European Association for the Study of Diabetes (EASD)

September 14-19, 2014: Vienna, Austria – Full Report – GLP-1 Agonists

Executive Highlights

GLP-1 agonists stole the show at EASD 2014 on the therapy front, due in large part to the recent and upcoming regulatory decisions for once-weekly agents within the class: GSK’s Eperzan/Tanzeum (albiglutide) was approved in March in the US and the EU, Lilly’s Trulicity (dulaglutide) is likely to be approved in the EU in a couple months following a positive CHMP opinion in late September (it was approved by the FDA during EASD), and a new pen for AZ’s Bydureon (exenatide) is also likely to receive EU approval soon following a positive CHMP opinion in July. During the four well-attended oral presentation sessions devoted to GLP-1 agonists, we saw full data for the first time from the phase 3 Harmony 1 trial for GSK’s new once-weekly candidate Tanzeum/Eperzan (albiglutide), as well as results from the DURATION-NEO-1 trial demonstrating greater A1c reductions and improvements in patient satisfaction with a once-weekly suspension formulation of AstraZeneca’s Bydureon (exenatide) delivered by auto-injector, compared to Byetta (exenatide twice daily).

Additionally, another key point that emerged during the conference was the increasing heterogeneity within the GLP-1 agonist class. We strongly believe that the rise of weekly products – AZ’s Bydureon, GSK’s Tanzeum/ Eperzan, Lilly’s Trulicity, and (in the future) Novo Nordisk’s semaglutide – will sharpen the distinction between short- and long-acting GLP-1 agonists in patients’ and providers’ minds. As noted, ease of administration should provide an important point of differentiation for patients (we found Trulicity’s single-use pen remarkably patient-friendly) and will likely set a higher precedent for the level of convenience expected with future products. We also suspect that we will see an increasing number of novel applications for GLP-1 agonists in the future as more options become available – at a Lilly corporate symposium, Professor Anthony Barnett (Heart of England NHS Foundation Trust and University of Birmingham, Birmingham, UK) expressed excitement about the potential of GLP-1 agonist/SGLT-2 inhibitor combinations, which we have heard discussed more and more at recent meetings, and Dr. Juris Meier (St. Josef Hospital, Bochum, Germany) highlighted expanded indications for type 1 diabetes and obesity as valuable short-term goals for the class.

As we saw through oral presentations and corporate symposia at EASD, GLP-1 agonists are increasingly being positioned as a potential alternative to basal insulin for patients not at goal on oral medications alone – this is an interesting dynamic given the wave of enthusiasm for the use of GLP-1 agonists and basal insulin in combination. Indeed, for type 2 patients, we think GLP-1 makes a lot of sense as the “first injectable” and as combination GLP-1/insulin products become available, they would make sense as the “first injectable” for patients that are far enough from their glycemic targets that GLP-1 alone may not be enough.

Turning to fixed-dose combinations, we saw substantial discussion of the favorable data on Novo Nordisk’s Xultophy (IDegLira; insulin degludec/liraglutide) and general enthusiasm for fixed-dose combinations of insulin and GLP-1 agonists. The focus on Xultophy conveniently accompanied the drug’s timely European approval on Day #4 of EASD. We also heard enthusiasm for insulin-incretin combination therapies from Drs. Stephen Bain (Diabetes Research Network, Wales, UK) and Sultan Linjawi (Coffs Endocrine & Diabetes Services, Coffs Harbour, Australia) at Novo Nordisk’s massive corporate symposium on Monday. Xultophy is the most advanced insulin-incretin combination, but it is not alone: in an oral presentation on Sanofi’s LixiLan (insulin glargine/lixisenatide), the combined benefits on A1c, weight, and hypoglycemia were termed “miraculous” by an audience member during Q&A. We have sensed excitement for this up-and-coming class of combinations for some time now (the esteemed Dr. John Buse [University of North Carolina School of Medicine, Chapel Hill, NC] extolled its benefits in a memorable presentation at last year’s Cardiometabolic Health Congress), and we are excited to see it (via Novo Nordisk’s Xultophy launch) finally nearing patients’ hands.

Titles highlighted in blue are new additions that were not mentioned in our daily updates from Vienna, and those highlighted in yellow represent what we felt were the most notable talks of the meeting.

Below you will find our detailed reporting on GLP-1 agonists.

Table of Contents 

GLP-1 Agonists

Oral Presentations: GLP-1 Analogues – Clinical Efficacy

Harmony 1 Year 3 Results: Albiglutide vs. Placebo in Patients with Type 2 Diabetes Mellitus Not Controlled on Pioglitazone (Pio) ± Metformin (Met)

Christopher Perkins, MD (Boston Medical Center, Norwood, MA)

Dr. Christopher Perkins presented full three-year data from the phase 3 Harmony 1 trial for GSK’s new weekly GLP-1 agonist Tanzeum/Eperzan (albiglutide). GSK ran a remarkably long phase 3 program (five of the phase 3 trials ran for at least three years), but studies that run for that long can be somewhat tricky to interpret due to trial dropouts. Harmony 1 tested the lower 30 mg dose of albiglutide vs. placebo in patients on background pioglitazone with or without metformin. The placebo-adjusted A1c difference between groups at one year was -0.8%, which narrowed to -0.4% at year three (from a baseline of 7.6-8.1%). This group was self-selected, as patients that had required hyperglycemic rescue (45% of the albiglutide group and 72% of the placebo group) were not included. In a full intent-to-treat analysis (rescued patients continued on their blinded therapy), the A1c benefit at week 156 with albiglutide was a much more respectable 0.6% (keeping in mind that this study tested the lower albiglutide dose). Consistent with other phase 3 results on albiglutide, there was not much of a change in weight, but also a relatively low incidence of GI tolerability issues, both of which are differentiating factors against the rest of the GLP-1 agonist class. Less nausea may make HCPs like it more since nausea is a pain in the next to deal with for HCPs. Tanzeum was recently launched in the US, and we learned at the exhibit hall that the first EU launches might now be expected early next year.

  • Harmony 1 randomized 299 type 2 diabetes patients on pioglitazone ± metformin to either albiglutide 30 mg (the lower of the two approved doses) or placebo. The study’s primary endpoint was after one year, but the full study ran for three years to investigate long-term efficacy and safety. We have to hand it to GSK for pursuing such an ambitiously long phase 3 program. Following the second week of treatment, patients could receive rescue therapy if they were still experiencing marked hyperglycemia. Uniquely, patients receiving rescue were allowed to remain on their blinded randomized therapy following rescue. As a result, there were two sets of analyses presented: an analysis of the pool of completers not including those that received rescue, and an expanded pool that included the patients that received rescue therapy.
  • At baseline, average patient age was 55 years, ~60% were male, and average BMI was 34 kg/m2. Slightly above one third of patients that were randomized discontinued active treatment during the three years of the full trial. The majority (80%) of patients were on pioglitazone + metformin at baseline.
  • Because of the long duration of the trial and the fact that patients requiring hyperglycemic rescue were kept in the efficacy pool, the results were presented in three ways: (i) An intent-to-treat, last-observation-carried-forward (ITT-LOCF) analysis at 52 weeks, the primary endpoint; (ii) An intent-to-treat, observed cases analysis (ITT-OC) out to year 3 in patients not receiving rescue therapy; and (iii) An intent-to-treat, observed cases (ITT-OC) analysis out to year 3, including patients that received rescue therapy.
    • ITT-LOCF at week 52 (primary endpoint): The albiglutide group achieved a mean A1c reduction of -0.8% from baseline (8.1%) whereas the placebo group achieved a reduction of -0.1%, for a placebo-adjusted difference of -0.8% (p<0.0001). A1c levels in both groups had stabilized by approximately week 12, and were relatively steady out to week 52. In this time period, 44% of the albiglutide group achieved an A1c goal of <7%, while only 15% of the placebo group achieved that goal.
    • ITT-OC (excluding rescued patients) at year 3: In this analysis, the efficacy difference between the two groups narrowed to -0.4%, although barely 25% of patients that were initially randomized remained in this pool by year 3.
    • ITT-OC (including rescued patients) at year 3: First, Dr. Perkins showed that 55% of albiglutide patients were rescue-free at year 3 compared to only 28% of the placebo group – this finding alone speaks to albiglutide’s improved efficacy over placebo. In this analysis, over half of patients initially randomized were in the trial, and the A1c difference was 0.6%.
  • In the pool that included rescued patients, body weight was unchanged in the albiglutide group and rose by 2 kg in the placebo group – this was likely due to the long duration of the trial, as well as pioglitazone’s weight gain effects.
  • Notably, consistent with other Harmony trials, the incidence of GI side effects was relatively low – 39% in the albiglutide group vs. 36% in the placebo group over the full three years. Discussion during GSK’s symposium on day #1 had touched upon the point that reported cases of nausea are usually high in the placebo group for GLP-1 agonist trials, given patients’ expectations.

Efficacy and Safety of Once Weekly Dulaglutide vs. Insulin Glargine in Combination with Metformin and Glimepiride in Type 2 Diabetes Patients (AWARD-2)

Francesco Giorgino, MD, PhD (University of Bari Aldo Moro, Bari, Italy)

Dr. Francesco Giorgino presented the results of the AWARD-2 trial, which compared Lilly’s newly-approved (in the US) once-weekly GLP-1 agonist Trulicity (dulaglutide) against Sanofi’s basal insulin Lantus (insulin glargine) added on to maximally tolerated doses of metformin and a sulfonylurea in type 2 diabetes patients for 52 and 78 weeks. These results were first presented at this year’s ADA. Dulaglutide 1.5 mg (the highest of the two doses tested) led to a superior reduction in A1c from baseline than insulin glargine at the 52 week primary endpoint (-1.1% vs. -0.6%) and at the final 78-week time point (-0.9% vs. -0.6% with insulin glargine). Dulaglutide 0.75 mg was non-inferior to insulin glargine at both time points. Dr. Giorgino also presented data of profile-point SMPG profiles at baseline and 52 weeks, showing that the higher dulaglutide dose provided better glucose control throughout the day compared to dulaglutide 0.75 mg and insulin glargine. Although we wish these had been CGM results, we think the eight-point profile reinforces a higher quality A1c for dulaglutide – because the measures of “time below zone” and “time above zone” will receive increased attention and focus in our view, we think the eight point profile will resonate with HCPs (although it’s harder to explain than “time in zone” in our view).  Both doses of dulaglutide resulted in weight loss and greater improvements in A1c with less hypoglycemia; safety-wise, dulaglutide had higher rates of GI-related adverse events compared to glargine, which is expected of the GLP-1 agonist class. We believe Trulicity will do very well commercially and think there is a battle ahead vs. Novo Nordisk’s well-established daily Victoza.

