European Association for the Study of Diabetes (EASD)

September 14-19, 2014: Vienna, Austria – Full Report – Insulin Therapy

Executive Highlights

It was a relatively insulin-heavy year at EASD, featuring new or expanded data on several novel insulin products. We saw full phase 3 data for the first time on Novo Nordisk’s Ryzodeg twice daily (pre-mixed Tresiba [insulin degludec]/Novolog [insulin aspart]) – the trial barely missed achieving statistical non-inferiority vs. basal-bolus therapy with Tresiba and Novolog, but the A1c reductions were numerically similar. We expect that the reduced injection burden with Ryzodeg will be a major advantage from the patient perspective, which we hope will lead to improved adherence to treatment. We also saw full results for the first time from the open-label EDITION IV trial comparing Sanofi’s Toujeo (concentrated insulin glargine U300) to Lantus (insulin glargine U100). Overall, the trial did not provide compelling evidence that Toujeo represents a major advance over Lantus, but it did suggest that the products are at least comparable, with possible modest benefits with Toujeo regarding weight gain and hypoglycemia. In addition, we were excited to see new data from a JDRF/AZ-supported phase 1 study comparing three fixed-dose combinations of pramlintide/regular human insulin to human insulin alone. As Dr. Matthew Riddle (Oregon Health and Science University, Portland, OR) put it, the results were “quite dramatic” – every pramlintide/insulin fixed-dose combination reduced incremental glucose AUC by >50% in the first three hours following a meal.

EASD also featured valuable high-level commentary on the coming wave of biosimilar insulins and the comparative advantages of various novel insulins in development. Dr. Melanie Davies (University of Leicester, Leicester, UK) said that biosimilar insulins will ultimately be very beneficial for patients by increasing access to a wider array of treatment options, but she reviewed several key challenges associated with the complex manufacturing process and uncertain regulatory requirements for biosimilars and cautioned that the cost reductions with these products will likely be less dramatic than anticipated.  Dr. Thomas Danne (Diabetes Center for Children and Adolescents, Hannover, Germany) and Dr. David Russell-Jones (University of Surrey, Surrey, UK) both provided enthusiastic overviews of new insulin preparations, with much discussion centered around Lilly’s novel basal insulin peglispro (BIL) and Novo Nordisk’s Tresiba (insulin degludec). There was also much excitement on the GLP-1/basal insulin fixed-dose combination front – see our EASD GLP-1 agonist report for full coverage of those products.

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

 

Table of Contents 

Insulin

Oral Presentations: Novel Insulin Formulations and Combinations

Treatment Intensification with IDegAsp BID vs. IDeg OD Plus IAsp in Insulin-Treated Patients with type 2 Diabetes: A Randomized, controlled Phase 3 Trial

John Cooper, MD (Stavanger University Hospital, Stavanger, Norway)

Dr. John Cooper presented results from a phase 3 trial in which Novo Nordisk’s Ryzodeg twice daily (pre-mixed Tresiba [insulin degludec]/Novolog [insulin aspart]) barely missed achieving statistical non-inferiority vs. basal-bolus therapy with Tresiba and Novolog, although the A1c reductions were numerically similar. We heard this trial’s topline results during Novo Nordisk’s 1Q14 update, but this was the first time we had seen the data in full. The open-label trial, which enrolled 274 patients with type 2 diabetes, very narrowly missed the threshold for non-inferiority (the upper bound of the 95% CI was 0.41% and the non-inferiority margin was 0.40%), but there was no statistically significant difference in average A1c reduction between the two groups (1.3% reduction with Ryzodeg vs. 1.5% with basal-bolus, both from a baseline of 8.3%). Compared to the basal-bolus therapy, treatment with Ryzodeg led to a 12% lower insulin dose and slight reductions in weight gain (~1 kg difference) and hypoglycemia (rates of overall and nocturnal confirmed episodes were 19% and 20% lower, respectively). Dr. Julio Rosenstock, (Dallas Diabetes and Endocrine Center, Dallas, TX) however, argued during Q&A that “whatever claims you make [regarding weight and hypoglycemia) cannot be sustained…when you don’t reach the primary objective of non-inferiority. In addition, Dr. Cooper shared data confirming that patients perceive additional injections as a burden and that mental health scores were significantly improved in the Ryzodeg group by the end of the study. We expect that from a patient perspective, the reduced injection burden is a major advantage of Ryzodeg, which will hopefully lead to better patient adherence.

  • The 26-week, open-label phase 3 trial randomized 274 patients with type 2 diabetes to receive either Ryzodeg twice daily (n=138) or full basal-bolus therapy with Tresiba and Novolog (n=136). All patients were on basal insulin and at least one oral agent at baseline. The mean baseline A1c was 8.3% and mean BMI was 32 kg/m2; the mean age of patients was ~60 years, with an average duration of diabetes of 11-13 years.
  • The average A1c reduction was 1.3% with Ryzodeg, statistically comparable to the 1.5% reduction achieved with basal-bolus therapy. Though the difference of 0.2% between the two groups was well below the 0.4% threshold for non-inferiority, the upper bound of the 95% confidence interval was 0.41%, meaning the trial technically did not meet its primary endpoint of non-inferior A1c reductions with the two treatment regimens.
  • Secondary endpoints, which included fasting plasma glucose (FPG), insulin dose, weight gain, and rate of hypoglycemia, were either comparable between the two groups or slightly in Ryzodeg’s favor. Reductions in FPG were statistically similar between the two arms (41 mg/dl with Ryzodeg vs. 32 mg/dl with basal-bolus). Rates of overall and nocturnal confirmed hypoglycemia were numerically but not significantly lower with Ryzodeg (reductions of 19% and 20%, respectively). There were significant differences in favor of Ryzodeg with regard to weight gain (~1.04 kg difference, or 2.2 pounds) and insulin dose (12% lower dose with Ryzodeg). However, Dr. Rosenstock argued during Q&A that it is impossible to make any claims about secondary endpoints when the trial failed to meet its primary efficacy endpoint, and Dr. Cooper himself admitted that the small weight difference may not be clinically significant.
  • From a patient perspective, improved quality of life due to fewer daily injections will likely be Ryzodeg’s most significant advantage over basal-bolus therapy. Dr. Cooper shared results from a survey demonstrating that patients’ perceived treatment burden increases linearly with the number of daily injections and that many patients on full basal-bolus therapy will often intentionally skip some injections. In addition, results of an SF-36 quality of life questionnaire demonstrated that patients’ self-reported mental health improved significantly in the Ryzodeg group over the course of the study. Therefore, though Dr. Cooper said he generally prefers basal-bolus therapy over premixed insulin, he believes Ryzodeg could represent an appealing alternative for those who find adherence to a basal-bolus regimen particularly challenging.

