International Diabetes Federation: World Diabetes Congress 2015

November 30-December 4, 2015; Vancouver, Canada; Full Report – Diabetes Drugs – Draft

Executive Highlights

This report includes our coverage of diabetes drugs from IDF 2015. Our themes are enclosed immediately below, followed by detailed discussion and commentary. Titles highlighted in blue are new additions that were not included in our daily updates, and those highlighted in yellow represent what we felt were the most notable talks of the meeting.

Table of Contents 


  • We heard plenty of commentary on cardiovascular outcomes trials (CVOTs) at IDF 2015, including discussions on the EMPA-REG OUTCOME results, other recently-reported or soon-to-report trials, and broader CVOT policy. Lilly/BI continued to highlight the cardioprotective benefit associated with Jardiance (empagliflozin) in a corporate symposium on the first day. During the symposium, Dr. Sanjay Kaul (Cedars-Sinai Medical Center, Los Angeles, CA) spoke from a regulatory perspective and argued that empagliflozin’s risk/benefit profile justifies a Class I recommendation (meaning the benefit greatly outweighs the risk and the treatment should be administered). In the same symposium, Drs. Silvio Inzucchi (Yale University, New Haven, CT) and Eugene Wright (Cape Fear Valley Health Affairs, Fayetteville, NC) noted in a panel discussion that the results are convincing enough to change their clinical practice, but noted the limitations of the trial in terms of minority representation and suggested that professional organizations may want to wait for the other SGLT-2 inhibitor CVOTs to report before changing treatment guidelines.
    • Part of this hesitation likely comes from the uncertainty behind the mechanism driving the cardioprotective benefit. Indeed, while acknowledging their strong validity, Dr. Kaul also stressed the need for replication of the results given their “unexpected and unprecedented” nature. In a separate session, nephrologist Dr. David Cherney (University of Toronto, Canada) pointed to the differences between empagliflozin and thiazide diuretics to suggest that sustained, reduced plasma volume associated with empagliflozin use could be behind the benefit. Most of the speculation we have heard about the mechanism of benefit has centered around empagliflozin’s volume effects, as its effects on glucose, weight, and blood pressure appear unlikely to be responsible. That said, we are still very partial to Dr. Kaul’s statement from EASD that “when you see ‘multifactorial’ or ‘multidimensional’ in a journal like NEJM, it’s usually a euphemism for we don’t know.”
    • While the EMPA-REG OUTCOME results proved that it is possible to demonstrate CV superiority under the current CVOT paradigm, some speakers urged caution about extrapolating the results to other upcoming trials. Professor Neil Poulter (Imperial College London, UK) in particular cautioned that, despite very positive pooled MACE analyses from pre-approval trials of Novo Nordisk’s Victoza (liraglutide), we should not get our hopes up too high that LEADER will demonstrate a cardioprotective benefit. While EMPA-REG OUTCOME has made us more optimistic that other SGLT-2 inhibitor trials will have positive results, it does not change our overall assumption that most of these trials are likely to be neutral. While there is certainly reason to believe that GLP-1 agonists could be cardioprotective due to their favorable effects on weight, blood pressure, and lipids, a benefit mediated through these mechanisms may not become apparent in relatively short trials that enroll very high-risk patients.
    • The divergent heart failure results in the three DPP-4 inhibitor CVOTs remained a source of much consternation and discussion. Dr. Stefano Del Prato (University of Pisa, Italy) suggested that head-to-head studies of DPP-4 inhibitors are needed to better understand the increased risk for heart failure seen with AZ’s Onglyza (saxagliptin), though Dr. Daniel Drucker (Lunenfeld-Tanenbaum Research Institute, Toronto, Canada) pointed out that the very rare nature of this adverse event makes the mechanisms behind it very difficult to study. Dr. Rury Holman (University of Oxford, UK) presented a meta-analysis of the three DPP-4 inhibitor CVOTs (SAVOR-TIMI, EXAMINE, TECOS) that demonstrated no statistically significant difference in the risk of hospitalization for heart failure for the class – this was largely driven by the neutral results from TECOS, the largest of the three studies. In the Absent absence of head-to-head studies or a clear mechanistic explanation of the SAVOR results, we assume the most likely practical consequence of these results is that clinicians will opt for Merck’s Januvia (sitagliptin) over Onglyza for patients at high risk of heart failure. Of course, Jardiance or other SGLT-2 inhibitors may become the favored option for these patients in any case following the EMPA-REG OUTCOME results.  
    • Several speakers proposed broader changes to how CVOTs are conducted. Dr. Eberhard Standl (Diabetes Research Group e.V. at Helmholtz Zentrum München, Munich, Germany) suggested that a large number of participants in CVOTs have undiagnosed heart failure. He argued strongly that future CVOTs should capture all cases of heart failure at baseline with BNP measurements to more carefully investigate the issue. Dr. Julio Rosenstock (UT Southwestern Medical Center, Dallas, TX) called for a closer look at the adjudication process for the pancreatitis results of ELIXA (the CVOT for Sanofi’s Lyxumia [lixisenatide]). Dr. David Matthews (Oxford Center for Diabetes, UK) argued that CVOTs may not be the most appropriate approach to answering questions in a huge epidemic disease such as diabetes, considering that they only enroll a tiny fraction of the affected population, only take into account the mean or median outcome, and may be too short to be useful. On the other hand, Dr. David Nathan (Harvard University, Boston, MA) argued that CVOTs still have a place since we don’t have adequate biomarkers for macrovascular risk in diabetes. While the criticism of the CVOT paradigm on the conference circuit has been muted to some extent post-EMPA-REG OUTCOME, these statements suggest that there is still plenty of room for improvement to ensure that these trials are as cost-effective and broadly valuable as possible.
  • We heard several speakers point to the cost of insulin as a major factor affecting clinical decision recommendations. Dr. Matthew Riddle (Oregon Health and Science University, Portland, OR) highlighted the 10-factor price differential between human insulins and newer insulin analogs and shared that some patients in his clinic are going back to human insulin due to cost issues. As a result, he suggested that, of the many newly available insulin formulations (“next-generation” basal insulins, concentrated insulin, inhaled insulin, and biosimilars), biosimilar insulin has the most potential to benefit the greatest proportion of patients. He argued that only a rather small percentage of patients will benefit from the relatively incremental efficacy or dosage improvements with the other new formulations, but upwards of 50% of patients would benefit from presumably lower-cost biosimilars. Going one step further, Dr. David Beran (Geneva University Hospitals, Switzerland) encouraged patients and health systems to move away from analogs and even insulin pens and switch back to human insulin via syringe delivery. He cited the comparatively high cost of insulin analogs and pens and the WHO’s position that analogs offer no clinical advantages beyond human insulin to support his point. While human insulin and syringes may have a lower upfront cost, we do not see a wholesale return to human insulin as an ideal solution due to the significantly increased risk of hypoglycemia and potential for lower patient adherence.
    • The issue of insulin cost in general is quite complex – while the list price of insulin has certainly gone up in recent years, rebating has also increased as pharmaceutical companies wrestle with pharmacy benefits managers over exclusive formulary contracts. Thus, it’s hard to say to what degree insulin costs have changed for payers and insured patients. Uninsured and under-insured patients are most directly impacted by higher list prices, though many companies do offer patient assistance programs for those in this situation. In addition, the manufacturing process for biologics like insulin is much more complex and expensive than small molecule oral diabetes drugs. As a result, even biosimilars are not likely to offer enormous discounts (Lilly/BI’s biosimilar insulin glargine Abasaglar is discounted at 15%-20% relative to Lantus in the European markets where it has launched, while small molecule generics are typically priced at a 50%-80% discount relative to their brand-name counterparts). A recent New York Times op-ed noted that it is 100 times more expensive to reverse-engineer a biologic than a small molecule – up to $200 million vs. about $2 million. Thus, it’s clear that insulin manufacturers are facing a variety of pressures and that more innovative, multidisciplinary solutions that engage industry, payers, providers, patients, and policymakers will be needed to reverse the current trend in insulin prices.
  • New data from ACCORDION and ADVANCE-ON reinforced the clear legacy effect of intensive glycemic control on microvascular outcomes and the lack of clarity around the connection to macrovascular outcomes. Dr. Hertzel Gerstein (McMaster University, Hamilton, Ontario, Canada) presented primary results from ACCORDION (observational follow-up to ACCORD; n=8,601) showing a neutral effect of previous intensive control on cardiovascular outcomes after a median of almost nine years of follow-up (HR = 0.95 for the primary MACE outcome; 95% CI: 0.87-1.04; p=0.27). Importantly, the increase in mortality that caused ACCORD to be stopped early resolved quickly after the end of randomized treatment, though there was still a significant difference in CV mortality by the end of ACCORDION. The most positive news from ACCORDION came from a retinopathy sub-study (n=2,856) showing a significant legacy effect of intensive control on retinopathy progression after eight years (HR = 0.42; 95% CI: 0.28-0.63; p<0.0001). We also saw a new analysis from ADVANCE-ON (observational follow-up to ADVANCE) showing that the legacy effect of intensive control on renal outcomes was greater for patients with lower systolic blood pressure and less advanced kidney disease at baseline.
    • Taken together, the two studies support the clear benefits of even a short period of intensive control on microvascular outcomes. We feel it is crucial for the field not to lose sight of this in the midst of the worthwhile debate over the relationship between glycemic control and macrovascular outcomes. Complications like blindness and end-stage renal disease are incredibly important from a patient quality-of-life perspective, and a treatment approach that reduces the risk of those complications while having a neutral effect on CVD could still be considered a success.
    • The results only add more fuel to the debate on the macrovascular outcomes side. Based on the totality of the evidence (from DCCT/EDIC, UKPDS, ACCORD, ADVANCE, and VADT), there appears to be a hint of a relationship between glycemic control and macrovascular outcomes, but it likely takes many years to emerge and is weaker than the link for factors like blood pressure and LDL cholesterol. Unfortunately, we do not expect to gain much more clarity on this subject in the near future given the expense and difficulty of these studies. We also do not anticipate a clear resolution of the most worrisome finding from these trials: the increased mortality in the intensive group in ACCORD. In his presentation, Dr. Gerstein stated that no explanation has emerged for that signal despite years of extensive analysis. We would speculate that the answer most likely lies in the fact that the patients who were treated intensively and did not achieve target were the ones at higher risk. This is supported by an analysis of SMBG data presented at ADA 2011 showing that mortality in both groups was linked with divergence from glycemic targets.
  • The concept of individualizing therapy – and the debate over how best to do so – remains a major theme in the diabetes drug arena. We were glad to see individualized care emphasized as a major priority in the new NICE guidelines for type 2 diabetes published during the conference. Of course, as was frequently acknowledged in a session on clinical practice guidelines, the real challenge lies in translating a statement like “individualized care” into effective clinical practice. Some of those dirty details were evident during an AZ-sponsored symposium in which panelists were asked how they would treat their own type 2 diabetes. While oral combination therapy seemed to be the consensus favorite approach, there was quite a range of views on the best specific agents, the benefits of initial combination therapy vs. aggressive sequential therapy, and the role of lifestyle intervention. We found it interesting that while lifestyle intervention often receives only a cursory mention in most type 2 diabetes drug-focused talks, a majority of panelists suggested that it would be a cornerstone of their own treatment plans. Farther down the treatment algorithm, Dr. John Buse (University of North Carolina, Chapel Hill, NC) raised several questions about the best place for GLP-1 agonist/basal insulin combinations vs. their individual components or basal-bolus insulin regimens. As with oral therapies, the answer is likely different for different patients. Both discussions illustrate the extent to which clinical judgment and the “art of medicine” often determine which patients are actually treated with which drugs. This will certainly continue for the foreseeable future given the wide range of (imperfect) choices in type 2 diabetes and the lack of comparative effectiveness research.
    • Another dimension of personalization is the question of how generalizable clinical trial results are to the larger patient population. This came up several times in the context of EMPA-REG OUTCOME, as several speakers cautioned against over-generalizing the results beyond the high-risk population enrolled in the trial but also acknowledged that the findings have already had an impact on their clinical practice. As one attendee put it, it is unlikely that there will ever be a properly powered outcomes trial of Lilly/BI’s Jardiance (empagliflozin) in lower-risk patients, and “are we really going to say you need to have a heart attack first?” before prescribing it.
  • PCSK9 inhibitors generated quite a buzz as a potential therapy (albeit an expensive one) for the many patients with diabetes who are not at their LDL cholesterol targets. The first two PCSK9 inhibitors – Sanofi’s Praluent (alirocumab) and Amgen’s Repatha (evolocumab) – were approved this past summer. While their indications are fairly narrow at the moment (individuals with familial hypercholesterolemia or clinical atherosclerotic cardiovascular disease [ASCVD] on maximally-tolerated statin therapy), the indicated population nonetheless includes a significant number of patients with diabetes. Dr. James McKenney (National Clinical Research, Richmond, VA) emphasized that both PCSK9 inhibitors are functionally the same, but the two companies chose to emphasize different options with their dosage strategies, with Sanofi emphasizing flexibility in efficacy and Amgen emphasizing flexibility in dosing schedule. Both Dr. McKenney and Dr. Harry Ginsberg (Columbia University College of Physicians and Surgeons, New York, NY) expressed excitement over the phase 1 RNA interference (RNAi)-based methods of PCSK9 inhibition that may support twice-yearly dosing. They also both noted that PCSK9 inhibitors can – and should – be used on top of statins, with additive (though not synergistic) effects. Dr. Christoph Wanner (University Hospital of Würzburg, Germany) emphasized that more and more individuals with diabetes and/or obesity are entering PCSK9 trials, suggesting that the drugs are appropriate for many within the diabetes patient population. Indeed, in a comparison, Dr. Wanner found the participant population of one of the phase 3 trials for Praluent to be quite similar to the participant population of EMPA-REG OUTCOME. If the ongoing cardiovascular outcomes trials for PCSK9 inhibitors produce positive results and lead to much broader indications, we would expect their relevance for patients with diabetes to increase even further.

Detailed Discussion and Commentary

Symposium: Cardiovascular Outcome Studies and Glucose-Lowering Therapies


Rury Holman, FMedSci (University of Oxford, UK)

Dr. Rury Holman presented an update on TECOS safety data, suggesting that the hint of renal dysfunction observed in sitagliptin-treated patients is “hardly of clinical significance.” He reviewed and expanded on the previously published renal findings: (i) the marginally improved mean urine albumin to creatinine ratio (UACR) values seen with sitagliptin compared with placebo (-1.0 mg/g) in the subset of 3,669 patients with data available; and (ii) the marginally worsened mean estimated glomerular filtration rate (eGFR) values seen with sitagliptin than with placebo (-1.3 ml/minute/1.73 m2) for the entire population. Dr. Holman suggested that both of these minor differences could be linked to the difference in A1c between the groups (glycemic equipoise was not quite achieved). To exclude the possibility that the small difference in mean eGFR values may mask larger differences in patients with reduced renal function at baseline, Dr. Holman shared new subanalyses indicating that the minor differences seen in mean eGFR values were similar across all chronic kidney disease (CKD) categories. Coupled with the fact that the changes were fairly weak to begin with, Dr. Holman concluded that it is unlikely that sitagliptin has a clinically significant impact on renal function. We fully concur. While in the aftermath of EMPA-REG, this information on renal risk is not the most compelling in terms of CVOT findings, we believe that it could still be somewhat useful in putting providers at ease when individualizing second-line therapies for patients with a history of chronic renal disease.

  • Dr. Holman also revisited the trial’s heart failure data to provide further clarity on the question of a DPP-4 inhibitor class effect; the study’s preliminary answer was reassuring with a neutral hazard ratio of 1.00. Dr. Holman stressed that further analyses of recurrent hospitalization for heart failure are just as encouraging, resulting in an unchanged hazard ratio of 1.00 after accounting for first+recurrent hospitalizations along with no evidence that patients in the sitagliptin arm had a differential outcome to heart failure (e.g., cardiovascular and all-cause morality in patients with one or more hospitalization for heart failure events did not differ). Notably, Dr. Holman also presented a meta-analysis of the three DPP-4 inhibitor CVOTs (SAVOR, EXAMINE, TECOS), demonstrating no statistically difference in the risk of hospitalization for heart failure when the studies are pooled (an overall hazard ratio of 1.14 [p=0.178]). He stressed the need to interpret these results with caution, echoing the lingering uncertainty around the mechanism behind the differential effects and suggesting that the increased heart failure result seen with saxagliptin may be due to chance, its trial’s patient population, adjudication process, or even a drug-specific effect.


Julio Rosenstock, MD (UT Southwestern Medical Center, Dallas, TX)

In this presentation, Dr. Julio Rosenstock reviewed the study design and outcomes of ELIXA, the CVOT for Sanofi’s Lyxumia (lixisenatide) – see full results (which demonstrated neutrality regarding CV risk) from our coverage at this past ADA and the recent publication in NEJM. Notably, when discussing the trial’s pancreatitis results, Dr. Rosenstock called for the need to more closely investigate the adjudication process for a more nuanced assessment of this safety signal. As a reminder, the results of ELIXA showed no significant increase in pancreatitis (8 events vs. 5 events). However, Dr. Rosenstock expressed skepticism of how the adjudication of these events was completed, noting that the process is based on only “classical criteria.” He highlighted that he would also like to see the “mild cases” reported, which are likely not captured through this adjudication process; specifically, he pointed to the value of identifying participants with higher lipase levels even if they do not technically represent a form of pancreatitis. Ultimately, we feel that acute pancreatitis is being increasingly acknowledged as a side effect for incretin-based therapies – although an acceptably small one – and a more granular analysis of this nature can help dissect the magnitude of the signal – we feel the tradeoff in using them is positive given how much benefit they give patients and given that the risk can be managed.

Questions and Answers

Q: I’m skeptical about this study since the efficacy was so modest – the weight reduction and blood reduction were very modest. Would you agree that the increase of heart rate is related to the very low dosage? Similarly, the neutrality around hospitalization due to heart failure may also be because dosage is so low. Would we see different results with higher doses?

A: First of all, 85% of people reached the standard dose of 20 micrograms. One of the important things of the study is that we’re testing the drug for CV events; we’re not testing for the impact on blood glucose. Its true efficacy is not that impressive but the intent is to get equipoise between the two groups. The placebo group was also receiving much more medication for glucose control. Also, I don’t think this drug’s purpose is to lose weight. This is something we’ll see and I look forward to other trials to see if weight loss can be maintained over two to three years. In terms of heart rate, here we didn’t really see an increase in heart rate. I don’t think this CVOT needs to be assessed for glucose control because the study design is to get equipoise between two groups.

