only is chronic kidney disease (CKD) easily detectable using eGFR or urine albumin-to-creatinine ratio (UACR), but it is also more treatable than heart diseases using anti-hypertensive treatment (e.g. losartan) or SGLT-2 inhibitors, for example. In fact, in EMPA-REG OUTCOME, empagliflozin slowed the rate of eGFR decline in patients regardless of renal disease at baseline, suggesting a preventative effect of these treatments as well. Treating renal disease has tremendous implications for long-term health, as Dr. Thomas suggested, reducing incidence of adverse drug reactions, heart failure, end-stage kidney disease, and mortality, among other events. As shown in the NHANES III US population-based study, the presence and severity of CKD was directly proportional with extent of excess mortality risk. Dr. Jarvis similarly highlighted the interplay between the heart and the kidney in her later presentation, showing that treating one organ may have the unintended benefit of treating the other.

Novartis Symposium Spotlights VERIFY Study; Improved Glucose Control & Beta Cell Function with Early Combination Therapy vs. Metformin Monotherapy

A Novartis-sponsored symposium offered a deep dive on VERIFY, a large RCT (n=2,001) investigating early combination therapy with DPP-4 inhibitor Galvus (vildagliptin) + metformin in type 2 diabetes. Results were first presented at last year’s EASD gathering, and were summarized this year by Professor Helmut Brath, who focused especially on beta cell function. All VERIFY participants were relatively early in their disease progression (≤2 years, drug-naïve, A1c <7.5%) and vildagliptin/metformin together seemed to preserve beta cell function: A mean change of +17.21 (SD=9.04) in HOMA-B was observed in the early combination group vs.  -2.02 (SD=9.02) in the metformin monotherapy group. VERIFY also showed 49% risk reduction for time to initial treatment failure (HR=0.51) and 26% risk reduction for time to second treatment failure (HR=0.74) with combination treatment vs. metformin alone (both p<0.0001). While Galvus itself makes up only ~13% of the global DPP-4 inhibitor market, the implications of VERIFY likely extend well beyond one product. Professor Matthias Blüher acknowledged the ongoing debate over whether early combination therapy should replace the stepwise approach to diabetes care. He contended that VERIFY could fill the evidence gap, as it shows clear benefits to dual therapy with a DPP-4 over metformin monotherapy early in the course of type 2 diabetes. We recall similar commentary from Professor Chantal Mathieu at EASD 2019. We suspect cost-effectiveness data from VERIFY (or a similar trial of sequential vs. stepwise treatment) will be helpful in changing diabetes care guidelines to favor initial intervention with a combination of drugs.

• VERIFY also provides insight on differences in outcomes in people with young- vs late-onset diabetes. Professor Julianna Chan emphasized the emerging epidemic of young-onset type 2 diabetes and the current gap in knowledge and evidence when it comes to treatment. VERIFY found that initial treatment failure is more likely in people with young-onset compared to late-onset type 2. In the young-onset group, 50.5% had initial treatment failure with combination therapy compared to 73.3% with monotherapy, whereas in the late-onset group, 42.9% had initial treatment failure with combination therapy compared to 61% with monotherapy. From this higher baseline risk, there was greater relative risk reduction in secondary treatment failure in the young-onset group (48%) compared to the late-onset group (24%).

• Translating VERIFY results into real-world practice will require an integrated strategy. Professor Chan highlighted the need for a team approach involving physicians and non-physician staff focused on early detection and personalized treatment, with an emphasis on people at high risk for type 2 diabetes.

Select Questions and Answers:

Q: Can you elaborate one the gender imbalance seen in clinical trials, particularly diabetes trials?

Professor Brath: This is one of the great strengths of the VERIFY study. The VERIFY study included 53% of women and 47% of men, and this reflects real life, and I am not aware of any other major diabetes landmark study which could include such a percentage of women. Even in contrast, the other studies included a much lower part of women in the majority of the historical studies. But also in the new studies, there are 30, 40% of women included which I think we should consider when interpreting the study. VIRTA-CV or DAPA-CKD are very new studies, published in the last weeks. They also included only 30, 40% of women. I think VERIFY shows that in a landmark study it is doable to include a well-balanced gender recruitment, and by doing so I think it is a model for future studies.

Q: Could you tell us how many patients dropped from the study because of failure of combination therapy with higher A1c above 8% or 9%? 

Professor Chan: There were a number of patients who were not put on insulin but were put on other drugs because this is a multinational study and there are inter-site differences in terms of access to insulin, coverage for insulin, and also initiation and acceptance of insulin by the practitioners and patients. However, we know once the patient’s A1c is repeatedly greater than 7%, they call for escalation of treatment. Also in the young onset diabetic group there are about 30% who continue treatment, compared to 15% in the late onset group. The higher discontinuation rates in the young-onset diabetes group reflects in part that many of these people probably need an additional third drug in order to maintain glycemic control but might not be accepting the insulin. Also, many of these individuals who dropped also because of personal, because of work reasons, or pregnancy. So, I think this really highlights the challenges in managing people with young-onset diabetes, it’s not just the medical need, sometimes the social and also the personal needs. 

Q: Why select the time to therapeutic failure as the first descent and not the effectiveness for the reduction of glycated hemoglobin as the primary outcome?

Professor Brath: Time to failure is referring to the statistical technique to measure the proportion of subjects out of control as time goes by during the study. So A1c is the main endpoint, but instead of measuring mean A1c values, the proportion of patients with an A1c value above 7% has been measured.

Q: In pharmacoeconomic terms, do you consider combination therapy to be cost effective?

Professor Chan: I think in most countries $1 in $8 of healthcare is actually spent in diabetes, but 80% of this healthcare cost are due to healthcare complications. Once complications develop, healthcare costs increase by 3-9% depending on the complication, particularly kidney failure and leg amputations are extremely costly complications. One of the driving reasons for escalating healthcare costs is lung disease deterioration and aging. And in that regard, I think the VERIFY results have clearly demonstrated the benefits of the early combination therapy to reduce that glycemic burden over time. Because of the glucotoxicity, the beta-cell failure would escalate, and so by controlling glucose earlier on, we will preserve beta-cell function, reduce glycemic burden, and reduce the long-term complications. So, I think this investment is going to be really important and this is particularly so in the young to middle age people because we are not just talking about direct health costs, but of course indirect healthcare costs. So, taking all these into consideration, I think early glycemic control is really important in terms of reducing long term complications. I think diabetes is easy to treat early. When you delay it, the treatment becomes more complex and less cost-effective. So, we now really have the evidence for us to identify these patients early and intervene early to get long term benefits.

Q: Our local truck agency has not wanted to accept this indication, despite the clear benefits shown by the VERIFY study. In addition to what is related to meta-analysis on the benefits of early control, they generally refer to us that metformin can achieve good results in a large majority of patients, so early combination is not necessary. What would you answer?

Professor Blüher: The VERIFY study clearly showed that early combination treatment halved the risk of time to initial treatment failure compared to metformin monotherapy. There are consistent data demonstrating that sequential intensification of OAD treatment in patients with type 2 diabetes leads to a delay in achieving HbA1c targets. Only one year of a failure to intensify anti-diabetes medications from metformin to combination therapies may be associated with and increased risk for myocardial infarction, stroke, heart failure, and other cardiovascular endpoints. It has now been clearly shown that, not only for patients who initially present with a high HbA1c value, early combination therapy may increase the durability of glycemic control if you start early enough.

GLP-1 Agonist and DPP-4 Inhibitor Poster Highlights

Title	Authors	Details and Implications
Glycaemic and body weight responses to oral semaglutide in the PIONEER trial programme	Tina Vilsboll, Erik Christiansen, Christin Hertz, Linda Mellbin, Morten Abildlund Nielsen, Vincent Woo, Kathleen Dungan	Post-hoc analysis evaluated the clinical response of type 2 patients to oral semaglutide (14mg) in comparison to empagliflozin, sitagliptin, liraglutide, and placebo in the PIONEER 1-5 and 8 trials (n=3,506). A higher proportion of patients on oral semaglutide (89-95%) exhibited reductions in A1c than those on placebo (51-64%), sitagliptin (82%), empagliflozin (86%), and liraglutide (88%). 72-86% of patients treated with oral semaglutide showed at least some reduction in A1c and body weight. The odds of this outcome were significantly higher with oral sema vs. placebo (p<0.0001) in PIONEER 1, 4, 5, and 8 and vs. DPP-4 inhibitor sitagliptin (p<0.0001) in PIONEER 3. Patients receiving oral sema (14mg) were more likely to achieve an A1c reduction of 1% and body weight loss of 5% than all other comparator therapies.
LIRA-PRIME: a randomised trial in primary care settings of liraglutide versus OAD for glycaemic control in patients with type 2 diabetes not in control on metformin	Marouan Zoghbi, Margit Kaltoft, Devayani Klhe, Jayant Panda, Mehmet Sargin, Benjamin Ole Wolthers, Jeff Unger	LIRA-PRIME was a phase 4 trial assessing liraglutide’s glycemic control maintenance (n=996) vs. oral drugs including SUs, TZDs,  DPP-4s, and SGLT-2s (n=995) in type 2 patients on background metformin. Liraglutide demonstrated longer maintained glycemic control than OADs (median 109 weeks vs. 65 weeks, p<0.0001) when added to metformin. Median time to premature discontinued treatment was longer for liraglutide than OADs (80 weeks vs. 52 weeks, p<0.0001). While the rates of adverse and hypoglycemic events were similar across liraglutide and OADs, more patients discontinued treatment due to GI side-effects with liraglutide vs. OADs (6% vs 1%).
Optimal treatment intensification for glycaemic control in patients with type 2 diabetes on two oral agents: real world comparison of GLP-1, OADs and insulin	Cyrus Desouza, Andreas Ross Kirk, Kamal Kant Mangla, Michael Lyng Wolden, Ildiko Lingvay	PATHWAY was a retrospective, observational study comparing treatment intensification with additional oral antidiabetic drugs (OADs), GLP-1s, or insulin in type 2 diabetes patients already on two OADs. Patients receiving a GLP-1 agonist (most commonly liraglutide, exenatide, or dulaglutide) were significantly more likely to achieve A1c <7% and meet all body weight targets than patients prescribed additional OADs (most commonly DPP-4s, sulfonylureas, or SGLT-2s) or insulin. Twice as many patients intensified with GLP-1s vs. OADs discontinued treatment with at least one of the baseline OADs, suggesting that GLP-1s offer greater glycemic + weight loss efficacy and can simplify medication regimens.
Efficacy and safety of once-weekly dulaglutide versus insulin glargine in Chinese patients with type 2 diabetes and different baseline glycaemic patterns	Rui Wang, Jianing Hou, Bin Zhang, Qifu Li	AWARD-CHN2 measured the effects of Lilly’s GLP-1 agonist Trulicity (dulaglutide) vs. Sanofi’s basal insulin Lantus (glargine) in Chinese patients with type 2 diabetes. Both dulaglutide and insulin glargine improved glycemic control regardless of baseline fasting glucose or postprandial glucose. In most baseline glycemic patterns (except the low fasting glucose/high postprandial glucose subgroup), dulaglutide (1.5mg) yielded significantly greater A1c reduction vs. glargine. The proportion of patients achieving A1c <6.5% was also greater with dulaglutide. In all subgroups, dulaglutide resulted in weight loss and insulin glargine resulted in weight gain, consistent with the known profiles of GLP-1 agonists and basal insulins.
Dulaglutide reduces HbA1c irrespective of antihyperglycaemic agents, duration of diabetes, BMI and weight loss: a post hoc analysis from the REWIND trial	Denis Xavier, Anita Kwan, Hertzel Gerstein, Jan Basile, Juan Maldonado, Sohini Raha, Manige Konig	This post-hoc analysis of REWIND found consistent A1c decline with Lilly’s GLP-1 agonist Trulicity (dulaglutide) irrespective of background diabetes medication, BMI, diabetes duration, comorbidities, and weight loss. A1c drop was significant at all time points between the treatment vs. placebo arms (p<0.001) and was durable for 72 months follow-up.
Effect of dulaglutide on kidney function-related outcomes in type 2 diabetes: post hoc analysis from the REWIND trial	Jonathan Shaw, Fady Botros, Raleigh Malik, Charles Atisso, Helen Colhoun, Hertzel Gerstein	This post-hoc analysis focused on renal outcomes from REWIND and found that dulaglutide (Lilly’s Trulicity) may slow DKD progression in type 2 patients with increased CV risk. Dulaglutide was linked to a 17% relative risk reduction of eGFR decline >40%, end-stage renal disease, or all-cause death. Dulaglutide also gave a 30% relative risk reduction for sustained eGFR decline >40%.
Patient-reported outcomes in patients with type 2 diabetes treated with investigational dulaglutide doses added to metformin (AWARD-11)	David Cox, Zhuoxin Yu, Angelyn Bethel, Kristina Secnik Boye, Reema Mody	This poster shared data on key patient-reported outcomes (PROs) from the AWARD-11 trial (n=1,842), Lilly’s investigation of high-dose dulaglutide (3 mg and 4.5 mg) compared to what’s currently on the market as Trulicity (1.5 mg). PRO scores on weight perception significantly improved for all doses of dulaglutide, with the change being higher in the 3 mg and 4.5 mg dulaglutide groups vs. the 1.5 mg group. The presenters suggested that higher-dose dulaglutide may positively influence patient perception. Patients reported a consistently positive experience with the dulaglutide injection device (designed by IDEO), regardless of dosage.
Cardioprotection with dulaglutide is not depending on baseline therapy with metformin. A subgroup analysis of the REWIND trial	Giulia Ferrannini, Hertzel C. Gerstein, Helen M. Colhoun, Gilles R. Dagenais, Rafael Diaz, Leanne Dyal, Mark Lakshmanan, Linda Mellbin, Jeffrey Probstfield, Matthew C. Riddle, Jonathan E. Shaw, Lars Rydén	A subgroup analysis of REWIND examined whether the effect of dulaglutide on CV events varies with background metformin. Analysis found no significant interaction, i.e. CV benefits of dulaglutide were consistent in patients with and without metformin as baseline therapy.
Cardiovascular outcomes in patients with type 2 diabetes and reduced eGFR and albuminuria: a REWIND post hoc subgroup analysis	Helen Colhoun, Raleigh Malik, Fady Botros, Charles Atisso, Hertzel Gerstein	This subgroup analysis of REWIND found consistent CV benefits to GLP-1 dulaglutide regardless of baseline eGFR or albuminuria status. Data suggests dulaglutide is safe for patients with type 2 and reduced eGFR or albuminuria.
Efficacy of investigational dulaglutide doses overall and by baseline HbA1c and BMI: exploratory subgroup analyses of the AWARD-11 trial	Enzo Bonora, Juan Frias, Luis Nevarez Ruiz, Zhuoxin Yu, Zvonko Milicevic, Raleigh Malik, Angelyn Bethel, David Cox	Various subgroup analyses of the AWARD-11 trial were summarized on this poster. Regardless of baseline A1c or BMI, dose escalation from dulaglutide 1.5 mg to dulaglutide 3.0 mg or 4.5 mg resulted in improvements to glycemic control and body weight. Safety/tolerability of high-dose dulaglutide was acceptable and similar between A1c and BMI subgroups.
Liraglutide and semaglutide reduce cardiovascular events in patients with type 2 diabetes and peripheral arterial disease	Subodh Verma, Søren Rasmussen, Hans A Saevereid, Maria Sejersten Ripa, on behalf of the LEADER and SUSTAIN 6 trial investigators	Post-hoc analyses of LEADER and SUSTAIN 6 found greater absolute risk reduction for three-point MACE in patients with baseline PAD vs. those without. Logical finding given the higher baseline CV risk associated with peripheral arterial disease; implications for GLP-1 agonist therapy in diabetes/PAD.
Safety and efficiency of the combination of basal insulin with liraglutide in hospitalised patients with type 2 diabetes	Styliani Papantoniou, Athanasia Papazafiropoulou, Maria Rallatou, Elias Georgopoulos, Eleutheria Kagkelari, Konstantina Anagnostopoulou, Panagiotis Kouvatsos, Alexandros Kamaratos, Andreas Melidonis, Stavros Antonopoulos	This study (n=78) investigated the safety and efficacy of combining basal insulin and GLP-1 liraglutide for hospitalized patients with type 2. Combination of basal insulin + liraglutide was safe and effective for achieving better glycemic control.
Superior effect of 1-year treatment with GLP-1 receptor agonist and exercise on weight loss maintenance and body composition after a very low-calorie diet: the S-LITE trial	Signe Soerensen Torekov, C. Janus, J. Lundgren, S.B.K. Jensen, C.R. Juhl, M.B. Blond, L. Olsen, R.M. Christensen, M.S. Svane, T.Q. Bandholm, K.N. Bojsen-Møller, J.-E.B. Jensen, B.M. Stallknecht, J.J. Holst, S. Madsbad	S-LITE (n=215) investigated the efficacy of Novo Nordisk’s obesity drug Saxenda (liraglutide 3.0 mg) delivered alongside a moderate-to-vigorous exercise program on weight loss maintenance. The combination of exercise + Saxenda led to greater weight loss (-9.5 kg) vs. exercise alone (-4.1 kg) and liraglutide alone (-6.8 kg). Fat mass decreased in all treatments, while lean mass increased in the exercise alone treatment only. See detailed write-up in our ADA 2020 coverage.
Effects of lixisenatide versus liraglutide on esophageal functions and motility in patients with type 2 diabetes	Daniel R. Quast, M.A. Nauck, N. Schenker, B.A. Menge, C. Kapitza, J.J. Meier	This study (n=57) compared a short- and long-acting GLP-1 agonist – Sanofi’s lixisenatide and Novo Nordisk’s liraglutide, respectively – on esophageal and gastric function in patients with type 2. Neither lixisenatide nor liraglutide had significant effects on gastroesophageal reflux, further supporting the GI safety profile of GLP-1s.
Short-acting exenatide and markers of cardiovascular disease in type 1 diabetes: a randomised double-blinded placebo-controlled trial	Nicklas J. Johansen, Thomas F. Dejgaard, Asger Lund, Camilla Schluntz, Emil L. Larson, Henrik E. Poulsen, Jens P. Goetze, Holger J. Moller, Tina Vilsboll, Henrik U. Andersen, Filip K. Knop	This study evaluated short-acting GLP-1 receptor agonist exenatide as an adjunctive to insulin therapy in type 1 diabetes patients with high cardiovascular risk. The study was a double-blinded, parallel-group trial in which participants were injected with exenatide or placebo in addition to insulin therapy three times a day for 26 weeks (n=108). Exenatide precipitated reductions in total fat mass by 2.6 kg and lean body mass by 1.1 kg. The researchers concluded that exenatide was an effective therapy for weight loss in type 1 diabetes without increasing cardiovascular risk.
Effect on beta cell function in newly diagnosed type 2 diabetes patients after treatment with vildagliptin and metformin: results from the VERIFY study	Päivi M Paldánius, Stefano Del Prato, Michael Stumvoll, David Matthews	VERIFY examined the effects of early combination therapy with DPP-4 inhibitor vildagliptin + metformin in recently-diagnosed type 2 diabetes. Compared with metformin monotherapy, the combination regimen showed improvement in beta cell function.
Effects of teneligliptin on continuous glucose monitoring-derived time-in range and glycaemic variability in patients aged 65 years and older with type 2 diabetes	GyuRi Kim, Ji Cheol Bae, Ji Eun Cha, Jong Chul Won, Soo Heon Kwak, BokJin Hyun, Jae Hyeon Kim	This RCT evaluated the effects of teneligliptin (Daiichi Sankyo’s DPP-4 inhibitor Tenelia) on time-in-range (TIR) and hypoglycemia in type 2 patients age 65+. After 12 weeks, teneligliptin (20 mg) gave significantly superior A1c reductions vs. placebo (-0.76%, p<0.0001). Fasting plasma glucose also declined more with the DPP-4 vs. placebo (-14.1 mg/dl, p=0.0088). CGM (iPro2, Medtronic) revealed that teneligliptin improved TIR without increasing time below range: (i) TIR increased from 62.7% to 82% with teneligliptin; (ii) time above range was significantly reduced in the teneligliptin group when compared to placebo (-8%, p=0.0016 in the Level 1 hyperglycemic range and -5.7%, p = 0.0039 in the Level 2 hyperglycemic range); (iii) there were no significant differences in time below range (hypoglycemia) between groups. The presenters suggest teneligliptin may be a beneficial therapeutic choice for older patients with type 2 diabetes.

SGLT-2 Inhibitor Highlights

Late-Breaking Analysis of VERTIS-CV Supports Consistent Renal Benefit of SGLT-2 Inhibitor Across Entirety of Class

University of Toronto’s Dr. David Cherney elegantly detailed results from a pre-specified exploratory analysis of VERTIS-CV for SGLT-2 inhibitor Steglatro (ertugliflozin), demonstrating the drug’s 34% relative risk reduction on composite (i) sustained ≥40% decrease in eGFR; (ii) dialysis/kidney transplant; or (iii) renal death (HR=0.66; 95% CI: 0.50 to 0.88, p<0.01). As a reminder, when full results for VERTIS-CV were first unveiled at ADA 2020, ertugliflozin did not meet statistical significance on its renal endpoint of composite (i) doubling of serum creatinine; (ii) renal dealth; or (iii) dialysis/kidney transplant, though the drug trended toward benefit (HR=0.81; 95% CI: 0.63 to 1.04, p=0.08). That being said, the majority of KOLs seemed confident in ertugliflozin’s consistent renal benefits with the overall SGLT-2 inhibitor (as shown by comparable eGFR declines). Nephrologist Dr. Per-Henrik Groop at ESC 2020 later elaborated that VERTIS-CV’s more rigorous composite renal endpoint (doubling of serum creatinine translates to a 57% reduction in eGFR) could be a driver of the statistical insignificance. Today’s results indeed confirm this hypothesis, and it is positive from our view to now have clear evidence of ertugliflozin’s equally strong renal benefits though we are concerned to some degree that the initial narrative prompts at least some fixed thinking. Other pre-specified analyses showed that relative risk reduction for the updated renal composite was consistent across CKD stages (1-3), UACR category, and KDIGO CKD risk score.

• In addition, UACR was significantly reduced across a range of albuminuria levels at baseline, with greater effects in patients with macroalbuminuria. While ertugliflozin conferred an acute decline in eGFR (as is common with the class), eGFR levels were better maintained over time vs. placebo.

• Following Dr. Cherney’s presentation, Dr. Darren McGuire (UT Southwestern) contextualized the updated renal findings in a series of valuable meta-analyses using data from EMPA-REG for Lilly/BI’s empagliflozin, CANVAS and CREDENCE for J&J’s canagliflozin, DECLARE-TIMI 58 for AstraZeneca’s dapagliflozin, and VERTIS CV. Dr. McGuire’s previous presentation at ADA 2020 established relative heterogeneity on 3P-MACE and consistency on hHF within the SGLT-2 inhibitor class. Using the “harmonized” kidney outcomes, heterogeneity between the different drugs on renal benefit goes to zero, demonstrating ertugliflozin’s equally attractive renal profile. Excitingly, Dr. McGuire shared during Q&A that the primary manuscript for VERTIS-CV will be published “very soon,” followed by publications on renal results, HF results, and the meta-analysis “in very short order.” One manuscript appeared today in NEJM as noted above – this, however, did not appear to have the new data or approach – see “Cardiovascular Outcomes with Ertugliflozin in Type 2 Diabetes” by Dr. Darren McGuire, et. al.

• The session chair, Professor Melanie Davies, conducted a fascinating Q&A, before which the attendees appeared to be generally rapt – we certainly were. The “new” analysis seems to have turned the tide on whether renal benefit was seen – this is a big turnaround and we’ll be eager to hear more expert views. The strength of the trial did already appear to be expert consensus when it was announced – see KOL views in both our Pfizer 2Q20 and Merck 2Q20 coverage.

Question & Answer (Select questions – more will be added in our full report)

Prof. Melanie Davies (Session Chair): In terms of the baseline characteristics, the risks are pretty similar to what we see in EMPA-REG, but if we look at the pre-defined primary endpoints, there clearly was not a positive endpoint in terms of MACE, CV death, or composite CV death/hHF. What are your thoughts on why we saw these differences?

Dr. Cherney: In terms of trying to understand the place of VERTIS-CV among the other trials, it is interesting to note that looking at eGFR loss over time, both in the ertugliflozin and placebo treated groups, compared to the other CVOTs – numerically, of course, there has been no formal statistical testing done to compare the trials – but just looking at the changes in GFR over time, it looks like loss of kidney function is somewhat attenuated in the VERTIS-CV group, suggesting that it may, for whatever reason, be a lower renal risk group than other trials. That may, at least in part, explain some of the difference in terms of outcomes. Because a high-risk renal group may lead to risk around CV factors as well, so that’s one potential reason, but really nothing else stands out as being differentiated from other trials.  

Dr. Davies: Just quickly, someone commented on ethnic risk. Does someone want to comment on the ethnic mix? 

Prof. Bernard Charbonnel: There is a small difference in the baseline race or ethnicity between EMPA-REG and VERTIS. There were 22% of Asians in EMPA-REG, and only 6% in VERTIS. In the subgroup analysis of EMPA-REG, the Asian subgroup showed greater benefit than the other ones. I don’t know if we can draw any conclusions from that, but there is a small difference in ethnicities. (Editor’s note – that seems a meaningful difference to us that we’d love to know more about – we hope to see more stratified data over time in major RCTs.)

Prof. Davies: David, around the renal outcomes and choice of doubling of serum creatinine vs. 40% reduction in eGFR. What’s your comment on the renal endpoints?

Dr. Cherney: Thanks for the question. I think that’s a really important point. Importantly, in terms of the 40% reduction in eGFR, that’s a very clinically important decline over time. It’s perhaps more amenable to a lot of the trials we have been discussing, involving cohorts of patients that are not renal cohorts. This is a cardiovascular cohort, who don’t necessarily have baseline CKD defined by GFR levels or albuminuria. Less than a third of patients had CKD at baseline. I think the use of a bit of a lower bar to define significant renal decline is very reasonable and important in this kind of trial and allows us to detect a level of eGFR loss that is important, but that can also be achieved in the relative limited time frame of two to four years, where one may not necessarily expect to see a loss of half of your kidney function or more. So, I think that’s a really important point, and I think 40% is a good level for these kinds of studies and is being used by other studies that are either ongoing or have been reported in this field.

Prof. Davies: There are a number of questions around how you choose an SGLT-2 inhibitor in routine practice, particularly in those with established CVD. Darren, would you like to start?

Dr. McGuire: Thanks for the question, and I very much like the tone. It’s not if you use an SGLT-2 inhibitor, but which one do you choose. So, we’re making progress around the world clinically. These are drugs that should be prescribed for patients with type 2 diabetes and ASCVD. I think it’s a little less important which one you choose. In the states, where payment for the drugs varies from one program to the other, for many, the one most important consideration is which one the patient can afford. I think the product labels give us some guidance, and the guidelines and society recommendations lean on those. The drugs should be preferentially used that have proven clinical benefits, and whether you’re targeting MACE, HF, kidney preservation, or all three, may inform the decision. My preference is to use one of the drugs, and we have evidence basis for all of them. I’m particularly enthused that the hHF is so consistent across the class and the kidney preservation is so consistent across the class, unlike the GLP-1s which primarily affect atherosclerotic outcomes.

EASD SVP Prof. Chantal Mathieu Gives Expert Commentary on DAPA-CKD: Calls for SGLT-2 Inhibitors Usage Down to eGFR of 25, Specialists to Unite and Create Clear Guidelines for PCPs; Full Results Published in NEJM

Thursday was headlined by an excellent overview of the DAPA-CKD trial for AZ’s SGLT-2 inhibitor Farxiga (dapagliflozin) in chronic kidney disease, with or without type 2 diabetes, featuring Profs. Peter Rossing, David Wheeler, Hiddo Heerspink, and Chantal Mathieu. In tandem with today’s presentation, the full results of DAPA-CKD were published in NEJM, entitled “Dapagliflozin in Patients with Chronic Kidney Disease” – a huge congratulations to the investigators and the many patients who took part in the trial.

