February 22-25, 2023; Berlin, Germany (+Virtual); Day #3 Highlights – Draft

The sun was shining through the large glass walls of Berlin’s CityCube on ATTD Day #2. Stay close to our resource hub for the latest happenings in Berlin, and see our preview for what’s in store to come.

• ATTD 2023 Day #1 Highlights (link here): OP5 RWD from >30k US users, with T2D pivotal commencing in few weeks; PRO consensus meeting; transmitter-free real-time CGM from Senseonics; three planned launches for Abbott AID in 1H23; Chantal Mathieu on “the new face of diabetes”
• ATTD 2023 Day #2 Highlights (link here): Dr. Bergenstal on CGM in T2D without insulin; Dexcom/TypeZero’s fully closed loop AID algorithm; JDRF CKM symposium; positive national data from UK AID users; Dr. Rosenstock on SGLT-2s/GLP-1s for T2D remission

2022 Yearbook

Virtual clinics for diabetes care

• Dr. Satish Garg (Barbara Davis Center) kicked things off with an overview of virtual clinics, highlighting rapid adoption of telemedicine during the COVID-19 pandemic with widespread acceptance and ongoing value, stating that “telehealth in some way, shape, or form is here to stay.” We were pleased to hear Dr. Garg mention the strong potential for cost savings with virtual care, especially in terms of reducing DKA and hypoglycemia. From the patient side, he highlighted a survey conducted in 2020-2021, which found that patient satisfaction with telehealth visits was equivalent to or greater than that for in-person visits in both a pediatric population at Children’s Hospital Los Angeles and an adult population at UC Davis (Crossen et al. 2022). Furthermore, he argued telehealth programs can be effective in reaching vulnerable populations, such as low-income people and people from racial/ethnic minority backgrounds (Vaughan et al. 2022). Of course, virtual care is not without its challenges, and it will be important to address these going forward. Dr. Garg touched on issues such as compensation for providers and reimbursement (Giani et al. 2022), access to technology and digital literacy (Forde et al. 2022), and the need to individualize frequency and incorporate some in-person visits (Kirzhner et al. 2022). In particular, a small study of people with type 1 diabetes in Israel revealed that telehealth may not be able to completely replace traditional in-person visits for elderly people with type 1 diabetes (Kirzhner et al. 2022).

Continuous and intermittent glucose monitoring

• Dr. Klemen Dovč (University Medical Center Ljubljana, Slovenia) discussed three recently published studies demonstrating the value of CGM in a broad range of populations. First, Dr. Dovč highlighted the Autoimmunity Screening for Kids (ASK) study (Steck et al. 2022), which showed a use case for CGM metrics in predicting the progression of type 1 diabetes. Dr. Dovč suggested that this data may provide a reliable screening tool for autoantibody positive children to replace oral glucose tolerance tests. Dr. Dovč then discussed a study on the uptake and outcomes of the use of CGM in youth with type 1 diabetes following universal CGM coverage in New Zealand (Johnson et al. 2022). The study found that following universal coverage, CGM uptake increased from 5% to 79%, and resulted in an odds ratio of 2.5 for achieving an A1c <7.5%. Dr. Dovč said that this demonstrates the power of this technology and the need for increased access to CGM. To close out his presentation, Dr. Dovč highlighted the extension phase of the landmark MOBILE study (Aleppo et al. 2021). The study’s extension phase found that people with type 2 diabetes that continued using CGM (n=53) achieved a Time in Range improvement from 44% at baseline to 56% at eight months (+2.9 hours/day) and 57% at 14 months (+3.1 hours/day relative to baseline). Likewise, the group who continued CGM use saw a maintenance of A1c from eight to 14 months. Dr. Dovč concluded that each of these studies supports the continued expansion of CGM beyond just people with type 1 diabetes.

Insulin delivery hardware: Pumps and pen

• Dr. Rayhan Lal (Stanford) once again presented the insulin delivery technology section of this year’s ATTD yearbook, pausing first to acknowledge all the people with diabetes and diabetes advocates in the audience. Dr. Lal highlighted smart insulin pens, infusion sets, and insulin pumps for type 1 diabetes as well as type 2 diabetes, along with some notable finds on tech safety. On smart pens, Dr. Lal pointed to a post hoc analysis of NovoPen 6 and CGM in Sweden that showed study participants had a 22% chance of missing a basal dose over two weeks with clinical effect on glycemia (Ekberg et al., 2022). On insulin infusion sets, Dr. Lal presented his own group’s ten-day pilot study of convatec’s infusion set with hydrophilic coating to reduce inflammation, with medical survival of 9.1 days (Lal et al., 2021); a rodent model study showing foreign body reactions to catheters with only diluent infusion (He et al., 2021); a University of Graz study that found insulin infusions over 7 days led to 21-fold higher tissue flow resistance (Regittnig et al., 2022); a post-hoc analysis of Control-IQ pivotal trials which showed hyperglycemic failure over three-day wear was more common in youth than adults and more common with a 90o Teflon set vs. steel or angled Teflon (Kanapka et al., 2022); and Medtronic’s extended-wear infusion set pivotal trial that led to its FDA approval (Brazg et al., 2022). On insulin pumps in T1D, Dr. Lal highlighted a retrospective review that showed inequities in pump/CGM use as well as glycemia in non-White individuals (Fantasia et al., 2021); a Canadian observational study found insulin pump use was more common in regions where government paid for it but that the programs favored individuals of higher socioeconomic status (Song et al., 2021); and a retrospective study that assessed for predictors of A1c lowering in adults who started pump therapy (Neves et al., 2022). On insulin pumps in T2D, Dr. Lal talked about a narrative review concluding that pump features should be as easy as possible and customized for type 2 diabetes (Freckmann et al., 2021). Last, Dr. Lal emphasized the importance of safety, highlighting a retrospective study showing individuals on Omnipod over 1-3 years had lower rates of DKA and severe hypoglycemia compared to those on MDI (Biester et al., 2021), and a French retrospective study that found discontinuation of pumps was mainly associated with recent severe hypoglycemia (Gargoui et al., 2022).

New insulins, biosimilars and insulin therapy

• Dr. Thomas Pieber (Medical University of Graz, Austria) discussed clinical literature on advances in insulin therapy in the last year, 101 years since the discovery of insulin, with a particular focus using technology and the best insulins to minimize hypoglycemia. Although there were no new publications on Novo Nordisk’s once-weekly insulin icodec and Lilly’s once-weekly basal insulin-Fc (BIF) in 2022, Dr Pieber expressed excitement for publications in the upcoming years, as the phase 3 ONWARDS and QWINT programs on BIF continue to progress, indicated by the announcement of several encouraging topline results on icodec and BIF. Dr. Pieber highlighted three studies demonstrating the safety of insulin degludec (Novo Nordisk’s Tresiba). First, the HypoDeg trial, published in Diabetes, Obesity and Metabolism, found that insulin degludec conferred a 28% relative risk reduction of nocturnal symptomatic hypoglycemia and a 35% relative risk reduction of all-day severe hypoglycemia compared with insulin glargine U100 in people with type 1 diabetes. These results, as well as a significant reduction in nocturnal asymptomatic hypoglycemia, were further confirmed by a CGM sub-study of the trial, published in DT&T. Similarly, the SWITCH PRO study in insulin-treated people with type 2 diabetes found that insulin degludec conferred greater Time In Range, greater Time in Tight Range, and reduced nocturnal Time Below Range. Finally, the results from the EXPECT trial, published in The Lancet Diabetes & Endocrinology, have established the safety of insulin degludec in pregnant women with type 1 diabetes.

Closed-loop and artificial intelligence-based decision support systems

• Dr. Revital Nimri (Shneider Children's Medical Center of Israel) highlighted eleven publications on AID systems and three on decision support systems. Dr. Nimri said that while the eleven publications on AID systems fell into a few broad themes, her presentation today would focus on two of them. Dr. Nimri began by highlighting five papers on AID systems in a wide population, highlighting (i) two studies on the use of AID systems in pediatric populations (Ware et al. 2022; Abraham et al. 2021); (ii) a real-world study on AID initiation (Messer et al. 2021); and (iii) studies from Latin America and Qatar (Proietti et al. 2022; Petrovski et al. 2022). Dr. Nimri utilized this swath of clinical evidence to demonstrate the potential for AID beyond “tech savvy, well-controlled” populations. Dr. Nimri then turned to recent studies on the effect of SGLT-2s in combination with AID systems in people with type 1 diabetes (Garcia-Tirado et al. 2022; Haidar et al. 2022), explaining that although the initial data holds promise, more studies are needed in this field. Turning to decision support systems, Dr. Nimri shared two studies on the use of decision support systems in people with type 1 (Ferstad et al. 2021) and type 2 (Oser et al. 2022) diabetes. Dr. Nimri then overviewed one study (Bisio et al. 2022) on a CGM-based insulin dosing decision support system for people with type 1 diabetes on MDI. While the study did not find any significant differences in Time in Range for those utilizing the decision support system compared to control, participants utilizing the tool (n=57) did achieve a significant decrease in hypoglycemia at twelve-weeks (p=0.03).

Using digital health technology to prevent and treat diabetes

• After a wonderful Yearbook presentation in 2022, Dr. Mark Clements (Children’s Mercy Kansas City) took to the stage to discuss the quickly evolving digital heath arena. Broadly, Dr. Clements explained that the literature over the past year in this field fell into two themes: (i) scaling mobile and digital health applications to reach broader populations; and (ii) using digital tools to identify risk and stratify care delivery based on this risk (e.g., population health management). In his signature “to-the-point” style, Dr. Clements highlighted two pieces of literature reflecting these themes. The first study (Ross et al. 2022) came from the NHS’s digital health prevention program and included over 3,600 individuals total, around 75% of whom were adults with prediabetes and eligible for the study. Dr. Clements emphasized that the NHS’s program included five distinct interventions, and that these collectively were able to drive modest but significant outcomes in stabilizing prediabetes with regard to weight and A1c. The second paper (Prahalad et al. 2022) was Stanford’s famous 4T study, data from which were presented at ADA 2022. Dr. Clements explained how using RPM allowed Stanford researchers to develop a “Timely Intervention for Diabetes Excellence” (TIDE) tool, which takes CGM data from patients, analyzes the data, and presents results in a dashboard that makes it easy to identify patients that may need more or less help. In the 4T paper, the value of CGM initiation following type 1 diabetes diagnosis is reaffirmed by LOESS-smoothed curves of A1c. In closing, Dr. Clements said that he hopes to “see more on population health in the next year.”

Technology and pregnancy

• Similar to last year’s yearbook update, Dr. Jennifer Yamamoto (University of Manitoba) presented on diabetes technology and pregnancy. Of the ten articles addressing key issues in technology during pregnancy in the Yearbook, Dr. Yamamoto highlighted two. The first, a 15-year follow-up of the TODAY (Treatment Options for type 2 Diabetes in Adolescents and Youth) study, investigated pregnancy outcomes in people diagnosed with youth-onset type 2 diabetes. Dr. Yamamoto highlighted the diversity of this cohort, saying 42% were non-Hispanic Back and 35% were Hispanic. On average, participants were 20.5 years old, had a diabetes duration of 7.3 years, and an A1c of 8.9%. As well, Dr. Yamamoto highlighted that 62% of the cohort were from low-income households. Dr. Yamamoto then turned to results, sharing that 16% of participants reported pre-pregnancy counseling and only 15% reported using any method of contraception prior to first pregnancy. Dr. Yamamoto then highlighted that 65% of this cohort had maternal pregnancy complications and 25% had a miscarriage or stillbirth, showing that more work must be done to improve counseling and access to safe/effective contraception in young people with type 2 diabetes. Turning to type 1 diabetes, Dr. Yamamoto said the field gained further insight into CGM metrics during pregnancy this past year, presenting the retrospective analysis (Scott et al. 2022) of two CGM cohorts (CONCEPTT & a Swedish cohort study) that demonstrated variation in mean glucose, Time in Range, and Time Above Range throughout pregnancy for participants with and without an LGA (large for gestational age) infant. Pointing to this study, Dr. Yamamoto stressed the clinical importance of weekly CGM targets as well as a need for CGM early in pregnancy. In conclusion, Dr. Yamamoto said more work must be done to support people with diabetes who are at risk of pregnancy as well as those that are pregnant, and strategies need to be developed to ensure practice and policies are safe and accessible for all people living with diabetes.

