September 14-19, 2014: Vienna, Austria – Full Report – Novel Therapies and Additional Topics

Executive Highlights

Though much of our coverage was focused on clinical research, EASD featured a number of notable lectures and presentations related to basic science and novel therapies. Three award lectures – by Dr. Anna Gloyn (University of Oxford, Oxford, UK) on the use of genetics in understanding diabetes pathogenesis, by Dr. Domenico Accili (Columbia University, New York, NY) on mechanisms of beta cell failure, and by Dr. Steven Kahn (University of Washington, Seattle, WA) on how clinical findings regarding beta cell failure can inform new studies in the lab – were some of the most well-attended sessions of the conference (a standing-room-only overflow hall was necessary to accommodate the audience for Dr. Accili’s opening keynote address). We also saw results from the TEDDY study (a large study on the environmental determinants of type 1 diabetes) demonstrating that early probiotic exposure is associated with reduced risk of islet autoimmunity; we heard a great deal about this study at the recent JDRF workshop on the early stages of type 1 diabetes and eagerly await more results in the coming years.

The conference also featured several presentations on additional topics including hypoglycemia, treatment adherence, and cost that did not fit neatly into our other categories but that we believe offered valuable insights on higher-level and less discussed topics in diabetes. One notable highlight was Dr. Jay Skyler’s (University of Miami, Miami, FL) comprehensive overview of his three ideal therapeutic goals in type 1 diabetes – beta cell replacement or regeneration, stopping immune destruction, and preservation of beta cell mass – and his proposed combination approach to treatment. Results from an impressively large (n=27,585) global study of self-reported hypoglycemia were also very illuminating: the investigators found that estimated rates of hypoglycemia were higher than previously reported, that underreporting was very common, and that (counter to conventional wisdom) there was no correlation between A1c and hypoglycemia.

Below, we include detailed reports on these topics and more. Presentation titles highlighted in blue represent talks that were not mentioned in our daily highlight reports during the conference, and those highlighted in yellow represent presentations that we found particularly notable.

Novel Therapies and Additional Topics

Oral Presentation: Metformin – New Insights into An Old Drug

Intestinal Glucose Uptake is Modulated by Metformin

Joost B. L. Hoekstra (University of Amsterdam, The Netherlands)

Dr. Joost Hoekstra presented a retrospective analysis of the determinants of high metabolic activity in the colon. As background, Dr. Hoekstra noted that metformin has previously been associated with increased metabolic activity in the colon, implying that the organ is a potential target for the treatment of obesity and type 2 diabetes. To investigate this hypothesis, the study reviewed 270 primary diagnostic positron emission tomography–computed tomography (PET-CT) scans to assess ¹⁸F-FDG (a marker of metabolic activity) uptake. Forward logistic regression analyses of maximum ¹⁸F-FDG uptake suggested patients could be grouped into four cohorts: grade 1 (lowest ¹⁸F-FDG uptake), 2, 3, and 4 (highest ¹⁸F-FDG uptake). Results indicated that patients with grade 4 uptake were significantly more likely to have type 2 diabetes relative to all other groups (p<0.05). Metformin was also determined to be the sole predictor of increased ¹⁸F-FDG uptake in every part of the colon when analyzed individually (see table below; one exception – sulfonylureas derivatives were a predictor of high ¹⁸F-FDG uptake in the ileum). Dr. Hoekstra concluded that the colon could be a promising target for therapeutics though we hold reservations, especially considering the fact that 46% of all patients had a high ¹⁸F-FDG, but only 12% were on metformin. We came away with the notion that much research remains necessary to elucidate this correlation, but salute Dr. Hoekstra for taking the first steps toward this goal.

• Metformin was largely the only predictor of increased ¹⁸F-FDG uptake in all part of the colon when analyzed individually:

Table 1: Predictors of high ¹⁸F-FDG uptake (only significant correlations shown)

• Further study is needed to confirm this correlation, as we noted that 46% of all patients had a high ¹⁸F-FDG, but only 12% were on metformin. During Q&A, an attendee asked Dr. Hoekstra to explain this disparity; to our disappointment, he did not directly answer the question.
  
  • Dr. Hoekstra also presented data indicating a slight but significant positive correlation between BMI and ¹⁸F-FDG uptake. This relationship introduces potential bias, as patients are more likely to be on metformin if they have progressed further toward diabetes, an endpoint associated with higher BMIs. Data on patient characteristics (e.g., BMI of patients on metformin) would have been particularly valuable in elucidating this question.

Questions and Answers

Q: How was ¹⁸F-FDG was administered? To the gut lumen?

A: All patients were administered ¹⁸F-FDG intravenously approximately one hour before the scan?

Q: You are measuring uptake from the plasma, not the gut. Given metformin’s inhibition of glucose uptake, what you are seeing may not represent the true glucose flux into gut.

A: That’s true.

Q: Does the level of plasma glucose affect the rates of uptake and did you correct for that?

A: There is a relationship between plasma glucose and ¹⁸F-FDG uptake. We did not correct for that. It is a good suggestion.

Q: Were these scans done in the fasting state or the postprandial state? Could food intake been seen as a competitor of ¹⁸F-FDG? Finally, ~15% of your patients had diabetes, while 12% were on metformin -- Were you able to detect independently the affect of metformin on ¹⁸F-FDG?

A: The patients were all fasted. We don’t know the effect of food, but it would probably influence it. We did not dissect the independent affect of metformin vs. diabetes, but that is a good suggestion. Most subjects with diabetes are on metformin, and you lose a lot of subject when you take that apart.

Q: You separated BMI and diabetes. It’s difficult to determine what numbers are associated with which. I think you need to make cuts within the data. Since 48% had high uptake, but only 12% were on metformin, what was the cause of the remainder of people with high uptake?