Though all presentations were stellar, Prof. Mathieu (KU Leuven) stole the show in her expertly delivered independent commentary, calling the trial “really a winner.” She noted, “To me, it is obvious that DAPA-CKD is a winner because it is the first trial to demonstrate in a large and vast number of people with CKD, with and without type 2, that indeed this class of agents, and dapagliflozin in particular, works.” To follow, Prof. Mathieu neatly delineated which features of the trial design and results were so striking: 

• First, DAPA-CKD included patients with eGFR ≥25 and ≤75 mL/min/1.73 m² and UACR ≥200 and ≤5000 mg/g. Prof. Mathieu praised the investigators for being “brave enough to bring down the eGFR range ever lower than CREDENCE” – the renal outcomes trial for J&J’s SGLT-2 inhibitor Invokana (canagliflozin). Patients in CREDENCE had eGFRs ≥ 30, and following approval of its novel renal indication, Invokana currently holds the lowest eGFR product label in the class. With these new data, however, Farxiga stands to take this designation and bring this very efficacious drug to an area of high unmet need.

• DAPA-CKD included a diverse patient population, with excellent participant retention (99.6%) and collection of vital status (99.9%). As a reminder (see ESC 2020 for more on study design), DAPA-CKD was a worldwide study with 52% White, 5% Black or African-American, 35% Asian, and 8% “other” participants in the active arm. While this split is relatively diverse compared to the broader landscape of clinical trials, Prof. Mathieu very aptly stated that she was “a bit disappointed” in the low prevalence of Black or African American participants, considering that CKD is extremely prevalent and a major driver of mortality in this population. Nevertheless, she concluded that screening 7,517 of these “very vulnerable” CKD patients and randomizing more than 4,000 in 20 months was quite a feat.

• The results justify an “OMG.” With regard to the 39% relative risk reduction seen on the primary endpoint of composite sustained ≥50% eGFR decline, end stage kidney disease (ESKD), and renal or CV death, Prof. Mathieu had three letters: “OMG.” She humorously highlighted the “0.00…whatever” p-value (p=0.000000028), demonstrating the extremely strong statistical significance of the study. Of course, the drug’s lack of statistical significance on CV death was disappointing, but we loved Prof. Mathieu’s pragmatic take on this: “Hey, who cares if you die less, which is shown by the all-cause mortality. Who cares if it’s cardiovascular or from another cause? The 31% less all-cause mortality is not trivial.” We would have to agree here, considering that dapagliflozin is now the first drug to show an all-cause mortality benefit in this population. We also note that other KOLs have speculated that DAPA-CKD’s early stop due to “overwhelming efficacy” may have contributed to the lack of significance on CV death. Importantly, Prof. Mathieu noted that the primary endpoint was found to be consistent across patients with or without type 2, UACR below or above 1,000 mg/g, and eGFR below or above 45.

On safety, Prof. Mathieu commented that dapagliflozin’s profile was “very nice and bland” (clearly a compliment when it comes to safety), noting the lack of signal on fractures, hypoglycemia, renal related adverse events, DKA, and presumably amputations. However, as a very close reader of the fine print, Prof. Mathieu pointed out the lack of data on genital infections being shared (see the slide here under “safety results). This key safety outcome was also not included in today’s publication in NEJM, and we hope to see further elaboration in future papers or presentations.

To end her presentation, Prof. Mathieu closed with two very important calls-to-action in regard to how SGLT-2 inhibitors are regulated and utilized. 

• First, a call to HCPs to encourage regulators and payers to allow usage of these medications in eGFRs as low as 25. Prof. Mathieu highlighted that, as an endocrinologist, she was particularly excited to see a consistent benefit at low eGFR levels known to have negligible effects on A1c. 

• Second, a call to endocrinologists, nephrologists, and cardiologists to unite and create clear and actionable guidelines for PCPs. Emphasizing the recent deluge of conference talks on ‘which specialty owns SGLT-2 inhibitors?’, Prof. Mathieu encouraged specialists to “be humble” and “come together” to realize that PCPs are the ones who treat the majority of patients with type 2 diabetes. Rather than “claim” the class of drugs, specialists should “come together and very clearly tell PCPs when and how to use these agents.”

On the last point of her summary slide, Prof. Mathieu shared a powerful message: “The challenge for learned societies (make clear guidelines), clinicians (follow guidelines), regulators (adapt labels), and companies (be wise in pricing) is to bring these agents to those patients who will benefit on the basis of the evidence available.”

• Before Prof. Mathieu’s presentation, Prof. Peter Rossing (Steno Diabetes Center) opened the session by reviewing the landscape of diabetic kidney disease and non-diabetic kidney disease clinical trials to date. Looking at the figure below, it is clear that DAPA-CKD is a landmark trial, as the only recent trial to effectively show a benefit in patients with CKD without diabetes. As these patients previously only had ACEs/ARBs to treat this condition (in addition to lifestyle), SGLT-2 inhibitors should continue to be used in a greater number of people.

• Prof. David Wheeler (University College London) followed Prof. Rossing with a helpful explanation of the study design and patient population. As a practicing nephrologist, Prof. Wheeler started his presentation with a patient case to demonstrate the level of unmet need. The patient, a 73 year old male with type 2, had end-stage kidney disease, had to come in 3x a week for dialysis, struggled with microvascular complications like retinopathy, and macrovascular ones like a non-ST-elevation myocardial infarction and recent transient ischemic attack. Rather dramatically, Prof. Wheeler stated, “I present this case because this man’s life has been destroyed by CKD and diabetes – a situation we want to avoid if at all possible. In order to do so, we want to pick up and manage CKD as effectively as possible and optimize diabetes control.” Within the overview (see ESC 2020 for a full breakdown of the patient population), Prof. Wheeler presented two particularly helpful slides: one showing investigator-reported causes of CKD between all participants and just those with type 2 and another contextualizing DAPA-CKD’s Kidney Disease Improving Global Outcomes (KDIGO) score compared to other SGLT-2 inhibitor trials.

  • On the latter, Prof. Wheeler pointed out that DAPA-CKD included a lower range of eGFRs than CREDENCE, though Lilly/BI’s EMPA-KIDNEY for Jardiance (empagliflozin), which has yet to readout, will include the lowest range down to an eGFR of 20. Interestingly, EMPA-KIDNEY will also include patients with all ranges of albuminuria, generating needed data on an important patient population – those with a low or falling eGFR but normoalbuminuria.

DAPA-CKD Lead Investigators Profs. Hido Heerspink and David Wheeler Talk Improving CKD Diagnosis and “Millions of Patients” Who Now Stand to Benefit from Dapagliflozin

In preparation for the DAPA-CKD full results presentation today, our team yesterday took part in a group conversation with DAPA-CKD lead investigators Profs. Hido Heerspink and David Wheeler. See below for highlights from that conversation, organized into responses on (i) Risk Stratification and Primary Prevention; (ii) Potential Label and Population Expansions; (iii) All-Cause Mortality and CV Benefits; (iv) Mechanism of Action; and (v) Patient Population.

On Risk Stratification and Primary Prevention

Q: How do you see this class of drugs, and dapagliflozin in particular, being integrated into a primary prevention usage in the future? 

Professor Heerspink: We all realize that primary prevention is very important because the earlier we identify chronic kidney disease, the earlier we treat and longer we can delay dialysis.

The DECLARE trial was published a few years ago and looked at the effect of dapagliflozin in patients with type 2 diabetes in early stages of kidney disease, but half of the population had an eGFR above 90, and even in that population, it was shown that it helped to delay kidney disease. We now extend these findings to patients with chronic kidney disease. So, when we look at these two trials in combination, they show that dapagliflozin is effective across the continuum of kidney function, and thus, allows for early treatment to delay the progression of chronic kidney disease to dialysis.

The problem is that we have to start to identify these patients. For example, in the US, still 50% of patients with chronic kidney disease remain undiagnosed. Another example, in the UK, 40% of patients with microalbuminuria did not have a record of it in the GPs’ database. So, there are still many patients with CKD who are still undiagnosed. So, it’s the responsibility of the physicians, and us as a community, to make people aware of the importance of early detection. To make people aware that we now have proper and appropriate, very effective, and safe treatments to slow the progression of kidney disease, but I completely agree that we should endorse early intervention and treatment with SGLT2 inhibitors. 

Q: I just wondered if there were any further thoughts about risk stratification. I think particularly in the US, many patients don’t know what their eGFR is. Is this addressable?  

Professor Wheeler: You know, we often get patients who are preparing for dialysis with eGFRs of less than 20, and they say to us, “why didn’t anyone tell me that my kidneys were failing.” We can’t always answer that question.

One of the reasons we introduced the concept of an estimated GFR was to actually help clinicians recognize how much kidney function had already been lost. You can use the eGFR as a rough percentage of kidney function – an eGFR of 36 is around ~36% of kidney function. You know, we’re trying to educate a broader group of physicians, primary care physicians, to actually pick up this kidney disease early, but it’s really good when the patients are aware. If you're a diabetic patient, and you're going to your diabetes clinic, then once a year you really ought to have your urine checked for albumin and have your blood taken to estimate your GFR. If we could activate patients so that they can help us with this, then I think we stand a better chance of picking up kidney disease early and starting these effective drugs earlier in these patients.

Q: Is the right question for a patient to ask just, “Doctor, how much protein is in my urine?”

Professor Wheeler: We had a campaign in the UK, where we encouraged patients to ask about their numbers: “What’s my eGFR?” and “What’s my urinary ACR?” For interested patients, we even have home test kits we can mail them to allow to test their own albuminuria levels and read the indicator on a smartphone and send it into the clinician. In reality, those aren’t the patients we need to find. It’s those who don’t engage with the healthcare system that we need to pick up and treat.

On Evolving Guidelines for CKD

Q: I know you can’t speak for the guidelines, but from the perspective of patients, we hear more advice and learn more when different therapies are included in guidelines. So, any speculation on what you would want to see changed there would be great.  

Professor Wheeler: So, let’s deal with the first issue, which is about unmet need and guidelines changing. What you do with a trial is publish the data obviously in the best journal you can, and you hope that those individuals who write the guidelines pick up on the evidence and use the evidence for future guideline development, and that’s certainly been the history in nephrology whereby trials done well – that have shown clear answers, particularly in terms of the clinical outcomes for the patients – have worked their way through to the guidelines and have ultimately changed practice.

But there’s a delay in that cycle between the trial being published and the guidelines changing. Clinicians sometimes short circuit that delay by saying, “there’s a really good trial out there. I want to treat my patients the best I can and therefore use the therapists,” so I believe we will see guidelines changing as a result of this and other trials.

On Potential Label and Population Expansions

Q: The trial included 14% of patients with an eGFR between 25-29. Current labeling for dapagliflozin and other drugs in this class do not currently include that population. To what extent do you think these findings establish safety and efficacy for patients with an eGFR of less than 30? And how broadly do you think that observation can be applied going forward?

Prof. Heerspink: This is a really relevant question. Until now, most of the clinical trials have enrolled patients with an eGFR down to 30. We are the first kidney trial with an SGLT-2 inhibitor that extends that range to an eGFR of 25. You are perfectly right, that there were 630 patients with an eGFR between 25-30 at baseline, i.e. 14% of the total cohort. In that population, it seems that dapagliflozin was as effective as in the other patients, and there were no excess adverse events in people with eGFR between 25 and 30. So, the overall results were consistent in patients with CKD Stage 4.

Can we then further extend the use of SGLT-2 inhibitors to eGFR of 20 or even lower? We do not know because we did not test those patients. There is one ongoing clinical trial, the EMPA-KIDNEY trial, which enrolled patients with an eGFR down to 20. That trial will read out in the next couple of years, and then we will know yif we can initiate an SGLT-2 inhibitor at even lower eGFR levels.

It is important to know that we continued dapagliflozin until patients with dialysis, and even the protocol did not mandate to stop dapagliflozin when patients reached dialysis. We did not observe an imbalance in adverse effects between the dapagliflozin and placebo groups, so we can advise physicians to continue SGLT-2 inhibitors, dapagliflozin in particular, until patients have reached dialysis. 

Q: In terms of initiation pre-dialysis. You mentioned that this is a modest experience with a little over 630 patients. Do you think, for say regulatory bodies or guidelines, that this is adequate evidence to recommend using an SGLT-2 inhibitor down to an eGFR of 25? Or do you think that’s kind of conditional right now?

Professor Heerspink: I can’t speak for guidelines or regulators right now. They have to make their own assessment based on the data, and it is too premature to speak for them.

Q: But in your opinion, as a clinician, how would you see this?

My experience is that the efficacy is present in those people. The drug is safe in those people. So, from my own perspective, I would treat a patient with a eGFR of 27 because I know that that individual will have a high chance to have a low risk of kidney failure, a lower risk of CV death, hospitalization for heart failure, and a prolonged survival. 

Professor Wheeler: I would also endorse that. My colleagues would also endorse that. We need treatments for these patients that stop the declining kidney function and that miserable life on dialysis. We will obviously have to work within the boundaries of the trial, but I think my colleagues will be keen to start those patients who have picked up late or who have already reached that stage, in a last-ditch effort to delay the need for dialysis. 

Q: Could one of you please give some characterization of the size of the new population that is now eligible for the dapagliflozin treatment based on these results – say relative to the size of the patients with diabetes caused CKD – the other patient without a diabetic etiology, but who meet albuminuria and eGFR criteria for the trial. Is that population about the same who would have a diabetic etiology, or is it small? 

Professor Heerspink: 700 million people in the world have chronic kidney disease. Many patients with CKD have some form of albuminuria that would qualify for participation for our clinical trial. I do not know the exact number of patients and the split of patients with diabetes status who would be eligible. Keeping in mind that 700 million people have CKD, perhaps some of them have lower albuminuria, but generally when you have CKD, you have higher levels of albuminuria.

I think that there are millions of people who would qualify for treatment with dapagliflozin. Then, as I already mentioned, we have to think about primary prevention. So, there will be another segment of patients who could be eligible. I think that this drug is potentially eligible to millions of patients. The only problem is that we have to start to identify these patients, so we know who to treat.

Professor Wheeler: We are talking about the broad spectrum of patients here who would benefit, but I think what you are asking is: “how many more would get in on the basis of our trial result?” Again, I don’t have numbers, but I think that these are, for the nephrologist, the most common patient we see and get referred. If you’re sitting in the nephrologist’s office, the most common referral is a patient with type 2 diabetes who has either increasing albuminuria or declining kidney function. There are plenty of these patients out there. I just couldn’t put numbers to the different subgroups, but I think that is something we will do now that we’ve got the trial data. We need to put the numbers in the boxes and establish how many extra patients would benefit as a result of this study. 

On All-Cause Mortality and CV Benefits

Q: We were excited to see the strong survival benefit, and if possible, I am wondering if there are any insights into what is driving that?

Professor Wheeler: Yes, it is a little bit unusual because if we saw a benefit in any type of death, we would have expected a benefit in cardiovascular death. I think this just proves that it is actually difficult to determine exactly what the cause of death is in these patients who have, you know, several different diseases that all come into play.

The final event may be a catastrophic infection or something along those lines. That because of the underlying heart diseases and diabetes and kidney diseases, that the patient succumbs to that infection, where they might not if they were fit and healthy. So, I think it’s a matter of the difficulties in actually deciding on cause of death. We see that in clinical practice as well. I think it’s difficult to determine exactly what the cause of death is and therefore, the cardiovascular didn’t reach statistical significance, but when you look at deaths overall, there was a benefit. 

Professor Heerspink: That’s why we pre-specified all-cause mortality as the secondary outcome because we knew already that it is difficult to classify death. All-cause mortality will capture that because it will count anybody that died. 

On Mechanism of Action 

Q: Do you have any numbers you could give us on the prevalence of a history of heart failure and a history of cardiovascular disease in the population that you looked at?

Professor Wheeler: Pretty common. You know, heart failure is one of the main modes of presentation of this patient population. So, we quite often see patients who present with breathlessness and heart failure, who turn out to have chronic kidney disease primarily, and poor heart function as a result of fluid retention resulting from the chronic kidney disease.

Cardiovascular disease is the most common cause of death in a CKD population, and I can tell you that far more patients with CKD die prematurely of cardiovascular complications than ever reach the dialysis machine. So again, it’s difficult to give you precise numbers but of most patients with CKD, premature death with cardiovascular disease is the more common outcome than is reaching end stage kidney disease and requiring dialysis, particularly in the diabetic population.

Q: And the reason why I asked, that is, I was wondering, the fact that none of the patients who are enrolled were selected for a cardiovascular disease or heart failure history, and yet, you saw this profound effect on that endpoint as well. I guess what you just told us is the reason for that observation…

Professor Wheeler: Yeah, the two are intermingled. You know the heart provides blood to the kidney. And the kidneys release fluid from the circulation that allows the heart to work properly and when you’ve got either kidney failure or heart failure, the opposite organ usually suffers as a result.

So, I think this is about cardiorenal disease and the concept of cardiorenal disease. And you know, Hiddo mentioned the DAPA-HF study, which was actually a heart failure study but also looked at kidney function in those patients with heart failure, and when you speak to the cardiologist, they say, “Oh, yeah, we often see patients with heart failure have reduced eGFRs, and we see it as part of the heart failure that we're looking after.” 

Q: Could you comment on how fast it seems that SGLT-2 inhibitors appear to “start”? And some of the other benefits that could prompt more interest at the PCP level such as outcomes like weight or Time in Range?

Professor Heerspink: With respect to the time course, the effect on albuminuria is directly present after a week.What you also see is a reduction in eGFR, and it may raise some safety concerns among physicians who are not familiar with the mechanism of action, but this drug works – just like an ACE inhibitor or ARB – causes an acute decline in eGFR that is directly related to the mechanism of action because it causes a reduction in the pressure in the glomeruli that causes a reduction in GFR.

The reduction in GFR is not a worsening of structural function, but it is a so-called hemodynamic effect, which is completely reversible when you stop the drug. So, the effects are directly present, and we saw that in our trial as well. The Kaplan-Meier curve for the primary outcome started to diverge already after three months, which is unique.

There are also other effects – bodyweight reductions, blood pressure reductions. All these effects are directly present after a week. After four days. you already see a reduction in body weight due to the diuretic effects of the drug. The reduction of blood pressure is also completely present after four days. We know from patients that they really like the drug because of the body weight reduction because they feel they have to lose body weight, particularly for those with diabetes. There are multiple benefits with these drugs. The only thing that physicians have to be aware of is that the drug also causes an acute fall in GFR, which is related to the mechanism of action.

On Patient Population 

Q: Do you think that it is important to have patients stabilized on a RAS blocker before starting an SGLT-2 inhibitor for this particular indication that was set up in the trial?

Prof. Wheeler: That is what we did in trial, and so that is what we should do in practice. These drugs do not substitute for ACEs and ARBs. We want patients to be on RAS blockade first, and we want to add this drug in, and that’s exactly what we did in the trial.

Q: I just wondered what kinds of kidney disease that people without diabetes had. What were the most common types?

Professor Wheeler: Glomerular diseases or glomerulonephritis – diseases that damage the glomerulus. Many of those diseases are autoimmune in nature, but often the autoimmune part of the disease burns out, and the patients just show gradual decline in kidney function. Also, what we call ischemic or hypertensive neuropathy, which is just kidney disease in the presence of hypertension, where the kidneys shrivel up, and we assume are not getting sufficient blood.

There were a few other rare chronic kidney diseases with damage to the tubulointerstitial kidney. Again, as we see in many of these populations, many patients got in with kidney diseases that we did not really understand the cause of. There is always a subset of patients that gets into CKD trials, where the actual cause of kidney disease is not completely obvious to the clinician looking after that patient, so we classify these as CKD of unknown cause. 

EMPEROR-Reduced/DAPA-HF Meta-Analysis Shows Consistent Cardio- and Renal-Protective Effects of the SGLT-2 Class in People with HF

A morning session reprised the striking results of the EMPEROR-Reduced trial, first reported just weeks ago at the virtual ESC 2020 conference, showing a compelling 25% relative risk reduction for CV death and hospitalization for HF in people with HFrEF with empagliflozin. The highlight of the session was a new meta-analysis published in The Lancet of both SGLT-2 inhibitor HF trials to date: EMPEROR-Reduced and DAPA-HF. Lead meta-analysis author Dr. Faiez Zannad (University of Lorraine) characterized the results of EMPEROR-Reduced and DAPA-HF as highly consistent, in fact “nearly superimposable.” For the meta-analysis primary outcome of first hHF or CV death, SGLT-2 inhibitors were associated with a 26% risk reduction (HR 0.74, 95% CI 0.68-0.82, p<0.001), and this was consistent between empagliflozin (HR 0.75) and dapagliflozin (HR 0.74). Importantly, these HF-protective effects were consistent in people with diabetes (HR 0.74, 95% CI 0.65-0.84) and without (HR 0.75, 95% CI 0.65-0.87). SGLT-2 inhibitors were further associated with a 14% reduced risk of CV death (HR 0.86, 95% CI 0.76-0.98, p=0.027) and all-cause death (HR 0.87, 95% CI 0.77-0.98, p=0.018), though notably these results were driven by dapagliflozin, with empagliflozin showing a non-significant trend toward risk reduction for CV death (HR 0.92, 95% CI 0.75-1.12) or all-cause death (HR 0.92, 95% CI 0.77-1.10). A similar phenomenon, but in the opposite direction, was observed for the renal composite endpoint (>50% sustained decline in eGFR, ESRD, or renal death): SGLT-2 inhibitors as a whole were associated with a 38% risk reduction (HR 0.62, 95% CI 0.43-0.90, p=0.0128) but this was driven by empagliflozin (HR 0.52, 95% CI 0.29-0.92) with only a trend toward renal protection with dapagliflozin (HR 0.71, 95% CI 0.44-1.16). It may be too premature to speculate on potential within-class differences in these outcomes with only two available trials, and Dr. Zannad underscored that the totality of the evidence indicates a protective effect of SGLT-2 overall against renal disease and CV/all-cause death. Indeed, statistical tests of heterogeneity showed no difference between the two trials. The difference on renal may have been due to populations being treated – we need to spend more time on this in order to fully understand it.

On a methodological note, Dr. Zannad pointed out that the extremely similar patient populations in EMPEROR-Reduced and DAPA-HF add further reliability to the results, as does the remarkably similar trial design in terms of intervention/control (10 mg empagliflozin or dapagliflozin, respectively, vs. placebo), median follow-up time (15.7 months vs. 18.2 months), and primary outcome (time to first CV death or hHF vs. time to first CV death, hHF, or urgent HF visit). Overall, this meta-analysis establishes a solid evidence base for the cardio- and renal-protective role of SGLT-2 inhibitors in people with HF.

• As the commentator on the EMPEROR-Reduced results, Baylor’s Dr. Milton Packer provided a fascinating deep-dive on the hypothesized mechanism underlying the cardio- and nephroprotective effects of SGLT-2 inhibitors: Mimicking the body’s starvation response. He argued that metabolic disease is characterized by a decrease in biological signals activated by nutrient deprivation (such as SIRT1 and AMPK) and an increase in those triggered by energy surplus (such as Akt and mTORC1). Together these signals “fool” the body into thinking it is in a nutrient surplus state, causing the body to suppress autophagy (“the cellular housekeeping process”), increase oxidative stress and organelle dysfunction, and activate pro-inflammatory pathways. This has the cumulative effect of accelerating the development of cardiomyopathy and nephropathy. Enter SGLT-2 inhibitors, which Dr. Packer argued reverse this process by decreasing SIRT and increasing mTORC1. In line with this framework, in addition to decreased SIRT and increased mTORC1, SGLT-2 inhibitor studies have shown a reduction in oxidative stress and pro-inflammatory signaling in the heart and kidneys, increased autophagy in the heart and kidneys, and the reversal of mitochondrial dysfunction. Altogether this provides an interesting hypothesis that SGLT-2 inhibitors’ cardio- and nephroprotective effects are unrelated to glucose-lowering (we believe that had already been accepted), and instead have to do with rebalancing cellular homeostasis. The session concluded with a rousing call to action: “With this evidence for SGLT-2 inhibitors in HFrEF with and without diabetes, we are ready to add SGLT-2 inhibitors as one of the cornerstones of HF treatment.” Dr. Packer further explained that he envisions a quadruple therapy paradigm for HF: angiotensin receptor neprilysin inhibitors, beta blockers, mineralocorticoid receptor antagonists, and now SGLT-2 inhibitors.

Cross-Specialty Roundtable Discussion Highlights Results of VERTIS-CV That Support Paradigm Shifts in Diabetes Care

Drs. Christopher Cannon (Brigham and Women’s Hospital), Richard Pratley (AdventHealth Translational Research Institute), and David Cherney (University of Toronto) all participated in a fascinating roundtable discussion on the effects of VERTIS-CV. Much of their conversation centered on what the study means for the SGLT-2 inhibitor and GLP-1 agonist classes more broadly and the clinical implementation of SGLT-2 inhibitor ertugliflozin. This includes paradigm shifts to have nephrologists and cardiologists begin prescribing the drugs to patients without diabetes that may benefit, as well as contemplating whether metformin should remain foundational therapy for patients with diabetes. See below for their conversation, which supports early implementation of SGLT-2 inhibitors or GLP-1 agonists across specialties and hope for the future of collaboration between endocrinologists, cardiologists, and nephrologists.

Panel Discussion

Dr. Christopher Cannon: How does ertugliflozin rank in safety and tolerability for what we use to treat diabetes?

Dr. Richard Pratley: We saw that the efficacy of ertugliflozin in terms of A1c lowering is comparable to other members of the SGLT-2 inhibitor class. The same is true for body weight change; what we saw is typical to what’s been seen in phase 3 trials. In VERTIS-CV, there were no new signals of concern – we saw the predicted increase in genital infections to people randomized to ertugliflozin. We also saw reductions in blood pressure, but no increase in adverse events related to blood pressure. There was no Fournier’s Gangrene, which is reassuring, and there was no increased risk of bone fracture. VERTIS-CV has added to the understanding of efficacy and tolerability of the SGLT-2 class.

Dr. Cannon: These started out as glycemia control drugs, and now, we see these broad effects. How much of a contribution do you think glycemic control plays in outcomes, and is that important?

Dr. Prately: That’s an important question and hard to answer from studies like this. What we know from UKPDS, ACCORD, and ADVANCE is that glycemia is important for microvascular complications, but in the short term has little effect on CV complications and no effect on heart failure. With other SGLT-2s, we have done mediation analyses to determine effects in A1c to address this point. If there’s glycemic equipoise, would we see a benefit? Any contribution of reduction in A1c has been small for this class and most of the others. If we looked at the same thing for changes in heart failure, we’d see no contribution of glycemic control. In some cases, reduction in A1c may reflect efficacy of the drug in other pathways. What it’s doing is a marker of the other effects of the drug. That’s why we see some small mediation effects.

Dr. Cannon: It’s terrific to see all this broad improvement. A lot of focus is on the kidney. How much new data does this share in terms of where this drug fits into the class of these different endpoints?

Dr. David Cherney: The question around class effect and how this compares to what comes before is really important. The data we talked about today, especially the 40% decline in kidney function, adds to what we know in VERTIS-CV about renal efficacy and benefits. It puts it on the same playing field and makes it comparable to the data published before. The 40% sustained decline in GFR was significantly reduced with ertugliflozin. The magnitude of effect, direction, and significance highlights the benefits of these therapies across the board regarding renal protection. We also have to keep in mind that this was a CV cohort, and some happened to have kidney disease at baseline. This isn’t a group like CREDENCE; this 40% decline meets the balance of being clinically relevant and important to physicians and patients, but also a sufficient bar in terms of showing clinically important effects. It’s a recognized marker and benchmark in therapy that’s protective overtime and leads to broad benefits for patients. That effect and benefit appears to be there across the board.

