American Diabetes Association 73rd Scientific Sessions

June 21-25, 2013, Chicago IL Report – Incretins – Draft

Executive Highlights

Incretin-based therapies remained a rich area of exploration in 2013, and this year the potential association between pancreatitis and incretins was a prevalent theme. Several talks during industry- sponsored sessions on GLP-1 agonists or DPP-4 inhibitors addressed the topic, reinforcing the opinion that current clinical data are insufficient to suggest taking patients off of these drugs. Most strikingly, though, Dr. Vanita Aroda (MedStar Health Research Institute, Hyattsville, MD) spoke at an independent session where, during Q&A an audience member asked if she herself would take her patients off of incretin-based therapies at this point. She turned the question around, and a poll of the very-packed audience found that no one would do this.

The most exciting data for GLP-1 agonists were for their potential in type 1 diabetes (liraglutide [1007- P]) and in combination with basal insulin (IDegLira [65-OR]). The poster on liraglutide for type 1 diabetes suggested that it could considerably reduce insulin requirement for people with type 1 diabetes (by 24-27%). The results for IDegLira (insulin degludec and liraglutide) were some of the most compelling we saw at the conference: Dr. John Buse presented data from IDegLira’s DUAL-1 (65-OR), during which over 80% of patients taking IDegLira achieved the A1c goal of 7.0% or below. IDegLira was better than degludec alone in terms of hypoglycemia (one third less than with degludec) and weight (slight weight loss, instead of weight gain with Tresiba). While Sanofi was supposed to share more information on the functionality of its Lyxumia/Lantus fixed-ratio combination device (and the problems that it encountered with the fix-flex device), the company disclosed disappointingly little during the conference call it held during ADA.

There was not much in the way of groundbreaking data for GLP-1 agonists as monotherapy for type 2 diabetes this year – highlights in this area included the presentation of detailed results for dulaglutide’s AWARD 1, 3, and 5, which expanded on the topline data for these trials disclosed in October 2012 (66- OR, 69-OR, 71-OR, 1004-P). The top dose of dulaglutide provided a ~1% additional A1c reduction over placebo, an additional 0.7% reduction over sitagliptin, and a 0.2% additional reduction (statistically significant) over metformin. Dulaglutide data seemed strong overall, with adverse event profiles similar to those of other long-acting GLP-1 agonists. Detailed results of GSK’s albiglutide’s HARMONY 8 were presented (68-OR) during which albiglutide provided 0.3% greater A1c reduction than sitagliptin in people with renal impairment. Also notably, a phase 1 study of the safety and efficacy of the oral GLP-1 agonist TTP054 (115-OR), found consistently greater declines in fasting plasma glucose and two-hour postprandial glucose versus placebo-treated patients, with no hypoglycemic events and with a similar rate of GI adverse events to the placebo arm. No other oral GLP-1 agonist has reported results yet.

DPP-4 inhibitors received the spotlight on Sunday morning in an oral session that included a range of topics. Pooled data from three phase 3 studies indicated that linagliptin (BI/Lilly’s Tradjenta) has a favorable effect on the composite primary endpoint of cardiovascular (CV) death, non-fatal stroke, and non-fatal myocardial infarction (HR: 0.78; 95% CI: 0.55-1.12; 376-OR). A study of linagliptin in type 2 diabetes patients on hemodialysis included the use of CGM and found that switching patients from insulin to linagliptin led to marked decreases in mean amplitude of glucose excursions (MAGE) while on hemodialysis (374-OR). On the preclinical front, Dr. Marco Bugliani (University of Pisa, Pisa, Italy) presented in vitro data showing that the DPP-4 enzyme is present in human pancreatic islet cells (primarily alpha cells) and that DPP-4 inhibitors such as Merck’s MK-0626 could have direct protective effects against glucotoxicity and lipotoxicity (378-OR). A notable poster by Dr. John Buse’s (University of North Carolina, Chapel Hill, NC) group investigated the question of DPP-4 inhibitors and pancreatic cancer and found no increased risk of pancreatic cancer or any cancer with DPP-4 inhibitors compared to SFUs or TZDS (111-LB).

With SAVOR-TIMI announcing topline results immediately prior to the start of ADA 2013 (BMS/AZ’s Onglyza revealed that the drug was non-inferior, but not superior, to placebo for reducing CV events), we would have expected to hear more discussion on DPP-4 inhibitors and cardiovascular effects. However, perhaps since SAVOR-TIMI did not achieve its ambitious endpoint or perhaps since most CVOTs are a few years away from reporting results, CVOTs were not a hot topic of discussion.

With regard to the future of DPP-4 inhibitors, we’re interested in the optimal approach for combining DPP-4 inhibitors with other therapies – we’re already seeing movement on this front with the BMS/AZ Alliance, the Lilly/BI partnership, and the Merck/Pfizer partnership (the latter which will aim to develop an ertugliflozin/sitagliptin fixed-dose combination), and likely with yet still others, and we’re curious about the possibility of combining a DPP-4 inhibitor with a GPR40 agonist such as Takeda’s TAK-875. (Takeda recently reported topline phase 3 data for the candidate – read our report at Lastly, we’ll be keeping a close eye on whether and how the prescribing patterns for DPP-4 inhibitors change as SGLT-2 inhibitors enter the market – while SGLT-2 inhibitors come with the added side effect of genitourinary infections, KOLs have yet to reach a consensus on whether this issue will be a significant barrier for patients; for now, it certainly seems to be considered a more manageable side effect than nausea, weight gain, and other side effects associated with other drugs.

Table of Contents 


Oral Sessions: GLP-1 Agonists in Practice


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

Dr. John Buse presented an excellent summary of the phase 3 DUAL-1 trial, which compared a fixed dose combination of insulin degludec and liraglutide (known as IDegLira) with insulin degludec (Tresiba) and liraglutide (Victoza) alone (all from Novo Nordisk). The concept of IDegLira is to combine the benefits of both therapies, with each agent mitigating the side effects of the other. The trial showed this approach was valid, yielding excellent results. In fact, the (nearly a third) lower hypoglycemia of the combination allowed participants to achieve 0.5% lower A1c levels with a lower total daily dose of insulin. GI side effects of the combination were also lower than with liraglutide alone and the combination yielded a small but sustained weight loss (of,0.5kg [1.1 lbs]) compared to a weight gain with insulin. The trial studied over 1,600 participants for a 26 week period titrated to target in the case of the IDegLira and insulin degludec arms, and titrated in the usual ramp-up fashion for liraglutide. After 26 weeks, the IDegLira group was taking 38U/day of insulin, compared to 53U/day in the degludec group. However, A1c reduction from a baseline of ~8.3% was 1.9% for IDegLira, compared to a reduction of 1.4% with degludec and 1.3% with liraglutide. The IDegLira group achieved and maintained a very low A1c of 6.4%. Even better was the result that 81% of the IDegLira group reached the target of a7.0% A1c (compared to 65% of the degludec group). IDegLira used a ratio of 50U of insulin degludec to 1.8 mg of liraglutide. Dr. Buse concluded that IDegLira resulted in better glycemic control with a lower risk of hypoglycemia and weight gain than either of its components alone.

  • The combination of a basal insulin and a GLP-1 agonist has the potential to combine their benefits while also mitigating the adverse effects of the two agents used individually. The post-prandial effects of GLP-1 can be combined with the fasting glycemic control of basal insulin. The weight loss of GLP-1 can mitigate any weight gain with insulin. The observed low hypoglycemia with GLP-1 can mitigate the potential for hypoglycemia with insulin and therefore allow patients to achieve lower A1c levels.
  • The open label DUAL-1 trial studied 1,663 people with type 2 diabetes randomized to IDegLira (Novo Nordisk), insulin degludec (Tresiba, Novo Nordisk) and liraglutide (Victoza, Novo Nordisk). IDegLira is Novo Nordisk’s once-daily injectable combination of insulin degludec and liraglutide. The IDegLira used in this study was fixed at a ratio of 50U degludec to 1.8 mg of liraglutide. Participants at baseline were inadequately controlled on metformin with or without pioglitazone (Takeda’s Actos). Liraglutide was titrated up from an initial 0.6 mg dose to a maximum dose of 1.8 mg as per the label, although the average dose at the end of the trial was 1.4 mg. IDegLira and degludec were titrated twice a week in two unit steps to a target fasting plasma glucose of 72-90 mg/dl (4.0-5.0 mmol/l). At baseline, BMI was roughly 31 kg/m2 and duration of diabetes was seven years. Eighty-three percent of the participants were taking metformin and 17% were taking metformin plus pioglitazone.
  • After 26 weeks, A1c reduction was 1.9% for IDegLira, 1.4% for degludec, and 1.3% for liraglutide from a baseline of 8.3%. Remarkably, IDegLira participants had an average A1c of 6.4%. In the IDegLira group, 81% of participants reached a target A1c of ≤7%, and 70% reached a target of ≤6.5%. While IDegLira and degludec provided an equivalent reduction in fasting plasma glucose (65 mg/dl) at 26 weeks, those in the IDegLira group were taking much less insulin (38 units/day, vs. 53 units/day for degludec). Nine point blood glucose profiles established that the post-prandial control of IDegLira was similar to that of liraglutide alone.
  • The rate of hypoglycemia in the IDegLira group was only 68% of the rate in the degludec group. Rates in the IDegLira group were around two events per patient year of exposure. The rates of hypoglycemia in the liraglutide group were ten times lower (around 0.2 events per patient year). Since hypoglycemia is usually the limiting factor in treating to target, Dr. Buse presented a model of the relationship between A1c and hypoglycemia rate for IDegLira and degludec, which supported the argument that a patient taking IDegLira can achieve a lower A1c at the same risk of hypoglycemia, (or that patients with the same A1c have a lower risk of hypoglycemia with IDegLira).
  • Participants taking liraglutide exhibited a weight loss of 6.6 lb (3 kg) after 26 weeks, compared to those taking degludec, who gained 1.5 kg (3.3 lb). Those taking IDegLira exhibited a slight weight loss of 1.1 lb (0.5 kg). This small weight loss was sustained to the end of the trial (and even out to 52 weeks in the DUAL-1 extension, according to Novo Nordisk).
  • Nausea for IDegLira was much better than liraglutide alone. For the first four weeks, nausea peaked at over 10% of subjects in the liraglutide group, but resolved to a sustained level of 3% after 13 weeks. In comparison, nausea for IDegLira declined steadily from roughly 3% of participants initially to 1% by the end of the trial. Dr. Buse commented that this reduction is probably related more to the titration program and the lower starting dose than to the total amount of exposure to liraglutide.
  • Compared to insulin degludec, IDegLira exhibited reduced hypoglycemia, leading to a lower A1c with better post-prandial control and no weight gain. Compared to liraglutide, IDegLira had a better reduction in A1c and fasting glucose, with a reduced amount of nausea.

Questions and Answers

Q [Dr. Julio Rosenstock]: The trial would have been nicer if it had been blinded…

A: Later this year, we plan to present the results from the DUAL-2 study, which is blinded. (Note that Novo Nordisk has already announced some top level results for DUAL-2, which compares IDegLira and degludec. In DUAL-2, IDegLira also achieved a 1.9% A1c reduction over baseline, which was reported to be statistically significantly better than degludec (A1c reduction not yet disclosed)..

Q: Do you have any data regarding nocturnal hypoglycemia?

A: I can’t remember it all off the top of my head, but it’s more or less equal across degludec and IDegLira.

Q: What was the washout protocol?

A: I don’t know what you are referring to. There was no washout period. I suspect that you were referring to the concern raised at the FDA Advisory panel on degludec on CVD events during the washout period. There was one adjudicated cardiovascular event in each of the three arms, but no issues with washout protocol.

Q: Can we combine U500 insulin with GLP-1?

A: I think that Wendy Lane from Asheville NC will be presenting on the combination with severely insulin resistant patients.

Q: There is an impressive reduction in hypoglycemia – what is the reason?

A: It’s related perhaps to the lower amount of insulin, but I remember the GWCO trial adding exenatide to glargine where we saw a similar thing. I suspect that it is also related to an improvement in alpha cell function.