Efficacy and Safety of Once Weekly Dulaglutide Versus Once Daily Liraglutide in Type 2 Diabetes (AWARD-6)

Santiago Tofé Povedano, MD (Clínica Juaneda, Palma de Mallorca, Spain)

Dr. Santiago Tofé Povedano gave the first full data presentation from the AWARD-6 head-to-head trial comparing Lilly’s once-weekly GLP-1 agonist Trulicity (dulaglutide) and Novo Nordisk’s once-daily Victoza (liraglutide) – we first saw results from this study presented as a poster at ADA. As a reminder, the study found comparable A1c reductions (-1.4%) from baseline (8.1%) with both Trulicity and Victoza. New from the poster, we saw data on pancreatic enzymes from the trial – mean lipase levels saw a significantly greater increase in the liraglutide group than in the dulaglutide group (p=0.012), and there was a trend towards more patients experiencing treatment-emergent abnormal changes in pancreatic enzymes (p=0.052), although there were no adjudicated cases of pancreatitis in the trial. As a reminder, the main difference between the groups seen in the original poster data (also included in this presentation, and discussed during Q&A) was a numerically modest but statistically significant (p=0.01) difference in weight loss: the Victoza group lost an average of 3.6 kg (~8 lbs) while the Trulicity group lost an average of 2.9 kg (~6 lbs).

Questions and Answers

Q: How do you explain the weight difference you observe between dulaglutide and liraglutide?

A: We have no explanation so far. Probably, there is a different mechanism of action in gastric emptying or maybe at the central nervous system, but this is something that should be explored.

Q: Do you have lipase data for dulaglutide?

A: There was a nearly significant increase in lipase in the liraglutide relative to dulaglutide, with a p-value of 0.052, but that did not lead to any pancreatitis.

Better Glycemic Control and Less Weight Gain with Once Weekly dulaglutide Vs. Bedtime Insulin Glargine, Both Combined with Thrice Daily Lispro, in Type 2 Diabetes (AWARD-4)

Johan Jendle, MD, PhD (Karlstad Central Hospital, Karlstad, Sweden)

Dr. Johan Jendle presented results, previously presented at ADA, from the phase 3 AWARD-4 trial comparing Lilly’s once-weekly GLP-1 agonist Trulicity (dulaglutide) to Sanofi’s once-daily basal insulin Lantus (insulin glargine), both in addition to Lilly’s Humalog (insulin lispro). The open-label trial (n=884 patients with type 2 diabetes) was the first to explore the use of a GLP-1 agonist with mealtime insulin – the therapeutic relevance of this trial is perhaps slightly less than that of AWARD-2 (also presented in this session), given that rapid-acting insulin is generally not patients’ first injectable therapy although this certainly would be relevant for patients that present with very high A1cs (diagnosed late). Both the 1.5 mg and 0.75 mg doses of Trulicity led to significantly greater A1c reductions (1.64% and 1.59%, respectively) compared to Lantus (1.41%) at 26 weeks, and superiority was maintained during the subsequent 26-week extension phase. Trulicity was also associated with less weight gain compared to Lantus (0.3 kg and 1.6 kg with the higher and lower doses of Trulicity vs. 3.7 kg with Lantus), and the rate of hypoglycemia was lower (44 events/patient/year) with the 1.5 mg dose of Trulicity and similar (53 events/patient/year) with the 0.75 mg dose compared to Lantus (63 events/patient/year).

  • As a reminder, Trulicity was recently approved by the FDA, and a US launch is expected later in 2014. The product will be the first ready-to-use once-weekly GLP-1 agonist to reach the market, and the single-use pen, which allows quick administration without a visible needle, stands to be a major leap forward in terms of patient comfort with injectable therapy for diabetes. The product is still under review in Europe.

Questions and Answers

Q: Do you have data on fasting and postprandial glucose, like seven-point profiles?

A: We have a sub-analysis with CGM data to explore further. Glargine was better in the early morning, but from midday and thereafter, there were significantly lower glucose values over the course of the day for both dulaglutide arms.

Efficacy and Safety of Liraglutide Vs Placebo When Added to Basal Insulin Analogues in Subjects With Type 2 Diabetes (LIRA-ADD2BASAL): A Randomized, Placebo-Controlled Trial

Jorma Lahtela, MD (Tampere University Hospital, Tampere, Finland)

Dr. Jorma Lahtela presented results, previously presented at ADA, from the 26-week LIRA-ADD2BASAL trial evaluating the safety and efficacy of Novo Nordisk’s Victoza (liraglutide) as an add-on to basal insulin (with or without metformin) in patients with poorly controlled type 2 diabetes. At 26 weeks, the Victoza group saw significantly greater reductions in A1c (1.3% vs. 0.1% from a baseline of 8.2-8.3%) and fasting plasma glucose (1.44 vs. 0.16 mmol/l [25.9 vs. 2.9 mg/dl]), perhaps unsurprising given that insulin dose was capped at baseline levels in both groups. Victoza also improved body weight (3.54 kg vs. 0.42 kg [8 lbs vs. 1 lb]) compared to placebo. Surprisingly, in addition to increases in GI adverse events and lipase levels, Victoza was associated with a higher incidence of documented symptomatic hypoglycemia (30.7% vs. 20.4%). A slight increase is to be expected even with the addition of a relatively hypo-friendly drug class such as the GLP-1 agonists, but we wonder if insufficient down-titration of basal insulin dose in the Victoza arm might have exacerbated the increase slightly.

Questions and Answers

Q: The number of patients with increased lipase was three times higher in the liraglutide group. Has this been seen in other studies, and is it concerning?

A: There was an increase in lipase values, but there was no pancreatitis and no clear explanation for the increased lipase. The results were the same as in other studies with liraglutide.

Q: Would it be best to add a short-acting or long-acting GLP-1 agonist to basal insulin? Especially with an A1c <8%, I know postprandial glucose is the greater problem.

A: According to this data, I can’t answer your question, but postprandial glucose decreased the same amount as fasting. But if your question is whether a short-acting GLP-1 agonist would be better, I can’t answer that.

Q: You showed that 59% of patients reached an A1c under 7%, but previous studies showed even more. If you escalated the basal dose more, could you get more people to target? You reduced it at the beginning and didn’t escalate it further, which suggests there was still room for improvement.

A: The basal insulin dose was reduced for those with an A1c <8% and titrated back according to blood glucose. It was a randomized study, and the people taking care of the patients didn’t know which group the patients were in. The insulin dose was less in the placebo group after titration than in the beginning, so it’s possible that if you titrate higher, we can get more people to target.

Q: Were there any non-adjudicated cases of pancreatitis?

A: There was no pancreatitis at all.

Q: I’m concerned about the choice of a placebo control group because they had no improvement of A1c – it was 8% at the start and the end, which might not be surprising. You’re comparing active treatment to no active treatment when you have many other options with efficacy. Why use placebo?

A: It was a placebo group and the insulin dose was not allowed to increase to higher than before the trial, so that explains why there was only a modest change in A1c.

Oral Presentations: Clinical Studies with GLP-1 Analogs

DURATION-1 Extension: Efficacy and Tolerability of Exenatide Once Weekly Over 6 Years in Patients With Type 2 Diabetes Mellitus

Eric Klein, MD (West Olympia Internal Medicine, Olympia, WA)

Dr. Eric Klein presented the six-year follow-up results from the DURATION-1 trial, which is currently the longest assessment of a GLP-1 agonist to date. Four-year results were presented at EASD 2012. The first year of the study compared Bydureon (exenatide once weekly) to Byetta (exenatide twice daily), but following one year all remaining patients were moved to Bydureon – as a result, the six-year results are a comparator-less look at safety and efficacy with Bydureon. In a last-observation-carried-forward analysis, Bydureon led to a 1.6% mean A1c reduction at year six from a baseline of 8.2% - for comparison, mean A1c change was -2.2% at year one and -1.8% at year five. These results are still notable and at first glance reinforce the long duration of GLP-1 (perhaps particularly Bydureon – at once weekly, the adherence may be higher, which may be reflected in the results) – although we learned in the presentation that providers could add on other agents (see below) so it’s a bit hard to tell the singular impact of Bydureon. Mean weight loss at year six was four kg (approximately nine pounds). The 3 bpm increase in heart rate seen at year one did not worsen during the trial, and in fact diminished slightly to 2 bpm at year six. The incidence of nausea during the extension period (week 30 – year six) was significantly lower than the rate seen during the initial 30-week controlled period (0.08 vs. 0.85 events/patient-year).

  • A few key factors impact the interpretation of the results: 59% of the intent-to-treat population discontinued Bydureon therapy before the end of year six, with nearly half of discontinuations due to withdrawal of consent (patients were not compensated during the last five years of the study – this is unfortunate) and about a fifth due to adverse events. Additionally, at year three, providers were allowed to add on other glucose lowering agents, which most did. This is likely to have inflated the efficacy seen in the trial from years three to six, possibly quite substantially. 

One-Year Efficacy and Safety of IDegLira in patients with Type 2 diabetes

Stephen Gough, MD, PhD (University of Oxford, Oxford, UK)

Dr. Stephen Gough presented results, previously presented at ADA, from a 26-week extension of the original 26-week DUAL-1 study, which compared the profile of Novo Nordisk’s recently-approved (in Europe) Xultophy (IDegLira; insulin degludec/liraglutide) to that of its individual components, Tresiba (insulin degludec) and Victoza (liraglutide). The 52-week results were very impressive and remarkably consistent with the original 26-week findings: compared to Tresiba, IDegLira was associated with significantly greater A1c reductions (1.8% vs. 1.4% with Tresiba and 1.2% with Victoza), 37% less hypoglycemia than Tresiba, and fewer GI adverse events than Victoza. There were no significant differences in adverse events between the two groups.

Questions and Answers

Q: Were there any cases of non-adjudicated pancreatitis?

A: Any case of pancreatitis went to an external adjudication committee. There were two adjudicated cases: one in a patient with acute renal failure and one in a patient who had a carcinoma that was clearly advanced at the time.

Effect of Exenatide on Postprandial Cerebral and Liver Glucose Metabolism: A Double-Blind Randomized Clinical Trial

Amalia Gastaldelli, PhD (Institute of Clinical Physiology, Pisa, Italy)

Dr. Amalia Gastaldelli presented data, previously presented at ADA, on the effects of exenatide (AstraZeneca’s Byetta/Bydureon) on cerebral glucose uptake, a rather poorly understood topic compared to GLP-1 agonists’ more extensively studied effects on the pancreas and GI tract. Results from the double-blind, randomized, crossover study (n=15, baseline A1c=5.7%) demonstrated that the rate of glucose uptake in the brain as a whole and in several regions implicated in food intake/appetite regulation was significantly higher with exenatide compared to placebo after an oral glucose tolerance test. Other data confirmed exenatide’s expected metabolic effects, such as delayed gastric emptying (which was strongly correlated with increased glucose uptake in the brain) and suppression of endogenous glucose production (which was not correlated with the cerebral effects). Dr. Gastaldelli closed by sharing data not included in the ADA presentation that demonstrated increased hepatic glucose uptake with exenatide in addition to the higher rates of uptake in the brain. These results confirm that GLP-1 agonists’ actions are distributed across a range of organ systems throughout the body. It would be interesting to know if cerebral glucose uptake, in addition to removing glucose from the blood, has any impact on neural circuits and behavior.

Questions and Answers

Q: The issue of brain metabolism of glucose is of interest in relation to animal papers demonstrating some effect of GLP-1 on Alzheimer’s. What explains the increased metabolism of glucose in the brain? Is it a faster transfer over the blood-brain barrier or is it just metabolism?

A: That’s a tough question. We were measuring glucose metabolism, so this shows an effect on glucose metabolism in the brain mediated by GLP-1. The GLP-1 has to pass through the blood-brain barrier, but it’s a difficult question to answer. We need more study. For sure, studies done in Alzheimer’s or Parkinson’s show an improvement in cognitive function using GLP-1.