Dr. Pratley: Following along the renal mechanism benefit, we’ve seen in other trials that there is an impact in some microvascular outcomes. Do you think glucose lowering is important [mechanistically], or is it a different mechanism?

Dr. Cherney: I think we’re looking at something different in terms of mechanism for kidney function based on five lines of research.

• One, glucose independence. If you control for the effect of A1c, those events are statistically independent of A1c lowering. That’s a surrogate marker.

• Two, if you compare the effect of SGLT-2s to another glucose lowering therapy like SUs, even though A1c changes overtime are the same, there’s still an anti-albuminuria effect that preserves kidney function, even in patients with good levels of function at baseline.

• Three is baseline kidney disease – we still see a kidney protective effect in in patients with low function and now A1c lowering.

• Four is that we see a renal protective effect reducing kidney disease via kidney function loss and renal death regardless of A1c at baseline and change during the trial.

• Five, data presented at ESC by Dr. Hiddo Heerspink showed that in patients treated with dapagliflozin, there was a significant reduction in the cardiorenal main endpoint, and those benefits were extended in people with and without diabetes. This shows there’s a benefit that is glucose independent.

Dr. Pratley: We should think of a paradigm shift. It’s no longer only glucose lowering for people who are not at goal, but instead, for people who also want to limit the progression of kidney or heart disease.

Dr. Cherney: I think that’s important for cardiologists and nephrologists like me; we aren’t trained to manage glycemic control or therapies. SGLT-2s are independent of glucose and easy to use in people with low kidney function. The message that it’s easy to use as non-endocrinologist is important for colleagues to know using the drug doesn’t create additional tasks.

Dr. Cannon: There’s an interesting notion of effect on other things than glycemia, which came through with the heart failure outcome. In DAPA-HF, patients without diabetes having a benefit woke us up to see this is a heart failure, renal, and diabetes drug. Consistency was seen across trials; each one has 30% reduction in hospitalization for heart failure, and early adopters are heart failure patients.   

Dr. Cherney: There’s a similar line of thought with ACE inhibitors. Don’t forget about glucose lowering, but the glycemic effects fall away from priority and end-organ effects come to the foreground. Going back to mechanisms, it probably has effects in the kidney that reduce hypertension, inflammation, and hypoxia – which a lot of ACE inhibitors do.

Dr. Pratley: One other aspect that makes them great drugs is that there’s low risk of hypoglycemia. It doesn’t matter if A1c is 6.5% or 6%, patients won’t have hypoglycemia because it works in an insulin-independent mechanism. Cardiologists and nephrologists should feel reassured about this and even if someone’s at goal, they could still benefit from a renal and CV perspective.

Dr. Cherney: I wanted to ask similar question regarding class effect from a CV perspective. You mentioned consistency around hHF, do you think it’s a class effect on the heart level?

Dr. Cannon: I think so. In VERTIS-CV, the risk of CV death was not significant, but in EMPA-REG, it was different. Different trials have more muted effects than the signal seen in EMPA-REG. Even in EMPEROR-Reduced, it was the same as in VERTIS. That’s where the meta-analysis is helpful – trials aren’t going to get the same results, so with this wealth of data, we see a consistent pattern with some trials having more effects but probably modest reductions in MACE, CV death, consistent HF signals, and now renal signals. This speaks to effects being there in all trials with some variability. It’s exciting that we can take these and apply them. From HF patients it’s clear they’re breaking through, when, would you think about this for a renal patient?

Dr. Cherney: Certainly, albuminuria is big flag right now because that marker is used to get patients in trials like CREDENCE and DAPA-CKD. We’ll have a huge wealth of data in that patient group. Many studies with several agents show benefits. They push albuminuria down a little bit and lower GFR thresholds down to 25 and 30 mL/min/1.73m² (in DAPA-CKD). We’re chipping away areas of the unknown in kidney protection. We still don’t know about function but no albuminuria. We don’t have any study of any drug that shows clear benefit in patient with normo- or micro-albuminuria in benefits of a hard outcome. EMPA-KIDNEY will show that with patients between 20 and 25 mL/min/1.73m². This will fill in some blanks. Patients with low eGFR are still unknown; that needs to be answered. You can extrapolate data, but there’s still more to come.

Dr. Pratley: From an endocrinologist perspective, we follow guidelines, and. It doesn’t take much to get people on a SGLT-2. If they have a decrease in eGFR or albuminuria, then they are candidates for SGLT-2s as a first-line therapy.

Dr. Cannon: When using this class, do you need to use metformin first? It’s been in the guidelines, but if these drugs are having effects in patients without diabetes, should you step through that hoop?

Dr. Pratley: Good question. This applies to people who come into the healthcare system with new diagnoses and comorbidities. Metformin remains a good first line therapy because it is efficacious, well tolerated, cheap, and there is some evidence that it’s safe from CV perspective. Important to note is that in most CVOTs, most patients have been on background metformin therapy. That doesn’t detract from the potential benefit of SGLT-2s or GLP-1s. It’s not clear whether background therapy’s needed, but most subgroup analyses have similar effects regardless of background metformin. Regarding new diagnosis of diabetes or established CVD or renal disease, I worry that starting metformin would lead to delays initiating other drugs with proven benefit. At the very least, metformin can be added, but shouldn’t delay other drugs with proven benefits.

KOLs Gather to Reflect on the Last Five Years of SGLT-2 Inhibitor CVOTs

In this comprehensive session that provided a review of the last five years of CVOTs, beginning with the revolutionary results of the EMPA-REG OUTCOME trial released at EASD 2015, KOLs reflected on the evidence base of CV/renal data and mechanisms of action.  As shared by Dr. Darren McGuire (UT Southwestern Medical Center), based on the EMPA-REG OUTCOME, CANVAS, DECLARE-TIMI 58, CREDENCE, DAPA-HF, EMPEROR-Reduced (topline results only), and VERTIS-CV trials, there is moderate heterogeneity on MACE and CV death outcomes, with consistency on hospitalization for heart failure. With this, six trials of four different SGLT-2 inhibitors show CV safety and incremental efficacy on these outcomes, indicating a shared efficacy across the class. According to Dr. Nikolaus Marx (University Hospital of Aachen), the cardio-protection conferred by SGLT-2 inhibitors is based various mechanisms such as diuretic effects, natriuresis as wel as effects on cardiac function and metabolism.

Similar attention was given to the renal data from the SGLT-2 outcome trials EMPA-REG OUTCOME, CANVAS, DECLARE-TIMI 58, VERTIS-CV, CREDENCE, DAPA-CKD, and EMPA-KIDNEY (which has yet to read out). Dr. Christoph Wanner (University Hospital of Würzburg) noted that (i) SGLT-2 inhibitor treatment induces initial declines in eGFR; and (ii) this initial eGFR dip raised concerns because of its ability to predispose patients to acute kidney injury. He wrapped up his talk with reference to the progression of the SGLT-2 inhibitor class. At their introduction, their glucose-lowering abilities were lauded, and after EMPA-REG OUTCOME results were announced in 2015, they gained momentum as drugs to promote heart health; later, DAPA-CKD results showed kidney protection in people both with and without diabetes. These results directly influenced the new 2020 KDIGO Guidelines - see compelling video from Dr. Ian De Boer as well as the webinar from late 2019 when the new guideline was launched. There have been about 2,700 views of the new guideline, which was KDIGO’s first guideline related to diabetes – unbelievably, before the guideline was finalized, which will be released in full next week, it went out for public review! There were about 150 people on the webinar, that is quite interesting. As we understand it, the guideline classifies SGLT-2 inhibitors as a primary therapy for organ protection and we suspect there will be significantly more information and insights shared. As aptly put by Dr. Wanner, “SGLT-2 inhibitors have a significant and clinically relevant impact across the spectrum of kidney function in DKD/CKD.” Dr. David Cherney (University of Toronto) stated that based on findings from hemodynamic, filtration, inflammatory, hypoxia-related pathway, and substrate effect investigations, sodium plays a role in kidney protection, and SGLT-2 benefits in the kidney seem glucose-independent. Dr. Cherney concluded that researchers should move past eGFR thresholds like 45 or 60 ml/min/1.73m² and focus on uncovering potential synergies with other novel agents.

• With reference to various professional societies’ guidelines, Dr. Silvio Inzucchi (Yale University) closed the session with a look to the future of diabetes treatment with SGLT-2 inhibitors. He stated that we are currently in the 3^rd era of type 2 diabetes treatment, which is truly evidence-based. Expressing excitement over future studies, he highlighted EMPEROR-Preserved, DELIVER, and EMPA-KIDNEY as the next revolutionary studies of SGLT-2 inhibitors. Many questions remain that could prompt even more expansion: these include (i) the actual underlying mechanisms underlying the observed CV and CKD benefits; (ii) the possibility of these benefits in primary prevention or in people without diabetes (the latter is no longer a question in our view); (iii) choosing between SGLT-2 inhibitors and GLP-1 agonists in people with ASCVD; (iv) considering if the type of ASCVD matters; and (v) determining whether SGLT-2 inhibitors’ CV benefits are additive or synergistic with GLP-1 receptor agonists. These questions will be extremely valuable as diabetes treatment progresses to SGLT-2 inhibitors used as first-line therapies in diverse situations and as early combination therapy approaches becoming standard care. For more information on how CVOTs have progressed throughout the years for SGLT-2 inhibitors and other drug classes, we suggest viewing our CVOT competitive landscape.

Real-World Study of ~9,000 Type 2s Suggests SGLT-2s Give Greater Risk Reduction for Three-Point MACE and Heart Failure vs. GLP-1s

Dr. Gian Paolo Fadini presented a real-world study (n=8,596) to illuminate how HCPs should choose between GLP-1 agonists and SGLT-2 inhibitors for patients with diabetes/high CV risk. He started with data on biomarkers: GLP-1s showed greater A1c-lowering efficacy vs. SGLT-2s, whereas SGLT-2s were significantly superior on systolic blood pressure reduction. Turning to hard outcomes, Dr. Fadini highlighted a directional signal that SGLT-2 inhibitors give greater protection against three-point MACE (non-fatal MI, non-fatal stroke, CV death) as well as heart failure, though neither finding reached statistical significance. We were surprised by this result, given the growing consensus among thought leaders that GLP-1s are more effective in reducing risk for atherosclerotic events (e.g. MI and stroke) while SGLT-2s are more effective in heart failure. Some guidelines explicitly recommend GLP-1s for patients with atherosclerosis, SGLT-2s for patients facing higher heart failure risk. Ultimately, the choice between a GLP-1 and an SGLT-2 should be personalized; studies like this one provide important insight, though it should be interpreted in the context of other real-world and RCT data as well. We should note that getting on either of these therapies will be meaningful for CV prevention, and we hope to see both drug classes expand substantially in the coming years.

Debating the Most Effective First- and Second-Line Therapies for Type 2 Diabetes: Insights from Nephrology, Cardiology, Endocrinology, and Primary Care

As part of Lilly/BI’s ACROSS T2D program – an interactive educational resource platform for HCPs who care for type 2 diabetes patients – this industry-sponsored symposium addressed three critically important medical questions: (i) should metformin remain the foundational therapy for patients with type 2? (ii) are SGLT-2 inhibitors preferred over GLP-1 receptor agonists for second-line therapy, or vice versa? And (iii) which organ is deserving of greater attention in the context of type 2, the heart or the kidney? Framed as a series of debates, this session gave endocrinologists Dr. Silvio Inzucchi (Yale School of Medicine) and Dr. Francisco Giorgino (University of Bari Lado Moro, Italy); cardiologist Dr. Nikolaus Marx (University Hospital Aachen, Germany); and endocrinologist and nephrologist Dr. Merlin Thomas (Monash University, Australia) the (virtual) stage to present their expert opinions on each topic. Dr. Sarah Jarvis (Patient Platform Limited, United Kingdom) wrapped up the symposium by providing her perspective on each of these questions as a primary care physician, paying particular attention to tremendously relevant topic of shared decision making between patient and provider. See below for the most significant take-aways from each debate.

• Metformin as foundational therapy. Dr. Giorgino cited data from VERIFY, UKPDS 34, and two smaller-scale CVOTs (by Kooy et al. and Hong et al.) to demonstrate metformin’s: (i) robust and sustained effects on glucose-lowering; (ii) potential CV and mortality benefit; and (iii) excellent safety profile alone or in combination with other therapies (e.g. with DPP4 inhibitor vildagliptin in VERIFY). He also pointed out various confounding factors within the large-scale GLP-1 receptor agonist and SGLT-2 inhibitor trials. Not only do they fail to enroll sufficient numbers of low-risk patients, but the majority of patients in both placebo and treatment arms of the trials are on background metformin therapy to begin with. According to Dr. Marx, however, this latter point is of little interest: across this library of trials, patients receive similar CV benefits irrespective of metformin background therapy. In CANVAS, the presence of metformin background actually blunted canagliflozin’s relative risk reduction for hospitalization for heart failure (hHF) and fatal and non-fatal stroke. Primary care physician Dr. Jarvis appeared to side with Dr. Giorgino in her commentary, believing there was “no question” that metformin gives rise to meaningful CV risk reduction and improved glycemic control. If used early on, metformin’s effect on glucose lowering can significantly reduce incidence of microvascular, macrovascular, and mortality events.

• SGLT-2 inhibitors versus GLP-1 receptor agonists as second-line therapy. While both agents have been shown to have significant benefits for glycemic control, CV risk, and cardiorenal health, among other characteristics, Dr. Inzucchi and Dr. Giorgino took opposing sides on which class should be preferred as second-line therapy. Dr. Inzucchi advocated for SGLT-2 inhibitors, paying particular attention to the drug class’ consistent benefits in hHF. Heart failure, as Dr. Inzucchi reminded the audience, is the complication associated with the single greatest excess risk in type 2 diabetes and is associated with a significantly worse clinical course (relative to that of diabetes without heart failure). SGLT-2 adherence was also shown to be superior to that of GLP-1 receptor agonists in a small BMJ study, presumably due to route of administration (oral versus injected, respectively). On the opposition, Dr. Giorgino pointed out than GLP-1 receptor agonists lower A1c to a greater extent than SGLT-2 inhibitors; as shown in a meta-analysis of GLP-1 CVOTs, the extent of A1c reduction has been directly and significantly related to the magnitude of 3P-MACE and stroke risk reductions. In his argument, Dr. Giorgino also advocated for the utility of GLP-1 receptor agonists in providing cellular protection (for beta cells and neurons) and in active COVID-19 infections. An opinion summary published in the The Lancet Diabetes & Endocrinology suggested that clinicians discontinue SGLT-2 inhibitors upon hospitalization for COVID-19, conversely. The adverse GI side effects of GLP-1s were brought up in both presentations, although Dr. Giorgino that these events can be minimized by dose titration and smaller meals.

  • On this question, Dr. Jarvis took the opportunity to highlight the importance of shared decision making in contemporary medicine. As both SGLT-2 inhibitors and GLP-1 receptor agonists are reasonable therapy options for type 2 diabetes patients, patients and clinicians should work together to reach a medical decision harmonious with the clinician’s knowledge and the patient’s experience with the condition.

Focusing on cardiac or renal risk reduction in type 2 diabetes. In advocating for a cardiac focus for type 2 diabetes care, Dr. Marx echoed many of the facts that Dr. Inzucchi also presented concerning the relationship between heart failure and diabetes. He showed that irrespective of age, heart failure is one of the most significant, frequent, and earliest CV complications associated with type 2 diabetes. Heart failure events reduce quality of life and are associated with future hospitalizations and continuous declines in cardiac function. On the opposition, Dr. Thomas advocated for a renal focus in type 2 diabetes. Not only is chronic kidney disease (CKD) easily detectable using eGFR or urine albumin-to-creatinine ratio (UACR), but it is also more treatable than heart diseases using anti-hypertensive treatment (e.g. losartan) or SGLT-2 inhibitors, for example. In fact, in EMPA-REG OUTCOME, empagliflozin slowed the rate of eGFR decline in patients regardless of renal disease at baseline, suggesting a preventative effect of these treatments as well. Treating renal disease has tremendous implications for long-term health, as Dr. Thomas suggested, reducing incidence of adverse drug reactions, heart failure, end-stage kidney disease, and mortality, among other events. As shown in the NHANES III US population-based study, the presence and severity of CKD was directly proportional with extent of excess mortality risk. Dr. Jarvis similarly highlighted the interplay between the heart and the kidney in her later presentation, showing that treating one organ may have the unintended benefit of treating the other.

MSD Symposium discusses new clinical and scientific breakthroughs in SGLT-2 inhibitors and cardiorenal metabolic health

Led by Dr. Christopher Cannon and Dr. Samuel Dagogo-Jack, a panel of leading cardiologists and endocrinologists discussed the evolving roles of SGLT-2 inhibitors as cardiorenal protective agents independent of glucose control considerations. The first half of the session summarized the results of the landmark VERTIS-CV trial. As a reminder, the trial examined ertugliflozin’s efficacy on CV endpoints, including MACE, heart failure hospitalization, and CV death. Ertugliflozin treatment met non-inferiority vs. placebo on the primary endpoint of MACE, but superiority was not met (HR=0.97; 95% CI: 0.85-1.11). Superiority was also not met on several key secondary endpoints evaluated in the trial, including (i) composite of CV death or hospitalization for heart failure (HR=0.88, p=0.11); (ii) CV death (HR=0.92); and (iii) a composite renal death, dialysis/transplant, or doubling of serum creatinine (HR=0.81). While on the surface, results from VERTIS appear to contradict past results from SGLT-2 inhibitor CVOTs, further examination and careful consideration of the totality of evidence from VERTIS (including a recent re-analysis of the renal data) showed that the trial squarely fits into the established effects that SGLT-2 inhibitors have on cardiorenal outcomes.

VERTIS disappointingly did not meet superiority on 3P-MACE vs. placebo – however, as a whole, the panel concluded from a meta-analysis of SGLT-2 inhibitor trials that SGLT-2 inhibitors favorably reduce MACE events as a class, with a 10% reduction rate. In a similar analysis looking at CV death outcomes, the results were more heterogenous, but produced an estimated 15% risk reduction. Markedly, the closest and tightest estimate across the board was for hospitalization for heart failure (HHF), with a homogenous pooled estimate of 32% risk reduction.

• Prof. Francesco Cosentino (Karolinska University, Sweden) discussed new insights into heart failure management from VERTIS-CV trial. In the meta-analysis discussed above, all SGLT-2 inhibitor trials consistently showed reduction in HHF. In the VERTIS-CV trial, ertugliflozin lowered the risk of first HHF by 30%, total HHF by 30%, and total HHF/CV death by 17%. Notably, ertugliflozin’s effect was consistent in patients with or without previous history of HHF. However, its effect was more pronounced on patients with a baseline eGFR<60mL/min/173m², patients with albuminuria, and patients who use diuretics. Dr. Constantino stressed that the VERTIS-CV trial supports the clinical use of SGLT-2 inhibitors as therapy for patients with type 2 diabetes to reduce risk of HHF.

• In a similar analysis, Dr. David Cherney (University of Toronto, Canada) emphasized that the VERTIS-CV trial is consistent with other SGLT-2 inhibitor studies in its renal effects. The renal composite outcome of the trial was renal death, doubling of creatinine, or need for dialysis. The trial found that ertugliflozin treatment trended in reduction in risk compared to the placebo group. The results of this particular trial were not statistically significant, although there was a trend shown in the initial renal analysis. While the data show a significant dip in eGFR at the beginning of treatment (an effect natural to SGLT-2 inhibitor’s mechanism of action), the ertugliflozin group was shown to have better preservation of eGFR compared to placebo over time – indicating that while there is an acute change to eGFR, this is not a lasting effect. When comparing VERTIS-CV to other SGLT-2 inhibitor trials, it is important to recognize that the renal composite endpoints were defined differently. This is the “caveat” to the VERTIS-CV results, and Dr. Cherney emphasized that further analysis (see here for an example) with different endpoint definitions is needed. Although ertugliflozin was overall well tolerated, there was an increased incidence of UTIs and GMIs compared to the placebo group

• Notably, the renal and cardiovascular protective mechanisms of SGLT-2 inhibitors seem to be glycemic control-independent. In a panel discussion with Dr. Samuel Dagogo-Jack (University of Tennessee, USA), Dr. Christopher Cannon (Brigham and Women’s Hospital, USA), Dr. Cherney, and Prof. Cosentino, the panel concluded that glycemic improvement “contributes minimally, if at all, to clinical outcomes of nephral or cardio protection.” Drawing on the results of the DAPA-CKD trial for AZ’s Farxiga (dapagliflozin), the panel emphasized that there was no difference in terms of benefit for people with and without diabetes – therefore, the protective effects of SGLT-2 inhibitors in the kidney and the heart are glucose-independent. While these are exciting results, Dr. Dagogo-Jack reminded the panel that adhering strictly to the data is important in clinical practice. Because the majority of these studies have used populations of patients with diabetes, there is a need for future studies to include a cohort of low-risk patients to glean a more complete picture.

The second half of the talk focused on the interplay between the kidney and the heart, with Dr. Kim Connelly (St. Michael’s Hospital, Canada) explaining the protective impact of SGLT-2 inhibitors on both systems. Dr. Connelly began by reviewing the physiology connecting the heart and the kidneys, underscoring the bidirectionality of cardiorenal syndromes. Activation of the renin-angiotensin-aldosterone system in particular was implicated as a cause for both renal and cardiac disease, with microvascular loss and fibrosis occurring in both organs as the conditions progress. In one key study, SGLT-2 inhibitors were shown to have direct impact on the heart (despite SGLT-2 receptors only existing in the proximal tubules of the kidney), indicating that there are unknown mechanisms for the cardioprotective effect observed. Future research is clearly required to understand the mechanisms of glucose-intendent organ protection mediated by SGLT-2 inhibitors.

AstraZeneca-Sponsored Symposium Offers a Deep Dive into the Cardio-Renal-Metabolic Mechanisms of SGLT-2 Inhibitors.

In a diversified and in-depth symposium, multiple KOLs gathered for a deep dive on the mechanisms of SGLT-2 inhibitors. Dr. John Buse began the session with an overview of the many mechanism of action hypotheses, which include diuresis, calcium leaking, and dampening of synthetic nervous system overdrive. Overall, Dr. Buse concluded that urinary glucose excretion, fuel switching, and restoring diurnal patterns of anabolism and catabolism play a role, while sodium metabolism may play a role. Dr. Angelo Avogaro then took the stage to discuss the potential of SGLT-2 inhibitors to reprogram cellular life history. This hypothesis is based on the balance between cell defense and dormancy stages defined by anabolism and catabolism, respectively. Dr. Hiddo Heerspink narrowed in on potential mechanisms of renoprotection, drawing primarily from the DAPASALT, DIAMOND, and CREDENCE studies to support that SGLT-2 inhibitors’ shunting sodium reabsorption from the proximal to distal tubules is responsible for many effects. These include restoring tubuloglomerular feedback and glomerular hyperfiltration and improving mitochondrial function, which then contribute to reducing albuminuria, long-term kidney protection, and reducing anemia. Dr. Javed Butler closed the session with an overview of different investigations that focused on connections to nerve activity and fasting. Overall, he stated that multiple cardioprotective mechanisms can be responsible for the SGLT-2 benefit, that this benefit appears to be glucose-independent based on strongest outcomes observed in (i) direct inhibition of NHE1; (ii) increases in EPO; (iii) reductions in sympathetic NS activity; and (iv) reno-cardiac signaling changes. All of these ideas were brought together in a panel discussion and Q&A, moderated by Dr. Jiten Vora. Read on for the conversation below!

Panel Discussion

Dr. Jiten Vora: What do you think results for SGLT2s will look like in HFpEF?

Dr. Javed Butler: If our prevalent way of thinking about HFpEF has any truth, then we think of it as a combination of left ventricular hypertrophy, hypertension, diastolic dysfunction, aortic stiffness and endothelial dysfunction, abnormal renal sodium and fluid handling, and cardiometabolic comorbidities associated adiposity and insulin resistance; all of these are favorably affected by SGLT2s. There’s a very good chance that SGLT2s might improve HFpEF. If it doesn’t work, then need to think about HFpEF differently.

Dr. Vora: Is losing calories good for the pancreas?

Dr. John Buse: I do think the same energetic concepts regarding mitochondria could have beneficial effects in the pancreas, as well in the beta cell in particular. It takes a different study design to demonstrate whether a drug is preserving beta cells than preventing hyperglycemia. I’m enthusiastic about that concept.

Dr. Angelo Avogaro: It is known since the early days by Dr. Ralph DeFronzo that SGLT-2s should improve insulin resistance, and they’ll probably delay diabetes onset. The problem is this is concerning DAPA-HF, but not EMPEROR. I wouldn’t stretch the result of the secondary endpoint. They probably magnify this effect in the prediabetes state.

Dr. Vora: SGLT-2 non-believers say the effect is all calorie loss and starvation. Why not just make people lose weight that way? This is also based on the DIRECT study in the UK, in which caloric reduction led to remission in type 2 as long as you were in the early few years of diabetes. What are your thoughts?

Dr. Buse: I think there is a kernel of truth. I do think calorie restriction is important in diabetes and that periods of fasting are important for metabolism. I do think there are things SGLT-2s do that go beyond. How important they are will take a long time to sort out in animal models. 2,500 gene transcription effects were seen in mice exposed to SGLT-2s; it’s not an either-or thing. Diet and fasting are important, and SGLT-2s are important in additional way, but it’ll take a while to prove that.

Dr. Avogaro: I think this is a slippery thing because there is no evidence that fasting decreases hard endpoints, so we shouldn’t confuse outcomes with intermediate outcomes. We can’t equate the action of SGLT-2s to prolonged fasting because the comparison doesn’t hold as a theory. When you fast, you increase ketones, but also increase free fatty acids, which are not good for your heart. SGLT-2 inhibitors selectively increase ketones, which is good.

Dr. Vora: From the kidney perspective, if a patient is failing on metformin, what do you give them first?

Dr. Hiddo Heerspink: From a purely kidney perspective, new diabetic kidney disease international guidelines will be published this week. It will state that metformin and SGLT-2s are the cornerstone for kidney disease. If A1c is not controlled in patients at high risk and with established CVD, then they recommend adding a GLP-1. This guideline recommendation comes from SGLT-2s showing protective effects for the kidney and heart failure. Remember, HF in DKD is most common cardiovascular complication. I think that’s a reasonable assumption, we know when GFR drops, there’s a higher risk of lactic acidosis, and metformin should be reduced or not used when GFR is below 30 ml/min. Now we have data that shows dapagliflozin can be safely initiated to eGFR of 25 ml/min. If you can’t start metformin because of GFR or side effects, SGLT-2s are the first choice for these patients.

Dr. Voda: Can you speak more on the renal effects in EMPEROR-Reduced?

Dr. Butler: There’s a bidirectional benefit. For patients with advanced CKD, the question is, do they get a cardiovascular benefit? Do patients with heart failure get CKD benefits? The signal we saw went both ways. HF patients get the same CKD benefit you’d get otherwise in diabetes patients and vice versa across the renal range for CV outcomes. It’s a win both ways.

Dr. Heerspink: I agree. This shows the interaction between the heart and kidney. If you connect CANVAS with CREDENCE or DECLARE with DAPA-CKD, and look across kidney function levels, effects are consistent on kidney prevention and HF prevention. This is different from what we’ve seen in ACEs and ARBs, which are only effective for people with high albuminuria and low eGFR. I feel that SGLT-2s are very strong, potent, kidney protective drugs. I’m interested in future head-to-head trials between SGLT-2s and ARBs

Dr. Butler: The issue with heart doctors is that most of the drugs we want to give effect GFR.

