American Diabetes Association 75th Scientific Sessions

June 5-9, 2015; Boston, MA; Day #4 Highlights - Draft

Executive Highlights

We’re in the middle of a big afternoon in Boston, having just seen the presentation of primary results from the TECOS and ELIXA cardiovascular outcomes trials (CVOTs) for Merck’s Januvia (sitagliptin) and Sanofi’s Lyxumia (lixisenatide), respectively. Below are our high-level takeaways from those presentations as well as a few selected highlights from earlier in the day.

1. TECOS full results showed neutrality for cardiovascular outcomes, including a spot-on 1.00 hazard ratio for heart failure; there was only a slight numerical imbalance in acute pancreatitis as a side finding. (Hundreds of hours on pancreatitis this and pancreatic cancer that, reduced to noise …)

2. No major surprises from ELIXA full results either: results showed CV neutrality, modest weight and blood pressure benefits, and comparable safety/tolerability with previous trials.

3. During a press conference, TECOS and ELIXA presenters touched upon some big-picture issues and questions regarding CVOTs, including the possibility of long-term follow up and the need for longer trials with active comparators. While some may see this as a positive, we see this as very surprising and disappointing that there is not at least passive follow up for all long-term outcomes trials – don’t we want to know, particularly after these drugs go generic, what the long-term impact is on heart and kidney disease? How much would this information be worth?

4. Intarcia’s exenatide mini-pump ITCA-650 led to a mean 1.4% A1c reduction from baseline (8.4%) among completers in the pivotal FREEDOM-1 trial, including a 1.7% mean reduction in the subgroup of patients not on a sulfonylurea; this was in addition to 2 kg (4.4 pounds) placebo-adjusted weight loss.

5. A late-breaking poster from Dexcom has demonstrated that the real-world incidence of hypoglycemia (glucose < 70mg/dl) is frequent and that consequent overtreatment (glucose > 180 mg/dl) is not at all uncommon.

6. Dr. Jun Yang (Lilly, Shanghai, China) presented a study showing superiority of Lilly’s once-weekly Trulicity (dulaglutide) over Sanofi’s Lantus (insulin glargine) in type 2 diabetes patients predominantly in Asia.

Top Highlights

1. TECOS Full Results

The headline results from TECOS, the CVOT for Merck’s Januvia (sitagliptin), featured no major surprises. (We note that this information is based on results verbally disclosed at a press conference – full results coverage will be coming your way soon).  The hazard ratio for the primary composite cardiovascular endpoint (MACE + hospitalization for unstable angina) was 0.98 (95% CI: 0.89-1.08). The secondary (more inclusive) cardiovascular composite was also neutral (HR: 0.99; 95% CI: 0.89-1.11). Of course, all eyes were on the heart failure results to provide more clarity on the question of a DPP-4 inhibitor class effect. The answer was resoundingly reassuring, with an almost impressively neutral hazard ratio of 1.00 (95% CI: 0.83-1.20) – notably the upper bound of the 95% confidence interval was 1.20, which is below the point estimate of 1.27 for hospitalization for heart in the SAVOR trial. Overall, these results from the fairly large and powerful trial were reassuring and a win for Merck and patients taking sitagliptin.

  • One of the few areas of non-neutrality in the trial was pancreatic adverse events, although the imbalances went both ways. Based on a small number of events, there were more cases of acute pancreatitis with sitagliptin. However, based on an even smaller number of events, there was an imbalance in pancreatic cancer in sitagliptin’s favor. The investigators did not seem particularly concerned, and we are inclined to agree with them.
  • The fact that Januvia emerged from TECOS unblemished should help the product retain its leadership in the DPP-4 inhibitor class, with AZ’s Onglyza (saxagliptin) and – to a lesser extent – Takeda’s Nesina (alogliptin) blemished by the heart failure question. However we see the neutral TECOS heart failure results as good for the class as a whole, and it increasingly looks like the signal in SAVOR was a matter of chance (like we noted with pancreatitis and GLP-1 – more noise).  In the context of a fairly homogeneous class of drugs, however, prescribers may choose to play it safe and go with Januvia. The commentary on heart failure in EXAMINE continues – during a press conference, TECOS investigator Dr. Eric Peterson (Duke Clinical Research Institute, Durham, NC) did cite the non-significant increase in hospitalization for heart failure with Takeda’s Nesina (alogliptin) and suggested that doctors will vote with their feet.