A: We couldn’t make any other conclusions. I don’t know what other factors could influence it.

Oral Presentations: Novel Compounds on the Horizon

Self-Reported Hypoglycemia: A Global Study of 24 Countries with 27,585 Insulin-Treated Patients with Diabetes: The HAT Study

K. Khunti, S. Alsifri, R. Aronson, M. Cigrovski Berković, C. Enters-Weijnen, T. Forsén, G. Galstyan, P. Geelhoed-Duijvestijn, M. Goldfracht, R. Kapur10, N. Lalic, B. Ludvik, E. Moberg,U. Pedersen-Bjergaard, A. Ramachandran

This large study examined the prevalence of hypoglycemia in an impressive 27,585 insulin-treated type 1 (n=8,022) and type 2 patients (n=19,563) in 24 countries. The experimental design included a six-month retrospective period and a one-month prospective period; patient diaries were used to assist recall and record hypoglycemic events. Severe hypoglycemia used the ADA definition, “Requiring third party assistance.” The poster had three important findings: (i) estimated rates of hypoglycemia were higher than previously reported; (ii) incidence rates rose in the prospective period, indicating significant underreporting of hypoglycemia; and (iii) there was no correlation between A1c and hypoglycemia (in line with data from the T1D Exchange, and running counter to the conventional wisdom from DCCT). The tables below summarize the hypoglycemia prevalence (% of patients) and incidence (events per patient-year) – we were particularly struck by the self-reported prevalence of severe hypoglycemia, which occurred in 27% of type 1 patients and 16% of type 2 patients in the six-month retrospective period, and 14% of type 1 patients and 9% of type 2 patients in the four-week prospective period. This translated to 2-5 severe hypoglycemia events per patient-year in type 1 and 0.9-2.5 events per patient-year in type 2. Overall, 83% of type 1s and ~49% of type 2s experienced one or more hypoglycemia events in a four-week period. Of course, the self-reported nature of the study is a potential limitation, but we still believe the data underscore a very critical point: hypoglycemia – both moderate and severe – still occurs far, far too often in type 1 and type 2 diabetes.

Table 1: Hypoglycemia prevalence (% of patients) in retrospective and prospective periods

*six-month period; **four-week period

Table 2: Hypoglycemia incidence (events per patient-year) in retrospective and prospective periods

*six-month period; **four-week period

Postprandial Effects of the Phosphodiesterase-5 (PDE-5) Inhibitor Tadalafil in Type 2 Diabetes Patients: A Randomized Controlled Trial

Lovisa Sjögren, MD, PhD (University of Gothenburg, Gothenburg, Sweden)

Dr. Lovisa Sjögren presented results from a study evaluating whether the PDE-5 inhibitor tadalafil could improve postprandial hyperglycemia and reduce the levels of proinflammatory markers in patients with type 2 diabetes. PDE-5 inhibitors are commonly used to treat hypertension and erectile dysfunction (tadalafil goes by the trade name Cialis for ED), and it has been hypothesized that PDE-5 inhibitors could help mitigate the endothelial dysfunction and impaired postprandial microvascular response associated with type 2 diabetes and obesity. This trial randomized 26 patients with type 2 diabetes to receive either a 20 mg dose of tadalafil or placebo following an overnight fast and 30 minutes before eating a mixed meal. Treatment with tadalafil did not lead to any significant differences in plasma glucose or insulin levels; the tadalafil group did see increases in forearm blood flow and capillary recruitment as well as lower levels of endothelin-1, but the differences were not statistically significant. In a post hoc analysis that excluded patients being treated with ACE inhibitors, there were significant differences in capillary recruitment, forearm glucose uptake, and endothelin-1 levels, but there were still no significant differences in plasma glucose or insulin levels. During Q&A, Dr. Sjögren indicated that she would like to conduct a longer-term trial to see if any more meaningful effects emerge, as she has seen examples of improvements in glycemic control with chronic use of tadalafil.

Questions and Answers

Q: Do you have a hypothesis as to why leaving out patients on ACE inhibitors changed the results?

A: In the literature, there are many reports of patients on ACE inhibitors having hyperglycemic events at high levels, but I don’t know the reason why.

Q: I don’t know the PK effect of PDE-5 inhibitors, but why did you choose a 30-minute interval before the study? Were you confident in the appropriate action of the drug?

A: It reaches peak value after two hours. It can take different amounts of time for different patients, so we had to choose the time we thought was good.

Q: Can you speculate on what longer-term use would do for the results, or will you do a longer trial?

A: That’s the dream question. Tadalafil has been given chronically and has shown positive effects. We’re planning a chronic study in our lab, and we’re on our way to funding and enrolling it. I would speculate, since in studies testing endothelial function, people have shown beneficial effects and suppression of endothelin-1 levels, that it is involved in insulin resistance.

A Novel Chemically Modified Analogue of Xenin-25 Exhibits Improved Glucose-Lowering and INsulin-Releasing Actions

Victor Gault, PhD (University of Ulster, Londonderry, UK)

Dr. Victor Gault presented preclinical data suggesting that a modified analog of the GI peptide xenin-25 can resist enzymatic degradation and lead to improved effects on insulin secretion and plasma glucose compared to the native peptide. Xenin-25 is a small peptide secreted by a subset of K cells in the mucosal lining of the GI tract that has been shown to stimulate insulin secretion, lower blood glucose, and reduce food intake in animal models; it is also known to be secreted in response to meals in humans. However (much like native GLP-1), its biological actions are short-lived due to enzymatic degradation, limiting its usefulness as a target for drug development. Dr. Gault presented results from several experiments using a novel xenin-25 analog with two amino acid substitutions developed in his lab; in vitro, the analog successfully resisted enzymatic degradation for up to 24 hours, increased insulin secretion from mouse beta cells 1.5 to 2.9-fold in a concentration-dependent fashion at blood glucose levels of 5.6 mmol/l (100.8 mg/dl) and 16.7 mmol/l (300.6 mg/dl), and augmented the insulinotropic effects of GIP. In vivo, treatment with the analog significantly reduced plasma glucose concentrations (56% vs. placebo and 38% vs. native xenin-25) and increased insulin secretion (3.4-fold vs. placebo) in mice fed with a high-fat diet. Notably, the analog also led to a 40% reduction in plasma glucose when it was administered eight hours before a glucose load in mice on a normal diet, and it reduced food intake by 31-39% up to 60 minutes after administration. Dr. Gault concluded from these results that this novel analog is more enzymatically stable and has a more effective action profile compared to native xenin-25 and that it would be a promising target for further research.