Q: Can you speculate on the heart rate aspect a little bit since the measure was taken one day after drug was given the last time? The drug has a half-life of three to four years. Can you really expect to find anything at this point?

A: There may be long-term effects of GLP-1 – maybe through the autonomic system or another mechanism, but we don’t know if it’s a direct action on the heart or if it’s related to the autonomic system. Part of Juris Meier’s study showed that it was clear that the increase of heart rate was sustained over 24 hours with liraglutide but over a shorter time with lixisenatide. It was maintained overnight without a dip of heart rate. Whether this has any implications on CV outcomes, we don’t know.

Cardiovascular Outcomes with Empagliflozin: Results from EMPA-REG OUTCOME

Bernard Zinman, MD (Mount Sinai Hospital, University of Toronto, Ontario, Canada)

Dr. Bernard Zinman reviewed the EMPA-REG OUTCOME results and offered some speculation on the potential mechanism of benefit. He stressed while explaining baseline characteristics that the trial participants had longstanding diabetes and received “what most of us would accept is state of the art” treatment for all CV risk factors. He also emphasized that the trial aimed to maintain glycemic equipoise – this seems to be one of the main points of confusion with this and other CVOTs for diabetes drugs, as some have interpreted the lack of an A1c difference between groups as a sign that the drug is not effective. Discussing the results, Dr. Zinman stressed that the dramatic effects on CV death and heart failure occurred very early in the trial. He expressed particular excitement about the significant reduction in all-cause mortality, which he described as the “holy grail” for any cardiovascular study. He also provided interesting granularity on how cardiovascular death is adjudicated in these trials, noting that a large portion of that category consists of deaths without a clear cause – “if you go to sleep and wake up dead, that’s called a cardiovascular death.” While this is standard practice in CVOTs, that lack of clarity around the cause of death could make investigation of the mechanism of benefit more complicated. After noting that he would “leave it to someone else” to refer to EMPA-REG OUTCOME as a landmark study, Dr. Zinman closed by naming a number of factors that could have contributed to the benefit, including sympathetic tone, weight and visceral fat reduction, oxidative stress, and others in addition to volume change. However, during Q&A he seemed to agree with the prevailing hypothesis that a hemodynamic effect was most likely responsible.

Questions and Answers

Q: Why not do a study on people with heart failure or heart conditions without diabetes?

A: That’s an excellent point. I’m sure various companies are thinking about that. It may be as effective in people with heart failure without diabetes. You’ll still get some glucosuria and an osmotic effect.

Q: If there’s not much difference between the 10 mg and 25 mg doses, why not use only the 10 mg dose?

A: That’s something the NEJM reviewers insisted we comment on. As far as cardiovascular outcomes go, there was no difference. If the metabolic targets aren’t met with the 10 mg dose we have the option of going to 25 mg, but I mostly use 10 mg.

Q: What do I tell my patients who ask why I didn’t give them this drug?

A: You didn’t know the results. Now that you do, you should give it.

Q: To all people with diabetes?

A: No, it should be people who meet the entry criteria for the trial, then you can extrapolate to some extent.

Q: Many patients don’t meet the inclusion criteria but have subclinical ischemia or heart failure. Could you consider earlier use?

A: Many patients have type 2 diabetes with underlying heart failure. DECLARE has a large subset that is basically primary prevention, so we’ll know more. For now it’s left to the individual.

Q: There was an impressive nephroprotective effect. Do you believe that was related or independent?

A: The relationship between renal and cardiovascular outcomes is an interesting concept. We will do a mediation analysis to see what relates to what. We know intra-glomerular pressure is reduced, which makes ACE inhibitors work much better. The development of doubling of serum creatinine was reduced nicely and progression from microalbuminuria to macroalbuminuria was reduced significantly. We don’t know the relationships but it’s nice to see it happening.

Q: It’s a landmark study to me, but I’m not sure I agree with the conclusion. Based on the statin trials I don’t think the mechanism was atherosclerotic, since there was no effect on MI and stroke.

A: I agree 100%. We chose MACE because that’s what the FDA wants. This effect was not atherosclerosis, it’s a hemodynamic effect. Eventually there may be benefits, but the immediate response occurred so quickly that it has to be a hemodynamic response, not an atherosclerotic effect.

Q: Could it be a direct effect on the heart?

A: It could be. It has to be studied. There are no SGLT-2 receptors on the heart.


Sophia Zoungas, MD, PhD (University of Sydney, Australia)

Dr. Sophia Zoungas first reviewed the primary results from ADVANCE-ON, the long-term observational follow-up of the ADVANCE study. Those results, presented at EASD 2014, found a significant legacy effect of glucose lowering on end-stage kidney disease (HR = 0.54; 95% CI: 0.34-0.85) but no impact on macrovascular endpoints or retinopathy. This latest analysis demonstrated that the effect on renal death was not significant, though the point estimate did go in the right direction (HR = 0.85; 95% CI: 0.45-1.62). It also found that the benefits on end-stage kidney disease were greater for patients with lower systolic blood pressure (<140 mmHg) and less advanced kidney disease (no CKD or CKD stage 1/2) at baseline. Importantly, unlike in ACCORD, there was no evidence of increased harm with intensive glucose control in patients with any stage of CKD at baseline. The relative risk of severe hypoglycemia was also comparable across CKD stages, though the absolute risk was greater with more advanced disease. As Dr. Zoungas noted, these results highlight the benefits of achieving intensive control earlier in disease progression. It also suggests that glucose-independent mechanisms of nephropathy may predominate over glucose-dependent mechanisms in later stages of the disease. As for why the overall results were less impressive than those from UKPDS or VADT, Dr. Zoungas pointed to lower baseline A1c (7.5% in this trial), an insufficient difference in A1c during the randomized period, or an insufficient trial duration as potential explanations. The point was also made at the original results presentation that UKPDS enrolled a much younger, more recently diagnosed cohort; the participants in ADVANCE may have passed the window in which intensive glucose lowering can have a benefit on outcomes.

Symposium: New Views On Incretins

GLP-1 and Type 1 Diabetes: Sufficient Efficacy and Safety?

Urd Kielgast, MD, PhD (Hvidovre University Hospital, Copenhagen, Denmark)

Dr. Urd Kielgast offered a middling take on GLP-1 agonists in type 1 diabetes: positive on safety and more pessimistic on efficacy. On safety, she concluded that there is no increased hypoglycemia risk, impairment of counterregulatory mechanisms, or other serious or unexpected side effects associated with the class in type 1 diabetes. However, she did note that gastrointestinal side effects are more common and that the long-term implications of the increase in heart rate and reduced systolic blood pressure are unknown at this point (we assume the blood pressure effect would be positive if anything but clearly need long-term trials to be sure). The question of efficacy is more complex. Dr. Kielgast pointed to promising early-stage data showing substantial reductions in postprandial glucose with GLP-1 agonists regardless of beta cell function. The drugs also led to A1c reductions, lowered insulin doses, and weight loss in uncontrolled proof-of-concept studies. As we have unfortunately learned over the past few months, the results from placebo-controlled trials – Novo Nordisk’s phase 3 ADJUNCT ONE and ADJUNCT TWO studies for Victoza (liraglutide) being the largest – were less encouraging. Novo Nordisk subsequently decided not to pursue a type 1 diabetes indication for liraglutide – a disappointing but understandable decision in our view although we do question what sub-group analysis in a very large trials . We imagine other GLP-1 agonist manufacturers will now be more reluctant to pursue type 1 diabetes indications as a result. However, Dr. Kielgast noted that she would like to see studies in new-onset patients and studies of short-acting GLP-1 agonists, which have a greater effect on gastric emptying. We also think benefits like weight loss and potentially reduced glycemic variability could be meaningful for at least some patients, but we are fairly pessimistic that there will be a regulatory path forward in the near future.

  • Dr. Kielgast shared results from an unpublished study (n=12) of liraglutide vs. placebo to support the safety of GLP-1 agonists in type 1 diabetes. The 12-week trial found no difference in gastric emptying rate or glucagon response during hypoglycemia between the liraglutide-treated and placebo-treated groups. There was also no difference in cortisol, growth hormone, adrenaline, or noradrenaline levels.
  • Dr. Kielgast presented results from two placebo-controlled trials suggesting modest efficacy with liraglutide in type 1 diabetes. A randomized, double-blind, placebo-controlled trial in normal-weight patients with type 1 diabetes (n=40) found no significant difference in A1c reduction, glycemic variability, 24-hour glucose profiles, hypoglycemia event rate, or diastolic blood pressure with liraglutide vs. placebo after 12 weeks. However, the liraglutide-treated group did experience a significant 3.1 kg mean weight loss (compared to a 1.1 kg weight gain in the insulin monotherapy group; p<0.0001), a significant reduction in bolus insulin dose (p=0.02), and a reduction in systolic blood pressure (p=0.04). Similarly, the Lira-1 study in patients with type 1 diabetes and obesity observed no improvement in 24-hour glucose profiles or glycemic variability among patients treated with liraglutide. While there was a significant difference in A1c reduction after three months, the effect was lost at six months due to the A1c plateauing in the liraglutide group and continuing to decrease in the insulin-only group. There was a significant difference in body weight reduction (p=0.015) and bolus insulin dose (p=0.02), however.

New Incretin Therapies: Is Longer Acting Always Better?

Jack Leahy, MD (University of Vermont College of Medicine, Burlington, VT)

Dr. Jack Leahy emphasized the differences between short-acting and long-acting GLP-1 agonists – particularly with regards to postprandial glucose control – and identified the patient populations who may benefit most from the less popular short-acting versions. He acknowledged that long-acting GLP-1 agonists (such as Novo Nordisk’s market leader Victoza [liraglutide]) are better in most cases due to their superior A1c-lowering efficacy in head-to-head studies against short-acting products. However, Dr. Leahy highlighted the effect of short-acting GLP-1 agonists (such as Sanofi’s Lyxumia [lixisenatide]) on mealtime glucose, calling the postprandial glucose suppression their “defining feature” and comparing them to native GLP-1. Delving into the potential mechanisms a bit, Dr. Leahy attributed the effect on postprandial glucose to delayed gastric emptying, which disappears with chronic administration of long-acting GLP-1 agonists but persists with short-acting agents. Ultimately, Dr. Leahy concluded that long-acting GLP-1 agonists are best for patients with both fasting and mealtime hyperglycemia, while short-acting versions are best for those who primarily have difficulty with postprandial glucose control. He suggested that patients who are on basal insulin already, are elderly, or eat high-carb diets are possible candidates for short-acting GLP-1 agonist therapy. He also speculated that individuals with prediabetes or early type 1 diabetes may benefit from short-acting GLP-1 agonists as well. We’ve previously heard from several speakers on the potential of short-acting GLP-1 agonists for postprandial glucose control, particularly in combination with basal insulin. This may offer a potential niche for the relatively late-to-market Lyxumia in the US and may be a point of differentiation between the two GLP-1 agonist/basal insulin fixed-ratio combinations Xultophy (insulin degludec/liraglutide) and LixiLan (lixisenatide/insulin glargine).

GLP-1 and Insulin: When and How?

John Buse, MD, PhD (University of North Carolina, Chapel Hill, NC)

Dr. John Buse expressed characteristic excitement about GLP-1 agonist/basal insulin combinations and raised several unresolved questions about their optimal use. He reviewed clinical trial data on Novo Nordisk’s Xultophy (insulin degludec/liraglutide) and Sanofi’s LixiLan (lixisenatide/insulin glargine), emphasizing their “clear dramatic effect” on glycemic control and significant hypoglycemia and weight benefits compared to basal insulin alone. He believes it remains uncertain how the two products will differ in clinical practice. We have generally assumed they will have fairly comparable clinical profiles, but Novo Nordisk indicated in its 3Q15 update that it believes Xultophy will be highly differentiated. Dr. Buse also suggested that GLP-1 agonist/basal insulin combinations (fixed-ratio or otherwise) come out on top in the arguably more relevant comparison to basal-bolus insulin. He noted that while the available evidence has suggested a modest advantage at best in terms of glycemic control, the significant weight and hypoglycemia benefits and the convenience of a single injection are very clinically meaningful. The major remaining questions for Dr. Buse are (i) where these combinations should be positioned in treatment guidelines; (ii) whether adding basal insulin when patients have failed on a GLP-1 agonist is as effective as the reverse; and (iii) whether GLP-1 agonists could be an alternative to basal insulin as an add-on to rapid-acting insulin. He believes the last two questions remain unresolved and require further study. While the answer to the first is complex, Dr. Buse believes that absent concerns about cost and the long-term safety of GLP-1 agonists, “you could make a very strong argument that the combined approach is clearly superior to basal insulin” as an add-on to oral therapy.

Symposium: Long-Term Effect of Intensive Glucose Control in Type 2 Diabetes – The ACCORD International Ongoing (ACCORDION) Study

Cardiovascular Outcomes

Hertzel Gerstein, MD (McMaster University, Hamilton, Ontario, Canada)

Dr. Hertzel Gerstein presented primary results from the ACCORDION study (observational follow-up to ACCORD) showing a neutral effect of previous intensive control on cardiovascular outcomes and an attenuation of the increased mortality risk seen during randomized treatment. ACCORDION followed 8,601 ACCORD participants (90% of those eligible) through up to seven phone calls and clinic visits following the end of the randomized treatment period, for a median total follow-up time of 8.8 years. As expected, the A1c difference between the intensive and standard control arms had disappeared by the end of follow-up, with an average A1c of ~7.7% in both groups by month 57 after the end of ACCORD. There was no difference between groups in the primary MACE composite outcome (CV death, non-fatal MI, non-fatal stroke) 14 years post-randomization (HR = 0.95; 95% CI: 0.87-1.04; p=0.27). Importantly, the increase in mortality that caused ACCORD to be stopped early quickly resolved after the end of randomized treatment, with a hazard ratio of 1.01 (95% CI: 0.92-1.10; p=0.91) after 14 years. There was no heterogeneity for either outcome in any subgroup analyzed. Unfortunately, the significant reduction in non-fatal MI at the end of ACCORD was not maintained through ACCORDION (HR = 0.89; 95% CI: 0.79-1.02; p=0.09). The results for non-fatal stroke were also neutral (HR = 0.87; 95%CI: 0.73-1.04; p=0.11). The increase in CV mortality was still significant (HR = 1.20; 95% CI: 1.03-1.40; p=0.02), but Dr. Gerstein stressed that it was attenuated over time and that there was no increased risk in the post-randomization period – this seemed to be a point of confusion during Q&A.

  • The original ACCORD results published in 2008 demonstrated no effect of intensive vs. standard glucose control on overall CV outcomes but a worrisome increase in mortality. The trial randomized 10,251 patients with type 2 diabetes and high CV risk to intensive (target A1c <6%) or conventional (target A1c 7-7.9%) glucose-lowering therapy. It also evaluated intensive vs. conventional blood pressure- and lipid-lowering therapy in a factorial design. Patients were followed for a mean of 3.7 years before the trial was stopped early due to increased mortality in the intensive group. The intensive group achieved a significantly lower A1c at one year (6.4% vs. 7.5% in the conventional group; baseline = 8.1%) and the difference was maintained throughout the study. The results for the primary composite outcome were neutral (HR = 0.90; 95% CI: 0.78-1.04; p=0.16). The risk of non-fatal MI was lower in the intensive group (HR = 0.76; 95% CI: 0.62-0.92; p=0.004), but there was an increased risk of both CV death (HR = 1.35; 95% CI: 1.04-1.76; p=0.02) and all-cause mortality (HR = 1.22; 95% CI: 1.01-1.46; p=0.04).
  • ACCORDION passively followed 8,601 ACCORD participants for a total median follow-up time of 8.8 years. Dr. Gerstein emphasized that this represents 90% of eligible participants, and 98% of patients who did not experience a MACE event or die during the trial, meaning it should provide a robust estimate of the long-term effect. As one might expect, patients who chose to participate in ACCORDION were somewhat healthier than those who did not (younger age, lower BP, less prior CVD, fewer smokers). Patients were treated at the discretion of their personal providers and followed up with study investigators by phone or in person up to seven times in a 3.5 year period. Outcomes were reported by study investigators, but a random 10% of outcomes were adjudicated for quality control.
  • A1c was comparable between groups by the end of ACCORDION, though a difference persisted for some time after the end of randomized treatment. A1c in the intensive group rose to ~7.4% but remained (non-significantly) below that of the conventional group (~7.8%) at the end of the five-year study period (randomization in the blood pressure and lipid arms was maintained even after the glycemic aspect of the trial was stopped). The difference had dwindled to ~0.2% by the beginning of ACCORDION and was essentially gone by the end of follow-up. The confidence intervals were quite wide at all points during and after the study.  
  • Results for the primary MACE composite and all-cause mortality were completely neutral, with no heterogeneity between subgroups. The hazard ratio for MACE (CV death, non-fatal MI, and non-fatal stroke) was 0.95 (95% CI: 0.87-1.04; p=0.27) 14 years after randomization. The hazard ratio for all-cause mortality at 14 years was 1.01 (95% CI: 0.92-1.10). We would note that the number of patients dropped off considerably after ~10 years. The neutral results held true when patients were stratified by sex, age, race/ethnicity, history of CVD, baseline A1c, blood pressure, and lipids.
  • Results for non-fatal MI and non-fatal stroke were both neutral, while the increase in CV death was attenuated but still significant. The hazard ratio for MI was 0.89 (95% CI: 0.79-1.02; p=0.09) and the hazard ratio for stroke was 0.87 (95% CI: 0.73-1.04; p=0.11). The risk of CV death was still significantly elevated in the intensive group (HR = 1.20; 95% CI: 1.03-1.40; p=0.02), but Dr. Gerstein stressed that this was attenuated over time and that the curves were parallel after the end of randomized treatment. This seemed to be a point of confusion during Q&A, as some questioners assumed this meant there was an increased risk in the post-randomization period alone.
  • Dr. Gerstein stated that no explanation has emerged for the excess mortality in ACCORD despite years of extensive analysis. He noted that the risk was not related to the degree of glucose lowering, cardiac autonomic neuropathy, hypoglycemia, weight change, or specific drugs. It was statistically correlated with neuropathy, aspirin use, and higher baseline A1c, but Dr. Gerstein argued that these are most likely surrogates for sicker patients. He also referenced the recent paper suggesting differences in hemoglobin glycation index as an explanation but argued that it did not provide a conclusive answer. Other than a chance finding (which Dr. Gerstein also raised as a serious possibility), we would speculate that the answer most likely lies in the fact that the patients who were treated intensively and did not achieve target were the ones at higher risk. This would suggest that providers should closely monitor high-risk patients and back off on treatment if they are not responding, but that those who succeed in reaching intensive targets are at no increased risk.