2. ELIXA Full Results

Results from the ELIXA CVOT for Sanofi’s Lyxumia (lixisenatide) were also resoundingly neutral. The hazard ratio for MACE was 1.02 (95% CI: 0.89-1.17), and the hazard ratios were also neutral for the expanded composite (MACE + unstable angina; HR 1.02 [0.89-1.17]), heart failure hospitalization (HR: 0.96 [0.75-1.23]), and all-cause death (HR: 0.94 [0.78-1.13]). No subgroup analyses deviated from neutrality for the primary MACE+ outcome. The Kaplan-Meier curves were fairly superimposable: reassuring for safety and study power, but with not even a hint of cardioprotection emerging towards the end of the trial. There were very modest average reductions in weight (0.7 kg [~2 lbs]) and systolic blood pressure (0.8 mmHg) with Lyxumia, with no effect on heart rate for most of the trial following a transitory 1 BPM increase over the first few weeks. Interestingly there was a slight attenuation in the rise in albuminuria over time (24% increase over two years with lixisenatide vs 34% with placebo; p<0.01), although it was far from a conclusive signal and there was no significant change in eGFR. Nausea and vomiting were increased roughly 3x but comparable with other lixisenatide trials, with discontinuation in just under 5% of lixisenatide patients due to GI side effects – this drove a very modest imbalance in overall adverse events and events leading to discontinuation, though severe adverse events were balanced. We were glad to see that there was no increase in pancreatitis or pancreatic cancer (neoplasms were actually imbalanced in lixisenatide’s favor).

3. Reflecting on CVOTs: A Missed Opportunity?

Commentary from trial investigators touched upon some of the big-picture issues and questions regarding CVOTs. Both sets of presenters mentioned that their respective trials were powered for superiority even though they were designed for safety, and ELIXA investigator Dr. Marc Pfeffer (Brigham and Women’s Hospital, Boston, MA) did not seem optimistic about a glucose-independent macrovascular benefit with Lyxumia given the neutral ELIXA results. However, discussant Dr. Silvio Inzucchi (Yale University, New Haven, CT) emphasized that short term CVD neutrality does not preclude a longer-term benefit, and that CVOTs as currently designed are likely to show MACE neutrality but are unlikely to show benefit. He directly posed the question of whether CVOTs are long enough and whether the trials as currently designed are answering all the right questions.

  • On the TECOS side, Dr. Rury Holman (University of Oxford, UK) acknowledged in a press conference that he would love for CVOTs to last longer and compare multiple therapies to each other. He expressed slight disappointment that there was not funding to allow for passive follow up of the study cohort (this was the case for ELIXA as well) – to us this is a bit of a missed opportunity to examine benefits that may take more time to emerge.
  • The presenters went to great lengths to emphasize that their respective trials were not glucose lowering studies. Consistent with previous CVOTs, the A1c delta between groups in both trials was only around 0.3%. Investigators encouraged a whole-body mentality and consideration of the known value of reducing A1c on microvascular complications. However given the questions we heard at the press conference, we worry that some may try to frame these results as failures for the respective drugs rather than valuable sources of reassurance on safety. Media re-education is needed.

4. Long-Awaited FREEDOM-1 Data on Intarcia’s ITCA-650: Impressive Efficacy

Results from Intarcia’s pivotal phase 3 FREEDOM-1 trial for the implantable exenatide mini-pump ITCA-650 were fairly strong, and even better in a pre-specified subgroup analysis. In FREEDOM-1 (n=460), the mean A1c reduction for the overall ITCA-650 group was 1.4% at the end of the trial (mITT analysis), and from the graph we would expect the placebo-adjusted efficacy to be around 1.1%. It is hard to draw fair comparisons vs. efficacy seen in other GLP-1 agonist phase 3 trials due to differences in study enrollment and design - we'll look forward to learning from Intarcia's impressive planned slate of head-to-head trials. Notably, a pre-specified secondary analysis of efficacy by concomitant therapy found that patients not on a sulfonylurea did far better in terms of A1c reduction from baseline to week 39 (1.7% from baseline; we estimate ~1.4% placebo-adjusted) than those on a sulfonylurea (1.2% from baseline; we estimate ~0.9% placebo-adjusted). It looked to us like the SFU group was beginning to see the deterioration of glycemic control that is characteristic of the drug class, with an upswing in A1c towards the end of the trial. Especially because it was prespecified, we see these stronger results in non-SFU-taking patients (likely to be earlier-stage diabetes patients on average) as pretty persuasive. In terms of efficacy, the higher 60 mcg/day dose appeared to barely beat out the 40 mcg/day dose without much of a safety/tolerability sacrifice. ITCA-650 drove 2% placebo-adjusted weight loss from baseline at the higher dose, and from the abstract it looks like weight was still decreasing at the end of the trial. Composite outcome analyses showed that ITCA-650 helped more patients achieve a 1% or greater A1c reduction plus a >5% decrease in weight (21% and 17% with the two ITCA-650 doses vs. 6% with placebo).