Questions and Answers

Q: There’s been an ongoing debate for years over whether premixed insulin or basal-bolus is preferred for type 2 diabetes. You said participants were taking two to four injections per day – do you know the percentage of patients who were taking two vs. three vs. four injections and the differences in their A1c control? You also didn’t show how many blood glucose measurements were done; was the correction for individual glucose levels measured?

A: I don’t have the breakdown of the percentage using two, three, or four injections on this slide deck. About 80% were using at least three injections of rapid-acting insulin.

Q: And how often were the blood glucose measurements done?

A: I don’t have the figures for that.

Q: With basal-bolus therapy, you have three injections of rapid-acting insulin for three meals. With the combination, how do you cover all the meals? What was the timing of injections?

A: The premixed insulin was given before breakfast or lunch and always before the evening meal. But you can’t cover all three meals; one will escape.

Q: Was there any difference between the patients who were on metformin and those who were not, in terms of behavior or insulin dose?

A: I don’t have the figures for that.

Q: Do you have data on patient satisfaction or quality of life?

A: We did do an SF-36 questionnaire at the start and the end of the trial. The results showed that there was no difference in physical scores between the groups, but there was a significant improvement in mental scores in the combination group, mainly driven by an improvement in social functioning.

Dr. Julio Rosenstock (Dallas Diabetes and Endocrine Center, Dallas, TX): The primary outcome of the study was non-inferiority. You said that you did not reach that margin. Therefore, your conclusion that they’re similar is not correct. In principle, you cannot claim that these two regimens are similar. The difference in A1c change of 0.2%, the doses of 1.1 U/kg vs. 1.3 – whatever claims you make cannot be sustained when two groups do not have the same dose and don’t reach the primary objective of non-inferiority.

A: Thank you for making that important point clear.

Glycemic Control and Hypoglycemia With New Insulin Glargine 300 U/mL in People With Type 1 Diabetes (EDITION IV)

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

Professor Philip Home presented full results for the first time from the open-label EDITION IV trial, which compared the safety and efficacy of Sanofi’s Toujeo (concentrated U300 insulin glargine formulation) to that of Lantus (insulin glargine U100) in type 1 diabetes patients on prior MDI treatment. We first saw data from this trial presented as a late-breaking poster at this year’s ADA, where we learned that both formulations achieved similar A1c reductions (-0.4% from a mean baseline of 8.1%). One of the major takeaways from the results was that there was significantly less nocturnal hypoglycemia with Toujeo during the first eight weeks of the trial, although Professor Home noted that the study was fairly underpowered to examine hypoglycemia and that the incidence of overall hypoglycemia was similar between the groups. Additionally, during the first four weeks of the trial, the Toujeo group experienced a transient but marked ~30 mg/dl deterioration in fasting plasma glucose control that could have contributed to the initial reduction in hypoglycemia – this was discussed in depth during Q&A (see below). Patients on Toujeo experienced significantly less weight gain (a difference of around half a kilogram) than did patients on Lantus. Overall, EDITION IV does not provide compelling evidence that Toujeo is a major advance over Lantus, but does suggest that Toujeo is (at the very least) comparable, with possible modest benefits on weight and nocturnal hypoglycemia.

  • EDITION IV randomized 549 type 1 diabetes patients on multiple daily insulin injections to either Toujeo or Lantus in an open-label fashion – the study could not be blinded, as there were differences in the insulin pens for each formulation. The minimum dose steps were 1.5 U for Toujeo and 1.0 U for Lantus. The study ran for six months (results from a six-month extension study will be presented at a later date) and 85% of the enrolled population completed the study.
  • At baseline, patients had an average age of 47, average BMI of 28 kg/m2, and average A1c of 8.1%. Over 80% of patients were on Lantus entering the study, with basal insulin accounting for slightly over half of their daily insulin requirements.
  • Before presenting the primary efficacy results, Professor Home provided some details on the insulin doses over the course of the study. Consistent with other EDITION studies, the Toujeo group had an average daily insulin dose ~10% greater than the dose in the Lantus group. In both groups, there was a slight reduction in daily insulin dose at study initiation followed by a subsequent rebound that fully stabilized by around week 12.
  • Both insulin glargine formulations achieved a mean A1c reduction of 0.4% from baseline, but SMBG measurements found a transient relative spike in fasting plasma glucose (~30 mg/dl) in the first few weeks of the trial in the Toujeo arm.  The spike occurred during week one of the trial, it began to wane by week two, and it was no longer present by the end of week four. This is understandable, as providers and patients who did not have longstanding familiarity with the new formulation likely erred on the side of caution in their initial titration protocol. However, this result affected our interpretation of the hypoglycemia data presented later in the trial.
  • In the first eight weeks of the trial, the Toujeo arm saw a statistically significant 31% reduction in confirmed or severe nocturnal hypoglycemia compared to Lantus (95% CI: -9% to -47%). There was no apparent difference in hypoglycemia (nocturnal or otherwise) beyond eight weeks. This was an intriguing finding, given that some other studies in the EDITION program (including EDITION I) only found a hypoglycemia benefit later in the trials, from months three to six. There were too few severe hypoglycemic events to draw a meaningful comparison.
  • Adverse events, including serious adverse events, hypersensitivity reactions, and injection site reactions, were fairly balanced between groups.
  • Toujeo’s efficacy and effect on hypoglycemia were statistically similar with morning vs. evening administration, although it looked to us like there was a hint of a positive trend with morning administration (A1c benefit of 0.15% [NS] between Toujeo morning and evening groups, ~5% less nocturnal and total hypoglycemia [NS]). From a Kaplan-Meier curve, it appeared that the difference in nocturnal hypoglycemia was more pronounced between Toujeo and Lantus than any difference between morning and evening administration of either formulation might have been.
  • We would be cautious in interpreting any major signs of benefit with respect to hypoglycemia in this trial, given that the differences were quite small numerically and some of the time ranges examined for hypoglycemia were not pre-specified endpoints. Additionally, even if there is a slight reduction in hypoglycemia for the first few weeks of treatment, we wonder whether it would be perceived as a major decision factor for patients and providers, given that the benefit did not to extend beyond eight weeks and patients are usually on insulin treatment for the long haul.

Questions and Answers

Q: Could the difference in nocturnal hypoglycemia during the first eight weeks be due to the strange dose reduction seen early in the study in the U300 group?

A: That dose reduction, which was partially determined by the study protocol, should not account for that much of a reduction. It may partially account for what is happening, but we have also seen a hypoglycemia advantage in some of the type 2 diabetes studies. I think the difference is real – to what extent it reflects dose or titration vs. differences in properties is debatable.

Dr. Julio Rosenstock (Dallas Diabetes and Endocrine Center, Dallas, TX): I have struggled to understand the value of this insulin. The nocturnal hypoglycemia benefit was only seen here during the first eight weeks, and you don’t see any difference subsequently. You also showed that during the first eight weeks of the trial, fasting glucose levels went up. Couldn’t that explain the hypoglycemia difference you saw? It is very hard to claim less hypoglycemia during the first eight weeks when you have that deterioration in fasting plasma glucose. Also, what definition did you use for nocturnal hypoglycemia?

