The American Diabetes Association (ADA) 77th Scientific Sessions was held in San Diego, CA from 9–13 June 2017. The highlight of this year's conference was back-to-back presentations from cardiovascular outcomes trials (CVOT) of diabetes drugs: previously unreleased data from SGLT-2 inhibitor Invokana's (JNJ, DSKYF, Mitsubishi, PDLI) CANVAS program and first conference details from long- acting insulin Tresiba's (NVO) DEVOTE. While Invokana’s data confirmed a CV benefit for its class, questions were raised about a smaller benefit on CV death than seen in Jardiance’s (Boehringer, LLY) EMPA-REG, as well as an increase in amputations. We provide initial results from our survey of primary care physicians, as well as key opinion leader (KOL) comments. Tresiba showed a substantial reduction in severe hypoglycemia compared to Lantus (SNY), but only non-inferiority on CV outcomes, with a slight numerical advantage on major adverse cardiovascular events (MACE). However, increased price competition will come from additional biosimilars to Lantus, though timing depends on patent litigation: first Phase III data presented at the conference for MYL-1501D (Mylan, Biocon) were reassuringly unexciting, showing it is indeed equivalent. There was also first data from Gan & Lee’s biosimilar. Other sessions at the conference had new data from previously released CVOTs, as well as comments on physician inertia in using drugs with CV benefits and the need for guideline changes. First conference presentations of Phase III studies for SGLT-1/2 inhibitor sotagliflozin (SNY, LXRX) in type 1 diabetes did not resolve questions about approval and subsequent use, due to modest efficacy and the side effect of diabetes ketoacidosis (DKA). Endocrinologists had mixed views, although there is already off-label use of the SGLT-2 inhibitor class in these patients. First data for additional Phase III trials of SGLT-2 inhibitor ertugliflozin (MRK, PFE), along with new extension study results, reinforced its lack of differentiation from currently marketed SGLT-2 inhibitors. Its one potential competitive advantage is a fixed-dose combination (FDC) with DPP-IV Januvia (MRK), but our analysis of insurance plans shows the first such FDC is still struggling with reimbursement. Using a combination of SGLT-2 inhibitor Farxiga and DPP-IV inhibitor Onglyza (AZN, BMY) only showed modest benefits over Januvia, raising questions about the approach, but the investigator still argued in favor of potential benefits from early combination use. We include a comment on fixed-dose/ratio combinations based on a special session at the conference. There were also first data or first conference presentations of new trials from a number of other drugs, including SAR342434 (SNY), a biosimilar to mealtime insulin Humalog; ultra-rapid mealtime insulin LY900014 (LLY); oral GLP-1 agonist TTP273 (VTVT); RG7992 (Roche), a bispecific agonist antibody to FGF-R1/KlothoB; activin receptor inhibitor bimagrumab (NVS, MorphoSys); and MetAP2 inhibitor ZGN-1061 (ZFGN). Shortly after the conference, a US Food and Drug Administration (FDA) advisory committee voted overwhelmingly in favor of adding a CV indication for NVO’s GLP-1 agonist Victoza, which could help slow share loss to Trulicity (LLY, Sumitomo Dainippon). We highlight these and other presentations below. The survey mentioned in the report will be available in BMT's special reports section in the future, available free of charge to KOL Insight subscribers. SGLT Inhibitors