Carol Wysham, MD (University of Washington, Spokane, WA)

Dr. Carol Wysham presented the results of the phase 3 AWARD-1 study, which compared dulaglutide at two doses (1.5 mg and 0.7 5mg) to exenatide twice daily (BMS/AZ’s Byetta) and placebo on a background of maximal doses of metformin and pioglitazone (Takeda’s Actos). After 26 weeks, dulaglutide 1.5 mg provided an A1c reduction of 1.5% from baseline (compared to a reduction of 1.0% with exenatide), which was roughly 1% lower than the placebo group. Nearly 80% of the patients in the high dose dulaglutide group reached an A1c target of 7%, compared to 54% in the exenatide group and 45% in the control group. Weight loss for dulaglutide 1.5 mg was similar to exenatide, while dulaglutide 0.75 mg was weight neutral.

  • Dulaglutide (Lilly) is a once-weekly GLP-1 agonist that links recombinant human GLP-1 to a human IgG4 Fc fragment in order to obtain longer plasma half-life (~5 days) in an injectable solution. This study, AWARD-1, is part of a series of five phase 3 trials, leading up to Lilly’s FDA submission, scheduled for 2013. Lilly has previously announced top-line results for the AWARD trials that are covered in this session (see our report at
  • The primary objective of the AWARD-1 trial is to prove superiority of once-weekly dulaglutide versus placebo in people with type 2 diabetes treated with metformin and pioglitazone (Takeda’s Actos). Metformin and pioglitazone were optimized to maximum doses before the start of the trial, and dulaglutide was investigated at two doses – 1.5 mg and 0.75 mg. The AWARD-1 trial also includes twice daily exenatide (BMS/AZ’s Byetta) as a comparator. The trial randomized 978 type 2 diabetes patients to dulaglutide 0.75 mg or 1.5 mg for 52 weeks, to exenatide for 52 weeks, or to placebo for 26 weeks followed by dulaglutide (0.75 mg or 1.5 mg) for 26 weeks. A1c at baseline was roughly 8.1% and participants had an average BMI of 33 kg/m2.
  • Dulaglutide 1.5 mg achieved an A1c reduction of 1.05% compared to placebo, demonstrating superiority. A1c reduction from baseline was 1.5% for dulaglutide 1.5 mg, 1.3% for dulaglutide 0.75 mg, 1.0% for exenatide and 0.4% for the placebo. In addition, 79% of participants taking dulaglutide 1.5 mg reached a target A1c of ≤7%, compared with 67% for the0.75 mg dose and 54% for exenatide. Both doses of dulaglutide were better at lowering fasting serum glucose than either placebo or exenatide.
  • Participants in the dulaglutide 1.5 mg group achieved a weight loss of 3.3 lb (1.5 kg) after 26 weeks, which was similar to exenatide. The weight loss was achieved after four weeks and was then maintained throughout the study. Dulaglutide 0.75 mg was weight neutral and the placebo group gained 3.1 lb (1.4 kg).
  • The majority of adverse events were GI related but reduced over time for dulaglutide. Dulaglutide 1.5 mg had the most nausea, at 28% of participants, dulaglutide 0.75 mg had a nausea rate of 16%, and exenatide was in the middle at 26%. There was one case of pancreatitis and one case of pancreatic cancer in the pooled dulaglutide group (five months of exposure).
  • The A1c reductions in all three GLP-1 groups were sustained out to 52 weeks. Questions and AnswersQ: Longer acting GLP-1 agonists tend to be associated with lower GI side effects, but that’snot the case here. Can you explain?

A: The answer is probably in the chemistry of the compound. We see a reduction in glucose earlier, this shows earlier efficacy, maybe just after first injected.


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

Dr. Guillermo Umpierrez treated us to the full results of the AWARD-3 trial, part of the phase 3 package for dulaglutide (Lilly) that was presented at this session. AWARD-3 compares dulaglutide to metformin in monotherapy, for patients with an average duration of 2.6 years of diabetes, and an A1c of 7.6%. Results indicate that dulaglutide has a superior A1c reduction at 26 weeks (around 0.2% better than metformin, but statistically significant), together with a higher percentage of patients reaching target. The dulaglutide 1.5mg dose had an identical effect on weight as metformin at 26 weeks, although the data suggested that it weight was trending less favorably than metformin at 52 weeks. Adverse events were comparable (mainly GI in nature), so Dr. Umpierrez was able to conclude that dulaglutide was superior to metformin in monotherapy. However, in the Q&A Dr. Rosenstock (chairing the session) admitted that while dulaglutide was statistically superior, he didn’t think it was clinically superior. Dr. Umpierrez countered that since a significantly greater percentage of people reach target, then by definition it has to be clinically superior.

  • This study, AWARD-3, is part of a series of five phase 3 trials, leading up to Lilly’s FDA submission for dulaglutide (a once-weekly injectable GLP-1 agonist) scheduled for 2013. Lilly has previously announced top-line results for the AWARD trials that are covered in this session (see our report at
  • This double blind randomized controlled study is designed to compare dulaglutide at two doses (1.5 mg and 0.75 mg once weekly) with metformin in monotherapy over 52 weeks. Participants were randomized to dulaglutide plus oral placebo (two doses) or metformin (1,500 or 2,000 mg/day according to tolerability) plus injectable placebo. The 807 treatment naïve participants were randomized, and at baseline, A1c was low at 7.6%, BMI was 33 kg/m2, average age was 56 years, and duration of diabetes a relatively short 2.6 years, since it was an early trial.
  • At 26 weeks, patients taking dulaglutide 1.5 mg experienced a mea 0.8% A1c reduction from baseline, which was 0.22% better than metformin (p<0.025). Dulaglutide was also superior at 52 weeks vs. metformin (A1c reduction for dulaglutide 1.5 mg was 0.7% lower than baseline and 0.17% lower than metformin). However, the lower dose dulaglutide was non-inferior, but not superior to metformin at 52 weeks. For the high dosedulaglutide, A1c showed a rapid reduction over the first 13 weeks, with a slight rise from week 26 to week 52. At 26 weeks, 62% of patients achieved a target of ≤7% A1c in the dulaglutide 1.5 mg arm, compared with 54% of metformin patients.
  • At 26 weeks, both the metformin and dulaglutide 1.5 mg arms showed the same weight loss, around ~4.4 lb (2 kg). By 52 weeks, dulaglutide patients appeared to be gaining weight compared to metformin patients, but the difference was non-significant. Overall adverse events were similar in all three arms – GI side effects were the largest component.

Questions and Answers

Q [Dr. Julio Rosenstock]: Dulaglutide maybe be statistically superior, but maybe not clinically superior – I would conclude that dulaglutide is effective, but you would never use it in monotherapy instead of metformin.

A: The number of people achieving A1c targets was significantly greater than metformin so it is clinically better. On A1c, dulaglutide is up to 0.4% better, and we think that this is clinically significant.

Q: Do you have any blood pressure data?

A: I didn’t show blood pressure data but there is 2mmHg difference in favor of dulaglutide - but it’s not significant.

Q: Did you ever consider the cost efficacy of this drug? Because metformin is very cheap…

A [Dr. Rosenstock]: Obviously!


Michael Nauck, MD, PhD (Diabeteszentrum Bad Lauterberg, Harz, Germany)

Dr. Michael Nauck presented the results of Lilly’s AWARD-5 study, which compared dulaglutide 1.5 mg and 0.75 mg to sitagliptin and placebo in 1,098 type 2 patients. At 52 weeks, both dulaglutide doses provided superior A1c reductions compared to sitagliptin (1.11%, 0.86%, and 0.39%, respectively), as well as larger improvements in body weight (-3.22 kg [-7.1 lbs], -2.7 kg [-6.0 lbs], and -1.63 kg [-3.6], respectively). Furthermore, a greater proportion of people in the dulaglutide 1.5 mg and 0.75 mg arms achieved the A1c goal of <7% (58% and 49%, respectively, vs. 33% for sitagliptin), as well as the A1c goal of 6.5% (42% and 29%, respectively, vs. 19% for sitagliptin). While dulaglutide was associated with a higher rate of GI side effects, Dr. Nauck noted that the rates fell within the range typically observed for long-acting GLP-1 agonists.

  • The doubled blind RCT enrolled type 2 patients who had been on either monotherapy (metformin or an oral anti-diabetic medication [OAM]) or combination therapy (metformin plus one OAM) with an A1c between 7.0% and 9.5%. Participants were first placed on metformin during the 11-week run-in period and then randomized to multiple doses of dulaglutide (ranging from 0.25 mg to 3.0 mg) for 13 weeks. Dr. Nauck explained that this phase mimicked a dose-ranging phase 2 trial and allowed Lilly to select the optimal doses for the subsequent phase 3 trial (AWARD-5). Patients not assigned to the selected doses were exclude from the phase 3 study. Lilly chose the 0.75 mg and 1.5 mg doses using a Clinical Utility Index, which takes into account efficacy and safety measures, with a focus on blood pressure and pulse rate effects – Dr. Nauck remarked that the full details of this process are being presented in poster 1045-P.
  • The phase 3 trial compared dulaglutide 0.75 mg and 1.5 mg (n=302 and 304, respectively) to sitagliptin (100 mg; n=315) and placebo (n=177). Patients randomized to placebo took the placebo capsule for 26 weeks (“for ethical reasons”) before switching to sitagliptin. At baseline, the participants had an average age of 55 years, weight of 86-87 kg (190- 192 lbs), and A1c of 8.1-8.2%. Dr. Nauck noted that while Lilly obtained both 52-week and 104- week data, his presentation would focus on the former. The 104-week data is detailed in poster 1004-P.
  • At 26 weeks, greater improvements in A1c were observed with dulaglutide 1.5 mg (- 1.22%) and 0.75 mg (-1.01%) compared to sitagliptin (-0.61%) and placebo (+0.04%). This pattern was also observed at 52 weeks, with larger A1c reductions observed for dulaglutide1.5 mg (-1.11%) and 0.75 mg (-0.86%) vs. sitagliptin (-0.39%; p<0.001 for both comparisons). At52 weeks, the three treatment groups had ending mean A1c values of 6.83%, 7.08%, and 7.57%, respectively. Both dulaglutide doses also provided greater reductions in fasting plasma glucose (- 43 mg/dl for 1.5 mg and -29 mg/dl for 0.75 mg) compared to sitagliptin (-16 mg/dl).
  • Not surprisingly, greater weight loss was observed with dulaglutide 1.5 mg (-3.22 kg [-7.1 lbs]) and dulaglutide 0.75 mg (-2.7 kg [-6.0 lbs]) compared to sitagliptin (-1.63 kg [-3.6]).
  • On the safety front, dulaglutide 1.5 mg and 0.75 mg were associated with slightly higher rates of adverse events (77% for both groups) compared to sitagliptin (70%). Dr. Nauck commented that while dulaglutide was also associated with higher rates of GI side effects, the numbers fall within the range typically observed with long-acting GLP-1 agonists. All three groups experienced a similarly low rate of injection site reactions (1%), as well as comparable rates of hypoglycemia (1.6% for dulaglutide 1.5 mg; 2.6% for dulaglutide 0.75 mg, and 1.1% for sitagliptin). No cases of severe hypoglycemia were reported in the study.
  • The trial also tracked cases of pancreatitis and pancreatic cancer over 104 weeks. The rate of pancreatitis was zero cases for the dulaglutide arms and 3 cases/1,000 patient-years in the sitagliptin arm. No cases of pancreatic cancer were reported.

Questions and Answers

Q: Can you tell us something about the biodistribution of the compound? It might be more concentrated in certain compartments such as the lymph nodes. We know that GLP-1 can act on T cells and B cells.

A: I can’t give you a detailed account of how this molecule is distributed in circulation. I can tell you that the half life is 5 days, so it’s in the circulation for quite a period of time. It’s thought to be degraded within the circulation. So I think the question you ask is very justified and it would be interesting to know such details – i.e., what is the concentration of dulaglutide in lymph vessels. I think we need to ask the company to do such details.

Q: Is it degraded in kidney?

A: It is not degraded in the kidney. All the large, once-weekly agonists are somewhat coupled to a big protein. The purpose is to circumvent the fact that the original peptide related to GLP-1 is cleared in the kidney. This is also another compound that could be tested in renally-impaired patients.