A: While I take your point, I don’t like the word “claim.” I am merely here to report the results, and the results do seem to support an effect. We used a definition of midnight to six AM, which was forced upon us by regulators that don’t understand the issues. I think that a more appropriate definition might be from bedtime to patients’ breakfast injection.

Q: What would be your explanation for the divergent doses of insulin between the two groups?

A: There is usually a roughly 10% higher dose with the U300 formulation. Some of us theorized that this might be because insulin is degraded more quickly if it sits around in the subcutaneous tissue, although not everyone agrees with that explanation. I do not think we know the full explanation in scientific terms.

The Effect of Insulin Degludec in Combination with Liraglutide and Metformin in Patients With Type 2 Diabetes Requiring Treatment Intensification

Vanita Aroda, MD (MedStar Health Research Institute, Hyattsville, MD)

Dr. Vanita Aroda presented new data from a 26-week trial evaluating Novo Nordisk’s Tresiba (insulin degludec) as an add-on to the GLP-1 agonist Victoza (liraglutide) and metformin, compared to a placebo insulin injection. Though this was not a study of Xultophy (insulin degludec/liraglutide fixed-ratio combination), adding Tresiba to Victoza produced efficacy results approaching what we saw in the DUAL I trial. Following a 15-week liraglutide run-in in which mean A1c fell from ~8.3% to ~7.5%, patients that added insulin degludec experienced an additional placebo-adjusted A1c reduction of 0.9%, bringing the group to a striking mean A1c of 6.5%. In addition, 78% of the insulin degludec group achieved a final A1c less than 7%, compared to 36% of the placebo group. Insulin degludec did cause nearly five times more confirmed hypoglycemia than was seen in the placebo group, but Dr. Aroda characterized the absolute incidence of hypoglycemia (17%) as relatively low for an insulin, and noted that there were no cases of severe hypoglycemia in the study. The insulin degludec group gained 2 kg (4.4 pounds) while the liraglutide group lost 1 kg (2.2 pounds) on average, but at baseline the insulin degludec group weighed around 3 kg (6.6 pounds) less, and both groups ended the trial with the same mean weight. More than anything, these very solid results increase our enthusiasm for the just-approved Xultophy fixed-ratio combination, which should preserve the efficacy seen in this trial and also offer greater convenience and less nausea (due to the slow up-titration protocol). See our report on Xultophy’s recent EU approval for more detail on why we are so excited about this compound.  

  • This randomized, placebo-controlled, 26-week study evaluated insulin degludec as an add-on to liraglutide in 346 type 2 diabetes patients. Notably, the study did not enroll patients who had been on longstanding liraglutide therapy. Rather, patients had to have been on metformin (possibly along with other oral medications or Byetta [exenatide]) for at least three months, and were subsequently started on liraglutide treatment for a 15-week run-in period before being randomized to add either insulin degludec or a placebo injection. It is not likely that the results would have been very meaningfully different if the study tested patients on longstanding liraglutide therapy, but we wonder how this might impact scientists’ and providers’ perception of the results.
    • Of the 970 patients that entered the run-in period, 624 were ineligible for randomization because liraglutide alone was able to bring their A1c below 7%.
    • In the randomized portion of the trial, the completion rate for the insulin degludec group (92%) was much higher than the rate seen in the placebo group (76%). A potential clue to this can be found in the graph Dr. Aroda displayed of the daily insulin doses over the course of the trial. The dose in the insulin degludec group reached 51 U after 26 weeks, but the placebo group (which continually up-titrated its dose, with no result) had a mean “placebo dose” of 105 U. Many patients in the placebo group (and providers that cared for them) may have been able to deduce what group they were in, which could have led to greater dropouts.
    • At baseline, patients had a mean age of 57 years, mean diabetes duration of just under 10 years, mean A1c of 7.5%, and mean BMI of 32 kg/m2.
  • Following a pre-randomization (liraglutide run-in) A1c decline from ~8.3% to ~7.5%, the group randomized to add-on insulin degludec lost an additional 1.0%, for a placebo-adjusted reduction of 0.9%. Strikingly, this left the insulin degludec group’s final mean A1c at 6.5%, with 78% of the group (compared to 36% of the placebo group) under an A1c goal of 7%. As expected, this improvement in overall glycemic control was likely driven by the 40-50 mg/dl reduction in fasting plasma glucose seen with insulin degludec + liraglutide relative to placebo + liraglutide.
  • There were no worrying or significant imbalances in adverse events in the trial.

Questions and Answers

Dr. Julio Rosenstock (Dallas Diabetes and Endocrine Center, Dallas, TX): I think this is a much better study than the one reported previously studying insulin detemir added to liraglutide, because in this trial you have a control injectable. In the detemir study, the A1c came down to 7.1% after six months, but the detemir dose in that study was 31 units compared to about 50 here. I am curious about whether the better results here were due to a higher dose or the different effects of the insulin. The best study for the future would be a head-to-head comparison of insulin detemir and insulin degludec.

A: How could I disagree with what you have to say, Dr. Rosenstock? By way of comparison, the dose achieved in this trial was comparable to other trials on insulin degludec, but I agree that the only way to know for sure is through head-to-head study.

Similar Efficacy and Safety with LY2963016 Insulin Glargine Compared with Insulin Glargine in Patients with Type 1 Diabetes Mellitus: The ELEMENT 1 Study

Robyn Pollom, NP (Lilly, Indianapolis, IN)

Ms. Robyn Pollom shared results, also presented at ADA, from the phase 3 ELEMENT 1 study demonstrating comparable efficacy and safety profiles with Lilly/BI’s biosimilar insulin glargine and Sanofi’s Lantus. The open-label trial randomized 535 patients with type 1 diabetes (baseline A1c = 7.8%; BMI = 25 kg/m2) to receive basal-bolus therapy with either the investigational glargine (n=268) or Lantus (n=267), both in combination with Lilly’s Humalog (insulin lispro). The study met its primary endpoint of non-inferior A1c reduction at 24 weeks (reductions of 0.35% with the investigational glargine vs. 0.46% with Lantus), though there was a significant difference between the groups at the 12-week mark, which Ms. Pollom suggested may have been due to less aggressive titration in the investigational glargine arm; non-inferiority was maintained at the end of the 52-week follow-up period (reductions of 0.25% with the investigational glargine vs. 0.28% with Lantus). The percentage of patients achieving an A1c >7% was also comparable between groups (35% of the investigational glargine group vs. 32% of the Lantus group), as were secondary endpoints related to fasting plasma glucose, body weight, daily insulin dose, and rate of adverse events. Based on these results, Ms. Pollom concluded that there were no clinically meaningful differences between the two insulin glargine formulations.