Dr. Voda: In people with Acute Kidney Injury (AKI), are outcomes better with SGLT-2 treatment?

Dr. Heerspink: By reducing sodium reabsorption, one can reduce the workload of the proximal tubule cell. This process is oxygen-dependent, so less oxygen required, the less hypoxic stress is in the proximal tubule. If there is a decrease in stress, SGLT-2s reduce the risk of AKI. It’s like a beta blocker for the kidneys in terms of that. A clinical trial is going to test the efficacy of this hypothesis.

Dr. Voda: Do peopel who recover from AKI do worse if they don’t reduce proteinuria?

Dr. Heerspink: Yes, there’s more insulin in trajectory of kidney function overtime, this is not totally linear, but there are many AKI episodes overtime. By alleviating those bumps, you may protect kidney function in the long term.

Dr. Voda: From mechanistic studies, how are the renal protective effects of SGLT-2s only seen in people with proteinuria reduction? Can we say that we are protecting the kidney from that?

Dr. Heerspink: We have to be careful because we can analyze these trials and see a mediation effect. We found that change in albuminuria is an important determinant of kidney protective effects. The AHA published a trial that showed that 40% mediation was explained by a reduction in albuminuria. 60% is not explained, which means there can be another mechanism. Multiple mechanisms may be coming together.

Dr. Voda: Do all patients respond by improvement of HFrEF with SGLT-2s, or is there a way to judge responders versus non-responders?

Dr. Butler: For every therapy, there are responders and non-responders. What is interesting is that maybe left ventricular remodeling with HFrEF is hallmark of response. Turns out, MRAs reduce mortality and don’t have much effect on cardiac remodeling, so it has a good positive, but not negative, predictive value. Also, there’s a study out that shows that if you stop therapy on HF patients, even if the heart has completely normalized, the EF goes down again- so there is a need to continue therapy. 

Dr. Voda: Do we need any more mechanistic studies? Or should we just get on and treat patients?

Dr. Buse: I think we do need to treat patients, but I think mechanistic studies will be important for future drug development, to tweak mechanisms more. The more it’s understood, the more likely the therapy’s used. It is a moral imperative to get patients with type 2 using these drugs that we know work.

Dr. Avogaro: I think that muscle is critical for heart and insulin resistance, and additional mechanistic effects in muscles and hemodynamics would be welcomed.

Dr. Heerspink: I agree with Dr Buse. It is important to understand the mechanism because it would improve uptake. Uptake is low despite compelling evidence; it’s our responsibility to implement findings in practice. Maybe more mechanism studies are needed to do that and determine which patient populations will benefit most.

Dr. Butler: I think mechanistic studies are needed because they can help us figure out that perhaps there are other disease states these drugs may treat as well. But we really do need to focus on implementation, as well. 

Mundipharma Symposium Gives New Perspectives on SGLT-2 Inhibitor CVOTs

We were excited to tune into the industry session, “Setting a new direction in managing type 2 diabetes – renal and CV perspectives,” sponsored by Mundipharma International Limited. This presentation featured three talks by Drs. Tina Vilsbøll (Steno Diabetes Center Copenhagn, Denmark), Oliver Schnell (Forschgruppe Diabetes e.V., Helmholtz Center, Germany), and David Wheeler (University College London, United Kingdom), who each reiterated the critical importance of treating cardiac and renal complications in type 2 diabetes. We were lucky to hear from Dr. Wheeler personally during ESC 2020, when we discussed implications and insights of the DAPA-CKD results in an exciting interview. During today’s presentation, special attention was given to the efficacy of SGLT-2 inhibitors in reducing the risk for hospitalization for heart failure and chronic kidney disease. As we have previously reported in our coverage of EMPA-REG OUTCOME, CREDENCE, and VERTIS CV, among other SGLT-2 CVOTs (including EMPEROR-Reduced and DAPA-CKD, most recently), these effects have been consistently demonstrated across members of the SGLT-2 inhibitor class. Presenters in this session went particularly deep into the details of CREDENCE, using the trial as an exemplar to demonstrate the tight correlation between improving cardiac health and improving kidney health with this single, innovative drug class. As the results from these large trials have become integrated more and more into diabetes guidelines, the presenters highlighted the potential for these drugs and others to recalibrate the focus of diabetes care in the clinic.

SGLT-2 Inhibitor Poster Highlights

Title	Authors	Details and Implications
Effect of empagliflozin on cardiorenal outcomes and mortality across BMI categories: subgroup analysis of the EMPA-REG OUTCOME trial with a focus on Asian patients	Qiuhe Ji, Linong Ji, Yiming Mu, Jiajun Zhao, Bernard Zinman, Christoph Wanner, Jyothis George, Isabella Zwiener, Kohjiro Ueki, Koutaro Yokote, Wataru Ogawa, Odd Erik Johansen	This study focused on a subset of data from EMPA-REG OUTCOME looking at cardiovascular outcomes in Asian populations (n=1517) across BMI categories of <25, 25 to <30, and ≥30 kg/m3. Empagliflozin reduced overall mortality, hypertensive heart failure, and cardiovascular deaths across all three BMI categories in both the total population and Asian subpopulation.
Empagliflozin facilitates sustained insulin dose reductions in patients with type 2 diabetes and cardiovascular disease: the EMPA-REG OUTCOME trial	Muthiah Vaduganathan, Naveed Sattar, David Fitchett, Anne Pernille Ofstad, Martina Brueckmann, Jyothis T. George, Subodh Verma, Michaela Mattheus, Christoph Wanner, Silvio E. Inzucchi, Bernard Zinman, Javed Butler	Subgroup analysis of EMPA-REG OUTCOME looking at type 2 patients with co-morbid cardiovascular disease who used insulin treatment at baseline. At four years, 12.4% of those treated with empagliflozin showed sustained reductions in insulin requirements, compared to just 6.5% of those in the placebo arm. The authors concluded that empagliflozin and perhaps other SGLT-2 inhibitors may precipitate meaningful, sustained, and appropriate reductions in insulin dose requirements.
Effects of intensive exercise, combined with dapagliflozin on body composition in type 2 diabetes: a randomised controlled trial	Ryotaro Bouchi, Yasuhiro Ono, Tatsuya Fukuda, Jun Itoh, Junji Kishimoto, Tetsuya Yamada, Noriyuki Sonoda, Takato Takeuch, Yoshihiro Ogawa	In a randomized trial, the authors analyzed the effect of exercise on preventing the loss of muscle mass in type 2 patients taking dapagliflozin (n=146). Loss in muscle mass typically accompanies reductions in body fat mass with SGLT-2 inhibitor treatment. Weight and A1c were significantly reduced in both exercise and usual treatment groups. Notably, participants who followed a moderate exercise regime (home resistance training) in combination with dapagliflozin showed greater reductions in fat mass than those who took dapagliflozin without exercising.
Patient preferences for newer oral therapies in type 2 diabetes	Gianluigi Savarese, Abhinav Sharma, Christianne Pang, Richard Wood, Jyothis T. George, Nima Soleymanlou	This study assessed the safety and efficacy of empagliflozin compared to DPP-4 inhibitor sitagliptin and oral GLP-1 receptor agonist semaglutide in patients with type 2 diabetes. 41% majority of patients preferred empagliflozin over the other treatment options. Despite evidence of its cardiovascular benefits, patients were concerned about empagliflozin’s genital infection risk.
Prominent effects of SGLT2 inhibitor canagliflozin combined with DPP-4 teneligliptin on postprandial hyperglycaemia at breakfast and dinner	Kyu Yong Cho, Hiroshi Nomoto, Shinichiro Kawata, Kazuhisa Tsuchida, Kazuno Omori, Akinobu Nakamura, Tatsuya Atsumi, Hideaki Miyoshi	A secondary analysis of CALMER (n=99) evaluated CGM metrics such as time-in-range and postprandial hyperglycemia in patients taking both SGLT-2 canagliflozin and DPP-4 teneligliptin. TIR increased from 71.2% to 82.7% (p<0.001) in the group receiving combination therapy of canagliflozin + teneligliptin. The reduction in time above range was significantly greater in the combination group vs. the monotherapy group that switched from teneligliptin to canagliflozin (-14.8% vs. -7.5%, p<0.05). ROC analysis showed that participants on combination therapy had significantly lower postprandial hyperglycemia after meal tolerance tests when compared to participants on monotherapy (p<0.05).
Acid-base changes during diabetic ketoacidosis in type 1 diabetes with/without SGLT2 inhibitor	Ines Mursic, E. Svehlikova, W. Regittnig, M. Urschitz, M. Wolf, M. Brunner, T. Augustin, C. Magnes, A. Eberl, T. Heise, O. Klein, H Sourj, T. Pieber	This randomized, cross-over trial analyzed the impact of SGLT-2 inhibitors on the development of diabetic ketoacidosis (DKA) in type 1 diabetes patients (n=16). During DKA development and recovery, individuals on SGLT-2 inhibitors demonstrated plasma glucose levels near normal range, potentially preventing the detection of DKA. Individuals on SGLT-2s demonstrated higher ketone levels, and increased baseline glucagon and free fatty acids. No significant effect on pH and acidosis was found.
Dapagliflozin induces kidney augmentation: potential mechanism for SGLT2 inhibitor induced nephroprotection	XingLin Wu, Yanling Zhang, Kerri Thai, Linda Nghiem, Kim A. Connelly, Richard E. Gilbert	Experiments were conducted on rats to determine if the SGLT-2 inhibitor dapagliflozin affected kidney size and function. Under increased workloads, kidneys have the ability to increase in size and function. The study found that the dapagliflozin increased cell proliferation in the kidneys in the first few days of treatment, leading to hypertrophy and enlarged kidneys. Dapagliflozin did not significantly affect eGFR as previously observed in kidney transplants and diabetes.
Glucose kinetics during oral glucose challenge following administration of exenatide and dapagliflozin alone and in combination in type 2 diabetes	Miriam Alatrach, Christina Agyin, Nitchakarn Laichuthai, Olga Lavrynenko, John Adams, Muhammad Abdul-Ghani, Curtis Triplitt, Ralph DeFonzo, Eugenio Cersosimo	This study analyzed the glucose kinetics of dapagliflozin alone in comparison with dapagliflozin/exenatide combination therapy in patients with type 2 diabetes (n=28). Plasma glucose levels increased when dapagliflozin was used on its own, but there was little increase in plasma glucose from baseline when the dapagliflozin/exenatide combination therapy was used. Dapagliflozin/exenatide combination resulted in lower glucose in peripheral circulation and a higher rate of urinary glucose excretion, suggesting that  long-term SGLT-2/GLP-1 combination therapy is more effective than SGLT-2 inhibitor monotherapy.
Luseogliflozin protects pancreatic beta cells via improving mitochondrial metabolism	Yuki Yamauchi, Akinobu Nakamura, Kiyohiko Takahashi, Shinichiro Kawata, Kazuhisa Tsuchida, Kazuno Omori, Hiroshi Nomoto, Hiraku Kameda, Kyu Yong Cho, Takashi Yokota, Toshihisa Anzai, Hideaki Miyoshi, Yasuo Terauchi, Tatsuya Atsumi	SGLT-2 inhibitors like luseogliflozin are known to protect pancreatic beta cells. This study aimed to elucidate the mechanism underlying such benefits. The expression of Nkx6.1 – which  is related to beta cell proliferation and maturation – was increased following luseogliflozin treatment. Luseogliflozin also reduced the generation of reactive oxygen species (ROS). Increased gene expression of Nkx6.1 can improve glucose uptake and mitochondrial metabolism, protecting pancreatic beta cells.
Differential effects of SGLT2 inhibitors on mitochondrial oxidative phosphorylation, glucose uptake, cell energy level and metabolism in HepG2 cells and HUVECs	Elmar Zügner, S. Hagvall, C.S. Elmore, H. Sourij, P. Kotzbeck, R. Esterline, S. Moosmang, H.-C. Yang, C. Magnes	The authors investigated the effects of canagliflozin, dapagliflozin, ertugliflozin, and empagliflozin on cellular metabolic function in human hepatoma cells (HepG2) and human umbilical vein endothelial cells (HUVECs). Canagliflozin significantly lowered oxygen consumption rate (by 60%) and extracellular acidification rate (by 45%) in HUVECs. In HepG2 cells, conversely, canagliflozin elevated extracellular acidification rate. In comparison, dapagliflozin, ertugliflozin, and empagliflozin produced non-significant effects on oxygen consumption rate and extracellular acidification rate. Metabolomics also revealed that canagliflozin worked through a distinct mechanism of action compared to the other SGLT-2 inhibitors investigated. It is clear that of the SGLT-2 inhibitor class, canagliflozin has a unique impact on cellular respiration. However, the clinical implications of these differences remain unclear.

Insulin Highlights

Lilly’s Ultra-Rapid Insulin Lyumjev Shows Superior PPG and Lower Hypoglycemia than Humalog in Type 1 Pump Users

New phase 3 data for Lilly’s recently approved Lyumjev (ultra-rapid acting insulin lispro) shows superior PPG control and lower hypoglycemia compared to Humalog (insulin lispro) when administered by pump in people with type 1 diabetes. The 16-week study enrolled 432 people with type 1 diabetes using insulin pumps (mean age 46 years, mean A1c 7.6%, mean diabetes duration 26 years), randomizing them 1:1 to Lyumjev or Humalog. On the primary outcome of A1c change, Lyumjev showed non-inferiority to Humalog (-0.7% vs. -1.0%, p=0.565). Furthermore, in standardized mixed meal tolerance tests, Lyumjev showed significantly greater reductions in 1-hour (difference: -24 mg/dl, p<0.05) and 2-hour post-prandial glucose than Humalog (difference: -28 mg/dl, p<0.05). Notably, participants wore blinded CGM prior to randomization and at weeks 8 and 16 of the study and time-in-range (TIR) was measured as a secondary endpoint. According to this CGM data, there was no difference in TIR between Lyumjev- and Humalog-treated participants (57.9% vs. 57.5%), but those on Lyumjev showed significantly less time spent in severe hypoglycemia (22 vs. 29 minutes/day, p=0.006) and moderate (63 vs. 77 minutes/day, p=0.006) hypoglycemia. As for adverse events, Lyumjev-treated participants had a higher rate of infusion-site related adverse events (60.5% vs. 44.7%) but the rate and incidence of severe hypoglycemia and DKA did not differ between the groups. Overall, this data adds further support for the strong clinical profile of Lyumjev, particularly in the context of pump therapy. Lyumjev’s indication in the EU, Japan, and Brazil is for both MDI and pump therapy, but it is only approved in the US for use with MDI. We certainly hope this latest phase 3 trial puts to rest any concerns over the use of this therapy with pumps.

Arecor’s Ultra-Rapid Insulin AT247 Shows Faster Onset/Offset and Superior Glucose-Lowering to NovoLog and Fiasp in Phase 1 Study

Dr. Eva Svehlikova (Medical University of Graz) presented encouraging phase 1 PK/PD results for Arecor’s ultra-rapid insulin candidate AT247 in people with type 1 diabetes. The trial enrolled 19 men with type 1 diabetes (mean age 35 years, mean A1c 7.1%, mean diabetes duration 20 years), each of whom received a single subcutaneous dose (0.3 units/kg) of either AT247, insulin aspart (Novo Nordisk’s NovoLog), or fast-acting insulin aspart (Novo Nordisk’s Fiasp) in a cross-over manner during a euglycemic clamp procedure. Overall, AT247 had a significantly faster onset of glucose-lowering action with a treatment difference of -23 and -9 minutes as compared with NovoLog and Fiasp respectively. Furthermore, the glucose-lowering effect in the first hour after dosing, measured in terms of area under the curve of the glucose infusion rate (AUC_GIR0-6omin), was significantly greater for AT247 than NovoLog or Fiasp (212 vs. 49 vs. 91 mg/kg, respectively). Finally, a measure of the offset of insulin exposure (Time to late 50% Cmax insulin) was significantly faster for AT247 with a treatment difference of -22.5 and -20 minutes as compared with  NovoLog and  Fiasp respectively. Overall, these results point to a very promising clinical profile for AT247, and we look forward to seeing how the candidate performs in larger, longer-term studies. In the initial announcement of these results, Arecor CEO Sarah Howell indicated that the company is eyeing AT247 for use in future artificial pancreas systems, in addition to use for traditional MDI insulin therapy. We agree that the full potential of next-gen ultra-rapid insulins may be best unlocked in AP systems – currently, ultra rapid-acting insulins such as Fiasp have struggled to make a large impact on the market.

Dr. Vivian Fonseca Touts Real-World Benefits of Early GLP-1/Basal Insulin Combination Therapy in Patients with Type 2

In a moderated poster, the brilliant Dr. Vivian Fonseca (Tulane University School of Medicine) presented real-world results supporting the simultaneous initiation of basal insulin and a GLP-1, over a sequential regimen. While current ADA/EASD guidelines state that patients who have not met their A1c goals on either therapy alone can consider adding the other injectable as a treatment intensification, there is little evidence on whether or not the timing of intensification impacts clinical outcomes. Using retrospective data on patients with type 2 and A1c ≥9% on OADs (n=6,339) from the Optum Humedica EMR database, Dr. Fonseca and his collaborators discovered that patients who initiated the therapies “simultaneously” (≤30 days apart) achieved the strongest glycemic outcomes at six and 12 months for all four primary endpoints: proportion of individuals achieving (i) an A1c <7%; (ii) an A1c <8%; (iii) ≥1% reduction in A1c from baseline; and (iv) ≥2%reduction in A1c, followed by individuals who started the therapies within 31-90 days of each other; Baseline A1cs for these two cohorts were 10.9% (±1.54%) and 10.7% (±1.51%), respectively.

• In terms of clinical implications, these findings suggest that first treating patients with either a basal insulin or GLP-1 and waiting to assess treatment efficacy before adding the other may not be the most effective strategy. These data harken back to results from VERIFY, presented at EASD 2019, which found that the early combination of metformin and Novartis’ DPP-4 inhibitor Galvus (vildagliptin) doubled the time to initial treatment failure vs. monotherapy and sequential intensification. As Dr. Dan Drucker convincingly put it, “Diabetes is a serious disease! Think more like an oncologist (early combination therapy) rather than a traditional treat to failure approach. Well-tolerated drug combinations with established safety profiles should be used earlier.” Although his words were in regard to Galvus and metformin, we think they certainly apply in this case as well.

Novo Nordisk Symposium on Insulin Therapy: Contemporary Strategies for Combatting Hypoglycemia and Clinical Inertia

Chaired by Global Chief Medical Officer Dr. Stephen Gough, Novo Nordisk today hosted a highly engaging and interactive symposium entitled “Exploring new paradigms in insulin therapy to improve outcomes.” Featuring a star-studded lineup of expert clinicians and researchers, this symposium addressed four key movements in contemporary insulin therapy: (i) recalibrating our understanding of hypoglycemia and its diverse health and economic consequences; (ii) incorporating digital technologies to optimize the amount and timing of insulin dosing; (iii) exploring fixed-ratio combination therapies (FRCs) to reduce treatment complexity; and (iv) investigating the potential of Novo Nordisk’s once-weekly insulin icodec to “redefine diabetes management.” This session called to attention the significant barriers faced by clinicians and patients alike to achieving optimal diabetes care, while offering practical and promising solutions – both in therapy and technology – to addressing them. See below for updates on some of the most exciting programs being developed by Novo Nordisk, including the newly released results of the company’s phase 2 insulin icodec program. Of note, the results from one trial in this program was published earlier this week in the New England Journal of Medicine.

• Dr. Simon Heller (The University of Sheffield, United Kingdom) discussed two exciting initiatives aimed at improving our understanding and management of hypoglycemia: The International Hypoglycemia Study Group (IHSG) and Hypo-RESOLVE (an Innovative Medicines Initiative project). IHSG was formed in 2013 and consists of 15 clinicians and clinical investigators. The initiative has made great strides in re-classifying hypoglycemia to include a third level encompassing major/serious hypoglycemia between 2.8-3.0 mmol/L (50-55 mg/dL); at this blood glucose level, patients are at increased risk of cognition impairment, cardiac arrhythmias, hypoglycemia unawareness, and health economic consequences, among other adverse outcomes. Previously, hypoglycemia had only been characterized as “highly relevant” (below 3.9 mmol/L or 70 mg/dL) or “severe.” This new categorization was incorporated into position statements written by IHSG in 2017, which were accepted by ADA/EASD. Since then, this reclassification effort has been similarly accepted by JDRF, ISPAD, and EMA, with FDA yet to provide their decision. ATTD also adopted the position statement, with a caveat for individuals using continuous glucose monitoring (CGM): as CGMs are sensitive to detect transient episodes of hypoglycemia (many of which may go undetected), ATTD issued a consensus statement declaring that only readings below threshold for greater than 15 minutes should be considered a hypoglycemic event. Akin to the goals of IHSG, Hypo-RESOLVE was developed with the intention of developing a better and more universal hypoglycemia classification system, to be incorporated into standard guidelines. Hypo-RESOLVE also aims to standardize clinical data collection, assess “value for patients” and funders, and better unravel the mechanisms and clinical implications of hypoglycemia. This initiative is comprised of 23 industry, academic, and associated partners across Europe.

• To address the issue of clinical inertia in insulin intensification, Dr. Martin Haluzík (Diabetes Centre, Institute for Clinical and Experimental Medicine, Czech Republic) highlighted the diverse benefits of Novo Nordisk’s IDegLira (insulin degludec/liraglutide) and IGlarLixi (insulin glargine/lixisenatide) – two fixed-ratio combination (FRC) therapies comprised of a GLP-1 receptor agonist and a basal insulin. Dr. Haluzík paid particular attention to the results of the DUAL VII trial evaluating the efficacy of IDegLira  versus basal-bolus insulin therapy in patients with type 2 diabetes. While IDegLira gave rise to a similar and non-significant reduction in A1c as basal-bolus therapy (only slightly less than a -1.5% reduction), patients in the IDegLira arm experienced significantly fewer hypoglycemic episodes and required lower insulin doses at the end of trial compared to those in the basal-bolus arm. Further, patients taking IDegLira lost weight over the course of the study while those taking basal-bolus insulin gained weight. Similar results have been achieved in a phase 3 study evaluating IGlarLixi versus basal-bolus insulin. From real world evidence (RWE) – the PAX-1 study in the US; the Swedish NDR; and EXTRA in Sweden, the UK, Switzerland, Austria, and Germany – IDegLira exhibited a similar efficacy profile as that shown in DUAL VII. At the end of each study, patients who shifted to IDegLira in EXTRA and Swedish NDR showed statistically significant A1c reductions from baseline. Further, in all three real world studies, all patients experienced statistically significant body weight reductions. Given these striking safety and efficacy results – both in controlled and real-world settings – Dr. Haluzík advocated for this FRC class as a way to reduce the complexity of therapeutic regimens for diabetes patients.

• Dr. Harpreet Bajaj (LMC Healthcare, Canada) concluded the symposium by presenting the results from Novo Nordisk’s encouraging phase 2 program for once-weekly insulin icodec. The program consisted of three trials in type 2 diabetes: one assessing icodec in insulin-naïve patients; one in patients switching from daily basal insulin to weekly icodec; and one assessing three distinct icodec titrations in insulin-naïve patients. As a reminder, the first of these trials was published earlier this week in the New England Journal of Medicine, and all phase 2 results were presented on the first day of EASD 2020 and earlier this year at ADA 2020. Across each study, icodec has showed comparable glucose-lowering efficacy and safety to insulin glargine. Of note, in the trial involving insulin-naïve patients, those in the icodec arm required significantly lower insulin doses at 26 weeks compared to those in the glargine arm. In a poster presented by Dr. Bajaj at the conference, icodec with 100% loading dose also significantly increased time-in-range (TIR) compared to glargine in the trial involve patients switching from daily basal insulin to icodec. In another poster presented at EASD, the middle of three icodec titration algorithms – which were distinguished by titration increments and targets – was shown to increase TIR to the greatest extent relevant to glargine. These encouraging and exciting results, as Dr. Bajaj suggested, may potentially “redefine” how we manage diabetes.

Sanofi Symposium Positions Toujeo as Best Basal Insulin for Elderly (>70) and Patients with Renal Impairment (eGFR <60)

Dr. Alice Cheng (University of Toronto) argued in favor of next-gen basal insulins for all type 2s and endorsed Sanofi’s Toujeo (insulin glargine U300) in particular for older individuals and patients with reduced kidney function. She summarized findings from BRIGHT (n=929), a Sanofi-sponsored head-to-head trial of Toujeo vs. Novo Nordisk’s Tresiba (insulin degludec). During the first 12 weeks of the study, the titration phase, participants randomized to Toujeo experienced significantly less hypoglycemia <70 mg/dl (OR=0.74, p=0.03) and <54 mg/dl (OR=0.63, p=0.044) than those on Tresiba. Dr. Cheng explained that this result is meaningful because baseline hypoglycemia risk is greatest during titration, when insulin dose is changing and A1c is falling rapidly. There was no significant difference in hypoglycemia during BRIGHT’s maintenance phase (week 12-24) or overall, and it remains an open question why Toujeo’s hypoglycemia benefit might fade. Dr. Cheng also acknowledged findings from CONCLUDE (Novo Nordisk’s equivalent head-to-head investigation) of no significant difference in symptomatic hypoglycemia with the two advanced basal insulins. We should note that Tresiba demonstrated a very compelling hypoglycemia benefit over Sanofi’s first-gen Lantus (insulin glargine U100) in DEVOTE, and those results are now included in the US product label. Even data showing similar hypoglycemia rates with Toujeo vs. Tresiba is therefore extremely positive for Sanofi, suggesting that both next-generation basals can greatly reduce the hypoglycemia that patients endure. This is the critical message in our view – that HCPs should be prescribing Toujeo or Tresiba over first-gen insulin analogs whenever possible – and Dr. Cheng highlighted this point throughout her talk.

Of course, as the diabetes field pushes toward personalized medicine, it will be helpful to understand which patients are ideal candidates for Toujeo vs. Tresiba. Dr. Cheng presented two subgroup analyses of BRIGHT that begin to answer this question. In patients with eGFR <60 ml/min/1.73m², Toujeo gave a greater magnitude reduction in A1c than Tresiba (mean difference 0.43%, p=0.02 for the interaction) without increasing hypoglycemia. There was no significant interaction between age above/below 65 and A1c decline in a pre-specified subgroup analysis. A post-hoc looked at age above/below 70 and did find greater A1c reduction with Toujeo vs. Tresiba in patients older than 70 (mean difference 0.34%), again without elevating hypoglycemia risk. Some will certainly argue that it’s a stretch to change the definition of “elderly” from 65 to 70 in order to find a favorable result for Toujeo. Dr. Cheng noted that post-hocs are only exploratory, but also defended that age 70 is a better biological definition of “elderly.” As she put it, we are accustomed to the 65 threshold because of the generally accepted retirement age, and even that is changing. From a commercial standpoint, it’s interesting to see Sanofi promoting Toujeo for specific pockets of the type 2 diabetes population (at least to the scientific community at EASD), as the product lags Tresiba in revenue and market share ($263 million 2Q20 sales vs. $318 million for Tresiba).

• Dundee’s Professor Rory McCrimmon opened this Sanofi-sponsored symposium by establishing the burden of hypoglycemia in the real world. Incidence of hypoglycemia is most certainly lower in clinical trials than in real-world practice, because on average, providers and patients are less engaged in the real world vs. an RCT. The implication, then, is that next-gen basal insulins could show even more dramatic benefits in the real world than we see in clinical trials – but only if they are titrated effectively, and only if HCPs assuage fear of hypo in their patients. Professor McCrimmon cited DAWN 2 results showing that 56% of diabetes patients fear hypoglycemia. He linked this fear to suboptimal insulin titration and to therapeutic inertia. In the UK, a patient with type 2 diabetes spends a mean 3.7 years above A1c target before initiating basal insulin (this is egregious, and downright sad). Professor McCrimmon emphasized the importance of patient education around hypoglycemia and how to mitigate it. He also shared strategies for optimal/effective titration. Dr. Cheng echoed this message in her remarks as well: “If you optimally titrate basal insulin, it works. That might seem like an obvious statement, but in the real world we don’t do that as well as we do in clinical trials.”