Questions and Answers

Q: You showed an effect at 5.6 mmol/l – was the effect glucose-dependent?

A: We’re currently performing these studies, and we’re looking at effects on islets as well.

Q: Did you have a chance to look at signaling pathways downstream of xenin?

A: No xenin receptor has been identified to date. We’re currently doing studies; we know that it might bind to an angiotensin receptor and at least one other, but we’re looking to confirm that data. I can’t speculate on what pathways might be affected.

Q: I think this is similar to the GLP-1 story, so it would be logical to know more about degradation. Is DPP-4 involved like it is with GLP-1? If so, what happens with xenin when it’s treated with DPP-4?

A: We believe it’s not due to DPP-4 because we’ve looked at that. There might be other enzymes that cleave the peptide. We haven’t looked at the concentration with DPP-4 because we don’t think that would be of use. We’re working on a good assay to measure xenin concentrations.

Q: Did you do any studies in humans?

A: Only in mice.

Q: Do you know whether the reduced food intake could be due to nausea?

A: Our studies were acute studies. We’ve just finished a long-term sub-chronic study but that data is not finalized. But when we injected it up to 28 days, there was no observable nausea.

Q: Was there data on weight loss in the new studies?

A: Yes; we didn’t see any noticeable effects.

Oral Presentations: Autoimmune Diabetes

Probiotic Use in Infancy and Islet Autoimmunity in the Environmental Determinants of Diabetes in the Young (TEDDY) Study

Ulla Uusitalo, PhD (University of South Florida, Tampa, FL)

Dr. Ulla Uusitalo presented results from the TEDDY study (a large study on the environmental determinants of type 1 diabetes) demonstrating that early probiotic exposure is associated with reduced risk of islet autoimmunity. This study (n=8,676) included children carrying type 1 diabetes associated HLA-DR-DQ alleles from three US centers and three European centers. Data was collected through questionnaires about both the mother and child and through monthly blood samples. Dr. Uusitalo showed that probiotic use was most common in Finland and Germany, with increasing trends in the US and Sweden. Approximately 22% of children in TEDDY received probiotics during the first year of life and 66% received them before the age of three months. Such use was also shown to be associated with diarrhea and antibiotic medication. Using the Cox proportional hazard model to analyze the statistics, Dr. Uusitalo concluded that probiotic use before the age of three months was associated with a 33% decrease in the risk of islet autoimmunity. Regarding areas for future research, she suggested further exploring factors that modify this association, the effect of different duration or dosage, and whether specific probiotic strains are more favorable than others.

Questions and Answers

Q: I wonder whether you have taken into account allergy status. In Sweden, people generally take probiotics for allergies. Have you corrected for that?

A: No, not here. But we’re planning on looking at other factors that are associated with probiotic use.

Q: One of the problems of probiotics is that they’re often not standardized – probiotics in different countries are not the same. Did you look at the composition of the probiotics and if they are similar?

A: It seems that they have similar composition, but we didn’t look specifically at structure or quantity.

Q: What was the definition of taking probiotics?

A: We only looked at probiotics from supplements or infant formula. No food was included here. 

Oral Presentations: Pharmacogenetics and Disease Progression

HbA1c Trajectories in Type 2 Diabetes Patients: The Diabetes Care System Cohort

Giel Nijpels, MD, PhD (Diabetes Care System West-Friesland, Hoorn, the Netherlands)

A modeling analysis presented by Dr. Giel Nijpels investigated the characteristics of type 2 patients in poor glycemic control. The study gathered A1c data from a cohort of 5,423 individuals treated in the Diabetes Care System West-Friesland over nine years, grouping these patients into four groups: (i) good glycemic control (average A1c < 7.0%; ~83% of population), (ii) fast responders (average A1c < 7.0% within two years of initiating treatment; ~8% of population), (iii) reduced glycemic control (A1c > 7.0%; ~5%), and (iv) non-responders (A1c did not respond to treatment; ~3%). The groups were compared using multinomial logistic regression analyses and, unsurprisingly, the subgroups with inferior glycemic control were characterized by a higher A1c at baseline. Dr. Nijpels also shared that non-responders and patients with poorer glycemic control reported a longer duration of diabetes at baseline (>1 year) and, surprisingly, were younger (<60 years old) than those who achieved good glycemic control. These characteristics seem incongruous, though Dr. Nijpels did not offer perspective on this paradox – this does appear to be a unique population of patients (~90% achieved A1c < 7.0%), so we might speculate that those not achieve glycemic goals are outliers to begin with. Ultimately, we are curious about the novel findings, but would like more data regarding patient characteristics to provide context.

Questions and Answers

Q: Did you take into account whether patients were adhering to their treatment?

A: We did not take that into account. However, we know that more than 95% take their anti-diabetic medications. It is slightly lower for statins.

Q: You found that the majority of patients responded to treatment. Perhaps the 10% that didn’t simply didn’t take their medications?

A: I can see what you are getting at. However, as the boss of this cohort, I am telling you that they all took their medications.

Q: Did you look at clinical characteristics at baseline?