Q: Regarding the adaptive design, it was very interesting, very nice. I’m guessing that because of the short timeframe, the reason for choosing the doses of the drug was the side effect profile. You have 17% nausea, 13% vomiting. What happened when you tested the higher doses?

A: I think we should go to poster 1045-P for these details. The major reason for not selecting the 3 mg dose was cardiovascular parameters – e.g., pulse rate and blood pressure. It wasn’t because of GI side effects.

Q: Was there a change in pulse rate in this trial?

A: We saw minor changes in pulse rate that are typical for most long-acting GLP-1 agonists.


Michael Trautman, MD (Lilly, Indianapolis, IN)

Dr. Michael Trautman presented three-year results from the DURATION 3 study, which compared the safety and efficacy of exenatide once weekly and insulin glargine in people with type 2 diabetes. Overall, treatment with exenatide once weekly was associated with greater reductions in A1c (7.3% vs. 7.5%; p=0.033) and weight (-2.49 kg vs. +2.0 kg; p<0.001) than with insulin glargine in the ITT groups. However, insulin glargine provided greater improvements in fasting glucose than exenatide once weekly (-47.4 mg/dl vs. -31.2 mg/dl; p<0.001). Expectedly, incidence of minor hypoglycemia was significantly lower with exenatide once weekly compared to insulin glargine throughout the three-year period. No new safety signals were detected, with one incidence of pancreatitis in both treatment groups. Finally, exenatide antibodies were found to decrease over time, with 80% of individuals anti- body negative at three years. There was no association between exenatide anti-body status and degree of clinical effect.

  • DURATION 3 was a three-year open label, randomized controlled study that compared the safety and efficacy of exenatide once weekly with titrated insulin glargine. Insulin glargine was continuously titrated following a treat to target algorithm. Exenatide once weekly was administered as a fixed 2.0 mg dose. The study randomized 223 individuals to receive insulin glargine and 233 individuals to receive exenatide once weekly. At three years, 140 individuals remained in the exenatide once weekly arm and 147 individuals remained in the insulin glargine arm. Both arms were balanced at baseline, with an average A1c of 8.3%, body weight of 90 kg, and average duration of diabetes of eight years. 70% of patients in both arms were inadequately controlled on metformin alone, whereas the other 30% were inadequately controlled on metformin and a sulfonylurea. These baseline drugs were continued throughout the trial. Importantly, there were no differences in baseline characteristics between the intent to treat and the completer populations.
  • Exenatide once weekly provided a greater average reduction in A1c over three years vs. insulin glargine. Rapid reductions in A1c through week 36 were observed in both arms. Following week 36, A1c steadily increased with both treatments at approximately the same rate. At three years, however, average A1c remained statistically significantly lower in the ITT group with exenatide once weekly treatment (7.3%) vs. insulin glargine (7.5%; p=0.033). The average insulin glargine dose at three years was 39 IU/day. In the completer population, a similar result was observed with an average A1c of 7.1% in the exenatide once weekly arm vs. 7.4% in the insulin glargine arm (p=0.022). A greater percentage of patients were found to achieve the A1c goals of 7.0% and 6.5% with exenatide once weekly than insulin glargine at three years. In the ITT population, this result was only significant for the 6.5% target (24% v 15%; p=0.02). The 7.0% target was achieved by 40% of patients in the exenatide once weekly arm and 33% of patients in the insulin glargine arm (p=0.12).
  • Insulin glargine was associated with greater reductions in fasting plasma glucose. At three years, there was a -47.7 mg/dl reduction in FPG in the insulin glargine arm vs. a -31.16 mg/dl reduction in the exenatide once weekly arm. Dr. Trautman noted that this result was not surprising and suggested that the much of the glycemic benefit associated with exenatide once weekly is due to its impact on post-prandial control.
  • Exenatide once weekly provided significantly greater weight loss at three years vs. insulin glargine. In the ITT population, there was a weight gain of 2.0 kg with insulin glargine vs. a weight loss of 2.5 kg with exenatide once weekly (p<0.001). The weight loss achieved by exenatide was rapid, with maintenance beginning around week 36 through the end of the trial.
  • No surprising new tolerability and safety findings were reported. Minor hypoglycemia was reported less frequently with exenatide once weekly treatment than with insulin glargine. In the metformin subgroup, 12% of patients treated with exenatide once weekly vs. 40% of patients treated with insulin glargine reported one or more minor hypoglycemic episodes over three years. There was one event of pancreatitis in both groups. Anti-exenatide antibodies were found to decrease overtime, with 80% of the exenatide once weekly treatment group anti-body negative at three years. Further, there was no association between anti-body status and degree of glycemic effect.


Lawrence Leiter, MD (University of Toronto, Toronto, Canada)

Dr. Lawrence Leiter presented 26-week results from the HARMONY 8 trial, which evaluated the efficacy, safety, and tolerability of albiglutide (n=246) compared to sitagliptin (n=240) in people with type 2 diabetes and renal impairment inadequately controlled on lifestyle modifications or other oral diabetes agents. Through week 26, albiglutide (-0.8%) provided significantly greater reductions in A1c than sitagliptin (-0.5%; p=0.0003). Further, numerically greater reductions in A1c were observed with albiglutide regardless of renal impairment severity, although the result only reached significance for the moderate renal impairment group. However, Dr. Leiter noted that the trial was not adequately powered to detect these differences. Greater weight loss was also achieved with albiglutide than with sitagliptin (-0.8 kg vs. -0.2 kg; p=0.028). Both treatments were generally well tolerated and safe. The most common adverse side effects included diarrhea (8.8% with albiglutide vs. 6.1% with sitagliptin), constipation (6% vs. 2%), and nasopharyngitis (5.2% vs. 6.5%). Notably, nausea (4.8% vs. 2.8%) and vomiting (1.6% vs. 0.8%) rates were low in both groups. Although a greater proportion of people experienced hypoglycemia with albiglutide than with sitagliptin (20.5% vs. 13.4%), the majority (>90%) of cases in both arms occurred in participants who were also using an SFU. Antibodies to albiglutide were found in 2.8% of participants, but none of the antibodies were determined to be neutralizing. There was one incidence of pancreatitis in the albiglutide group, but it was determined to be non-related to treatment.

  • HARMONY 8 is a 52-week randomized, double blind, active controlled phase 3 study that will evaluate the efficacy, safety, and tolerability of albiglutide compared to sitagliptin in people with type 2 diabetes and renal impairment (eGFR <90 and>15 ml/min/1.73 m2) inadequately controlled on lifestyle modifications (11%) or oral diabetes agents (metformin, TZD, and/or SFU). After a four-week run in period, participants were randomized to 30 mg albiglutide once weekly (n=246) or sitagliptin (n=240). Albiglutide was uptitrated to 50 mg beginning at week four if glycemic targets were not met. Sitagliptin was dosed by degree of renal impairment according to product label. The treatment arms were balanced at baseline, with an average age of 63 years, BMI of 30 mg/kg2, A1c of 8.1%, and duration of diabetes of 11 years. Further, 8% of each group had a prior MI, 51% in each group were determined to have mild renal impairment (eGFR of 60-89 ml/min/1.73 m2), 40% moderate renal impairment (eGFR 30-60 ml/min/1.73 m2), and 9% severe renal impairment (eGFR <30 ml/min/1.73 m2).
  • Through week 26, albiglutide (-0.8%) provided significantly greater reductions in A1c than sitagliptin (-0.5%; p=0.0003). Further, numerically greater reductions in A1c were observed with albiglutide regardless of renal impairment severity, although the result only reached significance for the moderate renal impairment group. Dr. Leiter noted that the trial was not powered to detect statistically significant differences between these subgroups. In the mild renal impairment group, albiglutide provided an average A1c reduction of -0.72% vs. a -0.66% reduction with sitagliptin. In the moderate renal impairment group, a -0.88% reduction with albiglutide was observed vs. a -0.37% reduction with sitagliptin. In the severe renal impairment group, a -1.08% reduction in A1c was observed vs. a -0.65% reduction with sitagliptin. With regards to FPG, albiglutide provided statistically significantly greater reductions than sitagliptin(-26 mg/dl vs. 4 mg/dl; p<0.0001). A greater proportion of participants were also able to achieve a target A1c of 7.0% with albiglutide (24.5%) than with sitagliptin (19.2%). 34% of participants in the albiglutide arm required uptitration to the 50 mg dose during the trial.
  • Greater weight loss was achieved with albiglutide than with sitagliptin. The albiglutide arm achieved an average weight loss of 0.8 kg vs. an average weight loss of 0.2 kg in the sitagliptin arm (p=0.028).
  • Both treatments were demonstrated to be generally well tolerated and safe. Drug related adverse effects were slightly more frequent in the albiglutide arm than the sitagliptin arm (19.7% vs. 13.0%). The most common adverse side effects included diarrhea (8.8% with albiglutide vs. 6.1% with sitagliptin), constipation (6% vs. 2%), and nasopharyngitis (5.2% vs. 6.5%). Notably, nausea (4.8% vs. 2.8%) and vomiting (1.6% vs. 0.8%) rates were low in both groups. Dr. Leiter attributed albiglutide’s low rate of nausea and vomiting to its long half-life, slow ramp up to therapeutic blood levels, and limited access to the CNS given its large size. With regards to hypoglycemia, there were no severe cases of hypoglycemia with albiglutide vs. two severe cases with sitagliptin. 20.5% of people in the albiglutide arm experienced hypoglycemia (94% of whom were also using an SFU) vs. 13.4% with sitagliptin (91% of whom were also using an SFU). Antibodies to albiglutide were found in 2.8% of participants, but none of the antibodies were determined to be neutralizing. There was one incidence of pancreatitis in the albiglutide group, but it was determined to be non-related to treatment. There were no incidences of pancreatitis in the sitagliptin group.

Questions and Answers

Q: How is albiglutide metabolized?

A: It is cleared by the reticuloendothelial system.

Q: What are your thoughts regarding the minimal GI side effects observed?

A: These results are consistent with previous data for albiglutide which showed a low incidence of nausea and vomiting. It is related in part to its once weekly injection, the gradual increased in blood drug levels, and the molecules large size, which possibly prevents CNS entry.

Q: Was there any difference in nausea rates with changing levels of renal impairment?

A: No, there was not.


Michaela Diamant, MD, PhD (VU University Medical Center, Amsterdam, Netherlands)

  • Dr. Michaela Diamant presented the results of a study directly comparing a prandial GLP-1 receptor agonist (exenatide BID) with a standard prandial insulin (insulin lispro T1DM). Following 12 weeks of basal insulin titration, participants who did not achieve A1c levels below 7.0% were randomized 1:1 to receive exenatide BID (5 μg, titrated up to 10 μg after four weeks, n=315) or insulin lispro treatment (n=312) for 30 weeks. This phase-3b 44-week open-label randomized, comparator-controlled interventional study found that both treatment arms achieved similar A1c reductions, but with lower fasting glucose, greater weight loss, less daytime hypoglycemia, and reduced systolic blood pressure observed among patients receiving exenatide BID.
  • In addition to establishing non-inferiority in A1c reduction (baseline after basal titration A1c 8.5%; exenatide 7.0% vs. 7.1% with insulin lispro at 44 weeks), while individuals receiving insulin lispro gained 2.1 kg (4.6 lbs), individuals receiving exenatide BID lost an average of 2.5 kg (5.5 lbs) after 30 weeks into treatment intensification. The change in fasting glucose was also 0.64 mmol/L (11 mg/dl) lower among patients receiving exenatide BID compared to patients receiving insulin lispro. Self-monitored blood glucose was also lower, except at lunch due to insulin lispro delivery at each meal, versus exenatide’s twice-daily delivery. At this time, approximately 15 more units of basal insulin were required daily by participants receiving exenatide BID (57 vs. 42).
  • Daytime hypoglycemia was found to be less frequent among individuals receiving exenatide BID than insulin lispro (15.2% vs. 33.7% incidence; p<0.001). Minor hypoglycemia was also less common; 40.7% incidence was reported among patients receiving insulin lispro, whereas only 29.5% incidence was observed among patients receiving exenatide BID (p<0.004). Given these results, Dr. Diamant believes that a short-acting GLP-1 receptor agonist may be a novel and efficacious treatment strategy for patients who fail on insulin and further intensification.