Diabetes technology and therapy in the pediatric age group

• Stanford’s Dr. David Maahs offered his review of diabetes technology and therapy in children over the past year, highlighting the potential of hybrid closed loop to improve glycemic control in youth with type 1 diabetes as well as the continued challenges facing youth with type 2 diabetes. For the youngest children with type 1 diabetes (ages one to seven), a multinational trial found that hybrid closed loop led to a 9% increase in Time in Range (+2.2 hrs/day) and a 0.4% reduction in A1c (Ware et al. 2022). Previously, a multicenter trial of a tubeless AID system found a 16% increase in TIR (+3.7 hr/day) alongside a 0.7% reduction in A1c (Brown et al. 2021). Turning to type 2 diabetes, Dr. Maahs noted that, while the DPP-4 inhibitor sitagliptin was well-tolerated with a similar safety profile as adults (Shankar et al. 2022, Jalaludin et al. 2022), it did not significantly improve glycemic control. These discouraging results underscore the significant challenges faced by youth with type 2 diabetes – namely, aggressive manifestation of the disease with rapid beta-cell function decline and early onset and accumulation of complications. At a bigger picture level, 10-year follow-up data from the SWEET Registry found reductions in A1c from 8.4% to 7.9%, reductions in severe hypoglycemia from 3.8% to 2.4%, and increased use of diabetes technology in youth with type 1 diabetes (Gerhardsson et al. 2022). Lastly, Dr. Maahs highlighted Australia as a trailblazer in advocacy for access to CGM for all youth with type 1 diabetes, with the country’s universal subsidized CGM program leading to a two-fold increase in the odds ratio for attaining target A1c (<7%), a three-fold decrease in the odds ratio for A1c >9%, and a two-fold reduction in the incidence of DKA (Johnson et al. 2022).

Advances in exercise and nutrition as therapy in diabetes

• This year’s remarks from Dr. Dessi Zaharieva (Stanford) around exercise and nutrition as therapy in diabetes focused on one nutrition study on modeling personal glycemic responses to food to aid in optimal insulin administration and one exercise study. The exercise study, published in Diabetes Care, was a randomized crossover trial comparing glucose control and potential additional signals during moderate intensity, high intensity, and resistance exercise with the MiniMed 670G hybrid closed loop insulin delivery system among 30 adults with type 1 diabetes. When following exercise consensus guidelines (e.g., setting a higher exercise target 2h before activity), the study found that hybrid closed loop was effective, with no difference in Time in Range (median of 81%) and 0% Time Below Range across exercise types. However, a recent publication (not included in the current yearbook chapter) from the Jaeb Center’s Type 1 Diabetes Exercise Initiative (T1-DEXI) published in Diabetes Care, was also highlighted, as it showed significant increases in Time Below Range with exercise versus sedentary days, despite clinically meaningful improvements in Time in Range in real-world settings. Reflecting on these results, Dr. Zaharieva concluded that while Time in Range is often excellent in controlled hybrid closed-loop exercise studies, real-world data suggest that challenges still exist around glycemic management and exercise in people with type 1 diabetes – we are curious to know what it would take to ultimately bridge this gap and to create new exercise guidelines for type 1 diabetes such that the risk of hypoglycemia is minimized.

Primary care and diabetes technologies and treatments

• Dr. Gregg Simonson (International Diabetes Center) presented on the primary care section of this year’s ATTD Yearbook, highlighting two themes: (i) the need to continue CGM use and view it as a long-term “therapy”; and (ii) the importance of optimizing the ABCs (A1c, blood pressure, and cholesterol). In an analysis of the MOBILE trial in adults with type 2 diabetes on basal insulin (Aleppo et al. 2022), participants who discontinued CGM lost about half of their initial Time in Range gain, largely due to an 11% increase in Time Above Range from 9% to 20% (+2.6 hr/day) in those discontinuing CGM. Ultimately, Dr. Simonson urged providers to view CGM not just as CGM, but as “CGM therapy” to emphasize the importance of maintaining this intervention to improve glycemia. Turning to the ABCS (A1c, blood pressure, cholesterol), Dr. Simonson emphasized the potential gains in life years ranging from one year of life with LDL optimization to nearly four years of life with attainment of target A1c in an analytical model of 421 adults with type 2 diabetes (Kianmehr et al. 2022). Considering all of these risk factors in a combined “risk calculator” reveals considerable differences in life expectancy for people meeting all three targets compared to those not meeting any targets – a difference of 10 years for the example 50-year-old female patient Dr. Simonson described. Ultimately, he argued that this data can give primary care providers the tools they need to talk to patients about the importance of attaining A1c targets, reducing blood pressure, and taking statins.

Real-world diabetes technology: Overcoming barriers & disparities

• Dr. Stuart Weinzimer of Yale presented on the real-world data section of this year’s ATTD Yearbook, which was renamed from “Real-world diabetes technology: What do we have? Who are we missing?” in last year’s update. Of the 300 articles published between July 2021 and June 2022, Dr. Weinzimer said 17 were included in this year’s Yearbook chapter and fell within three themes: (i) real-world outcomes across populations and regions; (ii) expanding diabetes technology access to novel populations; and (iii) methodological approaches to address diabetes technology disparities. Dr. Weinzimer presented one study from the last two sections. For expanding technology to novel populations, Dr. Weinzimer highlighted a retrospective analysis of CGM in people with type 2 diabetes on non-intensive treatment regimens (Miller et al., 2021). Presenting the study’s baseline characteristics, Dr. Weinzimer highlighted that 40% of the cohort were insulin users, and 60% of participants were on non-insulin therapy. Turning to results, Dr. Weinzimer highlighted that six months post-CGM start, participants had significantly fewer acute diabetes events and all-cause hospitalizations. For methodological approaches to address diabetes technology disparities, Dr. Weinzimer pointed to a study evaluating the Australian government’s policy change to fully fund CGM for youth with type 1 diabetes (Johnson et al. 2022). Dr. Weinzimer highlighted how only 5% of Australian youth were using CGM before the policy change, whereas afterward, ~79% of youth are now on CGM, illustrating that universal coverage dramatically increased CGM usage. Dr. Weinzimer also showed a graph of mean A1cs demonstrating that those with >75% use of CGM had lower A1cs than youth not on CGM or using CGM <25% of the time. In conclusion, Dr. Weinzimer said that ‘real world’ evidence should indeed include the real world and that increasing access to diabetes technologies is not only ethically imperative but also has societal and financial benefits.

Diabetes technology and the human factor

• Dr. Alon Liberman once again presented the ATTD Yearbook chapter on “Diabetes Technologies and the Human Factor,” reminding attendees that patients still play an “essential role” in the successful use of diabetes technology. This year, Dr. Liberman focused on the role of CGM in reducing diabetes distress and overall outcomes for children with diabetes and their caregivers. Dr. Liberman began by overviewing the TACKLE-T1D study (Youngkin et al. 2021), which found that while CGM can decrease hypoglycemia avoidance behaviors, it has little impact on caregiver’s hypoglycemia worry, mirroring a systematic review (Zhang et al. 2022) also presented by Dr. Liberman. Dr. Liberman then highlighted data from a study (Verbeeten et al. 2021) which found that parents and children had decreased worry scores after twelve months of CGM use compared to baseline (p<0.001). Dr. Liberman finished his presentation by overviewing a recent publication (Polonsky et al. 2022) which examined the impact of real-time CGM data on caregivers. The trial found that while 88% of parents found data sharing gave them “peace of mind,” the degree of benefit was influenced by the collaborative actions taken by caregivers to support their child's or partner's diabetes management. Dr. Liberman concluded by saying that due to the increasing use of diabetes technology, there must be more attention paid to the impact of this technology on diabetes distress for both people with diabetes and their caregivers.

Immune interventions and restorative therapies in type 1 diabetes

• Dr. Desmond Schatz (University of Florida) discussed type 1 diabetes prevention and cures, a topic that was expanded in this year’s Yearbook to include prevention following the approval of Tzield (teplizumab) in November 2022. Dr. Schatz honed in on two studies, one assessing low dose anti-human thymocyte globulin (MELD-ATG) and the other investigating calcium channel blocker verapamil. Following positive safety and efficacy two-year results for low-dose ATG (presented at Levine Riggs Symposium 2019), Dr. Schatz highlighted an ongoing RCT to determine the minimum effective dose by investigating changes in stimulated C-peptide responses (n=114). Dr. Schatz also touched on an exploratory study of verapamil, which demonstrated improvements in a mixed-meal tolerance test and a good safety profile; additional pediatrics data is being presented by Dr. Gregory Forlenza later on Friday at ATTD 2023. Turning to the ever-important question of biomarkers, Dr. Schatz noted that chromogranin A, a T1D autoantigen, was found to be a promising biomarker, as it was inversely proportional to C-peptide levels and normalized following verapamil administration. Of course, Dr. Schatz would have been in remiss not to highlight the tremendous advancements in immune protection of stem-cell derived islet cell therapy for type 1 diabetes in recent years and months (Tahbaz et al. 2021) – what a powerful way to close!

New medications for the treatment of diabetes

• Dr. Irl Hirsch (University of Washington) outlined recent advancements in diabetes therapies, primarily centering around SGLT-2s and GLP-1s. He noted that this past year has resulted in less information about new medications, but more new information about drugs already approved. This year’s chapter consists of 15 abstracts split across three “buckets”: (i) emerging adjuvant treatment for the management of diabetes; (ii) dual or triple incretins; and (iii) drugs focusing on obesity management. Dr. Hirsch highlighted a meta-analysis of nine studies (n=81,206) published in Diabetes Research and Clinical Practice showing that SGLT-2s may be superior to GLP-1s in renal outcomes, while there was no significant difference between the two drug classes in the risk of major adverse cardiovascular events, in people with type 2 diabetes with or without albuminuria. Dr. Hirsch also noted a new potential area of benefit with SGLT-2s, as a data base assessment of 16,566 and 2,746 people treated with SGLT-2s and GLP-1s respectively found a 27% relative risk reduction in new-onset atrial fibrillation with SGLT-2s vs GLP-1s. Another analysis found similar results, as SGLT-2s were associated with a significantly lower risk of incident atrial fibrillation, while DPP-4s and GLP-1s were not, in people with type 2 diabetes. On incretins, Dr. Hirsch reviewed a meta-analysis of seven trials (n=6,609) on tirzepatide published in Diabetologia showing tirzepatide’s dose-dependent superiority on glycemic and body weight outcomes compared with GLP-1s and basal insulin. We were also pleased to see Dr. Hirsch highlight a meta-analysis of GLP-1s among children and adolescents with obesity published in the Journal of Pediatrics, showing that GLP-1s are safe and effective in reducing body weight, BMI, A1c, and systolic blood pressure in youth. Following Wegovy’s approval in teens in January 2023 there are now five FDA-approved drugs for adolescents with obesity – this is night and day from as recently as a year ago! We hope to see these developments reflected in guidelines to encourage the use of these anti-obesity medications in youth, given worrying rises in childhood obesity seen in the past decade.

Obesity and diabetes

Dr. Viral Shah (Barbara Davis Center) closed this year’s Yearbook session with an overview of obesity and diabetes in 2022, shining light on the growing prevalence of obesity in traditionally overlooked populations, as well as the tremendous innovation in incretin therapies with especially promising benefits in NAFLD. While many studies have documented high obesity prevalence in developed countries, Dr. Shah drew attention to the rising rates of obesity and overweight in lower and middle-income countries (Teufel et al. 2021). In a similar vein, there is growing recognition of the need to treat obesity in type 1 diabetes (Van der Schueren et al. 2021), and research now suggests a causal risk for high childhood body size and diagnosis with type 1 diabetes (Richardson et al. 2022). Turning to incretin therapies, Dr. Shah highlighted that the benefits of semaglutide extend beyond obesity to NAFLD (Rubino et al. 2022, Newsome et al. 2021), and also touched on the potential of SGLT-2s in NAFLD (Phrueksotsai et al. 2021). These therapies will be especially relevant following AACE’s Clinical Practice Guideline recommending that all people with obesity, metabolic syndrome, prediabetes, and diabetes should be screened for NAFLD (Cusi et al. 2022).

AID Highlights

1. ADAPT study 12-month results (n=82) support use of AID earlier in treatment: switching directly from MDI to MiniMed 780G results in 1.4% A1c advantage, +6.5 hr/day TIR improvement

During a set of rapid-fire oral presentations, Dr. Ohad Cohen (Medtronic) presented full twelve-month results from the ADAPT study, demonstrating significant benefits from MiniMed 780G over MDI and further developing the evidence base for using more “advanced” technologies earlier. The six-month results from the randomized, prospective study were read out at ATTD last year and were published in The Lancet Diabetes & Endocrinology in late 2022. Today’s twelve-month results confirmed those initial findings. Underscoring the excitement around the results, Medtronic issued a press release this morning in conjunction with Dr. Cohen’s presentation.

As a reminder, the fourteen-center, multinational ADAPT study was designed to evaluate safety and effectiveness of MiniMed 780G in people with type 1 diabetes transitioning directly from MDI. Of course, this goes against the traditional paradigm of transitioning people with diabetes from injections to open-loop pump to hybrid closed loop therapy. Importantly, the study enrolled people not meeting glycemic targets, as all participants had an A1c ≥8%, and all participants were using FreeStyle Libre at baseline. Following a 14-day run-in, 82 subjects were randomized to six months of MiniMed 780G or control (MDI + FreeStyle Libre). After six months, the control group switched over to MiniMed 780G and the study group stayed on MiniMed 780G for the remaining six months of the study (i.e., all participants used MiniMed 780G).