A: Yes. That was what we tried to do.

Q: It’s pretty incredible that 90% of patients achieved an A1c < 7%. Do you want to set that achievement in context? Why are the patients doing so well?

A: It has not always been like that. Of course we have special area. General practitioners in the Netherlands are very aware of diabetes. We have a fairly strict system. We are fairly close to the patients. It’s a very special area. This study showed that it is really possible to treat most patients on target with very simple medication. We don’t need other medications that are very costly and, frankly, don’t have that many benefits.

Posters: Lifestyle and Delivery of Care

The Burden of “Serial Non-Adherence” in Patients with Type 2 Diabetes (Poster 1054)

C Frois, K Dea, D Ling, J Dunn, M Baron

A coalition of researchers affiliated with the consulting firm Analysis Group, the pharmacy VRx Pharmacy, and Intarcia (developing phase 3 ITCA 650 exenatide mini-pump) found that 58% of people initiating a second-line diabetes drug are not adherent to therapy, and that over 25% of these not-adherent patients are “serially non-adherent” (defined as a person being repeatedly non-adherent to multiple medications [for diabetes and other diseases] over time). Unsurprisingly, being “serially non-adherent” was a significant predictor for being non-adherent in the future, and non-adherence was associated with a higher risk of hypoglycemia and elevated medical costs. The authors reasonably suggest that strategies that can curb non-adherence early in the type 2 diabetes natural history may be particularly effective in improving the long-term outcomes and finances of people with diabetes. Given the high adherence rates inherent with ITCA 650 (since it is subcutaneously placed for up to one year delivery of exenatide), we think these findings lend support to the use of ITCA 650 in people with more recently diagnosed type 2 diabetes, particularly if they find adherence to other options challenging. Combining this suggestive evidence with the strikingly positive phase 3 results for ITCA 650 in people with poorly controlled type 2 diabetes (baseline A1c of 10-12%), we are seeing ITCA 650 as a strong therapeutic candidate throughout the natural history of type 2 diabetes.

• The retrospective study included 46,789 people who initiated a second-line anti-diabetic therapy and had at least 12 months of follow-up data recorded in the Truven Health MarketScan Commercial Claims and Encounters database, one of the largest claims-based databases in the US. More specifically, the study looked at treatment-naïve patients who initially started a first-line anti-diabetic therapy containing metformin, and subsequently added a second-line therapy. Patients under that age of 18 years were excluded. Medication adherence was measured by the proportion of days covered (the total number of non-overlapping days of medication supply one year following initiation, divided by 365 days). We note that this adherence measure does not account for people filling a prescription, and subsequently not taking the medication, and therefore likely underestimates non-adherence.
  
  • The Truven Health Market Scan Commercial Claims and Encounters database includes individual-level medical and drug information of enrollees in health plans offered by US employers. We note that this database therefore does not include older Americans on Medicare, low income/disabled Americans on Medicaid, or veterans. It includes about four million people with type 2 diabetes between 1Q08 and 4Q12.
• Fifty-eight percent of people were non-adherent to their prescribed anti-diabetic therapy during the year following the initiation of a second line therapy. The best predictors of non-adherence to second-line therapy were measures of prior adherence status, lead by serial non-adherence (OR: 4.67; 95% CI: 4.37-5.00). Serial non-adherence accounted for 65% of the variance in second-line adherence. The second and third best predictors were first-line non-adherence (OR: 2.29; 95% CI: 2.13-2.47) and non-adherence to baseline non-diabetic medication (OR: 1.38; 95% CI: 1.28-1.50).
  
  • Demographically, those non-adherent to the second-line therapy were on average four years younger (54 vs. 58 years), and were more likely to be female and residents of the southern US.

Table: Prior adherence among patient not adherent to second-line diabetes therapy

• In line with prior research, in this study, non-adherence to second-line therapy was associated with higher rates of hypoglycemia and higher medical costs. Non-adherence to second-line therapy was associated with a 20% increase in the rate of inpatient, ED, or outpatient visits for hypoglycemia (2.9% vs. 3.5%; p<0.001), and a 53% increase in the rate of visits for severe hypoglycemia (0.6% vs. 0.9%; p<0.001). Turning to the financial burden of non-adherence, average annual medical costs were 30% (or $2,432; p<0.001) higher for people who were non-adherent as compared to those who were adherent. These higher costs were largely driven by 70% (or $1,678; p<0.001) higher hospitalization costs – a finding that is also in line with prior research showing that inpatient care is a key factor in rising diabetes-related medical costs.  We think that these findings could greatly influence payers, as they assess the cost-effectiveness of user-friendly therapies.

Table: Medical Care Costs for the First Year After Initiation of Second-Line Diabetes Therapy

• In addition to the study limitations noted above (i.e., the study not including data on Medicare or Medicaid patients and not accounting for people not taking medications they fill), another one of its limitations is that it does not adjust for A1c. Furthermore, the study did not follow patients beyond one-year of second-line therapy initiation and therefore does not assess long-term non-adherence nor its burden on health and finances.

Posters: Pragmatic Prediction and Prevention of Type 2 Diabetes

Can Delaying Onset of Type 2 Diabetes be Cost-Effective? (Poster 273)

A Gray, J Leal, S Reed, O Rivero-Arias, K Schulman, R Califf, R Holman

Using NAVIGATOR trial patient data and the UKPDS Outcomes Model v1.3 (a computer simulation model for forecasting the first occurrence of major diabetes-related complications and death in people with type 2 diabetes), Dr. Alastair Gray et al. found that simulated hypothetical interventions to delay type 2 diabetes onset by one to seven years could potentially be cost-effective. Dr. Gray et al. considered hypothetical interventions with assigned annual costs of $75, $750, or $2,250 per person (for context, the National DPP costs $275-$325 per person). Logically, the study found that the longer type 2 diabetes onset was delayed by a hypothetic intervention, the more cost-effective said intervention would be. In line with prior research into the drivers of diabetes-related medical costs, Dr. Gray et al. found that an intervention’s cost-effectiveness would be driven by its ability to avoid or delay diabetes-related complications (as opposed to the cost of treating diabetes itself). The cost per QALY gained for the hypothetical interventions assessed ranged from almost $0 in the US (for an intervention costing $75 per person, per year and delaying type 2 diabetes onset by one to seven year) to ~$16,000 in the US (for an intervention costing $2,250 per person, per year and delaying type 2 diabetes onset by one year). Dr. Gray et al. considered interventions to be “cost-saving or highly cost-effective.”