Q: Did you explore any correlation between changes in the different brain areas to infer possible associations?

A: We haven’t done this yet because we just completed the analysis, but we will do this.

Impact of Baseline Gastric Emptying on Effects of Lixisenatide and Liraglutide in Type 2 Diabetes (T2DM) as an Add-On to Insulin Glargine

Juris Meier, MD (St. Josef Hospital, Bochum, Germany)

Dr. Juris Meier presented in-depth results examining how gastric emptying (both at baseline and the change in gastric emptying seen during treatment) impacted safety and glycemic parameters with Sanofi’s short-acting GLP-1 agonist Lyxumia (lixisenatide) and Novo Nordisk’s long-acting Victoza (liraglutide). These results are a follow-up to results presented as a poster at ADA demonstrating that lixisenatide had a more pronounced effect than liraglutide on slowing gastric emptying. The two main takeaways from this presentation, in our view, were that: (i) Baseline gastric emptying speed did not predict treatment response with either agent, and (ii) The delay in gastric emptying with lixisenatide was lower in patients with slower gastric emptying at baseline, suggesting that the risk of aggravating pre-existing problems with gastric emptying is relatively low.

Questions and Answers

Q: The results you see here are probably the results you see over the one meal when you inject the drug. Would you expect to see changes over the latter two meals?

A: We have focused here on breakfast, the meal for which lixisenatide was administered. We would expect that the mechanism of glucose lowering, especially with lixisenatide, would be different after subsequent meals. We are performing a study were we are looking at these phenomena, comparing what happens at the first vs. the last meal of the day. Hopefully next year we will have an answer.

Q: Was there any tachyphylaxis of the effect?

A: We believe that there is tachyphylaxis of the gastric emptying effects with long-acting GLP-1 agonists like liraglutide and exenatide LAR. We don’t think we see tachyphylaxis with the short-acting agents.  

Oral Presentations: GLP-1 Analogs – Non-Glycemic Endpoints

Effect of Liraglutide 3.0/1.8 mg on Body Weight and Cardiometabolic Risk Factors in Overweight/Obese Adults with Type 2 Diabetes: SCALE Diabetes Randomized, Double-Blind, 56-Week Trial

Bruce Bode, MD (Emory University, Atlanta, GA)

Dr. Bruce Bode presented positive data on cardiometabolic risk factors from the SCALE Diabetes trial of liraglutide 3.0 mg for obesity (Saxenda). Results showed that treatment with liraglutide 3.0 mg led to significant improvements in fasting lipids (4% reduction in total cholesterol, 14% reduction in triglycerides, p<0.05) compared to placebo. Both liraglutide 3.0 mg and liraglutide 1.8 mg were associated with significantly greater reductions in levels of C-reactive protein (hsCRP), a biomarker for cardiovascular risk; reductions were 33.5% and 33.3% with the 3.0 mg and 1.8 mg doses, respectively, compared to a 10.5% reduction with placebo (p<0.001 vs. placebo). Treatment with liraglutide 3.0 mg also led to a significantly greater reduction in urinary albumin to creatinine ratio (-18.4%) compared to both the 1.8 mg dose (-10.8%) and placebo (-2.3%). As we saw with the primary SCALE Diabetes results presentation at ADA, both liraglutide doses also significantly lowered systolic blood pressure (-2.8 mmHg and -3.5 mmHg for liraglutide 3.0 mg and liraglutide 1.8 mg, respectively, from a baseline of ~130 mmHg), but there were no significant changes in diastolic blood pressure. These results certainly offer cause for optimism, though they are not entirely unexpected based on what is known about GLP-1 agonists and what was seen at ICE/ENDO from the SCALE Obesity and Prediabetes trial.

  • Treatment with liraglutide 3.0 mg led to meaningful reductions in total cholesterol, HDL cholesterol, VLDL cholesterol, and triglycerides compared to placebo. From baseline, liraglutide 3.0 mg was associated with a 1.5% reduction in total cholesterol compared to placebo’s 3.8% increase. Liraglutide 3.0 mg had a 3.1% reduction in LDL-C levels while placebo had a 5.0% increase. VLDL-C levels were reduced by 8% with liraglutide 3.0 mg and increased by 0.5% with placebo. Similarly, triglycerides were reduced by 9.5% with liraglutide 3.0 mg and increased by 0.4% with placebo.
    • There was a significant difference between the two liraglutide doses only for VLDL-C and triglyceride levels. Changes in total cholesterol and HDL-C levels were relatively similar between liraglutide 3.0 mg and liraglutide 1.8 mg.
  • There was no significant difference in the number of cardiovascular events among the groups. The liraglutide 3.0 mg had five events and both the 1.8 mg and placebo groups had four events each. Notably, both arms of liraglutide experienced an increase in pulse rate of ~2 beats per minute.

Questions and Answers

Q: How did you arrive at the liraglutide 3.0 mg dose rather than 2.4 mg or 1.8 mg? 

A: The trial’s design was going toward 3.0 mg, showing that 3.0 mg was a dose that people could hopefully tolerate and would hopefully have greater weight loss and A1c lowering efficacy with no new safety signals.

Efficacy and Safety of Liraglutide vs. Placebo in Subjects with Type 2 Diabetes and Moderate Renal Impairment (LIRA-RENAL): A Randomized Trial

Guillermo Umpierrez, MD (Emory University, Atlanta, GA)

Dr. Guillermo Umpierrez presented favorable data from the recent LIRA-RENAL study on Novo Nordisk’s Victoza (liraglutide 1.8 mg) for patients with diabetes and moderate renal impairment (glomerular filtration rate 30-59 ml/min). 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; in addition, a significantly greater percentage of the liraglutide 1.8 mg group (53%) achieved an A1c <7% compared to the placebo group (20%). Those in the liraglutide 1.8 mg group also showed greater reductions in systolic blood pressure compared to placebo (-2.45 mm Hg vs. 0.33 mm Hg). There were no significant differences in diastolic blood pressure. 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. Regarding the drug’s safety profile, there were low rates of hypoglycemia and no unexpected safety/tolerability issues; the most common adverse events were GI side effects (nausea, vomiting, diarrhea), as is typical with GLP-1 agonists. This confirmation of safety is particularly valuable given the relative dearth of antihyperglycemic treatment options available to patients with renal impairment.

  • Liraglutide 1.8 mg was associated with higher lipase activity compared to placebo, but no cases of pancreatitis. At 26 weeks, the liraglutide group had a ~1.35 ratio to baseline of lipase activity and the placebo group had a ~1.0 ratio to baseline. Importantly, no acute pancreatitis events were observed. Liraglutide 1.8 mg also had slightly higher amylase activity, with a ratio to baseline of ~1.1 compared to placebo’s ~1.0 ratio.
  • While there were no significant differences in the incidence of adverse events, there were a total of five deaths throughout the trial, with four in the liraglutide group. Dr. Umpierrez stated that the deaths were not related to the study, but not did elaborate on their potential causes. 

Oral Presentations: GLP-1 Analogues – Novel Formulations

IDegLira, a Combination of Insulin Degludec and Liraglutide, Improves Both Pre- and Postprandial Plasma Glucose in Patients with Type 2 Diabetes

Allen King, MD (Diabetes Care Center, Salinas, CA)

Dr. Allen King provided additional data from the DUAL I and DUAL II trials demonstrating a significant improvement in “time in zone” with Novo Nordisk’s Xultophy (IDegLira; insulin degludec/liraglutide) – we are very pleased to see presentation of data that reinforces the importance of data within target zone and look forward to seeing more of this going forward. A post hoc analysis of nine-point glucose profiles from both trials revealed that a greater percentage of patients in the Xultophy group had preprandial, postprandial, and overall blood glucose levels within the target range compared to those in the comparator groups (Tresiba [insulin degludec] and Victoza [liraglutide] in DUAL I and just Tresiba in DUAL II). In addition, the reduction in the range of glucose values over 24 hours was significantly greater with Xultophy in both DUAL I (32.4 mg/dl vs. 19.8 mg/dl with Tresiba and 30.6 mg/dl with Victoza) and DUAL II (37.8 mg/dl with Xultophy vs. 18.0 mg/dl with Tresiba).

  • The DUAL I analysis demonstrated reduced glycemic variability with Xultophy compared to either Tresiba or Victoza. 48% of the Xultophy group had all four preprandial measurements within the target range of 70.2-129.6 mg/dl compared to 41% of the Tresiba group and 32% of the Victoza group; 51% of the Xultophy group had all three postprandial measurements under 162 mg/dl vs. 38% of the Tresiba group and 36% of the Victoza group; and 39% of the Xultophy group had all nine measurements within the target range of 70.2-162 mg/dl vs. 28% of the Tresiba group and 31% of the Tresiba group.
  • Results were similar in the DUAL II trial: 44% of the Xultophy arm had all four preprandial values within range vs. 27% of the Tresiba arm; 37% of the Xultophy arm had all three postprandial values under the target vs. 25% of the Tresiba arm; and 32% of the Xultophy arm had all nine measurements within range vs. 20% of the Tresiba arm.
  • We hope to see analysis of glycemic variability included more often as part of the evaluation process for new diabetes drugs, as we believe “time in range” has an enormous impact on patients’ day-to-day quality of life and long-term health. The flat glucose profile possible with Xultophy, along with its impressive A1c-lowering efficacy and balanced safety profile, makes us incredibly excited to see it finally close to reaching patients (see our report on yesterday’s European approval for more details).

Questions and Answers

Q: What were the final doses in the combination compared to the individual groups?

A: The liraglutide dose was 1.8 mg alone and 1.6 mg in the combination. The dose of insulin was around 69 U in the degludec group and around 36-38 U in the combination.

Q: As in many trials, some people responded and some didn’t. Were there any obvious predictors of who responded?

A: We don’t have that information, but individual variability of people is a keen interest of mine. Other studies have shown intrinsic variability in some people but not others that could have an impact.

DURATION-NEO-1: Greater HbA1C Reductions with Exenatide Suspension Once Weekly by Autoinjector Pen vs Exenatide Twice Daily in Inadequately Controlled Type 2 Diabetes

Carol Wysham, MD (Rockwood Clinic, Spokane, WA)

Dr. Carol Wysham presented results from the DURATION-NEO-1 trial, demonstrating greater glucose-lowering efficacy and improvements in patient satisfaction with a once weekly suspension formulation of AstraZeneca’s Bydureon (exenatide) administered by auto-injector compared to Byetta (exenatide twice daily). Patients who received the once weekly ready-to-use suspension achieved significantly greater A1c reductions (1.4% vs. 1.0% after 28 weeks from a baseline of 8.5%) and significantly greater reductions in fasting plasma glucose (32.4 mg/dl vs. 23.4 mg/dl). Weight loss and hypoglycemia rates were comparable between the two groups. With regard to adverse events, the once weekly group experienced fewer GI side effects (rates of nausea, vomiting, and diarrhea were all ~50% lower from a relatively high baseline) but a higher rate of injection site nodules. Perhaps most notable was the impact on patient-reported treatment satisfaction (measured by the Diabetes Treatment Satisfaction Questionnaire): patients treated with the once weekly formulation showed significantly greater improvements in overall satisfaction (total scores increased by 4.8 from baseline with the Bydureon suspension vs. 2.9 with Byetta) and individual items including treatment convenience and flexibility. While we would have liked to see a head-to-head comparison of the Bydureon suspension/auto-injector with the recently launched Bydureon pen (which simplifies but does not eliminate the reconstitution process), these results still paint a favorable picture of the product, which should be submitted in the US and the EU in 2015.