• Positive efficacy results from an integrated analysis of Invokana’s two CANVAS CVOTs helped confirm a CV benefit for the class, first seen with Jardiance’s EMPA-REG.
  • The reduction in the MACE primary endpoint was similar to EMPA-REG, but there was a much lesser reduction in CV death.
    • Some KOLs attributed the difference in CV death to the inclusion of patients without CV disease or to chance, though others would not rule out a difference between the drugs.
    • The difference could lead to questions about the mechanism for the CV benefit with the class and which patients are most suitable for the drugs.
  • There was no benefit in the subgroup without CV disease, which may consequently be excluded from a label indication.
    • There are other questions that could come up as the FDA picks through the data. For example, due to unblinding of data for the initial Invokana approval, the statistical plan was to test mortality endpoints after a cutoff date, and in that analysis, the hazard ratio for CV death was closer to unity.
  • On safety, the CANVAS program found a doubling of amputations, which shortly before the conference led to a confirmatory public health advisory from the FDA and, a couple months before, EMA warnings to the labels of all the SGLT-2 inhibitors, noting that a class effect cannot be excluded.
    • Most cases were in patients with a prior amputation.
  • Endocrinologists we spoke with at the conference were not too concerned about the amputation issue, and KOLs did not generally make too much of the difference in CV death between Invokana and Jardiance.
    • On the other hand, a survey we conducted of US primary care physicians suggested that Jardiance will continue to make substantial gains, though Invokana and Farxiga should see some growth as well, pending results from Farxiga’s CVOT.
    • Several speakers decried physician inertia in adopting diabetes drugs with a CV benefit and urged the ADA to adopt stronger recommendations, perhaps with a position statement on treating patients with CV disease.
  • Farxiga’s DECLARE-TIMI 58 will have a higher percent of high-risk patients without CV disease to help answer the question on the CV benefit of these drugs in a primary prevention subgroup, as well as the impact of risk groups on the magnitude of the benefit for CV death. Invokana's CREDENCE trial in diabetic nephropathy patients also includes CV death in its primary endpoint, which will provide another view on the topic.
  • Confirmation of the heart failure and renal benefits in CANVAS was also quite encouraging for the class, but more definitive data on that will come from dedicated heart failure and renal trials.
    • During the meeting, Boehringer and Lilly finally also announced plans for their own renal trial with Jardiance.
• First conference presentations for several Phase III sotagliflozin studies in type 1 diabetes demonstrated a modest glycemic benefit, and the focus of presenters seemed to be on the benefit for other unmet needs, which may become the focus of commercialization efforts if the drug is approved.
  • Approval in type 1 diabetes, and physicians’ subsequent decision on whether to prescribe the drug, will depend on Lexicon’s and Sanofi’s ability to convince physicians and regulators that the benefits outweigh the risk of DKA.
    • A number of physicians we queried at the conference did not feel that the reductions in A1C, while clinically significant, were substantial enough to outweigh the increased risk of DKA or that the benefit could be achieved by proper use of insulin.
    • Others had a more positive outlook, due to benefits beyond glycemic control. Based on Datamonitor Healthcare's proprietary survey, SGLT-2 inhibitors are actually already used off-label in 2–6% of drug-treated type 1 patients in developed markets.
• New data for ertugliflozin, from new Phase III trials and extension studies, confirmed previous findings, suggesting that ertugliflozin has an easy path to FDA approval, but not much differentiation to help it against established competitors.
  • Ertugliflozin’s one competitive advantage is the potential for fixed-dose combinations (FDC) with market leading DPP-IV inhibitor Januvia (Merck). However, per an analysis of US insurance plans, such branded combinations are having difficulty with reimbursement (in Europe, they are still being assessed), though it remains to be seen how competition will impact that.
    • We include a comment on fixed-dose/ratio combinations based on a special session at the conference.
• First data from a study of Farxiga plus Onglyza (which make up the FDC Qtern [AZN, BMY]) versus Januvia, in patients failing metformin with a high baseline A1C, only showed a modest difference in A1C or those reaching goal, raising questions about the approach.
  • Nevertheless, proponents believe earlier control with combination therapy can reduce comorbidities long term, though there is not a great deal of data to show that yet.