• Curiously, the study did not appear to find a very large difference in cost-effectiveness between interventions of the annual cost but variable effectiveness (i.e., delaying type 2 diabetes onset by seven years vs. one year). For example, while the cost per QALY of an intervention costing $2,500 and delaying onset by one year was found to be ~$16,000 in the US; the cost per QALY for an equally expensive intervention delaying onset by seven years was found to be ~$14,500. The poster did not explain this trend.
• The study found the cost-effectiveness of interventions with equal costs and equal effectiveness to be greater when implemented in the UK than in the US. The poster authors did not explain what drove this difference in cost per QALY.

Award Lecture: 49th Minkowski Lecture

Unravelling Causal Mechanisms in Diabetes Pathogenesis

Anna L. Gloyn (University of Oxford, Oxford, UK)

Dr. Anna Gloyn discussed the utility of genetics in the study of disease mechanisms and the generation of more clinically valuable data on predictive biomarkers and on new therapies. She drew heavily from genome-wide association studies (GWAS) that have identified risk alleles common among type 2 patients, emphasizing that these studies have informed our understanding of disease pathophysiology. Dr. Gloyn highlighted studies of the P446L variant (which codes for the GKRP protein) that indicated the gene was linked to both reduced glucose levels (and therefore, reduced type 2 diabetes risk) and increased levels of triglycerides – the latter finding stemmed from a large-scale genetic analysis of patients with rare accumulations of GKRP who were found to be at risk for developing hypertriglyceridemia. In this sense, she emphasized that genetics is able to inform the effects of target perturbation on eventual phenotypes and inform us of the potentially harmful side effects of manipulations. Ultimately, Dr. Gloyn asserted that genetics are a useful tool for unraveling question about diabetes, both identifying effective targets and adverse side effects. In our view, this area of research is a somewhat underappreciated and we salute Dr. Gloyn’s effort to raise the level of conversation regarding its promise.

Award Lecture: 46th Claude Bernard Lecture

The New Biology of Diabetes

Domenico Accili, MD (Columbia University, New York, NY)

A standing-room-only overflow hall was necessary to accommodate the ~5,000 individuals who attended Dr. Domenico Accili’s stirring opening keynote address on the new biology of diabetes. In a presentation that echoed his award lecture at this year’s ICE/ENDO, Dr. Accili focused on the underlying mechanism of beta cell failure in the pathophysiology of type 2 diabetes, describing how underexpression of the Fox01 gene leads to defective beta cells with impaired metabolic flexibility (meaning the cells are unable to properly choose between glucose and lipids as a fuel source). He highlighted that this beta cell failure stems from de-differentiation as opposed to apoptosis; he shared data from human studies in which people with type 2 diabetes were shown to have the same total number of endocrine islet cells, but have a smaller percentage that produce insulin (5-10%) than those without the disease (~40%). Last, Dr. Accili provided a unique look at type 1 “cure” research, highlighting that the genetic removal of the Fox01 gene from endocrine cells (i.e., gut cells) has been shown to result in insulin-producing cells. Dr. Accili harbored cautious optimism regarding these findings, though he emphasized numerous advantages of gut insulin-producing cells over embryonic stem cell-derived beta cells in potential “cure” therapies.

• “This is nuts!” exclaimed Dr. Accili, referring to findings that human gut cells can be transformed into insulin-producing cells. In particular, these studies inhibited the Fox01 gene in human “gut-oids” yielding functional insulin-producing cells. In mice, similar experiments have demonstrated that animals with defective beta cells can spontaneously recover from hyperglycemia. Among the many advantages of developing gut insulin-producing cells over embryonic-derived beta-cells, Dr. Accili highlighted possible protection from autoimmune attack.

Award Lecture: 8^th Albert Renold Lecture

The Beta Cell in Type 2 Diabetes: Lessons Starting at the Bedside

Steven Kahn, MB, ChB (University of Washington, Seattle, WA)

In his lecture for the Merck-sponsored Albert Renold Award, Dr. Steven Kahn reversed the bench-to-bedside directionality usually cited in research, instead discussing how clinical findings on the loss of beta cell function in type 2 diabetes can inform new studies in the lab. He began by emphasizing the importance of the decline in beta cell function in type 2 diabetes: in the ADOPT study (for which Dr. Kahn was a lead investigator), glyburide was the most effective agent in terms of A1c-lowering at six months, followed by metformin and then rosiglitazone. However, at four years, the efficacy trends were completely reversed, as glyburide’s efficacy had deteriorated sharply with increased beta cell burnout, while rosiglitazone’s efficacy held remarkably steady. In his conclusion, Dr. Kahn suggested that amyloid polypeptide could be a driver of beta-cell apoptosis, and highlighted the early study currently being conducted in this area.