3. Sanofi Symposium Highlights the Promise of Second-Generation Basal Insulin in Glycemic Control and Avoiding Hypoglycemia

In one of the first therapy symposia of EASD 2020, speakers discussed the wide-reaching benefits of Toujeo in comparison to first-generation insulins (insulin degludec) and fellow second-generation insulins (Lantus). While one portion of the symposium, moderated by Dr. Thomas Danne, focused on making insulin management relevant to individuals and their goals, while another session, moderated by Dr. Julio Rosenstock, focused on the promise of basal/GLP-1 fixed ratio combination treatment.

• Drs. Eric Renard, Thomas Danne, Alice Cheng, and Luigi Meneghini spoke on increasing the relevance of insulin management. Dr. Renard began the session with a talk on the current unmet needs in the type 1 diabetes population. According to data from T1D Exchange, only 17% of youth and 21% of adults achieved target A1c levels of 7.5% or below. Based on the results of the SAGE study, only 24.3% of people with type 1 are reaching these targets globally. Even though second-generation basal insulins are being adopted, education is still needed to ensure effective use (especially in terms of dose titration), with technology holding great potential to assist.  Dr. Danne continued this theme with his discussion of generational change in the use of basal insulin among children and adolescents. This population faces unique challenges in glycemic control, with A1c targets ranging from 6.5% to 8%, depending on risk or awareness of hypoglycemia. Based on the EDITION JUNIOR and BEGIN YOUNG RCTs, Toujeo showed similar glycemic efficacy and hypoglycemia incidence rates to Lantus, as well as lower rates of severe hypoglycemia and hyperglycemic DKA. Dr. Danne commented that these trends may be related to the level of supervision and care by clinical trial staff in these studies.

• Building off the importance of confidence in insulin use, Dr. Alice Cheng took the virtual stage to discuss the necessity of building confidence in insulin self-administration. Fear of hypoglycemia is one of the major barriers to insulin initiation and titration. With this, most insulin dose and glycemic control goals are met early in treatment (during the first 12 weeks), making confidence in self-titration early in one’s diabetes journey an integral step to most efficiently using second-generation products to their full capacity. She stated that the difference between expected and actual titration (shown in the graphic below) could be the root cause of the rise in complications in people with diabetes. Still, the type of insulin used remains important; according to the BRIGHT study, Toujeo is non-inferior in reducing A1c and incidence of hypoglycemia compared to Tresiba. Moreover, results of the TAKE CONTROL study showed that patients can effectively self-titrate Toujeo to levels comparable to that of an HCP, as well as integrate dose-support technology to increase the accuracy of self-titration. Dr. Luigi Meneghini discussed similar topics in his talk, which centered on promoting glycemic control in the elderly and those with renal disease. His main takeaway was that diabetes management in these populations requires a personalized approach. With this, Toujeo may be a promising insulin selection, as results from subgroup analyses of the BRIGHT study show glycemic control without increased risk of hypoglycemia.

• Basal/GLP-1 fixed-ratio combinations can offer a solution to the need for early treatment intensification and combination therapy for patients needing more glycemic control. In his opening session, Dr. Julio Rosenstock stated that there is a need for earlier treatment with combination metformin, SGLT-2 inhibitor, and/or GLP-1 agonist combination therapy. He added that basal/GLP-1 FRCs should be implement more often because they are an easy and effective treatment option. Drs. Francesco Giorgino and Leigh Perreault continued to make the case for FRCs, citing data from DISCOVER that highlight the need for global improvements in type 2 diabetes management. To accomplish this, treatment paradigms must shift to support early sustained glycemic control and personalized treatment intensification. According to Dr. Perreault, treatment of type 2 diabetes also needs a multifactorial approach to achieve control, which basal/GLP-1 FRCs can support according to data from treatment with iGlarLixi. To conclude the symposium, Dr. Juan Frias provided an overview of how to select patients for FRCs, which requires ample patient education and managing expectations.

Dr. Julio Rosenstock Praises Potentially Transformative Nature of Once-Weekly Insulin Icodec for Patients with Type 2

Closing out the Novo Nordisk-sponsored symposium, the illustrious Dr. Julio Rosenstock (Dallas Diabetes Research Center) delivered a powerful case for why once-weekly basal insulin candidate insulin icodec “will transform how we treat people with type 2 diabetes.” In brief, fewer injections for the weekly insulin may increase adherence and compliance for patients with type 2, along with reducing barriers to initiation of insulin therapy that currently contribute to therapeutic inertia. As a reminder, encouraging phase 2 outcomes were presented at ADA 2020, in which insulin icodec showed comparable glucose-lowering efficacy and safety to insulin glargine U100. Out of the impressive results, Dr. Rosenstock emphasized that: (i) rates of Level 2 or 3 hypoglycemia were low for both insulin treatments; and (ii) the insulin up- or down-titration algorithm for insulin icodec was quite simple to use. Dr. Rosenstock also noted that although he is frequently asked if hypoglycemia with insulin icodec is longer in duration, there is no evidence suggesting that hypoglycemia duration differs between weekly or daily insulin. Promisingly, a phase 3 trial for the candidate is slated to begin before year’s end.  

• Along with Dr. Rosenstock’s presentation, two posters highlighted the Time-in-Range benefits and possible dosing schemes for the candidate. First, a poster from Dr. Julio Rosenstock’s group showed encouraging safety and efficacy data for insulin icodec in comparison to insulin glargine. Participants (n=154; mean A1c 7.9%, mean BMI 29.8 kg/m²) were randomized 1:1:1 to icodec with a loading dose (i.e. doubling the initial icodec dose, a measure that early PK/PD studies indicated would prevent an initial increase in blood glucose as icodec reached steady-state), icodec without a loading dose, or insulin glargine for 16 weeks. Very notably, time-in-range (TIR) was the study’s primary endpoint, measured using Dexcom G6 CGMs. While there was no difference in baseline TIR between the three study groups (55-59%), within the final two weeks of the study participants on icodec with a loading dose showed significantly higher TIR (73%) than those on icodec without a loading dose (66%) or insulin glargine (65%). This translates to nearly 2 hours (1 hour and 53 minutes, to be precise) of additional TIR with icodec with a loading dose. This icodec dosing scheme was further associated with a numerically greater reduction in the secondary outcome of A1c (-0.77%), but this was not significantly different from the A1c reductions seen with icodec without a loading dose (-0.47%) and insulin glargine (-0.54%). Furthermore, rates of combined moderate and severe hypoglycemia were similar for icodec with a loading dose and insulin glargine (78.0 and 79.4 events per 100 patient years, respectively) and numerically lower for icodec without a loading dose (14.8 events per 100 patient years). Overall, this study confirms that 1) switching to insulin icodec is well-tolerated and provides effective glycemic control without additional risk of hypoglycemia; and 2) when switching from a once-weekly basal insulin to icodec, using an initial loading dose helps prevent transient “highs” without increasing hypoglycemia risk, thereby maximizing TIR. This second point could have important implications for the design of future phase 3 trials of the candidate, and we will be watching this closely.

• While Dr. Rosenstock expressed extreme confidence in the potential benefits of insulin icodec in the type 2 population, his statements on patients with type 1 were far more guarded. He noted that any investigation on this front would have to be performed in a “very carefully designed study,” assessing patients with a very stable basal insulin dose, who primarily rely on prandial insulin. His final words on this? “There is potential.” 

Insulin Poster Highlights

Title	Authors	Details and Implications
Dosage profile of insulin glargine U100 and U300 in patients with type 2 diabetes in real clinical practice: DosInGlar study	Natalia Duque, Esther Artime, Irene Romera, Jeremie Lebrec, Silvia Díaz, Antoni Sicras, Jesús Reviriego	Observational retrospective cohort study compared mean dosage of Insulin Glargine (InGlar) and mean A1c in 1112 patients (n=556 in the matched cohorts) with type 2 diabetes who initiated treatment with InGlar U100 or U300 in Spain. After 6, 12, and 18 months of follow-up, the mean dose of InGlar (U/kg/day) was higher in patients treated with InGlar U300 compared to InGlar U100 (10.3% greater at 12 months and 12.8% greater at 18 months). Dosage increase was accompanied by similar reductions in A1c, and at equal insulin glargine price/unit in Spain, increased dose requirements of InGlar U300 would result in higher costs.
Therapy trends in initial 6 months of the first large-scale longitudinal nationwide study on management and real-world outcomes of diabetes in India (LANDMARC)	Hemant Thacker, Romik Ghosh, Ashok Kumar Das, Ambrish Mithal, Shashank Joshi, KM Prasanna Kumar, Sanjay Kalra, AG Unnikrishnan, Bipin Sethi, Senthilnathan Mohanasundaram, Shalini K Menon, Arjun Nair, Nadeem Rais, Subhankar Chowdhury, LANDMARC Study group	Reported on first six months of a three-year study (March 2018 to March 2021) observing therapy patterns of 6236 patients with type 2 diabetes throughout India. At baseline, OADs were the most commonly used diabetes therapy with 74.5% participants taking only OADs, 24.4% taking both OAD and insulin, and 0.4% taking only insulin. After six months, the proportion of participants taking only insulin increased to 19% and the proportion taking only OADs reduced to 62.8%. Concluded that insulin regimen was more effective than OADs since improvement was more evident in the insulin subgroup compared to insulin naïve subgroup (p<0.0001). Intensifying treatment with more that 3 OADs may not be effective in improving glycemic parameters.
Effectiveness of premix insulin in type 2 diabetes: a retrospective UK cohort study	Edward Jude, Amar Ali, Rifat Emral, Nithya Nanda, Robert Lubwama, Karen Palmer, Alka Shaunik, Philip Raskin, Fernando Gómez Peralta, Carlos Trescoli	Investigated the effectiveness of premix insulin regimens for achieving glycemic control in UK adults with uncontrolled type 2 diabetes (A1c 9%) who already take at least two OADs. For the 974 people who initiated premix between 2010 and 2016, the probability of achieving glycemic control (A1c %) was highest at 3-6 months (18%), 15% at 9-12 months, 14% at 6-9 months, and lowest from 21-24 months (0%). This low probability suggests a high unmet need for early and timely therapy intensification. Limitations include lack of information on certain baseline characteristics, preventing analyses to determine whether there was an association between these characteristics and glycemic outcomes.
A retrospective, observational, cohort study of the use and effectiveness of basal-bolus or premixed insulin in Japanese people with type 2 diabetes	Mike Baxter, Hideaki Miyoshi, Takeshi Kimura, Masakatsu Hattori, Yukiko Morimoto, Masami Tamiwa, Dion Marinkovich, Takahisa Hirose	Study analyzed A1c outcomes, treatment changes, and hypoglycemia of Japanese patients initiating basal-bolus regimens (BB) or premixed prandial and basal insulins (premixes) after failure of OAD drugs to control A1c. In BB and premix cohorts, respectively, 45.8% vs. 79.7% of people had no change in therapy, of whom 46.2% vs. 37.0% achieved A1c < 7% within 3 months. In the BB cohort, 45.9% switched to a lower injection frequency regimen. Among people who did not switch therapy, 27% of the BB cohort and 16% of the premix cohort experienced hypoglycemia. BB patients were younger, had higher A1c, and had shorter disease duration than premix patients. Although A1c reductions were higher for BB patients, more BB patients experienced hypoglycemia, suggesting that BB regimens may be used for a limited duration with early improvements in pre- and post-prandial glycemic control before treatment is simplified.
Effectiveness and safety of Gla-300 vs. IDeg-100 evaluated with continuous glucose monitoring profile, in adults with type 1 diabetes in routine clinical practice in Spain: OneCARE study	Ignacio Conget, Margarita Gimenez, Mireia Borrell, Juan Caro, Cristobal Morales, Elias Delgado, Miguel Mangas	This observational, retrospective cohort, cross-sectional, multicenter study compared the effectiveness and safety of insulin glargine 300 U/ml (Gla-300) with degludec 100 U/ml (IDeg-100) by analyzing the percentage of time in range (TIR, 70-80 mg/dl, 3.9 mmol/l) during a four-week period, measured using continuous glucose monitoring (CGM) for people with type 1 diabetes. Concluded that for controlled type 1 diabetes patients switching from the first-generation basal insulins, Gla-300 and IDeg-100 are equally effective at controlling glucose. There is no significant difference in TIR or glycemic variability between the two treatment groups, except for at night, when people on Gla-300 spent more TIR than people on IDeg-100.
Insulin glargine 300 U/ml vs. first-generation standard-of-care basal insulin analogues in adults with type 2 diabetes: impact of renal function in the ACHIEVE Control study	Luigi Meneghini, Alice Cheng, Pierre Evenou, Jasvinder Gill, Ayad Mohamed, Guillermo Umpierrez	Study performed a post hoc analysis of the ACHIEVE Control study, exploring outcomes in subgroups with variable renal function. Analyzed the subgroups of estimated glomerular filtration rate (eGFR) of <60, 60 to <90, and 90 ml/min/1.73 m2 at 6 and 12 months after beginning medication. Within subgroups, OR suggest a beneficial trend of Gla-300 vs. SOC-BI for avoiding hypoglycemia in adults with impaired renal function (OR > 1.40).
Efficacy and safety of fast-acting insulin aspart in adults with type 2 diabetes with different insulin requirements	Giorgio Sesti, Peter Senior, Magnus Ekelund, Edward Franek, Esteban Jódar, Wendy Lane, Naveen Rathor, Olöf Thórisdóttir, Alejandra Oviedo	This post-hoc analysis of Onset 9 assessed the safety and efficacy of fast acting (FA) vs. insulin aspart (IAsp), both with insulin degludec, in adults with long-standing type 2 by insulin requirement. Treatment difference in change from baseline in PPG increments for FA vs. IAsp was unaffected by bolus insulin dose. In adults with T2D and high bolus insulin requirements, risk of hypoglycemia may be lower with FA vs. IAsp.
Treatment satisfaction in people with type 2 diabetes receiving basal insulin: results from real-world and randomised controlled studies with insulin glargine 300 U/ml	Stewart Harris, Frank Snoek, Kamlesh Khunti, Luigi Meneghini, Jukka Westerbacka, Aude Roborel de Climens, Felipe Lauand	This study (n=8,014) investigated patient satisfaction following basal insulin (BI) treatment in studies of insulin glargine U300 (Gla-300) vs. other BIs Participants who switched to Gla-300 had improved overall patient satisfaction and reduced perceived frequency of hyperglycemia after six months of treatment, irrespective of baseline treatment
Dual I China: improved glycaemic control with IDegLira versus its mono-components in Chinese patients with type 2 diabetes uncontrolled on oral antidiabetic drugs	Weiquing Wang, Qui Zhang, Bue Ross Agner, Guang Ning, Shen Qu, Lei Liu, Yongde Peng, Bin Luo, Ming Liu, Guixia Wang, Karolina Stachlewska	Insulin degludec/liraglutide (IDegLire) is effective and well-tolerated in patients with T2D on uncontrolled oral antidiabetic drugs This data is consistent with results from the global DUAL trials.
Dual II China: superior HbA1c reductions and weight loss with IDegLira vs insulin degludec in Chinese patients with type 2 diabetes inadequately controlled on basal insulin	Yu Pei, Xiangjin Xu, Bue Agner, Dongmei Li, Xiaolin Dong, Tomoyuki Nishida, Jun Liu, Yibing Lu, Yiming Mu, Lei Liu, Bin Luo, Ming Liu	Insulin degludec/liraglutide (IDegLira) resulted in superior glycemic control, weight loss benefits, lower daily insulin dose, and a lower rate of hypoglycemia vs. insulin degludec (IDeg) in patients with T2D inadequately controlled with BI.
Comparison of insulin degludec / insulin aspart co-formulation therapy twice-daily with free combination of liraglutide plus IDeg	Yoshimasa Aso, Toshie Iijima, Teruo Jojima, Masaaki Sagara, Takuya Tomaru, Isao Usui	Treatment with once daily dual injection IDegLira and twice daily single injection of IDegAsp showed no significant difference in glycemic control. IDegLira resulted in a slightly larger reduction of A1c and superior weight loss.
Trends in U.S. insulin use for people with diabetes: 2009-2017	Magaly Perez-Nieves, Ludi Fan, Elizabeth Eby, Eric Meadows, Maureen Lage, Rattan Juneja	This Lilly-sponsored study examined trends in insulin usage for patients with T1D and T2D from 2009-2017 IBM MarketScan Commercial Claims and Encounters and Medicare Supplemental database data was used to assess the trends Use of short/rapid insulin and analogue insulins has increased, and premix insulin has decreased for T1D and T2D. Use of insulin pumps and disposable pens increased, while use of vials decreased for both cohorts More patients with T1D use continuous glucose monitoring, but use is increasing for both T1D and T2D patients (4x greater over the period)

Insulin Delivery Highlights

Control-IQ Pivotal Funding for Very Young Children (Ages 2-6) Secured, Recruitment to Begin “January, February” 2021; Six Months of Adult RWD (n=5,891) Show +2.2 Hours/Day in Time in Range, 95% Time in Closed Loop

EASD 2020 closed out this afternoon with an action-packed symposium from Tandem. The valuable symposium was hosted by Tandem’s Ms. Steph Habif with Drs. Paul Wadwa (Barbara Davis Center, Denver, CO) and Marc Breton (University of Virginia) as speakers. There was a lot to talk about, including the announcement of a Control-IQ pivotal in young children and a large set of real-world data from early Control-IQ adopters.

• Towards the end of his presentation, Dr. Breton shared that NIH funding for a Control-IQ pivotal trial in young children (ages 2-6) had just been secured last week. With funding in place, recruiting is set to begin in “January, February” of 2021; however, as Dr. Wadwa noted, because of the specialized population, recruitment is expected to take until the “end of 2021” across three clinical sites. The trial is hoping to recruit a total of ~102 participants, one-third between the ages of 2-4 and two-thirds between the ages of 4-6. The primary study will be run as an RCT (similar to the adult and older pediatric pivotals) for three months, followed by a three-month extension phase. Altogether, Dr. Wadwa shared that they were tentatively hoping to have results “sometime in 2022,” though he agreed with Dr. Breton that “we’re still excited just to hear about funding from the NIH right now.”

  • As a reminder, positive results from a small-scale study of Control-IQ (n=12) in this young pediatric population was presented at ADA 2020. Thought it went unreferenced during today’s symposium, presenters at ADA noted that some “modifications” were made by Tandem to run the study in patients that young (unclear if the modifications were made to the algorithm, hardware, or both). Tandem already has pediatric indication for Control-IQ in ages 6-13 (June 2020) following strong data from the iDCL5 pivotal trial and, of course, was initially cleared for ages 14+ in December 2019 following the landmark iDCL3 pivotal trial. Earlier this month, Medtronic received FDA approval for its MiniMed 770G, the first ever AID system available in the 2-6 years age group in the US.

• Dr. Breton also shared some very positive real-world data from ~6,000 Control-IQ users – representing “~10%” of the total Control-IQ user base. The anonymized, retrospective data came from Tandem’s t:connect database and included users 6+ years with at least six consecutive months of data. Average age of the sample was 42 years old and 89% of the users had type 1 diabetes, 4% had type 2 diabetes, with the remaining 7% presumably unknown or (possible but unlikely in our view) with another type of diabetes. The positive results mirror those from a smaller, ~2,000-3,000 person real-world data set presented by Tandem at ADA 2020 – this is great to see, and simply reinforces the great outcomes were the same in the “real world”. As of 2Q20, we estimated ~76,000 US t:slim X2 users on some sort of automated insulin delivery (Basal IQ or Control-IQ), with the vast majority on Control-IQ. The data from Dr. Breton is compared with results from the adult pivotal trial in the table below.

  • Time in Range increased by nearly 2.5 hours/day within the first two weeks of Control-IQ use and was sustained through at least six months. At baseline, mean Time in Range in the sample was 65%, quickly rising to 75% after just two weeks and remaining there through the 30-day, 90-day, and 180-day time points. It’s worth noting that 65% Time in Range is already a fairly high baseline, though this is not surprising given much of the sample is made up of early adopters of the Control-IQ technology. At the same time, we imagine that even the 75% could increase with coaching and input on hyperglycemia.

  • Mean glucose dropped by ~14 mg/dl after just two weeks of Control-IQ use and remained stable through at least six months. At baseline, mean glucose was 164 mg/dl, falling to 150 mg/dl after two weeks and remaining steady for the remainder of the time points. For context, 164 mg/dl corresponds to an estimated A1c of 7.3% (already quite low), while a mean glucose of 150 mg/dl corresponds to an estimated A1c of 6.9%. Regardless of baseline, real-world data from ~6,000 people achieving a mean estimated A1c of 6.9% for six months is unlike other experiences we’ve heard. It’s not so much that a 6.9% isn’t possible – it’s that from a population perspective, this many with a “standard deviation” this low is terrific.

  • Similar to the data set presented at ADA, real-world time in closed loop was an outstanding 95%. Overall, we admit that we continue to think this metric is a bit overblown in all closed loop trials – while it’s terrific that people wore the CGM and pump and stayed “tuned in,” that is certainly what what would hope – that’s why patients get the technology! The “time in zone” has nothing to do with the technology working, just on the fact that patients are “plugged in.” Still, this is an impressive result, driven by the ease of use of the t:slim X2 pump, Control-IQ system, and the Dexcom G6 CGM – whether or not they wear it is a vestige of the JDRF CGM trial back in the day, driven by Jeffrey Brewer and Aaron Kowalski (an assistant vice president back then!). As we noted previously, 95% time in closed loop is actually not even “similar” but a bit higher than the 92% time in closed loop seen in the pivotal trial, though this can likely be attributed to the difference in the characteristics of the groups.

	Pivotal		Real-world data
	SAP at Six Months (n=56)	Control-IQ at Six Months (n=112)	Baseline (n=5,891)	180 days of Control-IQ (n=5,891)
Time in Range	59%	71%	65%	75%
Time <70 mg/dl	1.9%	1.4%	1.9%	1.6%
Mean glucose	170 mg/dl	156 mg/dl	164 mg/dl	150 mg/dl
Time in closed loop	--	92%	--	95%

 

Insulet Omnipod 5 Hybrid Closed Loop Pivotal Trial Complete, Set to Read Out at ATTD 2020 in February; Reaffirmed Commitment to FreeStyle Libre 2 Integration

Insulet closed things out on EASD Day #2, with SVP and Medical Director Dr. Trang Ly hosting an hour-long symposium. The biggest news from the symposium came towards the end, with Dr. Ly sharing that the pivotal trial for Insulet’s Omnipod 5 hybrid closed loop system has now completed. Dr. Ly also maintained the 2Q20 plans to read out results from the pivotal trial at ATTD 2021 in February and launch the system in the US in the “first half of 2021” – it’s incredible that this timeline is holding, given COVID-19 uncertainty. The three-month trial initially began at the end of December 2019, temporarily paused in March due to a “software anomaly,” and resumed again in June. During today’s presentation, Dr. Ly mentioned that a “majority” of pivotal participants have already moved on to the extension phase of the trial. A few months ago at ADA, Stanford’s Dr. Bruce Buckingham shared results from the pre-pivotal trial (n=18 adults and 18 children), focusing in on strong system usability results. He also stressed high time in closed loop (97%) – we are not surprised by this. During the symposium, Dr. Ly made similar comments, noting that the Insulet design team spent a lot of time thinking about “alarms, usability, simplicity, and ease-of-use.”

• In the final slide of her presentation, Dr. Ly re-affirmed Insulet’s plans to integrate Omnipod 5 with Abbott’s FreeStyle Libre 2, sharing, “We haven’t announced any dates yet, but it’s happening.” The integration between Omnipod 5 (then known as Horizon) and FreeStyle Libre 2 was announced back in February, becoming the first AID system with finalized partnerships with two CGM manufacturers. The planned integration between FreeStyle Libre 2 and Omnipod 5 became unclear in June, when FreeStyle Libre 2 was cleared by the FDA as an iCGM, but came with a surprising “warning/limitation” against use with “automated insulin dosing (AID) systems, including closed loop and insulin suspend systems.” During our conversations with Abbott, the company seemed confident that the “warning/limitation” would eventually be removed and today’s re-commitment from the Insulet side seems to suggest both parties have confidence in ultimately getting an FDA-cleared Omnipod 5 + FreeStyle Libre 2 system to market. However, given the two companies are still in their very early stages of integration, Omnipod 5 will certainly come to market first with Dexcom G6 integration, the system set-up used in the pivotal trial. 

• As a reminder, Omnipod 5, will be Insulet’s first-generation hybrid closed loop system. The system will launch with Dexcom G6 integration and importantly, smartphone control. We’ve previously expected Insulet to initially launch smartphone control on Samsung Galaxy phones with support for other phones coming later – for those not using smartphone, Omnipod 5 can also be controlled by a dedicated device (similar to the Omnipod Dash PDM). Insulet’s talking points on Omnipod 5 have remained remarkably consistent for a long time, focusing on system usability and having multiple glucose set points. Dr. Ly offered some interesting perspective on the variable set points (targets can be set to 110, 120, 130, 140, or 150 mg/dl), noting that new to AID users could “ease” into the therapy by beginning with higher set points and working their way down.

• Dr. Ly also briefly commented on yesterday’s announcement for broader international Omnipod Dash rollout. The system is now available in eight countries: the US, UK, Netherlands, Italy, Sweden, Finland, Norway, and Denmark. In the “coming weeks,” the system will be rolled out to France, Belgium, Germany, Austria, Switzerland, and Israel.

Two Early Feasibility Studies (n=20 total) for Lilly AID Show System in Safe Under Multiple Challenges (Overbolus, Delayed Bolus, and Unannounced Exercise); Time in Range of 85% Outside of Challenges

In a poster, Lilly’ Ms. Amy Bartee presented findings from two small, early in-patient feasibility studies using Lilly AID in adults with well-controlled type 1 (n=10 in each study), which established the safety and efficacy of the system under several challenges. Following a 10-day lead-in period using blinded Dexcom G5 (wow this must’ve been an old study!), participants engaged in 2.5 days of system testing under close supervision; the studies included three real-world self-management challenges: (i) a 30% over-bolus (we thought this one was particularly brilliant and “real-world”; (ii) a delayed bolus given ~ one hour after the start of a meal; and (iii) unannounced 60 minutes of moderate exercise (this is very common and also a brilliant one to show). In the first study, participants (n=10, 20% male) averaged 52 years old, had long-standing, fairly well-managed diabetes (35 years diabetes duration and baseline A1c of 7.1%). In the second study, participants (n=10, 30% male) were younger (37 years) and also had long standing, well-controlled diabetes (26 years diabetes duration and an A1c of 7.3%). Based on identical baseline demographics, we believe that separate meal-challenge data from study 1 was presented at ADA 2020, though we admit this is partial speculation. As a reminder, the Lilly AID system uses Dexcom CGM, a DEKA-developed patch-style pump, and an MPC algorithm from Montreal’s ClassAP group. While Lilly has never shared specific timing for development of its AID system, at the unveiling of the idea in November 2017, launch timing of “two to three years” was given; the same timing was given to former associates at Close Concerns and diaTribe, Brian Levine and Jeemin Kwon, when they visited Lilly in Boston in mid-2018. As such, broadly, we interpret their remarks to convey that Connected Care products will launch in stages over 2019-2021; at this point, we’re assuming this timeline has been pushed back though that is speculation.