Oral Sessions: New Information on DPP-4 Inhibition


Odd Erik Johansen, MD, PhD (Boehringer-Ingelheim, Ingelheim, Germany)

Dr. Odd Erik Johansen discussed a study that compared the incidence of CV events for linagliptin (Tradjenta) vs. comparator drugs in ~9,500 patients with type 2 diabetes enrolled in 19 double-blind RCTS (“the most comprehensive trial database available”). He dedicated nearly half of his presentation time to background information, ending with a swift review of the data. The study used a prospectively defined adjudication process and measured time to first event for a composite primary endpoint of CV death, non-fatal stroke, non-fatal myocardial infarction, and hospitalization for unstable angina pectoris (4P-MACE). The primary analysis indicated that linagliptin has a favorable effect on CV risk (HR: 0.78; 95% CI: 0.55-1.12). Further analysis shows that this result was mediated mostly by linagliptin’s effect on non-fatal stroke (HR: 0.34; 95% CI: 0.15-0.75) – data for CV death (HR: 1.04; 95% CI: 0.42-2.60) and non-fatal myocardial infarction (HR: 0.86; 95% CI: 0.47-1.56) were not as compelling. For these three endpoints combined (strict MACE composite), the hazard ratio was positive (0.74) and with the upper bound of the 95% confidence interval over one (95% CI: 0.49-1.13). Dr. Johansen ended his presentation by highlighting that further CV data will emerge from linagliptin’s cardiovascular outcomes trial (CVOT) CAROLINA (comparing the drug to glimepiride; n=6,103), as well as from the CARMELINA study (a placebo-controlled CV and renal outcomes study; n=8,300). As noted previously, BMS/AZ just reported topline results from Onglyza’s CVOT SAVOR-TIMI showing that the drug was non-inferior but not superior to placebo in reducing CV risk (full details are in our report at

  • Dr. Johansen explained the rationale for this analysis of CV risk. He began by reminding the audience that type 2 diabetes confers a roughly two-fold excess risk for a wide range of vascular disease, independent of other cardiovascular (CV) factors. Studies show that people with type 2 diabetes exhibit characteristics of CV biopathology distinct from those without the disease. Dr. Johansen cited an independent CV assessment of the DPP-4 inhibitor class vs. other drugs that included 70 trials (n= ~42,000) and found no CV harm associated with DPP-4 inhibitors (Monami et al., Dibaetes Obes Metab 2013). In addition, mechanistic trials have indicated that DPP-4 inhibitors have several potential CV benefits independent of their effects on glycemia (e.g., decreasing blood pressure and triglyceride levels, increasing left-ventricle function, decreasing inflammation and oxidative stress, increasing endothelial function, decreasing myocardial infarct size).
  • The pooled analysis included 5,847 patients on linagliptin (5 mg and 10 mg; 4,421 patient years in total) and 3,612 patients on the comparators (placebo, glimepiride, voglibose; 3,255 patient years of exposure). As baseline, the participants had an average age of 59 years, A1c of 8.1%, BMI of 29-30 kg/m2, and Framingham 10-year CV risk score of 9.7- 10.2.
  • Hazard ratios for secondary and tertiary endpoints:

Secondary Endpoints


Tertiary Endpoints

Strict MACE

HR: 0.74

(95% CI: 0.49-1.13)


CV death

HR: 1.04

(95% CI: 0.42-2.60)

All adjudicated CV events

HR: 0.82

(95% CI: 0.61-1.09)


Non-fatal MI

HR: 0.86

(95% CI: 0.47-1.56)

FDA-custom MACE (not defined)

HR: 0.70

(95% CI: 0.45-1.08)


Non-fatal stroke

HR: 0.34

(95% CI: 0.15-0.75)


Transient ischemic attack

HR: 0.09

(95% CI: 0.01-1.75)


Unstable angina pectoris w/hospital.

HR: 1.08

(95% CI: 0.56-2.06)


Total Mortality

HR: 0.89

(95% CI: 0.45-1.75)


Satoshi Funakoshi, MD, PhD (Sakuramachi Clinic, Nagasaki, Japan)

Dr. Satoshi Funakoshi presented evidence that replacing insulin with the DPP-4 inhibitor linagliptin (Lilly’s Tradjenta) in people undergoing hemodialysis can reduce the glycemic variability associated with hemodialysis. In his study, six relatively well-controlled patients with diabetes (A1c <7%) who were on insulin were monitored by CGM on one hemodialysis-free (HD-free) day and one hemodialysis (HD) day. Subsequently, they were switched off of insulin onto linagliptin 5 mg and again monitored by CGM on one HD-free and one HD day. Dr. Funakoshi did not specify how long the patients underwent linagliptin treatment prior to the second round of HD and CGM monitoring. He presented two case studies where mean amplitude of glucose excursions (MAGE) was markedly decreased on HD days after switching to linagliptin (~20 mg/dl decrease in one case and ~100 mg/dl decrease in the other). Overall, when he showed the composite CGM traces for all six patients, it was very clear that switching to linagliptin flattened out the glycemic variability at the time of HD. As such, MAGE on the HD days significantly decreased after switching from insulin to linagliptin (mean value not provided; p<0.05). No changes in hypoglycemia or A1c were observed after switching to linagliptin. We found it quite striking that patients needing insulin could achieve adequate control after switching to linagliptin – it was unclear whether Dr. Funakoshi was suggesting that patients should only switch to linagliptin treatment on days when they need HD or if they should switch to linagliptin treatment long-term. This seems like a creative approach to solve a key problem for this very sick population.

  • Hemodialysis (HD) is often associated with glycemic variability. Dr. Funakoshi proposed a number of potential reasons: 1) glucose transport into erythrocytes may change due to cytoplasmic pH changes during HD; 2) the HD solution contains glucose to buffer against acute hypoglycemia; 3) removal of insulin from the blood during HD could result in an insulin shortage afterwards; 4) clearance of uremic toxins from the blood improves insulin resistance.
  • Dr. Funakoshi hypothesized that, with its glucose-dependent mechanism of action, linagliptin could reduce glycemic variability during and after HD. Linagliptin is the only DPP-4 inhibitor cleared hepatically rather than renally, so is the prime incretin option for patients with renal impairment. Additionally, it has a high protein binding rate (>90%), so it is minimally removed by HD therapy.
  • The study enrolled six relatively well-controlled patients with diabetes on hemodialysis (A1c <7%) and on background insulin therapy. Patients monitored by CGM on one hemodialysis-free (HD-free) day and one hemodialysis (HD) day. Subsequently, they were switched off of insulin onto linagliptin 5 mg and again monitored by CGM on one HD-free and one HD day. Dr. Funakoshi did not specify how long the patients underwent linagliptin treatment prior to the second round of CGM monitoring.
  • Switching from insulin to linagliptin significantly reduced mean amplitude of glucose excursions (MAGE) on HD days (mean value not specified; p<0.05). A composite graph of all six CGM tracings visually supported this finding – tracings on the HD day were much flatter and much more resembled the HD-free day after switching to linagliptin.


Marco Bugliani, PhD (University of Pisa, Pisa, Italy)

Dr. Marco Bugliani presented in vitro evidence that 1) the DPP-4 enzyme is present in human islet cells (particularly in the alpha cells); 2) that DPP-4 inhibitors exert a direct protective effect against gluco- and lipotoxicity; and 3) that these effects may be mediated by actions of a local GLP-1 system in the islet cells. Using immunohistochemistry, Dr. Bugliani’s group found that human pancreatic tissue stained positively for DPP-4; using immunofluorescence, they co-localized the DPP-4 expression with glucagon but not with insulin or GLP-1, thus concluding that DPP-4 was present primarily in alpha cells. They then treated islet cells with the DPP-4 inhibitor MK-0626 and found that MK-0626 treatment increased active GLP-1 concentration in islets; induced insulin release; preserved beta cell glucose sensitivity after 24-hour exposure to 11 mM glucose or 0.5 mM palmitate; and conferred a partial preservation of insulin granules. MK-0626 produced a significant increase in DPP-4 gene expression in islets cells, but no major changes were observed for insulin or GLP-1 gene expression.


James Foley, PhD (Novartis Pharmaceuticals Corporation, East Hanover, NJ)

Dr. James Foley presented evidence that baseline fasting plasma glucose predicts the weight change a patient experiences on vildagliptin (Novartis’ Galvus). Type 2 diabetes patients frequently have plasma glucose levels above the renal threshold for glucose reabsorption (approximately 10 mmol/L or 180 mg/dl), resulting in glucose loss through urine. Thus, any anti-hyperglycemic agent may reduce glycosuria, thereby increasing caloric retention. Dr. James Foley’s group hypothesized that weight change experienced by patients on the DPP-4 inhibitor vildagliptin is dependent on baseline fasting plasma glucose (FPG) levels, with high initial FPG corresponding to a higher potential for accruing a positive caloric balance due to increased caloric retention. They compiled data on 2,863 type 2 diabetes patients drawn from a database of vildagliptin clinical monotherapy studies. The average age was 54 years, and the mean duration of their diabetes was two years. The results showed that patients’ change in body weight was positively and significantly correlated with their baseline FPG levels, with a coefficient of approximately 8.5 g/(mg/dl). A FPG level of 230 mg/dl was the break-even point in terms of weight gain or loss. The data demonstrated that, when glycosuria is taken into account, vildagliptin treatment results in negative caloric balance. To explain the phenomenon, Dr. Foley noted that DPP-4 inhibitors likely reduce intestinal triglyceride absorption, increase lipolysis in adipose tissue, and increase fat oxidation in muscles. Despite the limitations of database studies, we believe these results provide valuable information on which patients can expect to see weight loss or gain on vildagliptin therapy.

Questions and Answers

Q: Have you been able to do any studies on thermogenesis?

A: No, we haven’t yet done a formal thermogenic study.

Q: I assume the same would hold true for measurements of glycosuria?

A: No weren’t able to obtain those data either.


Noriko Satoh-Asahara, MD (Kyoto Medical Center, Kyoto, Japan)

Dr. Noriko Satoh-Asahara discussed a study evaluating the effect of sitagliptin on M1 and M2-like phenotypes, an indication of the drug’s cardiovascular (CV) effects. To begin, Dr. Satoh-Asahara noted that GLP-1 has pleiotropic effects that could favorably impact CV risk (e.g., decreasing body weight, glucose levels, and silent inflammation, oxidative stress, endothelial dysfunction, and blood pressure). She further explained that obesity induces a phenotypic switch in the macrophages of adipose tissue – from an anti-inflammatory (M2) to a pro-inflammatory (M1) polarization – which contributes to inflammation and insulin resistance. This switch is detected by measuring M1 and M2 markers in peripheral blood monocytes (a type of white blood cell). This open-label trial randomized 52 Japanese type 2 patients (at baseline, mean age of 60 years, BMI of 26 kg/m2, A1c of 8.2%) to either diet alone (n=26) or to diet plus sitagliptin (50 mg daily; n=26) for three months. As expected, sitagliptin therapy decreased fasting plasma glucose, A1c and plasma GLP-1 levels. In addition, sitagliptin decreased levels of the M1 pro-inflammatory marker TNFα in both serum and monocytes (p<0.01 and p<0.05, respectively) and increased levels of the M2 an-inflammatory marker IL-10 in both serum and monocytes (p<0.01 and p<0.05, respectively). Dr. Satoh-Asahara noted that multivariate regression analysis revealed that sitagliptin was the only factor independently associated with an increase in monocyte IL-10. The therapy also decreased serum levels of two other inflammatory markers – oxidized LDL and C-reactive protein – further suggesting its positive effect on the CV system.