  • As a reminder, Lilly/BI’s glargine formulation received European approval on September 10, making it the first insulin ever approved under the EMA’s biosimilar pathway. Sanofi’s patent protection for Lantus does not expire until mid-2015, so Lilly/BI’s formulation cannot be launched until then. The product also received “tentative” approval from the FDA on August 18, but a US launch is likely delayed until mid-2016 due to ongoing patent litigation by Sanofi. Still, we believe global regulatory agencies have certainly made clear their interest in and approval of (so to speak) biosimilars. The approved brand name for the formulation in Europe is Abasria, while the provisional trade name in the US is Basaglar; however, the companies plan to announce a single global trade name at a later date.

Questions and Answers

Q: Why were there so many studies on antibody formation, since they’re so low and have no influence on the metabolic situation? What are the differences?

A: In general, the biosimilar guidance is clear that immunogenicity is an area of special interest around the development of biosimilars, since the amino acid sequence is the same but there are differences around the manufacturing process. We did several tests to confirm that, so we would have a totality of evidence to provide data that would leave no open questions in people’s minds.

Q: Some of the patients switched from Lantus to the biosimilar. What was the behavior in terms of efficacy of those who switched? Could that account in some way for the difference in the initial period?

A: 85% of the patients came in on Lantus, so most of the data is from that group. We have looked at the subgroup of just those on pre-study Lantus and found no difference in this population relative to the total population in terms of efficacy and glucose lowering. We saw a significant difference at 12 weeks. We have a number of reasons for that; one that we looked at was dose titration, and although the doses were similar, there was less aggressive titration in the LY arm, which may have been somewhat due to the open-label nature of the study. Titration was supposed to occur in 12 weeks; it was a little less aggressive, and then it came together from 12-18 weeks and we no longer saw a significant difference.

Recombinant Human Hyaluronidase Pretreatment Of CSII Cannula Sites Provides Comparable Glycemic Control With Reduced Hypoglycemia In T1dm Compared To Usual CSII: Results of the CONSISTENT-1 Study

Jay Skyler, MD (University of Miami, Miami, FL)

Dr. Jay Skyler presented results from the CONSISTENT-1 trial testing Hylenex pretreatment in insulin pumpers (n=342) vs. standard pump therapy with rapid-acting insulin alone (n=113). Topline data was released in April and full results were in shared in a late-breaking ADA 2014 poster (85-LB). As a reminder, the trial met its primary endpoint of a non-inferior A1c reduction (0.4% margin) at six months between the treatment and control groups; A1c fell 0.14% with Hylenex pretreatment vs. 0.18% with standard pump therapy, both from a baseline of 7.7%. The rate of overall hypoglycemic events (≤70 mg/dl) dropped a modest 12% with Hylenex pretreatment (p=0.11). More notably, hypoglycemic events <56 mg/dl dropped by 23% (p=0.02); nocturnal hypoglycemic events (≤70 mg/dl) declined by 21% (p=0.02); and the rate of severe hypoglycemic events (requiring assistance) dropped by a notable 61% (p=0.08). Dr. Skyler did an excellent job of explaining that bolus timing impacted the results quite meaningfully, a finding that was presented rather confusingly in the ADA poster. He also emphasized that Hylenex pre-treatment enabled consistent insulin absorption over the three days of infusion set life, enhancing predictability for patients. As of Halozyme’s 2Q14 update, the company was still talking to the FDA about a label update. We continue to wonder about the hassle factor and reimbursement of Hylenex relative to the clinical benefit. Tough to say how payers and patients will weigh the pros and cons. A co-formulation would be most ideal, though Dr. Skyler’s opening slides suggested this is being considered for an MDI indication. It was notable in Q&A to hear Dr. Julio Rosenstock quite focused on “time in zone” and conveying the importance of this measure.

  • Impact of bolus timing: Patients who took boluses within 15 minutes before meals saw a statistically significant reduction in hypoglycemia with Hylenex, while those who gave a bolus within 15 minutes after meals saw no significant hypoglycemia benefit. However, those who bolused within 15 minutes after meals did see a ~40% lower postprandial excursion (p=0.11) with Hylenex relative to the control group.
  • CGM curves demonstrated a significant postprandial advantage to using Hylenex at all three meals. Hylenex pre-treatment led to lower postprandial excursions at breakfast (~15 mg/dl lower at 90 minutes), lunch (~10 mg/dl lower), and dinner (~5 mg/dl lower). The Area Under the Curve was substantially reduced at breakfast, in particular, while the benefit at lunch and dinner was fairly marginal.
    • In Halozyme’s 2Q14 call, management said there was no difference in mean 90-minute postprandial glucose excursions between the Hylenex and control groups (+19 mg/dl with Hylenex and +20 mg/dl with standard pump therapy). We assume this was SMBG data, as the CGM profiles showed by Dr. Skyler did show a benefit and did not match this data. As a reminder, ~30% of trial participants wore CGM in the study. This served as yet another reminder of why CGM is critical for assessing the clinical benefit of ultra-rapid-acting insulin approaches.
  • Hylenex made insulin absorption more consistent and predictable over the three days of infusion set wear. The day one, day two, and day three postprandial glucose excursion curves were nearly identical in the Hylenex group, while the control group saw a consistent acceleration in insulin absorption from day one to day three. The implication is that more consistent and predictable absorption could help patients maintain better control and avoid hypoglycemia.
  • Adverse events were balanced between treatments except for “generally mild infusion site events” – 20% of Hylenex patients and 8% of control group patients experienced “any general disorder or administration site condition.” The biggest difference came in two categories: pain, discomfort, or paresthesia (16% vs. 6%) and hematoma/bruising/hemorrhage (4% vs. 0%).
  • In CONSISTENT 1, patients used an infusion set connector/syringe to administer Hylenex at each infusion set change, which added ~three minutes to the site change process. The trial also reported that >98% of study participants reported that the infusion site change process was either “very easy” or “easy” for both treatment groups.

Questions and Answers

Q: We recommend our patients change their infusions et after two days, since they get more hypoglycemia on the third day of wear.

Dr. Skyler: The higher hypoglycemia rate on the third day is due to improved insulin action. By standardizing it with PH20, hopefully patients can achieve more consistent control and not need to change their set every two days.

Dr. Tim Heise (Profil, Neuss, Germany): With conventional CSII, you might see an improvement in efficacy due to inflammation. Could the effect of hyaluronidase be due to an inflammation signal?

Dr. Skyler: The frequency of injection site reactions was low. We saw some minor changes in skin reactions, but the inflammatory component was not something we measured.

Dr. Julio Rosenstock (Dallas Diabetes & Endocrine Center, Dallas, TX): You’re talking about a different paradigm in CSII to help improve glucose control. I agree. But I don’t think we can judge it by changes in A1c. A1c is not going to change, but you’re going to eliminate the ups and downs. It will be more uniform – the same A1c, but with less variability.

Dr. Skyler: That may very well be. The goal here was to get consistency in excursions. Before we started, I didn’t think we would. I was actually surprised by the results.