• Of the 82 subjects randomized, 67 remained through the full twelve-month study. The MDI and MiniMed 780G groups were well-matched at baseline with an average age around 40 years and average diabetes duration around 19 years. Mean A1c was 9.1% for the control and 9.0% for the study group and both groups were averaging nine FreeStyle Libre scans per day at baseline.
• After switching from MDI to MiniMed 780G at the six-month point of the trial, participants saw mean A1c improvement of 1.4%. The improvement seen by the crossover group in the last six months exactly matches the improvement reported in the initial six-month results by the MDI-to-MiniMed 780G switchers. For the study group who switched to MiniMed 780G during initial randomization, there was no change in A1c from months 6 to 12, indicating that the benefits from MiniMed 780G are sustained.

• Time in Range results tell a similar story – subjects who switched from MDI to MiniMed 780G saw a Time in Range improvement of 6.5 hours/day. After switching from MDI to MiniMed 780G, Time in Range improved from 44% to 71%, just meeting the consensus target of >70% Time in Range. Time in hyperglycemia was significantly reduced from 54% to 28%, or -6.3 hours/day. Time in hypoglycemia was only slightly lower with MiniMed 780G at 2.6% vs. 1.8%, though this result was statistically significant for superiority.
  
  • For the group that switched to MiniMed 780G at randomization, the Time in Range improvements seen at six months were maintained out to twelve months. Those participants saw Time in Range improve a drastic 6.6 hours/day (36% to 71%) during the initial six-month period and at the end of twelve months, Time in Range was maintained at 70%.
• In his conclusion, Dr. Cohen encouraged providers to consider moving toward AID systems earlier in diabetes management. Particularly as CGM and insulin pumps become easier and more usable (for example, no fingersticks and longer wear duration for CGMs, as well as smartphone control and more automation for insulin pumps), we may see growing uptake of “advanced” technologies like AID utilized earlier after diagnosis. (This train of thought aligned with the conversations held outside of this oral session during Medtronic’s sponsored symposium.) Furthermore, to that end, the first results from the JDRF-sponsored CLVer study were also read out this afternoon. That study placed newly diagnosed youth with type 1 diabetes directly onto an AID system to see if the early intensive management may help preserve beta cell function – read on for those results.

2. Meta-analysis of three Control-IQ pivotals (n=369; ages 2-72) shows remarkably strong performance across all ages (2.8 hour/day TIR gain on AID vs. control); RCTs align with real-world data

The famed Dr. Roy Beck (Jaeb Center) presented a meta-analysis of the three Control-IQ pivotal trials during an early-morning Tandem symposium. Control-IQ is the only AID system with FDA approval based on the results of randomized controlled trials: one in adults (14-72 years), one in children (6-13 years), and one in preschoolers (2-<6 years). Outside the US, CamDiab’s CamAPS FX algorithm received CE-Marking based on RCTs as well (in 2020). In this meta-analysis, Dr. Roy Beck explained that outcomes from these trials were pooled (n=369; 76% on pumps at baseline; 24% on MDI at baseline). The pooled cohort had a mean baseline A1c of 7.5% (33% with A1cs ≥8%) and a mean Time in Range of 57%. The table below, which we published when the Control-IQ PEDAP preschool pivotal was presented at ISPAD 2022, summarizes the outcomes from the three Control-IQ pivotals.

• Across the entire group of Control-IQ’ers (ages 2-72), Time in Range improved 3.1 hours/day to 70% and A1c fell ~0.5% to 7.0%. Comparatively, Time in Range hovered at 56%-57% for the control group and A1c also stayed the same at 7.5%-7.6%. Dr. Beck stressed that time in closed loop remained high for Control-IQ users (median 93%) and that 52% of AID users met CGM targets during follow-up (vs. 20% in control; p<0.001). Treatment group differences for CGM metrics are shown below – the difference in time >250 mg/dL is particularly impressive, as is of course the difference in Time in Range.

• As is usual with AID systems, Control-IQ was particularly effective at nighttime control across a broad age group. Dr. Beck highlighted the consistently good early-morning glucose control specifically, seen in the figure below by the narrowing of the interquartile range on AID. The treatment group difference in Time in Range was 21% or five hours (!) at nighttime, and 8% or two hours during the day.

• Moreover, the pooled data illustrate just how quickly Control-IQ can confer glycemic benefits: think days, not weeks. Dr. Beck presented the figure below to demonstrate that Time in Range improvements are seen with Control-IQ just after the first day of use.

• Dr. Beck then dove into subgroup analyses that reinforced how AID can help a broad range of participants. Stratified by age, socioeconomic status, baseline insulin delivery modality (pump or MDI), and baseline A1c, Dr. Beck explained that every cohort achieved a remarkable improvement in Time in Range and Time Below Range. Every group also saw improvements in A1c, except for the cohort with an A1c <7% at baseline. Dr. Beck noted that it’s “interesting” how there was no significant overall improvement in A1c for this cohort after starting Control-IQ, but there still was a Time in Range improvement favoring Control-IQ vs. control. The image below is slightly hard to digest, but a quick glance shows that the dots always land on the “favors Control-IQ” side as opposed to the “favors control” side, with the exception of the one cohort mentioned above – that dot lands directly in the middle!

• Encouragingly, the symposium underscored how real-world outcomes from Control-IQ users align with those seen in the meta-analysis. Dr. Beck cited one-year real-world outcomes in DT&T from Breton et al. 2021, as well as a massive dataset from Kovatchev et al. 2022 in Diabetes Care. Later in the session, Dr. Nick Oliver (Imperial College Healthcare NHS Trust, UK) presented pediatric and adult outcomes from real-world Control-IQ users at his institution (n=55; mean age 38 years). Baseline A1c was 7.4%. As shown in the table below, the cohort, which was predominantly composed of people with problematic hypoglycemia or people who had previously struggled to meet glycemic targets despite therapy intensification, still managed to hugely benefit from AID. Most participants, according to Dr. Oliver, came from sensor-augmented pump therapy or Basal-IQ at baseline. In a subgroup of individuals over 18 years old (n=35), Time in Range increased 2.3 hours/day to 70%, driven by a reduction in time >180 mg/dL. As Dr. Oliver so eloquently stated, people don’t need to be “embedded within a highly resourced RCT” to achieve these kinds of outcomes with Control-IQ.

• Dr. Partha Kar (NHS England, UK) also provided a salient update: CGM penetration (is-CGM and rt-CGM) has reached 83% of type 1s in the UK. Dr. Kar also overviewed the NHS audit data from AID users shared during Day #2 of ATTD. NICE’s most recent draft recommendations (published January 2023) include reimbursement for hybrid closed loop for all people with type 1 diabetes with an A1c >8%, importantly including pregnant people with diabetes.

3. MiniMed 780G real-world data (n=61,481): Strong TIR of 73% (71% in pediatrics and 74% in adults); Mr. David Dunleavy expands on next steps for Medtronic AID

During Medtronic’s jam-packed symposium, Dr. Ohad Cohen (Medtronic Diabetes) presented real-world outcomes from individuals using MiniMed 780G (n=61,481). This sample size is 50% larger than the most recent large real-world outcomes update we saw from Medtronic at ISPAD 2022 in October 2022, and nearly 1.5-times the size of the sample shared one year ago at ATTD 2022. The strong YOY growth in MiniMed 780G users and the availability of the system in 90+ countries as of Medtronic 4Q22, we imagine, enables the team at Medtronic to pull more and more profiles to add to this ever-growing dataset.

This updated impressive cohort includes longitudinal outcomes from participants who uploaded their data voluntarily to CareLink between August 2020 and December 2022. Notably, the entire cohort boasts a Time in Range of 73% and a GMI of 6.9%, both of which meet consensus targets for a Time in Range >70% and an A1c <7%. Stratified by age, participants ≤15 years old (n=14,577) saw a Time in Range of 71% on MiniMed 780G and a GMI of 6.9%, vs. a Time in Range of 74% and a GMI of 6.9% in adults (n=42,781). As we wrote yesterday, 64% of users had a GMI <7%, 65% had a Time in Range >70%, 87% had a Time Below Range of <4%, and 50% achieved all three of these metrics.

• In a subset of users who opted for the most aggressive system settings (n=4,392), Time in Range reached 79%. Recall that MiniMed 780G allows users to customize their glucose target and active insulin time. For those choosing the most aggressive combination (target of 100 mg/dL and active insulin time of 2 hours), outcomes are shown below stratified by age. This data closely mirrors the topline takeaway from the Omnipod 5 real-world dataset presented on Day #1: glycemic outcomes seem to improve with increasing aggressivity, at least as far as the real-world is concerned. While  there are a variety of reasons why people with diabetes might opt for less aggressive settings, and while we appreciate Insulet and Medtronic for providing customizable targets to enable people with diabetes to adapt their AID systems to their own liking, data at this meeting certainly has implied that there may be more reasons than PWD may realize to think a bit more in depth about their targets.

• Closing out the session, Mr. David Dunleavy (VP of Global Portfolio, Medtronic Diabetes) highlighted the future direction for the company’s AID franchise. We were impressed how Mr. Dunleavy stressed that while the real-world data presented by Dr. Cohen are definitely impressive, there is so much more that can be done for AID users. While conversation abounds on how to further increase Time in Range into the 80%-90% range for existing AID users, it was music to our ears to hear Mr. Dunleavy envision how Medtronic will focus on making it easier to reach a 70% Time in Range – think about the opportunity for market expansion. As one such example, Mr. Dunleavy noted that 100 years of insulin has meant 100 years of waiting to decouple the administration of insulin with carb counting. As such, Mr. Dunleavy explained that the “next big step” for Medtronic will be an AID system with an automated food bolus, that fully closes the loop – the company’s next-gen Simplera CGM was pictured as a part of this system. As Mr. Dunleavy was wise to point out, reducing therapy complexity requires innovation beyond algorithmic innovation, and thus Medtronic is committed to expanding closed loop technologies to new pumping systems and its “MDIS platform” – we assume this was about InPen.

• On Mr. Dunleavy's closing slide, some points of note: the company’s Simplera CGM (submitted for FDA approval and CE-Mark) was pictured alongside InPen (which included the company’s Guardian 4 Sensor), the seven-day Extended Infusion Set (launched in 19 countries), and also a picture of an Apple Watch and smartphone app.

4. Dr. Irl Hirsch presents DERMIS results (n=30) revealing increased inflammation at pump infusion sites, which correlated with Time in Range; “deep itching” from sets likely due to eosinophils

During a packed morning session on dermatology in diabetes, Dr. Irl Hirsch (University of Washington) presented data from the DEfining, Reviewing, and Monitoring skIn pathologieS (DERMIS) study in type 1 diabetes. The study sought to characterize skin changes in insulin pump infusion sites among people with type 1 diabetes using traditional histopathology and novel non-invasive skin imaging methods. The study involved taking skin samples from three sites: (i) a normal skin site; (ii) an active pump site; and (iii) a recent (72 hours) pump site. Study participants (n=30; mean age 48 years) had an average diabetes duration of 30 years and an average A1c of 6.6%. In terms of the participants’ technology use, ten were on a Medtronic pump, five were on an Insulet pump, and 15 were on a Tandem pump. For those on traditional insulin pumps, two utilized steel infusion sets and 23 utilized Teflon infusion sets. 

• The study found a significant difference between inflammation levels at control infusion sites compared to current/recovering infusion sites. Dr. Hirsch shared that the study utilized a novel, non-invasive method of investigating skin changes that occur from long-term insulin pump use. The method, called optical coherence tomography, revealed that at current/recovering infusion sites, there was (i) significantly increased mean blood vessel density; (ii) significantly higher mean optical attenuation coefficient, which is indicative of increased inflammation; and (iii) increased mean epidermal thickness compared to control (p<0.05 for all).
• The study also examined common inflammatory markers (e.g., fibrinogen, IGF1), which revealed significantly higher levels of all markers at current/recovering sites compared to control sites (p=0.001). Dr. Hirsch shared that these results demonstrate the inflammatory impact of pump infusion sites on the skin, a phenomenon that can result in pump discontinuation. Given the rising use of infusion sets driven by increasing insulin pump penetration, Dr. Hirsch said an understanding of the underlying causes of this problem is critical to find solutions to help people with diabetes. 
• Interestingly, inflammation was found to be positively associated with insulin dose and negatively associated with Time in Range. Dr. Hirsch noted that as participants’ average daily insulin dose increased, so did skin inflammation score (p=0.009). The study also looked at the association between Time in Range and inflammation scores, finding that as Time in Range increased, inflammation scores decreased (p=0.01). The association between increased Time in Range and improved inflammation was fascinating to see – we wonder the mechanisms behind this trend and we believe this will be a point of much discussion moving forward. Dr. Hirsch also shared that all participants experienced skin symptoms throughout the study, with 93% experiencing “deep” skin itching and 63% experiencing skin pain and flat redness on the skin.
• Dr. Hirsch explained that there was no difference in dermatologic characteristics in long-term vs. short-term pump users. Dr. Hirsch shared that the study was unable to elucidate any differences in inflammatory markers in those who used a pump for >20 years compared to those who had used a pump for <10 years. Dr. Hirsch said that this meant the researchers were unable to determine why some people with diabetes have a higher propensity to skin issues related to infusion sets, while others do not. Dr. Hirsch said that the DERMIS study team will be conducting more studies to continue to better understand this issue.
• Dr. Hirsch shared that the most surprising finding of the study was the high levels of inflammatory eosinophils in skin samples. As background, eosinophils are a type of inflammatory cell that are commonly activated in allergic reactions. The study found that eosinophilic inflammation was significantly more common at current infusion set sites (73% of sites; p<0.01) and recently used infusion sites (75% of sites; p<0.01) compared to control sites. Concerningly, the inflammatory cells were located deep in the dermis, which Dr. Hirsch suggested as a reason for the “deep” itching experienced by 90% of study participants. Dr. Hirsch also said that this result suggests that the issue may be an allergen delivered at the tip of the catheter, which he speculated could be a preservative in the insulin, plastics in the infusion sets, or glue in various components.