Oral Sessions: Novel Therapeutics


Stephanie Gustavson, PhD (TransTech Pharma, High Point, NC)

Dr. Stephanie Gustavson presented the results of a four-week trial examining the safety and efficacy of TTP054, a phase 2 oral GLP-1 agonist currently in development by TransTech Pharma. In the trial, patients with type 2 diabetes on background metformin were randomized to receive either TTP054 400 mg QD (n=9), TTP054 200 mg BID (n=10), TTP054 200 mg QD (n=11), or placebo (n=18). TTP054-treated patients demonstrated consistently greater declines in fasting plasma glucose (-15 mg/dl with 200 mg QD, -18 mg/dl with 200 mg BID, and -20 mg/dl with 400 mg QD vs. -5 mg/dl with placebo) and two-hour postprandial glucose (-30 mg/dl, -32 mg/dl, and -35 mg/dl vs. no change with placebo) versus placebo-treated patients; while not long enough in duration to produce change in A1c, predictive models estimated A1c change at three months at about 0.6-1.0% across doses, which Dr. Gustavson suggested was a conservative estimate given the study design. Notably, TTP054-treated patients demonstrated no hypoglycemic events, with incidence of GI adverse events similar to those observed in the placebo arm. Dr. Gustavson noted that all phase 1 studies with TTP054 have completed, with one 12- week phase 2 study just recently completed. Given the reduction in GI adverse events and convenience, we will be interested to follow TTP054 as further clinical data is released. The company currently has another oral GLP-1 candidate, TTP273, in development; other companies known to have oral GLP-1 candidates include Zydus Cadila (ZYOG1; phase 1), Arisgen (preclinical), Heptares (preclinical), and Novo Nordisk (NN9924, NN9926, NN9927; all phase 1).

  • TTP054 is a phase 2 oral GLP-1 agonist currently in development by TransTech Pharma. In addition to the convenience of oral administration, given administration is through the intestine, it is expected the drug will have a lower incidence of GI adverse events versus injectable GLP-1; additionally, the oral drug is a small molecule agonist rather than a peptide, so it is believed it will not incur antibody formation. Dr. Gustavson noted that TTP054 has completed phase 1 studies, with one 12-week phase 2 study recently completed. An additional candidate TTP273 is in development with proof of mechanism demonstrated and phase 2b trials to begin in early 2014.
  • In this trial, patients with type 2 diabetes on background metformin were randomized to receive either TTP054 400 mg QD (n=9), TTP054 200 mg BID (n=10), TTP054 200 mg QD (n=11), or placebo (n=18) for four weeks. Evidencing proper absorption, TTP054 had a median Tmax of ~3 hours and mean terminal half-life of ~6 hours.
  • TTP054-treated patients demonstrated consistently greater declines in fasting plasma glucose (-15 mg/dl with 200 mg QD, -18 mg/dl with 200 mg BID, and -20 mg/dl with 400 mg QD vs. -5 mg/dl with placebo) and two-hour postprandial glucose (-30 mg/dl, -32 mg/dl, and -35 mg/dl vs. no change with placebo) versus placebo-treated patients. As the trial was too short in duration to show A1c change, Dr. Gustavson relied on various models to estimate A1c change at three months using these effects on glucose variables. Across four models of predicted change, declines in A1c were roughly -0.7% with 200 mg QD, -0.6% with 200 mg BID, and -1.0% with 400 mg QD versus -0.2% versusplacebo (baseline A1cs 8.1%, 8.1%, 8.3%, and 7.7%, respectively). Additionally, she suggestedthese estimates were conservative, given the meal challenge in this trial was small with no stabilization of diet prior to assessment and glucose lowering was continuing to decline with treatment at the time of assessment.
  • TTP054-treated patients demonstrated no hypoglycemic events, with incidence of GI adverse events similar to those observed in the placebo arm (2/11 patients [18%] with 200 mg QD, 0/10 patients [0%] with 200 mg BID, and 1/9 patients [11%] with 400 mg QD vs. 1/18 [6%] with placebo). Dr. Gustavson suggested all GI adverse events were mild in severity with investigators indicating adverse events were due to diet rather than the study treatment. While not powered to assess changes in lipids, Dr. Gustavson suggested that the maximum change in triglycerides with TTP054 was -50 mg/dl versus -10 mg/dl with placebo and that the maximum change in total cholesterol was -15 mg/dl versus +5 mg/dl with placebo, indicating a potential benefit.

Questions and Answers

Q: Do you have evidence of effect on GI motility?

A: We did look at GI motility in other studies and saw reduction. We also saw reduction in gastric emptying, so we think the drug is acting in the same way as the mimetics.

Oral Sessions: Diabetic Dyslipidemia


Daisaku Masuda, MD, PhD (Osaka University, Osaka, Japan)

Presenting a study of linagliptin in mice, Dr. Daisaku Masuda explained how the drug reduces post- meal triglyceride levels. He opened by reminding the audience that recent studies have demonstrated postprandial hypertriglyceridemia (PHTG) as a risk factor for atherosclerotic cardiovascular diseases. The cause of PHTG is accumulation of chylomicrons, including triglycerides, Oral fat loading (OFL) tests have shown that people with type 2 diabetes have high triglyceride peaks, and reports have shown that DPP-4 inhibitors may improve PHTG in these individuals. Using eight-week old mice (n=20), Dr. Masuda’s group investigated the possible mechanisms of this effect. After four weeks on chow alone or chow plus linagliptin, the mice underwent an overnight fast and then an OFL test. The linagliptin treatment group showed significantly reduced postprandial triglyceride levels four hours after the OFL compared to the control group (320 mg/dl vs. 464 mg/dl, p<0.01), and significantly reduced free fatty acid levels six hours post-OFL. Treatment with linagliptin also decreased the postprandial mRNA expression of apoB and FATP-4, which are involved in the intestinal formation of chylomicrons. Dr. Masuda concluded that oral administration of linagliptin improves postprandial triglyceride and triglyceride-rich lipoprotein metabolism and decreases the intestinal production of chylomicrons. In other words, the research demonstrated a possible pathway by which DPP-4 inhibitor therapy might improve PHTG in people with type 2 diabetes.



Simon Heller, Stefan Korsatko, Jamala Gurban, Lene Jensen, Erik Christiansen, Fumiaki Kiyomi, and Thomas Pieber

This randomized, double-blind, crossover study investigated the effects of liraglutide treatment as an adjunct to insulin in patients with type 1 diabetes (n=45). Patients were randomized to one of three doses of liraglutide – 0.6 mg (n=15), 1.2 mg (n=14), or 1.8 mg (n=16) – and to one of two sequences (placebo/liraglutide or liraglutide/placebo) for a four-week treatment period, two-to-three-week washout period, and four-week crossover treatment. At baseline, patients across groups had comparable age (30-39 years), A1c (7.5-7.8%), body weight (mean 72-75 kg [158-165 lbs]), duration of diabetes (14-18 years), and daily insulin dose (42-49 U). Daily insulin dose was significantly reduced by 27% in the 1.2 mg and by 24% 1.8 mg liraglutide group versus placebo (p<0.001), but there were no significant reductions in the 0.6 mg liraglutide group. This indicates a pharmacologic effect that may contribute to glycemic control when liraglutide is added as an adjunct to insulin to treat type 1 diabetes. All liraglutide groups significantly reduced weight (p<0.001). Adverse events were higher during liraglutide treatment compared to placebo, with the highest number due to gastrointestinal adverse events. No differences in the number of hypoglycemic episodes or patients with hypoglycemic episodes were observed across treatments, and there were no severe hypoglycemic episodes reported.

  • After four weeks, liraglutide 0.6 mg, 1.2 mg, and 1.8 mg brought about significant weight reductions versus placebo. Patients lost 2.0 kg (4.4 lbs), 3.7 kg (8.1 lbs), and 3.3 kg(7.3 lbs) versus placebo (p<0.001) in the 0.6 mg, 1.2 mg, and 1.8 mg arms, respectively. Therewere no significant changes in blood pressure, pulse, or body temperature for liraglutide versus placebo.
  • Adverse effects were common: 67%, 86%, and 93% of patients in the 0.6 mg, 1.2 mg, and 1.8 mg liraglutide groups experiencing GI adverse events. Nausea occurred in 53%, 78%, and 78% of patients in the 0.6 mg, 1.2 mg, and 1.8 mg liraglutide groups, respectively. Atleast 86% of patients among all groups experienced documented symptomatic hypoglycemic episodes, and at least 93% of patients among all groups experienced an ADA-classified hypoglycemic event during treatment. However, no severe hypoglycemic episodes occurred.


Ruth Weinstock, Guillermo Umpierrez, Bruno Guerci, Michael Nauck, Karen Boleyn, Zachary Skrivanek, and Zvonko Milicevi

This 104-week, randomized, multi-center, double-blind trial examined the efficacy and safety of dulaglutide (DU) versus sitagliptin (SITA) in patients with type 2 diabetes treated with metformin. Patients were randomized to 1.5 mg DU (n=304), 0.75 mg DU (n=302), 100 mg SITA (n=315), or placebo (n=177) (to 26 weeks). At baseline, subjects had average BMI of 31 kg/m2, age of 54 years, A1c 0f 8.1-8.2%, and weight of 87 kg (191 lbs). At the end of 104 weeks, DU provided significantly greater A1c reductions than SITA (p<0.001). In the 1.5 mg DU, 0.75 mg DU, and SITA arms, A1c decreased an average 1.0%, 0.7%, and 0.3%, respectively, over 104 weeks. The 1.5 mg DU group was associated with greater body weight loss than SITA (1.1 kg [2.4 lbs] beyond placebo) (p<0.001) at 104 weeks. There was a higher rate of GI adverse events in the DU arms compared to SITA, including nausea, vomiting, diarrhea, constipation, and abdominal distension. The incidence and rates of hypoglycemia were similar across treatment groups and there were no reports of severe hypoglycemia episodes.

  • At the end of 104 weeks, A1c decreased in the 1.5 mg DU, 0.75 mg DU, and 100 mg sitagliptin groups by an average 1.0%, 0.7%, and 0.3%, respectively. The percentage of patients who achieved an A1c <7.0% at the end of the study was 54%, 45%, and 31%, with a weight reduction of 2.9 kg (6.4 lbs), 2.4 kg (5.3 lbs), and 1.7 kg (3.8 lbs), in the 1.5 mg DU, 0.75 mg DU, and 100 mg SITA arms, respectively. We note that a high number of people dropped out in each treatment arm: 37% in the 1.5 mg DU, 39% in the 0.75 mg DU, 41% in the SITA, and 46% in the placebo arm, with the primary cause due to adverse events. We also note that serious adverse events were reported in 12% of patients in the 1.5 mg DU arm, 8% of patients in the 0.75 mg DU arm, and 10% of patients in the 100 mg SITA arm. No difference across treatment groups in systolic or diastolic pressure was observed.
  • There were no reports of severe hypoglycemic episodes at any point during the study. The incidence of hypoglycemia was, however, prevalent and similar across treatment groups: 13% in the 1.5 mg DU arm, 9% in the 0.75 mg DU arm, and 9% in the 100 mg SITA arm. The authors concluded that the safety profile and superior A1c-lowering efficacy of DU vs. SITA after 104 weeks indicate an acceptable benefit/risk profile; however, it is important to note that significantly more patients in the DU arms experienced adverse events during the first 24 weeks of the study, with the highest incidence among the 1.5 mg DU group.


Mugdha Gokhale, John Buse, Virginia Pate, Christine Gray, Alison Marquis, and Til Stürmer

Recent evidence has suggested a potential association between pancreatic cancer and DPP-4 inhibitors, but to date, the human populations studied have been limited. This poster explored the incidence of pancreatic and all other cancers among new DPP-4 inhibitor users compared to those initiating sulfonylureas or TZDs. The source of data was a 20% random sample of all available Medicare claims from 2007-2010. All patients were >65 years of age and inclusion criteria required at least 6 months enrollment prior to drug initiation. Patients with any cancer diagnosis prior to initiation or between the first and second prescription were excluded. The study found no increased risk of pancreatic cancer or any cancer post initiation of DPP-4 inhibitors as compared to sulfonylureas or TZDs. For pancreatic cancer, the adjusted hazard ratio (HR) for the as-treated patient group was 0.52 (95% CI=0.28-1.00) for DPP-4 inhibitors vs. SFU and 1.11 (95% CI=0.67-1.83) for DPP-4 inhibitors vs. TZDs. The results for crude HRs and the intent-to-treat population followed the same pattern whereby the point estimate for DPP-4 vs. SFU was below 1.0 but the 95% CI did not exclude 1.0, and the point estimate for DPP-4 vs. TZD was slightly over 1.0 but with the 95% CI crossing 1.0. To exclude the possibility of reverse causality, the authors repeated the analysis excluding first 6 months of the follow-up. However, results remained the same. To address concerns about detection bias, the authors noted that there was little evidence for differential diagnostic work-up in DPP-4 inhibitor initiators. There was also no difference in any-cancer incidence when comparing DPP-4 inhibitor to SFU or TZD initiators.