  • Delivery devices and ease of administration continue to be significant points of improvement and differentiation for GLP-1 agonists. We expect significantly more user-friendly devices like those for this Bydureon suspension as well as (to an even greater extent) Lilly’s just-approved Trulicity (dulaglutide) to lead to greater uptake for the class.

Questions and Answers

Q: Was there any difference in antibody formation with the new preparation? Did you measure antibodies in this study?

A: Antibodies were measured, but I’m not privy to that information.

Q: If you look at the A1c lines, there appears to be a change in slope over time with the once weekly formulation. Is that real?

A: That was similar to a meta-analysis of DURATION-1, where there was a change in slope over time.

Q: Was the rate of nodules the same as other preparations?

A: It was very similar, yes.

Benefits of a Fixed-Ratio Formulation of Once-Daily Insulin Glargine/Lixisenatide (LixiLan) vs Glargine in Type 2 Diabetes Inadequately Controlled on Metformin

Julio Rosenstock, MD (Dallas Diabetes and Endocrine Center, Dallas, TX)

Dr. Julio Rosenstock presented the results from a 24-week phase 2 study on Sanofi’s LixiLan fixed-ratio combination of its market-leading basal insulin Lantus (insulin glargine) and short-acting GLP-1 agonist Lyxumia (lixisenatide). Dr. Rosenstock also presented these results at this year’s ADA. Highlights from the results included an A1c reduction of 1.8% with LixiLan and 1.6% with Lantus from a baseline of 8% - the difference was modest (perhaps due to a stronger-than-expected performance in the Lantus group) but still statistically significant. The LixiLan group also saw significantly less weight gain. Dr. Rosenstock emphasized the importance of examining composite endpoints combining efficacy and safety such as “percentage of patients with final A1c<7% with no weight gain or symptomatic hypoglycemia,” for which LixiLan beat Lantus (46% and 29% of patients, respectively) – again we see effective “time in zone” types of measurements becoming more common (while CGM wasn’t used here, a “seven point profile” was – effectively, using this, “time below zone” and “time above zone” were calculated and analyzed).

Questions and Answers

Q: You have seen many studies using treat-to-target with insulin glargine, and even the best studies in the past achieved a mean A1c of 7% with several-fold higher rates of hypoglycemia. Can you explain whether a different definition of hypoglycemia was used here, or how you can achieve a final A1c of 6.5% with several-fold less hypoglycemia and less weight gain? That seems miraculous.

A: We don’t do miracles; we just report the data. These are people that I presume have early stage diabetes, and who were only on metformin. In the previous studies you refer to, many utilize sulfonylureas on board, which might explain the hypoglycemia. People with early-stage diabetes have better defenses against hypoglycemia. (Editor’s note – he did not say exactly how hypoglcyemia was defined).

Q: Your seven-point profile showed a fall in fasting glucose in both groups, but the postprandial difference was really only at breakfast. When did you do the meal tolerance test?

A: It’s almost impossible to do meal tolerance tests for lunch or dinner. There was a study that was done by Bo Ahrén comparing lixisenatide given at breakfast or the main meal of the day, and you see a postprandial effect on the meal immediately after injection with some carry-over to the subsequent meal. If you give it at breakfast, you may get some effect on lunch.

Q: Could LixiLan be used twice-daily?

A: I don’t think you need that, as in this trial patients already got down to a mean A1c of 6.5%. I think it is critical to give it once a day – it facilitates the widespread usage of this type of compound. Some patients might see a benefit using it twice-daily, but I predict that at least 80% will be able to be controlled with just one shot a day.

Efficacy and Tolerability of ITCA 650 (Continuous Subcutaneous Exenatide) in Poorly Controlled Type 2 Diabetes With Baseline A1C >10%

Robert Henry, MD (University of UCSD, La Jolla, CA)

Dr. Robert Henry presented the full results from the FREEDOM-1 HBL trial on Intarcia’s implantable exenatide mini-pump ITCA-650. These results were also presented as a poster at this year’s ADA. Highlights included a mean A1c reduction of 3.2% at week 26 from a high baseline of 10.7% for the entire cohort. In his conclusion, Dr. Henry emphasized that ITCA-650 delivers 100% adherence to therapy, a way to frame the implantable device that certainly caught the attention of many providers in the room – obviously, this won’t appeal to all patients but for those it does, this is certainly a major advantage (especially from an HCP perspective).

Questions and Answers:

Q: Did you face any problems when removing the mini-pumps?

A: Initially when the devices were first implanted, in some cases they were implanted too deeply, leading to some difficulty in finding and removing them. Subsequently the sponsor perfected an introducer that delivers the mini-pump just underneath the skin so that it is very easy to locate, remove, and replace. 

Posters: GLP-1 Based Therapies: Efficacy II

Harmony 4: 3 Year Efficacy of Albiglutide (albi) vs Insulin Glargine (glar) in Patients with Type 2 Diabetes Mellitus (Poster 837)

P Weissman, M Stewart, D Cirkel, J Ye, P Ambery

Harmony 4 was a three-year, randomized, open-label phase 3 trial that compared GSK’s once-weekly GLP-1 agonist Tanzeum (albiglutide) head-to-head against Sanofi’s long-acting basal insulin Lantus (insulin glargine). 52-week data had previously been published showing noninferiority between albiglutide and glargine with regards to A1c lowering and it was terrific to see detailed results out to three years. This poster presented 156-week results demonstrating that both albiglutide and glargine maintained significant A1c declines from baseline out to three years. In the roughly half of patients who did not require hyperglycemic rescue during the three-year trial, the A1c lowering at three years was statistically similar between the albiglutide and glargine groups (-0.83% and -1.00%, respectively). However, when the patients who required hyperglycemic rescue were counted in the analysis, it appeared that the albiglutide group had a marginally smaller A1c decline (exact values not shown but the change from baseline in the albiglutide group appears to be roughly -0.6-0.7%, while the change from baseline in the glargine group appears to be closer to -0.8%). Not surprisingly, patients in the albiglutide group experienced substantial weight loss (-7.7 lbs) compared to patients in the glargine arm who gained 2.0 lbs. One albiglutide-treated patient was identified as having probable pancreatitis possibly related to study medication, while there were no adjudicated cases of pancreatitis in the glargine group. Overall hypoglycemia, unsurprisingly, was less frequent in the albiglutide group (30%) than the glargine group (43%). No other adverse events were notable or surprising. This data reinforces that for patients in the “real world,” adherence to a once weekly therapy is probably easier than a once daily therapy; presumably, adherence to a once weekly therapy that prompts significantly less hypoglycemia is likely easier than a once daily therapy associated with more hypoglycemia.

  • Harmony 4 was a three-year, randomized, open-label phase 3 trial that compared GSK’s once-weekly GLP-1 agonist Tanzeum (albiglutide) head-to-head against Sanofi’s long-acting basal insulin Lantus (insulin glargine). 52-week data had previously been published showing noninferiority between albiglutide and glargine with regards to A1c lowering. This poster presented the 156-week results.
  • The trial randomized 779 patients in a roughly 2:1 ratio to albiglutide and insulin glargine, respectively. The completion rates were similar: 60% in the albiglutide group and 62% in the glargine group. At baseline, patients had a mean A1c of 8.3-8.4%, BMI of 33 kg/m2, 8.9 and 8.4 years diabetes duration for the albiglutide and glargine groups, respectively. Most (82% in both groups) were on background metformin + sulfonylurea therapy, with the remaining 18% on background metformin therapy only.
  • The study allowed individual titration of insulin glargine dose to reach a target goal of 7% A1c. The median insulin glargine dose at baseline was 10 units/day (range 2-40), and at week 156 the median was 34 units/day (range 0-216).
  • They study also allowed titration of albiglutide dose. Patients in the albiglutide group started at the lower 30 mg dose and titrated up to the 70 mg dose if necessary (presumably to reach the target A1c of 7%, although this was not explicitly stated in the poster). Mean time to uptitration was 31.7 weeks for the 77% of albiglutide-treated patients that uptitrated.
  • Throughout the 156 weeks, a slightly higher, but non-significant, percentage of patients in the albiglutide group required hyperglycemic rescue (56% vs. 48% in the glargine group; p=0.1515). Hyperglycemic rescue was administered if patients in either group exceeded a predefined (unspecified) A1c threshold. A1c change results (next bullet) were stratified by rescue status.
  • Both albiglutide and glargine produced a significant A1c decline that was maintained for three years. A similar percentage of patients in each group also achieved the A1c goal of 7% (48% for albiglutide and 52% for glargine).
    • Of the patients who completed the three-year trial and did not require hyperglycemic rescue (albiglutide n=123; glargine n=88), albiglutide and glargine produced statistically similar A1c changes (-0.83% for albiglutide and -1.00% for glargine).
    • In the entire three-year cohort (n=273 for albiglutide and n=162 for glargine), both treatments still provided a significant A1c drop from baseline, although it appears that the albiglutide group may have achieved less of an A1c decline than the glargine group (exact values and significance not specified on the poster, although the change from baseline in the albiglutide group appears to be roughly -0.6-0.7%, while the change from baseline in the glargine group appears to be closer to -0.8%).
  • Not surprisingly, the glargine group ended the trial with better fasting plasma glucose (FPG) values, suggesting that postprandial glucose lowering paid a larger role in the A1c-lowering for albiglutide.
  • Also not surprisingly, albiglutide produced weight loss (-3.5 kg [-7.7 lb]) while glargine produced weight gain (+0.9 kg [2.0 lb]), with a treatment difference of 4.4 kg [9.7 lb] in favor of albiglutide. The weight loss, impressively, seemed to be maintained pretty steadily over the course of the entire three years without evidence of rebound. In patients who did not require hyperglycemic rescue, weight seemed to still be declining after three years.
  • Overall adverse events (AEs) were similar between the groups, although AEs leading to withdrawal were higher for albiglutide (9.5%) compared to glargine (4.1%). Not surprisingly, nausea and diarrhea were more pronounced in the albiglutide group, although vomiting was similar between groups.
    • One albiglutide-treated patient was identified as having probable pancreatitis possibly related to study medication, while another albiglutide-treated patient was identified as having possible pancreatitis unlikely to be related to study medication. No likely cases of pancreatitis were identified in the glargine group.
  • The rate of any hypoglycemia was lower for albiglutide (30%) than glargine (43%), which was also unsurprising.