Dr. Rosenstock: People expect to see better A1cs. But even if you have the same A1c, but less peaks and valleys, that has value.

Dr. Skyler: I think it will. These were long-term pump users with five to 10 years of pump therapy experience. They know how to subtly adjust doses. Despite that, we were able to generate the consistency of insulin absorption.

A Novel Concentrated Recombinant Human Insulin Formulation with Improved Ultra-Rapid Action for Continuous Subcutaneous Infusion Therapy

Roderike Pohl, PhD (Biodel, Danbury, CT)

Dr. Roderike Pohl presented findings from a preclinical study investigating the pharmacokinetic and pharmacodynamics profile of Biodel’s U400 ultra-rapid-acting human insulin (BIOD-531). As a reminder, we have already seen positive topline phase 2 data from the compound, which showed superior postprandial control and better time in range compared to Lilly’s Humalog Mix 75/25 and Lilly’s Humulin R 500. Dr. Pohl acknowledged that the rapid absorption profile and concentrated formulation make BIOD-531 an ideal candidate for pump therapy in insulin-resistant type 2 patients, as well as for possible use in closed-loop systems (e.g., the U400 formulation would enable a smaller reservoir). Ongoing studies are evaluating lower doses of BIOD-531 for these potential uses – the key for pump use would be to conserve the ultra-rapid profile, but avoid the extended duration of action seen at larger doses.

Questions and Answers

Q: The results indicated multiple humps in pharmacokinetic curves for BIOD-531. What’s going on there?

A: We’re still investigating the nature of the insulin during this absorption phase. What I believe is happening is that some insulin is monomeric, while some is hexameric. These monomeric units are absorbed immediately and then we see the sustained absorption of hexameric units that are broken down over time. I think this combination is of great benefit for those patients in basal-bolus-type situations.

Q: How is this compound different from the Viaject formulation?

A: The insulin concentration has been increased, and EDTA has been increased for stability and for a faster profile. Viaject was related to this, but this is a completely different profile.

Q: What was the rationale for incorporating magnesium sulfate into the compound?

A: Magnesium sulfate is provided to reduce the injection site discomfort that was seen with the Viaject formulation. The very first formulation of this compound did not have magnesium sulfate in it. Subsequently, we have added it and have done studies to show no difference in the pharmacokinetic and pharmacodynamic profile.

Oral Presentations: Novel Compounds on the Horizon

Pramlintide-Insulin Fixed-Dose Combination: A Phase 1 Dose Ratio-Finding Study In Patients With Type 1 Diabetes Mellitus

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

Dr. Matthew Riddle presented new data from a JDRF- and AZ-supported phase 1, dose-finding, meal challenge study comparing three fixed-dose combinations of pramlintide/regular human insulin (at 6, 9, and 12 mcg/unit of insulin) to regular human insulin alone. The insulin and pramlintide were administered by separate injections. The randomized, single masked, four-way crossover study enrolled 19 type 1 patients (mean A1c 7.8%, mean BMI 26 kg/m2) who ate a standardized 600-calorie meal for breakfast (insulin glargine the night before); glucose and glucagon AUC, along with blood glucose, were followed for three hours post meal. Notably, Dr. Riddle called the results “quite dramatic” – every pramlintide/insulin fixed-dose combination reduced incremental glucose AUC0-3 h by >50%, and all dose ratios reduced incremental glucagon AUC0-3 h by >50%. The results were highly statistically significant, and were even more notable given that the mealtime insulin dose was reduced by 30% in the pramlintide arms. Interestingly, there were no significant differences between any of the ratios. A plot of postprandial glucose showed the most dramatic improvement with the fixed-dose combination 30-120 minutes post-meal (e.g., ~155 mg/dl [insulin+pramlintide] vs. ~260 mg/dl [insulin alone] at 60 minutes), though by three hours, postprandial glucose was not statistically different between the arms of the study (~260 mg/dl). Dr. Riddle did not say it specifically, but that could imply that slowed gastric emptying was largely driving the improved glycemia with the insulin/pramlintide combination (which had worn off by three hours). 

  • Dr. Riddle covered the rationale for a fixed-dose combination of pramlintide and insulin. He explained that amylin is a second beta cell hormone that is co-secreted with insulin, implying that a fixed ratio would mimic physiology. Amylin has three modes of action: suppressing elevation of glucagon (especially following meals); slowing gastric emptying; and reducing food intake and enhancing satiety.
  • Pramlintide (Symlin) is a stable injectable analog of amylin currently sold by AstraZeneca. It reaches a peak after subcutaneous injection at 20 minutes, with a total duration of approximately three hours. Pramlintide is approved in the US at a fixed dose of 60 mcg in type 1 or 120 mcg in type 2, irrespective of the insulin dose. Dr. Riddle noted that pramlintide is not extensively used in real-world practice, in part due to the challenges of dosing it correctly.
  • Based on prior work (Wayer et al., Diabetes Care 2003), this study elected to use regular human insulin instead of a rapid-acting analog. The earlier study found that plasma glucose response was flatter when pramlintide was combined with regular human insulin vs. insulin lispro. It’s unclear if future studies will try using a rapid-acting analog or not.
  • All dose ratios were well tolerated, and no treatment related hypoglycemia was reported. One participant reported nausea in all three pramlintide/insulin dose ratios, and one reported abdominal pain and diarrhea. The positive safety data, while of short duration, was encouraging to see at this early stage – real-world patient experience dosing pramlintide is often hampered by hypoglycemia. We wonder if future studies will back off the 30% reduction in bolus insulin, since all doses were well tolerated.
  • The insulin and pramlintide were administered as separate subcutaneous injections; we would be very interested to see a co-formulation tested in the future.
  • Further studies are planned in ongoing basal-bolus treatment. The next goal is to see what happens when there is fixed-ratio delivery of pramlintide with insulin during the basal state, as well as with meals. The hypothesis is that this may work better than when delivery is with a single meal. Q&A implied these will take place over 24 hours. We believe this fixed-dose approach has particular promise in a pumped formulation for the artificial pancreas, where meal excursions are still a challenge for control algorithms, and where bolus dosing can be more creative.
  • This study was supported by JDRF and AstraZeneca. As a reminder, JDRF and Amylin partnered back in 2011 to co-formulate pramlintide and insulin. BMS/AZ purchased In June 2012 for ~$7.0 billion, and AstraZeneca acquired BMS’ diabetes business in December 2013 for $2.7 billion in cash plus milestones. The acquisition included Symlin (pramlintide). Little further information has been given on this partnership.

Questions and Answers

Dr. Tim Heise (Profil, Neuss, Germany): Nice presentation. Do you have data for more than three hours? It seemed that blood glucose was still rising at the end of three hours with the fixed-dose combination.