  • The comparison between DPP-4 inhibitors and sulfonylurea initiators compared 11,602 and 48,746 individuals, respectively. The comparison between DPP-4 inhibitors and TZDs followed 18,991 DPP-4 inhibitor initiators and 22,441 TZD initiators. Populations were weighted to balance distribution of baseline covariates. The mean age of DPP-4 initiators was 75 years with 35% men and 65% women. Insulin use 6 months prior to index date was 15.2-19.5% among all groups.
  • Limitations in this study include the short treatment from actual treatment (as treated analysis), homogeneity of age (all subjects over 65), and data availability (intent-to-treat analysis).


Silvio Inzucchi, Michael Nauck, Maximilian von Eynatten, Uwe Hehnke, Hans-Juergen Woerle, and Robert Henry

Hypoglycemia remains a prominent concern for elderly insulin-requiring patients with type 2 diabetes. This study demonstrated that adding linagliptin to basal insulin could both improve glycemic control and improve hypoglycemia compared to placebo. Linagliptin (Lilly/BI’s Tradjenta), is one of few oral agents that is not renally excreted, and thus does not require dose adjustments for people with impaired renal function. This study was a pooled analysis of data from two phase 3, double-blind, randomized, placebo-controlled trials that investigated hypoglycemia risk with linagliptin in patients above the age of 70 receiving basal insulin. Baseline characteristics of subjects included a mean age of 74 years, A1c of 8.2%, BMI of 30 kg/m2, and basal insulin dose of 36 U/day. Treatment with linagliptin resulted in significantly reduced A1c levels from baseline (-0.77%) and no relevant changes in daily basal insulin requirements. Despite this improvement in glycemic control, adding linagliptin substantially decreased overall (-37%; OR 0.63) and confirmed symptomatic (-34%; OR 0.66) hypoglycemia risk compared to placebo. We think that now the even more interesting follow-up study would be to compare linagliptin to prandial insulin or a GLP-1 agonist as a basal-insulin add-on.

  • In two phase 3 trials of linagliptin, patients with type 2 diabetes were randomized to receive either 5 mg once-daily linagliptin (n=126) or placebo (n=121), in addition to stable doses of basal insulin. Baseline characteristics of subjects included a mean age of 74years, A1c of 8.2%, BMI of 30.3 kg/m2, and basal insulin dose of 36 U/day. One study was ≥ 52- weeks comprised of patients ≥ 18 years taking basal insulin alone or in combination with metformin and/or pioglitazone. The other study was a 24-week study consisting of patients ≥ 70 years of age receiving metformin and/or sulfonylurea and/or basal insulin. The glycemic efficacy and tolerability of linagliptin relative to placebo was assessed using data pooled from the subgroups of patients ≥ than 70 years of age in each study. Basal insulin doses did not change notably during the study.
  • A significant reduction in incidence of overall and confirmed (blood glucose70 mg/dl) hypoglycemia was found in patients treated with linagliptin. A downward trend in incidences of hypoglycemia with linagliptin vs. placebo was observed for patients with mild- moderate baseline hyperglycemia (A1c 7.5-<9.0%; OR 0.41), with baseline A1c <7.5% (overall OR 0.77), and subgroups receiving insulin glargine, insulin detemir, or NPH insulin (overall OR 0.74, 0.59, 0.49, respectively). The mechanism for reduced risk of hypoglycemia with linagliptin was speculated to involve the glucagon counterregulatory response, though this hypothesis requires further study to confirm.


Samuel Engel, Lei Xu, Gregory Golm, Edward O’Neill, Keith Kaufman, and Barry Goldstein

This post-hoc analysis of pooled data from three double-blind studies compared the efficacy and safety of sitagliptin (Merck’s Januvia) to sulfonylureas in patients with type 2 diabetes (n=1,180) and mild renal impairment (eGFR=60-89 ml/min/1.73 m2). The primary endpoint was change from baseline A1c, and special focus was placed on differences in rates of hypoglycemia between the two treatment arms. From a baseline A1c of 7.6%, sitagliptin and sulfonylurea provided similar A1c reductions (0.62% and 0.68%, respectively). However, treatment with sitagliptin provided less symptomatic hypoglycemia (7% vs. 26% of patients on sitagliptin vs. sulfonylurea, respectively, reported at least one event; p<0.001), and loss in body weight compared to SU (0.9 kg [2 lb] weight loss on sitagliptin vs. 1.4 kg [3 lb] weight gain on sulfonylurea; p<0.0001). Most notably, 40.6% of patients on sitagliptin’s met the composite endpoint of an A1c decrease of greater than 0.5% with no symptomatic hypoglycemia and no weight gain vs. only 16.8% of patients treated with SU.

  • Data from 1,180 patients with type 2 diabetes and mild renal insufficiency were pooled from 3 randomized, double-blind studies. Subjects were administered either 100 mg/day of sitagliptin (SITA n=584) or sulfonylurea (SU n=596) in titrated doses over a 25-30 week study period. In both treatment arms, subjects were similar in terms of baseline characteristics. The mean age was 58 years, and mean BMI was 30-31 kg/m2. Baseline A1c levels were 7.6% in both groups, while baseline fasting plasma glucose (FPG) levels were 154 mg/dl in the SITA group and 156 mg/dl in the SU group.
  • SITA and SU demonstrated comparable improvements in glycemic control. In terms of change in A1c from baseline, SITA achieved a 0.62% reduction vs. 0.68% reduction in SU. Furthermore, 60% of patients treated with SITA reached A1c levels of less than 7%, compared to 64% of patients on SU. The percentage of patients with greater than a 0.5% reduction in A1c in SITA vs. SU were 56.7% and 56.5%, respectively. Change in FPG levels from baseline were also similar, with a decrease of 16.6 mg/dl in SITA and 18.0 mg/dl in SU. None of these small differences were statistically significant, suggesting equivalent glycemic efficacy in both drugs.
  • SITA was associated with a lower incidence of hypoglycemia. Compared to 26.2% of patients on SU, only 6.8% of patients on SITA experienced at least one symptomatic hypoglycemic event. None of the patients treated with SITA reported an episode of severe hypoglycemia (defined as requiring assistance), while one patient receiving SU reported at least one event.
  • Treatment with SITA resulted in greater weight loss compared to SU. In addition to the lower prevalence of adverse events of symptomatic hypoglycemia, patients treated with SU experienced weight gain from baseline (+1.4 kg [3 lb]), whereas those on SITA observed a 0.9 kg (2 lb) loss in body weight.

Symposium: Non-Glycemic Effects of Incretin-Based Therapy – Glucagon-Like Peptide-1 (GLP-1) and Dipeptidyl Peptidase-4 (DPP-4) (Supported by Boehringer Ingelheim and Eli Lilly)


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

Upwards of 1,000 people flocked to hear Dr. Vanita Aroda present on incretin safety. Dr. Aroda attended the June 5-6 NIDDK/NCI workshop on pancreatitis, diabetes, and pancreatic cancer, and her presentation largely recapitulated highlights from the discussion that took place there (for our coverage of the meeting, please see our reports for day one and day two). The conclusions she drew, in her fast-paced and well- organized presentation, resonated largely with prevailing sentiments on the subject: 1) potential mechanisms have been identified in animal models to suggest a potential pancreatitis risk; in her view, the real debate is over which of these models is relevant and representative since different models have produced differing results; 2) current clinical evidence is insufficient to support altering the risk/benefit profile for incretin-based therapies; 3) recent studies (e.g., Singh et al., JAMA Int Med 2013 and Butler et al., Diabetes 2013) have not prompted substantial changes in clinical recommendations; and 4) there is a great need to pool together the appropriate expertise and data based on rigorous methodologies to address these questions.

  • Dr. Aroda reviewed the many factors that complicate the investigations into pancreatitis risk: type 2 diabetes increases risk of pancreatitis and pancreatic cancer by 82% (Huxley et al., British Journal of Cancer 2005). Additionally, there is the chicken and the egg problem of reverse causality – individuals recently diagnosed with diabetes (<4 years) have a 50% greater risk of pancreatic cancer compared to those with longer diabetes duration (odds ratio of2.1 vs. 1.5; Huxley et al., British Journal of Cancer 2005).
  • Mechanistically, animal models have provided conflicting results on the effects of the exocrine pancreas. Dr. Aroda emphasized that the real question is discerning which models were appropriate and representative. Liraglutide did not induce pancreatitis in mice, rats, or monkeys at exposure levels greater than 60 times the levels used in humans (Nyborg et al., Diabetes 2012). Meanwhile, in the Pdx-1 Kras mouse (a model primed to develop chronic pancreatitis and pancreatic cancer), 12 weeks of exendin-4 treatment produced expansion of pancreatic duct glands; additionally, premalignant pancreatic intraepithelial lesions (PanINs) were identified in these animals. The FDA’s re-examination of toxicology and carcinogenicity studies has not provided any additional clarity.
  • Recent clinical evidence of elevated pancreatitis risk for incretin-based therapies comes from two largely flawed studies. Dr. Aroda relayed criticisms of Dr. Butler’s “cadaver study” (Diabetes 2013) and Dr. Singh’s retrospective insurance claims cohort analysis (JAMA Int Med 2013) that were discussed at the NIH pancreatitis meeting. The three groups from which Dr. Butler collected pancreata (nondiabetic, type 2 diabetes + incretin therapy, and type 2 diabetes without incretin therapy) were very small and poorly matched for background treatment, diabetes duration, age, gender, and BMI. Dr. Aroda referenced Dr. Steven Kahn’s commentary stating that the study also did not control for the effects of prolonged life support. In the Singh et al., study, the treatment group had higher background rates of risk factors for pancreatitis (e.g., obesity, alcohol use, hypertriglyceridemia, tobacco abuse, etc.). She reviewed methodological limitations of observational studies, including inability to control for covariates/confounders or a number of biases (e.g., reporting bias, notoriety bias, channeling bias [preferential prescribing to specific patient populations], selection bias, and reverse causation).
  • For our previous coverage of the diabetes community’s reactions to these studies when they were initially published, please see our Closer Looks at and
  • The NIH workshop also discussed the FDA’s Adverse Events Reporting System (AERS) database, concluding that additional data mining of AERS is unlikely to shed more light on these safety signals. Given the nature of the voluntary, spontaneous reporting, these data can only be hypothesis generating rather than hypothesis confirming.
  • So far, randomized controlled trials (RCTs) of incretin-based therapies have not detected a pancreatitis signal. For liraglutide, Dr. Aroda stated that the rate of pancreatitis has been 1.8 cases/1,000 patient-years of exposure, which is comparable to the background rate in diabetes. Three cases of pancreatic cancer have been reported on liraglutide – one in a patient treated on liraglutide for 152 days, one in a patient treated for seven days and was then diagnosed at stage 4 (suggesting it was not related to the seven days of liraglutide treatment); and one was reported prior to randomization. Similarly, she stated that no significant difference has been observed between exenatide and control groups in RCTs for exenatide. In RCTs for sitagliptin, the rate of pancreatitis was found to be 0.05 events/100 patient-years and 0.06 events/100 patient- years in the sitagliptin and comparator groups, respectively.
  • Looking forward, Dr. Aroda expressed optimism that the EMA’s Safety Evaluation of Adverse Reactions in Diabetes (SAFEGUARD) to assess the CV, cerebrovascular, renal, and pancreatic safety of currently marketed non-insulin glucose lowering agents will provide better quality clinical data – it will provide up to 240 million patient years of exposure.
  • During this presentation, Dr. Aroda also briefly addressed hypoglycemia, CV safety, and thyroid concerns for incretin-based therapies. As we have heard before, she stated that thyroid safety concerns stem from rodent models that do not match humans’ thyroid c-cell response to GLP-1 receptor agonism.

Questions and Answers

Q: That was a beautiful description done in a scholarly way. However if tomorrow morning you have a patient in front of you, and the patient is currently on an incretin and says, Doctor, I have heard a lot of things on television. What is your advice going to be with regard to that patient?

A: I guess I would flip that question to the audience and ask how many would say to stop the drug?