A Network Meta-Analysis to Compare Once Weekly Dulaglutide Versus Other GLP-1 Receptor Agonists in Patients with Type 2 Diabetes (Poster 839)

A Padhiar, J Thompson, J Eaton, N Hawkins, K Norrbacka, M Reaney, R Shaginian, K Boye, N Varol

In a poster presentation, Dr. Padhiar and colleagues presented results from a network meta-analysis comparing the efficacy of Lilly’s once-weekly GLP-1 agonist dulaglutide (Trulicity) with other GLP-1 agonists. In the study, systematic literature review identified 44 randomized controlled trials of at least 16 weeks duration evaluating the efficacy of a GLP-1 agonist therapy. Based on these studies, three networks (16-36 week trials as add-on to metformin, 16-36 week trials as add-on to metformin ± sulfonylurea [SFU] ± thiazolidinedione [TZD], and 37-56 week trials as add-on to metformin ± SFU ± TZD) were created and compared to the results of the AWARD trials program for 1.5 mg dulaglutide. Results from the 16-36 week networks indicated improved decline in A1c from baseline with dulaglutide versus twice-daily exenatide (-0.76% as add-on to metformin, -0.68% as add-on to metformin ± SFU ± TZD) and lixisenatide (-0.62% as add-on to metformin, -0.66% as add-on to metformin ± SFU ± TZD). Results from the 37-56 week network indicated improved decline in A1c with dulaglutide versus once-weekly albiglutide (-0.56%) and twice-daily exenatide (-0.57%). Results from the remaining comparisons were non-significant. As a reminder, dulaglutide was recently approved in the US with Lilly planning to launch the product later this year.

  • This was a network meta-analysis designed to compare the efficacy of once-weekly dulaglutide with other GLP-1 agonists. In contrast to traditional pair-wise meta-analyses, network meta-analyses create networks using common comparator arms such that estimation of relative efficacy for various treatments can be performed when no head-to-head trials exist. In this study, systematic literature review identified 44 randomized controlled trials of at least 16 weeks duration from which three networks (16-36 week trials as add-on to metformin, 16-36 week trials as add-on to metformin ± sulfonylurea [SFU] ± thiazolidinedione [TZD], and 37-56 week trials as add-on to metformin ± SFU ± TZD) were created and compared to the results of the AWARD trials program for 1.5 mg dulaglutide.
  • Results from the network meta-analysis indicated significant improvement in A1c decline with once-weekly dulaglutide in seven comparisons. All other comparisons were non-significant:

Dulaglutide 1.5 mg QW versus:

Add-on to metformin 16-36 weeks

Add-on to metformin ± SFU ± TZD 16-36 weeks

Add-on to metformin ± SFU ± TZD 37-56 weeks

A1c decline from baseline (%)

Albiglutide 30 mg QW




Albiglutide 50 mg QW




Exenatide 10 mcg BID




Exenatide 5 mcg BID




Exenatide 2 mg QW




Liraglutide 1.2 mg QD




Liraglutide 1.8 mg QD




Lixisenatide 20 mcg QD




  • Study authors noted that methods to adjust for heterogeneity such as stratification and meta-regression may not always overcome differences between patient populations and study designs, with sensitivity analyses designed to quantify the heterogeneity ongoing. While results based on indirect evidence may be subject to greater uncertainty, the study suggests comparable efficacy of dulaglutide with others in the GLP-1 agonist class.

EASD/ADA Symposium: Incretin-Based Therapies

Summary of Recent Studies

Clifford Bailey, PhD (Aston University, Birmingham, UK)

Before a crowded hall of attendees, Dr. Clifford Bailey opened the symposium with a comprehensive review of recent studies on incretin-based therapies. In exploring the glycemic efficacy and weight effects of incretin therapies, Dr. Bailey reviewed several meta-analyses and head-to-head comparisons. He demonstrated the different GLP-1 agonists’ effects on A1c levels and body weight and showed how DPP-4 inhibitors are more weight-neutral compared to GLP-1 agonists’ weight loss. He also highlighted what we don’t yet fully understand, mentioning the several hypotheses for response variability as well as our lack of knowledge on whether GLP-1 agonists can preserve human islet beta-cells. Looking to the future, Dr. Bailey pointed out the potential in liraglutide 3.0 mg for obesity (see our coverage of the FDA Advisory Committee meeting for Saxenda), chimeric incretin peptides, and once-weekly DPP-4 inhibitor omarigliptin (first phase 3 data presented at EASD). Concluding, Dr. Bailey ventured outside of diabetes, mentioning incretin therapies’ pleiotropic effects in treatment for Parkinson’s disease as well as post-myocardial infarction.

  • Dr. Bailey presented a meta-analysis (Aroda et al., Clinical Therapy 2012) that showed incretin therapies’ significant reductions in A1c levels, but demonstrated that DPP-4 inhibitors tended to have slightly lower efficacy. Exenatide BID and QW along with liraglutide had A1c reductions between 1.1% and 1.6%, while alogliptin, linagliptin, saxagliptin, and sitagliptin had reductions around 0.7%. Results were similar with weight loss, as the GLP-1 agonists led to weight reductions of ~2.2 kg and the DPP-4 inhibitors led to weight reductions of ~0.4 kg.
  • Dr. Bailey presented data on the efficacy and durability benefits of exenatide. One head-to-head comparison (Bergenstal et al., Lancet 2010) showed that exenatide led to greater A1c reductions (-1.7% from baseline) compared to pioglitazone (-1.4%) and sitagliptin (-1.0%). In addition, exenatide had greater weight loss with a 2.8 kg reduction, compared to sitagliptin’s 0.9 kg reduction and pioglitazone’s gain of 3.4 kg. Another study (Gallwitz et al., Lancet 2012) demonstrated that exenatide added onto metformin had more durable glycemic benefits compared to glimepiride added onto metformin. Exenatide’s effects on A1c, weight, and fasting plasma glucose were also shown to be relatively constant over six years of treatment.
  • Regarding the role of incretin therapies in treatment, Dr. Bailey highlighted the effective combination of GLP-1 agonists and basal insulin. He explained how the two drug classes have complementary effects in type 2 diabetes and pointed to IDegLira data as an example. In addition, Dr. Bailey noted incretin therapies’ role in in type 1 diabetes, as liraglutide has been shown to reduce insulin dose and glycemic variability in this population.

Questions and Answers

Q: So are these DPP-4 inhibitors the same or different?

A: You can answer that both ways. They are different structurally, but not by much. They’re similar in action as in they sit on the same place on the enzyme. They have different pharmacodynamics properties. But overall, their pharmacological properties can be adjusted by dose to have much of the same effect.

Do We Need GLP-Based Therapies?

David Nathan (Harvard Medical School, Boston, MA)

Dr. David Nathan delivered a frank and at times skeptical presentation on the “value added and value subtracted” with incretin-based therapies. Dr. Nathan did not dispute the fact that GLP-1 agonists and DPP-4 inhibitors are “here to stay,” but he urged providers to take a clear-eyed view of their limitations and avoid falling victim to excessive hype. While he listed several important benefits of incretin drug classes, including strong A1c-lowering efficacy (with GLP-1 agonists), minimal hypoglycemia risk, and weight neutrality/loss, he argued that incretin-based therapies also have significant disadvantages. For example, Dr. Nathan highlighted the classes’ cost, the fact that GI side effects are often under-acknowledged (we would argue that lots of this depends on HCP education), and that the products’ A1c-lowering efficacy is relatively modest (presumably he was referencing DPP-4 inhibitors). He also noted that there is very little data conclusively demonstrating that incretin therapies improve long-term outcomes or that they are more effective than other anti-hyperglycemic agents. While we certainly agree that longer-term comparative effectiveness data would be valuable, we also believe that incretin therapies, especially GLP-1 agonists, have some of the most appealing efficacy profiles out of the currently available type 2 diabetes drug classes, with a positive impact on patient adherence due to the low risk of weight gain and hypoglycemia. And, regarding cost – there is, of course, no way that therapies like these could ever be generic without them generating a profit first; we don’t think the “high cost” should dissuade development. Overall, while Dr. Nathan is, of course, a pre-eminent researcher and clinician, we found these remarks reductive overall given the extent to which patients have benefited from incretins.  

The Side Effects of GLP-Based Therapies: Should There Be Concern?

Baptist Gallwitz, MD (Eberhard-Karlz University, Tubingen, Germany)

Dr. Baptist Gallwitz provided a review of safety concerns and side effects associated with incretin-based therapies, characterizing his goal as turning the “pitch black” of uncertainty into nuanced shades of gray. He does not see much real reason for concern about pancreatitis and thyroid safety issues, although he considers cardiovascular benefit or risk to be more of an open question. In his conclusion, he categorically stated that there is no reason to reconsider use of GLP-1 agonists in the overall population based on the data currently available. In his view, the benefits (low intrinsic hypoglycemia risk, straightforward dosing, good patient acceptance) outweigh the potential risks by far. He noted that the class’ potential risks must be viewed in the context of the risks associated with other add-on therapy alternatives. Until we receive data from the full set of GLP-1 agonist CVOTs, Dr. Gallwitz suggested that the scientific community can pool results from observational studies, conduct mechanistic studies, and investigate responder and non-responder patient subgroups.

  • On pancreatitis, Dr. Gallwitz focused on the limitations of observational and epidemiological studies: they can suffer from the misclassification of cases and controls, underreporting of important potential confounders (i.e.: alcohol, tobacco, obesity), and cannot differentiate between the effects of different therapies. In randomized control trials of incretin-based therapies such as SAVOR (for AZ’s DPP-4 inhibitor Onglyza [saxagliptin]), most patients that had pancreatitis recovered, and many stayed on the study drug. Even if there is an effect, Dr. Gallwitz emphasized that the magnitude may be so small that it will not be a major factor. Citing data from a pooled analysis of phase 3 trials (Meier & Nauck, Diabetologia 2014), which found a non-significant 39% increase in pancreatitis with GLP-1 agonists, Dr. Gallwitz noted that 2000 patient-years of treatment would be needed to cause one excess case of pancreatitis even if the 39% increase was in fact a real signal.
  • Dr. Gallwitz dispatched the thyroid safety issue relatively quickly, noting that the concern stemmed largely from preclinical models. He believes that this issue is most likely not clinically relevant.
  • On cardiovascular safety, Dr. Gallwitz acknowledged the presence of some key unanswered questions. In his view, the heart failure signal seen in SAVOR was completely explained. The implications of the increase in heart rate seen with GLP-1 agonists are also unclear. Looking ahead, Dr. Gallwitz expressed optimism that currently ongoing CVOTs will provide robust data on these and other questions. He highlighted the LEADER trial (for Novo Nordisk’s Victoza [liraglutide]) for its large size, as well as the CAROLINA trial (for BI/Lilly’s Tradjenta [linagliptin]) for using an active comparator (glimepiride). Despite the current uncertainty, Dr. Gallwitz does believe that there is a chance for the GLP-1 agonist class to demonstrate cardioprotection.

Symposium: Safety of Novel Diabetes Drugs from the SAFEGUARD Project (Sponsored by Charles University in Prague)

Introduction and Overview

Miriam Sturkenboom, PhD (Erasmus University Medical Center, Rotterdam, Netherlands)

Before a modest audience on a Sunday pre-conference session, Dr. Miriam Sturkenboom provided a brief overview of the SAFEGUARD project, a consortium funded by the European Commission and involving 14 institutions from seven countries that aims to better understand safety concerns associated with various drug classes for type 2 diabetes, including incretins. The SAFEGUARD group is conducting original observational and mechanistic studies as well as gathering evidence from spontaneous adverse-event reports and meta-analyses of existing data. Pertinent outcomes include cardiovascular events, cerebrovascular safety, pancreatitis, pancreatic and bladder cancer, and overall mortality. The group will report its eventual findings directly to the EMA, which could in turn precipitate a change in regulatory practices (depending on the direction of the findings). As a result, it is an initiative well worth watching. Over the afternoon’s presentations, discussion of some potential signals emerged; overall, the preliminary and observational nature of the data was emphasized, which seems in order.