Symposium: Risks and Benefits of New Diabetes Treatments

Hypoglycemia: Importance of the Cut-Off Point

Stephanie Amiel, MD (King’s College London, London, UK)

Dr. Stephanie Amiel focused on the need to modify the ADA’s hypoglycemia guidelines, as she believes that using 70 mg/dl (3.9 mmol/l) as the hypoglycemia definition overestimates the incidence of clinically important hypoglycemia. To most comprehensively gauge a clinically relevant definition, Dr. Amiel pointed out the need to understand which glucose concentrations are associated with the effects of hypoglycemia, the glucose threshold that predicts subsequent severe hypoglycemia, as well as the impact of glucose measurement technology on all of these factors. Currently, the definition of hypoglycemia is simply the lower boundary of a target range; Dr. Amiel believes that these two set points can and should be different. She demonstrated that 70 mg/dl does not predict many clinically important consequences, whereas concentrations of 65 mg/dl (3.5 mmol/l) and 54 mg/dl (3.0 mmol/l) are more closely associated with consequences such as cognitive impairment and symptomatic stress responses. Therefore, Dr. Amiel proposed compromising with the ADA guidelines by using 70 mg/dl as an “alert value” for impending hypoglycemia but using a lower limit between 65 mg/dl and 54 mg/dl as a pre-prandial glucose target to define hypoglycemia.

• While Dr. Amiel acknowledged that 70 mg/dl yields a reduction in insulin secretion and the stimulation of glucagon in normal physiology, she argued that this concentration is too high to relate to clinically important consequences. She noted that 70 mg/dl is still within normal range and that insulin and glucagon is irrelevant in insulin deficient diabetes. In addition, it does not affect surrogates of symptoms, increases prevalence of impaired awareness, and is a poor predictor of subsequent severe hypoglycemia.
• Dr. Amiel reviewed the advantages and disadvantages of using 65 mg/dl and 54 mg/dl as the hypoglycemia definitions. She concluded 65 mg/dl as a lower limit of normality that can define hypoglycemia and that 54 mg/dl should be recorded as it is associated with important clinical consequences.
  
  • In favor of 65 mg/dl, she explained that this concentration is the lower limit of normal glucose and is associated with the beginning of symptomatic stress responses. However, 65 mg/dl is not associated with cognitive impairment.
  • Therefore, she pushed for also recording 54 mg/dl since it is associated with cognitive impairment and gives prevalence of impaired awareness that better matches clinical experience. This concentration also better predicts severe hypoglycemia and its avoidance leads to restored hypoglycemia awareness.

Questions and Answers

Q: For a patient who’s not diabetic, and especially a woman, who is complaining of postprandial hypoglycemia and you do a prolonged GTT, what should we define as postprandial hypoglycemia?

A:  This is not really relevant to the discussions of hypoglycemia in diabetes but obviously very important. OGTT is not a good way to elicit postprandial hypoglycemia because a lot of the time, you will see biochemical hypoglycemia that is not even defined as less than 3.5 mmol/l (63 mg/dl) and it’s sometimes going down below 3 mmol/l (54 mg/dl) in completely asymptomatic people. I wouldn’t wish to make a diagnosis on that basis. I think for the diagnosis of reactive postprandial hypoglycemia, you have to look at symptomatology. As we were told earlier in this meeting, one of the problems is that those people lack a greater symptomatic awareness response to their hypoglycemia. As for glucose less than 3.3 mmol/l (59 mg/dl), if we want to be really technical, glucose less than 3.5 mmol/l is hypoglycemia in the circumstance you’re describing. I think you have to make clinical judgment.

Q: The point that really gets confused is one of terminology. I think it’s very important to differentiate between asymptomatic and hypoglycemia unawareness. People can have glucose levels that are low and also have severe hypoglycemia. They are aware that they are low and then they go on and have a severe hypoglycemia episode. This is very different from someone with low glucose levels and has a severe hypoglycemia episode without warning. I think there’s confusion between these terms, which is a very dangerous thing.

A: I think the preferred term actually now for condition of what you just referred to as hypoglycemia unawareness is impaired awareness of hypoglycemia. I didn’t make the point clear and thank you for bringing it up, but impaired awareness of hypoglycemia is a condition for the patient and asymptomatic hypoglycemia is a hypoglycemia episode which they have no symptoms. It’s really important to recognize that under 3 mmol/l, we can detect prolongation of reaction time in every setting we’ve looked at. The patient who says that I feel fine with concentrations of around 2 mmol/l (36 mg/dl), and it’s not often that I say this, but in that case, the patient is wrong and the person who knows that is the patient’s family.

Q: What level do you use when you’re removing a driver’s license? Do you use someone with current episodes with 3.9 mmol/l (70 mg/dl)? Or do you wait until it drops down until 3 mmol/l?

A: Happily in the UK, I never remove any driver’s licenses as that decision happily lies in the driver’s licensing authority’s hands. But that would just avoid your question. I worry that people, especially non-specialists, will apply a diagnosis of hypoglycemia less than a 3.9 mmol/l and they will say that the patients have hypoglycemia unawareness and write to the licensing authorities and licenses would be removed. That is one of our worries of this use of 3.9 as a cutoff. We would argue because of the evidence that if you don’t feel hypoglycemic until under 3 mmol/l, it does seem as a group, to protect against hypoglycemia. But happily at least in the UK, the licensing authority only removes licenses if the patient is completely unaware. If the patient doesn’t know that they are hypoglycemic in trying to treat it, to me, they’re completely unaware. Patients who recognize hypoglycemia robustly at 3 mmol/l has impaired awareness, but I would not describe it as completely unaware. So again, you do have to apply clinical judgment. But to me, the important issue is that we need to agree on two things: (i) what is not hypoglycemia and I would argue that 3.8 or 3.9 mmol/l is not; and (ii) what is important hypoglycemia for us to know about in assessing treatment and to that, I would say less than 3 mmol/l.

Q: Some physicians have said that they have seen symptoms suggestive of hypoglycemia but when blood glucose is measured, they are above 3.9 mmol/l. This often happens in my own experience and then their blood glucose is suddenly reduced. What is your treatment for this?