Dr. Riddle: We do not have data after three hours. It’s an important question. As expected, there was a rather strong suppression of postprandial glucose, with all three doses tested. There were essentially no nausea and hypoglycemia concerns. We feel safe proceeding over a longer 24-hour period. Another reason for the failure over time to continue to suppress glucose and glucagon levels was a dosage reduction in the pramlintide arm. Also, basal insulin was provided by glargine, and in these insulin sensitive people, blood levels of glargine were probably waning at that time. We don’t know the answer.

Dr. Heise: As you pointed out, these were insulin sensitive subjects with relatively low doses. Would you develop another dose ratio for type 2 diabetes patients? Or is this intended for type 1 only?

Dr. Riddle: We’re starting with type 1 diabetes. There is the greatest potential benefit in type 1, particularly in a nice closed-loop system. I am very interested myself in type 2 diabetes. We have done previously, a head to head comparison of a rapid-acting insulin vs. pramlintide on top of glargine. It was in this same dosage, 120 mcg, with meals. It turns out to about the same ratio. The results were interchangeable – non-inferiority. So in type 2 diabetes, you can get a quite considerable prandial effect with pramlintide.

Oral Presentations: Insulin – Clinical Decision Making

The INITIATOR Study: Real-World Treatment Patterns and Outcomes in Patients with Type 2 Diabetes Initiating Insulin Glargine or Liraglutide

Philip Levin, MD (MODEL Clinical Research, Baltimore, MD)

Dr. Philip Levin presented full results from the Sanofi-sponsored INITIATOR trial, a real-world observational longitudinal cohort study comparing baseline characteristics, safety, efficacy, and cost for patients that started on either Lantus (insulin glargine) or Novo Nordisk’s GLP-1 agonist Victoza (liraglutide). As Dr. Levin made very clear, the study was not intended to compare safety or efficacy for the two agents, but rather provide insights into the differences in the patient populations and reported outcomes between groups that opted for either drug. Patients that initiated injectable therapy with Victoza had a lower baseline A1c (~8.1% vs. ~9.7%) than those that began on Lantus and were around 10 kg heavier – weight loss was cited as a reason favoring initiation on Victoza in 21% of patients in that group. After one year follow up, patients beginning Lantus experienced a mean A1c reduction from baseline of 1.2% compared to 0.5% with Victoza, although the very different baseline A1c values color these results. Interestingly, Lantus showed an advantage over Victoza with regards to persistence to treatment (64% vs.50%), although persistence was indirectly measured through claims data. Lantus also led to less of a gain in drug-related cost borne by payers ($300-$1000 per six months), but given that data was reported as change from baseline rather than absolute cost, and with no data on what patients paid out-of-pocket, it is hard to interpret this finding.

  • INITIATOR was a large-scale real-world observational longitudinal cohort study that extracted data from the large Optum and HealthCore databases. Patients included had type 2 diabetes and were on at least one oral diabetes drug and no injectable therapy at baseline, and subsequently initiated treatment with either Lantus or Victoza (both via pen).
  • We would be interested in knowing to what extent this sort of data holds sway with payers. The cost data, for example, may not be very interpretable from a patient perspective due to the lack of information on baseline costs and out-of-pocket costs, but a finding that patients that began Lantus did not see much of an increase in costs (as opposed to patients that began Victoza) would appear to be meaningful from a payer perspective.

Questions and Answers

Q: In your cost data, you looked at the change in cost from baseline rather than absolute cost. I’m wondering whether, given that insulin was started in patients that were likely more ill, whether this is partially a regression to the mean from being higher-cost at baseline.

A: I don’t know that I can provide a complete answer to that. [Dr. Levin displays supplementary slide] If you look at the summated total healthcare costs, not just focusing on diabetes, there was an increase of $400 with insulin glargine and $1,200 with liraglutide.

Q: How did you define persistence to therapy?

A: If patients did not refill their prescription within 90% of the time, that was considered failed persistence.

Q: Were both medications equally reimbursed for all patients in the cohort?

A: The costs presented here were costs to the healthcare plan, not costs to the patient. It’s hard to get data on the cost to the patient in these various plans. Some plans have different cost tiers, and some drugs have cards that might discount the cost.

Posters: Glucose Variability in Insulin Treatment

Least Glucose Variability is Observed With the Combination of a GLP-1 Agonist and Basal Insulin Among Four Commonly Used Insulin Regimens in type 2 Diabetes (VARIATION Study) (Poster 955)

H Bajaj, R Aronson, C Ye, K Venn, A Patrick

The VARIATION study was a prospective, cohort study designed to assess glucose variability and hypoglycemic episodes in patients on one of four commonly-used insulin regimens: 1. Basal insulin and oral hypoglycemic agents (OHA); 2. Basal insulin, GLP-1 agonist, and permitted OHAs; 3. Pre-mixed insulin and permitted OHAs; 4. Basal-bolus insulin and permitted OHAs. The cohort included both men and women with an average age of around 60-65 for each group and no significant difference in BMI or duration of the disease. All the patients had type 2 diabetes that was well controlled (HbA1c ≤ 7.5%) within the prior three months. When analyzing glucose variability, there were narrow ranges (1.8 mmol/L – 2.2 mmol/L) for all four groups. The basal insulin and GLP-1 regimen produced the lowest variability with the basal-bolus insulin regimen producing the highest variability. Interestingly, the basal-bolus insulin group also had the highest incidence of self-reported hypoglycemia (34 episodes) over the six day study period. This value was significantly higher than the other three groups: basal insulin (5 episodes), basal insulin and GLP-1 (3 episodes), and pre-mixed insulin (8 episodes). Overall, the data suggests there are improved benefits for using a basal insulin and GLP-1 regimen to minimize glucose variability and hypoglycemic episodes in patients with type 2 diabetes. This data was derived from an interim analysis of the first 80 patients, so additional investigations will need to be conducted to confirm the findings from this study. 

Symposium: Hot Topics in Diabetes

New Insulin Preparations

David Russell-Jones, MD (University of Surrey, Surrey, UK)

Dr. David Russell-Jones provided an enthusiastic overview of new insulin preparations, speaking particularly positively about Lilly’s novel basal insulin peglispro (BIL) and Novo Nordisk’s Tresiba (insulin degludec). He began his presentation by suggesting that the goal of insulin (which is, in essence, a replacement therapy) should be to mimic the PK/PD and physiology of endogenous insulin, which acts preferentially at the liver. This point set him up to speak about Lilly’s BIL, which has demonstrated hepato-selective activity. He highlighted that BIL appears to have a longer half-life than nearly any other insulin currently available (including Tresiba), and expressed enthusiasm about its superior A1c lowering, attractive weight profile, and reduction in nocturnal hypoglycemia vs. insulin glargine. Dr. Russell-Jones also spoke positively about Tresiba, highlighting what he sees as very concrete and fairly robust hypoglycemia benefits vs. insulin glargine as well as the identical efficacy achieved with flexible dosing. Pulling back from specific insulins, Dr. Russell-Jones ended his talk by expressing optimism that there are a number of companies working to develop better insulins, which is healthy for science and good for patients.