Kevin Blake, MD (EMA, London, UK)

Dr. Kevin Blake provided background on the EMA’s perspective on evaluating safety with diabetes drugs. He noted that the agency is shifting from the traditional paradigm, which is dependent on trials conducted by manufacturers, to a new model that also involves the collection of data from other sources such as EMA-sponsored studies and pharmacovigilance networks. Another ongoing EMA shift, based on the history of diabetes drug safety controversies, is to more proactively and holistically assess cardiovascular, cerebrovascular, and pancreatic safety for entire diabetes drug classes – this process is what gave birth to the SAFEGUARD project. In addition to pulling together a wider database on the safety of diabetes drug, SAFEGUARD also aims to investigate potential mechanisms underlying the cardio/cerebrovascular and pancreatitis-related safety questions.

Mechanistic Basis of Potential Safety Issues Around Incretin Based Therapies

Martin Haluzík, CSc (Charles University in Prague, Prague, Czech Republic)

Dr. Martin Haluzík offered a balanced assessment of the safety concerns associated with incretin therapies, ultimately concluding that more long-term data is needed to understand any potential class effects. Dr. Haluzík does not believe that the GI side effects commonly seen with GLP-1 agonists represent any serious safety concern, as they typically decrease over time and are not a major obstacle to tolerability for most patients. We would also point out that they appear to vary depending on the comfort level of the provider and their skill at titration. On the “very hot topic” of cardiovascular safety, he characterized the DPP-4 inhibitor profile as neutral to slightly positive and the GLP-1 agonist profile as mostly positive, due to beneficial effects on cardiovascular risk factors like body weight, blood pressure, and lipids. Dr. Haluzík acknowledged concerns about the slight increase in heart rate with GLP-1 agonists and the possible heart failure signal seen in some studies of DPP-4 inhibitors, but said that there is not currently enough data to draw any meaningful conclusions. Similarly, he believes that there is not sufficient evidence at this point to infer a link between incretins and a higher risk of pancreatitis, pancreatic cancer, or medullary thyroid cancer (MTC), as all data thus far has come from animal studies or trials whose methodologies have been questioned.

Meta-Analysis of Published Clinical Trials on Incretin-Based Therapies

Giorgia de Berardis (Consorzio Mario Negri Sud, S. Maria Imbaro, Italy)

Dr. Giorgia de Berardis presented the result of a meta-analysis of 213 randomized clinical trials on incretin-based therapies (56 on GLP-1 agonists and 157 on DPP-4 inhibitors), which sought to investigate the risk of cardiovascular and pancreatic adverse outcomes. The pooled results for DPP-4 inhibitors were disproportionately influenced by the results of the two cardiovascular outcomes trials (CVOTs) that have reported to date, SAVOR for AZ’s Onglyza (saxagliptin) and, to a lesser extent, EXAMINE for Takeda’s Nesina (alogliptin). As expected, pooled results from these and other DPP-4 inhibitor placebo-controlled trials were neutral for most cardiovascular outcomes with the exception of heart failure (hazard ratio: 1.20, 95% CI: 1.02-1.40). In active-controlled trials, DPP-4 inhibitors appeared to show a reduction in ischemic stroke (HR: 0.33, 95% CI: 0.14-0.79) and a CV adverse event composite (HR: 0.60, 95% CI: 0.42-0.88), but Dr. de Berardis pointed out that these results were driven by comparisons against glimepiride and that the difference seen could be due to cardiovascular harm with the SFU rather than benefit with DPP-4 inhibitors. (On this basis, should SFUs be used less frequently?) For GLP-1 agonists, the fact that no CVOTs have been completed meant that there were too few events to show a meaningful effect on CV adverse events. There were too few events to rule in either direction for pancreatitis and pancreatic cancer with either drug class. Dr. de Berardis concluded on an optimistic note, suggesting that there is currently no evidence of a connection between incretin therapies and either pancreatitis or overall CV adverse events (with the exception of heart failure with DPP-4 inhibitors).

Systematic Literature Review of Observational Studies on Pancreatic Outcomes

Lorna Hazell (Drug Safety Research Unit, Southampton, UK)

Complementing the meta-analysis of randomized control trials presented during the previous talk, Ms. Lorna Hazell presented a systematic review of case-control and cohort studies conducted specifically on incretin therapies and pancreatitis/pancreatic cancer. Between the initial literature review in 2011 and the updated review conducted in July 2014, the number of observational studies on incretins and pancreatic adverse events tripled to 38, indicating the recent level of interest in pancreatic adverse events with incretin therapies. The results Ms. Hazell presented did not provide any new cause for alarm, with a fairly consistent trend towards the null but based on very few events. Only three of the 38 studies reported provided useful data on pancreatic cancer, and the number of events was far too small to draw any meaningful conclusions. Ms. Hazell did not make any concrete predictions on the risk of a risk or benefit based on the results of the literature review. Rather, she focused on the fact that observational studies are limited by a number of biases, short exposure time, and incomplete reporting of data.

  • Although Ms. Hazell’s presentation focused on incretins, she also dedicated a slide to show pooled observational study data on pancreatic adverse events with biguanides and sulfonylureas. Biguanides (presumably mostly metformin) appeared to lean towards a protective effect whereas SFUs appeared to lead to an increase in pancreatic cancer, although Ms. Hazell suggested that these findings may be due to an exposure bias (she did not elaborate on that suggestion).

Safety Signals from Spontaneous Reports on Incretin Based Therapies

Lorna Hazell (Drug Safety Research Unit, Southampton, UK)

Ms. Lorna Hazell summarized results from an analysis of spontaneous adverse event reports for incretin therapies, though she cautioned that this type of reporting is only one component of drug safety evaluation and carries many limitations.  The SAFEGUARD analysis aimed to identify cases where a particular adverse event was disproportionately reported for certain drugs in FDA and EMA databases and to examine the circumstances in which the disproportionate reporting was most detectable. Each drug-outcome combination was analyzed four times, using two different comparator groups (all drugs and all diabetes drugs), and two different definitions of the outcome (narrow and broad). A proportional reporting ratio (PRR) was calculated for each combination and the level of evidence for an association was characterized as high, intermediate, low, or inconclusive. The most notable finding from the preliminary results was a high level of evidence for a link between incretin therapies and acute pancreatitis and between some incretin therapies and pancreatic cancer. While these results appear concerning at first glance, Ms. Hazell noted that at least part of the effect could be due to stimulated reporting, as there were more reports of acute pancreatitis events in the years after the FDA issued warnings about a potential association with incretin drugs. The investigators are currently developing an algorithm to calculate an “uncertainty score” for each association, which should allow for better interpretation of the results.

User Patterns of Incretin-Based Therapies

Gwen Masclee, MD (Maastricht University Medical Center, Maastricht, Netherlands)

Dr. Gwen Masclee provided big-picture perspective to color the epidemiological findings presented by her colleagues. The most notable takeaway, in our view, was that despite the massive size of the data pool that SAFEGUARD can draw from, the project will only be able to detect a 100% increase in risk for acute pancreatitis with GLP-1 agonists and a 50% increase in risk with DPP-4 inhibitors. This limitation is a product of how rare pancreatitis is, in terms of absolute incidence. We imagine it is unlikely that the real signal is anywhere near those values, which raises questions about how to interpret an eventual finding that might lean one way or the other but lacks the power to definitively prove a risk.

Comparative Assessment of the Cardiovascular and Pancreatic Safety of Incretin Based Therapies from SAFEGUARD Epidemiological Studies

Silvana Romio, PhD (Erasmus University Medical Center, Rotterdam, Netherlands)

Dr. Silvana Romio presented preliminary results from SAFEGUARD’s ongoing epidemiological study intended to identify associations between type 2 diabetes drugs and cardiovascular, cerebrovascular, and pancreatic adverse events. Investigators are collecting data from nine databases on a total of over 55 million patients, 1.6 million of whom have type 2 diabetes, and will report relative risk estimates of associations between specific drugs and adverse outcomes. Notably, the two analyses performed thus far, on data from the PHARMO and BIFAP databases, both found a significant association between incretin therapies and acute pancreatitis (odds ratios of 2.41 and 2.65) in a cohort of subjects with newly diagnosed type 2 diabetes; the reference group consisted of a matched cohort treated with metformin. Of course, these are only preliminary results, but if confirmed in the final analysis (which is expected within the next year), they would represent some of the most solid evidence to date of a link between incretin therapies and pancreatitis.

  • It is important to note that diabetes itself is a risk factor for pancreatitis (a general analysis of all the SAFEGUARD databases confirmed a higher rate among patients with type 2 diabetes compared to the general population) and that while acute pancreatitis is certainly a serious condition, it is chronic pancreatitis that is most likely to lead to pancreatic cancer. Therefore, even if solid evidence emerges for a higher risk of pancreatitis with incretin therapies, there may still be countless patients for whom the benefits of these drug classes outweigh the risks.

Corporate Symposium: Perspectives on GLP-1RA Therapy – Advancements in T2DM (Sponsored by Lilly)

A Review of the GLP-1RA Therapy Landscape

Anthony Barnett, MD (University of Birmingham, Birmingham, UK)

Professor Anthony Barnett characterized the GLP-1 agonist class as an appealing therapeutic option for type 2 diabetes and highlighted Lilly’s once-weekly Trulicity (dulaglutide) as an “important addition to the armamentarium” of treatment options. We agree with this, having seen the new therapy recently. In particular, Professor Barnett stressed that Trulicity would be the first once weekly GLP-1 agonist on the market that does not require reconstitution and the only one that has demonstrated equal A1c-lowering efficacy to Novo Nordisk’s Victoza (once daily liraglutide). He urged providers to “think about the value of the drug, not just the cost of the drug” when making prescribing decisions, as he believes benefits like weight loss and a reduced risk of hypoglycemia are crucial for patients but often not taken seriously enough by the medical community – certainly, hypoglycemia prevention has value as hypoglycemia is very costly. We were disappointed to hear (though perhaps not wholly surprised) that weight loss isn’t taken that seriously by all HCPs – that is very unfortunate. During Q&A, Professor Barnett described the limitations that high costs have placed on GLP-1 agonists’ uptake, particularly in the UK, where providers are encouraged to use the class only as a third line therapy or in combination with insulin. He also acknowledged the significant benefits of SGLT-2 inhibitors and expressed interest in the future potential of SGLT-2 inhibitor/GLP-1 agonist combination therapies, though he stressed that for many patients, GLP-1 agonists alone can be a very appropriate second-line treatment option. We’ll be eager to see how the “second line” option shakes out when GLP-1/basal insulin combos are available.

  • As a reminder, Trulicity was recently approved in the US. It remains under review in the EU.

Questions and Answers

Q: What are the benefits of SGLT-2 inhibitors vs. GLP-1 agonists?

A: SGLT-2 inhibitors are a bit cheaper in Europe, so the thought is, “Why give a more expensive injectable rather than a cheaper oral medication?” There’s a significant place for SGLT-2 inhibitors. If you want to avoid injections and lose weight, then they’re a great choice. They’re [SGLT-2s] not as efficacious generally, and there are issues with genital infections, use in elderly patients, and other safety concerns. There’s room for both classes; they’re not going to displace GLP-1 agonists.

Q: If dulaglutide comes to the market, will there be restrictions for renal impairment?