	Time in Range during Study Periods
Glucose Ranges (mg/dl)	Outside of challenges	Nocturnal (12am – 6am)	Postprandial	30% Over-bolus	Delayed Bolus	Exercise
<70	3%	0.8%	5%	3%	4%	1.3%
70-180	85%	97.2%	75%	65%	57%	89.6%
>180	12%	2%	21%	32%	37%	9.1%
>250	2%	0%	4%	5%	10%	0%

 

• Overall, Time in Range outside of the challenges was a very impressive 85%. This was surprising particularly given that the sensor used was the G5. The system also easily achieved the consensus targets for CGM metrics (<4% time <70 mg/dl) and hyperglycemia (<25% time >180 mg/dl), spending just 3% time <70 mg/dl, 12% time >180 mg/dl, and 2% time above 250 mg/dl. With the Lilly AID, overnight Time in Range was an astounding 97%; 0.8% of overnight time was spent in hypoglycemia and 12% of time was spent above range (we assume this was over 180 mg/dl – we are not sure how much over 250 mg/dl it was). Looking at mealtimes, the system performed well after regular meals, achieving 75% postprandial Time in Range. These are very positive results, even with the obvious disclaimers that the participants were a very well-managed group to begin with and the study took place under quite supervised settings. The baseline CGM metrics, such as Time in Range, were also not provided, making it even more difficult to draw strong conclusions from the data.

• No adverse events occurred during the short study, suggesting the system was safe, even with the over-bolus, delayed bolus, and exercise challenges. Of course, the study has limited generalizability: (i) it was conducted in an in-patient setting; (ii) only the study staff interacted with the device; (iii) the investigators could mitigate hypo- and hyperglycemia for severity; and (iv) the small cohort (n=20) had strong glycemic control (7.2% baseline A1c). Per Dr. Bartee, plans for an out-patient study are “ongoing” – we assume they have not been solidified yet though we await more news. There is no doubt that an out-patient study would provide more clarity into the real-world efficacy and safety of the Lilly AID system and we look forward to seeing this as well as finding out more about the pump and specifically how much hassle it has.

• In the over-bolus challenge, participants with a pre-meal glucose above 120 mg/dl delivered a 30% over-bolus 15 minutes before a meal. Time in Range in the post-prandial four hours remained impressive, at 65%; more importantly, the over-bolus did not significantly increase time in hypoglycemia – time <70 mg/dl was 3%. Presumably this was because a portion or all of the insulin shut off.

• In the one-hour delayed bolus challenge, participants were asked to deliver their meal bolus one hour after finishing their meal (breakfast in one study, dinner in the other). Even with the delayed bolus, Time in Range during the four hours post-meal remained fairly high at 57% - unsurprisingly, time in hyperglycemia was a bit high at 37%.

• Finally, the Lilly AID system performed well in the unannounced exercise challenge. Participants were asked to take part in 60 minutes of moderate exercise on a bike or treadmill (before breakfast in one study and after dinner in the other). Time in Range during exercise was very high at 90% and encouragingly, very little time was spent in hypoglycemia (1.3%). This was good to see although this was not too much of an exercise challenge – more problems with exercise come up with extended, unplanned exercise.

The Free-Life Kid AP Study Shows 24/7 Control-IQ Use Increases Time in Range by 3.5 Hours/Day for Children with Type 1 (Ages 6-12); Nighttime-Only Use Increases Time in Range by 2.3 Hours/Day

Montpellier University Hospital’s Dr. Eric Renard read out 18-week (~4-month) results of the Free-Life Kid AP study, which assessed the efficacy and safety of 24/7 versus dinner/night-only use of Control-IQ in kids (ages 6-12) with type 1. Participants (n=120, 8.6 years, 40% female) were trained and initiated on t:slim X2, Dexcom G6, and Control-IQ and then randomized to either: (i) 24/7 use of the Control-IQ algorithm (n=60), or (ii) use of Control-IQ only during dinner and nighttime, with the system manually shut off during the day by parents (n=60). All participants had had type 1 for longer than a year (mean 5.2 years), had used a pump for longer than six months (~4.6 years on average), and had A1c’s <10% (baseline: 7.7%). For both groups, all glycemic metrics were significantly better at 18 weeks relative to baseline. Those using Control-IQ 24/7 saw a significant 15% percentage-point increase in Time in Range (+3.5 hours/day) to 68% at week 18. Those using Control-IQ only during dinner/nighttime saw a smaller 2.3 hours/day improvement in Time in Range, reaching 65% at week 18. As we saw in the pivotal studies of Control-IQ in both adults and pediatrics, the most significant Time in Range improvements with the AID algorithm come at night; however, it’s worth noting that Control-IQ use during the day incurred an additional 1.2 hours/day benefit in Time in Range relative to the night-only group (p=0.001). As expected, there were no significant between-group differences in nighttime glycemic improvements, including Time in Range, mean glucose, time below 70 mg/dl, time below 54 mg/dl, and time above range. Notably, for both the 24/7 and nighttime-only groups, all baseline Time in Range and A1c ranges saw improvements in Time in Range at 18 weeks. Likewise, for both groups, time in closed loop was sustained over the 18-week period: the 24/7 Control-IQ group averaged 94% time in closed loop, and the nighttime-only group achieved an average 50.9% time in closed loop (i.e., they did not use it during the day). There were no cases of DKA, severe hypoglycemia, hospitalizations, suggesting that Control-IQ was safe for both nighttime-only and 24/7 use.

• As a reminder, the study was announced at ATTD 2018, and we heard 12-week interim results (n=30) from Dr. Renard at ADA 2019. The researchers intend to extend the study through day 252 (~8 months); however, after 18 weeks, all participants switched to 24/7 Control-IQ use, so it will no longer compare 24/7 and nighttime-only use of Control-IQ. We look forward to seeing the long-term results, which hopefully will echo the positive US pivotal results read out at ATTD 2020 which drove Tandem’s pediatric indication for 7-13-year-olds announced in June.

More on MiniMed 780G Pivotal Trial: Shorter Active Insulin Times Achieve Better Time in Range Regardless of Setpoint; With 100 mg/dl Setpoint, 75% of Participants Achieve TIR Target vs. 45% at Baseline

Dr. Anders Carlson (International Diabetes Center) highlighted new data on the impact of active insulin time and basal setpoint on glycemic outcomes from the MiniMed 780G pivotal trial, which was first read out at ADA 2020. Specifically, Dr. Carlson showed that with both glucose setpoints (100 mg/dl and 120 mg/dl), a shorter active insulin time incrementally increased Time in Range. Perhaps more notably, with the 100 mg/dl setpoint, a shorter active insulin time also incrementally decreased time below range. With the higher setpoint (120 mg/dl), a shorter active insulin time actually slightly increased time below range. These outcomes seem to go against intuition that more “aggressive” settings might increase time in hypoglycemia and are worth exploring. Dr. Carlson also showed that a lower setpoint (100 mg/dl) saw improvements in percentage of participants achieving consensus goals for Time in Range (>70% time 70–180 mg/dl) and time below range (<4% time <70 mg/dl). Specifically, participants achieving Time in Range >70% increased from 45% at baseline to 69% overall (both 100 mg/dl and 120 mg/dl setpoints) to 75% with only the 100 mg/dl setpoint. Likewise, participants achieving the target for time below range increased from 70% at baseline to 80% with the 100 mg/dl basal setpoint. All of this suggests that together the more aggressive basal setpoint and the shorter active insulin times result in better glycemic control, an exciting update from what we learned in June. The data also further highlights the fact that the MiniMed 780G system is primarily designed to be used with a set point of 100 mg/dl, though some people will have their reasons for going with the 120 mg/dl setpoint. At ADA 2020, we also saw impressive overall results, showing a +1.4 hour/day gain in Time in Range, a 0.5% A1c improvement on MiniMed 780G relative to baseline (SAP or MiniMed 670G), and 95% time in closed loop. As a reminder, the US pivotal was a ~3.5-month, single-arm (n=157, ages 14-75) study that compared Medtronic’s “Advanced Hybrid Closed Loop” (aka MiniMed 780G system) with sensor-augmented pump therapy and MiniMed 670G.

Two Posters on MiniMed 670G: Nine-Month, Sustained Positive Outcomes for Youth Transitioning Straight from MDI; >30% Rise in Percentage of Type 1s Achieving A1c and Time in Range Targets After Six Months

Tuesday afternoon’s poster sessions included two posters with promising findings on the sustained glycemic improvements associated with MiniMed 670G use. While there is much excitement around MiniMed 780G and its potential to further improve glycemic control (see the Medtronic symposium on 780G from Day #1), it is still great to see the improvements which patients can achieve with MiniMed 670G, which had a quarter of a million users as of May. In particular, these two studies are especially promising in that they suggest that MiniMed 670G can help people with diabetes achieve sustainable glycemic improvements and are effective tools for improving glycemic outcomes in children and adolescents with diabetes.

• Following up six-month data presented at ATTD 2020, Dr. Goran Petrovski (Sidra Medicine) shared nine-month data on youth (n=30, ages 7-18) transitioning directly to MiniMed 670G from MDI after a 10-day initiation. Through month six: participants’ mean A1c fell from 8.2% to 6.9% and their mean Time in Range increased from 46% to 73% (+6.5 hours/day!). The data from months seven to nine show that participants sustained these positive results, achieving an A1c of 6.8% and a Time in Range of 74%. Similarly, the reductions in time above and below range that were achieved after three months were also sustained after nine months. Also impressive, time in auto mode was 87% in month one, which was sustained through month nine when time in auto mode was 89%. The data from Dr. Petrovski is not only very important because it represents a traditionally poorly-controlled group (adolescents), but also because it demonstrates the ability to “accelerate” patients straight from MDI to MiniMed 670G.

• Dr. Pilar Beato-Vibora (University Hospital of Badajoz) presented six-month clinical outcomes for children (n=22) and adults (n=36) with type 1 started on MiniMed 670G after being on predictive low glucose suspend (60%), CSII + BGM (19%), MDI + BGM (12%), or MDI + CGM (9%). Overall, participants’ mean A1c was reduced by 0.4% at three months (baseline 7.4%), a drop that was sustained at the six-month endpoint and quite significant from a low baseline A1c. Mean Time in Range increased significantly from 63% to 73% after three months on MiniMed 670G (+2.4 hours/day), remaining even at 72% at six months. Likewise, time above 180 mg/dl and time above 250 mg/dl improved significantly, dropping from 35% to 26% (-2.2 hours/day) and from 9% to 5% respectively at month six. The cohort experienced zero hospitalizations due to DKA or hypoglycemia during the trial. Particularly noteworthy, after six months, the percentage of participants achieving target A1c of <7% rose 30% to 61%, and percentage of participants achieving the target Time in Range of >70% rose a whopping 36% to 57%. This improvement was already seen after three months (53% meet A1c target and 60% meet Time in Range target) and was sustained through six months. Participants spent 86% and 88% of time in auto mode at three and six months respectively, and experienced 0.6 exits per day.

18-Month Real-World Study of MiniMed 670G with Intensive Start-Up and Follow-Up Protocol Shows Consistently High Time in Auto Mode (>80%) and Sustained Glycemic Improvements in Young Adults with Type 1

During a session on updates in AID, Dr. Dick Mul (Diabeter, Rotterdam, Netherlands) shared findings from a real-world study which introduced young adults with type 1 (n=104, mean age 21) to Medtronic’s MiniMed 670G system with a comprehensive and intensive start-up and follow-up protocol. The study was particularly focused on Time in Auto Mode, as some studies have seen stark drop-offs in long-term time in Auto Mode, for example, dropping from 74% one week after starting on MiniMed 670G to 35% after a year on the system (Lal et al. 2019). It’s worth noting that other data sets have shown less drastic results on time in Auto Mode; for example, real-world data from Barbara Davis Center found time in Auto Mode at ~77% after 3 months and 73% at 6 months. Excitingly, the Diabeter study showed that time in Auto Mode remained stable over the 18-month study period, averaging over 80%. Likewise, the distribution of participants’ time in Auto Mode stayed consistently high-leaning throughout the trial, and by month 15, fewer than 25% of participants were not achieving >70% time in Auto Mode. Time in Range also increased and remained stable, and time below range and time above range decreased and remained stable throughout the trial. Dr. Mul attributes this success to their comprehensive education and extended support and follow-up program, which encouraged patients to maintain or move toward a high time in Auto Mode value. Of course, the sample size was small for this study, especially by month 18 (n=10 time in Auto Mode data points). However, the results are promising, especially for an age group that has been shown to have such difficulty with glycemic control.

Dana Lewis and Dr. Henk-Jan Aanstoot on Open Source or Commercial Closed Loop: “The Question Is Not Open Source or Commercial, But Rather: What Can We Learn From Open Source to Improve What’s Commercially Available”

In a debate on open source versus commercial closed loop systems, Ms. Dana Lewis (OpenAPS) and Dr. Henk-Jan Aanstoot (Diabeter) redirected the question, instead emphasizing that open source and commercial systems are not mutually exclusive and can be used to meet different patient needs. Ms. Lewis, one of the founders of the DIY diabetes movement and creator of OpenAPS praised the ability of open source closed loop systems to provide users flexibility and personalization in their insulin delivery due to open source code and algorithms that can be adapted to meet their personal diabetes management needs. Specifically, Ms. Lewis highlighted open source algorithms, like the one she uses (OpenAPS), that no longer require bolus or meal announcements from users, allowing her to spend less time managing her diabetes and more time living her life. Open source algorithms sometimes also offer features not yet available on commercial system, such as “auto-sensitivity assessments” that review data from the last 8-24 hours to determine if the user’s insulin sensitivity has changed and adjust insulin delivery accordingly. Two other features of open source algorithms that Ms. Lewis and other open source users have found to be especially useful are temporary targets and net insulin on board (net IOB), which calculates insulin on board from both basal and bolus doses (most commercial insulin pumps make insulin on board calculations from bolus doses only). Though many DIY closed loop users love the open source systems, Ms. Lewis recognizes that it is not for everyone presenting a “one-nine-ninety” understanding of people using either open source or commercial closed loop technology. In this “one-nine-ninety” model, 1% of users are involved in developing and improving the technology for all users (both in open source and commercial systems), 9% of users utilize open source technology and repurpose it to meet their needs, and 90% of users choose to optimize ease of use and access utilizing commercial closed loop systems. In this understanding, both open source and commercial closed loop systems have a place in diabetes care. As Ms. Lewis expressed, “the question is not open source or commercial, but rather: what can we learn from open source to improve what’s commercially available …” – presumably the reverse is true as well. With over 29 million estimated patient-hours of DIY closed loop experience among users, there is substantial data to draw from to continue to improve both open source and commercial closed loop technology.

• Building on Ms. Lewis’s presentation, Dr. Aanstoot recognized the place of DIY closed loop technology in diabetes care, but argued commercial closed-loop systems offer the most value to the most people. This is not something we felt Ms. Lewis would necessarily argue with, though the commercial systems may not have reached the place they are without knowledge from DIY. Dr. Aanstoot discussed the scalability of commercial closed loop systems, noting they are much more likely to achieve widespread penetration than open source systems. Additionally, Dr. Aanstoot raised questions about the ethical and legal ramifications of open source systems where it can be more challenging to know who is liable and responsible for the system’s functioning, e.g., customer support. Use of open source systems is also complicated by the reality that in most countries, providers are unable to recommend open source closed loop technology because not officially approved. With these limitations in place, Dr. Aanstoot highlighted the incredible value patients have seen on commercial closed loop systems and urged all diabetes stakeholders to keep working to make these systems available to more people. Though Dr. Aanstoot agreed with Ms. Lewis that DIY and commercial closed loop systems are not mutually exclusive, he said “my major concern is how I can add value to many more people, which, in my country and healthcare system will primarily be by new commercial systems.”

• Ms. Lewis highlighted data from a number of retrospective and observational studies on the efficacy of open source systems. Since Ms. Lewis and Scott Leibrand created OpenAPS in 2015 some critics have been concerned there is not enough peer-reviewed evidence in support of the systems’ efficacy. To counter this claim Ms. Lewis cited studies indicating sustained blood glucose improvements and gains in Time in Range while using the OpenAPS system. Additionally, Ms. Lewis cited a study presented at EASD 2019 that demonstrated 80% Time in Range and only 1.7% time <70mg/dl. Additionally, Ms. Lewis drew viewers’ attention to the ongoing CREATE RCT comparing the OpenAPS algorithm + Dexcom G6 + DANA-i pump with sensor-augmented pump therapy.

• During Q&A, Ms. Lewis and Dr. Aanstoot answered a question about whether open source or commercial AID algorithms are more advanced. Both Dr. Aanstoot and Ms. Lewis, who acknowledged her bias toward open source algorithms, believe commercially available algorithms lag behind what is available from the DIY community. Anecdotally, Ms. Lewis cited her ability to forgo meal announcements with her open source AID algorithm and still achieve well over 70% Time in Range. She also asserted that most users of open source systems achieve 80%-90% Time in Range whereas commercial users are closer to 70% Time in Range – we believe most of this data is anecdotal as there is not as yet source of “Time in Range” results. That said, Ms. Lewis also commented that it is hard to compare across commercial and open source systems because there is significant variability in how users choose to interact with their closed loop systems that also impacts their time in range and health outcomes.

Small Study (n=36) on 670G Users Shows “Low” Carb Intake Days Have 2% Higher Time in Range (+32 minutes/day) Compared to “Normal” Days, 7% Higher Time in Range (+1.7 hours/day) Compared to “High” Carb Days

In a compelling study from the University of Bern, Dr. Vera Lehmann presented data from MiniMed 670G users, finding an association between lower carb intake and higher Time in Range. The small (n=36), single-center study looked at retrospective data from MiniMed 670G pump readouts, including CGM and carbohydrate intake,  and medical and diabetes history from EHR. Normalizing daily carbohydrate intake by patient, days were divided into “low” (≤80% of mean individual daily carb intake), “medium” (81%-120% of mean daily intake), and “high” (≥120%) carb days. The participants skewed young (mean age 37 years), male (72%), and were well-controlled at baseline (mean A1c 7.1%). Average carb intake isn’t known.

• Time in Range on “low” carb intake days was 7% higher (+1.7 hours/day), compared to “high” carb intake days (p<0.001). The difference between “low” and “medium” carb intake days was smaller, with a 2% difference favoring “low” carb (+32 minutes/day). As would be expected, all of the improvement for “low” carb came from a reduction in time in hyperglycemia, while time below 70 mg/dl remained statistically identical across all days. Overall, a 10% reduction in daily carb intake was associated with a 1.1% increase in Time in Range (+16 minutes/day) and a 1.2% reduction in time >180 mg/dl. Notably, the percent time spent in Auto Mode was significantly higher on “low” carb intake days – on “low” days, 82% of participants spent >90% of time in Auto Mode, compared to just 73% of participants on “high” days.

Insulin Delivery Posters

Title	Authors	Details + Takeaways
Patient-reported outcomes for 2,335 adults with type 2 diabetes using the Omnipod Insulin Management System show glycemic improvement over the first 90 days of use	Lauren Huyett, Anders Carlson, Albert Chang, Jay Jantz, Trang Ly, Todd Vienneau	Retrospective study characterizes patient-reported outcomes for large cohort of adult type 2’s (n=3,952, age 57, 14 years of diabetes, 74% previously on MDI) before and 90 days after initiating use of Omnipod or Omnipod Dash Average A1c fell significantly by 1.3% to 7.9% 90 days after pump initiation; prior MDI users saw a greater change than did prior CSII users (-1.4% vs. 0.9%); higher baseline A1c was correlated with higher drop in A1c Percentage of users with A1c >9% dropped significantly by 30% to 19% after 90 days on an Insulet pump 22% achieved ADA target of <7% at follow-up (+12% from baseline); 58% met HEDIS criteria for adequate control (<8%) at follow-up (+28% from baseline) Also saw significant -0.55 event/week reduction in hypoglycemic events and -33 U/day reduction in insulin use; greater reduction in daily dose for MDI (-35 U/day vs. -20 U/day)
Glycemic control improves over 4-month use of closed loop insulin delivery in school-age children with type 1 diabetes	Marc Breton, Craig Kollman, Eda Cengiz, Roy Beck, Daniel Cherñavvsky, Gregory Forlenza, Bruce Buckingham, Mark DeBoer, Laya Ekhlaspour, Lauren Kanapka, R Wadwa, Melissa Schoelwer, Stuart Weinzimer, Katrina Ruedy	4-month study compares Control-IQ (+G6, +t:slim X2) to sensor-augmented pump in children (ages 6-13) with type 1 (n=100, 40% A1c >8%; 20% pump-naïve, 8% CGM-naive) Control-IQ group saw a ~15% increase in Time in Range to high 60s/low-70s within 1-4 weeks whereas no change was seen in the SAP group; a difference maintained throughout the 4-month period Time in Range improvement primarily achieved during night Control-IQ group saw 0.6% reduction in A1c to 7% while SAP group saw 0.3% reduction to 7.6% The majority of participants on Control-IQ spent 85%-100% of Time in Auto Mode, which was consistent across study period
Clinical outcomes after 6 months of use of hybrid closed loop system in children and adults	Pilar Beato Vibora, Francisco Javier Arroyo-Díez, Fabiola Gallego-Gamero, Lucía Lázaro-Martín, María del Mar Romero-Pérez	Evaluation of 6-month clinical outcomes for children and adults with type 1 (n=58, age 28, 38% <18yo) who start Medtronic 670G A1c significantly dropped by 0.4% to 7.0% at 3 months, which was sustained at 6 months 10% increase in Time in Range from 63% at baseline to 73% at 3 months and 72% at 6 months (p<0.05) Time above range + time >250 mg/dL significantly decreased (-10% and -3.7% respectively);  time below range did not significantly change Patients achieving A1c <7% increased to 53% at 3 months (+22%) and to 61% at 6 months (+30%); at 3 months and at 6 months, >57% of participants achieved >70% Time in Range compared to 21% at baseline Time in Auto Mode was 85% and 88% at 3 months and 6 months respectively
Six-months at-home hybrid closed-loop vs manual insulin delivery with finger-stick blood glucose monitoring in adults with type 1 diabetes: a randomized controlled trial	David O’Neal, Steven Trawley, Leon Bach, Peter Colman, Neale Cohen, Christel Hendrieckx, Jane Holmes-Walker, Sybil McAuley, Timothy Jones, Joey Kaye, Alicia Jenkins, Roland McCallum, Sara Vogrin, Steven Stranks	6-month RCT (n=120, age 43, 7.4% A1c) compares at-home MiniMed 670G (n=60) and manual insulin delivery (MDI or pump) with BGM (n=60); used blinded CGM to measure glycemic outcomes MiniMed 670G group’s Time in Range increased by 15 percentage-points relative to the control group (p<0.001) to 70% Time in Range All other secondary CGM-based glycemic outcomes significantly favored the MiniMed 670G group at 6 months (p<0.001); as did A1c (-0.4% to 7.0% in the MiniMed 670G system) Participants using MiniMed670G were more satisfied with their treatment and felt more positive about their diabetes
Evaluating the long-term cost-effectiveness of introducing a smart insulin pen in standard-of-care treatment of type 1 diabetes in Sweden	Barnaby Hunt, Åsa Ericsson, Jens Gundgaard, Johan Jendle, Jonas Møller, William Valentine	Study assesses the long-term cost-effectiveness of a smart insulin pen (NovoPen6) vs. standard care for simulated cohort of type 1s in Sweden (n=_) using IQVIA CORE Diabetes Model to project clinical outcomes and healthcare costs over lifetime Model used data from observational study with 94 adults with type 1 for whom smart pen use increased Time in Range by 1.89 hours/day, translated to a 0.62% A1c reduction, and resulted in 33 fewer non-severe hypoglycemia events/patient/year Smart insulin pen usage was associated with a €10,815 reduction in mean discounted direct healthcare costs/patient and a +1.13 increase in QALY relative to standard care Higher treatment cost of smart pen use was offset by lower costs and frequency of complications
First home evaluation of the Omnipod Horizon Automated Glucose Control System in adults with type 1 diabetes	Sue Brown , Gregory Forlenza, Amy Criego, Bruce Bode, Carol Levy, Bruce Buckingham, Trang Ly	Multicenter, single-arm study  assesses the safety and efficacy of Omnipod Horizon in outpatient setting over 14 days of use in adults (n=18, ages 14-70, average age 35, 7.1% A1c) During 14 days, had 3 days w/130 mg/dL target, 140 mg/dL target, and w/150 mg/dL, followed by 5 day with free choice (110-150) and 3 months of the pivotal study During free choice period, participants spent 1.7 hour/day more Time in Range compared to standard care baseline Most participants did not experience hypoglycemia during Omnipod 14-day period while most did with standard care There were no serious adverse events, DKA, or severe hypoglycemia during the 14-day Omnipod period Participants spent 97% Time in Auto Mode
Predictive low glucose suspend (PLGS) necessitates less carbohydrate supplementation to rescue hypoglycemia:  need to revisit current hypoglycemia treatment guidelines	Richard Jones, Amy LaLonde, Amy Bartee, Eyal Dassau, Howard Wolpert, Michelle Katz, Jordan Pinsker	Early in-patient feasibility study (n=10, age 39, 100% white, 7.2% A1c) with Lilly AID system (+Dexcom G6) investigates when need for rescue carbs can and cannot be avoided with predictive low glucose suspend (PLGS) System’s PLGS responded appropriately (suspend when glucose falls, resume when glucose rises) With the PLGS system, participants experienced postprandial hypo but not overnight hypoglycemia even when triggered PLGS by gradually increasing basal rate overnight <15 g of carbs may be needed to treat hypoglycemia in adults using Lilly’s PLGS system, which should be reflected in treatment guidelines to minimize rebound hyperglycemia
Marked improvements in HbA1c levels following insulin pump therapy initiation in people with type 1 diabetes: a nationwide observational study in Scotland	Anita Jeyam, Helen Colhoun, John McKnight, Luke Blackbourn, Fraser Gibb, Brian Kennon, Stuart McGurnaghan, Joseph O'Reilly, Paul McKeigue	Used data from Scottish Diabetes Collaboration database (n=4,684, 59% female) to compare A1c 2 years prior to pump initiation (baseline, prior to structured education prep) with A1c following pump initiation Median A1c reduction of 0.5% during first year after pump initiation; reduction was sustained for 5 years other than in teenagers and somewhat in children (ages <13)  Greater reductions in A1c seen for those with higher baseline A1c levels; ex. >9.8% A1c at baseline had 1.9% A1c reduction in first year of usage, the highest reduction seen No variation in A1c reduction by socioeconomic “deprivation” Compared to non-pump users, pump users saw reductions in A1c
Patient-reported outcomes reveal the potential positive impact of hybrid closed-loop systems on users‘ emotional well-being	Sara Suhl, Richard Wood, Jonathan Rost	Survey-based study (n=927) assesses associations between use of AID systems and emotional well-being as measured by WHO-5 with adult type 1s (n=878) and type 2s (n=49) Control-IQ and DIY system users were significantly less likely to report poor well-being than were non-AID users (23% vs. 18% vs. 38% respectively) MiniMed 670G users and Basal-IQ PLGS system users had WHO-5 scores that were statistically similar to non-AID users
Postprandial glucose control using the Medtronic Advanced Hybrid Closed Loop System: faster-acting insulin aspart vs insulin aspart	Melissa Lee, Barbora Paldus, Benyamin Grosman, Richard MacIsaac, Alicia Jenkins, Jean Lu, Hannah Jones, Natalie Kurtz, Dale Morrison, Anirban Roy, Emma Netzer, Catriona Sims, David O'Neal, Sara Vogrin, Dessi Zaharieva	16-week randomized crossover study (n=12) compares postprandial glucose and overall glucose control using FiAsp vs. Aspart using MiniMed 780G in free-living conditions in pump-experienced adults (age 49, 7.1% A1c, 7 years of pump) Each stage included 2 weeks in open loop followed by 6 weeks in closed loop using MiniMed 780G algorithm, active insulin time of 4 hours, and 100 mg/dL basal target FiAsp was safe and effective when delivered via MiniMed 780G; may confer advantage over Aspart for postprandial glucose control after breakfast Trends in favor of FiAsp for all glycemic metrics but not significant
Ultra-Rapid Lispro (URLi) demonstrates similar time in target range to Lispro with the Medtronic MiniMed 670G hybrid closed-loop system	Bruce Bode, Scott Morrett, Richard Bergenstal, Anders Carlson, Jennifer Boyd, Thomas Hardy, Debra Ignaut, Rong Liu	4-week cross-over RCT (n=42, age 41, baseline 7% A1c) compares Time in Range for MiniMed 670G with ultra-rapid lispro (URLi) vs. Lispro When used with MiniMed 670G, URLi and Lispro achieved comparable (i) 24-hour glucose control; (ii) 24-Hour, daytime, nighttime, postprandial, and pre-prandial Time in Range values; (iii) basal, bolus, or total insulin dose; (iv) safety profiles Might need an AID system with automated meal and correction boluses to see benefits of URLi
First results from PRO Solo: patient reported outcomes from a clinical trial comparing a new patch pump with MDI and an established patch pump	Julia Mader, Katharine Barnard-Kelly, Thomas Künsting, Nick Oliver, Iris Vesper	Roche-funded 3-arm RCT compares patient-reported outcomes for Accu-Chek Solo micropump system, Insulet Omnipod, and MDI with pump-naïve adult type 1s (n=181, age 29, 8% A1c) 26 weeks on three arms, then all switched to Accu-Check Solo micropump system for 3 weeks Patients switching from MDI to Accu-Chek Solo saw sustained improvements in all psychosocial outcomes (per PAID-5 scores), but the same was not true for those switching to Omnipod Accu-Chek Solo pump has slightly lower ratings for device acceptance and satisfaction than Omnipod
Usage and handling of insulin pump features by individuals with type 1 diabetes	Selina Dürrbeck, Nadine Kuniss, Christof Kloos, Nicolle Müller, Guido Kramer, Gunter Wolf	Structured interviews on use, ability to use, and frequency of use of advanced pump functions with type 1 adults (n=72, age 52, 8.0% A1c) All participants knew ≥4 of 7 functions, 1/3 correctly specified all functions of the pump, 42% could explain dual wave/combo bolus function correctly Bolus calculator = most commonly used feature; only 29% of those with low-glucose suspend feature used it 1/3 were able to properly handle all surveyed features without help Bolus calculator + e-logbook (>5x/week) users were more satisfied with their treatment
Efficacy and safety comparison between U-100 human regular insulin and rapid acting insulin when delivered by V-Go insulin delivery device in an older type 2 diabetes	David Sutton Jr. , John Sink II, Beverley Adams-Huet, Rebecca Goldfaden, Bantwal Baliga, Ashwini Gore, Carla Nikkel, Pablo Mora	14-week RCT evaluates safety and efficacy of human regular insulin (HRI, n=28) compared to rapid acting insulin (RAI, n=25) when administered through V-Go wearable insulin delivery device in older adult type 2s (n=, 72 years, 7.9% A1c) Overall found that for an older population, HRI delivered by V-Go was as safe and effective as RAI Compared to baseline, the HRI group saw a statistically significant 0.35% reduction in A1c while the RAI group’s A1c did not significantly change Total daily dose of insulin increased by 0.78 U/day for HRI and by 1.49 U/day for RAI, resulting in a nonsignificant -0.71 U/day reduction for HRI HRI cohort reduced their weight by 2.34 kg compared to the RAI cohort (p=0.0028)
The ROTO Track® device improves rotation of insulin injections in type 1 diabetes: a proof-of-concept study	Rasmus Tjalk-Bøggild, Yasmin Hamid, Peter Kristensen, Carina Klarskov, Lise Tarnow	Partially blinded, 12-week cross-over study investigates ability of ROTO Track to reduce # of insulin injections at the same site compared to non-aided technique in adult type 1s (n=35, age 55, 8.9% A1c) ROTO Track is a new medical device that automatically guides PWD to rotate abdominal insulin injections, which is associated with lower incidence of lipohypertrophy and reduced A1c Proof-of-concept study showed that ROTO Track device is safe, improves rotation, reduces lipohypertrophy, and reduces glycemic variability
The burden of mealtime insulin dosing in adults and children with type 1 diabetes	Wendy Lane, Emma Lambert, Jesso George, Naveen Rathor	Online multinational survey (n=2,711) assesses attitudes, behaviors, and overall impact of mealtime insulin dosing (specifically 15-20-minute injection-t0-meal interval) for adult type 1s (n=1401), parents of children with type 1 (n=350), and HCPs (n=960) A majority of both adult patients (96%) and parents (94%) understood the importance of accurately administering mealtime insulin, but only 35% of adults and 47% of parents felt very confident in estimating the amount of insulin required; only 16% of HCPs believed that their patients felt very confident 82% of adult type 1s and 93% of parents felt that the 15-20-minute interval had a negative impact on their/their child’s daily life; 91% of HCPs agreed that this is a burden for their patients 73% of adults and 67% of parents would prefer to administer insulin immediately before or immediately after a meal
Do “looper” have better glycemic control?	Timm Roos, Dominic Ehrmann, Bernhard Kulzer, Lutz Heinemann, Norbert Hermanns	Survey-based study (n=2,249) assesses the self-reported A1c of people using closed-loop system (6%) vs. CSII (43%) vs. MDI (51%) Those using closed loop had lower A1c (6.3%) than did MDI (median 7.0%) and CSII (median 7.1%), which was true for both adults and children (all same averages except pediatric CSII A1c average of 7.2%) The study is limited by the fact that it is based on self-reported data