ACCORDION Eye Study Results

Emily Chew, MD (National Eye Institute, Bethesda, MD)

Dr. Emily Chew presented results from the ACCORDION Eye Study (n=1,268) demonstrating a significant legacy effect of intensive glycemic control on retinopathy progression after eight years. The original ACCORD Eye Study (n=2,856) found that a significantly lower percentage of patients in the intensive group experienced progression of retinopathy compared to the conventional group after four years (7.5% vs. 10.4%). Outcomes were also better in patients treated with fenofibrate in addition to statins compared to those treated with statins alone. There was no difference between the intensive and conventional blood pressure arms. ACCORDION aimed to evaluate whether the beneficial effects of intensive glucose control and fenofibrate persisted after the end of randomized treatment and whether blood pressure had more of an effect with long-term follow-up. The legacy effect of intensive glycemic control was clear: 5.8% of patients in the intensive group experienced progression of retinopathy after eight years vs. 12.7% in the standard group (HR = 0.42; 95% CI: 0.28-0.63; p<0.0001). The effect of fenofibrate disappeared after treatment discontinuation, and no significant blood pressure effect emerged. Given the fairly low retention in the study (58% of eligible patients participated), Dr. Chew presented sensitivity analyses bolstering the validity of the results. An analysis restricted to sites with >80% retention also demonstrated a favorable hazard ratio, though the benefit was not significant due to the small numbers. A post-hoc analysis using a composite outcome of retinopathy progression or death (a way to account for competing risks) also demonstrated a significant benefit.

Brain Structure And Function

Anne Murray, MD (Hennepin County Medical Center, Minneapolis, MN)

Geriatrician Dr. Anne Murray presented results from the ACCORDION MIND study demonstrating no significant legacy effect of intensive control on cognitive decline or total brain volume. The original ACCORD MIND study (n=2,977) found no significant difference between groups in change in DSST scores (an assessment of cognitive function) or other cognitive measures. It did find a significantly lower decline in brain volume with intensive glycemic control and standard blood pressure treatment, but it also found significantly higher abnormal white matter volume with intensive control, leaving the overall conclusions unclear. Participants in ACCORDION MIND (n=1,328) underwent a fourth cognitive assessment and a third brain MRI at 80 months post-randomization. Results showed no significant difference in mean change in DSST score (p=0.32), total brain volume (p=0.91), or abnormal white matter volume (p=0.41) between groups at 80 months. Dr. Murray suggested that patients’ long duration of diabetes, the short and transient difference in A1c between groups, healthy survivor bias, and younger age could be possible explanations for the lack of an effect. Our sense is that efforts to understand the link between diabetes and cognitive dysfunction remain at an early stage, and we hope that better mechanistic understanding will inform more targeted studies and therapies in the future.

Panel Discussion

Q: It’s difficult to understand why there is still increased mortality since patients who died already are no longer patients. How can you exclude the complexity of drug treatment in the intensive group as a factor behind the increased mortality?

Dr. Hertzel Gerstein (McMaster University, Hamilton, Ontario, Canada): There was absolutely no long-term mortality effect at all. Mortality from cardiovascular causes persisted but was attenuated. It’s impossible to disentangle the many different drugs and combinations and figure out whether one or the other was responsible for the mortality signal that was originally observed or the persistent CV mortality signal. There have been many analyses looking at specific drugs and combinations and nothing emerged as significant. That doesn’t mean there is no answer. We just can’t find it. A few papers suggested it was maybe a random finding. It was not seen in other studies. We just don’t know.

Q: Were the drugs used in the two arms similar or different?

Dr. Gerstein: The drug cocktail in the intensive group was much more complex. During follow-up it was the same.

Q: Then cardiovascular deaths should also be the same?

Dr. Gerstein: It was attenuated. The curves diverged at two years and stayed exactly the same after. That may be why we see an attenuation of the long-term effect from 35% to 20%.

Q: Should we change practice according to these results that show no big difference in mortality in all studies? Do patients need all these medications we’re giving?

Dr. Gerstein: That’s a philosophical question. We know that ACCORD had a short period of treatment of 3.7 years. We know from all the trials that long-term legacy effects on CV events may take longer. VADT had five years of treatment and at the end there was no effect on CV outcomes. That required another five years. Metabolic interventions probably take longer to have an effect if they have an effect. There is a clear unequivocal eye benefit. During the active period there was also a benefit on brain volume that went away.

Dr. Emily Chew (National Eye Institute, Bethesda, MD): That’s an important point. You can read lots of conclusions on what to do after ACCORD, but the eye findings are important. If you look at the public and ask what senses they value, 70-80% would give anything up for vision. Vision is incredibly important for quality of life. To prevent people from getting an injection in their eyes or laster photocoagulation every month is a big deal. Retinopathy is not to be forgotten. A legacy effect of only 3.7 years is pretty powerful. This is very important for care and quality of life.

Q: It’s reassuring that total mortality was resolved. For CV mortality, during the follow-up did they look at the total time duration or just the follow-up duration?

Dr. Gerstein: All of these trials are analyzed from the day of randomization to the end of follow-up. They look at whether a specific period of intensive control has a ten-year effect on whatever. If you analyzed just the follow-up that’s an epidemiological study. If it’s a long-term effect you’ll see it that way. If it goes away, it’s not a long-term effect.

Q: It’s disappointing that the original signal of improved nonfatal MI was not continued. What does this tell us about the role of intensive therapy on atherosclerosis?

Dr. Gerstein: My personal belief is that the data in total say you maybe need three to five years of intensive intervention to see any long-term effect on outcomes like ischemic heart disease. In ACCORD there was a benefit but it was modest and probably takes a long time. In the DCCT it took a long time.

Q: If we have no mortality difference but still an increase in CV mortality, what is it protecting against?

Dr. Gerstein: There was slightly less non-CV mortality but the numbers change up and down. The difference in numbers is small, 10-20 people. I wouldn’t over-interpret it. You have to look at the totality of the evidence. The ADA got it right with their recommendations.

Symposium: Do We Need New Insulins for the Management of Diabetes?

Why Do We Need New Insulins? An Overview

Philip Home, DM DPhil (Newcastle University, Newcastle upon Tyne, UK)

Dr. Philip Home discussed why we need new insulins and which advances on the horizon are most promising. He believes the main flaw of current insulins is the high variability of action resulting from subcutaneous administration – they cannot achieve the minute-to-minute control of blood glucose that endogenous insulin can. This means that new therapies that provide feedback control, such as smart insulin, closed loop systems, or transplanted islets (including those derived from stem cells), will represent the largest advances. However, he stressed that even the closed loop will have limitations because the glucose sensing and insulin delivery still occur subcutaneously. One therapy Dr. Home believes will not solve the problem is oral insulin, as the poor bioavailability and erratic absorption will make it very difficult to replicate the correct physiological profile in his view. The same constraints apply to insulins with buccal, nasal, or pulmonary absorption. While we believe some patients will see oral insulin as a major success even if it is just able to match the efficacy of current analogs, it sounds like that will be a high bar according to Dr. Home. In terms of more incremental improvements, Dr. Home argued that we have likely optimized basal insulin analogs as much as possible, as Novo Nordisk’s Tresiba (insulin degludec) and Sanofi’s Toujeo (insulin glargine U300) have essentially flat profiles. He believes there is still some room for improvement with mealtime insulin. Products that offer a faster onset of action like Novo Nordisk’s faster aspart (just submitted in the EU and the US), Lilly/Adocia’s BioChaperone Lispro, and Biodel’s BIOD-531 will be improvements over current options, though still unlikely to match physiological control.

New Formulations: How Will They Change Therapy?

Matthew Riddle, MD (Oregon Health and Science University, Portland, OR)

Dr. Matthew Riddle suggested that potentially lower priced biosimilar insulin will benefit more patients than other new insulin formulations for type 2 diabetes. Dr. Riddle estimated that over 50% of patients, for whom cost is a leading concern, could benefit from biosimilar insulin glargine. He shared that he sees some patients coming off insulin analogs and returning to normal human insulin due to cost issues, as he noted the 10-factor price differential between human insulins and newer analogs, which is likely to only go up. As a result, he believes that cost considerations will drive use of the biosimilar version of insulin glargine. We’ve been hearing increasing concerns over the high cost of insulin and it seems that many providers are putting great hope in biosimilar insulin glargine as a potential source lower priced insulin. That said, biosimilar insulin glargine has been priced at only a 15%-20% discount in the markets it has launched in, begging more and bigger questions on how we can resolve this growing challenge (i.e. patents, who’s funding innovation, etc.).

  • In contrast, Dr. Riddle estimated only about 25% of patients would benefit from the longer and flatter profiles of next-generation basal insulins such as Novo Nordisk’s Tresiba (insulin degludec) and Sanofi’s Toujeo (U300 insulin glargine). He argued that while these products can reduce hypoglycemia risk, the advantage over standard of care Lantus (insulin glargine) is not as great as the hypoglycemia risk difference between older insulin analogs and human insulin. Thus, he suggested that since the relative benefit is more incremental (an unsurprising view given that Lantus in particular was a “once a generation” product in terms of improvements for patients, in our view), providers need to know who will get the most benefit. Specifically, he speculated that, based on data showing especially marked improvements in hypoglycemia and body weight with Toujeo vs. Lantus in Japanese populations, people with lower BMIs and lower insulin requirements may benefit the most.
  • Dr. Riddle estimated that 5% of patients would benefit from concentrated insulin and suggested that the main appeal of concentrated insulins such as Lilly’s Humulin U500 is the possibility of fewer injections for individuals with very high insulin doses, which may promote better adherence. More specifically, Dr. Riddle felt that concentrated insulins only made sense for patients with a dosing requirement >80 units per day (at which point it becomes impossible to deliver other forms of insulin with a single injection). Lilly’s increased attention to Humulin U500 over the past year and the launch of the Humalog (insulin lispro) U200 KwikPen in May suggests that the company clearly believes there is a significant market for these products, though we have heard some complaints from HCPs about the price increases for Humulin U500 that have been part of that increased attention.
  • Finally, Dr. Riddle rather pessimistically suggested that inhaled insulin (Afrezza) would only benefit his estimated 1%-10% of patients with true, complete needle phobia. Inhaled insulin’s (Afrezza) main strengths, in Dr. Riddle’s view, include its injection-free delivery mechanism and rapid onset, which he suggested may translate into better adherence and mealtime control. However, he noted that he doesn’t think there are that many patients with complete, true needle phobia. While this is likely true, we have assumed there would be a broader population of patients who would find these advantages meaningful. We hope a continued focus on educating HCPs, improving reimbursement, and reducing administrative hurdles can help reveal how broadly appealing Afrezza’s clinical profile is.

Yes, If We Are to Avoid Hypoglycemia and Mortality

Simon Heller, MD (University of Sheffield, UK)

Dr. Simon Heller delivered his signature talk on the dangers of hypoglycemia, focusing on the connections to CVD and mortality. He opened with a very sobering discussion of the “dead in bed” syndrome, citing the many epidemiological studies showing a link between hypoglycemia and sudden death in people with type 1 diabetes. While the mechanism of this phenomenon is not entirely clear, Dr. Heller suggested that arrhythmias, particularly long QT intervals, could be responsible. Dr. Heller also discussed the ever-controversial question of the increased mortality in the intensive group in ACCORD. He notes that rates of hypoglycemia were higher in the intensive group but cautioned that there are “undoubtedly other things going on.” For example, he suggested that hypoglycemia could function as a marker for overall frailty that can be used to identify people at risk. Dr. Heller then turned to the many potential mechanisms by which severe hypoglycemia can increase the risk of CVD, including blood coagulation abnormalities, inflammation, endothelial dysfunction, and sympathetic activation. He also stressed that the risk seems to be present in type 2 diabetes as well, citing studies that found high rates of asymptomatic nocturnal hypoglycemia in type 2 diabetes that was associated with arrhythmias. Returning to the central topic of the symposium, Dr. Heller argued that better insulin delivery is necessary to overcome these serious risks. While the exact hypoglycemia benefits of the newest insulins are not entirely clear, we are always glad to see the importance of hypoglycemia thrown into such sharp relief, as we fear the issue is sometimes overlooked in the current A1c-centric, cost-conscious environment.

Questions and Answers

Q: We’ve discussed the realities of NPH use in your country and the US because of cost. With the lack of CGM availability for type 2 diabetes, there’s a high incidence of nocturnal hypoglycemia that we presumably don’t identify. Are there any biomarkers you can think of to identify people at cardiac risk?

A: We haven’t done that. My hypothesis is that C-peptide might be a biomarker. We speculated in a previous study that it would be a marker for risk of severe hypoglycemia. The problem is it’s asymptomatic. If we are using NPH or even analog insulin, it would behoove us to persuade our patients to at least once or twice a month set an alarm and measure their glucose at 3 am. There might well be a considerable benefit. It could drive changes in insulin prescriptions if people are having hypoglycemia.

Dr. Irl Hirsch (University of Washington, Seattle, WA): We presented a poster from the T1D Exchange showing that after 40 years of diabetes, quite a few patients were C-peptide positive, and it did protect them from severe hypoglycemia. Your point about the lasting effects of inflammation from hypoglycemia is almost like talking about metabolic memory for hypoglycemia. I’m not exactly sure of the mechanism, but do you agree with the metabolic memory analogy?

A: Clearly studies in newly diagnosed patients have demonstrated that early intensive control produces benefits. That’s clearly indicated from DCCT and UKPDS. If we can get it right early, there are benefits. Whether repeated hypoglycemia – this isn’t severe, this is what all patients experience every day – whether those effects accumulate, I have no idea. It’s a great hypothesis and extremely difficult to test. It would explain the surprising results with intensive therapy.

Novel Ways to Deliver Insulin: Which Routes for Which Patients?

William Cefalu, MD (Pennington Biomedical Research Center, Baton Rouge, LA)

Dr. William Cefalu discussed the challenges facing alternative modes of insulin delivery, including those in the pipeline and recent arrivals to the market. He primarily touched on intranasal, buccal, oral, and inhaled insulin delivery, though he noted that sublingual and intraperitoneal insulin are also being investigated. Dr. Cefalu noted that intranasal and buccal insulins have faced significant bioavailability challenges due to the thicker mucosa and lower surface area of the upper airways. Dr. Cefalu was more optimistic about oral insulin, but noted that there are still significant hurdles to overcome within the GI tract and stated that he would withhold judgment until larger trials are complete. As a reminder, Novo Nordisk has an ongoing phase 2a trial for its primary oral insulin candidate expected to complete in January 2016; Biocon and Oramed are also in the oral insulin game. In contrast to other delivery methods, Dr. Cefalu was more positive about the theory behind inhaled insulin delivery, noting the inherent advantages of delivering insulin through the thin, high-surface-area alveolar membrane. Regarding Sanofi/MannKind’s Afrezza in particular, Dr. Cefalu suggested that the product’s distinct PK/PD properties, more convenient and patient-friendly delivery devices, diminished lung toxicity concerns, and greater acceptance of aerosolized medicines may allow it to succeed where previous inhaled insulins have failed. That said, Dr. Cefalu concluded that while inhaled insulin does offer distinct advantages, it still cannot fully challenge the current injectable insulins. He suggested that the most appropriate patients for inhaled insulin might be: (i) those with type 1 diabetes who want to improve glucose control but face issues with hypoglycemia; (ii) individuals with type 2 diabetes with needle phobia who are otherwise reluctant to take insulin; and (iii) patients with either type 1 or type 2 diabetes who are on  basal-bolus therapy but want to reduce their number of daily injections. However, Dr. Cefalu noted that in this last case, replacing a rapid-acting insulin with an inhaled insulin may prove unpredictable since the PK/PD profiles are not exactly the same. For more insights on inhaled insulin, please see our coverage of Dr. Anne Peters’ (USC, Los Angeles, CA) thoughts on the “hybrid” use of this mode of insulin delivery.

Symposium: Type 1 Diabetes – Avenues for New Approaches to Therapy

Disease Modifying Therapy in Type 1 Diabetes

Carla Greenbaum, MD (Benaroya Research Institute, Seattle, WA)

Dr. Carla Greenbaum delivered a passionate talk on the need for disease-modifying therapies for type 1 diabetes – and for endocrinologists to overcome their fear of immunotherapy. To those who argue that advances in current treatment will make disease-modifying therapies unnecessary, Dr. Greenbaum responded that “we’re not as good as we think we are.” She pointed to T1D Exchange data showing “pretty lousy” glucose control across the board, significant rates of severe hypoglycemia, and declining but still prevalent rates of complications. She argued that the best place for disease-modifying treatments is before clinical diagnosis, at stage one of the recently published JDRF/ADA staging system (multiple autoantibodies with normal blood glucose). She also emphasized that children need to be considered as an independent group, as they experience a much faster rate of disease progression than adults. Dr. Greenbaum has been a staunch advocate of this point and co-chaired a consensus conference on the topic earlier this year. In terms of specific therapies, Dr. Greenbaum highlighted teplizumab, rituximab, abatacept, and alefacept as the four candidates that have demonstrated efficacy and a reasonable safety profile in clinical trials. She made the important point that the level of efficacy seen in those trials (~20-25% benefit vs. control) is comparable to that produced by approved immune therapies for other diseases – the trick is figuring out how best to use them. The issue of expectations is perhaps one of the biggest challenges facing this field: there has been so much hope and hype around a “cure” for type 1 diabetes that an immunotherapy or cell replacement approach could be perceived as a disappointment even if it produces real clinical benefits.