A: I don’t know. Regulatory agencies tend to be cautious at the beginning, so there may be restrictions. Over the long term, it looks safe for patients with renal impairment.

Q: What are your thoughts on the difference in weight loss between liraglutide and dulaglutide?

A: It is probably due to differences in the duration of action, or PK/PD differences, but we don’t have a good explanation. It’s a very small difference; it’s statistically significant but clinically very small.

Q: Which GLP-1 agonist is best for elderly patients?

A: That’s a difficult question. It depends on the license. Many elderly patients have renal impairment. A big advantage of the class for the elderly is the low hypoglycemia risk. The elderly are the most adversely affected; that group is where most of the deaths and hospitalizations occur, especially in type 2 diabetes. You have to look at the labeling. There’s no particular reason that this class shouldn’t be useful for elderly patients.

Q: You mentioned many differences between different GLP-1 agonists, but what is the most important difference?

A: There are a lot of differences. Once weekly dosing vs. once or twice daily is a big differentiator. Reconstitution is a big problem, but we’re going to see this issue overcome with products like dulaglutide. It won’t be the only one; we will see other products, but people have difficulties using the two agents on the market now.

Q: Could GLP-1 agonists be useful for the treatment of prediabetes?

A: Possibly. We would need to see data.

Q: Where do you position GLP-1 agonists in the treatment algorithm?

A: Because of absolute costs, in the UK we’re discouraged from using them as a second-line treatment. They’re usually used as an alternative to insulin or with insulin. For some patients, it would be entirely appropriate to use them second line. If a patient is failing on metformin, is overweight, and has a comorbidity from obesity like sleep apnea, I think a GLP-1 agonist is the treatment of choice because you will get the benefits of weight loss and glycemic control. The key is individualization. We certainly use GLP-1 agonists as a second-line treatment, as a third-line treatment, and with insulin.

Q: Does dulaglutide have as much of an effect on postprandial glucose as daily exenatide?

A: The effect is the same.

Q: Is there any potential for a GLP-1 agonist/SGLT-2 inhibitor combination therapy?

A: I’m not aware of any in development, but those would be very sensible studies.

Comment: An article was recently published that showed an increase in glucagon and hepatic glucose production with SGLT-2 inhibitors, so that combination might be better compared to a combination with DPP-4 inhibitors.

A: The big question is whether the weight loss would be additive or even synergistic with the different mechanisms.

Q: Will using a GLP-1 agonist delay insulin initiation?

A: It may; we’ve seen GLP-1 agonists used as an alternative to basal insulin. You can see the advantages if you look at the package of efficacy: not just A1c but weight loss and hypoglycemia. We’re using GLP-1 agonists before insulin, then we bring in insulin later if needed. I don’t think it’ll be the demise of insulin.

Q: In some countries, GLP-1 agonists are not as widely used because they’re not reimbursed for all patients. What’s the situation in the UK?

A: It’s mostly used as part of triple therapy and with insulin. One third of the use is with insulin.

Q: You mentioned that weight loss can be so different among patients taking GLP-1 agonists, with some people losing enormous amounts and others not at all. Are there any predictors of who will do well?

A: I don’t think there are any good predictors in an individual patient. At the population level, there have been studies, but with an individual patient it’s very hard to predict. People with a higher BMI tend to lose more weight, but I can’t say in an individual case who will lose more. In a clinical situation, I have patients who have lost 20-30 kg (44 – 66 pounds); one guy looks no different – he’s so overweight that a 20 kg weight loss isn’t noticeable – but he has lost that much weight. It’s difficult to say in an individual case who will lose the most. There are some practical issues in the UK because the NICE guidelines say that patients need to have an A1c reduction of 1% and lose 3% of their body weight for a GLP-1 agonist to be considered effective. That’s an issue because some people lose massive amounts of weight and don’t see an A1c drop, while others have an A1c drop of 1.5% and don’t lose weight. Theoretically in either of those cases, you should stop the drug, but that would be ridiculous. The guidelines do say it’s an individual decision. (Editor’s note – 1% drop of course is more meaningful or less meaningful depending on the baseline.)

Q: Are GLP-1 agonists as effective in non-obese patients?

A: The weight loss not as great – people won’t waste away to nothing by taking a GLP-1 agonist. The percentage weight loss is essentially the same, but more obese people lose more weight.

Future Perspectives for GLP-1 Receptor Agonists

Juris Meier, MD (St. Josef Hospital, Bochum, Germany)

Dr. Juris Meier ended the Lilly GLP-1 agonist symposium with an optimistic presentation on the class’ future potential, although he also mentioned some more practical recommendations for the present such as scenarios for using short acting vs. long acting GLP-1 agonists (see the table below). Dr. Meier began with Intarcia’s matchstick-sized implantable osmotic mini-pump ITCA-650, characterizing its efficacy as comparable with other long-acting GLP-1 agonists and highlighting the unique lack of injections. He displayed data on the strong efficacy seen with fixed-ratio GLP-1 agonist/basal insulin combinations (namely Novo Nordisk’s Xultophy [liraglutide/insulin degludec] and Sanofi’s LixiLan [lixisenatide/insulin glargine]), and also noted that the slow up-titration protocol with these combinations appears to attenuate the nausea experienced by patients on GLP-1 agonist monotherapy. He spoke quite positively on the use of GLP-1 agonists in type 1 diabetes, citing the improvements in weight, daily insulin dose, postprandial glucose, and insulin sensitivity seen in early studies (Sarkar et al., Diabetes Care 2014). He also mentioned GLP-1 agonists’ use for obesity, posting data on Novo Nordisk’s liraglutide 3.0 mg for obesity (which recently received a positive FDA AdComm vote), although he did not discuss this application in great detail.

  • Dr. Meier characterized neuroprotection as an exciting new frontier for the GLP-1 agonist class. This suggestion is supported by a modest body of preclinical work (McClean et al., J Neurosci 2011) as well as very early clinical studies (Aviles-Olmos et al., JCI 2013). Neurodegenerative diseases such as Alzheimer’s have been theorized to have mechanistic connections with insulin resistance, although the body of knowledge in this area is quite limited.
  • Intravenous GLP-1 agonist administration appears to preserve the class’ dose-responsive efficacy while largely abolishing nausea and vomiting seen with subcutaneous administration. Dr. Meier noted that intravenous administration could allow the class to achieve its full potential, although with current technology there is not a very patient-friendly way to achieve chronic IV drug administration.

Table: Recommended GLP-1 agonist subtype for different therapeutic applications



Patient with high fasting glucose levels

Long-acting GLP-1 agonist

Patient with high postprandial glucose levels

Short-acting GLP-1 agonist

Patient with susceptibility to GI disorders

Long-acting GLP-1 agonist

Patients with a preference for simplicity

Long-acting GLP-1 agonist

Combo with basal insulin

Short-acting GLP-1 agonist

Initiating GLP-1 Receptor Agonist Therapy As First Injection

Francesco Giorgino, MD, PhD (University of Bali Aldo Moro, Bari, Italy)

Dr. Francesco Giorgino began a two-part mini-debate by suggesting that GLP-1 agonists could serve as patients’ first injectable therapy – Lilly’s Trulicity (dulaglutide) is being positioned as a possible initiator injectable. Dr. Giorgino surveyed a range of data comparing GLP-1 agonists to DPP-4 inhibitors, TZDs, and SFUs, demonstrating that GLP-1 agonists are generally able to provide at least comparable glucose-lowering efficacy along with benefits on weight and hypoglycemia. He characterized once-weekly administration as a definite convenience advantage for many patients, which might influence providers’ choice of GLP-1 agonists when all else is roughly equal (as he suggested is the case with Trulicity and Novo Nordisk’s Victoza [liraglutide]). Dr. Giorgino made a compelling case that GLP-1 agonists (especially longer-acting products in the class) can be a clinically appropriate and patient-friendly way to intensify treatment beyond oral agents.

  • Dr. Giorgino provided suggestions on which injectable therapy (short-acting GLP-1 agonists, long-acting GLP-1 agonists, or basal insulin) was best suited to which clinical phenotype:
    • Basal insulin: Patients with elevated fasting glucose levels, a strong insulin secretory defect, and/or rapid progression of beta-cell failure (i.e.: with LADA).
    • Short-acting GLP-1 agonist: Patients with elevated postprandial glucose, obesity/overweight, and/or metabolic syndrome.
    • Long-acting GLP-1 agonist: Patients with elevated fasting plasma glucose and postprandial glucose, obesity/overweight, and/or metabolic syndrome.
  • In studies of GLP-1 agonists vs. oral antihyperglycemic agents, GLP-1 agonists generally come away with benefits on efficacy, weight, and/or hypoglycemia. This has proven the case in DURATION-2 (AZ’s Bydureon [exenatide] vs. Merck’s Januvia [sitagliptin] and Takeda’s Actos [pioglitazone]), Harmony 3 (GSK’s Tanzeum/Eperzan [albiglutide] vs. Januvia and glimepiride), and AWARD-5 (Trulicity vs. Januvia).
  • A smaller number of more recent studies suggest that GLP-1 agonists have concrete clinical advantages vs. basal insulin. Dr. Giorgino mentioned that in AWARD-2, Trulicity demonstrated superior A1c lowering vs. Sanofi’s Lantus (insulin glargine), along with less hypoglycemia. Harmony 4, the only other study comparing a once weekly GLP-1 agonist to insulin glargine, was presented as a poster later at EASD – in that study, at three years, albiglutide achieved non-inferior (but not superior) A1c lowering vs. insulin glargine, but with no weight gain and significantly less hypoglycemia as well.
    • The results from AWARD-2 (A1c superiority vs. insulin glargine with less hypo and weight gain) were quite impressive, but some have suggested that insulin was not titrated aggressively enough in the insulin glargine group. Dr. Giorgino (an AWARD-2 investigator) addressed that point in Q&A, noting that achieving better A1cs would have been possible in the insulin glargine arm, but at the expense of too much hypoglycemia (especially given that all patients were on background sulfonylurea therapy). He also noted that the 30 U daily insulin dose seen in the trial is comparable to the final mean daily dose achieved in other clinical trials.
  • Dr. Giorgino mentioned the results of AWARD-6, a phase 3 trial comparing Trulicity (once-weekly) and Victoza (once-daily). Based on the drugs’ clinical profile, Dr. Giorgino suggested that providers and patients can consider using either the daily or weekly agent, but noted that the advantage of once-weekly administration may be a key decision factor for many patients.

Questions and Answers

Q: In AWARD-2, how was insulin glargine titrated?

A: Insulin glargine was titrated aggressively. Of course, I guess that there might be some questions about the titration given that insulin glargine did not match dulaglutide’s efficacy with the 1.5 mg dose, but I can give you some details. The 30 unit-per-day dose for insulin glargine was similar to what we have seen on other studies comparing a GLP-1 agonist to insulin glargine. The number of patients that were at less than 120 mg/dl fasting plasma glucose was around 55%, indicating that we were having fairly good success with the insulin glargine titration. Finally, if you look at the fasting plasma glucose curves in glargine-treated patients, the fasting glucose levels were reduced very rapidly, within two to three weeks, and stayed at that level through the rest of the study. We think that in this study, insulin glargine was effectively titrated. Of course, there was room to do better, but it would have been at the expense of hypoglycemia. In this study, patients were on maximally tolerated doses of glimepiride, and some patients had to reduce their glimepiride dose. When we titrated the basal insulin, we could have perfectly hit the targets, but we have to watch out for hypoglycemia.