A: In 2005, the ADA referred to that as relative hypoglycemia and now it’s called pseudohypoglycemia, which I find an unkind term. So in my experience, that is usually the situation for people with chronically poor glycemic control who are used to a much higher blood glucose level and then they feel hypoglycemic. They do indeed have a stress response at what is frankly a normal or slightly increased glucose level. The only thing you can do is to explain to them and gradually get them through it so they can start achieving better glycemic control assuming that that’s appropriate at their clinical stage.

Symposium: East-West Forum at EASD 2014

Characteristics of Diabetes and its Complications in Japan

Hirohito Sone, MD, PhD (Niigata University, Niigata, Japan)

Dr. Hirohito Sone presented follow-up data from the Japan Diabetes Complications Study (JDCS), which evaluated the impact of lifestyle intervention on clinical parameters and complications in 2,033 Japanese patients with type 2 diabetes from 1995-1996, to highlight key differences in the manifestation of diabetes between Japanese and Western populations. He explained that there is much less of an association between diabetes and obesity in Japan than in the West – JDCS participants had a very low average BMI of 23.1 kg/m² – and that insulin secretion appears to be less correlated with insulin resistance in many East Asian patients. With regard to complications, while diabetes increases the relative risk of cardiovascular disease in both Japanese and Western populations, absolute event rates are much lower in Japan. Indeed, cardiovascular disease was not the leading cause of mortality in the JDCS cohort, accounting for only 20% of the deaths in the study vs. 34% that were attributed to cancer (this could also be attributable to high-quality care for patients that experience non-fatal CV events). Based on this data, along with an analysis showing differences in the relative impact of risk factors like LDL and triglycerides on cardiovascular risk in JDCS patients compared to Western populations, Dr. Sone concluded that the long-term risk profile for Japanese patients with diabetes is likely very different compared to that for patients in Europe or the United States, and that risk calculators based on data from Western studies like the UKPDS may not be applicable to Japanese patients.

Questions and Answers

Q: Basically, you said the risk of dying of cancer is 50% higher than dying from cardiovascular disease. Was there any impact on malignancy from lifestyle intervention?

A: We didn’t analyze the cause of death in relation to the intervention, but it was analyzed in relation to exercise. There was a relatively lower incidence of cancer in patients who exercised the most. It’s already evident in the general population that exercise can lower cancer risk, and the same thing happened in our diabetic population.

Q: There turned out to be such a strong association between triglycerides and cardiovascular disease in the Japanese patients. Was it fasting triglycerides?

A: Yes.

Q: What are your thoughts on the differences in risk compared to other populations? Why was the association so strong?

A: We still haven’t perfectly analyzed or understood that. One hypothesis is that triglyceride levels are a representative marker of visceral obesity. As you know, in Japanese/East Asian patients, even slight visceral obesity can affect a lot in terms of cardiovascular disease. Another hypothesis is the effect of alcohol. We lack an enzyme to metabolize alcohol, so we tend to have higher risks in that area. But still, this is all speculation.

Symposium: Face Diabetes – Therapeutic Strategies (Sponsored by Österreichische Diabetes Gesellschaft)

Mortality Trends in Type 1 Diabetes – Progress Made, More to be Done

Marietta Stadler, MD (King’s College London, London, UK)

Dr. Marietta Stadler discussed the trends in mortality in type 1 diabetes, examining mortality rates, causes of death, and “non-classical” risk factors. Regarding the rates of mortality, she highlighted that rates are generally declining but continue to be two to six times higher than the background population. Notably, she presented data that showed that females have much higher standardized mortality ratios (SMR of 13 vs. 5 in the US). In addition, mortality rates are higher in people diagnosed with type 1 diabetes after 15 years of age. Looking at causes of death, more recently diagnosed cohorts appear to suffer from greater acute but fewer chronic complications. Diabetic ketoacidosis and hypoglycemia remain as the most common adverse events; however, mental-health-related deaths are on the rise, perhaps a consequence of patients making it to greater and greater ages. After acknowledging that “classical” risk factors for mortality are typically glycemic control, kidney function, and cardiovascular risk factors, Dr. Stadler explored the more “non-classical” risk factors. She emphasized that severe hypoglycemia, social deprivation, and depression are also all associated with mortality and should be researched more extensively, and we would certainly agree on all fronts.

Questions and Answers

Q: Why is there more death in females?

A: I wish I knew the answer. I think it is partly a calculation problem. Social factors may also be involved.

Hospital Inpatient Care – Improvements and Needs

Thomas Pieber, MD (Medical University of Graz, Graz, Austria)

Dr. Thomas Pieber introduced the GlucoTab Hopsital Care System, a decision support system, as a method of improving the quality of diabetes care in the hospital inpatient setting. He opened by highlighting how hyperglycemia in hospitals is a risk indicator for poor clinical outcomes and that there are very few randomized control trials on diabetes care in hospitals, outside of ICU settings. Dr. Pieber reviewed the many challenges of routine care including factors influencing insulin requirements, provider time limitations, the lack of a standardized workflow, as well as insufficient knowledge (missing diabetes training and teams). Regarding recommendations to improve this care management, he explained his team’s work on the GlucoTab Hospital Care System, a decision support and workflow management system. The system consists of a tablet computer that automatically suggests the correct insulin dosage based on basal-bolus therapy. Dr. Pieber then presented data from an open, single-center controlled trial comparing between decision support vs. standard care that showed that decision support led to better glycemic control and fewer hypoglycemic events. Concluding, Dr. Pieber emphasized that decision support and workflow management systems have the potential to improve quality of diabetes care. Inpatient diabetes management is an area with significant room for improvement, and given the limitations in provider time along with the shift towards technology in healthcare, we see tools such as GlucoTab as a great way to streamline and standardize diabetes care in this challenging setting.