Symposium: Hypertension and Diabetic Complications

SGLT-2 Inhibitors and the Regulation of Vascular Function in Diabetes

David Cherney, MD, PhD (University of Toronto, Canada)

Dr. David Cherney speculated on the potential mechanisms driving the heart failure benefit in EMPA-REG OUTCOME. While the suggestion that Lilly/BI’s Jardiance (empagliflozin) may be acting as a diuretic is a popular theory, Dr. Cherney emphasized that there are significant differences in the clinical effects of thiazide diuretics and empagliflozin. Like empagliflozin, thiazide diuretics are associated with lower diastolic and systolic blood pressure and reduced plasma volume. However, the volume contraction effects are attenuated over time, with plasma volume returning to normal by eight weeks. Thus, Dr. Cherney hypothesized that the blood pressure lowering observed may be related to systemic vasodilation rather than plasma volume. In contrast, SGLT-2 inhibitors appear to produce long-term, sustained reductions in plasma volume. The finding that the hospitalization for heart failure risk curves separated very early on in the trial – within 3-4 months – suggests to Dr. Cherney that an acute hemodynamic or volume-mediated effect is at play. He also suggested that the reduced plasma volume could contribute to reduced myocardial stretch, which in turn can reduce arrhythmias. Arrhythmias could have been counted within the category of sudden cardiac death in the trial, so Dr. Cherney hypothesized that the reduced plasma volume could contribute to the reduction in cardiovascular mortality risk. Another hypothesis Dr. Cherney raised included a potential role for metabolically active epicardial adipose tissue (which has been shown to suppress myocardial contractility). Dr. Cherney also suggested that SGLT-2 inhibitors could increase renin levels, thereby activating the renin-angiotensin system. Since many of the participants in EMPA-REG OUTCOME were already on an ACE inhibitor, the angiotensin I from the system may be shunted over to the vasodilatory pathways. Dr. Cherney also hypothesized that the risk reduction could be due to multiple pathways or a potential renal protective effect of SGLT-2 inhibitors.

Symposium: SGLT-2 Inhibitors in Treating Diabetes

Clinical Role

Per-Henrik Groop, MD (University of Helsinki, Finland)

Dr. Per-Henrik Groop focused on the profound nature of the EMPA-REG OUTCOME results, using the HOPE study of the ACE inhibitor ramipril as a point of comparison. He first synthesized the pre-EMPA-REG data on the SGLT-2 inhibitor class, highlighting the durable weight loss and blood pressure reductions and expressing slight concern about the increase in LDL cholesterol. He argued that in marked contrast to the situation for DPP-4 inhibitors (a lot of hope for cardioprotection followed by neutral outcomes trials), the EMPA-REG OUTCOME results completely exceeded expectations for the SGLT-2 inhibitor class. While we would argue that there was some optimism in the field for cardioprotection with SGLT-2 inhibitors, we agree that the magnitude of the results and the specific parameters affected were a complete surprise. Dr. Groop highlighted the similarity of the EMPA-REG OUTCOME results to those in HOPE, which led to the establishment of ACE inhibitors as standard treatment for people with diabetes. With the caveat that this was not intended to be a head-to-head comparison, Dr. Groop noted the similar relative risk reduction for all-cause and CV mortality in both trials. He also noted that the risk reduction in EMPA-REG OUTCOME occurred in the context of much lower absolute risk, which he attributed to a higher standard of care. Several questioners pushed back against Dr. Groop’s excitement during Q&A, noting that HOPE included a much broader patient population than EMPA-REG OUTCOME and stressing the need to avoid over-extrapolation of the results. Dr. Groop agreed that the data should not be over-generalized but stressed that he made the comparison to emphasize the landmark nature of the study and the prediction that it will lead to a similar paradigm shift.

Questions and Answers

Q: I hope people are applauding your graphics, not your conclusion. We have a very high-risk population and your extrapolation is out to generality. The HOPE trial had a mix of people. I think we’re moving out into unsafe waters with enthusiasm for this idea that this should be generically differentiated.

A: These were not head-to-head comparisons. I’m trying to show what we’ve learned from big studies. We learned from HOPE to treat people with diabetes with ACE inhibitors. Now we have new medications that on top of standard of care have similar effects. That’s the take-home message to bring to the clinic.

Dr. Ralph DeFronzo (University of Texas Health Science Center, San Antonio, TX): We have a very select population who’s incredibly sick and very high risk. I’m not convinced at all that if you take newly diagnosed diabetics you’ll see the same results. If you have patients who fit these characteristics, they should be on empagliflozin. To extrapolate to other populations, who probably make up 80-90% of the total population, is hazardous.

A: I agree, but I’ll give the same answer. I’m trying to show more of a paradigm shift. The data for empagliflozin was shown in a high-risk population. We don’t know if it can be translated to those without established CVD. DECLARE will hopefully show that. I totally agree with you. I’m just saying we should consider that this was a landmark study, and I think it will lead to a paradigm shift.

Q: We saw the comparison between SGLT-2 inhibitors and ACE inhibitors. Do we have a comparison with metformin?

A: I tried to squeeze out what I could from HOPE. I don’t know how many were on metformin. This was not head-to-head; I’ll stress that again. I’m saying that with 15 years of time difference, the two trials showed a very similar effect. The absolute risk reductions were the same and on top of standard of care. Don’t go home and try to translate everything to a patient not shown to have a benefit.

Q: HOPE studied people out to 4.5 years. EMPA-REG was two years. What are your thoughts on the implications of two years vs. five years?

A: Again, EMPA-REG enrolled very sick patients. In HOPE you also had people with risk factors. In EMPA-REG everyone had established disease. In HOPE, 31% just had risk factors. If you have established CVD, then the time to treat is shorter.

Clinical Safety

Stefano Del Prato, MD (University of Pisa, Italy)

Dr. Stefano Del Prato reviewed the various safety parameters of SGLT-2 inhibitors, providing commentary on the class’ various suggested safety concerns. After demonstrating that the class can provide good durable glycemic control, he presented data around hypoglycemia, concluding that the drugs’ risk is low in various treatment regimens (added on to insulin and other medications) likely due to the class’ indirect mechanisms of action on insulin sensitivity and secretion. With regards to urinary tract infections, Dr. Del Prato noted that this has become less of an issue while mycotic genital infections remain the most common side effect in females, although events tend to diminish over time and episodes are mostly mild and manageable. However soon after this presentation, the FDA added new warnings and precautions on DKA and urinary tract infections to all SGLT-2 inhibitors’ labels – the recent reporting of serious UTI cases may thus bring this signal back to the forefront. Regarding kidney function, he noted that no evidence yet suggests that the drug class can accelerate loss of kidney function, although certain patient populations (i.e. older age, use of loop diuretics) may be at higher risk for hypovolemia. Dr. Del Prato also reviewed the evidence around bone fractures, as canagliflozin but not dapagliflozin has shown an effect on bone mineral density, thus begging the question as to if this is a drug effect or class effect. However ultimately, Dr. Del Prato pointed to the EMPA-REG results when assessing the class’ risk/benefit balance, concluding that the class’ potential risks are worth taking in exchange for cardioprotection. However, he shared the important message that amidst the great enthusiasm surrounding these CVOT findings, we must not forget to remain wary of any safety signals that emerge – in light of EMPA-REG, DKA, and recent additional FDA warnings, we believe this class has certainly received attention from all sides.

  • On DKA, Dr. Del Prato shared his belief that SGLT-2 inhibitors must be properly used by their indications – a view that we hear shared by many other KOLs. He stated that SGLT-2 inhibitors may predispose individuals to ketoacidosis with most of published cases resulting from insulin deficiency in type 1 diabetes while type 2 diabetes cases were typically with patients already on insulin or under stressful conditions.
  • Interestingly, he pointed to SGLT-2 inhibitors’ possible role as a technique to identify tumors in cancer. Dr. Del Prato highlighted that the data in diabetes show a non-significant increased risk, but noted that recent research has utilized the functional expression of sodium-glucose transporters in studying cancer – an intriguing area that we’ll keep our eye out for to see if this brings any insights to type 2 diabetes.

Questions and Answers

Q: Why should there be increased bone fracture risk in one drug but not the others?

A: I have no clear-cut explanation why the risk of bone fracture should be increased with one of the drugs in the class but I have to acknowledge that FDA has revised the label for canagliflozin. Data recently published have shown a decrease in bone density with canagliflozin, while similar studies have not found such an effect with dapagliflozin. No increase in bone fractures has been reported over the 4-year follow-up of the EMPA-REG with the use of empagliflozin.  However, there’s no way to get a final answer unless head-to-head trials are carried out.

Q: Looks like this mostly has to do with bone density, which may possibly be explained by lack of calcium?

A: No imbalance in calcium or magnesium concentration have been reported in the registration trial. And data on biomarkers of bone absorption and formation have not been found to be significantly affected by the use of SGLT2 inhibitors.

Symposium: Diabetes in Youth

The SEARCH for Diabetes in Youth Study: Overview and Results

Elizabeth Mayer-Davis, PhD (University of North Carolina, Chapel Hill, NC)

After a review of the published data from the first two phases of the SEARCH for Diabetes in Youth Study (SEARCH 1 and 2; latest results published in 2014), Dr. Elizabeth Mayer-Davis noted that the investigators are currently analyzing the latest data from SEARCH 3 and have submitted an abstract to ADA for presentation at ADA 2016. The SEARCH 3 data will cover incidence trends for type 1 and type 2 diabetes from 2002-2012. It will be the first SEARCH dataset to report incidence trends for type 2 diabetes and for racial/ethnic groups other than non-Hispanic whites. The next prevalence data will be collected in 2017. The group is also launching the SEARCH 4 registry and cohort studies, which will complete in 2020. The registry study will investigate incidence and prevalence trends for type 1 and type 2 diabetes. It will also look at whether the alarmingly high rates of DKA seen in previous SEARCH studies (28-29% at diagnosis for type 1 diabetes) have declined and whether the surveillance process itself can become more cost-efficient. The cohort study will evaluate the incidence and progression of complications in a group of youth (n=1,464 with type 1 diabetes, n=384 with type 2 diabetes) who had in-person visits early in the study (2002, 2006, and 2008). It aims to identify the risk factors driving development of complications and understand why complications seem to occur earlier in youth with type 2 diabetes compared to type 1.

Symposium: Islet Transplantation vs. Embryonic Stem Cells – Where Does The Future Lie?


Kevin D’Amour, PhD (ViaCyte, San Diego, CA), Timothy Kieffer, PhD (University of British Columbia, Vancouver, British Columbia, Canada), James Shapiro, MD, PhD (University of Alberta, Edmonton, Alberta, Canada), Ian Rogers, PhD (Lunenfeld-Tanenbaum Research Institute, Toronto, Ontario, Canada)

Speakers in this symposium offered a strong vision of a future in which stem cell-based therapies are a mainstay of type 1 diabetes treatment. Dr. Kevin D’Amour first reviewed the preclinical studies of ViaCyte’s VC-01, currently the most advanced cell replacement therapy in development. He noted that investigators are still enrolling the first cohort of patients in the ongoing phase 1/2 trial of the device. Ten patients have been treated to date, and a few have now reached their one-year anniversary with the product. Dr. D’Amour declined to provide specifics on how those patients have responded but emphasized that this first phase is intended to evaluate safety and optimize the device – in other words, impressive efficacy results were not expected at this point. Dr. Tim Kieffer then discussed his group’s protocol with BetaLogics to produce functional beta cells (more mature than the pancreatic progenitor cells produced by ViaCyte’s protocol) from embryonic stem cells. That project remains preclinical, as investigators are exploring how environmental factors like diet or thyroid hormone levels could influence the cells’ efficacy. Dr. James Shapiro discussed recent advances in protocols for islet transplantation.  He concluded on an optimistic note, saying it is “clear that cell transplantation will become mainstream treatment in the future…I firmly believe the nuts and bolts of a cure are in play.” Finally, Dr. Ian Rogers explained the science behind iPS cells, which could potentially serve as a less controversial replacement for embryonic stem cells as a cell source for these treatments.

Symposium: The Alpha-Cell – Learning from the Past and Building for the Future

Blocking Glucagon Action: Ready for Prime Time?

David Kelley, MD (Merck, Rahway, NJ)

In a detailed presentation on glucagon receptor (GCGR) antagonist therapy, Dr. David Kelley stated that the worrisome dose-dependent elevation in plasma LDL cholesterol associated with the class is mechanistic, and can therefore be mitigated pharmacologically. For example, he shared preclinical results demonstrating that co-administration of the cholesterol drug ezetimibe inhibited the rise in LDL cholesterol associated with Merck’s discontinued GCGR antagonist MK-0893. Dr. Kelley also contextualized current GRA therapy research by providing background on MK-0893, which was discontinued in 2011 due to safety concerns (weight gain and increased liver transaminases in addition to elevated LDL cholesterol). Indeed, this drug class has encountered a few recent stumbling blocks, though the companies with candidates still in development have expressed great excitement about the potential efficacy. Lilly recently discontinued its development of its phase 2 candidate LY2409021 after a phase 2 blood pressure trial failed to meet its primary endpoint. Further, at this year’s AACE, Dr. Ralph DeFronzo (University of Texas Health Science Center, San Antonio, TX) expressed skepticism about the class, cautioning that though GCGR antagonists boast exciting glucose-lowering efficacy, the long-term impacts are largely unknown. However, Isis and Ligand both have GCGR antagonists in early clinical development and have offered very positive commentary on the class. Isis recently initiated a new phase 2 trial of its candidate ISIS-GCGRRx and Ligand shared in its 3Q15 update that it plans to initiate phase 2 for its candidate LGD-6972 in mid-2016.

Symposium: Pipeline Therapies For Type 2 Diabetes

GPR Agonist: Ready for Prime Time?

Fiona Gribble, DPhil, BM, BCh (University of Cambridge, UK)

Dr. Fiona Gribble discussed G-protein coupled receptors (GPCRs) as potential therapeutic targets in type 2 diabetes. For context, she opened by explaining that GPCRs are a massive class of receptors found in hormone-secreting enteroendocrine cells (EECs) that can produce GLP-1 and promote insulin secretion while sending satiety signals to the brain. However, as opposed to other targets, GPCRs are particular attractive because they are easy to activate (or block) with specificity. Dr. Gribble also pointed to the great diversity of GPCRs – GPR30, GPR40, GPR119, GPR120 – that are often found only in particular tissues, thus reducing the potential for off-target side effects. As such, she positioned interest in GPCRs as appropriate despite relatively poor results seen with some recent candidates (e.g., Takeda’s GPR40 agonist) and argued the lack of efficacy seen in clinical trials is more likely to reflect our relatively limited understanding of the class vs. the insignificance of GPCRs as a whole. She noted that the co-action of various GPCRs may offer the best potential solution in the future or that combinations with our drug classes, such as DPP-4 inhibitors, are needed for greater efficacy. Ultimately, Dr. Gribble suggested that there are far too many unanswered questions for the field to lose interest in this approach. Indeed, we felt that she remained very optimistic as she stressed that raising gut hormone levels is one of the hallmarks of gastric bypass surgery – in her view, drugs that boost gut hormone levels must reasonably be expected to mimic some of the beneficial effects of bypass surgery. She did acknowledge that the challenge is mimicking the scale (a 10-fold increase in hormone secretion) of the effect we see after surgery, suggesting that perhaps targeting the distal gut in addition to the proximal gut offers one avenue for increasing hormone secretion more significantly.

FGF-21 as a Potential Therapeutic Approach

David Moller, MD (Lilly, Indianapolis, IN)

Dr. David Moller summarized research examining the promise of fibroblast growth factor-21 (FGF-21) as a therapeutic approach to treating type 2 diabetes. He began by walking through the history of the molecule’s discovery and development, speaking in glowing terms about the vast potential. He noted, though, that some of this enthusiasm has diminished in recent years given clinical findings from Lilly and Pfizer that the analog did not show statistically significant reductions in glucose. Despite these negative results to date, Dr. Moller remained convinced that the net relevance of these potential effects in humans has yet to be fully described. He suggested that other lines of evidence also point toward indications in fields other than diabetes or obesity (such as in dyslipidemia or NASH) as possible alternative therapeutic uses. Indeed, we heard Dr. Hans-Ulrich Haring (University of Tubingen, Germany) discuss the importance of fatty liver disease as a driving factor within diabetes at this past EASD. Several critical questions remain to be addressed, though overall, Dr. Moller seemed cautiously optimistic that FGF-21 based therapies could potentially be harnessed in the future as novel approaches to prevent and treat a variety of metabolic disorders.

  • FGF-21 mechanism of action: Unlike other members of the FGF family, FGF-21 is not mitogenic (does not promote cell division), which suggests that it would not be carcinogenic. Preclinical studies suggest that FGF-21 exerts its positive metabolic effects by increasing metabolic rate (increasing energy expenditure). With regards to its molecular mechanism, FGF-21 binds to FGF receptors, of which FGF-1R is of primary interest. β-Klotho (KLB) is an obligatecofactor for FGF-21 binding to FGF-1R, and is also sufficient for FGF-21 stimulation of glucose uptake in adipocytes where it is not normally expressed. Downstream, FGF-21 increases adiponectin levels, reduces levels of free fatty acids, increases leptin sensitivity, and potentially causes “browning” of white adipose tissue (a very exciting prospect).
  • Adverse events: Notably, Dr. Moller acknowledged that there has been a signal of bone-density reduction associated with FGF-21 treatment in mouse studies though stressed that the relevance of this data in humans has yet to be examined.

IDF Award Lecture: Basic Science

Cellular Communication, Insulin Resistance and the Pathogenesis of Type 2 Diabetes

C. Ronald Kahn, MD (Joslin Diabetes Center, Boston, MA)

IDF Award Recipient Dr. C Ronald Kahn delivered an in-depth presentation on the pathogenesis of type 2 diabetes and obesity, highlighting several key underlying mechanisms: protein kinase C (PKC) δ expression, adipose tissue inflammation, concentration of brown/beige fat, and external (diet) and internal (microbiome) environmental factors. He reviewed relevant data from a series of studies on two mouse models, the B6 mouse (prone to obesity and type 2 diabetes) and the 129 mouse (resistant to obesity and type 2 diabetes). One study found that the disease-prone mice had a higher concentration of liver PKCδ, an enzyme known to promote insulin resistance, than the resistant mice. Further, the overexpression of the PKCδ gene induced fatty liver disease in mice, while knocking it out reduced signs of ectopic fat in the liver. Dr. Kahn noted that humans with obesity and obesity-associated diabetes also have higher levels of PKCδ. He described additional research on the same mouse models, which revealed that leukocyte markers of inflammation in white adipose tissue are higher in B6 mice compared to 129 mice, while 129 mice are resistant to inflammation. 129 mice also have more inducible brown/beige fat than B6 mice, which may also play a role in their resistance to obesity. Next, Dr. Kahn discussed outcomes of two studies using separate 129 mouse cohorts from different vendors, echoing comments he made at last year’s Research Symposium on Diabetes and the Microbiome. The two groups of 129 mice were raised in different environments, leading to significant changes in the gut microbiome that altered their propensity for glucose tolerance and obesity: both displayed strong glucose tolerance, but one was obesity resistant while the other was obesity prone. Dr. Kahn’s lab found that many of the differences were eliminated when the two lines were bred for three generations in the same environment, but each strain still retained a unique microbiome. Notably, after three generations the two 129 strains had very similar and low levels of weight gain. This suggests that the difference in microbiome contributed to the discrepancies in weight gain observed in early generations, highlighting how simple changes in the external environment can meaningfully alter the the microbiome and influence metabolic outcomes.