GLP-1 Agonists

• Panelists in an FDA advisory committee meeting on Victoza’s (NVO) LEADER CVOT, held just a week after the conference ended, voted overwhelmingly for adding an indication of reducing the risk of MACE in type 2 diabetics.
  • Most panelists, however, felt the indication should be restricted to those with the highest risk, namely CV disease and chronic kidney disease, since a subgroup analysis of others did not show a benefit. The FDA is likely to follow this advice.
  • Panelists had reservations about the lack of a benefit in the US subgroup, but testimony from an outside expert changed the tone of the discussion and appeared to assuage a number of panelists.
  • The label addition should help slow share loss to Trulicity, pending data from the latter's CVOT. The trial has a high percent of patients without CV disease, so it will be interesting to see whether that impacts results.
• A new analysis of semaglutide’s (NVO) SUSTAIN 6 CV safety study showed that the increase in the side effect of retinopathy seen in the semaglutide group only occurred in those with prior retinopathy and, in both the semaglutide and placebo groups, it occurred more often in patients with greater A1C drops. This suggested it was related to large A1C drops rather than the drug.
  • Semaglutide did appear to have a slightly higher rate of retinopathy than placebo in patients with lesser A1C changes, but the chair of the session felt one could not make too much of the small difference between the groups, and felt reassured by the data.
• First conference data from oral GLP-1 agonist TTP273’s (VTVT) Phase IIb LOGRA study showed an A1C reduction that was reasonably good for an oral drug, but numerically not as strong as would be expected for a long-acting injectable GLP-1 agonist. The results are also likely less than what will be seen for the two highest doses of NVO's oral semaglutide being tested in Phase III.
  • A reverse dose response complicates interpretation and conclusions on what the best dose is. A poster on potential reasons for such a response was fairly hypothetical.
  • TTP273 also showed only modest weight loss, but on the positive side, unlike other GLP-1 agonists, did not increase nausea or vomiting. More details are also needed on a case of elevated liver enzymes.
  • The company will not be taking the drug forward themselves, but will work on preclinical studies to gain more clarity on the reverse dose response. Without a better understanding, prospects for the drug appear weak.
• In a trial of diabetics with moderate to severe chronic kidney disease, Trulicity was associated with a smaller eGFR decline compared to insulin glargine over 26 weeks, albeit mainly in those with a high urine albumin to creatinine ratio (UACR).
  • This reinforced an emerging theme at the meeting that GLP-1 agonists may be reno-protective in some patients, though the evidence has not been as strong as for the SGLT-2 inhibitors. More data from the Trulicity trial, however, will be presented at the American Society of Nephrology meeting.