Glucose Monitoring Highlights

Medtronic: Guardian Sensor 4 Will Have Zero Cal, Non-Adjunctive, and Improved Hypo Accuracy; Small Study (n=14) Shows Skipping One in Ten Meal Boluses Does Not Reduce Time in Range or Raise A1c With 780G

As expected, Medtronic’s symposium focused on MiniMed 780G, which received its CE-Mark in June. The biggest news of the session was a quick slide on Guardian Sensor 4 within Medtronic Diabetes VP of R&D Ali Dianaty’s presentation on the features of the MiniMed 780G system. On his presentation’s final slide, Dr. Dianaty announced that “a new sensor algorithm will become available very, very shortly” (Guardian Sensor 4) and that the design goals for Guardian Sensor 4 include: (i) “zero calibrations,” (ii) non-adjunctive indication, and (iii) improved accuracy in the hypoglycemia range. The hardware on the slide appears to be the same disposable clamshell form factor used in the current Guardian Sensor 3 and the planned “Zeus” iCGM. As a reminder, the publicly disclosed goals for “Zeus” were to keep the same form factor, reusable transmitter, and 7-day wear as Guardian Sensor 3, but will be day 1 calibration only (vs. 2 fingersticks/day) and meet iCGM criteria. Based on this, we’d guess that Guardian Sensor 4 is actually some variant (1 cal/day vs. zero cal) of “Zeus.” As of 1Q20, “Zeus” was expected to be submitted to the FDA by the “end of summer” after completing the pivotal trial in May. The non-adjunctive indication is particularly important for Medtronic, as it will enable Medicare reimbursement for the Guardian Connect and MiniMed 670G and 770G systems (770G approved in September) and enable the automatic correction boluses that are part of the 780G system. For the past ~year, we’ve heard Medtronic earnestly and openly acknowledge the CGM as the “weak link” in its offerings and a focus on accelerating the CGM pipeline. Certainly, improving from two calibrations/day to no-calibrations and getting non-adjunctive labeling will be a big step forward from Guardian Sensor 4 – while it’s easy to say the feature set still lags behind Dexcom’s and Abbott’s sensors, some would say considerably, it’s one foot in front of the other and we look forward to watching how else they can move faster, with the warchest.

Other this Guardian Sensor 4 update and a small (n=14) study on meal announcements presented by Dr. Amir Tirosh (Sheba Medical Center, Israel) to close out the session (see more below on this), little of the presentation was new. Dr. Ohad Cohen (Director of Medical Affairs at Medtronic Diabetes EMEA) reviewed real-time data on MiniMed 670G in EMEA which was presented at ATTD 2020, and Dr. Tadej Battelino (University Children’s Hospital, Slovenia) reviewed data on the MiniMed 780G system presented at ADA 2020.

• Dr. Amir Tirosh (Sheba Medical Center) shared preliminary data from a small (n=14) trial that evaluated the performance of MiniMed 780G when meals are announced or unannounced, showing that missing one in ten in meal boluses does not negatively impact glycemic control. The nine-month, prospective, single-center study tested the impact of announced and unannounced meals on MiniMed 780G system’s ability to maintain glycemic control in adults with type 1 (age 44, 71% male, previously on the MiniMed 640G predictive low glucose suspend system, baseline 6.8% A1c and 68% Time in Range). The system was set with a glucose target of 100 mg/dl and active insulin time of 2 hours. After a 7-day run-in with manual mode and suspend before low settings on, the participants entered a four-day villa period with three pre-defined unannounced meals. Then the participants took part in three three-month periods: the first with announced meals, the second with unannounced meals, and the third with the choice to announce or not announce each meal (“announcement at will”). Using MiniMed 780G, participants’ Time in Range improved from 68% to 78% (+2.4 hours/day), and their mean A1c improved by 0.4% to 6.4% (p>0.05 due to small sample size). Not only did glycemic control improve with use of MiniMed 780G, but also participants were more satisfied with the system according to a post-study questionnaire.

  • Across the three three-month study periods, the “announcement at will” period had the highest Time in Range (79%) and only a 10% reduction in bolusing compared to the announcement period. When asked which study period they preferred, 93% of participants preferred the “announcement at will” period, while 7% preferred the announcement period, sharing comments about their preference for choice with “more or less the same glucose control.” While only a small trial, this investigation highlights two important facts: (i) even when given the choice to announce or not announce, there is only a 10% reduction in meal bolusing and (ii) that reduction in bolusing does not reduce Time in Range or glycemic control with the MiniMed 780G system. Of course, this trial was very small (n=14) and the participants already had strong glycemic control at baseline (average A1c of 6.8% and 68% Time in Range), so further investigation with a larger cohort would be very helpful to validate these findings. We did find the difference, however, from unannounced certainly meaningful, particularly cutting in half the time over 250 mg/dL.

• During his presentation, Dr. Tadej Battelino (University of Ljubljana) compared the data from the US pivotal trial (n=157), New Zealand study (n=59), and FLAIR trial (n-113), which were all read out at ADA 2020. Although the studies evaluated different populations based on age and baseline A1c and Time in Range, all the studies achieved strong glycemic control with two of the trials achieving an average A1c <7% and >70% Time in Range (see below for a side-by-side). Likewise, all study cohorts achieved <4% Time Below Range. In all studies, average time in closed loop was >85%, with the US pivotal trial and the New Zealand study achieving 97% and 96% respectively. (From our view, although these percentages need to be high, we really don’t’ give too much credit for this “achievement” – we’d certainly hope that time in closed loop would be close to 100% and it says nothing about the technology, only that the trial subjects are consistent with wearing the technology.) Of the total 329 participants across the three studies, there was only one severe hypoglycemia event that was unrelated to the MiniMed 780G system (we’ve asked for details how they know, just to be thorough) and a mild case of DKA which was related to a related site occlusion and viral illness. The best outcomes were achieved when using a target of 100 mg/dl and a setting for active insulin time of 2 hours (in the US pivotal trial).  As a reminder, MiniMed 780G is slated for launch in Europe “this fall” with launch expected by April 2021 in the US.

DEVOTE Post-Hoc Analysis Uses Eight-Point SMBG Profiles to Link “Derived Time in Range” With MACE and Severe Hypo Events: 10% Increased TIR Associated With 6% RRR in MACE and 10% in Severe Hypo

In another excellent study validating Time in Range, the seemingly omnipresent Dr. Richard Bergenstal (International Diabetes Center) presented results from a post-hoc analysis of the DEVOTE trial linking “Time in Range” with major adverse cardiovascular events (MACE) and severe hypoglycemia events. The DEVOTE trial was a large cardiovascular outcomes trial for people with type 2 diabetes, originally presented at ADA 2017, comparing Novo Nordisk’s insulin degludec (Tresiba) with Sanofi’s insulin glargine (Lantus). At three points in the trial, the end of year 1, the end of year 2, and the end of treatment, participants in the study had 8-point SMBG glucose profiles taken. Of the nearly 8,000 patients randomized in the original trial, 5,774 had glucose profiles with at least six measurements and were used for analysis. “Time in Range” was derived from these 8-point glucose profiles (more on this below). 

• Every 10% increase in Time in Range was associated with a 6% relative risk reduction for three-point MACE (non-fatal MI, non-fatal stroke, CV death). As seen in the figure below, the incidence of MACE across Time in Range bins does not follow a smooth curve, but shows a clear trend towards higher MACE incidence with lower Time in Range. It’s worth noting that some of the bins, particularly those with lower Time in Range, have a relatively small number of subjects, explaining some of the curve’s choppiness. Compared to people with derived Time in Range ≤50%, a derived Time in Range between 50%-70% was associated with a relative risk reduction of 9% for MACE (HR=0.91, 95% CI: 0.66-1.24) and a derived Time in Range >70% was associated with a 31% relative risk reduction (HR=0.69, 95% CI: 0.52-0.91).

• Every 10% increase in Time in Range was associated with a 10% reduction in the risk of severe hypoglycemia events. Shown in the red line below, the trend towards higher severe hypoglycemia incidence with lower Time in Range is even more pronounced that for MACE. Compared to people with derived Time in Range ≤50%, a derived Time in Range between 50%-70% was associated with a relative risk reduction of 13% for severe hypoglycemia (HR=0.87, 95% CI: 0.63-1.2) and a derived Time in Range was associated with a 46% relative risk reduction (HR=0.54, 95% CI: 0.4-0.73). On this, Dr. Bergenstal commented that the data clearly dispel the notion that “tighter control” (i.e., higher Time in Range) would lead to more hypoglycemic events.

• In addition to MACE and severe hypo, the analysis looked at incidence of microvascular complications, again finding a link between derived Time in Range and rate of complications. Compared to people with derived Time in Range ≤50%, a derived Time in Range between 50%-70% was associated with a relative risk reduction of 27% for severe hypoglycemia (HR=0.73, 95% CI: 0.49-1.09) and a derived Time in Range was associated with a 40% relative risk reduction (HR=0.6, 95% CI: 0.43-0.85).

• As noted above, the “Time in Range” values used in the analysis are not a true, CGM-measured Time in Range, but rather a proxy derived from the 8-point SMBG profiles. This is a very similar analysis to the DCCT post-hoc (which Dr. Bergenstal was also a part of) published in Diabetes Care in 2018. That study similarly used 7-point glucose profiles from the DCCT study to link “Time in Range” with rates of retinopathy and microalbuminuria. When asked during Q&A, Dr. Bergenstal admitted some limitations to the “derived Time in Range” approach, calling for more large-scale CVOTs and other trials to include CGM and Time in Range as part of their analysis – we couldn’t agree more.

• It’s been a big week for Time in Range at EASD 2020. In addition to Dr. Bergenstal’s excellent presentation, Dr. Anass El Malahi (University of Antwerp) presented results on Tuesday from a RESCUE sub-study linking Time in Range with many microvascular complications, particularly retinopathy. On Wednesday, we also caught a poster showing higher Time in Range (but notably, not A1c) was independently correlated with improved quality of life. 

Retrospective Sub-Study of Belgium RESCUE Trial Shows Association Between Time in Range and Microvascular Complications, Especially Retinopathy

University of Antwerp’s Dr. Anass El Malahi presented powerful data strengthening the argument for Time in Range as a primary metric for assessing glucose control. Using data from Belgium’s RESCUE trial (n=515 adult type 1 insulin pump users), Dr. El Malahi demonstrated a strong correlation between Time in Range and several microvascular complications, particularly retinopathy. The sub-study from Dr. El Malahi builds on a growing evidence base validating Time in Range – see a 2018 Diabetes Care paper from Drs. Roy Beck and Rich Bergenstal showing a link between lower Time in Range and higher rates of microvascular complications and a 2019 DT&T paper linking Time in Range with carotid intima-media thickness, a CVD risk biomarker. As CGM has become more common, Time in Range has become a very popular metric for both patients and providers to assess glycemic control and to manage diabetes. However, some have been less open to trusting Time in Range as a primary metric due to a number of things, including the absence of a long-term outcomes “DCCT-like” trial for Time in Range validating the link between Time in Range with complications.

• Notably, higher Time in Range was significantly associated with fewer microvascular complications in this trial. For example, half of participants with a Time in Range <40% had at least one microvascular complication, i.e., peripheral or autonomic neuropathy, retinopathy, or nephropathy. In contrast, just 27% of those with a Time in Range >70% had at least one microvascular complication. We will be back on p values.

• Retinopathy was very strongly correlated with lower Time in Range, with the average retinopathy rate in the Time in Range > 70% group of 22% far lower than the rate of retinopathy of Time in Range <40% group – fully 50% of those with Time in Range less than 40% had retinopathy.

• Notably, just 8% of those with Time in Range over 70% had nephropathy, and 10%-13% had nephropathy between 40% and 69% Time in Range – the big change came between 40% Time in Range, where 30% of those with Time in range under 40% had nephropathy.

• Notably, the link between microvascular complications and Time in Range existed independent of diabetes duration. The 22% retinopathy rate seen in the Time in Range >70% group is considerably higher than the 5% retinopathy rate reported by the DCCT group with estimated Time in Range >70%. We will be asking experts their views on this.

• The correlation between macrovascular complications and Time in Range was less clear. About one in five participants in the Time in Range <40% group had presence of at least one macrovascular complication, while just 3% of participants in the Time in Range >70% group had at least one macrovascular complication. However, the rates of macrovascular complications were low once Time in Range crossed 40% - nephropathy, above, is similar to this as well and some would term nephropathy itself as closer to a “macrovascular” benefit rather than a microvascular one.

• Higher Time in Range was also linked to fewer diabetes-related hospitalizations. Mean Time in Range for participants who had no hypo- or DKA-related hospitalizations in the last twelve months was 64%, compared to 59% for participants who had been hospitalized in the last twelve months (p=0.02).

• Interestingly, the data from the RESCUE study also showed that while A1c was significantly improved after initiating CGM, Time in Range remained very similar. The mean A1c in the group improved just four months after initiating CGM from 7.6% to 7.2%. After 24 months, mean A1c had was reduced by 0.3% compared to baseline. In contrast, mean Time in Range at baseline 63% and remained in the low 60% range through every time period. This was somewhat surprising and we look forward to getting more expert commentary on this.

• As a reminder, the RESCUE data set includes outcomes from national reimbursement of CGM in Belgium for type 1s on insulin pump therapy. The registry includes data from September 2014 to January 2017. The first data from the registry were presented at EASD 2017, demonstrating a 75% reduction in hypoglycemia/DKA-related hospitalizations in the first 12 months following CGM initiation.

Prospective “LIBERATES” RCT (n=141): FreeStyle Libre Use Significantly Reduces Time <70 mg/dl (-1.3 hour/day) and Time <54 mg/dl (-25 min/day) in Type 2s with Recent Myocardial Infarction Relative to BGM

During a 90-minute session on Wednesday, Drs. Robert Storey, Deborah Stocken, and Ramzi Ajjan (University of Leeds) presented the findings of LIBERATES, a multi-center, open-label, prospective, randomized controlled trial which compared the glycemic control of type 2s following a heart attack using either FreeStyle Libre (intervention) or BGM (control). Within 5 days of myocardial infarction, participants (n=141, 63 years, 73% male, 90% white, mean BMI of 31 and A1c of 8.9%) were randomized to the BGM control group (n=72) and the FreeStyle Libre intervention group (n=69) – the groups were demographically statistically identical. All participants were either on insulin therapy or sulfonylurea (55% and 45% respectively) with 86% on additional glucose-lowering medications. More than 90% of participants were on anti-anginal/hypertensive therapies, anti-thrombotic therapies, and/or anti-hyperlipidemic therapies. Unsurprisingly, participants had a slew of diabetes comorbidities: 19% of participants had coronary artery disease, 11% heart failure, 7% atrial fibrillation, 13% cerebrovascular disease, 7% peripheral vascular disease, 19% retinopathy, 20% neuropathy, 9% eGFRs <60 ml/min/m², and 34% albuminuria. No participants had active malignancies or required dialysis. Researchers compared a series of intervention and control group outcomes, including Time in Range, time below range, time <54 mg/dl, A1c, adverse events including MACE, quality of life, and cost-effectiveness. To assess glycemic measures, the control group wore blinded FreeStyle Libre Pro during data collection periods.

• More than anything, we were intrigued by the study’s design and population, particularly Dr. Stocken’s reference to the trial as “like a phase 2b trial,” terminology usually associated with medication trials, rather than devices. Specifically, Dr. Stocken explained the next step would be a “phase 3,” larger, longer-term, international study which compares the impact of FreeStyle Libre and BGM on MACE and cardiovascular outcomes in type 2s following a heart attack (if funding can be acquired). Similar to the large-scale CVOTs run for SGLT-2s and GLP-1s, we wonder if a large-scale trial using FreeStyle Libre (or other CGM) would find effects beyond improving glycemia (e.g., reducing A1c and improving Time in Range), and also reducing incidence of CV events and mortality. As noted by Dr. Storey in their presentation, the interaction between hypoglycemia and MACE is likely mediated through endothelial cell dysfunction, increased inflammation, and increased thrombosis. Given CGM can reduce rates of hypoglycemia, it certainly seems possible that a trial like this could make CGM a useful not only in diabetes, but also in cardiology. This is speculation, of course, although we’d be interested in seeing prospective observational trials on this front as well – in this age, “standard of care” would presumably have to include an SGLT-2 or GLP-1 for most people with type 2 or most trial participants may have to be newly diagnosed or without CV risk, a hard population to find and a harder one to show results in, in a CVOT.

• The improvement in hypoglycemia, especially for those on sulfonylureas, was the most significant finding from the trial. At day 30, the FreeStyle Libre group trended towards reduced time in hypoglycemia, spending ~45 fewer minutes/day below 70 mg/dl relative to the control group. This hypoglycemic improvement actually grew to a significant 1.3-hour/day reduction in time below <70 mg/dl for the FreeStyle Libre arm relative to the control group by three months. The improvement is particularly stark among those using sulfonylureas at baseline (n=45). At day 30, sulfonylurea users in the intervention group saw a ~1 hour/day reduction in time <70 mg/dl relative to their control group counterparts (not statistically significant due to low sample size). By day 90, that relative improvement climbs to a statistically significant ~2 hour/day reduction in time below range. This suggests that the three-month hypoglycemia-reducing benefit of wearing FreeStyle Libre is particularly strong among sulfonylurea users – presumably this has .

• The results were also striking for severe hypoglycemia, which as Dr. Storey noted, is linked to long-term incidence of MACE. A month into the trial, the intervention arm saw a ~25 minute/day reduction in severe hypoglycemia (<54 mg/dl) relative to the control group, with the baseline insulin and sulfonylurea users seeing similar benefits. This effect was maintained at ~25 minute/day relative to the control group at three months. Once again, at day 90, FreeStyle Libre users on sulfonylureas at baseline saw a significant ~35 minute/day reduction in time <54 mg/dl relative to their counterparts in the control group; those who were on insulin at baseline observed a nonsignificant ~10 minute/day relative improvement in time <54 mg/dl. Overall, the severe hypoglycemia reduction associated with FreeStyle Libre use is hugely exciting, especially given that previous research has tied hypoglycemia to long-term negative outcomes, including MACE and death, as explained above. With a larger sample and longer study, one might expect to see a relative risk reduction in MACE from using FreeStyle Libre.

• To close out, Dr. Ajjan shared preliminary data on adverse events, including severe hypoglycemia, mortality, and MACE. The intervention group saw zero severe hypoglycemia events whereas the control group saw two (2.8%), a minimal between-group difference, although the sample size (n=141) and a 90-day study length make it hard to draw any definitive conclusions. The intervention group and control group had two and three deaths, respectively. Surprisingly, the preliminary MACE data indicates that the FreeStyle Libre group saw higher incidence of MACE over time, although as Dr. Ajjan noted, the data is still very preliminary; also of note, most MACE occurred after day 90 when neither group was on FreeStyle Libre.

• Time in Range differences between the control and intervention group were nonsignificant, but trends suggest that the intervention arm incurs a benefit in Time in Range both at day 30 and day 90. The researchers found that relative to the control group, FreeStyle Libre users saw a non-significant 48 minute/day increase in Time in Range between days 76 and 90, an improvement that was primarily driven by insulin users. (For context, a 48 minute/day improvement in Time in Range translates to about 3.3%.) Looking at data early in the trial (days 16-30), the Time in Range improvement is greater: the intervention nonsignificantly increased Time in Range by ~1.5 hours/day, and among baseline insulin users (n=55), Time in Range increased significantly by 3 hours after 30 days. In addition, older age and higher baseline A1c values were correlated with a lower end-of-study Time in Range.

• After the three-month study period, there was no difference in A1c between the FreeStyle Libre intervention group and the BGM control group, although both groups’ average A1c trended downward, dropping by 0.6% (baselines of 8.8% and 9% for the control and FreeStyle Libre groups, respectively). Of course, given the length of the trial (90 days), we’d be more interested to see the mean glucose (and GMI) measures for the two groups, rather than just A1c.

• Overall, satisfaction and engagement with FreeStyle was high in the intervention group. At day 90, treatment satisfaction (measured with DTSQ) was greater for the FreeStyle Libre group than for the BGM group (34 vs. 32), although both groups saw an improvement relative to baseline. Although Dr. Ajjan didn’t share data on this, he did comment that the intervention group had greater awareness of hyper- and hypoglycemia and improved understanding of their own diabetes and scored higher than the control group in terms of treatment convenience and flexibility. In terms of engagement, in the intervention arm, average scanning frequency was six scans/day and did not significantly change over the course of the three-month trial.
  

Real-World Data from 9,216 US Dexcom Users: 59% Time in Range, 2% Time <70 mg/dl; Rebound Hyper- and Hypoglycemia Much Greater in Type 1s; CGM Use Highly Persistent

During a morning session, Dexcom data scientist Robert Dowd read off results from a real-world data set of Dexcom users. The anonymized data was collected from 9,216 Dexcom G6 users who began using the device before December 1, 2019. The dataset only included users with ≥30 days of uploaded CGM data. The large sample included both adults and pediatrics (G6 is indicated for use down to 2 years old) and people diagnosed with type 1 and type 2 diabetes (n=5,426 type 1s and 3,790 type 2s). Not surprisingly, the type 2 cohort was significantly older than the type 1 cohort (mean age of 33 years vs. 52 years).

• Mean Time in Range for type 1s was 57% (13.7 hours/day), significantly lower than that of type 2s (62%; 14.9 hours/day). Of course, neither group is meeting the consensus guidelines for >70% Time in Range (16.8 hours/day). Most of the improvement needs to come from time spent in hyperglycemia: in the Dexcom data set, type 1s spent 14.3% time >250 mg/dl and type 2s spent 13.8% time >250 mg/dl. These are both far above the <5% time >250 mg/dl consensus targets. More encouragingly, time in hypoglycemia (<70 mg/dl) was low in both groups: 2% for type 1s and 1.9% for type 2s – easily meeting the consensus 4% time <70 mg/dl goal.