  • Additional research attempting to modify the gut microbiome revealed that transplanting a B6 microbiome into a germ-free mouse led to weight gain and high blood glucose. Moreover, administration of antibiotics caused dramatic changes in the microbiome, radically decreasing diversity across all mouse strains. In B6 mice, antibiotics decreased the expression of liver inflammatory genes and improved insulin resistance, thought these changes were not observed in the 129 mice. Dr. Kahn emphasized the need to understand how the whole bacteria community contributes to metabolic syndrome phenotypes, rather than the contribution of individual bacterial species.

CDA 2015 Young Scientist Award

Toward a Complications-Free Life: Early Detection and Intervention Strategies in Type 1 Diabetes

Bruce Perkins, MD (University of Toronto, Canada)

In the heartfelt CDA Young Scientist Award lecture, University of Toronto’s Dr. Bruce Perkins shared optimism for use of SGLT-2 inhibitors in type 1 diabetes. His team plans to test SGLT-2s as an adjunct to dual-hormone closed-loop. Dr. Perkins specifically mentioned the ongoing phase 3 study of empagliflozin in type 1 diabetes (EASE2), which follows positive proof of concept studies of SGLT-2s in type 1. He admitted that euglycemic DKA is a concern in type 1, but it should not prevent use of the drug in his view – we just need to be proactive about preventing later-stage DKA. Dr. Perkins shared that his team is planning to add an SGLT-2 inhibitor on top of insulin-glucagon closed-loop, which could really helpful attenuate postprandial spikes. There was no timeline on when this study could begin, but we certainly can’t wait to see those results, especially because adjunctive liraglutide and pramlintide have shown only modest benefits in the closed-loop setting. Cost is also a concern with any adjunct to insulin-only closed loop ... (As an aside, Dr. Perkins seemed less optimistic on insulin-only closed-loop – he noted a larger improvement in time-in-range with the Bionic Pancreas vs. the Cambridge system, and highlighted Dr. Ahmad Haidar’s data from ADA 2015 showing an advantage insulin+glucagon to insulin-only overnight. This debate will have much better data once the Bionic Pancreas team completes its head-to-head insulin-only vs. insulin+glucagon studies next year).

  • More broadly, Dr. Perkins talked about ongoing validation of Corneal Confocal Microscopy (CCM) to predict future neuropathy, and work on nephropathy through the PERL study. The former was new to us and is currently in a 1,040-patient study with 6-8 years of follow-up. The goal is to establish the diagnostic thresholds for future prediction of neuropathy before patients show any signs of the complication. That would be big indeed. Notably, CCM could be combined with retinal screening visits, offering a more integrated approach to complications screening.

Debate: Do We Still Need More Outcome Trials In Diabetes?

Yes, Nothing Else Truly Suffices

David Nathan, MD (Harvard University, Boston, MA)

Arguing in favor of outcomes trials, Dr. David Nathan asserted that while observational studies can raise important questions that inform future trials, only controlled experimental studies can sufficiently control for bias and confounding, properly assess risk, and demonstrate causal relationships. This is evidenced by the list of prominent discrepancies between epidemiological and experimental studies in the past. Dr. Nathan did acknowledge the concerns about cost and generalizability with RCTs and noted the recent push for “pragmatic trials” to help bridge the gap to real-world practice. He did not explicitly endorse this approach, but we believe it could be a very promising way to strike the right balance between scientific rigor and clinical relevance. Dr. Robert Califf, President Obama’s nominee for FDA Commissioner, has been an advocate of pragmatic trials in the past, suggesting that this could become a priority for the FDA if he is confirmed. On the more specific question of whether outcomes trials are needed for diabetes drugs, Dr. Nathan argued that while there is strong evidence in favor of A1c as a surrogate for microvascular benefit, the lack of clarity in the macrovascular arena means that outcomes trials are still required. Even if that is the case, we believe there is still plenty of room for revision of the current guidelines for these trials, potentially including more flexible endpoints, longer durations, or use of active comparators. We look forward to watching results emerge for GRADE, a trial that Dr. Nathan is helping lead.

No, There Are Other Good Alternatives

David Matthews (Oxford Centre for Diabetes, Oxford, UK)

On the opposing side, Dr. David Matthews argued that while RCTs are the gold standard for answering explicit, limited questions, “the truth from an RCT can be very thin.” He showed a pie chart illustrating the tiny fraction of the total type 2 diabetes population enrolled in most CVOTs and argued that this is often not sufficiently acknowledged when interpreting results. He also noted that trial results only refer to the mean or median outcome of an intervention, and the wide distribution of individual responses is often forgotten when interpreting the data. As many others have noted, he argued that most current CVOTs may be too short to provide useful information: UKPDS would also have been neutral if it lasted 18 months to three years.  Most importantly, Dr. Matthews believes that conducting RCTs of specific drugs in patients with advanced diabetes is not the smartest approach to tackling the epidemic. He argued that it is very clear from epidemiology that rising obesity prevalence is driving rising diabetes prevalence and that it would be more prudent to focus on large-scale prevention efforts comparable to those in the infectious disease field. Dr. Matthews also made the oft-cited point that these trials are very expensive, estimating their cost at $8-$16 billion by 2020.

  • “Could we do something more useful with $16 billion?” We thought this was an excellent thought-provoking question – while the information gained from CVOTs has clearly been valuable, it is less clear whether it has been worth the cost. Dr. Matthews’ proposed alternative topics for investigation:
    • Scaling up community prevention interventions
    • Primary pathology of the beta cell
    • Effects of calorie labeling and soda taxation
    • Optimizing retinopathy screening and prevention of blindness
    • Biological solutions like stem cells and transplantation

Debate: Incretin Therapies and Cancer

We Still Need To Be Cautious

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

Dr. Daniel Drucker urged caution in the debate over cancer risk with incretin therapies, focusing primarily on the potential risk for colorectal cancer. He argued that there is a “paucity of excellent data” on any type of cancer with incretins, as clinical trials (including CVOTs) are too short to properly evaluate the risk and retrospective analyses can’t correct for confounding variables. He did not express an enormous amount of concern over the risk of medullary thyroid carcinoma (MTC), as the incidence is extremely low (1 in 25,000) and the only concerning signals have come from rodents, which have very different physiology from humans in this area (only rodents have GLP-1 receptors on thyroid C-cells). Dr. Drucker predicted that registry studies should provide more color on whether any clinical imbalance exists. He also offered fairly reassuring commentary on pancreatic cancer, noting that the only hints of an imbalance in CVOTs of incretin therapies have favored the active treatment group. As he has in other recent talks (most extensively at Rachmiel Levine in March), Dr. Drucker expressed the greatest concern about the risk of colorectal cancer, as preclinical studies have consistently shown that GLP-1 acts as a growth factor in the bowel. While there is no conclusive human data on the risk, there are some concerning signals. Human biopsy data has shown increased proliferation of epithelial cells in the colon following bariatric surgery (when GLP-1 levels rise), colorectal cancer was the only type of cancer with an increased incidence after bariatric surgery in the Swedish Obese Subjects study, and the ELIXA CVOT of Sanofi’s Lyxumia (lixisenatide) found a 17-10 imbalance in colorectal tumors with the drug. Dr. Drucker was careful to emphasize that this does not constitute conclusive evidence but reminded the audience that “the absence of evidence is not the evidence of absence.”

There is Little to Worry About

Michael Nauck, MD (Diabeteszentrum Bad Lauterberg, Germany)

Dr. Michael Nauck argued that concerns about cancer with incretin-based therapies have been overstated. He cautioned that the FDA adverse event reporting system, from which most concerns about pancreatic and thyroid cancer have arisen, is highly susceptible to reporting bias. He also noted that the best animal studies – involving long-term exposure and careful histological analysis in non-human primates – have found no signs of increased risk of either pancreatic or thyroid C-cell cancers with incretins. Dr. Nauck shared data from a new meta-analysis he conducted (not yet published), which he believes constitutes the most reliable human data on the subject. Using clinical trial data from over 40,000 people on GLP-1 agonists and over 80,000 on DPP-4 inhibitors, it found a risk ratio of 0.96 with a tight confidence interval (0.90-1.03) for all neoplasms. The incidence of MTC was too low to draw any statistical conclusions, but the results for all other types of thyroid carcinomas were neutral (HR = 0.67; 95% CI: 0.37-1.21). The same was true for colorectal cancer (HR = 0.84; 95% CI: 0.57-1.24) and pancreatic cancer (HR = 0.82; 95% CI: 0.57-1.19). Dr. Nauck’s overall practical conclusions were not all that different from Dr. Drucker’s; he even dug up a quote from Dr. Drucker stating that the hypothesis of increased cancer risk with incretins was not supported by the evidence. The main point of disagreement seemed to be on the degree of concern around colorectal cancer, though when pressed during Q&A, both agreed that frequent screening is the most appropriate clinical response to the potential risk at this point.

Panel Discussion

Q: What level of increased cancer risk would be tolerable with drugs that haven’t demonstrated a benefit on cardiovascular or total mortality?

Dr. Michael Nauck: You’ve mentioned the right way to assess risk and benefit. We’re just in the beginning of understanding the benefits. We’ll see results reporting from CVOTs with other GLP-1 agonists soon. One can question whether the design of the studies is suited to bring out the benefits the drugs have under conditions where you allow glucose to be reduce. I think it’s too early now to say given this or that benefit, we can live with this or that increment in cancer risk for rare tumors. We’re not in a position to make that calculation now.

Dr. Daniel Drucker: A lot of drugs don’t show a clear reduction in CV risk in the context of studies. You may be familiar with a drug called insulin; ORIGIN showed no reduction in CV risk but gave valuable information on its negligible effects on tumor formation. We’re not in a position to say yet what is the true risk/benefit ratio. We’re still waiting for most CVOTs to report, and even those studies are likely not sufficient to provide definitive answers for malignancies, even colorectal. We’ll get a dozen or so cases per arm.

Q: Dr. Drucker, can you explain the quote from the paper he dug up?

Dr. Drucker: I still feel the same way for most biology. The reality is that the bowel growth data is newer than two years old. It was just published, and it’s data that was previously replicated in Copenhagen. It’s mouse data. To me it generates a hypothesis and nothing more. In order to test it, you need a much larger, rigorous long-term controlled trial or perhaps a meta-analysis.

Dr. Nauck: I think we have a strategy to reduce risk for colorectal cancer in general since patients at higher risk really should use screening programs. The typical interval between two colonoscopies is three years. I think the duration of exposure that’s been covered comes close to three years. If you do things right and screen, we should be on the safe side.

Q: What about SGLT-2 inhibitors and cancer? Preclinical data did show some signals of renal carcinoma and blood cancer. Is there any potential risk there with long-term clinical use?

Dr. Nauck: That’s an unexpected question. I remember the discussion about dapagliflozin and bladder tumors. Now empagliflozin has reported some beneficial effects on renal outcomes. I think we have to continue to monitor it, but today I don’t know of a strong signal that would lead us to expect increased rates.

Dr. Drucker: I think that data is pretty unconvincing.

Q: The drugs we’re talking about are not first line. Can you comment on the difference in risk with GLP-1-based therapies combined with metformin vs. used separately?

Dr. Drucker: Not really; the dataset is insufficient. There’s huge literature suggesting we should use metformin as an anti-cancer agent. There’s an NIH-sponsored trial with breast cancer that should report soon. Most metformin cancer data is retrospective, not controlled. We will get good high-quality data to see if we need to put metformin in the drinking water.

Dr. Nauck: My clinical conclusion is that these drugs are best used in combination with metformin. That should take care of the cancer risk to the best of our knowledge today.

Q: Do you feel comfortable prescribing incretins for a patient with a history of pancreatitis?

Dr. Drucker: No, I don’t think you should. There’s a very slight risk with incretin-based therapy and there are lots of other ways to treat diabetes.

Q: What about the risk of benign colonic adenomas?

Dr. Drucker: I’ll answer your question with Dr. Nauck’s reminder that if you’re concerned, you should have regular colonoscopies. I can’t answer the question. We don’t make clinical decisions based on mouse data. There’s no compelling human data, so it’s up to the physician.

Q: Given the rarity of events and the difficulty to distinguish them in clinical trials, what is the answer to determining risk? Will registries give us the data we need or will people avoid treatment of those with chronic polyps who are at the highest risk so we’ll never know? Will we rely on post-marketing CV trials that recruit a different population to look for tumors?

Dr. Nauck: I only recommend that kind of activity if there is a real signal in humans that there might be an increased risk. Then to nail that, registries are one way of keeping our eyes open. Adenomas can be managed, and sending continuous data to a database is helpful. I don’t think that’s necessary if there’s no real signal that we have to be afraid of a real increase in risk. I don’t see that now.

Dr. Drucker: I don’t see it either because we don’t have the data.

Non-Industry Satellite Symposium: Joint Symposium of the Diabetes & Cardiovascular Disease EASD Study Group and the Munich Diabetes Research Group e.V.

Antihyperglycemic Drugs and Heart Failure

Eberhard Standl, MD (Diabetes Research Group e.V. at Helmholtz Zentrum München, Munich, Germany)

Dr. Eberhard Standl tackled the complex subject of diabetes drugs and heart failure, calling for greater attention to the topic in clinical trials. Echoing recent comments from CODHy, he noted that a large number of patients in recent CVOTs for diabetes drugs likely had undiagnosed heart failure. For example, he suggested that 25% of participants in SAVOR and EXAMINE should have been worked up for acute heart failure and over 50% for chronic heart failure based on biomarker measurements, yet the actual percentage of diagnosed patients was only ~12.8% and 27.8%, respectively . He argued that future CVOTs should attempt to capture all cases of heart failure at baseline using BNP measurements and include a composite of heart failure hospitalization and CV death as a key secondary outcome. Even if the FDA requirements do not change, we expect there will be increased attention to heart failure in future trials given the growing consensus around its relevance in diabetes and the unexpected results from recent studies. In terms of specific drug classes, Dr. Standl suggested that the story for the TZDs is a complicated one. The shift away from insulin resistance and fatty acid metabolism induced by the class should be beneficial, but the adverse effects of increased fluid retention may outweigh those benefits at least in patients with existing heart failure. He offered no silver bullet explanation for the much-debated discrepant heart failure results from the three DPP-4 inhibitor CVOTs or for the unexpected benefit in EMPA-REG OUTCOME. He did suggest that the contribution of heart failure to the mortality benefit in EMPA-REG OUTCOME may be “getting overrated” because heart failure only accounted for 25% of those deaths. However, as noted in the morning panel, that figure may be a significant underestimate given the potentially high rate of undiagnosed heart failure and the difficulty ascertaining cause of death in many cases.

Corporate Symposium: Getting to the Heart of Cardiovascular Risk in People with Type 2 Diabetes (Sponsored by Lilly/BI)

Take Heart: Contemporary CVOTs and EMPA-REG OUTCOME

Silvio Inzucchi, MD (Yale University, New Haven, CT)

Dr. Silvio Inzucchi reviewed the now-familiar EMPA-REG OUTCOME results, emphasizing their dramatic and unprecedented nature. He particularly stressed the importance of heart failure as an outcome in diabetes trials now that more patients are surviving heart attacks and living longer with cardiac damage. He also suggested that many EMPA-REG OUTCOME participants likely had subclinical heart failure at baseline. The importance of heart failure in CVOTs for diabetes drugs has emerged as one of the strongest points of consensus in the field since the first trials have begun reporting results. Some have suggested that heart failure should be incorporated into the primary composite endpoint for these trials or at least be evaluated as a key secondary endpoint. The challenges involved in diagnosing and adjudicating heart failure seem to be the biggest practical hurdles, but we expect that this will become a higher priority in future trials given the surprising results (positive in EMPA-REG OUTCOME and negative in SAVOR) that have emerged thus far. On the question of the hour – what drove the benefit in EMPA-REG OUTCOME – Dr. Inzucchi echoed the most common hypothesis that osmotic diuresis seems to be the most likely explanation, though he stressed that there are plenty of other options that need to be investigated. He did argue that the benefit was almost certainly not mediated through atherosclerosis or glucose lowering given the extremely small difference in glycemic control between groups and the lack of a significant effect on MI or stroke.

The Heart of the Matter: CV Risk Reduction in Type 2 Diabetes

Sanjay Kaul, MD (Cedars-Sinai Medical Center, Los Angeles, CA)

Dr. Sanjay Kaul contextualized the implications of the EMPA-REG OUTCOME results from a regulatory standpoint. He began by presenting the compelling benefit-risk profile demonstrated in the trial: 17 fewer MACE events, 22 fewer cardiovascular deaths, seven fewer MIs, 25 fewer overall deaths, and 14 fewer hospitalizations for heart failure vs. 53 excess genital infections and five excess strokes for every 1,000 patients treated with empagliflozin (Lilly/BI’s Jardiance). He argued that this benefit-risk profile justifies a Class I recommendation (meaning the benefit greatly outweighs the risk and the treatment should be administered). The next question is whether further studies must be performed to justify a Class IA designation (sufficient evidence from multiple randomized trials or meta-analyses, vs. a Class IB designation where the evidence comes from a single randomized trial or nonrandomized studies). He argued that the highly significant p-value of p<0.0001 is sufficiently persuasive since it means there is a 91% chance that future studies would replicate the findings. For comparison, the likelihood is only 50% for a p-value of 0.05. Dr. Kaul admitted that when he first encountered the EMPA-REG OUTCOME results, he was skeptical and wondered if the findings were “too good to be true.” However, he now believes that the results are likely to be valid based on the large and early benefit observed, the highly statistically persuasive findings, the large number of events in the trial, and the consistent effect seen with both doses. That said, Dr. Kaul believes that the data do need to be replicated given how “unexpected and unprecedented” they were. He concluded on an uplifting note by stating that if the results hold up to regulatory scrutiny, the long-sought “holy grail” of a diabetes drug that improves cardiovascular outcomes may have been achieved.