  • Dr. Russell-Jones also discussed other insulin candidates in development:
    • Sanofi’s Toujeo (insulin glargine U300): Dr. Russell-Jones covered results from EDITION I and II, demonstrating comparable A1c reductions with Lantus (insulin glargine U100) and benefits on nocturnal hypoglycemia at certain time points. During Q&A, he noted that the early hypoglycemia benefit seen in certain trials might be a potency issue.
    • Lilly/BI’s biosimilar insulin glargine formulation (LY2963016): Dr. Russell-Jones’ section on this candidate was brief. He mentioned that it behaves comparably to Lantus, and that the EMA recently granted it marketing approval.
    • Biodel’s BIOD-531: This candidate, Dr. Russell-Jones noted, benefits from both a quick onset of action and an extended 18-hour duration.
    • Novo Nordisk’s FIAsp (faster-acting insulin aspart): Dr. Russell-Jones expressed particular curiosity about whether this candidate will demonstrate meaningful benefits for patients on insulin pumps.
    • Adocia’s Biochaperone insulin lispro: Adding oligosaccharides to insulin lispro (Lilly’s Humalog) allows it to be absorbed more quickly.

Questions and Answers

Q: If I was back at the lab, designing an insulin, what proportion of peripheral actions vs. hepatic selectivity do I want?

A: Insulin levels in the portal vein are about five times higher than they are elsewhere. Even with hepatopreferential analogs, I don’t think we’re getting anywhere near mimicking that. People get worried about this, but I don’t think we’re likely to mimic normal physiology. Getting the balance right is very important.

Q: There was no word in your talk about price, and price is very important in many countries, especially in the developing world.

A: My view is that price is price – we are doctors and we are our patients’ advocates. I personally try to get the best for our patients. Price aside, what I’m interested in is physiology and advancing new therapies, which is why I have not mentioned price.

Q: Could you comment on the temporal incidence of hypoglycemia in the glargine U300 trials? In some trials it appears that there is a benefit in the first few weeks of titration.

A: It appears from the presentation given by Philip Home earlier in the week that there did seem to be an effect in the early stages. That is during the titration phase, so might it be a potency issue? I don’t know. When all the trials are out we can do a proper meta-analysis and see. It appears that the principal advantage is at night, which is presumably because of the differences in the pharmacokinetic profiles.

Q: You did not mention the recently approved Technosphere insulin [MannKind’s Afrezza], which has amazingly different kinetics than any insulin that was covered today. Is there a reason you ignored it?

A: No reason other than time and expertise on my behalf.

Corporate Symposium: New Insulin Therapies on the Horizon (Sponsored by Lilly)

Development of New Insulin: Why Do We Need Them, How Different Are They and How Do We Prove Their Value in Diabetes Care?

Thomas Danne, MD (Diabetes Center for Children and Adolescents, Hannover, Germany)

Dr. Thomas Danne discussed new basal insulin analogs on the horizon, including Novo Nordisk’s insulin degludec, Lilly’s PEGylated insulin lispro, and Sanofi U300 glargine. He explained that the currently available basal insulins do not optimally mimic endogenous insulin secretion, and the newer analogs can improve on current options. Toward the end of his talk, Dr. Danne pointed out that biosimilars may enter clinical practice soon – he called for careful testing of these molecules.

  • Introducing Novo Nordisk’s degludec as a second-generation acylated insulin, Dr. Danne noted that it has a flatter pharmacokinetic profile (compared to first-generation analogs) that exceeds 24 hours. The analog can be administered once-daily at any time of the day, which may also improve adherence.
  • Regarding Lilly’s PEGylated insulin lispro, Dr. Danne explained how its larger hydrodynamic size leads to a more hepatopreferential mode of action, which has the potential to restore a more physiological insulin gradient. This reportedly can lead to improved glycemic control, lipoprotein composition, and hepatic IGF-1 and IGFBP-3 generation in those with type 1 diabetes. This makes it a potentially favorable option for those who are vulnerable to hypoglycemia or require high daily insulin doses.

Panel Discussion

Q: Is it mainly the size of the insulin that influences its hepatoselectivity or are there other factors as well?

Dr. David Russell-Jones (University of Surrey, Surrey, UK): There are other factors as well, such as the affinity to albumin. Detemir is slightly hepatopreferential. But when it binds to albumin, the insulin is not available for the receptor and its affinity for albumin comes into play. It comes off and is hepatopreferntial. Degludec’s affinity for albumin is much higher. I would suspect it’s not hepatopreferential.

Dr. Thomas Danne (Diabetes Center for Children and Adolescents, Hannover, Germany): I absolutely agree that there are other factors. I think the weight loss effects must also be affected by something. Detemir has that weight loss effect. And it’s greater with BIL from what we know now. We have to see how it’s going to play out. Degludec, so far, is not showing any weight loss.

Q: We’ve been talking about expanding the potential benefits of hepatoselectivity. It seems right to pursue the physiological argument but does it actually matter in clinical practice as long as A1c improves?

Dr. Russell-Jones: You’re restoring normal physiology. Nature never does anything without a reason, so it’s not an accident - there are advantages. If insulin is delivered through the portal route, you expect hypoglycemia should be easier to overcome because the counter regulatory hormones have an effect on the liver. You would expect fewer hypoglycemic episodes and less weight gain. And another area that hasn’t been discussed but should be is that patients with type 1 diabetes hypersecrete growth hormone, and it’s been hypothesized that it’s partly involved in microvascular complications. It’s thought to be due to a lack of portal insulin, which leads to not enough IGF-1 being produced, which leads to growth hormone hypersecretion. It’s especially apparent in kids. If you develop an insulin that restores all that, you could argue that this insulin might have an effect on the development of microvascular complications. That’s my personal opinion, but I would love to do studies.

Q: Just on the lack of protraction of hypoglycemia dated with CGM in the BIL Type 2 Study, does this suggest that hepatopreferential action does not adversely affect hypoglycemic recovery?

Dr. Danne: I’m thinking it is too early to judge that. Hepatoselectivity generally restores many pathways, and these could potentially be improved. I still know of times when we’ve had children where we’ve seen glucagon responses. Basically, neomorphological insulin should benefit you in many ways, like in hypoglycemic recovery.

Q: Beyond larger molecular size, are there other ways being explored to get more insulin into the liver?

Dr. Russell-Jones: Delivery is one area. In the past, there were pumps that did it but they’ve got problems. Delivery of pump insulin through the umbilicus may create hepatopreferential effects. Another one is oral insulin; lots of companies are trying to do that, and that will of course go into the portal system. There are multiple ways to do it, and lots of excitement.