5. Diabeloop AID system with no meal announcements for adolescent type 1s, enrollment to complete in March; RWD from Diabeloop’s DBLG1 shows TIR of 69%-72% across five EU countries (n=6,658)

Diabeloop announced plans for a study in Europe for its next-gen AID feature, which is intended to make meal announcements optional. The clinical study is expected to enroll 50 people with type 1 diabetes, ages 12 to 18 years old, across hospitals in France, Belgium, and Germany. Impressively, the study is expected to complete enrollment by “the end of March.” The study will evaluate Diabeloop’s “Unannounced Meal Management” module, which is designed to detect meals whenever a glucose spike is detected. Combined with Diabeloop’s “auto-learning” capabilities for personalizing insulin-to-carb ratios and insulin sensitivity factor over time, Diabeloop believes it can make meal announcements optional without significantly sacrificing post-prandial glycemic control.

• Diabeloop is initially testing the feature in adolescents with diabetes, as this age group generally struggles with glycemic management compared to other age groups. As Diabeloop notes in its press announcement about the study, just 17% of adolescents are achieving an A1c <7.5% and that missed meal boluses may be a very significant part of this. As access and availability of AID systems grows, it’s been encouraging to see manufacturers continue to strive for better. Just yesterday at ATTD, Medtronic and Dexcom/TypeZero both shared details on their work towards fully closed loop and eliminating meal announcements.
• In a separate presentation, Diabeloop shared real-world data from 6,658 users of its commercially available DBLG-1 algorithm. The users came from across five European countries (Germany, Spain, Italy, Switzerland, and the Netherlands). As a reminder, the DBLG1 system currently consists of Diabeloop’s DBLG1 algorithm, Dexcom G6, and Roche Accu-Chek Insight. Across these five countries, real-world Time in Range was fairly consistent, with Italy on the low end by achieving a Time in Range of 69% and Spain on the high end by achieving a Time in Range of 72%. Time below range was also 1.2% or lower across all countries. Using A1c, Dr. Alice Adenis (Diabeloop) also tried to estimate the change in Time in Range before the participants were on DBLG1 using Dr. Roy Beck’s model published in 2019. Using this estimated Time in Range prior to DBLG1 and actual Time in Range after DBLG1, Time in Range improved by 17% or 4.1 hours/day.

6. *NEW* Real-world Control-IQ in 936 people with type 2 diabetes in the US: Time in Range improves 1.9 hrs/day to 71% following transition from Basal-IQ after three months

During a morning session on AID, we got our largest look to-date at real-world data from people with type 2 diabetes using Tandem’s Control-IQ hybrid closed loop system. Dr. Chiara Fabris (University of Virginia) read out real-world results for 936 people with type 2 diabetes in Tandem’s database who had at least one month of data using Basal-IQ and three months of data using Control-IQ.

• Three months after switching from Basal-IQ to Control-IQ, mean Time in Range increased from 63% to 71%. This translates to an additional two hours/day in Range. Not surprisingly, the quartile with the lowest Time in Range at baseline saw the biggest improvement, going from 31% to 53% after initiating Control-IQ (+5.3 hours/day). Similarly, time >180 mg/dL fell significantly from 36% to 32% (-1 hour/day), with the lowest quartile group reducing time >180 mg/dL from 69% to 47% (-5.3 hours/day). Time in hypoglycemia did not appear to significantly change between Basal-IQ and Control-IQ use periods.
• Mean glucose was reduced from 170 mg/dL to 159 mg/dL with use of Control-IQ. The best quartile at baseline saw no significant change with a mean glucose of 136 mg/dL at baseline and a 139 mg/dL mean glucose after three months of Control-IQ use. Both of these values translate to a GMI of 6.6%. However, the lowest quartile at baseline saw a large reduction in mean glucose from 221 mg/dL to 186 mg/dL. This translates to GMI values of 8.6% and 7.8%, respectively.

• During Tandem’s 4Q22 update, we learned that the company is “preparing to start a pivotal” to expand Control-IQ’s indication to include people with type 2 diabetes. A feasibility study in this population has already been presented, and in 1Q22, we heard an update that Tandem already has 20,000 people with type 2 diabetes using Control-IQ off-label.

7. *NEW* Diabetes and dermatology: Dr. Laurel Messer overviews common skin reactions caused by diabetes technologies and gives practical pearls for treatment and prevention

Dr. Laurel Messer (Senior Director, Translational Research & Applied Science, Tandem) highlighted tips and tricks to preserve skin integrity for the long-term use of diabetes technology. Dr. Messer began by overviewing data from an international survey of healthcare providers, conducted by ISPAD’s JENIOUS group. The survey found that while only 50% of HCPs are aware of harmful allergens in adhesives found in diabetes technology, 98% of respondents felt that practical knowledge of skin reactions should be a part of mandatory training. Dr. Messer then said that since 2019, there have been 54 studies on CGM-related skin reactions, but the studies are not consistent with respect to how reactions are reported, leading to a lack of knowledge on the real toll of adverse skin events – and that’s to say nothing of pump-related skin reactions. Dr. Messer then dove into some practical knowledge on the subject:

• Dr. Messer overviewed common skin reactions and how to distinguish among them. She said that contact irritation, or irritant contact dermatitis, is the most common type of skin reaction and is a non-immune inflammatory response caused by direct contact with chemical or mechanical irritants. Conversely, Dr. Messer said, allergic contact dermatitis is an immune response that results in a permanent sensitization to certain adhesives. She said that due to the similar presentation of irritant contact and allergic contact dermatitis, it can be difficult to distinguish between the two.
  
  • This issue is further complicated by a lack of diagnostic tools for these reactions, as an allergy can only be diagnosed with an intensive and time-consuming patch test. However, Dr. Messer said that approximately 80% of the time, a reaction is a presentation of irritant contact dermatitis. She said that while irritant reactions are associated with burning, allergic reactions typically can be distinguished by intense itching and papules (raised cyst-like spots on skin). Dr. Messer said that unfortunately, avoidance of the allergen is typically the only option for allergic contact dermatitis, meaning that discontinuation of the device is most likely necessary. However, she suggested trying other devices that may use different adhesives, such as Senseonics' Eversense implantable CGM (silicone-based adhesive), or ones that do not use adhesives altogether.
• Turning to treatment and prevention, Dr. Messer said that reminding device users of the basics can be the most impactful way to prevent skin reactions. She urged providers to remind device users to (i) select a site where the skin does not fold; (ii) pinch the skin to make sure there is enough available tissue; (iii) wash the area with soap to remove oils in skin; and (iv) dry the area of all prepping material. Dr. Messer encouraged trial and error with the process, sharing that many of her patients use different methods; for example, she said that some users find skipping the alcohol wipe during the insertion of their CGM to be helpful.
  
  • Dr. Messer then said that preventing reactions is also possible through liquid and physical barriers. That said, she also warned that with some of these liquid barriers (e.g., DuoERM and Matisol), there may be exposure to more chemicals, which may increase the risk of new hypersensitivity. Dr. Messer said that healthcare providers and users should weigh these options as they decide the best solution for skin irritation.
    
    • Last, Dr. Messer said that careful removal of a diabetes device, and taking care of skin, is key to healing the skin. She suggested using removal agents such as TacAway and Detachol, along with using a “slow, low angled” tape removal strategy that doesn’t further irritate the skin. Dr. Messer also championed using several different infusion/CGM sites, giving skin time to heal. She also said that to prompt faster healing, she recommends the use of hydrocortisone and skin emollients between site usage. Dr. Messer concluded by saying that healthcare providers have a responsibility to be aware of off-label use of topical steroids and other “hacks” for skin prep, even if they themselves do not agree with their use, as it is necessary to be able to provide support and guidance for the choices patients make.

8. *NEW* Potpourri of MiniMed 780G studies on meal management and exercise reflect Medtronic’s goal to simplify reaching 70% Time in Range on AID

Medtronic’s Day#3 symposium at ATTD 2023 focused on simplifying meal management and exercise with the MiniMed 780G AID system. The presentation of these four studies aligned nicely with Mr. David Dunleavy's (VP of Global Portfolio, Medtronic Diabetes) presentation elsewhere during the symposium, during which he stressed that Medtronic’s AID pipeline will not focus on breaking the mid-70% Time in Range barrier that most commercial AID users experience (including MiniMed 780G users, per real-world data). Instead, Mr. Dunleavy's explained that Medtronic make it easier to achieve a Time in Range over 70%, enabling more people with diabetes to benefit from AID and meet consensus targets. While Medtronic’s late-stage pipeline will tackle this lofty goal through a fully closed loop AID system that minimizes user interaction, the potpourri of MiniMed 780G studies presented during this symposium reflect the aspirations of the near-term AID pipeline for Medtronic: simplifying meal management and exercise, which the company coined as the “final frontier” of hybrid closed loop therapy in type 1 diabetes.

• The first study evaluated MiniMed 780G’s performance when meals are unannounced, mirroring data we’ve seen from Omnipod 5 and Control-IQ. As a huge caveat, faculty stressed that MiniMed 780G is designed to be used with full meal announcements. That said, given that HCPs might tend to view certain individuals as non-ideal candidates for AID (e.g., inconsistent carb counting, struggle with tech, new to pumps), the presenters stressed that these data are important to show that preconceived notions around a patient’s bolus adherence should not preclude them from being prescribed AID. The open-label, single-arm study enrolled 14 adults with type 1 diabetes on sensor-augmented pump therapy at baseline with an A1c ≤10%. Even with no boluses, Time in Range was 68%, improving 2.4 hours/day to 78% for those announcing meals. Meals with carbs ≤80g made no difference in safety between the two arms, and for meals ≤20g carbs, there was no clinically significant difference in Time in Range.

• The second study, an RCT (n=34) already presented at ISPAD 2022, evaluated a meal management simplification scheme in adolescent type 1s using MiniMed 780G. Participants were randomized to either a “flex” meal management arm (traditional carb counting), or a “fixed” meal management arm in which participants could report either a “regular meal,” a “large meal,” or a “snack” at mealtime (like the insulin-only iLet). The fix group achieved an impressive +6.2 hour/day Time in Range improvement from 48% to 74% after three months (p<0.001), entirely from a reduction in Time Above Range. As expected, the flex group using traditional carb counting also achieved an absolutely massive +7.5 hour/day Time in Range improvement from 49% at baseline to 80% after three months (p<0.001). The flex group had a slight 1.6 hour/day Time in Range advantage compared to the fix group, but this advantage was not visible when looking at mean sensor glucose (147 mg/dL in fix group and 145 mg/dL in flex group; p>0.05) and A1c (6.8% vs. 6.6%; p>0.05).

• Coming in third was a study (n=15 type 1s) investigating a “open-loop bolus” on MiniMed 780G to handle high fat, high protein meals. Recall that in large quantities, fats and proteins do impact on blood glucose; the impact usually isn’t as “quick”-acting as it is for glucose, making it sometimes challenging to handle meals where carbs, protein, and fat are all involved. Participants ate a pizza with 60g of carbs exactly, but factoring in protein (38g) and fat (41g), the carb count was announced as 78g. Traditional carb counting led to a Time in Range of 86% at 10 hours post-meal, but using a modified “open-loop bolus” with carb, protein, and fat counting in a dual-wave bolus led to a Time in Range of only 65% 10 hours post-meal. Using the simplified “fixed” meal announcement as outlined in the previous study, Time in Range was comparable to traditional carb counting at 78%. The investigators concluded that the results suggest “integration of protein and fat content is not required.”
• Last on the docket was the “SMART” exercise trial (ClinicalTrials.gov). In the study, participants (n=55) wore MiniMed 780G and partook in three 45-minute instances of moderate intensity continuous exercise, 90 minutes after a breakfast (standardized based on body weight). Three strategies for insulin management were used: (i) 100% of insulin delivered with announcement immediately at exercise onset (SE cohort); (ii) 25% reduced bolus with exercise announcement 90 minutes prior to exercise onset (AE90 cohort); and (iii) 25% reduced bolus with exercise announcement 45 minutes prior to exercise onset (AE45 cohort). The results were staggering: in the post-exercise period (50-105 mins), time <70 mg/dL was 56% (!) for those in the SE group, vs. only 2% in the AE90 group and 29% in the AE45 group – see the full results below. In summary, the 90-minute pre-exercise announcement was shown to lead to significantly reduced hypoglycemia in the post-exercise period.