  • Dr. Kaul presented a Bayesian analysis supporting the validity of the EMPA-REG OUTCOME all-cause mortality results (32% reduced risk). Even when using a skeptical prior (which assumes a large treatment effect is unlikely), Dr. Kaul found that there is a 92% probability of a 15% risk reduction in all-cause mortality. Using the baseline non-informative prior (which makes no assumptions about the expected treatment effect), the probability of a 15% risk reduction increased to over 99%. Dr. Kaul emphasized that even with skepticism injected into the data, they still hold up to scrutiny and stated that he is very comfortable using this 92% probability when making clinical decisions.
  • Dr. Kaul suggested that it’s a “slippery slope” to attempt to explain the mechanism behind the risk reduction. Like most speakers we have heard over the past few months, he ruled out glycemic control, blood pressure lowering, and weight loss as likely factors. He also dismissed the idea that background medications could have exerted a harmful effect in the placebo group or a cardioprotective effect in the empagliflozin group because there was no significant difference in additional medications between the two groups. He suggested that that volume depletion and antiarrhythmic effects were potential drivers of the risk reduction and repeated his quip from EASD that “multifactorial” is often a euphemism for “we simply don’t know.” Ultimately, he emphasized that future mechanistic studies are required to pick apart the exact mechanism behind the risk reduction.
  • Dr. Kaul suggested that a beneficial SGLT-2 inhibitor class effect cannot be assumed and that the current evidence more strongly favors a cardioprotective benefit for AZ’s Farxiga (dapagliflozin) than for J&J’s Invokana (canagliflozin). Dr. Kaul argued that a class effect is still an unanswered question since the cardioprotective benefit of empagliflozin is unlikely to be mediated by cardiometabolic factors. He emphasized the importance of continuing the ongoing cardiovascular outcomes trials for SGLT-2 inhibitors in order to provide a more definitive answer to the question of a class effect.

Panel Discussion

Dr. Sanjay Kaul (Cedars-Sinai Medical Center, Los Angeles, CA): Dr. Wright, is the evidence generated generalizable? Do you think it’s practice changing? What do you think about African-Americans in this trial?

Dr. Eugene Wright (Cape Fear Valley Health Affairs, Fayetteville, NC): Clinical studies like this are informative and potentially actionable to the extent that they represent patient populations we see. In my area, we have a large population of African-Americans, Hispanics, and Native Americans, who were clearly underrepresented in this study. That said, the results we see are very encouraging. I must admit when I first saw them, I was very excited about the benefit. Now we not only have a therapy that gives glycemic benefit, improves markers like blood pressure and reduces weight (which makes patients happy), but now it has an indication that it may actually be saving lives. As the ongoing studies are looked at, I hope we do get to see if this benefit extends to special populations.

Dr. Kaul: What do you make of the mortality data? What’s driving the effect?

Dr. David Fitchett (St. Michael’s Hospital, Toronto, Canada): I can say what does not contribute. I don’t think it’s an atherothrombotic effect and I don’t think it’s the reduction of MI or stroke. I don’t think it’s the anti-hypertensive effect. An early separation of mortality curves was seen in heart failure studies. If we look at eplerenone and beta blocker trials, we see mortality benefits with under three months of treatment. My feeling is it has to do with the myocardial effect or the effect on heart function or metabolism. Certainly the osmotic diuretic effect is probably a reasonable hypothesis. We know that in heart failure patients if you treat to increase hematocrit and reduce circulating plasma volume, patients do better in terms of heart failure symptoms and have reduced mortality.

Dr. Kaul: Dr. Inzucchi, as a practicing endocrinologist, what role do you see for this drug in the therapeutic armamentarium? You wear many hats and one is a writer of guidelines. What impact do you feel this has on future guidelines? Do you feel vindicated that you preempted this to some extent by including SGLT-2 inhibitors in the latest iteration of the ADA/EASD guidelines?

Dr. Silvio Inzucchi (Yale University, New Haven, CT): In 2015, when the ADA and EASD asked us to update the guidelines, we included SGLT-2 inhibitors. There was some discussion among us of whether this was too soon. At that point, the drugs had only been on the market for less than three years. We felt based on the safety data, tolerance, and popularity with patients, they deserved the second-line therapy status and, in retrospect, we were right. I’m certainly glad we included them. I’m not sure if our panel will be asked to write the next iteration; we’re at the behest of the ADA/EASD committees. However, most organizations are conservative in their guidelines. They want good evidence and perhaps a multiplicity of evidence. For me, the data are good enough to change my practice for patients who meet the criteria for inclusion in EMPA-REG OUTCOME. However, it would be major mistake to be overenthusiastic and extrapolate the data to patients who are lower risk. I don’t know the timeline for next set of guidelines, but perhaps the committee will want the data to percolate before rushing to change the guidelines.

Dr. Kaul: If you have a mortality benefit from EMPA-REG OUTCOME, is it ethically justified to conduct another trial with empagliflozin to replicate the mortality results? In terms of an evidence level of A or B: if the FDA allows a superiority claim for empagliflozin for cardiovascular mortality then I think it’s a no-brainer – the evidence should be level A, not B.

Dr. Inzucchi: The ADA/EASD document is a position statement, it’s actually not a set of guidelines. “Guidelines” means it incorporates all levels of evidence. Ours is a position statement because we realized there is no data to make recommendations and no comparative trials. It’s essentially expert opinion, which is the lowest form of evidence we have.

Dr. Wright: In terms of changing practice, from a primary care perspective, it’s hard to deny the benefits vs. the risks. It’s hard to make the argument for not wanting to incorporate empagliflozin into practice. We need to understand the limits of who was included in the study and who was not represented, but when you see this kind of benefit analysis, it’s something I think we all get pretty excited about.

Q: How do you explain the increase in non-fatal stroke 30 days after stopping empagliflozin? Did the data break up in terms of hemorrhagic vs. ischemic stroke?

Dr. Inzucchi: There were few hemorrhagic strokes. They were mostly ischemic, which is what tends to occur in diabetes.

Dr. Kaul: If you look at the pharmacokinetics of the drug, there is very little of it around after day 3, it persists perhaps up to day 7. It’s hard to explain why stroke could be attributable to this drug. There could have been an impact during the phase of active therapy that manifested later. Whenever we have sparse data, the point estimates are very unstable and fragile – they dance around the null. I‘m not putting too much on that, but nonetheless it’s going in the wrong direction and needs to be addressed. The effect was also observed in the canagliflozin program, especially in the early stages. At this point, we’re not sure how to explain it and can only offer an explanation post-facto. If you have volume depletion in a person who is already on a loop diuretic, perhaps it induces watershed infarcts that result in strokes. That information was not captured in the early canagliflozin studies. The short answer is I can’t explain it. If we see a similar effect in other outcome trials, we’ll have to take it seriously.

Q: I’m unfortunately a cardiologist that got involved in preventive medicine. The FDA is mandating billions of dollars in trials in high-risk CV patients. These drugs are easily used and are taking off like crazy. Are we really going to say you need to have a heart attack first to use them? Or how about a stroke or a bypass? At the end of the day, we won’t have a properly powered preventive trial in diabetics. I understand the importance of evidence-based medicine but I also understand logic and imperative of preventing these things. Thoughts?

Dr. Wright: The drug has an indication for glycemia. How you tend to use it and in whom is a decision you and the patient make. The study fuels a different conversation about why you’re giving this drug. Many patients I’ve had on SGLT-2 inhibitors before the outcomes findings had great results. Now, in the appropriate patients, I’m able to have a different conversation about why it’s beneficial.

Dr. Kaul: There’s an art of medicine and a science of medicine component. From a regulatory perspective, the science of medicine trumps art. But as physicians, once a drug is approved, we’re free to use it for the appropriate indications. Here’s where it’s tricky: what is driving the benefit? Is it possible that more than 10% of patients had heart failure at baseline? Because it was not captured in a systematic way, my guess is a fairly larger number of these patients, who had to have prevalent ischemic heart disease, had heart failure. That’s the reason why we can only at this point say people with established heart disease are appropriate candidates because the mechanism of benefit might be mediated through that.

Q: Can you elaborate on the causes of cardiovascular death? You talked about heart failure and sudden death. There was also a significant percentage of unspecified deaths – do we have data on this?

Dr. Kaul: As best as I know, it’s un-assessable. We really don’t know what they were. By convention, “other causes” is anything that has ruled out cardiovascular causes. I’m not aware of any more granularity of deaths.

Q: But there were still a number of deaths that were included under cardiovascular deaths but marked “other.”

Dr. Inzucchi: In all large CV trials, if there is no known cause of death such as hit by a bus or cancer, they are all apportioned to “cardiovascular death – other. Dr. Fitchett, what is your underlying sense regarding heart failure outcomes?

Dr. Fitchett: Certainly patients with pre-existing heart failure had a high event rate and a six-fold increase in mortality. Patients admitted to the hospital with heart failure had something like 18% mortality. There’s no question that patients with pre-existing heart failure or who developed heart failure had a high event rate. Sudden death and heart failure death constituted about 50% of mortality. The others were unspecified and classified as cardiovascular death. I think heart failure, either recognized or unrecognized, is a very important part of it. I want to emphasize that heart failure associated with diabetes is very common. In many trials heart failure events are far more common than vascular events like MI or stroke. It’s an outcome that has been under-recognized until recently.

Dr. Kaul: In a non-heart failure trial it is very difficult to adjudicate heart failure outcomes. That’s one reason why we don’t typically include it as an endpoint. One question is regarding the definition of heart failure at baseline and adjudication. It needs to be scrutinized.

Q: The 10 and 25 mg doses appear equivalent. When do you use 25 mg?

Dr. Wright: I start with 10 mg for glycemic effects. If there’s no effect, I go to 25 mg.

Q: In both the placebo and empagliflozin arms statin use was less than 75%. Were you surprised by that? I would have thought it would be more than 90%.

Dr. Kaul: That’s consistent with other non-ACS trials. Only there do you see compliance rates above 85% or 90%. In heart failure or diabetic trials, like all the DPP-4 inhibitor trials, it’s 70-80%.

Dr. Inzucchi: When we looked at the heterogeneity of response we didn’t find any. It didn’t seem like people on statins had an attenuated response.

Q: The results seem to be driven by people with heart failure. Should it only be used in people with heart failure?

Dr. Kaul: I can only address the results for CV death or hospitalization for heart failure. History of baseline heart failure was not an effect modifier. The results were statistically indistinguishable. I haven’t seen the data for MACE yet but I would not be surprised if history of heart failure was not an effect modifier. That’s probably to some extent because it was not recognized. That feeds into my speculated mechanism.

Dr. Fitchett: I don’t think we should say the drug should only be used in people with a history of heart failure or to prevent heart failure. A lot of people with diabetes have subclinical heart failure, so we shouldn’t target just people with a history. There’s heterogeneity there.

Dr. Inzucchi: I would bet $10 that 30-40% of patients had diastolic dysfunction if we had been smart enough to get an echocardiogram at baseline. They were older, there was lots of insulin, and they all had cardiovascular disease. I can’t imagine 90% had normal ventricles.

Dr. Kaul: We concur.

Changing Heartbeats: Challenges and Patients' Emotional Struggles at Insulin Initiation in Type 2 Diabetes

Lawrence Fisher, PhD (UCSF, San Francisco, CA)

Psychologist Dr. Lawrence Fisher gave an actionable talk on adherence to insulin, suggesting that traditional adherence boosting strategies are too narrowly focused. “Medication use is driven by patient feelings, beliefs, and expectations,” he said, and it all starts with more empathetic provider-patient conversations. Indeed, Dr. Fisher recommended everyone read a short NEJM publication on adherence from earlier this year (Rosenbaum et al., 2015), which concludes, “I want to believe that if patients knew what I know, they would take their medicine. What I have learned is that if I felt what they feel, I’d understand why they don’t.” With no change in adherence after 40 years of research (the conclusion of a depressing 2014 Cochrane Review), Dr. Fisher advocated for taking a step back and asking patients why they stop using insulin. His slide summarizing four such studies really supported the psychological underpinnings of adherence to insulin (e.g., “Makes me feel like a failure”; “It overwhelms me”; “It makes me feel hopeless”). Dr. Fisher concluded with eight actionable tips for improving adherence to basal insulin, including (i) a different, ongoing conversation about insulin from diagnosis; (ii) framing the insulin message accurately (not failure!); (iii) addressing fear of hypos and injection phobia; and (iv) conveying the good news about insulin (more energy, better sleep). See below for all eight strategies and the summary of what basal insulin stoppers said – as well, see Q&A between Dr. Fisher and the very respected Dr. Silvio Inzucchi (we were very impressed by his question and obvious care for patients).

  • Dr. Fisher summarized patient responses from four studies to questions like “What are the very good reasons you have for not taking this medication?” and “What is it like for you to take this medication?” (Polonsky et al., Diabetes Care 2005; Karter et al., Diabetes Care 2010; Fisher et al., Pt. Educ Counseling 2012; Rosenbaum NEJM 2015).

Summary of Common Patient Responses Across Studies

  • “Reminds me I’m sick, that I am not normal”
  • “I don’t like putting unnatural things in my body”
  • “No real benefit”
  • “I can do it myself by exercise and diet” (control)
  • “Make me feel like a failure”
  • “It make me feel ashamed that I have this disease”
  • “I am embarrassed by having to take it”
  • “I am not sure I really trust what they tell me”
  • “It makes me feel hopeless”
  • “It overwhelms me”
  • “I am afraid of the future (complications) – it scares me”
  • “I can’t manage it – it’s too complicated”
  • “I don’t like the idea that I HAVE to take it”
  • Dr. Fisher provided eight practical tips for enhancing insulin use, mostly focused on the language providers use. The strategies definitely came from the perspective of a psychologist, but were based in both scientific literature and Dr. Fisher’s clinical experience. We wonder if scripts could be developed to help HCPs use the right words when talking about insulin.  

1. Have a different, ongoing conversation about insulin from diagnosis

Be an active listener and reflect (hard for HCPs). Label, acknowledge, and accept beliefs, attitudes, and feelings – repeat and summarize. Highlight changes over time. Normalize. Don’t try and change how someone feels or believes – you’ll fail. Ask questions like: “Do you know anyone who is taking insulin? How has it been for them? How would you feel about taking insulin? What troubles you the most about insulin? What might it mean to you to take insulin?”

2. Restore a sense of personal control and choice

Provide options. Try time limited experiments (e.g., measure BGs for one week, then start taking insulin for one week and compare the blood glucose results). Ask how patients might want to test the accuracy of a belief. Distinguish between fears and realities. Point out both sides of the ambivalence (“I can see that you are worried about your diabetes. But you also have concerns about taking insulin. It seems like you are caught in the middle here”). Allow patients to direct the interaction.

3. Enhance self efficacy

Try things in the office. Schedule frequent follow-ups for selected patients. Allow patients to meet with other patients. Send reminders (they work in the short-term).

4. Frame the insulin message accurately

So often insulin has to do with failure and blame. But it’s an expected part of having diabetes, and it’s no reflection on a patient’s management. Tell patient it may be odd at first (anticipate how they might feel). Normalize concerns. Link to A1c or other health-related targets and goals to concretize the experience.


5. Address fear of hypos

Address real or imaginary fears. Normalize. Try a brief experiment as a test. Use low doses to start.

6. Address injection phobia

True injection phobia is ~2-3%. What seems like injection phobia is often a cover for other things: low self-efficacy, other beliefs, feelings, and expectations. Do office demos.

7. Convey the good news about insulin

Emphasize potential improvements in sleep, energy, highlight changes in target values, A1c, blood glucose levels, clock changes in mood over time.

8. Repeat frequently because things change

This process is key after diagnosis, when the number or dose of orals increase, and when insulin is first suggested.

  • Dr. Fisher said traditional approaches to improving adherence have relatively small effect sizes, low bang for the buck, and don’t work as well as we would like (“necessary but not sufficient”). These include behavioral (CBT), education, team-based care, social and peer support, different packaging, reminders (successful early on, but waning effectiveness over time), reducing co-pays and cost (“modest results”), and increasing the time between refill or using mail order. Dr. Fisher lamented that there has been no change in adherence after 40 years of research (per a 2014 Cochrane Review).
    • The most effective adherence strategies employ very complex multi-components target interventions. Unfortunately, Dr. Fisher said these are “impractical” and “too expensive” in real-world use.
  • No patient characteristics consistently distinguish between medication takers and non-medication takers across measures or studies (age and gender are possible exceptions). Even within the same subjects, different ways of measuring adherence yield different results (self report vs. medication monitoring with a chip in the pill bottle vs. claims data). Studies generally find a ~20% higher rate of adherence with self-report vs. medication monitoring or claims data. Medication use is also highly variable both within and across subjects, making prediction difficult. Dr. Fisher further noted that most predictors that are significant are not actionable (age, gender).

Questions and Answers

Dr. Silvio Inzucchi: There has been progress in diabetes towards less frequent administration of drugs. We have once-weekly GLP-1 and even once-weekly pills. Is there data that improves adherence?

Dr. Fisher: It improves adherence in some patients, but makes it worse in others. A number of patients wonder if the strength is sufficient to be maintained over the course of a week. They just don’t trust it. So it’s a mixed blessing – it’s helpful for some, and not helpful for others – which speaks to the need for patient-centered care.

Corporate Symposium: Evolving Perspectives in Treating Diabetes and Obesity (Sponsored by Novo Nordisk)

Cardiovascular Outcome Trials in Diabetes: What Have We Learned and Where Are We Heading?

Neil Poulter, FRCP (Imperial College London, UK)

Professor Neil Poulter offered his very candid take on the “misinterpretation and disappointment” that have characterized diabetes cardiovascular outcomes trials (CVOTs) to date. He began by contextualizing the body of epidemiological evidence that has demonstrated a causal relationship between levels of blood glucose and major cardiovascular events, citing the strong dose-response relationship that has been observed in multiple studies (e.g., EPIC-NORFOLK). However, with the exception of the recently published results of EMPA-REG OUTCOME, Professor Poulter noted that the benefits of lowering blood glucose on major cardiovascular events have been disappointing – from the results of ACCORD to more recent CVOTs – and have prompted an important question: How can we rationalize the paradoxical adverse effects of glucose lowering on cardiovascular outcomes? Professor Poulter discussed a number of possibilities: Were trial results wrong due to power or chance? Could harm been caused by other off-target damage? Were interventions too short and too small? Were the wrong populations studied? He noted that the reasons behind the underwhelming effect of glucose lowering on major CV events remain unclear and that the even more unexpected, cardioprotective results of EMPA-REG have only add to the confusion. Professor Poulter labeled the findings “a TOTAL sea change,” stressing more broadly that the unknown mechanism should encourage caution in thinking about the coming results of LEADER – much more on that below!