[No one raises hand]

A: How many would try to educate the patient on balancing risk and continue? [Scattered hands go up]

A: I am overwhelmingly surrounded by scholars much more scholarly than I. I think we need to spend

more time with patients educating them on what the comparative benefits and risks are.

Q: I agree with that answer


Q: Can you speak about levels of GLP-1 induced by surgery?

A: GLP-1 levels increase after bariatric surgery, and we have not seen an increase in pancreatic cancer. Obesity surgery is an area where you see decrease in these cancers.

Q: We can capture numbers of acute pancreatitis. My bigger question is, what about numbers of people with asymptomatic pancreatitis that we won’t know about for a long time? Studies may not show us that we have increased risk of pancreatic cancer until we have a whole lot of people already at greater risk. That’s my concern with short-term risk.

A: Do you mean chronic pancreatitis?

Q: Yes.

A: Some have asked whether it is worth it to look at enzyme fluctuations. There hasn’t been good correlation. All studies have been monitored by Data Safety Monitoring Boards, so if there were signals we would be alerted.

Corporate Symposium: Impacting Type 2 Diabetes and Optimizing Patient Outcomes with GLP-1 Agonists (Sponsored by Novo Nordisk)


Lawrence Blonde, MD (Ochsner Medical Center, New Orleans, LA)

Dr. Lawrence Blonde, co-author of the new AACE guidelines, maintained that minimizing the risk and magnitude of hypoglycemia and/or weight gain should be a high priority for patients with type 2 diabetes. He presented data demonstrating that hypoglycemia may be more common and have a more profound impact on patient outcomes than providers might expect: a patient survey found that at least 60% of type 2 diabetes patients experience at least one episode of non-severe hypoglycemia/month, with 35% experiencing episodes once-daily to once-weekly (Brod M et al., Value Health 2011); in addition, hypoglycemia is associated with increasing healthcare costs and reduced long-term survival in type 2 diabetes (Williams SA, et al., J Diab Complications 2012; Hsu et al., Diabetes Care 2013). Dr. Blonde presented GLP-1 agonists and DPP-4 inhibitors as effective options for avoiding hypoglycemia and weight gain, highlighting their prominence in recent algorithms (e.g., ADA/EASD 2012 and AACE 2013). He presented data demonstrating that adding a GLP-1 agonist to metformin provides better glucose control and greater weight loss than adding a DPP-4 inhibitor. In addition, he argued that GLP- 1 agonists may “prime” a patient for eventual progression to insulin (he noted that most patients with type 2 diabetes will end up needing insulin). In DeVries et al. (Diabetes Care 2012), adding basal insulin detemir when people failed to achieve A1c goals on metformin plus liraglutide provided a further 0.5% improvement in A1c without causing patients to regain the weight they had lost by adding liraglutide to metformin. He interpreted this to mean that optimizing one’s own endogenous glucose excretion prior to adding exogenous insulin may have a benefit.

  • Notably, Dr. Blonde referenced results from a poster he presented earlier today suggesting that A1c reduction derived from Bydureon was independent of weight lost. He emphasized that even if patients do not lose weight on GLP-1 agonist therapy, they can still experience tremendous glycemic benefits.


Michael Nauck, MD, PhD (Diabeteszentrum Bad Lauterberg, Harz, Germany)

To begin the symposium, Dr. Michael Nauck discussed the use of GLP-1 agonists in the treatment of early-stage type 2 diabetes. After discussing incretin mimetics’ mechanism of action, he began discussing the efficacy of GLP-1 agonists compared to placebo, metformin, and DPP-4 inhibitors. Compared to (or in addition to) the three alternatives, GLP-1 agonists led to equal or better glycemic control and weight loss. He added that some studies even suggest that GLP-1 agonists are slightly more effective than basal insulins and have better effects on weight. Dr. Nauck noted that the most recent ADA/EASD treatment algorithm facilitates shared decision-making, allowing healthcare providers narrow down the list of therapy options depending on patients’ preferences and limitations. He used a case study to illustrate that GLP-1 agonists are optimal for those seeking to avoid weight gain or hypoglycemia. Dr. Nauck confronted concerns over GLP-1 agonists and thyroid tumors, noting that concerns arose from data from rodent models, which process GLP-1 differently in the thyroid. He cited a year-long human trial that showed that high doses of liraglutide did not lead to a significant increase in thyroid tumors. Dr. Nauck also allayed concerns over incretins and pancreatitis. He ended by proposing that GLP-1 agonists may have potential in individuals with only impaired glucose tolerance (IGT), citing data that demonstrated that GLP-1 administration improved insulin response in patients with IGT

Questions and Answers

Q: Given the concerns you mentioned, is it still appropriate to use GLP-1 agonists and DPP- 4 inhibitors earlier in the course of diabetes?

A: Those familiar with these studies say that there is no need to change the clinical practice regarding these drugs, and that they should be used as they have been. Psychologically speaking, the only reason to stop therapy is if the patient feels uncomfortable about using them. Otherwise, I would not be comfortable recommending that patients go off these medications.

Dr. Vivian Fonseca: I’d like to remind audience that these agents are not approved for use in prediabetes.


Vivian Fonseca, MD (Tulane University Health Sciences Center, New Orleans, LA)

In his presentation, Dr. Vivian Fonseca discussed the treatment of well-established type 2 diabetes patients. He noted that many such patients are already on basal insulin, and that if more intensive therapy is needed, providers generally need to choose between initiating the patient on a prandial insulin or adding another agent. Dr. Fonseca noted that prandial insulins require more careful management, are less convenient, and come with an increased risk of hypoglycemia. In contrast, he suggested that incretin mimetics can be effective and safe when added to basal insulins. He argued that basal insulin and GLP-1 agonists have complementary actions, with the former contributing more to fasting and nocturnal control and the latter acting more postprandially. He discussed experimental findings showing that combined therapy with a basal insulin and GLP-1 agonist results in better glycemic control (especially postprandially) and more weight loss. Notably, some companies are starting to explore head to head comparisons between adding a GLP-1 agonist or prandial insulin on top of basal insulin (Dr. Fonseca mentioned GSK’s albiglutide). He added that DPP-4 inhibitor/basal insulin combination therapy also resulted in improvements, albeit with less efficacy and weight loss than the GLP-1/basal insulin combination. Dr. Fonseca concluded that the joint use of incretin-based therapy and basal insulin is a potent combination, able to significantly and safely improve glycemic control. Looking to the future, he expressed optimism about co-formulations of GLP-1 agonists and basal insulins. He cited an early study demonstrating that IDegLira (a combination of insulin degludec and liraglutide) reduced A1c and provided 2 kg more weight loss than the insulin component alone, and expressed hope that future studies will confirm these promising findings.


Michael Nauck, MD, PhD (Diabetes Zentrum bad Lauterberg, Harz, Germany) Lawrence Blonde, MD (Ochsner Medical Center, New Orleans, LA), Vivian Fonseca, MD (Tulane University, New Orleans, LA)

Dr. Fonseca: Monitoring lipase and amylase – do either of you recommend that in clinical practice?

Dr. Nauck: Not at all. We now know a significant proportion of patients with type 2 diabetes have elevated lipase if you just do a random blood sampling. That does not predict pancreatitis. One must know that the positive predictive value of an elevated lipase concentration is really restricted to the ER setting when severe abdominal pain appears. In that situation you can count on a sever lipase concentration to diagnose. In the asymptomatic patient, it doesn’t tell you anything

Q: Is GLP-1 expressed in other tissues in humans?

Dr. Nauck: It certainly is. Mainly the brain and nervous system are equipped with GLP-1 receptors. Some important metabolic organs apparently are not, such as the liver, muscle, and adipose tissue. There have been occasional reports of binding sites if you look at radioactively labeled material, but no GLP-1 receptor has ever been identified in this tissue.

Dr. Fonseca: Larry, there are a lot of outcomes trials going on, all looking at the same composite endpoint for MACE, etc. Should we be a bit more adventurous in this area? Are there novel endpoints we could explore? What else would you like to see, especially now that SAVOR TIMI reported a neutral outcome?

Dr. Blonde: I’m hopeful these trials will in fact show whether or not there is an increased risk of pancreatitis. Other safety endpoints including pancreatitis are being adjudicated or collected. As SAVOR is reported, we may learn about these data. If individual trials are not enough, then we’ll pool data. I think we will learn more about safety in a rapid period of time than we have before.

Dr. Fonseca: The ADA has asked for patient level data for all of these trials to be pooled, or else individual trials will never show a signal because it is so rare. If there is an increase, what will the level be? From the data we have now, we can’t tell.

Dr. Blonde: I agree completely. In these discussions we lose any discussion of benefits vs. risk. There is a risk to not treating a patient. The absolute risk in the Singh study was relatively low, whether or not it was done correctly.

Q: In clinical trials comparing GLP-1 receptor agonists and insulin, what was the mean insulin dose?

Dr. Fonseca: In most studies I recall around 25-30 units.

Q: Dr. Nauck do you know what predisposes somebody to respond well to treatment? Are there genetic markers to look at?

Dr. Nauck: There is a publication on a variant of the GLP-1 receptor that at least determines response to an experimental administration of GLP-1. It could well be that such polymorphisms exist, but there’s no test you can order so it’s something for the future. It’s not applicable today.

Dr. Blonde: And as I showed before, there’s more consistency for A1c reduction than weight loss, though in that study most people had a reduction in both A1c and weight.

Q: How should you consider these agents for children vs. adults?

Dr. Nauck: Small trials have reported, and it works. But safety considerations are different from treating adult populations, so you need a large database to judge safety. The interference of growth in adolescents means we don’t have the data right now.

Q: A point of clarification on the thyroid issue – have there been primate studies of C-cell proliferation?

Dr. Nauck: There is some discussion. Some groups use antibodies or ligands for GLP-1 receptors that have been found to be not specific by other groups. There may be erroneous data out there claiming the existence of GLP-1 receptors where there are none. We’ll have to see with better methods. Regarding thyroid and primates, we know Novo Nordisk has published a large study looking at the pancreas. I would be surprised if they have not looked at the thyroid as well.

Dr. Fonseca: You did mention prediabetes, though it is obviously not an approved use right now. Is a trial being done? What kind of endpoints would be used in such a trial?

Dr. Nauck: The kind of endpoint is under discussion. Is it just the prevention of diabetes? Is that sufficient to support clinical use of these agents? The answer is probably no. So you want so need long-term benefit to be proven. Also the bar is raised pretty high with regards to safety because basically then you would be treating “healthy” people.

Dr. Fonseca: But of course many of these people are obese, and you have an obesity trial ongoing.

Dr. Nauck: In the end it’s about the prevention of diseases that are associated with high glucose. So diabetic complications. You would have to do extremely large trials and observe patients from early stages of prediabetes to stages where you expect diabetes complications to draw conclusions that the event rate would certainly be much lower.

Corporate Symposium: GLP-1 in Focus: The Knowledge Challenge (Sponsored by Novo Nordisk)

Charles Reasner II, MD (University of Texas Health Science Center, San Antonio, TX); Melissa Magwire, RN, CDE (Shawnee Mission Medical Center, Shawnee, KS)

In this highly interactive corporate symposium, Dr. Charles Reasner and Ms. Melissa Magwire discussed the use of GLP-1 agonists for the treatment of type 2 diabetes in a game-show format, complete with Jeopardy music and the occasional audience spotlight. They focused on issues such as the incretin effect, differences between GLP-1 agonists and DPP-4 inhibitors, and the role of glucagon in the pathophysiology of type 2 diabetes. In the discussion of Novo Nordisk’s Victoza, they discussed clinical trials showing that Victoza was more effective than sitagliptin at reducing patient A1c levels and improving the proinsulin/insulin ratio, while also helping patients achieve greater weight loss compared to sitagliptin. They noted that nausea, Victoza’s most common side effect, generally decreases over time. Dr. Reasner attempted to alleviate certain safety concerns about Victoza therapy, such as the thyroid cancer tumor increase seen in rodent studies. He acknowledged that other concerns, such as the potential association between GLP-1 agonists and pancreatitis highlighted at the recent NIDDK summit, are still being studied. During the symposium, the audience of ~400 was divided into four teams that competed with each other – though our team (Team 4) had the initial lead, Team 1 ultimately won the day’s event. Given the changes in pharma marketing guidelines in recent years, Dr. Reasner congratulated the victors jokingly, “You win absolutely nothing!”