9. *NEW* Adult T1D technology uptake at Barbara Davis Center was 83% in 2021 (n=1,839), with ~40% on AID (70% Control-IQ, 30% 670G); just over half of adults on AID meeting A1c <7%, significantly higher share than T1DX

During a Tandem-sponsored symposium, Dr. Viral Shah (Barbara Davis Center) presented real-world outcomes from his center in Colorado. Specifically, this impressive longitudinal dataset (2014-2021) includes adults with type 1 diabetes and tracks the uptake of CGM (MDI or sensor-augmented pump [SAP]) and AID over time. Participants were indexed by visit year, which is why, as shown in the table below, the number of patients in each year fluctuates and might decrease sequentially (for example, from 2019 to 2020 the cohort decreased in size, likely due to COVID-19). As shown in the table below, the penetration of CGM in this cohort increases from ~27% in 2014 to ~83% in 2021 (this figure in 2021 includes the percentage of people on CGM/SAP and AID as well). In 2021, the penetration of AID was just under 40%, with around 70% of users on Control-IQ and the other 30% on “first-generation AID” (i.e., MiniMed 670G).

• Dr. Shah first presented a table showing that diabetes technology uptake was accompanied by a decrease in A1c. Specifically, as the overall uptake of technology in this cohort increased from 27% in 2014 to 83% in 2021, the mean A1c of the cohort fell significantly from 7.7% to 7.5% Perhaps more strikingly, Dr. Shah showed a graph (pictured below) revealing a significant A1c disparity among tech users and non-users. In 2014, this disparity was only 0.3% (7.5% vs. 7.8%, respectively). However, by 2021, this disparity was nearly one full percentage point (7.5% vs. 8.4%, respectively).

• The percentage of adults on CGM (MDI/SAP) or AID meeting the A1c goal of <7% significantly outpaced that of adults in the T1DX registry. Dr. Shah displayed a graph (shown below) revealing that while 51% and 44% of adults on AID and CGM, respectively, were meeting their targets for A1c in 2021, this figure was only 21% in the T1DX registry. Dr. Shah specifically highlighted how the introduction of Control-IQ in 2020 boosted the percentage of AID users meeting A1c goals.

Therapy and Big Picture Highlights

10. JDRF-funded CLVer RCT shows verapamil partially preserves stimulated C-peptide levels at 52 weeks in newly diagnosed pediatric type 1 diabetes; no significant difference in stimulated C-peptide levels with AID, despite staggering glycemic benefits (TIR 78% vs. 64%)

Dr. Gregory Forlenza (Barbara Davis Center) presented the highly anticipated full CLVer results of AID (Control-IQ or MiniMed) and verapamil on the preservation of beta cell function in newly diagnosed pediatric type 1 diabetes. JAMA simultaneously published the results in two papers, titled “Effect of Tight Glycemic Control on Pancreatic Beta Cell Function in Newly Diagnosed Pediatric Type 1 Diabetes” and “Effect of Verapamil on Pancreatic Beta Cell Function in Newly Diagnosed Pediatric Type 1 Diabetes.” See JDRF’s press release here.

Funded by JDRF and in collaboration with Dexcom, Medtronic, and Tandem, CLVer was a randomized, double-blind clinical trial including 113 youths aged 7 to 17 years within one month of diagnosis of stage 3 type 1 diabetes (T1D)[1] across six centers in the US. The primary outcome was mixed meal tolerance test-stimulated C-peptide secretion, as a measure of beta cell function, 12 months (52 weeks) after diagnosis. The study also assessed two questions: (i) could intensive diabetes management (with AID) achieve near-normalization of glucose concentrations versus standard therapy (CGM and no AID) 52 weeks after diagnosis; and (ii) could once-daily oral verapamil achieve near-normalization of glucose concentrations versus placebo?

For background, verapamil is a calcium channel-blocker that is administered either orally or intravenously and is indicated for hypertension, chronic stable angina, unstable angina, supraventricular tachycardia, and paroxysmal supraventricular tachycardia. It modulates the influx of calcium ions across arterial smooth muscle and myocardial cells, and it reduces the expression of thioredoxin-interacting proteins, which are involved in glucotoxicity-induced beta cell death.

Overall, stimulated C-peptide levels in the verapamil group were 30% higher compared to placebo at 52 weeks (p=0.04), showing that verapamil had partially preserved beta cell function. Meanwhile, there was no significant difference in stimulated C-peptide levels between AID users and non-users (i.e., intensive care vs. standard care).

• Rationale. A 1989 study by Shah et al. found that use of a Biostator, a glucose-controlled insulin infusion system, (n=12) was associated with beta cell preservation one year after diagnosis compared with controls (n=14). However, other studies, including Buckingham et al. and Boughton et al. have failed to confirm this. On verapamil, a phase 2 study published in Nature Medicine in 2018 reported that once daily oral verapamil promoted endogenous beta cell function while reducing insulin requirements and hypoglycemia in adults with recent-onset T1D, as C-peptide levels in the verapamil group (n=11) were 35% higher compared to placebo (n=13) after 12 months. This study sought to evaluate both hypotheses.
• Trial Design. Participants were evaluated across two cohorts and 12 months of follow up. In Cohort A (n=88), participants with a body weight ≥30 kg (due to drug dosing restrictions) were randomly assigned to one of four groups: (i) intensive care and placebo; (ii) intensive care and verapamil; (iii) standard care and placebo; or (iv) standard care and verapamil. In Cohort B (n=25), participants with a body weight <30 kg were randomly assigned 2:1 to either intensive or standard care, so as to not exclude younger participants with a lower body weight. Participants assigned to standard care in both cohorts were given Dexcom G6. Participants in the intensive group were randomly assigned to use either the Tandem t:slim X2 with Control-IQ or a Medtronic AID system (MiniMed 670G or 780G). Intensive care also included 35 visits in addition to five follow-up visits, while standard care only included five follow-up visits. Dosing was weight-dependent and was escalated at two-to-four-week intervals. The primary outcome was C-peptide levels – a marker of endogenous insulin production and beta cell function – in response to a two-hour mixed meal tolerance test at 52 weeks. Time in Range and A1c were secondary outcomes.

• Baseline characteristics. Across the verapamil and placebo groups, overall mean age was 12.7 years and A1c was 10.3% at baseline. Mean C-peptide levels at baseline were 0.66 pmol/mL and 0.60 pmol/mL in the verapamil and placebo groups, respectively. Across the intensive and standard care groups, mean age was 11.8 years and A1c was 10.3% at baseline.
• Intensive vs. standard diabetes management: Intensive control with AID had no impact on the decline in C-peptide secretion at 52 weeks, despite excellent glycemic benefits. At 52 weeks, the intensive care group had a mean Time in Range of 78%, compared to 64% for standard care (3.4 hour/day difference). A1c in the intensive care group was also significantly lower compared to the standard care group at 52 weeks (6.5% vs 7.1%). For the Europeans in the audience, Dr. Forlenza remarked that this is “phenomenal for American children.” However, AID had no impact on stimulated C-peptide preservation, aligning with data from the CLOuD RCT. Dr. Forlenza said that the results put the AID/beta cell preservation hypothesis to rest.
• Verapamil vs. placebo: Verapamil significantly reduced the decline in C-peptide secretion at 52 weeks. Impressively, consistent use of treatment was 94% in the verapamil group and 93% in the placebo group. At 52 weeks, the mean C-peptide level was 0.65 pmol/mL in the verapamil group (from a baseline of 0.66 pmol/mL) and 0.44 pmol/mL in the placebo group (from a baseline of 0.60 pmol/mL). This amounts to a 30% higher C-peptide level at 52 weeks with verapamil (p=0.04). The full results across the study duration are shown below. In line with the standard vs. intensive care results, there was also no interaction between verapamil and the management arms on C-peptide at 52 weeks. In other words, AID had no bearing on the efficacy of verapamil or placebo. Interestingly, as shown below, despite apparent beta cell preservation with verapamil, there was no difference in Time in Range and A1c between the veramapil and placebo groups at 52 weeks. Dr. Forlenza suggested that a longer study duration might reveal significant differences in future studies. Excitingly, JDRF’s press release indicates that it is funding several clinical trials to validate the results of this study and to see if verapamil is effective when used in conjunction with other disease-modifying therapies, such as Tzield.

• Safety results. Verapamil has a robust safety profile, having been first approved in the 1980s, and this trial was no different, with only a few adverse events. There were four serious adverse events in the verapamil group vs five in the placebo group. Additionally, there were three ECG abnormalities (zero in placebo), one hypotension event (zero in placebo), and two instances of elevated liver enzymes (two in placebo) in the verapamil group.
• Questions remain on the clinical implications of these results. We are curious to see what these results mean on a patient-level for youth who are newly diagnosed with type 1 diabetes. Particularly, we wonder whether verapamil would significantly impact glycemic metrics like A1c and Time in Range if taken for a longer duration. We also wonder whether C-peptide preservation would continue following the termination of treatment, as well as how verapamil treatment might impact DKA risk and insulin use. We were pleased to hear from Dr. Forlenza that the study group is currently conducting an extension trial with the same cohort. Almost everyone stopped verapamil after 12 months, so while the trial might provide insight into a potential legacy effect, we would also be curious to see studies examining a longer treatment duration. Dr. Forlenza hypothesized that the extension trial might reveal differences in Time in Range or hypoglycemia events, although he mentioned that the trial may not be powered for this. He also emphasized that the results are especially exciting given the low cost and well-established safety profile of verapamil (it’s even on the WHO’s List of Essential Medicines).

11. Irl Hirsch wins Children with Diabetes President’s Award for Innovation

Dr. Irl Hirsch (University of Washington) was bestowed with the second annual Children with Diabetes President’s Award for Innovation by CWD President Mr. Jeff Hitchcock. In his award lecture, “Innovation in Diabetes: What I Have Seen in my Lifetime,” Dr. Hirsch chronicled the remarkable innovation in diabetes from the time of his diagnosis in 1964 to modern day and invited attendees to appreciate how far the field has come. From BD’s Plastipak syringes to the invention of home blood glucose monitoring and the introduction of CGM, Dr. Hirsch noted that many of the innovations that today are viewed as commonplace were initially controversial. Ultimately, we left Dr. Hirsch’s presentation feeling inspired by the history of profound improvements in glucose monitoring, insulins, and prevention and treatment of complications; grateful for all those who contribute and continue to work towards new innovations; and with a new appreciation for the many tools available to patients and providers in today’s treatment landscape. Following a standing ovation for Dr. Hirsch’s moving and highly reflective presentation, representatives from Israel-based Hagar performed a live demonstration of the company’s non-invasive glucose monitoring technology, GWave. We were honored to witness the device’s first use outside of Israel (!) and we look forward to following its development and testing in larger populations. Dr. Hirsch concluded with a look to the future stating, “What I do know, is that we are not far away from non-invasive glucose monitoring and we are a long way from urine glucose testing. Whether it’s this company or another company, I do think we will be seeing, at some point in the future, automated insulin delivery with non-invasive glucose monitoring.” Bravo to Dr. Hirsch on this stirring talk and demonstration!

12. Dr. Viral Shah unveils investigator-initiated ADJUST-T1D trial of semaglutide in type 1 diabetes and emphasizes the importance of early CVD risk reduction type 1 diabetes

Dr. Viral Shah (University of Colorado) shared that he is leading a JDRF-funded, investigator-initiated trial (ADJUST-T1D) of GLP-1 semaglutide in people with type 1 diabetes, which is expected to initiate in April 2023. Dr. Shah provided an overview of its design – see slide below and see ClinicalTrials.gov. The study will enroll 72 participants with type 1 diabetes on AID who will be 1:1 randomized to either once weekly semaglutide 1 mg or placebo. Dr. Shah said that the primary outcome will measure semaglutide’s efficacy in improving time-in-range >70% without increasing time below range <4% and weight loss of at least 5% or more, and secondary outcomes will assess improvement in surrogate cardiovascular measures, using cardiovascular magnetic resonance (CMR), the carotid intima-media thickness test (CIMT), and pulse wave velocity (PWV). Cardiovascular assessments will be conducted on a subset of 36 participants at baseline and after 26 weeks of treatment. Dr. Shah said that he is optimistic about the trial results and hopes that this work can support further research and industry-sponsored phase 3 trials that could lead to the approval of GLP-1s in type 1 diabetes. We are particularly interested in the safety results from this trial, since previous research has suggested that GLP-1s are associated with slight increases in hypoglycemia risk in type 1 diabetes. 

• Dr. Shah emphasized the need to go beyond A1c and target multiple cardiovascular risk markers early in the course of type 1 diabetes. While the DCCT/EDIC study showed that intensive A1c lowering leads to significant reduction of cardiovascular events, A1c lowering does not eliminate the elevated CV risk that people with type 1 diabetes experience. Even people with type 1 diabetes with an A1c <7% have a nearly three-fold increased CVD mortality risk compared to the general population. To further lower CVD risk, Dr. Shah pointed to blood pressure, cholesterol, and insulin resistance. While a glucose clamp is the gold standard for measuring insulin resistance, it is not practical to do on every patient, so Dr. Shah highlighted the triglyceride/HDL ratio, which is a surrogate marker for insulin resistance and is a strong predictor of CV events in people with type 1 diabetes, much better than LDL. GLP-1s and GIP/GLP-1 tirzepatide have been shown to reduce insulin resistance, though their use in type 1 diabetes has not been well-investigated. During Q&A, Dr. Shah noted that he would focus on investigating GLP-1s and SGLT-2s in type 1 diabetes to target insulin resistance before TZDs, in part due to the weight loss associated with GLP-1s and SGLT-2s.