Initiating GLP-1 Receptor Agonist Treatment as Intensification of Therapy

Johan Jendle, MD, PhD (Central Hospital, Karlstad, Sweden)

Dr. Johan Jendle discussed a different application for GLP-1 agonists: as an add-on to insulin (either basal or, occasionally, rapid-acting insulin). The majority of the presentation was given to a straightforward presentation of clinical data from trials such as GetGoal-Duo 1 (Sanofi’s Lyxumia [lixisenatide] added on to Lantus [insulin glargine]), the 4B trial (AZ’s Byetta [exenatide] vs. Lilly’s Humalog [insulin lispro] as add-on to insulin glargine), and AWARD-4 (Trulicity vs. insulin glargine in patients on insulin lispro). These and other studies suggest that adding a GLP-1 agonist to insulin can lead to a modest but significant improvement in A1c along with other benefits (weight, hypoglycemia, insulin dose).

Corporate Symposium: Debating the Next Step for Long-Term Control with GLP-1 Receptor Agonists – What’s New, What’s Next, What Now? (Sponsored by GlaxoSmithKline)

Panel Discussion: Weekly vs. Daily GLP-1 RAs – Convenience, Tolerability, Efficacy, and Safety

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

Professor Philip Home moderated a panel discussion discussing the advantages and disadvantages of once-weekly vs. daily GLP-1 agonists. Panelists considered the acute vs. chronic side effects and the adherence associated with injections and the timing. Overall, the panelists seemed generally comfortable with the side effect profile and saw potential adherence benefits of weekly agents. Below are some quotable quotes from the panelists:

  • “Regarding safety and tolerability, if you give an injection on one day and see an adverse reaction, then that’ll disappear. But if you give an injection that lasts a week, that adverse reaction will last several days. The good news is that with these agents, there doesn’t appear to be any particular issues with the injection or injection site. The concern is over acute vs. chronic adverse reactions; but with these agents, we have not seen any particular issues.” – Dr. Cliff Bailey (Aston University, Birmingham, UK)
  • “I’m concerned about the acute side effects, especially nausea, so I would favor once-daily.” – Dr. Hans DeVries (University of Amsterdam, Amsterdam, The Netherlands)
  • “The data suggest people marginally prefer once-weekly. I’m all in favor of patient preference. We thought people would hate injections but once a drug has weight loss, people don’t care. Weight loss way outweighs injections. I would vote once-weekly.” – Dr. David Matthews (University of Oxford, Oxford, UK)
  • “I believe that there is not the same answer for every patient. But among companies, everyone is going after once-weekly and hopefully once-monthly injections. There can be other avenues for improving efficacy and tolerability of these agents – intravenous delivery is the best. Ask companies to not forget about other developments, not just for longer-term injections.” – Dr. Michael Nauck (Diabeteszentrum Bad Lauterberg, Harz, Germany)
  • “If we’re getting similar efficacy, I would lean towards once-weekly. If people throw up a lot, I would do once-daily and see where they move.” – Dr. Jane Reusch (University of Colorado Denver, Denver, CO)

Debate: GLP-1 RAs or Insulin? Weighing Up an Individual’s Options: GLP-1 Agonists

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

To begin the mini-debate,Professor Philip Home was assigned to make the case for GLP-1 agonists over insulin. A pre-debate poll showed that the majority of attendees (64%) would advise GLP-1 agonists for type 2 diabetes patients that need to intensify treatment beyond oral agents (30% voted to advise insulin therapy and 6% voted to advise something else). Professor Home argued that GLP-1 agonists can offer similar glucose control as basal or as prandial insulin without the risk of weight gain and hypoglycemia. He stressed that GLP-1 agonists require less self-monitoring as these drugs do not require dose adjustments and do not need as many injections as insulin therapy does. Professor Home acknowledged that the disadvantages of GLP-1 agonists are GI side effects and injection site problems, although he pointed out that they have low rates of occurrence (1 in 10 people affected by nausea and 1 in 20 affected by injection site problems). Notably, Professor Home led a discussion on the “Butler hypothesis” in Q&A.

Debate: GLP-1 RAs or Insulin? Weighing Up an Individual’s Options: Insulin

Hans DeVries, MD (University of Amsterdam, Amsterdam, The Netherlands)

Dr. Hans DeVries then followed by defending insulin therapy. He argued that the GI side effects of GLP-1 agonists are problematic, showing data of high proportions of patients (up to 18%) dropping out of studies due to nausea. In addition, Dr. DeVries pointed out the lack of long-term data of GLP-1 agonists compared to insulin, which has had over 90 years of safety and efficacy data. However, Dr. DeVries acknowledged that GLP-1 agonists together with insulin is likely to be better than insulin (or a GLP-1 agonist) alone. Indeed, several years out, we do expect to see GLP/insulin combinations as a standard of care “first injectable” appropriate for many patients (presumably Sanofi’s lixi/Lantus and Novo Nordisk’s IDegLira will be approved globally then). The post-debate poll showed that attendees continued to prefer GLP-1 agonists over insulin, although slightly more attendees moved to the insulin side (57% for GLP-1 agonists and 34% for insulin).

Panel Discussion

Dr. Home: Can we go down the table and share our positions on the so-called Butler hypothesis and the overall issues of pancreatitis with these agents?

Dr. Hans DeVries (Academic Medical Center, Amsterdam, Netherlands): I think that pancreatitis is a very low-risk side effect with GLP-1 agonists.

Dr. David Matthews (Oxford Center for Diabetes, Endocrine, and Metabolism, Headington, UK): I think that the Peter Butler data has been pretty discredited based on where the samples were drawn from. I think that the reality is that cancer looks to be a very low likelihood event for GLP-1 agonists, probably negligible. I would have thought that the pancreatitis story has a tiny bit of evidence, but incredibly small risk. I think it’s a signal that is around, and you just need to be careful.

Dr. Jane Reusch (University of Colorado, Denver, CO): Diabetes is a very common disease, with over 300 million patients worldwide, and so pancreatitis has enough of a signal that you would want to have providers on the lookout rather than dismissing it. However, the benefit of optimizing glucose control is positive in other ways. You want to go ahead and treat patients’ diabetes, and if there are potential side effects, be vigilant and watchful. With cancer, you want to keep your eyes open as well, but I don’t think that there is any compelling evidence.

Dr. Cliff Bailey (Aston University, Birmingham, UK): It is extremely difficult to get the diagnosis exact in all of these studies between acute and chronic pancreatitis. The evidence from the Drucker lab is that exocrine ductal cells don’t actually express the GLP-1 receptor, so there probably isn’t a direct effect. I think that the product label is very well written for this class of agents, and provides us with the appropriate amount of caution.

Dr. Michael Nauck (Diabeteszentrum Bad Lauterberg, Harz, Germany): It’s a fact that treatment with GLP-1 agonists raises serum lipase levels to a certain degree. There is some interaction. Juris Meier and myself recently asked the companies producing these agents to tell us about the numbers of pancreatitis cases they observed in phase 3 trials, and if you look at the pattern, it certainly is not a ten-fold elevation like Peter Butler suggested, but it might be in the range of 20%, based on small numbers and with wide confidence intervals. I would be prepared for a slightly elevated risk with these agents, but we cannot be sure at this moment.

Q: How can we explain the differences in gastrointestinal profiles amongst the different GLP-1 receptor agonists?

Dr. Nauck: The observation is that if you expose someone to rapidly increased levels of a GLP-1 agonist as with a short-acting agent, this will provoke a lot of nausea and vomiting. If you increase the concentration more slowly, as with a long-acting GLP-1 agonist, then this prevents a lot of the problem. In addition, there may be some variation in how the different GLP-1 agonists penetrate into the brain area where nausea is sensed, but we do not have very strong research in this area.

Dr. Bailey: It’s always difficult to work out from each of the studies what it is that patients are reporting as nausea.

Dr. Nauck: To build on that, you often times see big incidences of nausea in the placebo groups for GLP-1 agonist studies because of the expectations. You don’t see that in trials of DPP-4 inhibitors.

Dr. David Matthews: We need to do a better job of setting patients’ expectations. Some patients go home and think that now that they are on this drug, they can eat massive meals. They learn, but over weeks. Part of the attenuation of nausea is patients learning about their meal capacity. You need to help patients learn that this drug acts to improve their satiety, and that they should move to smaller portion sizes.

Dr. Philip Home (Newcastle University, Newcastle upon Tyne, UK): Would you suspect that a GLP-1 agonist attached to a large molecule like albumin would cause less nausea?

Dr. Nauck: You do see less nausea with albiglutide, so that stands to reason, although we don’t have head-to-head studies.

Dr. Home: What are patient-reported outcomes like with these agents?

Dr. Nauck: Patient-reported weight loss is often a cause for patient satisfaction with these agents.

Corporate Symposium: Addressing Challenging Clinical Questions in Type 2 Diabetes Mellitus (Sponsored by Sanofi)

Debate: How Should Basal Insulin Therapy Be Intensified? With Rapid-Acting Insulin vs. With GLP-1 RAs

Stewart Harris, MD, MPH (Western University, Ontario, Canada)

In a packed session, Dr. Stewart Harris opened a mini-debate on how to best intensify basal insulin therapy with the assigned role of promoting rapid-acting insulin. He highlighted that “insulin is a drug we know and trust” and that we have over 90 years of clinical experience with it. Dr. Harris pointed out that insulin is recommended as an option for first-, second, or third-line therapy by all major national and international guidelines and has been shown to be safe in various vulnerable patient populations including infants, pregnant women, and critically ill patients. He also discussed how prandial insulin therapy does not need to be complicated, presenting the results of the START and AUTONOMY studies, which showed that patients can safely and effectively self-titrate. In addition, he referred to the FullSTEP study, which demonstrated that the stepwise approach is an effective alternative as this approach was shown to be just as efficacious as the full basal-bolus treatment approach. Overall, Dr. Harris drove home the ideas that insulin can be personalized and is simple to use and argued against GLP-1 therapy by noting that this drug class lacks widespread access and long-term data.

Debate: How Should Basal Insulin Therapy Be Intensified? With GLP-1 RAs

Julio Rosenstock, MD (Dallas Diabetes and Endocrine Center, Dallas, TX)

Dr. Rosenstock followed with a case for GLP-1 agonists, by first admitting that while rapid-acting insulin can effectively reduce A1c, it is a declining treatment option. To back up this view, he argued that insulin’s glucose-lowering efficacy is highly inconsistent and user-dependent and comes with the high costs of hypoglycemia and weight gain. Dr. Rosenstock called GLP-1 agonists simpler to use (as insulin requires the “complex, costly intervention of SMBG” to strictly manage dose titration) and noted that they may offer a more tailored post-prandial glucose intervention. He also stressed that this drug class leads to A1c reductions without increased hypoglycemia risk and is either weight-neutral or leads to weight loss. Dr. Rosenstock did, however, agree that GLP-1 agonists can be limited by cost, GI intolerance, and uncertain long-term safety. Therefore, he concluded his end of the debate highlighting that basal insulin analogs and GLP-1 agonists have complementary effects and that the “future is going to be GLP-1 agonists in fixed combination with insulin.”



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

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