  • Professor Poulter cautioned that the positive findings from EMPA-REG OUTCOME should not be generalized to other CVOTs, specifically Novo Nordisk’s LEADER trial for Victoza (liraglutide). Indeed, in spite of the very positive pooled MACE analyses of liraglutide to date [HR = 0.56 (95% CI: 0.34, 0.93), Professor Poulter was strikingly adamant. He positioned the remarks not as a critique but as a realistic take based on history of disappointing and paradoxical findings. Professor Poulter pointed in particular to the initial CV outcomes analysis of sitagliptin prior to TECOS that seemed to hint at cardioprotection – an exposure-adjusted hazard ratio of 0.83 – but eventually showed “not even the slightest hint” of a benefit. The remarks came as a valuable reminder of the importance of managing sky-high expectations and resisting the temptation to overgeneralize from a single study. As a reminder, Novo Nordisk Chief Scientific Officer Dr. Mads Thomsen has repeatedly expressed cautious optimism regarding the trial’s ability to show superiority. That said, we’ve heard from many KOLs that the trial is powered for non-inferiority, not superiority, and the greater patient exposure relative to other CVOTs (mandated minimum exposure of 3.5 years per patient and total exposure of over 30,000 patient-years) may not be enough to illuminate a subtle benefit.
  • On the topic of intensive glucose control, Professor Poulter disagreed with the ADA’s decision to change the recommended A1c target in adults to >7.0% from >6.5%. Said Professor Poulter, “I suspect that this is NOT the right thing to do.” He pointed out that VADT and ADVANCE both showed no increase in mortality with intensive control, suggesting instead that the ADA has reacted too hastily to a single study. He drew an analogy to the FDA and EMA’s response to rosiglitazone adverse event data, suggesting that we may be making the same mistake twice. Certainly, we think there is still much to learn about the tradeoffs of tight vs. lenient glycemic control – e.g., how does glycemic variability play a role? – and we appreciated Professor Poulter’s effort to bring this conversation into the spotlight.
  • We found Professor Poulter’s impressions of FDA clinical trial assessments particular fascinating as he suggested that the Agency has become increasingly lenient (in an ironic sense) in the way it thinks about safety post-rosiglitazone – see image below. He opened by reviewing the various clinical trial result scenarios and the resulting likelihood of approval: (i) upper HR limit of 95% CI<1.0 -> approvable (no need for post-marketing study); (ii) 1.0<upper HR limit<1.3 -> approvable (need for post-marketing study); (iii) upper HR limit>1.8 -> not approvable. He noted that these cutoffs (HR= 1.0 vs. HR=1.3 vs. HR=1.8) are arbitrary, noting that that resulting misguided goal is to ensure no extra harm. However, he reminded us that our original thinking was that “lowering glucose should reduce CV events,” impressing upon the audience how much we have sacrificed in getting to our current conception: “If it doesn’t increase the likelihood of a CV event by more than 29% - probably – then we will take it.” Very interesting food for thought to be sure!

Managing Diabetes in Special Populations

Melanie Davies, MD (University of Leicester, UK)

Dr. Melanie Davies summarized data from clinical trials, demonstrating that GLP-1 agonists and DPP-4 inhibitors enable efficacious and well-tolerated treatment options in two specific populations of patients with type 2 diabetes: those who have concomitant mild-to-moderate renal impairment and those who are fasting during Ramadan. As expected, she drew heavily from the results of two major liraglutide trials – LIRA-RENAL and LIRA-RAMADAN. We reported results from the former at EASD 2014: the 26-week study randomized patients (n=279) 1:1 to receive either liraglutide 1.8 mg or placebo as an add-on to oral agents and/or insulin. The results showed significantly superior A1c reductions (-1.0% vs. 0.4%) and weight loss (-2.4 kg vs. -1.1 kg) with liraglutide 1.8 mg compared to placebo. Very importantly, treatment with liraglutide did not result in worsening of renal function, as there were no significant differences between the groups on change in eGFR and urinary albumin/creatinine levels. As Dr. Davies noted, the confirmation of safety is particularly valuable considering the relative lack of antihyperglycemic treatment options available to patients with renal impairment. Dr. Davies then turned to LIRA-RAMADAN, which – as a reminder – found liraglutide to be superior to SFUs in terms of A1c and hypoglycemia for individuals fasting for Ramadan (see our ADA 2015 coverage for more). The results are especially important in the context of the growing diabetes prevalence in the Middle East, and we applaud Novo Nordisk for considering the needs of this special population in a full-scale clinical trial.

Mechanisms in Cardiovascular Effects of GLP-1

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

In this presentation, Dr. Daniel Drucker illustrated the complexity of dissecting the mechanisms of cardiovascular effects of GLP-1 agonists. He opened by stressing that the class’ “single most important” risk factor is its reduction of blood pressure, as he demonstrated that such reductions have been consistently demonstrated in clinical trials of GLP-1 agonists. However, Dr. Drucker noted that the intricacy of the class’ direct and indirect actions on the heart and blood vessels makes it challenging to identify mechanisms or predict the effects, as he touched on research of GLP-1 agonists’ effects on apolipoproteins, vasodilation, diuresis and natriuresis, and more. While GLP-1 receptors are known to be expressed in cardiomyocytes, blood vessels, and platelets, the localization of these receptors remains ambiguous and it is unclear whether some of the class’ benefits seen in rodents can translate into humans. In addition, Dr. Drucker mentioned that while we can collect human data from CVOTs, the patient population studied is very well-defined and not generalizable in fully comprehending GLP-1 agonists’ cardiovascular effects. While CVOTs may provide valuable information, they certainly leave many questions unanswered as seen by the many circulating theories regarding the mechanism of cardioprotection seen in the EMPA-REG trial for Lilly/BI’s Jardiance (empagliflozin).

Corporate Symposium: New Evidence from the TECOS CV Safety Trial and the Role of DPP-4 Inhibitors in the Treatment of Patients with Type 2 Diabetes

Celebrating 20 Years of DPP-4 Inhibitor Research: The Latest Science and Where Will It Take Us?

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

Dr. Daniel Drucker reviewed the complex mechanistic evidence on DPP-4 inhibitors. He explained that the drugs’ glycemic effects are clearly mediated through GLP-1 and GIP and that they likely do not act directly on the heart, despite demonstrating cardioprotective effects with acute administration under certain circumstances. Intriguingly, he also cited recent preclinical studies suggesting promise for DPP-4 inhibitors in immuno-oncology: the drugs prevent degradation of a substance called CXCL10, which increases the number of activated immune cells directed toward tumors. This is an interesting positive counterpoint to the concerns about increased cancer risk that have plagued this class and the GLP-1 agonists over the years. Dr. Drucker himself has suggested in the past that the potential for a long-term risk of colorectal cancer is something to watch with incretin-based therapies.

Lessons Learned From TECOS Cardiovascular Safety Trial: The Role of Sitagliptin in the Treatment of Cardiovascular Patients with Type 2 Diabetes

Lawrence Leiter, MD (University of Toronto, Canada)

Dr. Lawrence Leiter reviewed the design and results of TECOS, the CVOT for Merck’s Januvia (sitagliptin). See the full results of this trial from this past ADA and an update on the safety findings from EASD, which were all presented again by Dr. Leiter. In addition, Dr. Leiter briefly walked attendees through the results of other DPP-4 inhibitor CVOTs including SAVOR-TIMI and EXAMINE, noting that they both demonstrated safety but not superiority. Concluding, he highlighted that the TECOS study was similar to SAVOR and EXAMINE, having demonstrated overall CV safety but not superiority of the DPP-4 inhibitor class in high CV risk patients with type 2 diabetes. However as expected, he noted the apparent difference in risk of hospitalization of heart failure amongst the trials as SAVOR saw a small increased risk (HR of 1.27) and EXAMINE saw a numerical increase (HR of 1.19) while TECOS saw no signal. Looking forward, he pointed to ongoing trials that may help provide additional color on these safety parameters and more for other antihyperglycemic agents.

Questions and Answers

Q: How was the diagnosis of heart failure made and who made the decision to be hospitalized?

A: The decision for hospitalization was at the treating physician’s discretion. The decision was not mandated but adjudication criteria was standardized across the study. It’s very similar across all three CVOTs. They virtually had an identical definition across the trials.

Q: Can you comment on the decline of eGFR?

A: There was a very small diff in eGFR in sitagliptin-treated patients. It’s probably not clinically significant. At the moment, we have no additional info.

Q: There were patients on sitagliptin in EMPA-REG – what did we find in those results?

A: In the EMPA-REG results, 10%-11% of patients had background therapy of DPP-4 inhibitors. The clinical benefit was similar to those on DPP-4 background vs. those who are not.

Q: We saw in EXAMINE that although there was a trend of increased heart failure, it wasn’t statistically significant in those with pre-existing heart failure. Patients without heart failure had a statistically significant increase with heart failure risk.

A: You’re correct. It’s an interesting observation, which you wouldn’t expect. But note of caution – we’re talking very small numbers and we don’t know what that means. In the SAVOR study, the relative risk was the same whether they had prior heart failure or not.

Q: TECOS was designed to be a non-inferiority trial so was it powered to prove superiority?

A: What’s interesting and confusing was that SAVOR was always advertised as a superiority trial and TECOS was advertised as non-inferiority trial. But the statistical analysis plan was identical in both studies because of the regulatory requirement of hierarchical analysis. The first analysis is safety, and the second is superiority. So neither study showed superiority.

Clinical Discussion: What Role Can CV Safety Trials Play in Helping Clinicians to Best Treat Patients with Type 2 Diabetes?

Stefano Del Prato, MD (University of Pisa, Italy); Daniel Drucker, MD, PhD (Lunenfeld-Tanenbaum Research Institute, Toronto, Canada)

The top-rate team of Drs. Stefano Del Prato and Daniel Drucker discussed how to elucidate the lingering uncertainty around heart failure risk in DPP-4 inhibitors. During the panel discussion, the audience was informally polled regarding their sentiments on the class’ heart failure signal: most attendees expressed a belief that there may be differences in the class’ products regarding this risk while no attendees voted that all DPP-4 inhibitors had increased risk or that none of the DPP-4 inhibitors had no increased risk. Comparing data from dQ&A’s Diabetes in Primary Care surveys from 2014 and 2015 shows a slight decline in PCPs’ enthusiasm for DPP-4 inhibitors as a second-line therapy after metformin, concurrent with growing enthusiasm for SGLT-2s. Addressing the differences seen in the various CVOTs for the class, Dr. Del Prato stressed the need for better data as he noted that it is very difficult to make any statements without a head-to-head trial of the class’ drugs, as he suggested that the increased risk seen with AZ’s Onglyza (saxagliptin) may be due to chance or its trial’s different patient population or adjudication process. Regarding how to make clinical decisions without this “better data,” he referred to the use of Merck’s Januvia (sitagliptin) as TECOS demonstrated clear neutrality. Meanwhile, Dr. Drucker expressed that he is unsure if even a head-to-head trial can elucidate this uncertainty, as he pointed out that the very rare nature of this adverse event makes it incredibly difficult to delineate its mechanisms. Ultimately however, providers’ moderate concerns over the class’ safety signal appear to be reassured by Januvia specifically with TECOS’ spot-on neutrality – for more on Januvia’s recent growth, please see our coverage of Merck’s 3Q15 update.

Corporate Symposium: Insulin Resistance – Core Defect in Type 2 Diabetes and Implications for Cardiovascular Disease (Sponsored by Takeda)

Combination of DPP-4 Inhibitors and TZDs: A Synergistic Treatment Approach

Guntram Schernthaner, MD (University of Vienna, Austria)

Dr. Guntram Schernthaner discussed the potentially synergistic therapeutic targets of using DPP-4 inhibitors and TZDs in combination. He argued that the two classes are complementary mechanisms of action. DPP-4 inhibitors increase the incretin effect, increase insulin secretion, and suppress glucagon secretion. TZDs, on the other hand, increase insulin sensitivity, suppress hepatic glucose production, and decrease plasma free fatty acids by suppressing lipolysis. Furthermore, there is minimal risk of hypoglycemia with the two drugs and the combination can be administered with once-daily dosing. Overall, Dr. Schernthaner concluded that the combination addresses multiple pathologies of type 2 diabetes, including pancreatic dysfunction and insulin resistance. Dr. Ralph DeFronzo’s (UT Health Science Center, San Antonio, TX) famous triple therapy involves a TZD and another incretin in the form of a GLP-1 agonist. It’s no coincidence that Dr. DeFronzo is also one of the biggest proponents for using a cocktail of drugs to address the multiple pathologies involved in type 2 diabetes. The concept appears to gaining mainstream acceptance, as evidenced by Dr. Schernthaner’s presentation (and the many other presentations at IDF) advocating combination approaches.

  • TZDs have largely fallen out of favor in the last few years, but Dr. Shernthaner defended the class on the grounds of their effect on improving insulin sensitivity and the durability of their effect (which he argued is unmatched by any other class of diabetes drugs). In addition, he highlighted that the ADA/EASD treatment guidelines continue to include TZDs on equal footing with other drug classes as potential second- and third-line options. In addition, he argued that pioglitazone has demonstrated protective effects on multiple organ, including the heart, the brain, the kidneys, and the liver.
  • Looking to the cardiovascular outcomes trial results for pioglitazone and DPP-4 inhibitor alogliptin, Dr. Schernthaner suggested that the two drugs have complementary effects and are a good option for patients with high cardiovascular risk. Alogliptin demonstrated cardiovascular safety with no increase in heart failure while pioglitazone significantly reduced the combined cardiovascular endpoint of death, MI, and stroke and the associated increase in heart failure was not associated with increased mortality.

Corporate Symposium: The Science Behind Early Combination Treatment for Type 2 Diabetes: Current Insights and Perspectives (Sponsored by AstraZeneca)

What is the Role of Glucagon in Type 2 Diabetes?

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

The highly respected Dr. Daniel Drucker discussed the dysregulation of glucagon in type 1 and type 2 diabetes and how new type 2 diabetes drug classes can potentially target these defects. He cited studies showing dysregulation of glucagon suppression very early in the progression of diabetes and others demonstrating an abnormal glucagon response in type 1 diabetes even when glycemia is normalized. With regard to current drug classes, he noted the paradoxical increase in glucagon and hepatic glucose production that occurs with SGLT-2 inhibitors and suggested that combining them with incretin-based therapies could help ameliorate those effects. The latter portion of his talk focused on the somewhat counterintuitive idea of increasing glucagon to treat type 2 diabetes. Dr. Drucker explained that glucagon can exert favorable effects on energy expenditure mediated through FGF-21, but he cautioned that many basic questions (like dose response, tachyphylaxis, and the differences between humans and rodents) remain to be addressed. He cited a recent preclinical study showing very impressive weight loss and glycemic improvements with a GLP-1/GIP/glucagon receptor triagonist but stressed that it is currently unknown whether this approach is safe or translatable to humans. Dr. Drucker’s greatest safety concern with this compound and other glucagon receptor agonists is the potential for adverse cardiac effects, as there are glucagon receptors on the heart and preclinical studies have shown harmful effects of acute glucagon administration during ischemia. Quite a few companies are developing GLP-1/glucagon dual agonists for type 2 diabetes. The class has demonstrated impressive potential in early clinical trials, but cardiovascular safety will certainly be critical to watch as they move through the pipeline.

Is Early Combination Therapy Essential in Type 2 Diabetes?

Stefano Del Prato, MD (University of Pisa, Italy)

Dr. Stefano Del Prato advocated for the use of combination therapy early in the progression of type 2 diabetes. Dr. Del Prato characterized type 2 diabetes as a progressive disease, marked by both insulin resistance and ongoing loss of beta cell function. As such, he argued that treatment intensification is needed throughout the course of the disease, but noted that “ treatment intensification” often turns into “treatment procrastination” and clinical inertia. He pointed to UKPDS data to show that no matter which diabetes drug you start with, monotherapy will always eventually fail. As an alternative, he suggested combining two or more diabetes drugs that, together, target more of the dysfunctional organs in type 2 diabetes. Dr. Del Prato emphasized the need for a “rational” drug combination that works together to address more than one pathogenic mechanism. An ideal combination would reduce hyperglycemia, reduce hypoglycemia, and either reduce or maintain body weight. In addition to providing sustained efficacy and a low rate of side effects, Dr. Del Prato suggested that a good combination would reduce clinical inertia and improve adherence, resulting in even greater real-world A1c reductions. In terms of timing, Dr. Del Prato asserted that previous trials suggest intensive combination therapy later in the progression of type 2 diabetes is less helpful since patients often already have complications at that point. In contrast, Dr. Del Prato was optimistic that early use of combination therapy could bend the association curve between duration of disease and risk for complications, with the added benefit of preventing beta cell failure. Use of combination therapy in early-onset type 2 diabetes is gaining real traction among clinicians – at a debate during EASD 2015, the vast majority of attendees voted “yes” to the question of whether or not we need triple- and quadruple-therapies in type 2 diabetes. That said, it’s unclear if payers and more cost-conscious health systems are able and/or willing to finance combinations of fairly expensive therapies.

What is the Physiological Rationale for Early Combination Therapy?

John Buse, MD, PhD (University of North Carolina, Chapel Hill, NC)

Dr. John Buse argued that the rationale for early combination therapy in type 2 diabetes is self-evident given the large and growing number of physiological abnormalities associated with the disease. He began by citing his “favorite study of all time”: a paper by Weyer et al. that identified lack of compensatory insulin secretion as the main differentiating factor between high-risk patients who progressed to type 2 diabetes and those who did not. Insulin resistance increased in both groups over time. For the purposes of this talk, Dr. Buse’s main conclusion was that type 2 diabetes by definition involves at least two physiological defects and that loss of beta cell function seems to be the key driver of progression. He also noted that as new drug classes have been developed, it has become clear that type 2 diabetes almost always involves additional defects like a decreased incretin effect, increased lipolysis, and increased glucose reabsorption in the kidneys. Dr. Buse celebrated the availability of so many new drug classes with complementary mechanisms and advocated for aggressive intensification of therapy if the approach fails. He did caution that more clinical trials are needed to validate the benefits of aggressive combination therapy right at diagnosis. Interim results from Dr. Ralph DeFronzo’s triple therapy study likely represent the best evidence to support this approach at the moment, though it is not directly assessing initial combination therapy vs. aggressive sequential therapy with the newest drugs. In terms of specific combinations, Dr. Buse spoke favorably about the potential synergistic effects of metformin/DPP-4 inhibitor combinations on GLP-1 secretion and noted that combining an SGLT-2 inhibitor with a DPP-4 inhibitor and/or metformin can improve efficacy by blunting the increase in hepatic glucose production.