Corporate Symposium: Looking Longer Term – New Options for Insulin Therapy (Sponsored by Lilly/Boehringer Ingelheim)

Emerging Options in Insulin Therapy – Development of Biosimilars

Melanie Davies, MD (University of Leicester, Leicester, UK)

Dr. Melanie Davies shared her perspective on the hot topic of biosimilar insulins, a subject that she believes can potentially be very daunting for many clinicians. She reviewed the obstacles posed by the complex manufacturing process and uncertain regulatory requirements for biosimilars, using Lilly/BI’s data package for its biosimilar insulin glargine (which recently became the first biosimilar insulin approved in Europe) as a case study to demonstrate the level of evidence required to assure regulatory authorities that a biosimilar insulin is clinically similar to the reference product. Key challenges she identified include robustly demonstrating similarity in clinical trials, overcoming the less-than-stellar reputation of biosimilars in some regions due to previous negative experiences, and clarifying the regulatory criteria for interchangeability. Dr. Davies also cautioned that although the introduction of biosimilars should lead to some cost reductions, the savings will likely be lower than anticipated and certainly not as dramatic as the discounts seen with generic small molecule drugs. Despite these challenges, Dr. Davies believes that biosimilar insulins will ultimately be very beneficial for patients, as they will help address the “genuine need” to increase access to a wider range of treatment options. For more background on the manufacturing and regulatory challenges surrounding biosimilars, see our coverage of Professor Philip Home’s (Newcastle University, Newcastle upon Tyne, UK) presentation on the topic at this year’s Keystone conference.

Questions and Answers

Q: Why does one develop biosimilars? One expectation is to save money, but that is assuming that the cost of a biosimilar would be less than the cost of the approved molecule. You’ve shown how tightly regulated the process is and how much investment of time and money is necessary. What is driving this rush to biosimilars?

A: I don’t know, I’m sure someone from the company would have a better perspective on that. Clinically, it’s not just about the actual product, but with a new product comes a device and new support programs as well. There is a genuine need to increase access to good therapies. Glargine is a good therapy, so increasing access is a good thing. This is uncharted territory in terms of cost. (Editor’s note – we think the assumption that biosimilars will be at least slightly less expensive than branded insulin is a pretty safe one to make.)

Closing Comments

Ele Ferrannini, MD, PhD (University of Pisa School of Medicine, Pisa, Italy)

Dr. Ele Ferrannini called on the diabetes community to raise its standards for success, stressing the need for providers to overcome clinical inertia and accept the fact that more conservative treatment options like metformin monotherapy will eventually fail for the vast majority of patients. He noted that with other chronic conditions like hypertension and dyslipidemia, most patients successfully reach targets well within the normal range with available medications, but the majority of patients with diabetes fail to reach A1c targets that are significantly above the expected range for people without diabetes. He suggested that this gap may be responsible for the field’s inability thus far to demonstrate a link between improved glycemic control and reduced cardiovascular risk, and he argued that the current trend toward individualized therapy ought to include more stringent targets (with composite endpoints including parameters like weight and hypoglycemia) for some populations in addition to more relaxed targets for others. From a patient perspective, we would ask whether more relaxed targets are needed or whether different treatment paradigms for those experiencing hypoglycemia might be needed.

Corporate Symposium: Building a Brighter Future – Addressing Challenges in Patient Care (Sponsored by Novo Nordisk)

Building Together: Insulin and Incretin Combination Therapy

Stephen Bain (Diabetes Research Network, Wales, UK) and Sultan Linjawi (Coffs Endocrine & Diabetes Services, Coffs Harbour, Australia)

Drs. Stephen Bain and Sultan Linjawi discussed the potential of insulin-incretin combination therapies to improve glycemic control without compromising safety. The highlight of the presentation was our first look at new data from BEGIN, Novo Nordisk’s phase III program for Tresiba (insulin degludec). This 26-week study compared the effect of insulin degludec in combination with liraglutide and metformin in insulin-naïve type 2 patients requiring treatment intensification (also presented in OR-145). Results indicated that from a baseline A1c of ~7.5%, patients in the intervention cohort (insulin degludec + liraglutide + metformin; n=174) experienced significantly greater reductions in A1c (-0.9%, p<0.001) relative to those who received liraglutide and metformin alone (n=172). Low rates of overall hypoglycemia were seen in both groups (though significantly higher in the intervention group, p=0.0002). Ultimately, Dr. Bain shared optimism that the insulin-incretin combination represents a more efficacious and comparably safe intensification of treatment relative to GLP-1 receptor agonist therapy alone. This suggestion was borne out by a subsequent summary of efficacy and safety data from the DUAL I and DUAL II trials that compared insulin degludec alone with a fixed-ratio combination of degludec and liraglutide (IDegLira). As we saw at ADA 2014, one-year findings from DUAL I showed that the substantial A1c reductions, weight benefit, and hypoglycemia benefit with IDegLira after 26 weeks were preserved through to 52 weeks; DUAL II confirmed these results. Given this strong efficacy and safety profile, it is no wonder Drs. Bain and Linjawi were so positive on combination therapy.

  • The BEGIN trial investigated the efficacy and safety of insulin degludec in combination with liraglutide and metformin. The study included patients with type 2 diabetes (≥ 6 months) who were on metformin monotherapy with A1c 7.5-10.0% or on metformin and a second agent (sulfonylureas, DPP-4 inhibitors, glinides) with A1c 7.0-9.0%. If eligible, patients entered a 15-week run-in period during which the second agent was stopped (metformin was continued) and liraglutide was titrated up to 1.8 mg. Patients who still required treatment intensification at the end of 15 weeks (A1c ≥ 7.0%) were randomized to receive either degludec or placebo for 26 weeks at a starting dose of 10 units titrated to a target of 71-90 mg/dl.
  • Results indicated a significant difference in efficacy between the intervention and control cohorts at 26 weeks. Patients in the insulin-incretin therapy group achieved an average A1c reduction of -1% (baseline 7.5%) vs. -0.1% in the incretin-alone group (baseline: 7.6%; p<0.001). Notably, the intervention group experienced an immediate and linear reduction in A1c for ~14 weeks before stabilizing at 6.5%; those in the control group experienced a more slight reduction (~-0.2%) that also persisted for ~14 weeks before regressing to 7.5%. Notably, patients in the intervention cohort also achieved lower fasting plasma glucose (-46 mg/dl; p<0.0001).
    • Though rates were low, there were significantly more episodes of hypoglycemia in the intervention group relative to the control group (0.57 vs. 0.12 episodes per patient, p=0.0002). Dr. Bain acknowledged that this was not unexpected given the mechanism of the drugs. Notably, he highlighted that the disparity was confined to daytime, as nocturnal rates of hypoglycemia were similar in both groups (0.05 vs. 0.03 episodes per patient) – however, we would note that the small number of events (2 vs. 3 events) limits the power in interpreting this data. Last, insulin-incretin therapy led to a slight increase in body weight (2.0 kg), while placebo therapy led to a slight decrease in body weight (-1.3 kg).

-- 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