13. Once-weekly insulins are the “wave of the future”: KOLs dive into once-weekly insulins, with an eye toward type 1 diabetes

Drs. Julio Rosenstock (Velocity Clinical Research at Medical City Dallas and University of Texas Southwestern), Juan Pablo Frias (Velocity Clinical Research), and Steven Edelman (University of California San Diego) took to the stage to discuss progress on once-weekly insulins in both type 1 and type 2 diabetes – Lilly’s basal insulin Fc (BIF) and Novo Nordisk’s insulin icodec. As a side note, we learned from Dr. Frias that the preferred name for Lilly’s basal insulin Fc (BIF) is its generic name, Efsitora Alfa, though the two can be used interchangeably. We were especially intrigued by Dr. Edelman’s talk, which explored the potential of once-weekly insulins in type 1 diabetes. Ever the comedian, Dr. Edelman astutely commented, “Efsitora sounds like an Italian sportscar, BIF sounds like a brand of peanut butter.” Throughout his presentation, we were pleased to hear Dr. Edelman call for greater use of Time in Range as a primary endpoint – particularly as many of these trials included blinded or unblinded CGM. He made a particularly on-the-ball shoutout to Sanofi for incorporating Time in Range in the InRange trial, presented at ATTD 2022. Overall, panelists highlighted the transformative potential of once-weekly insulins for type 2 diabetes given the strong phase 2/3 evidence and noted that these insulins may also be useful for the management of type 1 diabetes. However, data in type 1 diabetes is currently limited and there is a substantial learning curve to understand initiation, titration, and patient and provider education. Following the somewhat lackluster results for BIF and icodec in type 1 diabetes, we were pleased to hear Dr. Edelman address the potential of these agents in type 1 diabetes and we look forward to further research, especially on the titration and patient/provider education front. For more, see our overview of the progress and future of once-weekly insulins in our 2022+2023 Reflections.

• Dr. Edelman discussed the potential of once-weekly insulins to help address unmet needs in type 1 diabetes, citing their safety and efficacy in a clinical trial setting and emphasizing that the greater convenience can significantly reduce the burden of self-management and improve quality of life. In the phase 2 study for BIF in type 1 diabetes (presented at EASD 2022), BIF led to a Time in Range of approximately 60%, while Time Below Range met the goal of <4% and Time Above Range was 20%. Overall, Dr. Edelman said this is what he’d expect to see in this population of people with type 1 diabetes on MDI. Additionally, while BIF showed higher fasting plasma glucose values than insulin icodec, Dr. Edelman said this can be attributed to limitations of the prespecified dosing algorithm. Turning to the ONWARDS 6 trial for icodec in type 1 diabetes, for which topline results were released in June 2022, Dr. Edelman noted the 0.4% reduction in A1c as well as the elevated rate of severe or clinically significant hypoglycemia with icodec. Dr. Edeleman said the field needs to wait and see how the 26-week extension phase of the study plays out and how the titration regimen affects the results. Ultimately, he emphasized that the initiation and titration process will require education to understand dose equivalents (particularly important from switching from daily to weekly basal insulin) and loading doses. In terms of a target patient population, Dr. Edelman said that once-weekly insulins may be especially valuable for people in need of self-care assistance or for people who face difficulty with the technology burden of “dumb pumps” (i.e., those with no communication with CGM) and hybrid closed loop systems.
• Dr. Rosenstock characterized once-weekly basal insulins as having transformative potential in type 2 diabetes as the “way of the future”, focusing on Novo Nordisk’s insulin icodec in the phase 3 ONWARDS program. In particular, he highlighted ONWARDS 1 as the longest duration randomized trial versus glargine U100 to confirm the efficacy and safety of weekly insulin icodec with HbA1c as primary outcome but with blinded CGM as key secondary outcomes to assess multiple target metrics. He also emphasized the translational designed of the ONWARDS 5 as an important study because it incorporates real-world elements to mimic a routine care setting. ONWARDS 5 has more liberal inclusion and exclusion criteria (insulin-naïve type 2 diabetes with an A1c >7%; no limits on oral antidiabetic agents), allowing for the enrollment of a broader patient population that is likely to be more representative of the routine clinic setting. With a streamlined visit schedule, simplified titration via a dosing app, and insulin initiation/titration of daily basal insulin according to clinical practice and local labelling, the study design gives HCPs more flexibility to tailor diabetes management to the individual patient. Recall that topline results announced in October 2022 found that icodec conferred a 0.4% greater A1c reduction vs. insulin degludec and insulin glargine U100/U300, with no difference in severe or clinically meaningful hypoglycemia. Across the wider ONWARDS program in type 2 diabetes (ONWARDS 1, ONWARDS 2, ONWARDS 3, ONWARDS 4, and ONWARDS 5), Dr. Rosenstock emphasized icodec’s consistent topline A1c reductions and low hypoglycemia risk of less than 1 event/patient/year.
• Reviewing the phase 3 QWINT program for Lilly’s BIF, Dr. Frias agreed with Dr. Rosenstock that once-weekly insulins will be “the way of the future” and also highlighted important considerations around patient and provider education. Briefly reviewing phase 2 results (as no phase 3 data has been released yet), Dr. Frias highlighted BIF’s noninferior reductions in A1c with similar percent Time in Range, hyperglycemia, and hypoglycemia in a trial presented at ADA 2021. He additionally noted that BIF had lower within-day glucose variability – likely a function of BIF’s very plat PK profile throughout the week – and strong reductions in nocturnal hypoglycemia. Likewise, in the second phase 2 trial (presented at ADA 2022 and accepted by Diabetes Care this week!), BIF led to non-inferior reductions in A1c, with participants spending >75% Time in Range and 60% achieving an A1c <7%. Now that all five phase 3 QWINT trials are underway, with expected completion by 2024, Dr. Frias said he imagines regulatory filing will happen sometime in 2024, with BIF becoming available in the clinic in 2025.

14. diaTribe Solvable Problems on the future of Time in Range and CGM with key takeaways on early initiation, EHR integration, Time in Tight Range, and more

Professor Chantal Mathieu (Katholieke Universiteit, Leuven, Belgium), Professor Ananta Addala (Stanford University, USA), Professor Elaine YK Chow (Chinese University of Hong Kong, Hong Kong SAR), and Professor Emma Wilmot (University of Nottingham, UK) took to the stage in the beautiful Französischer Dom to discuss the future of Time in Range and CGM with moderator Professor Tadej Battelino (University Medical Center Ljubljana, Slovenia) at the diaTribe Foundation’s Solvable Problems in Diabetes. The conversation covered a massive array of topics relating to technology use in people with type 1, type 2, gestational, and prediabetes, elucidating a number of key considerations for management moving forward. Salient themes from the evening included the importance of initiating technology use as early as possible following a diagnosis, the growing base of cost-effectiveness data supporting CGM use across a range of populations, the importance of integrated device data with electronic medical records, considerations around Time in Tight Range, and the fundamental need for access to these technologies, particularly as they become integral to the standard of care. We will certainly be back with more on this wonderful event, but for now, please enjoy these “quotable quotes” from the esteemed panel.

• Professor Mathieu on the rapidly growing use of diabetes technology: “What will happen with CGM in diabetes? I think it will not be five years before we look back and say, ‘do you remember the day we pricked the finger?’ It’s like the revolution of pens in Europe. Do you remember when we had syringes? I think it will be the same with CGM and that will go very fast. So the concept of TIR it’s already there. In the clinic, it’s a tool we use to discuss glucose control with people living with diabetes, and they dig it, they get it. It’s going very fast and everything will depend on the affordability of the well-performing sensors. We have some cheap sensors, but if it’s cheap, and it’s not accurate, we don’t want it. So we need the good ones to become more affordable.”
• Professor Wilmot on Time in Tight Range: “Diabetes is really complicated. It is one of the most challenging chronic conditions to manage and we want to support people and to help them feel like they are achieving their goals. Moreover, we have to think about different populations in diabetes and consider what is achievable. It may be different for a person with newly diagnosed type 2 diabetes to reach target than a person with type 1 diabetes with a disease duration of 30+ years. So there are lots of things to consider and we’re still trying to figure out where we are.”
• Professor Chow on Time in Range in Clinical Trials: “Time in Tight Range will become increasingly important, especially as we use Time in Range as an endpoint in clinical trials. For now, we need to build more evidence and more data so that regulators feel comfortable approving therapies on this more reliable measure of blood glucose.”
• Professor Addala on actionable items to address disparities: “Sometimes when I do my session on disparities, people get disheartened because it looks like there’s disparities everywhere. While that is true, what’s nice is that actually means there are actionable items wherever you are in the diabetes community. If you are in industry, think about how devices can be more culturally and linguistically accessible. If you are a funder, are you ensuring that the diabetes in clinical trials are well-represented so that when the trials results are published, then they’re representative of the population.”Professor Battelino on the society-wide benefits of CGM: “CGM doesn’t only benefit people with diabetes, it benefits society. You use resources more efficiently, and we dont discuss this angle enough with the public, perhaps we should. The benefit of CGM TIR is also the self-assurance people may get from it. People come to me and I see their A1c, which I really dislike because this is the history of the last three months. I’ve seen it, the A1c isn’t good, but can I change the A1c of the last 3 months? Even if I were a God, I couldn’t, and neither could the person with diabetes. So, I dislike this backward-oriented discussion of diabetes.”

CGM and Connected Care Highlights

15. CGM as first-line therapy for all people with type 2 diabetes in primary care in rural Findlay, Ohio: A1c improves from 9.8% to 7.2% after six months; no additional education or clinical visits provided

In one of the more profound presentations in our recent memory, Dr. Thomas Grace (Blanchard Valley Health System) shared promising initial results of CGM as a “first-line” tool for all people with type 2 diabetes, regardless of insulin use. In a study specifically designed to test CGM (Dexcom G6) in people with type 2 diabetes who do not have coverage, mean A1c fell from 9.8% to 7.2% for non-insulin users (n=11) and from 10.0% to 7.1% for insulin users after six months (n=13). While these results alone are impressive, Dr. Grace’s story and the context around the study were equally powerful.

As Dr. Grace described, he began his career as a hospitalist in the ICU in Findlay, Ohio, a rural town with a population around 40,000. Seeing a high rate of people showing up to the ICU with DKA, Dr. Grace began spending more and more time working upstream as the diabetes specialist in town. This was also motivated by Dr. Grace’s personal connection to the disease, as a person with type 1 diabetes. When Dexcom G6 was launched in the US in 2018, Dr. Grace began using the device with his type 2 patients. Not surprisingly, there were a lot of barriers early on and Dr. Grace noted the hours he spent writing letters to payers for nearly every CGM prescription he wrote for his type 2 patients. Knowing this wasn’t sustainable and recognizing the impact in his community, Dr. Grace approached Dexcom to support a small, six-month study (n=38) in Ohio, enrolling people with type 2 diabetes with an A1c >7.5% using either basal insulin only or non-insulin therapy and providing those participants access to CGM. Those results, presented at ADA 2021, showed a remarkable A1c drop from 10.1% to 7.1%. Motivated by these initial results, Dr. Grace began an ambitious project, which he referred to as “Community Glucose Monitoring (CGM),” aiming to make Findlay, OH, the first community in the US “where everyone with diabetes has access to CGM.”

• The study was open to all residents of Findlay, OH with type 2 diabetes and an A1c ≥7.4%. Participants were recruited through the local news, word of mouth, and primary care physicians. As the study is ongoing, Dr. Grace only presented the interim results from the first 32 participants in a project that will reach more than 500 people in the near future. These participants averaged 55 years old and 91% were White. Not surprisingly, BMI and A1c were quite high at 37 kg/m² and 9.9%, respectively, at baseline. In terms of medication, 18 people were using insulin and fourteen were on non-insulin therapy. Metformin was the most common non-insulin medication, though GLP-1, SGLT-2, and sulfonylurea utilization was also common. The majority of participants (n=24 of 32) were taking between 2-3 medications.

• Importantly, primary care physicians were not given any guidance on CGM use or interpretation. Initially, some basic operation and CGM set-up was done with the subjects at the health department. Then, participants effectively were left “to their own devices,” wearing CGM continuously but otherwise had no changes to their standard care. Every three months, the subjects would return to the health department to re-supply on CGMs and get their A1c, weight, and height measured.
• For the 24 subjects with full six-month data, average A1c fell from 9.9% to 7.1%. In fact, the A1c after three months was 7.2% and that improvement was maintained out to six months. Broken into insulin users and non-users, results were very similar, as A1c fell from 10.0% to 7.1% for insulin users and 9.8% to 7.2% for non-insulin users over six months. Similarly, A1c improvements were seen across all baseline A1c levels, though the participants with the highest baseline A1cs saw the biggest improvements.