How Do You Select the Right Early Combination Approach for Best Patient Outcomes?

Julio Rosenstock, MD (University of Texas Southwestern Medical Center, Dallas, TX)

Dr. Julio Rosenstock offered his thoughts on fixed-dose combinations in the early treatment of type 2 diabetes. He focused on the four available combinations: DPP-4 inhibitor/metformin, SGLT-2 inhibitor/metformin, SGLT-2 inhibitor/DPP-4 inhibitors, and GLP-1 agonist/basal insulin. He showcased clinical data for AZ’s Kombiglyze (saxagliptin/metformin) and Xigduo (dapagliflozin/metformin) demonstrating greater A1c reductions for the combination than for each component alone. He also pointed to the only SGLT-2 inhibitor/DPP-4 inhibitor combination on the market, Lilly/BI’s Glyxambi (empagliflozin/linagliptin). However, Dr. Rosenstock noted that Glyxambi is the combination of two expensive drugs, compared to metformin combinations, in which metformin is a relatively inexpensive drug. Thus, he felt that combinations with metformin were more appropriate for the early treatment of type 2 diabetes, but suggested that Glyxambi may be an option for those with metformin intolerance. Glyxambi generated significant buzz for its clinical efficacy, however Lilly has barely mentioned the product since its launch in March. Along with the fact that the product’s sales have not been broken out in the company’s earnings calls since its launch, it seems that Glyxambi has yet to gain significant market traction. Finally, Dr. Rosenstock touched on GLP-1 agonist/basal insulin fixed-ratio combinations, emphasizing that the combination can mitigate the weight gain often seen with initiation of insulin therapy.

  • Dr. Rosenstock emphasized that the ideal diabetes medication changes depending on which factor is most important to a particular patient and his/her provider. He noted that diabetes drug selection should take into account a variety of factors, including glucose-lowering efficacy, safety profile, effect on body weight, cost, and cardiovascular impact. As examples, Dr. Rosenstock suggested sulfonylureas and metformin are best to avoid high costs and GLP-1 agonists and SGLT-2 inhibitors are best for weight loss.

Management of Hyperglycemia: Prospects for Therapeutic Advances

Philip Home, DM DPhil (Newcastle University, Newcastle upon Tyne, UK)

Dr. Philip Home argued that future type 2 diabetes therapies will need to (i) alter calorie balance to achieve greater glycemic efficacy or (ii) offer significant non-glycemic benefits. He argued that most current therapies have been able to achieve only modest A1c reductions (“could I call them pathetic?”) because they either flood the system with insulin or otherwise push glucose into the liver. The liver then “fights back” by reducing postprandial glucose uptake and increasing glucose production, limiting the effect of the therapy. Dr. Home argued that only approaches that affect calorie balance will be able to achieve greater glycemic efficacy. This could include GLP-1/glucagon dual agonists (which he sees as particularly promising), SGLT-1/2 inhibitors, fat/carbohydrate absorption inhibitors (if side effects can be reduced), and potentially metabolism enhancers that target the mitochondria. Therapies Dr. Home believes are “unlikely to get anywhere” are G-protein receptor drugs, insulin receptor agonists, 11-beta HSD1 inhibitors, glucokinase activators, glucagon antagonists, and biguanide mimetics. He also suggested that a drug could still be considered successful if it matched the efficacy of current therapies and offered additional benefits like weight reduction (GLP-1 agonists), heart failure benefits (at least one SGLT-2 inhibitor), blood pressure or LDL reductions, or possibly anti-inflammatory or anti-thrombotic effects. We loved hearing this much granularity on what it will take to meet the increasingly high bar for new type 2 diabetes therapies. The role of next-generation versions (such as oral or longer-acting formulations) of existing drug classes was one aspect of innovation not addressed in this talk. While it is unlikely that such products can offer the same sort of dramatic clinical improvements, they may be able to achieve meaningful quality-of-life and adherence benefits that can themselves improve outcomes.

Corporate Symposium: Type 2 Diabetes – Is There a Benefit in Tailoring Treatment to The Asian Phenotype? (Sponsored by Bayer)

What Will The Acarbose Cardiovascular Evaluation (ACE) Trial Tell Us?

Rury Holman, FMedSci (University of Oxford, UK)

Dr. Rury Holman headlined this Bayer symposium by providing an informative overview of the possible role of the alpha-glucosidase inhibitor acarbose in reducing risk of glycemia-related CV disease. He opened by discussing the association between postprandial hyperglycemia and CV disease (STOP-NIDDM study) – “the strength of that data is stronga significant reduction was seen in CV risk but with very few endpoints ” – though acknowledged that we have yet to understand the relationship in its totality. In particular, he raised the findings of the NAVIGATOR trial (that demonstrated that treatment with the short-acting insulin secretagogue nateglinide nateglinide did not significantly reduce the incidence of either of the trial’s two co-primary cardiovascular outcomes), noting two major objections to this study’s protocopossible confounders: the relatively short duration for a glycemic intervention and the use of an insulin-promoting agent. With these lurking variables in mind, Dr. Holman suggested that there remains a gap in our understanding of acarbose as a therapeutic agent and introduced attendees to the Acarbose Cardiovascular Elevation Study (ACE), a secondary CV intervention trial which is currently evaluating the effect of acarbose treatment on CV outcomesin patients with impaired glucose tolerance. For context, Dr. Holman shared a study by Kato et al. (Cardiovascular Diabetology 2010) that suggested that acarbose has more beneficial effects on postprandial endothelial function than nateglinide and that these effects are mediated by an improvement of postprandial hyperglycemia without accompanying postprandial hyperinsulinemia. As such, the purpose of this trial is to specifically address the role of postprandial metabolism in CV disease. As a reminder the trial began in 2008 with a goal of enrolling 7,500 patients, though Dr. Holman shared that completed recruitment included only 6,526 individuals. Results are expected in 2018.

  • The Acarbose Cardiovascular Evaluation (ACE) Trial is now evaluating the effect of acarbose on CV outcomes. This multicenter, randomized, placebo-controlled, double-blind secondary prevention CV outcomes study is being sponsored by the University of Oxford, is funded by Bayer Healthcare, and is being conducted entirely in China because of the rapidly increasing incidence of diabetes in China. It will examine addition of acarbose or matching placebo three times a day to fully optimized CV disease care over a period of four years and has enrolled 6,.526 patients. Its primary endpoint is time to first occurrence of 5-point MACE (CV death, nonfatal MI, and nonfatal stroke, hospitalization for heart failure or hospitalization for unstable angina).

Corporate Symposium: Every Patient Matters – Towards Improving Population Health Care in Diabetes (Sponsored by Sanofi)

Understanding Diabetes Through a Patient and Population-Based Approach

Juliana Chan, MD (Hong Kong Institute of Diabetes and Obesity, Hong Kong)

Dr. Juliana Chan discussed the role of observational studies as a complement to randomized controlled trials, arguing that they can help bridge the gap from investigating efficacy to determining the real-world effectiveness of interventions – we certainly agree. As an example, she focused on her group’s International Diabetes Management Practice Study (IDMPS), a large observational study of diabetes care in the developing world. The ongoing study includes over 72,000 patients and over 5,000 healthcare providers from 48 countries where the diabetes epidemic is expected to explode in the near future. It began in 2005 and includes six “waves” of investigation in different areas: (i) achievement of targets for screening, education, and treatment; (ii) health resource use; (iii) barriers to insulin use; (iv) the experience of hypoglycemia; (v) mental health issues including depression, anxiety, and stress; and (vi) self-management. The group has published a number of papers with results from the first phases of the study. For example, the first phase found that long duration of diabetes and younger age were key risk factors for not achieving A1c targets and that diabetes education led to improvements in almost all parameters measured. The second phase found, as one might expect, that complications and poor control were associated with higher health resource use. The third phase identified several factors associated with severe hypoglycemia in type 1 diabetes, including lower education, complications, and use of premix insulin. Interestingly, it also found that severe hypoglycemia was associated with higher A1c values – the opposite from the relationship in many clinical trials. The fourth phase identified dosing difficulties, risk of hypoglycemia, and side effects as main barriers to insulin initiation; it also noted that patients’ main concerns about insulin change once that initiation barrier has been crossed.


Lori Laffel, MD (Joslin Diabetes Center, Boston, MA)

Dr. Lori Laffel reviewed results from the TEENS study of type 1 diabetes in US adolescents. She began by describing the difficult transition all adolescents with diabetes face, from a stage where the entire family essentially has diabetes to one where the patient is primarily responsible for daily management. She noted that even the “haves” – children with good access to resources and very involved parents – face significant emotional challenges meeting the demands of diabetes management. Of course, the “have nots” – those who have to worry about basic resources and may not have access to tools like blood glucose meters – face a different sort of emotional burden. These challenges are illustrated by the fact that very few people with type 1 diabetes between the ages of 18 and 25 achieve glycemic targets. In the TEENS study, 72% of patients were not at their target A1c, and one in five had an A1c >10%. The study found that factors associated with a higher A1c included female gender, puberty, and longer duration of diabetes. Relevant family factors included living in a single-parent home, low parental education, high family conflict, and parents who had to stop working to care for the child (though that is more likely an effect rather than a cause of the poor control). Not surprisingly, lower A1c was associated with better quality of life in the study, and many of the same factors associated with better glycemic control were also associated with better quality of life. Many of the risk factors identified were not modifiable, but Dr. Laffel singled out diabetes-specific family conflict as the one that should be the most effective to target.

Corporate Symposium: Novel Mechanisms and Advancing Therapeutic Paradigms for Optimizing LDL-focused Management and Cardioprotection in the Diabetic Patient (Sponsored by CMEducation Resources)

Evolving Clinical Paradigms for LDL Lowering: Focus on Efficacy, Implications, Side Effects, and Safety Results of Landmark Trials Evaluation PCSK9 Inhibitions, APOB Suppression, and Other Novel Targets – Implications for the Diabetic Patient

James McKenney, PharmD (National Clinical Research, Richmond, VA)

Dr. James McKenney reviewed the mechanism of action and phase 3 clinical data for PCSK9 inhibitors to familiarize the non-cardiologists in the room with this exciting new drug class and its potential implications for their patients with diabetes and cardiovascular disease. Dr. McKenney noted that the FDA indication was narrower than the EU indication and suggested that the FDA’s aim was to “get its feet wet” and see what the cost to the health system looks like before expanding to a wider indication. We assume positive results from the ongoing cardiovascular outcomes trials would almost certainly lead to expanded indications. He emphasized that the two PCSK9 inhibitors on the market – Sanofi’s Praluent (alirocumab) and Amgen’s Repatha (evolocumab) – are identical in terms of efficacy. Praluent is slightly behind in the battle for formulary access at the moment, and this will likely be the major determining factor in the two products’ success given the lack of clinical differentiation. However, Dr. McKenney highlighted the different dosing options each company decided to offer. Praluent is available in a 75 mg dose and a 125 mg dose, both taken bi-weekly, and the 125 mg dose has greater LDL-lowering efficacy. On the other hand, Repatha offers flexibility in its dosing schedule with either 140 mg bi-weekly or 420 mg (taken as three 140 mg injections) once-monthly. Looking to the future, Dr. McKenney also whetted the audience’s appetite with a peek at a small interfering RNA (siRNA)-based PCKS9 inhibitor that has demonstrated persistent LDL lowering after four to six months in phase 1 studies, raising the possibility of twice-yearly dosing.

  • Dr. McKenney also emphasized that PCSK9 inhibitors have additive – though not synergistic – effects on top of statins. In a different presentation, Dr. Harry Ginsberg (Columbia University College of Physicians and Surgeons, New York, NY) echoed this sentiment and, like Dr. McKenney, drove home the point that the new class should be used on top of current standard-of-care statin therapy.
  • While PCSK9 inhibitors are not a “diabetes drug” in the strictest sense, Dr. Christoph Wanner (University Hospital of Würzburg, Germany) argued in favor of their appropriateness for many patients with diabetes in a separate talk. He emphasized that more and more individuals with diabetes and/or obesity are entering PCSK9 trials, suggesting that the drugs are appropriate for many within the diabetes patient population. Indeed, in a comparison, Dr. Wanner found the participant population of one of the phase 3 trials for Praluent to be quite similar to the participant population of EMPA-REG OUTCOME.


Liraglutide vs SGLT-2 Inhibitors in People with Type 2 Diabetes: A Network Meta-Analysis (0266-P)

M Lorenzi, U Ploug, J Langer

In a meta-analysis, Victoza (liraglutide 1.8 mg) demonstrated greater A1c reductions than SGLT-2 inhibitors in patients with type 2 diabetes on metformin. The analysis (n=17 RCTs) found that both doses of Victoza demonstrated greater mean reductions in A1c than any dose of any of the three approved SGLT-2 inhibitors. Victoza was also associated with a greater percentage of patients reaching an A1c goal of ≤7%. However, weight loss was similar between Victoza and the SGLT-2 inhibitors, with the higher 300 mg doses of J&J’s Invokana (canagliflozin) and 25 mg doses of Lilly/BI’s Jardiance (empagliflozin) producing larger weight reductions that the lower 1.2 mg dose of Victoza.

Insulin Degludec/Insulin Aspart Lowers Fasting Plasma Glucose and Hypoglycemia Rates Independent of Baseline HbA1c, Disease Duration, and Body Mass Index (0336-P, 0337-P, 0338-P)

H Rodbard

Three posters showed that Ryzodeg (insulin degludec/insulin aspart) produced lower rates of overall and nocturnal hypoglycemia than NovoMix 30 and/or basal-bolus insulin degludec and insulin aspart. This was true regardless of baseline A1c, disease duration, or BMI. Patients treated with Ryzodeg also had lower fasting plasma glucose vs. both comparators. Ryzodeg-treated patients with a BMI ≤30 or a disease duration longer than 10 years saw significant differences in insulin dose at the end of the trials as well.

Patient/Physician Inertia in Insulin Intensification for Patients with Uncontrolled Type 2 Diabetes Using Basal Insulin; Perceptions of Control Among Type 2 Diabetes Patients Treated with Basal Insulin (1081-P; 0742-P)

M Brod, KM Pfeiffer, AH Barnett

The Perceptions of Control study identified barriers to insulin therapy intensification in patients with poorly controlled type 2 diabetes on basal insulin from the patient and physician perspective. The most common barriers cited by patients in the web-based survey were fear of weight gain (45%), the perception that intensification would signal disease progression (44%), fear of hypoglycemia (41%), and reluctance to add more injections (41%). Overall, 57% of patients surveyed stated they were only somewhat willing or not at all willing to initiate  bolus insulin. The primary source of physician reluctance to intensify was the belief that their patients would not agree to intensification (noted by 62% of physicians). Other top reasons included hypoglycemia concerns (46%), mental illness (48%), poor cognitive skills (46%), and concerns about patient adherence (41%) as barriers to intensification. This study dovetailed nicely with Dr. Lawrence Fisher’s (UCSF, San Francisco, CA) actionable talk on barriers to insulin initiation yesterday. The two together emphasize the need for better provider education on how to talk about insulin, and perhaps better patient-provider communication in general.

Post-Prandial Hyperglycemia and Impact on Healthcare Resource Use, Diabetes Management, Missed Work Time and Reduced Productivity (0720-P, 1074-P, 0956-P, 0721-P)

Four posters examined the physical, emotional, and economic implications of post-meal hyperglycemia. 27% of respondents with recent post-meal hyperglycemia in one survey reported missing work time and a whopping 71% reported decreased work productivity. Another survey found that people with post-meal hyperglycemia had significantly more healthcare visits (5.5 visits in the past year vs. 4.4 visits for patients without post-meal hyperglycemia). The same study found that those with post-meal hyperglycemia also measured their blood glucose more frequently (1.9 extra measurements per day vs. 1.2, p<0.001) and were more likely to report diabetes-related complications. A third study found that contributing factors for post-meal hyperglycemia included stress (in 27% of respondents), eating out at a restaurant (25%), being busy (21%), and feeling tired (19%). Physical and emotional consequences included tiredness, dizziness, and feelings of demoralization, lack of sociability, and irritability.

Efficacy and Safety of Once Weekly Dulaglutide Added on to Sulfonylurea in TYpe 2 Diabetes (AWARD-8) (0219-P)

K Dungan, R Weitgasser, FP Manghi

Lilly presented a new poster from its phase 3 AWARD-8 study, demonstrating that Trulicity (dulaglutide) as an add-on to sulfonylurea is more effective than a sulfonylurea alone. The randomized, double-blind, placebo-controlled trial (n=300) evaluated the efficacy and safety of treatment with dulaglutide vs. placebo in sulfonylurea-treated patients over 24 weeks. From a baseline A1c of 8.4%, participants in the dulaglutide-treated group experienced a 1.3% greater A1c reduction than those in the placebo group (p<0.001). In addition, those treated with dulaglutide were more likely to reach a target A1c <7% (55% of those in the dulaglutide-treated group vs. 19% in the placebo group; p<0.001). Dulaglutide treatment significantly improved fasting plasma glucose over placebo (p<0.001) as well. The dulaglutide group showed some weight reduction (-0.9 kg), although the between-group weight change was not statistically significant. As expected, the most common adverse events associated with dulaglutide treatment were gastrointestinal-related (10.5% nausea and 8.4% diarrhea), leading to six discontinuations in the dulaglutide group vs. none in the placebo group. Hypoglycemia rates were also higher in the dulaglutide-treated group (p=0.025), although the overall incidence remained low at 11.3%. Overall, the trial demonstrated that dulaglutide is a safe and effective treatment when added on to sulfonylureas – while this finding isn’t particularly surprising, it could lend confidence to providers who may have been previously hesitant to prescribe the two together.


-- by Melissa An, Adam Brown, Helen Gao, Varun Iyengar, Sarah Odeh, Emily Regier, Ava Runge, and Kelly Close