  • There was no hypoglycemia data presented, but we learned later that rates were less than 10% across the three groups and were higher in the sulfonylurea patients – no surprise there. We look forward to seeing full hypoglycemia results whenever they are presented or published.
  • ITCA-650 was associated with GI tolerability issues for some patients, although most of the risk was right around initiation and dose increases. Around 30% of patients on ITCA-650 40 mcg/day and 31% of patients on the 60 mcg/day dose experienced nausea vs. 10% on placebo; 19%-24% of ITCA-650 patients experienced vomiting vs. 2% on placebo. These issues were the primary drivers behind an imbalance in adverse events leading to discontinuation between the ITCA-650 group (9%-12%) vs. placebo (4%) – keep in mind that “discontinuation” means device removal in this case, not just the cessation of taking a pill or injection. Providing some reassurance, presenter Dr. Julio Rosenstock (Dallas Diabetes and Endocrine Center, Dallas TX) did display data indicating that the vast majority of nausea cases occurred during initiation or dose escalation, returning to much lower background levels (<5%) afterwards. This fact should make the issue manageable, though we wish there was some way to more gradually step up the dose as can be done with some injectable GLP-1 agonists.
  • Importantly for the patient experience, administration site events (bruising, irritation) were generally mild and transient. Intarcia has gone to great lengths to focus on the patient experience, including totally reworking the insertion procedure to make it much quicker and easier, a push that appears to have paid off.

5. Dexcom Poster: Share Data on Hypoglycemia

A late-breaking poster from Dexcom has demonstrated that the real-world incidence of hypoglycemia (glucose < 70mg/dl) is frequent and that consequent overtreatment (glucose > 180 mg/dl) is not at all uncommon. The retrospective study analyzed 50,000 hypoglycemic events identified in 1,177 users of the company’s SHARE platform and found a striking hypoglycemia rate of 0.96 event/day/patient. Of more concern, the proportion of hypoglycemic events followed by rebound hyperglycemia was 18% within 60 minutes and 26% within 90 minutes. Talk about glycemic variability! It certainly makes us look forward to the results of FLAT-SUGAR (coming tomorrow! Irl Hirsch in the last presentation of the day – very smart, ADA!) to know that at least one-fourth of hypoglycemic events in this real-world population were followed by hyperglycemia. Notably, rebound hyperglycemia occurred more frequently during the day than the night (p < 0.001), which may be partially related to meals. Ultimately, the findings are a strong reminder of the value of cloud-based data platforms in facilitating population tracking and broader learnings about real-world diabetes management.

6. Trulicity vs. Lantus in Asian Patients

Dr. Jun Yang (Lilly, Shanghai, China) presented a study showing superiority of Lilly’s once-weekly Trulicity (dulaglutide) over Lantus (insulin glargine) in type 2 diabetes patients predominantly in Asia. In a 26-week study of predominantly Asian participants with type 2 diabetes (n=165), participants were randomized to either Trulicity or Lantus in combination with metformin and/or a sulfonylurea. The findings showed that Trulicity led to greater reductions in A1c with weight loss and less hypoglycemia. Specifically, A1c reductions from a baseline A1c of 8.4% were 1.7% for the 1.5 mg dose of Trulicity, 1.3% for the 0.75 mg dose of Trulicity, and 1.2% for Lantus. The proportion of patients who reached target A1c levels (<7%) were 65%, 54%, and 41% for Trulicity 1.5 mg, Trulicity 0.75 mg, and Lantus, respectively. In addition, Trulicity was associated with greater weight loss from a mean baseline of 78 kg, with the 1.5 mg and 0.75 mg doses losing a mean of 1.5 kg (~3 lbs) and 0.9 kg (~2 lbs), respectively; the Lantus group, on the other hand, gained 1.0 kg (~2 lbs). On the safety front, Trulicity appeared to have a lower hypoglycemia risk, with the proportion of patients with hypoglycemia being 18% for the Trulicity doses vs. 29% for Lantus (this would have been even lower without the SFU usage). Regarding adverse events, Trulicity was relatively well-tolerated, with the exception of having more GI side effects compared to Lantus – 15% of Trulicity 1.5 mg and 8% of Trulicity 0.75 mg experienced diarrhea compared to 2% of Lantus. Lilly has worked hard to compare Trulicity to Lantus in a variety of applications (see AWARD-2 and AWARD-4), and the benefits definitely appear to be there for A1c, hypoglycemia, and weight.


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