  • This large data set from Dexcom provides another valuable reference point for estimating Time in Range in the broad population of people with diabetes. The data from Dexcom matches well with Abbott’s much-larger real-world data set (n=592,328 readers) showing a mean Time in Range of 56% (see poster at ADA 2019). However, both of these data sets (Abbott’s and Dexcom’s) are obviously a self-selected group of CGM users who likely have better glycemic control than the broader population. While Dexcom didn’t report mean A1c, the estimated A1c in Abbott’s data set ranges from 6.7%-8.2% - lower at the midpoint than the diabetes population’s average A1c.

• The frequency and severity of rebound hyper- and hypoglycemia events were much higher for type 1s than type 2s. Rebound hyperglycemia events were defined as CGM values >180 mg/dl less than two hours following a reading <70 mg/dl. Rebound lows were defined as the inverse (hyperglycemic values shortly followed by hypoglycemic values). Rebound lows were much more common than rebound highs. In 2019, type 1s experienced a mean of 162 rebound hypoglycemia events, compared to 60 for type 2s (~one every 2.25 days vs. one every six days). Type 1s experienced a mean of 45 rebound hyperglycemia events, while type 2s experienced a mean of 14 events (one every eight days vs. once per month). Severity of the events, as measured by area under the curve, was also significantly higher for type 1s than type 2s.

• Device use persistence was very high in both cohorts at 87%-89%. Additionally, the urgent low soon and Dexcom Clarity features were both very popular. On a related note, Dexcom’s Dr. Giada Acciaroli presented real-world data from Dexcom users who transitioned from Dexcom G5 to G6 at ADA 2019. Those results suggested the urgent low soon feature with Dexcom G6 helped significantly reduce rebound hyperglycemia events. Finally, rates of CGM data sharing in type 1s was higher (63%) than in type 2s (40%).

Selected CGM Posters: Higher Time in Range and Use of Data Sharing Correlated with Higher Quality of Life; Promising Animal Study of Implantable Multi-Analyte Sensor (Glucose, Ketones, and Lactate)

This afternoon, we rounded up a few posters that caught our eyes in the (virtual) poster hall. The first poster evaluates relationship between glycemic variability and Time in Range on quality of life. The second, led by Scripps’ Dr. Addie Foreman and Behavioral Diabetes Institute’s Dr. Bill Polonsky, evaluates how CGM data sharing affects quality of life. Finally, we wanted to highlight an (admittedly) early-stage implantable multi-analyte sensor from Belgian startup Indigo Diabetes.

• In another very interesting sub-analysis from the Belgian RESCUE trial, KU Leuven’s Dr. Pieter Gillard presented a poster linking higher Time in Range with improved quality of life. In case you missed it, Dr. Anass El Malahi’s presentation on RESCUE yesterday linked Time in Range with microvascular complications. Dr. Gillard’s sub-analysis evaluated the relationship between glucose variability, Time in Range, and quality of life (HFS-Worry and PAID-SF questionnaires) in RESCUE’s 515 adult type 1 pump users. Shorter Time in Range was independently correlated with “more limitations due to emotional problems” (p=0.02), “lower vitality” (p<0.01), and more diabetes-related emotional distress (p<0.01). Surprisingly, glucose variability (%CV), age, diabetes duration, and A1c were not independently associated with quality of life measures. While quality of life is perhaps not talked about enough, the analysis is another important validation of Time in Range as the primary diabetes management metric.  

• Using data from an online questionnaire (n=302), Scripps Whittier Diabetes Institute’s Dr. Addie Fortmann and BDI’s Dr. Bill Polonsky, identified behaviors around CGM data sharing that resulted in higher quality of life. The survey included adults with type 1 diabetes using Dexcom G5 or G6 and the Share feature. The participants were mostly female (60%) and college graduates (58%) with a mean age of 43 years. The vast majority (81%) reported that their chief follower was their spouse/partner. 70% of the group had discussed with their follower how to best respond to the CGM data; still, nearly 18% of the group said that their follower was “bugging them” too much and 24% reported that they felt more judged by their follower because of Share. Most actionable, the survey results also indicated that more frequent “encouraging” and “celebrating” by the person following CGM data were linked to improved clinical outcomes (decreased A1c and frequency of severe hypo), as well as improved quality of life (as measured by the diabetes distress scale, hypoglycemic confidence scale, and WHO-5 well-being scale).  

• In our final selected CGM poster, Dr. David Stocker (Indigo Diabetes) presented data from animal proof-of-concept studies for the Belgian startup’s implantable multi-analyte sensor. As a reminder, the company recently closed a ~$45 million Series B to support development and study of its miniature spectrometer-based implantable sensor. Two Indigo prototype sensors were implanted on each side of four adult female Gottingen minipigs for two months. The pigs underwent IV challenges with glucose, ketones (beta-hydroxybutyrate), and lactate with sensor readings compared to reference samples analyzed on-site. With 765 paired data points, MARD for glucose readings (between 40 and 400 mg/dl) was ~6.5% compared to analyzer reference. For ketones, the mean absolute difference between sensor readings and a commercial ketone meter (A. Menarini) was ~0.23 mM (n=125 paired points). Finally, for lactate, the mean absolute difference between sensor readings and analyzer readings was ~0.5 mM (n=288 paired points). Indigo is aiming for market approval by 2023, initially for six-month wear, extended to two-year wear by 2025. The device is expected to be non-adjunctive with no on-body transmitter required.

Swedish National Diabetes Register: 84% of Adult Type 1s, 94% of Pediatrics Now Using CGM (Mostly FreeStyle Libre); Massive Population-Level Outcomes Improvements from 1996 to 2018

At an Abbott-sponsored symposium, Dr. Katarina Eeg-Olofsson (University of Gothenburg) shared a number of interesting data points from the Swedish National Diabetes Register, highlighting a significant improvement in population-level A1c outcomes for adult type 1s in Sweden from 1996 to 2018. In 2018, nearly 18% of adults type 1s achieved an A1c ≤6.5% (≤48 mmol/mol), more than doubling from the 7% of adults achieving the same A1c in 1996. Looking at the opposite end, ~18% of adults in the registry had an A1c >8.6% (>70 mmol/mol) in 2018, a massive improvement from the 42% of adults in 1996. Perhaps more telling, in 1996, there were about six-times more adults in the A1c >8.6% bucket compared to the A1c ≤6.5% bucket; in 2018, these two groups were about equal in size – wow! This was certainly very good to see. It’s worth noting that the data from Sweden (that includes more than 90% of type 1s in Sweden) is very encouraging, but contrasts with the population-level results seen in the US. T1DExchange data, which comes from the top diabetes centers in the US, showed just 21% of adult type 1s achieving an A1c <7% in 2016-2018. Even more concerning, the mean adjusted A1c in the US rose from 7.8% in 2010-2012 to 8.4% in 2016-2018, though, admittedly, much of this rise was driven by adolescents and young adults.

• While the outcomes improvements in Sweden no doubt come from a variety of places, Dr. Eeg-Olofsson stated in no uncertain terms that CGM is a significant driver. Again, looking at registry data, the percentage of adult type 1s in Sweden using CGM was estimated at 84% in 2020. As seen in the graph below, CGM uptake has been incredibly quick Sweden, rising from less than 1% just five years ago! Of the CGM users, more than three-quarters (77%) are using Abbott’s FreeStyle Libre, while the remainder are using other CGM, presumably Dexcom. Dr. Eeg-Olofsson contrasted CGM uptake with that of insulin pumps, which have been flat-to-slowly-rising: in 2014, pump utilization in the register was ~21% compared to ~25% in 2020. While these figures are impressive, we’d love to see measures like

  • Retinopathy (% in 1996 vs. % in 2018);

  • Severe hypoglycemia (% in 1996 vs. % in 2018);

  • Kidney disease (% in 1996 vs. % in 2018);

  • Heart disease (% in 1996 vs. % in 2018); and

  • DKA (% in 1996 vs. % in 2018).

• In another very encouraging result, CGM utilization has grown across Sweden’s geographical regions. As shown in the graph below, the variability in CGM utilization across regions was massive just three years ago: in 2017, four areas had utilization rates >70%, while six regions had utilization below 50%. In 2019, this gap had shrunk greatly, with all regions achieving CGM utilization rates >50% and all but one area above 70%. This data suggests that Sweden has done a good job of making CGM (and thus, better diabetes management) accessible to all, regardless of geographical location.

• In yet another exciting result from the National Diabetes Register, Dr. Eeg-Olofsson showed significant A1c improvements for both type 1s and type 2s initiating FreeStyle Libre. The retrospective analysis looked at 8,316 type 1s and 538 type 2s in the registry with A1c readings before and 12-months after starting FreeStyle Libre (study period between January 2014 and June 2019). For type 1s, A1c was reduced from 8.1% to 7.8% after 12 months of FreeStyle Libre; for type 2s, A1c was reduced by from 8.6% to 8.1% after 12 months. Unsurprisingly, the A1c reductions were even larger for those naïve to CGM. For type 1s in this group, A1c was reduced from 8.2% to 7.8% (0.4% reduction); for type 2s, A1c was reduced from 8.7% to 8% (0.7% reduction). Of course, we’d love to know the respective Time in Range measurements and will ask for more information on this front.

• Later in the day at a Dexcom-sponsored symposium, Dr. Peter Adolfsson (University of Gothenburg) shared equally impressive figures on pediatric outcomes and CGM use in Sweden. Based on the Swedish National Diabetes Registry, 95% of children with type 1 use CGM, after a sharp rise from 4% in 2015 t0 78% in 2016. Within his own clinic in Kungsbacka, ~94% of children with type 1 use real-time CGM, ~6% use FreeStyle Libre, and <1% are on SMBG. This stands in contrast to the adult data, which shows the vast majority of type 1s using FreeStyle Libre; we’d guess the popularity of real-time CGM in the pediatric group reflects the alarm and real-time remote monitoring features. Perhaps even more impressive, as of August, the average pediatric Swedish patient had an A1c of 7.0%, meeting the ADA’s target for pediatric patients with diabetes; for context, T1D Exchange data from 2016-2018 showed a mean A1c in US children of ~8.1% at 5 years old, peaking at ~9.3% between ages 15-18.

CGM for All? Dexcom-Sponsored Symposium Highlights Studies on CGM Efficacy for Pregnant Women, Teens, and Seniors

In its second industry symposium of EASD 2020 (see the first here), Dexcom highlighted data on CGM use in historically underpenetrated populations including pregnant women, teens and young adults, and seniors. During her presentation, the eminent Dr. Helen Murphy (Cambridge) encouraged providers and pregnant women to focus more Time in Range and average glucose than A1c targets, as A1c is affected by pregnancy hormones in all women, with or without diabetes. Consensus Time in Range goals for pregnant women were included in the 2019 updated ADA Standards of Care. During the early pregnancy (8-16 weeks), Dr. Murphy warned that hypoglycemia can occur quickly with little to no warning, advising women to cut back on insulin during this time period and use “micro-carb” doses to avoid hypoglycemia (in addition to having rescue glucagon, if necessary). From 14-18 weeks, Dr. Murphy noted women will start to feel better and should begin to pay more attention to prandial insulin doses, potentially increasing them. In the late second trimester, from 20-24 weeks, Dr. Murphy explained women tend to experience higher glucose levels for longer periods of time, displaying more signs of growing insulin resistance. At this point, Dr. Murphy suggested working with a dietician to adjust meals, increasing pre-meal dosing, taking pre-meal doses earlier, and including physical activity after eating. Additionally, Dr. Murphy acknowledged that many women struggle to meet the >70% Time in Range goals, especially early in pregnancy when most women report Time in Range closer to 40%-55%. Despite this, Dr. Murphy encouraged providers and patients to remember that every 5% increase in Time in Range has a clinical and statistically significant impact on maternal and neonatal outcomes. As a reminder, Dexcom’s G6 (since February) and Abbott’s FreeStyle Libre (since 2017) are the only CGMs with a pregnancy indication in Europe, while no CGM has yet been approved for pregnancy use in the US. However, given the challenge of managing glycemia during pregnancy, off-label use of CGM is quite common.

• To start, Dr. Murphy recapped results from the landmark CONCEPTT RCT (n=215; presented at EASD 2017). Using Medtronic’s older Guardian CGM, the CGM arm showed a small 0.2% advantage on the primary outcome, A1c, as well as higher Time in Range (68% vs. 61%; ~100 minutes/day) compared to the control group. Perhaps more importantly, neonatal outcomes for the women in the CGM group were significantly better: CGM drove a significant reduction in the incidence of large for gestational age (OR=0.51, p=0.02), fewer NICU admissions lasting 24+ hours (OR=0.48, p=0.02), fewer incidences of neonatal hypoglycemia (OR=0.45, p=0.03), and one-day shorter length of hospital stay (p=0.01).

• Cost-effectiveness analysis from CONCEPTT have also backed up the use of CGM in pregnancy. As Dr. Murphy plainly stated at ADA 2019, “for publicly funded health systems like the NHS, they cannot afford not to provide CGM to pregnancy women with diabetes.” Given the low numbers needed to treat in the CONCEPTT trial – about 6-8 women on CGM to prevent one adverse neonatal outcome – and the ~ £330 incremental cost for using CGM, Health Technology Wales estimated “cost savings of £1,029 [~$1,300] per pregnancy.”

• Later on, Dr. Hood Thabit (Manchester University) highlighted efficacy of CGM in teens and young adults drawing on data from the SENCE, CITY, and MILLENIAL studies. Across these three studies, CGM reduced time spent >180 mg/dl, improved Time in Range, and reduced hypoglycemia. All three studies also demonstrated participants spent the majority of the time wearing their sensors and Dr. Thabit emphasized this point as a key metric of the studies’ successes. Additionally, in the SENCE study (children ages 2-7), the families of children with type 1 diabetes were randomized to receive family behavioral interventions. These interventions were shown to reduce parental burden, increase CGM satisfaction, and reduce fear of hypoglycemia indicating educating those around young people with diabetes may be a key factor in improving diabetes management in young people. However, it’s still worth noting that SENCE (and other studies) are encouraging, but a reminder of how much there is left to improve: even after a year of CGM use, young children with type 1 diabetes in the rigorous study treated at top clinical centers still spent over half of their days with glucose levels >180 mg/dl and just ~40% in-range.

• Dr. Thabit also presented data from the WISDM trial in seniors >60 years with diabetes. Over the six-month study, participants were randomized to wear a Dexcom G5 or SMBG with a masked G5 CGM. After 6 months, participants on CGM spent two more hours in range per day compared to those on SMBG (15 hours/day vs. 13 hours/day; p<0.001). Additionally, patients using CGM saw a significant reduction in severe hypoglycemia events and an overall reduction in time spent in hypoglycemia especially in patients who were previously hypoglycemia unaware. High CGM utilization was maintained 12 months after initiation due to high levels of trust and satisfaction. The CITY and WISDM trials were published in JAMA back in June.

• During Q&A, both Dr. Murphy and Dr. Thabit emphasized that CGM should be made accessible to all people with diabetes regardless of socioeconomic background and technological literacy. Dr. Thabit spoke from his personal experience at the Manchester Diabetes, Endocrinology, and Metabolism Center, citing their great success starting many CGM naïve patients on sensors. Additionally, Dr. Thabit expressed that CGM education up front is the key to improved patient outcomes. Dr. Murphy and Dr. Thabit also answered a question on using A1c and Time in Range, agreeing that they both found Time in Range much more accessible for patients and much more actionable for providers.

Dr. Rich Bergenstal on Using CGM in Non-Insulin Type 2s and Primary Care: Know the Targets, Observe the Activity, and Learn; Many Patients Want to Make Lifestyle Changes, Not Take More Medications

Closing out Abbott’s morning symposium, the extraordinary Dr. Rich Bergenstal (International Diabetes Center) gave a very important talk on utilizing CGM in primary care settings, particularly for type 2s not using insulin. Setting the stage for his presentation, Dr. Bergenstal noted that clinic leadership in primary care settings often focus on the “big picture,” namely, the ABCs (A1c, blood pressure, and cholesterol), smoking cessation, and aspiring. Dr. Bergenstal critiqued this “big picture” approach to diabetes management, citing a study demonstrating improved blood pressure and cholesterol outcomes, yet flat-to-worsening A1c outcomes (Fang 2020). The results largely mirror those of a 2019 JAMA article showing slight improvement in blood pressure + cholesterol control, yet a worsening rate of those meeting combined A1c, blood pressure, and cholesterol targets from 2005 to 2016. Finally, Dr. Bergenstal showed data from his own center in Minneapolis (n>60,000), showing high rates of patients meeting blood pressure, cholesterol, tobacco-free, and aspiring goals, but A1c goal achievement lagging considerably behind. Dr. Bergenstal argued for a more glucose-centric focus on diabetes management, including CGM.

• To help explain CGM to new patients, Dr. Bergenstal’s group developed a handout with an easy-to-remember three-step mantra: know, observe, and learn. The first step (“know”) is straightforward, explaining CGM and the consensus Time in Range goals. To “observe,” CGM users are encouraged to take mental note of their behaviors and their effect on CGM values; these include sleep, stress, medications, physical activities, foods, etc. Finally, users are encouraged to write their observations down or internalize them in the “learn” step, helping make behavior changes that keep their blood glucose in range. This three-step mantra is similar to Adam Brown’s “Bright Spots and Landmines” paradigm, identifying behaviors that drive blood glucose in and out of range – ultimately, these sorts of easy-to-remember paradigms will be helpful in making CGM more accessible and useful to more people.

• In another important point promoting the use of CGM in type 2s not on insulin, Dr. Bergenstal noted that many people already on oral medications (e.g., metformin, SUs, SGLT-2s) are reluctant to start more medications and would rather pursue lifestyle changes. Of course, CGM is not only a powerful learning tool, but also a powerful behavior modification tool. Using a case study, Dr. Bergenstal presented one patient with an A1c of 8.2% and CKD on metformin and an SGLT-2 inhibitor. Rather than initiate a new pharmacological therapy, the patient asked for three months to make lifestyle changes. With CGM, the patient’s A1c dropped to 7% with a Time in Range of 78%; the CGM had helped the patient give up sweet tea, increase physical activity, improve medication adherence, and make other lifestyle changes that improved glycemic control.

• Explaining another series of easy-to-remember mantras, Dr. Bergenstal briefly discussed his MGLR (“More Green, Less Red”) and FNIR (“Flat, Narrow, and In-Range”) mnemonics. The first, MGLR, refers to the time in ranges bar on a standardized AGP, with Time in Range shown in green and time in hypoglycemia shown in red. The FNIR, of course, refers to an ideal glucose profile in the AGP: a flat and narrow glucose line (low glucose variability) and in-range (mean glucose between 70-180 mg/dl).

Sanofi-Sponsored Symposium: Dr. Rich Bergenstal Advocates for CGM in Drug Trials, Dr. Pratik Choudhary on Using Time in Range in the Clinic

In a powerful sign of the growing importance of CGM and diabetes technology in pharmaceutical therapies, Sanofi dedicated the final hour of its symposium to discussing CGM, Time in Range, and “integrated care.” The busy Dr. Richard Bergenstal (International Diabetes Center) gave a popular presentation, giving an overview of CGM, but also calling for more use of CGM in clinical trials for drugs. After discussing the limitations of A1c, Dr. Bergenstal noted that CGM-derived metrics (e.g., Time in Range) could be much more powerful at understanding and evaluating various therapies and intervention strategies, rather than relying exclusively on topline A1c results. Dr. Bergenstal then provided two examples of clinical trials assessing Time in Range. Firstly, the inRange study (ClinicalTrials.gov page), currently recruiting, will evaluate CGM outcomes for ~340 type 1s randomized to either Toujeo (insulin glargine) or Toujeo (insulin degludec) basal insulins. The second, the OneCARE study, which took place in Spain and will be presented as a poster at EASD 2020, also compared insulins glargine and degludec using both CGM-derived metrics and patient-reported outcomes.

• Leicester Diabetes Research Center’s Dr. Pratik Choudhary nicely made the case for using Time in Range in clinical practice, saying, “[Time in Range] works because it incorporates both mean glucose and variability into the same number.” Additionally, the metric “makes more sense” to patients – a sentiment we’ve heard from both providers and people with diabetes over the last few years. Notably, Dr. Choudhary also shared data gathered from his center, showing that the proportion of fingerstick results in range correlated well with A1c and in a similar way to CGM-derived Time in Range. In other words, BGM-derived Time in Range metrics can likely be used in a similar way to CGM-derived Time in Range, further establishing the importance of building in connectivity and well-designed reports for the majority of people with diabetes still on BGM.

Glucose Monitoring Posters

Title	Authors	Details + Takeaways
Assessment of sensor performance of a blinded professional continuous glucose monitoring system in hospitalized patients  with type 2 diabetes	Hans C Haberl, Felix Aberer, Petra Baumann, Peter Beck, Lirie Bytyqi, Astrid Fahrleitner-Pammer, Daniel Hochfellner, Julia Mader, Tina Pöttler, Judith Samonigg, Haris Ziko	Inpatients with type 2 diabetes (n=30) participated in the study comparing the accuracy of FreeStyle Libre CGM compared to point of care testing in a hospital setting; mean participant age was 73, 60% of participants were female Sensors were used for an average of 7 days; overall MARD of 8.4%; hypoglycemic MARD of 11.7%; euglycemic MARD of 8.9%; hyperglycemic MARD of 7% Results suggest CGM accuracy during hospitalization on a general ward is comparable to data under normal circumstances; potential for hospitals to use CGM to monitor glucose levels for inpatients
Hospitalisations for acute complications before and after FreeStyle Libre® system initiation in people with type 1 and type 2 diabetes  in France	Bruno Guerci, Jean-Pierre Riveline, Bruno Detournay, Gérard De Pouvourville, Corinne Emery, Fleur Levrat-Guillen, Ronan Roussel, Eric Vicaut	Longitudinal retrospective cohort study in patients using FreeStyle Libre with one full year of follow up; 33,165 patients with type 1 diabetes, 74,076 patients with type 2 diabetes 6.4% of patients with type 1 experienced hospitalization for diabetes complications in the year prior to starting FreeStyle libre vs. 3.3% of patients with type 1 in the year following initiation (48% relative reduction) 2.7% of patients with type 2 experienced hospitalization for diabetes complications in the year prior to starting FreeStyle Libre vs. 1.6% of patients with type 2 in the year following initiation (41% relative reduction)
Diagnosis of gestational diabetes in the first trimester of pregnancy is associated with differences in CGM night profile	Ondřej Krystyník, Lubica Cibickova, Dominika Goldmannova, David Karasek, Jan Schovanek	54 women participated in the study; 21 were not pregnant (control), 15 were pregnant but did not have gestational diabetes, and 18 were pregnant with gestational diabetes; participants wore blinded CGM for 7 days between 9-15, 24-28, and 34-38 weeks of gestation Participants with gestational diabetes were further divided into two groups based on method of diagnosis as either oral glucose tolerance test or fasting glycemia Women diagnosed with gestational diabetes based on fasting glucose levels had significantly higher nocturnal glucose levels compared to healthy pregnant women (p=0.04) Elevated nocturnal glycemia among pregnant women with gestational diabetes was apparent as early as the first trimester
Marked improvements in HbA1c following flash monitor initiation  in people with type 1 diabetes: a nationwide observational study  in Scotland	Fraser Gibb, Helen Colhoun, John McKnight, Luke Blackbourn, Brian Kennon, Anita Jeyam, Stuart McGurnaghan, Paul McKeigue	Using data from the Scottish Care Information—Diabetes Collaboration database researchers conducted a retrospective cohort study comparing outcomes before and after initiating flash glucose monitoring 9,655 patients contributed data to the study; in 2019, prevalence of flash glucose monitoring in Scotland was 36%; median age at flash monitoring initiation was 38; 26% of flash monitor users were on pump therapy; flash glucose monitoring use was disproportionately lower among patients with low socioeconomic resources After initiation of flash glucose monitoring population-level A1c decreased from 8.2% at baseline to 7.8% at one year and 7.5% at two years Within-person A1c reduction were 0.3% at one year after initiation and 0.4% at two years; the effect of flash glucose monitoring varied significantly depending on baseline A1c, age at initiation, and prior pump use (p≤0.01), but did not vary by sex DKA events were significantly reduced following initiation of flash glucose monitoring; mixed models indicated an increasing trend in DKA events prior to initiation of flash glucose monitoring
Associations between variability in glycated haemoglobin (HbA1c) and glycaemic control with severe hypoglycaemia in adults with type 2 diabetes	Giat Khee, Yong Mong Bee, Hua Heng Cheen, Su-Yen Goh, Paik Shia Lim, Yan Zhi Tan, Ming Ming Teh, Julian Thumboo	1,197 patients with type 2 diabetes who presented with severe hypoglycemia, mean age 65, 52% female, at a Singapore academic medical center were included in a retrospective case control study and matched with controls 1:1 using baseline demographics Results indicated patients with greater glycemic variability were associated with greater risk of severe hypoglycemia Patients with higher A1c levels were also associated with higher risk for severe hypoglycemia
Continuous glucose monitoring record length and minimum number of daily observations for clinical interpretation	Jędrzej Chrzanowski, Wojciech Fendler, Przemysław Kucharski, Beata Mianowska,  Arkadiusz Michalak, Konrad Pagacz, Agnieszka Szadkowska	Retrospective analysis of point-of-care CGM records in children (n=494) with type 1 diabetes duration >1 year between 2015-2019; 290 participants used Medtronic CGM, 204 participants used Abbott flash glucose CGM Results indicated minimum record duration for reliable CGM interpretation were 7 and 35 days for short- and long-term assessment, respectively; mean glucose variability was more robust than glucose standard deviation and coefficient of variation Current consensus guidelines (>14 days, >70% sensor utilization) only satisfy short term assessment to reliable CGM analysis; for reliable interpretation of short CGM records, higher sensor utilization (>80%) is required
Tracking haemoglobin A1c from CGM data via personalised model of haemoglobin glycation and clearance	Chiara Fabris, Roy Beck, Boris Kovatchev	Using data from the DIAMOND trial with 153 participants with type 1 diabetes and 151 participants with type 2 diabetes researchers developed a dynamic model of hemoglobin glycation and clearance CGM data was collected from 6 months from all participants; A1c was measured at month 3 and month 6 Developed first order differential equation driven by time in target range; model was trained on data from 50 participants with type 1 and 50 participants with type 2; trained model was validated using data from remaining 103 participants with type 1 and 101 participants with type 2; model calibrated with A1c values at 3 months and prospectively tested with A1c values at 6 months Model produced correlation factor between estimated A1c and laboratory A1c of ~0.9 indicating one-time calibration with lab A1c allows for high agreement between CGM-based eA1c and lab A1c
How are the relationships between hypoglycaemia over 2 weeks and glycaemic variability different for different measurement durations of glycaemic variability?	Soichi Takeishi, Kyosuke Miura	Cross sectional study analyzing within week Time in Range and glycemic variability using FreeStyle Libre Pro CGM for 13 days among hospitalized patients (n=168) with type 2 diabetes Mean absolute glucose, coefficient of variation, and glycemic variability percentage were all positively correlated with time ≤70 mg/dl; mean absolute glucose was more strongly correlated with time in hypoglycemia than coefficient of variation or glycemic variability percentage The correlations between mean absolute glucose, coefficient of variation, and glycemic variability percentage with time spent in hypoglycemia held under different patterns of glucose traces and trends
Examination of time in range as a correlate of subjective well-being in glucose sensor users with type 1 diabetes	Kira Wang, Caterina Florissi, Christianne Pang, Richard Wood	Members of the dQ&A European Patient Panel completed a biennial survey from April to May 2019; cross-sectional analysis included respondents with type 1 diabetes who reported CGM use (n=1,502). Participants self-reported estimated Time in Range; participants self-reported mental well-being using the WHO-5 Well-Being Index Analysis indicated a positive trend between increasing Time In Range and 4 of the 5 well-being metrics on the WHO-5 Index: cheerfulness/good spirits, relaxation, activity/vigor, restfulness.