• While these are just interim results, Dr. Grace shared some qualitative observations from the study. Primary care physicians who were initially unfamiliar with CGM began getting comfortable with Dexcom Clarity. Additionally, patients were able to start, use, and learn from CGM without much oversight from clinicians. Dr. Grace also speculated that as more data comes in, it’s likely that there will be evidence of patients reducing their number of medications, an important outcome not just for health economics but also patient quality of life. As this “Community Glucose Monitoring” project continues to enroll participants, we look forward to seeing some of these additional data points in the future.
• As Dr. Grace noted, the conversations around CGM use in type 2 diabetes are changing. In this year’s ADA Standards of Care, CGM is now strongly recommended to people on basal insulin only. In the US, Medicare appears to be headed in a very similar direction with its draft guidelines published in October. Perhaps driven by some of Dr. Grace’s work, Ohio recently became the first state whose Medicaid program covers CGM for all people with diabetes.
• Dr. Grace encouraged providers to think about using CGM as a “first-line therapy” for all people with diabetes. While he noted there are differences between the needs of intensive insulin users where CGM is a “safety device” for avoiding hypoglycemia and non-intensive insulin users where CGM is a behavior modification tool, CGM is effective in both populations. As Dr. Grace noted, the majority of the text and space for most diabetes treatment algorithms is focused on A1c and adjusting medication regimens. However, at the very top, the first-line therapy is “lifestyle therapy and ongoing glucose monitoring.” With the continuous feedback provided by CGM, Dr. Grace believes that lifestyle-based therapy can be much more effective. In a word, this symposium was moving, and we can only hope that the diabeterati will take note of this story as a sign of where the field could one day be.

16. Sanofi-sponsored symposium elevates person-centered approaches to care: using connected technologies and digital tools to foster effective self-management

During a Sanofi-sponsored symposium chaired by Dr. Stewart Harris (Western University, Canada), Dr. Thomas Danne (Children's Hospital Auf der Bult, Germany), Tadej Battelino (University Medical Center Ljubljana, Slovenia), and Partha Kar (NHS England) discussed barriers to and opportunities for effective diabetes self-management, with an emphasis on digital tools. Dr. Danne first dove into the importance of diabetes self-management, highlighting that if people with diabetes spend three hours per year with healthcare providers (HCPs), that leaves 8,757 hours per-year of self-management. (Though this is our speculation, we’d imagine three hours a year with HCPs is actually more than most people with diabetes receive.) This is especially important to consider, and Dr. Danne emphasized the need for HCPs to support self-management, as optimal self-management behaviors correlate with improved glycemic outcomes. Dr. Battelino explained that CGM and CGM-based metrics like Time in Range have facilitated collaborative decision making between patients with diabetes and HCPs, breeding an era of greater engagement with self-management. Dr. Kar closed the session with a discussion of strategies for improving adherence and satisfaction with insulin therapy through novel technological advances like smart connected pens and caps. Across each presentation, we were pleased to see the speakers’ emphasis on using technology to address the personal needs and goals of each person with diabetes.

• Dr. Danne emphasized that current approaches to support self-management may not fully meet the needs of people with diabetes, as hypoglycemia rates and A1c levels remain high at a population level. Dr. Danne noted that CGM has been shown to significantly reduce clinically meaningful hypoglycemia compared with BGM. Additionally, new findings from the InRange study, presented at EASD 2022, showed that CGMs catch five to six times more biochemical hypoglycemic events than fingersticks. Beyond technology, Dr. Danne also highlighted the importance of using positive language and the potential of fixed-ratio combinations to encourage effective self-management.
• Dr. Battelino’s presentation centered around improving communication between people with diabetes and HCPs through CGM metrics. He explained that metrics like Time in Range allow patients and HCPs to collectively examine the data and create an action plan. He also expressed his excitement for upcoming trials that may further support the clinical utility of TIR. For instance, the TRENT trial (expected to complete in March 2024) will evaluate insulin degludec and insulin glargine U300 in insulin-naïve people with type 2 diabetes and renal impairment, and will include a CGM sub-study looking at TIR, Time Below Range, Time Above Range, GMI, and glucose variability. Another upcoming study, GP-Detect, estimate the occurrence of hypoglycemia in people with type 2 diabetes on insulin glargine U100 in primary care, with TIR as a secondary endpoint. Dr. Battelino also presented results from two ATTD 2023 posters. The first, a sub-analysis of the LixiLan-L trial (fixed-ratio insulin glargine U100 and lixisenatide), demonstrated a greater increase in TIR with iGlarLixi vs just insulin glargine. The second, a pooled analysis of the LixiLan program, confirmed results of the individual LixiLan studies. Overall, although TIR can be enormously useful in glycemic management, Dr. Battelino also cautioned that as CGM use increases, there are several factors beyond glycemic control to consider when determining its value for a patient, like hypoglycemia unawareness, personal preference, digital literacy and numeracy, and access to education around CGM use.
• Dr. Kar underscored the importance of building relationships free of judgement when using digital tools to support diabetes self-management. He explained that through digital tools, people with diabetes are now trusting HCPs with a view of their lives, so HCPs must prioritize empowering self-management without personally judging the data. He explained that digital healthcare solutions can foster a more person-centered system by encouraging effective clinical conversations, improving concordance with therapies, and enhancing self-management. After providing this context, Dr. Kar dove into recent developments in connected devices. Primarily, he highlighted SoloSmart, a smart pen attachment, first unveiled at ATTD 2022, that is compatible with all Sanofi SoloStar insulin pens. Additionally, Dr. Kar presented results from an ATTD 2023 poster on the real-world effectiveness of a connected system with Bicorp’s connected pen cap, Mallya (which received FDA 510(k) clearance in December 2022), and an app (Health2Sync) in adults with type 2 diabetes on basal insulin in Taiwan. The real-world study (n=83) found significant decreases in A1c among both new and previous insulin users. Specifically, 66% of new and 29% of previous insulin glargine U300 users achieved A1c reductions >0.5% after 120 days. Dr. Kar said the core of it all is “improved communication” and emphasized a need to address barriers to digital healthcare, including integration into EMR, training, and digital literacy.

17. Time in Tight Range? Drs. Tadej Battelino, Thomas Danne, and Peter Adolfsson advocate for the metric for people with both type 1 and type 2 diabetes

In an evening session, Dr. Tadej Battelino (University Medical Center Ljubljana, Slovenia), Thomas Danne (Hanover Medical School, Germany), and Peter Adolfsson (University of Gothenburg, Sweden) all advocated for utilization of Time in Tighter Range. As background, Time in Tight Range (TITR) is a newer proposed metric to evaluate glycemic management, with a defined “in range” of 70-140 mg/dL, lower than the range of 70-180 mg/dL for Time in Range (TIR). Dr. Danne said that he believes TITR is a more useful metric as it allows clinicians to understand the amount of time an individual spends in normoglycemia. To make his point more salient, Dr. Danne said that he “doesn’t want to lie anymore” when he is asked whether children with diabetes will have a normal life expectancy, as he believes normoglycemia, ultimately, is what’s needed to reduce complications. Dr. Danne then overviewed the clinical relevance of Time in Tight Range for both type 1 and 2 diabetes, taking a deep dive into why he believes the more aggressive definition is the ideal goal to achieve better outcomes. Dr. Danne said that while TIR is gaining acceptance, he sees AID as the “game changer” that can propel people with diabetes to achieve TITR. To back this claim up, Dr. Danne shared data demonstrating that while there are differences in average A1cs across countries, the same countries all have comparable average TIR from AID systems. Dr. Danne said that now that these systems are becoming the standard of care for many people with diabetes, achieving a tighter glycemic goal is increasingly doable and should become the standard goal.

• Dr. Adolfsson then advocated for the use of TITR for people with type 1 diabetes. Dr. Adolfsson shared that he believes TITR is a relevant metric to use as a target since it is both an understandable and achievable metric for people type 1 diabetes. Dr. Adolfsson accredited the fast pace of technology innovation as an enabler of “greater and higher” goals for people with diabetes, including TITR. Additionally, he reminded the audience that current technologies, particularly CGM, enable more aggressive glycemic management without significantly increasing risk of hypoglycemia, so people with diabetes should feel less risk setting a more aggressive goal. Dr. Aldofsson then shared data demonstrating that a TITR of 50% corresponds with an A1c level of 6.5% and reminded the audience that the lower the A1c level, the better the overall outcomes, suggesting that all people with type 1 diabetes can benefit from a tighter target range. However, the presenters did agree for the need for larger data sets and more trials on the correlation of TITR with long-term outcomes.
• Dr. Battelino then took the stage to advocate for TITR for people with type 2 diabetes. Dr. Battelino shared that he believes there are multiple use cases for TITR for people with type 2 diabetes, especially in newly diagnosed type 2 diabetes. For newly diagnosed type 2 diabetes, Dr. Battelino argued that TITR should be an achievable glycemic goal, as at the early stages of disease, this target should be easier for people to achieve. In addition, Dr. Battelino used data from a study to illustrate that for people with newly diagnosed type 2 diabetes, achieving a TIR of 90% did not result in significantly more cases of remission, suggesting a more aggressive metric is necessary. Dr. Batellino also made the case that TITR has the potential to reduce complications for people with type 2 diabetes. To back up this point, he overviewed glycemic data that demonstrates the value of lower average blood glucose levels in decreasing risk of both cardiovascular disease and dementia. Although the jury is still out on the use of this metric on a routine basis (giving everyone a choice), we appreciated the data and fervent advocacy from the speakers on this topic and we surely understand that some people with diabetes would like to make sure they are in a tighter range. That’s true for many reasons – one could, of course, have an average of 179 mg/dL and 100% Time in Range. As we understand it, this is largely what AGP (average glucose profile) is for, although we are not yet at the point where patients themselves use this routinely – though it’s surely becoming easier to do so! Some doctors also just have an eye on both “TIR” as well as “average glucose” and ask PWD to try to keep averages in the 130s … other clinicians try other strategies! More to follow on this front – we consider “TITR” one of the biggest discussion points of the meeting and we applaud the work to help patients actively do as well as possible while also helping clinicians talk in a standardized way about patient progress. 

18. Drs. Peter Jacobs and Leah Wilson discuss insulin dosing decision support system, DailyDose; Dr. Revital Nimiri shares real-world data on DreaMed Advisor Pro

In an afternoon session, Oregon Health and Science University’s Dr. Peter Jacobs and Dr. Leah Wilson presented physician validation data of DailyDose, an insulin dosing decision support tool for people with diabetes on MDI. For background, DailyDose provides on-demand, real-time mealtime insulin dose advice via a mobile app connected to Dexcom G6 and Medtronic’s InPen. The system analyzes glucose patterns and provides recommendations for carb ratios and correction factors each week, along with an updated suggestion for basal insulin dosing regimens. As we learned from the research group at DTM 2019, DailyDose’s recommendations are specifically programmed to prioritize improved Time in Range, driven by a postprandial hyperglycemia prevention algorithm, and reduced hypoglycemia, aided by decision support around exercise. Dr. Jacobs began by overviewing eight-week results from the DailyDose system which were originally published in DT&T in September 2022 and evaluated the clinical efficacy of the support tool in 25 people with type 1 diabetes. Overall, the study found that accepting and following more than half of DailyDose’s dosing recommendations within a week was associated with a +1.5 hour/day increase in Time in Range, which Dr. Jacobs explained translates to a clinically significant 6.3% increase. The improvement jumped to a +4.1 hour/day improvement in Time in Range during weeks when participants followed the meal bolus dosing recommendations provided by the DailyDose bolus calculator within 0.5 units (on average) compared with weeks when they did not follow the bolus calculator recommendations, with no increased likelihood of hypoglycemia.

• Dr. Wilson shared non-inferiority data on DailyDose’s dosing recommendations compared to those from expert providers. To validate DailyDose’s recommendations, Dr. Wilson shared that three endocrinologists scored weekly CGM and insulin reports and reviewed the recommendations provided by the decision support tool. The endocrinologists rated DailyDose’s recommendations based on: (i) their agreement with the system’s recommendation and (ii) their perception of the recommendation’s safety. The analysis found that the majority of physicians agreed with DailyDose’s recommendations 67% of the time and only 6% of the time did all three physicians disagree with the system’s recommendations. Interestingly, the study demonstrated that the agreement rate between DailyDose and expert providers was higher than internal agreement between providers. This would suggest that DailyDose’s recommendations tend to lie somewhere in the middle of providers’ recommendations. Turning to safety data, the physicians generally scored the system’s recommendations as safe, with all three physicians rating the recommendation as safe 85% of the time and unsafe only 3% of the time.
• Dr. Wilson then overviewed the design of the next clinical study to investigate DailyDose. The multicenter randomized controlled trial, which is sponsored by the Helmsley Charitable Trust, will investigate DailyDose in 93 adults for 38 weeks. The trial will include updates to the DailyDose app based on feedback from the original study, with added features based on user feedback, such as (i) allowing user-driven customization of recommendations; (ii) adding a flagging feature, which allows users to remove selected data from the recommendation algorithm; and (iii) adding a recommendation timing feature and collecting more data about declined recommendations. Dr. Wilson shared that the researchers hope these features increase the interpretability of the app and also increase the number of accepted recommendations across all users. Unlike the original clinical trial, Dr. Wilson shared that the upcoming trial will also include diabetes education and behavioral health interventions. We look forward to seeing the outcomes from this next study, which Dr. Wilson mentioned will be readout in ~2026.
• We have seen similar validation study designs from other decision support systems, including endo.digital’s DreaMed Advisor Pro. Dr. Revital Nimri (Schneider Children’s Medical Center, Israel), who also presented during today’s session, read out DreaMed Advisor Pro’s Advice4U validation study at ATTD 2020. The study showed that use of the Advisor Pro yielded non-inferior results to expert physician advice in young people with poorly controlled type 1 diabetes on pump therapy. During today’s presentation, Dr. Nimri shared satisfaction data from people with type 1 diabetes who utilized the DreaMed Advisor Pro decision support system. The ten-week study enrolled participants (n=16) with type 1 diabetes on MDI. Dr. Nimri shared that these data represent the results up to February 1, 2023 (see graphic below). We look forward to seeing the full results from this trial and hope to see satisfaction correlated with improved outcomes.