Questions and Answers

Q: Do you think you could use this in patients’ daily lives as well?

A: This device is mostly designed for diabetologists. We’re actually thinking about using this in nursing homes because there is even less diabetes knowledge there. We could move into home care but more data is needed for that.

Corporate Symposium: Addressing Diabetes Challenges Across the Continuum of Care (Sponsored by Sanofi)

What Does the Future Hold for Diabetes Management?

Jay Skyler, MD (University of Miami, FL)

Dr. Jay Skyler provided a comprehensive research overview of the three ideal therapeutic goals in type 1 diabetes: beta cell replacement or regeneration, stopping immune destruction, and preservation of beta cell mass. Most interesting were his positive thoughts on using combination therapy, as evidenced by the ATG/GCSF findings presented at ADA 2014. He concluded with a proposed combination therapy that included seven different classes of therapies, aimed at targeting different defects in type 1 diabetes and administered at different time points.

• To conclude his presentation, Dr. Skyler presented a schematic of a potential comprehensive combination approach to treating type 1 diabetes. The slide is summarized in the table below, with time moving from left to right. Certain therapies would be administered over time (Anti-IL1B or Anti-TNF; Oral Insulin; GLP-1 Receptor Agonist), while others would be administered at single time points (Anti CD3 or Anti-CD20 or Co-Stimulation Blockade; GAD; IL2 or GCSF; T-regs).

• Dr. Skyler highlighted the “remarkable” ATG/GCSF combo pilot study presented at ADA 2014. At one year, the combination therapy maintained beta cell function in patients with recent onset type 1 diabetes (four months to two years) – the treatment arm’s two-hour C-peptide level was preserved over the 12 months at ~2 ng/ml (baseline was 2.14 ng/ml). In contrast, the placebo group experienced a significant C-peptide decline to ~1 ng/ml (p-value for the difference between the two groups = 0.05). Dr. Skyler noted that this is one of first combination therapies tested, “and it works.” Notably, neither of the two components worked when it was used alone. As a result, the study questioned the regulatory paradigm that in order to combine therapies, the individual constituents must show an effect alone. Dr. Skyler said there are plans in place to take this study and do it in patients within three months of diagnosis.
• Sources of islet cells are currently a challenge, though many approaches have completed animal testing and are moving into human trials. Transdifferentiation, which takes patients’ own liver cells and differentiates them into beta cells is now moving into human studies. Viacyte has also recently entered a clinical trial. Meanwhile, Dr. Camillo Ricordi and colleagues are testing transplantation scaffolds, which encapsulate cells in a pouch in the omentum. They are protected against auto-immunity and can include oxygen generation components and localized drug delivery.
• “Human T1D Prevention studies have had very limited impact.” Dr. Skyler briefly summarized findings from a number of trials:

• There are currently four prevention studies in TrialNet: Anti-CD3, Abatacept, Oral insulin, and a treatment TBD (the desired treatment is no longer available).

Corporate Symposium: Modern Type 2 Diabetes Management – The Experts’ Guide to the Universe of Choices (Sponsored by AstraZeneca)

Glucagon Suppression in Type 2 Diabetes: Is It Important?

Daniel Drucker, MD (Lunenfeld-Tanenbaum Research Institute, Toronto, Ontario, Canada)

Diabetologist extraordinaire Dr. Daniel Drucker discussed the somewhat overlooked contribution of glucagon to the pathophysiology of type 2 diabetes, explaining that despite the availability of drug classes like GLP-1 agonists and DPP-4 inhibitors that suppress elevated glucagon production, the field still lacks a comprehensive understanding of the mechanisms underlying glucagon dysregulation. He pointed out that there are significant differences in pancreatic islet physiology between species, limiting the applicability of findings from animal studies to clinical practice. He also presented data demonstrating that abnormal glucagon production is seen very early in the progression of type 2 diabetes – even patients with mildly impaired glucose tolerance have noticeable defects in both insulin and glucagon secretion. Dr. Drucker encouraged the use of diabetes drug classes that suppress glucagon secretion and consequently reduce hepatic glucose production, and he characterized glucagon receptor antagonists as an “extremely attractive new approach” to type 2 diabetes pharmacotherapy. However, he also acknowledged that impressive A1c reductions are possible even with an increase in glucagon and endogenous glucose production, as is evident in patients treated with SGLT-2 inhibitors or who have undergone bariatric surgery. He ultimately concluded that suppressing elevated glucagon production is one attractive option for addressing hyperglycemia in type 2 diabetes, but that far more research is needed to fully understand glucagon’s role in the pathophysiology of the disease.

Questions and Answers

Q: Do we know anything about the relationship between sulfonylureas and glucagon?

A: There’s not much data in that area. We tend to think that agents that target the beta cell should indirectly suppress glucagon, but the data showing potent suppression is more limited than we would like.

Q: Does glucagon have any effect on renal glucose production?

A: In animals, you can show under some circumstances that intestinal and renal glucose production rise in response to glucagon. In animal models of bariatric surgery, in preclinical studies, there’s a hint that it is regulated by glucagon. It’s very difficult in humans to show that renal glucose production is regulated by glucagon or that it contributes significantly to glycemia.

Q: Based on the bariatric surgery data, should I treat my bariatric surgery patients with a DPP-4 inhibitor?

A: There are several studies looking at that. It makes sense that since GLP-1 and GIP are elevated after bariatric surgery, so we might get a disproportionately wonderful benefit. We don’t have rigorous randomized controlled trials yet; it’s a great idea but we need evidence.

-- by Melissa An, Adam Brown, Varun Iyengar, Emily Regier, Manu Venkat, and Kelly Close

-- The authors thank Eric Chang, Hannah Deming, Jessica Dong, Nina Ran, and Melissa Tjota for additional help on conference writing and editing