Melissa Magwire, RN, CDE (Shawnee Mission Medical Center, Shawnee, KS)

Ms. Melissa Magwire reviewed the incretin effect and the key defects associated with type 2 diabetes. She explained the incretin effect, noting that the insulin response is enhanced after an oral glucose load versus intravenous glucose (Nauck et al., Diabetologia 1985). Subsequently, Ms. Magwire highlighted eight core defects in type 2 diabetes – the infamous “ominous octet” – decreased insulin secretion, decreased incretin effect, increased glucagon, increased glucose absorption and production, decreased glucose uptake, neurotransmitter dysfunction, increased lipolysis and inflammation, and increased hepatic glucose production (DeFronzo et al., Diabetes 2009). In addition, Ms. Magwire emphasized that the brain is essential in glucose metabolism, as it processes information from neural, hormonal, and nutrient signals, and modulates glucose output in the liver and glucose uptake in peripheral tissues. Finally, she noted that excess weight contributes to the development of type 2 diabetes: 1) insulin resistance is related to abnormal partitioning of fat among adipose, hepatic, muscle, and pancreatic tissues; 2) secretion of excess free fatty acids and proinflammatory cytokines may damage beta cells; and 3) increased levels of free fatty acids decrease glucose uptake and tolerance.


Charles Reasner II, MD (University of Texas Health Sciences Center, San Antonio, TX); Melissa Magwire, RN, CDE (Shawnee Mission Medical Center, Shawnee, KS)

Dr. Charles Reasner and Ms. Melissa Magwire reviewed the role of glucagon in type 2 diabetes and the impact of GLP-1 on decreasing glucagon secretion. Whereas people with type 2 diabetes have a ~25-50% deficiency in beta cell mass compared to individuals without diabetes (we’d love to see an end, by the way to the “healthy individuals” moniker, since so many with diabetes are very healthy), they have the same alpha cell (glucagon-producing cells) mass as healthy individuals. Thus, people with type 2 diabetes have a relative surplus of glucagon. Dr. Reasner and Ms. Maguire presented data suggesting that this hyperglucagonemia in type 2 diabetes is responsible for about 50% of the excess postprandial increase in glucose levels (with lack of sufficient insulin accounting for the other 50%; Shah et al., J Clin Endocrin Metab 2000). Thus, they portrayed hyperglucagonemia as a significant underlying defect of type 2 diabetes. Finally, they discussed how GLP-1 addresses hyperglucagonemia by inhibiting glucagon secretion in a glucose-dependent manner: over a six-hour post-meal GLP-1 infusion, GLP-1 caused insulin to increase and glucagon to decrease until blood glucose levels reached about 100 mg/dl, at which point the insulin and glucagon secretion levels returned to baseline despite continued GLP-1 infusion (Nauck et al., Diabetologia 1993).


Melissa Magwire, RN, CDE (Shawnee Mission Medical Center, Shawnee, KS)

Ms. Melissa Magwire discussed the differences between GLP-1 and DPP-4, emphasizing the therapeutic effects GLP-1 agonism provides beyond DPP-4 inhibition. To start, Ms. Magwire highlighted that GLP-1 is a hormone that directly stimulates insulin release from the beta cell in a glucose-dependent manner, while DPP-4 is an enzyme responsible for degrading native GLP-1 in the body. She noted that both GLP- 1 and GIP are nutrient-stimulated intestinal incretin hormone, with GLP-1 having a much more robust effect on glucagon suppression than GIP (Vilsbøll et al., JCEM 2003; Nauck et al., JCEM 1993). She then emphasized that though the insulin response to physiologic levels of native GLP-1 is severely impaired in type 2 diabetes, pharmacologic levels can elicit a much more robust and improved response. To conclude, Ms. Magwire commented that though both physiologic GLP-1 levels for people without diabetes, and DPP-4 inhibition can bring about increased insulin secretion and decreased glucagon secretion, only GLP-1 receptor activation confers the full pharmacologic effects of delayed gastric emptying, increased satiety, decreased energy intake, decreased weight, and increased nausea.


Charles Reasner II, MD (University of Texas Health Science Center, San Antonio, TX)

In this section, the presenters focused on Novo Nordisk’s GLP-1 agonist Victoza (liraglutide). Dr. Reasner attempted to allay concerns about potentially increased thyroid tumor risk by noting that the association was only found in rodent studies. A number of differences between mice and humans exist that led Dr. Reasner to suggest that thyroid tumor risk may not be a concern for humans: the target for GLP-1 is the thyroid, not the pancreas, in mice; GLP-1 receptor stimulation causes release of calcitonin, which rodents need for calcium absorption and without which they will die; humans do not need calcitonin to absorb calcium, and don’t exhibit an increase in calcitonin when administered GLP-1 agonists. Dr. Reasner did acknowledge that the potential connection between GLP-1 agonists and pancreatitis will likely come up frequently during the conference. He echoed the sentiment that no causal relationship has been established. The presenters then discussed the positive results of Victoza clinical trials, namely that Victoza is twice as effective as sitagliptin (100 mg) in helping type 2 diabetes patients achieve A1c levels under 7% (the ADA recommendation), and that it leads to weight loss (approximately 6 lbs lost compared to 2 lbs lost on sitagliptin). Of note, Victoza was able to reduce glucose with a concomitant decrease in the ratio of proinsulin to insulin. A high proinsulin/insulin ratio is indicative of pancreatic stress and suggests that the pancreas is producing more proinsulin than it has the capacity to appropriately cleave. Thus, a decrease in the proinsulin/insulin ratio suggests that Victoza’s glycemic effects are not solely attributable to increasing insulin production. Finally, the presenters also noted that the nausea associated with Victoza use decreases with time; we would also point out that it is easier for patients to manage, albeit off-label, since there are many doses one can take outside the recommended three.


Melissa Magwire, RN, CDE (Shawnee Mission Medical Center, Shawnee, KS)

During this portion of the symposium, Ms. Melissa Magwire encouraged audience members to change the injection conversation with their patients (highlighting Victoza’s user-friendly pen), emphasized that there are generally no dose adjustment requirements for special populations, and stressed that Victoza can be taken any time of day. Subsequently, she provided a brief overview of the ADA/EASD position statement and AACE diabetes management algorithm, noting that incretin-based therapies, in particular GLP-1 agonists, are the preferred second-line option following metformin in the AACE algorithm.

  • Ms. Magwire encouraged audience members to change the injection conversation, noting that most patients surveyed are open to self-injection when provided an explanation of the clinical benefits. In an online market research survey of US adults with type 2 diabetes (n=797), 80% were open to self-injection (data on file at Novo Nordisk, 2011). In a randomized, multicenter, open-label single visit study, 90% of participants (n=90) considered the Victoza pen easy to learn how to use (data on file at Novo Nordisk). In addition, Ms. Magwire emphasized that most experience little or no pain when using a 32-gauge, 6 mm needle for Victoza injections. She noted that the Victoza pen can be used with NovoTwist needles, which are designed with user-friendly features: a “Just Twist” injection needle, quick and easy attachment, and an audible and tactile confirmation click to signal correct attachment.


Charles Reasner II, MD (University of Texas Health Sciences Center, San Antonio, TX); Melissa Magwire, RN, CDE (Shawnee Mission Medical Center, Shawnee, KS)

Q: Have you observed decreased nausea when taken right before bedtime?

Ms. Magwire: Studies didn’t look at timing, but if you think about the mechanism of action, overeating with delayed gastric emptying may exacerbate nausea, so we tend in our practice to dose at bedtime to minimize that. You can educate patients, if they dose first thing in morning, to listen to their bodies and not overeat. That will help with nausea. If you know someone has a tendency to overeat, potentially bedtime is a better choice.

Q: Are there GLP-1 receptors on alpha cells, or is glucagon suppression an indirect effect?

Dr. Reasner: You can find both schools of thought. Some feel that there are indirect effects, while others think there are receptors on the alpha cell.

Q: Can you use DPP-4 and GLP-1 together?

Ms. Magwire: If you think about the mechanism of action, it doesn’t make a lot of sense. You don’t get that much additional response from adding a DPP-4 inhibitor. If you add on the GLP-1 agonist, there’s other things like appetite suppression, potential decrease in weight, and a satiety effect – you’ll get more of that with a GLP-1 so it doesn’t make a lot of sense to use the two in combination.

Q: Can you talk about the latest safety issues with incretins and DPP-4 inhibitors?

Dr. Reasner: My understanding of the safety issues revolves mainly around pancreatitis and pancreatic cancer. If you look at clinical trials with Victoza, there are more patients that develop pancreatitis with Victoza than with placebo. However, the number of cases did not exceed what you would have predicted with historical controls. The increase is small. I think the big concern recently has been over a study done by Dr. Peter Butler. He is a pathologist who specializes in studying pancreases. He looked at 34 individuals who died of various causes, and of these, 20 had diabetes. Of the 20 who had diabetes, seven were on sitagliptin, and one was on Byetta. When he looked at the histology of patients with diabetes on incretin-based therapies, he found evidence of increased inflammation. It is not clear to me how extensive the inflammation was, and we don’t have any data on the patients while they were alive complaining about pain. I don’t know if they had other causes of pancreatitis. Also, there was evidence of hyperplasia in some of the samples. Some suggest that it is a good thing, whereas others feel that it may be a precursor to cancer. There is also evidence of ductal metaplasia. Some think it was a common finding, and some think it could be a precursor of cancer. There was no cancer detected, but there were suggestions that hyperplasia and metaplasia may be precursors. This has gotten a lot of interest. The FDA has become interested, and there was an NIDDK symposium last week (editor’s note – see our coverage at and I’m sure we’ll hear more about this in the future.

Q: What about the combination of premixed insulin and Victoza?

Ms. Magwire: Currently the package insert is for combination with basal insulin only. At this point it’s not an indication. We can’t comment more than that.

Q: Do sulfonylureas increase glucagon levels?

Dr. Reasner: Not to my knowledge. If anything it should suppress glucagon. Anything that increases insulin should decrease glucagon levels.

Q: What is the main defect in type 2 diabetes – a decrease in production of GLP-1, or GLP-1 resistance?

Dr. Reasner: In an early study, it was shown that those with type 2 diabetes had a lower post-meal GLP-1 response than those without diabetes. Initially the thinking was that once you have diabetes, you have a decrease in GLP-1 secretion, which was likely to track with beta cell function (just like we see a decrease in insulin secretion). In more recent studies, it has been shown that most people with type 2 diabetes have normal levels of GLP-1. If you measure production and it’s normal, and it’s binding to its receptors normally, but the insulin response is abnormal, then there is a post-receptor defect in the pancreatic beta cell. You don’t get the insulin secretion you would expect. People with type 2 diabetes are resistant to

GLP-1 the same way they are resistant to insulin.

Q: Comparative difference between GLP-1 and DPP-4 inhibitors on glucagon?

Dr. Reasner: I think the best study that was done was by Michael Nauck. Glucagon suppression at physiologic levels of GLP-1 it was about 38% and at a pharmacologic level it was about 55%. So that’s pretty similar to DPP-4 inhibitors vs. GLP-1 agonists.

Product Theater


Eugenio Cersosimo, MD, PhD (University of Texas Health Science Center, San Antonio, TX)

In this lightly attended product theater, Dr. Eugenio Cersosimo reviewed the efficacy, safety, dosing, and administration of Nesina (alogliptin), Kazano (alogliptin/metformin), and Oseni (alogliptin/pioglitazone). He emphasized that alogliptin is effective as monotherapy, in dual and triple therapy, and in fixed-dose combinations, and that Nesina, Kazano, and Oseni all address multiple defects of type 2 diabetes. In addition, he highlighted that the Takeda Diabetes Advantage Program offers support programs for patients; eligible patients who are enrolled in the program pay no more than $4 per month for up to 12 months for Nesina, Kazano, or Oseni.

-- By Eric Chang, John Close, Poonam Daryani, Jessica Dong, Benjamin Kozak, Nina Ran, Phaedra Randolph, Joseph Shivers, Tony Thaweethai, Manu Venkat, Katrina Verbrugge, Vincent Wu, and Kelly Close