19. *NEW* Novo Nordisk-symposium on digital health highlights importance of connectivity and patient-reported outcomes in diabetes management; age and day/time of insulin dosing found to be significant predictors of concordance in new connected pen data in T1D (n=1,336)

Dr. Thomas Danne (Hannover Medical School, Germany) moderated this primetime Novo-Nordisk sponsored session with Dr. Julia Mader (Medical University of Graz, Austria), Dr. Partha Kar (NHS England), and Dr. Tadej Battelino (University Medical Center Ljubljana, Slovenia). The conversation focused on digital health and technology use to improve outcomes for people with diabetes, specifically homing in on connected pen use and continuous glucose monitoring in people with type 2 diabetes. Dr. Danne framed the presentation by noting a high disease burden is partially related to limited data surrounding the timing and dosing of insulin. He noted that nearly one in four meals are associated with either a late or missed insulin bolus, and that insulin use is associated with more medication errors than any type of drug. Moreover, these missed boluses are associated with worse glucose outcomes, signifying the need for greater support for patients, which primarily comes through technology use. Patients have numerous options to manage their diabetes, and although we definitely hear significant support for AID, particularly in patients with type 1 diabetes, this option may not be viable for some patients. Thus, much of the session reviewed how connected pens, combining smart insulin pens with continuous glucose monitoring, can be used to improve outcomes for these patients. Throughout the session, there was a strong theme of individual care, creating judgment-free lines of communication, and the importance of patient-reported outcomes, which speak to a broader shift toward patient-centered care in recent years.

• Dr. Mader used a case study of her patient, Tina, to outline the major benefits of connected pens. Tina was recently pregnant and was eager to maintain tighter glucose control but was hesitant to go on an AID system because of her position as a kindergarten teacher and her hobby as a Lindy Hop dancer. Thus, the connected pen was an intriguing option, because it helps with (i) missed insulin doses; (ii) mistimed insulin doses; (iii) insulin dosing errors; and (iv) bolus dose misjudgment, which enables better glucose control through more accurate and timely insulin dosing and improved and more informed caregiver-patient interactions.
  
  • The benefits of connected pen use are undisputed. Studies have found a 1.9 hour/day increase in Time in Range and a 1.5 mmol/L reduction in mean blood glucose, as well as a 43% decrease in missed bolus injections and 5.4 fewer mistimed insulin doses/month. Importantly, patient reported outcomes were also improved, with patients stating greater confidence in managing daily injections leading to improved diabetes self-management and overall well-being. That said, these improvements may still not be quite on the same level as is seen with AID systems, so the tradeoffs between control and personal preference should be considered carefully.
  • Part of the issue is that it can be quite challenging to administer insulin doses in the midst of busy lives, and even with a connected pen, the problem is not entirely resolved. In new data based on 1,336 individuals with T1D using a connected pen for basal insulin therapy resulting in data for 171,378 injection days, the mean daily probability of missing a basal insulin dose was 3.5%. Interestingly, age, day of the week, and time of injection were factors significantly associated with treatment concordance, with the likelihood of missing an injection increasing for people in their 20s-30s and increasing when taking insulin on Saturday or in the middle of the day. However, more frequent data uploads were associated with a lower frequency of missing a basal dose.

• Turning to people with type 2 diabetes, Dr. Kar said that there is a strong need for technology in this population as well. Diabetes technology offers a visualization of the patient’s daily life, and while this can certainly improve dialogue between patients and providers, Dr. Kar said that these data must be used with compassion and humility. “If there is one thing I want you to take away from this lecture,” he said, “we must never judge anyone based on this data.” Much of the conversation focused on concordance with treatment regimens, especially complex ones, that are unfortunately lower than should be. For example, a 2012 study found that while 78% of patients taking one injection were in concordance with their suggested management, this number dropped to 61% for people taking 4 injections.
  
  • Dr. Kar was adamant that there is no one “typical person” for whom a connected pen might be beneficial. In particular, he cautioned providers from making snap judgments about tech or ruling out certain populations. While it may be easy to think that older people may not want or be able to handle technology, a study shared at DTM 2022 found that individuals 65+ actually had a higher Time in Range (61% vs 57%, +1 hour/day) compared to younger users due to better basal/bolus concordance, higher calculator use, and better report generation. Thus, these data reiterated that engagement with the connected pen data may improve outcomes and strongly underscored that all patients should be offered this technology.

• Continuing the theme of type 2 diabetes management, Dr. Battelino highlighted the benefits of CGM use for these individuals. Starting with the growing support for CGM in T2D, he highlighted the 2023 Standards of Care and ADA/EASD Consensus Report on the Management of Hyperglycemia in T2D, which strongly advocate CGM for those on insulin while supporting adaptation of effective technology use in the clinic to support self-monitoring. Turning to the evidence for these claims, Dr. Battelino highlighted a 2021 study (n=175 participants) that showed CGM use conferred an additional 0.5% reduction in A1c compared to BCM, and participants spent 43% less time in hypoglycemia, but he noted that the support is flooding in, with “a new paper published almost every week.” Beyond the glycemic benefits, Dr. Battelino also shared treatment satisfaction data showing that CGM use allowed for more flexible diabetes management and that a significant number of participants would recommend the system to their counterparts over BGM. Moving past the device itself, Dr. Battelino argued for greater connectivity between CGMs, connected pens, and pumps with electronic medical records (EMR) and for digital clinics to create expanded access to care for people with diabetes..
• While the session underscored the many benefits of diabetes technology, there is certainly work to be done. When asked what each of the panelists will look for in future devices, Dr. Kar said, “cheaper and smaller,” Dr. Mader said, “cheaper and smarter,” and Dr. Battelino said, “connected and kinder to the individual that uses it.” These are certainly compelling visions, and we hope to see the field continue to shift in these directions.

20. *NEW* Dr. Irl Hirsch questions current in-hospital glycemic targets and calls for an RCT with CGM to investigate individualized in-hospital targets

Dr. Irl Hirsch (University of Washington) suggested that optimal in-hospital glycemic targets may vary depending on an individual’s A1c at admission and past insulin use, and hypothesized biological mechanisms for this variability. Specifically, Dr. Hirsh highlighted a study of 5,567 ICU patients from 2011 to 2019  in Critical Care Medicine, which spurred his interest in this topic: “The Interaction of Acute and Chronic Glycemia on the Relationship of Hyperglycemia, Hypoglycemia, and Glucose Variability to Mortality in the Critically Ill.” Unexpectedly, this study found that people admitted with an A1c >8% had a three-fold greater mortality if their glucose levels were between 80-140 mg/dL, compared to >180 mg/dL (16% vs. 5% mortality, respectively). In contrast, people with an A1c <6.5% had 3.5-fold lower mortality if glucose levels were between 80-140 mg/dL, compared to >180 mg/dL (11% vs. 36%, respectively). In line with previous studies, such as the landmark NICE-SUGAR study, this study also found that hypoglycemia was associated with increased mortality for all patients, and this association was strongest for those with an A1c <6.5%. Dr. Hirsch said these findings lead to more questions than answers: Why is a high A1c with hyperglycemia protective, while the same A1c with euglycemia is associated with an increased in-hospital mortality? Can this observation be reproduced in a multi-center RCT? Why is stress hyperglycemia and hypoglycemia more dangerous in people with A1c <6.5%? What’s the impact of pre-admission insulin therapy, GLP-1, SGLT-2? Diving into these questions, Dr. Hirsch shared his and others’ ongoing research on the impact of prior glucose control and insulin therapy on ICU mortality and provided hypothesized mechanisms their surprising findings. Ultimately, Dr. Hirsch called these findings hypothesis generating and currently insufficient to change ICU targets, and he said that further study through an RCT with CGM is required to replicate these findings.

• In-hospital hyperglycemia has been associated with reduced mortality among people with an A1c <6.5% on insulin and among people with an A1c >8%. Dr. Hirsch highlighted this interesting finding from a 2021 study that he co-authored: “Acute and Chronic Glucose Control in Critically Ill Patients With Diabetes: The Impact of Prior Insulin Treatment.” This study of 5,245 ICU patients examined ICU mortality rates stratified by A1c level and insulin use, as shown in the image below. As expected, people without diabetes experienced a five-fold increased risk of mortality of their in-hospital glucose levels were >180 mg/dL. However, people with an A1c <6.5% on insulin experienced a two-fold increased risk of mortality if they had in-hospital euglycemia (80-140 mg/dL), as opposed to hyperglycemia (>180 mg/dL). Furthermore, as highlighted by the red circles, people with an A1c >8% with or without insulin experienced reduced mortality if their in-hospital glucose levels were >180 mg/dL, compared to between 80-140 mg/dL. Dr. Hirsch shared two potential biological mechanisms for these results.

• • Dr. Hirsch hypothesized that people with an A1c <6.5% on insulin are protected by in-hospital hyperglycemia due to decreased expression of the GLUT-4 glucose transporter. For background, Dr. Hirsch noted that the GLUT-1 transporter is dependent solely on glucose concentration, while the GLUT-4 transporter is regulated by insulin. He suggested that people with an A1c <6.5% on insulin have decreased GLUT-4 expression due to insulin resistance, so greater blood glucose concentrations are required for glucose to enter cells. As a result, in-hospital hyperglycemia does not result in glucose toxicity and is associated with reduced mortality in this population. In contrast, in people without diabetes, in-hospital hyperglycemia leads to increase glucose flux into cells, greater glucose toxicity, and increased mortality. Dr. Hirsch summarized this hypothesis by saying, “Systemic exogenous hyperinsulinemia with low HbA1c protects from hyperglycemia.”
  • Dr. Hirsch hypothesized that people with an A1c >8%, with or without insulin, are protected by in-hospital hyperglycemia due to increased expression of the GLUT-1 glucose transporter. Since the GLUT-1 transporter is dependent solely on glucose concentration, Dr. Hirsch suggested that people with an A1c >8% may experience upregulation of the GLUT-1 transporter. This GLUT-1 upregulation may result in improved glucose flux with hyperglycemia in the ICU. However, when these individuals have in-hospital glucose between 80-140 mg/dL, GLUT-1 transporters on cell membranes are reduced and glucose flux into cells is reduced, compared to these individuals’ baseline. As a result, these individuals may experience relative hypoglycemia, which Dr. Hirsch defined as glucose levels between 70-110 mg/dL for those with an A1c >8%. As Dr. Hirsch and other demonstrated in a February 2022 publication, people with an A1c >8% who spent more time in relative hypoglycemia in the hospital experienced greater mortality. Dr. Hirsch added that an A1c >8% may lead to a new glucose setpoint, such that rapid exposure to euglycemia in the hospital results in tissue pathology, which has been seen in diabetic retinopathy.   
• Dr. Hirsch also suggested that the protection from in-hospital hyperglycemia could be due to insulin-independent brain regulation of glucose, though it’s unclear exactly how this may occur. Dr. Hirsch described two models for how the brain can regulate glucose levels independent of insulin in a January 2023 Diabetes Care article: “Brain Glucose Sensing and the Problem of Relative Hypoglycemia.” At high level, he noted that an individual’s basal glucose level is predominately regulated by insulin-independent mechanisms of glucose disposal. Additionally, in mouse models of diabetes, euglycemia can be induced by targeting the brain via mechanisms that are insulin independent. Dr. Hirsch also pointed out that obesity and diabetes in humans are associated with brain pathology, particularly in areas involved in glucose homeostasis, which may contribute to hyperglycemia. Ultimately, Dr. Hirsch said that while GLUT transporters or the central nervous system may play roles in the etiology of these in-hospital mortality findings, many questions remain and an RCT with CGM is needed to see if these findings can be replicated.   

­--by Katie Lingen, Talia Pikounis, Elaine Young, Mahima Chillakanti, April Hopcroft, Ashwin Chetty, Claire Holleman, Albert Cai, Armaan Nallicheri, and Kelly Close