Insulins

• In details from the DEVOTE CVOT, long-acting basal insulin Tresiba (NVO) reduced the risk of severe hypoglycemia in type 2 diabetes by 40% compared to Sanofi's Lantus.
  • While severe hypoglycemia is associated with worse outcomes, Tresiba only showed non-inferiority on cardiovascular events, though with a slight numerical advantage.
    • The absolute difference in severe hypoglycemia was not large, limiting the potential impact on MACE. There will be more data at the European Association for the Study of Diabetes (EASD) meeting on the relationship.
  • Physicians we spoke with at the conference (albeit before updated DEVOTE results were released at the end of the conference) had mixed views on Tresiba.
    • Some were impressed with topline results from DEVOTE and wanted to prescribe Tresiba as a first-line option in type 1 diabetics and in type 2 diabetics at high risk for hypoglycemia, but were often hindered by insurance coverage. Others felt the longer-acting insulins only offer a modest benefit, not enough to get excited about.
  • An actual label indication for less hypoglycemia, which NVO is seeking, is difficult to predict, as there is not much precedence. However, they may be able to secure data from the trial on the label.
  • Particularly as more biosimilar glargine competition enters the market, convincing insurers that a benefit on hypoglycemia should allow preferential formulary placement may not be easy without substantial rebates.
• Phase III data from Mylan/Biocon’s biosimilar insulin glargine were reassuringly unexciting – without major differences from Lantus (glargine) in type 1 or 2 patients. Phase II data from Gan & Lee’s candidate in type 1 patients showed bioequivalent PK, but did not meet criteria for one of the pharmacodynamic endpoints, though presenters attributed this to the small trial size.
  • Datamonitor Healthcare expects MYL-1501D to be priced at a greater discount than LLY’s Basaglar, significantly eroding Lantus revenue, though timing may depend on patent litigation.
    • After the conference, Merck's insulin glargine candidate, was tentatively approved in the US (where such insulins are approved under the 505(b) (2) regulatory pathway), but due to patent litigation it cannot launch until February 2019 (unless the case is decide before).
      • Mylan just recently filed for an inter partes review of a couple Sanofi patents with the US patent office.
    • Of note, most US endocrinologists we talked to at the conference did not care which company was making the biosimilar or which country, as long as it was approved by the FDA.
  • We also comment on cost-effectiveness studies, more payor considerations, and a session on the rising cost of insulin.
• Sanofi’s SAR342434, a biosimilar to LLY’s mealtime insulin Humalog (lispro), likewise showed equivalent results in Phase III studies.
  • A small pilot pump study in type 1 diabetes patients did not find a statistically significant difference from Humalog in infusion set occlusions, but the rate of failures was notably numerically higher for SAR342434. A larger study will be needed for more definitive results.
  • Datamonitor Healthcare expects the launch of biosimilar insulin lispro to significantly slow the growth of the fast-acting insulin market, though dynamics are different than in the basal segment.
• In PK/PD studies, LLY’s more rapid-acting formulation of insulin lispro, LY900014, which uses treprostinil and citrate as excipients, appeared to have as rapid an effect as seen with NVO’s FIAsp and Adocia’s BioChaperone lispro (the latter also had new data at the conference), but a slower drop off subsequently – with the caveat that it is quite difficult to compare across such trials.
  • LLY had terminated its license agreement for Adocia’s product earlier in the year in favor of pursuing LY900014.
  • It is not clear if the differing profile will lead to clinical differences, but the LLY presenter felt there could be better control without a substantial increase in hypoglycemia. On the other hand, the Adocia presenter noted their product could have an advantage for insulin pump use.
  • While treprostinil, a pulmonary arterial hypertension drug, can lower blood pressure, the presenter said concentrations used were not strong enough to do this, and there was no increase in injection site reactions or pain.
  • Both Adocia and Lilly are hoping to avoid having to meet requirements for the FDA’s CV safety meta-analysis.

Other Mechanisms

• Weight loss for obesity drug ZGN-1061 (ZFGN) at the four-week time point in a Phase I multiple ascending dose study was numerically slightly lower than what had been seen with beloranib, Zafgen's first generation drug with side-effect issues.
  • However, there are signs that the newer drug may have a stronger safety profile. The investigator acknowledged, though, that they do not yet fully understand the mechanism for how ZGN-1061 could still be efficacious yet avoid the side effects. Larger clinical studies are needed to be confident of safety.
• NVS is taking a novel approach with bimagrumab, which increases lean body mass and potentially energy expenditure, thereby reducing insulin resistance. A Phase I study showed signs of decreased fat mass and insulin resistance, though overall weight did not change. More data will come from an ongoing Phase II trial in obese diabetics.
• Hyundai sought to demonstrate a higher margin of liver safety for GPR40 agonist HD-6277 compared to Takeda’s suspended fasiglifam, but questions remain.

Devices

• Shortly before the conference, Medtronic launched in the US the first “artificial pancreas,” the MiniMed 670G. “Real world” data presented at the conference from a customer training phase showed results comparable to the pivotal trial.
• Updated data from Dexcom’s DIaMonD study comparing continuous glucose monitoring (CGM) with fingersticks now included a type 2 diabetes cohort, but curiously, data specifically for that cohort was not broken out from the type 1 patients.
  • If one works backwards from averages to estimate the type 2 results, it appears that those patients did not benefit nearly as much as those with type 1, and may have had increased hypoglycemia with CGM.
  • This raises questions about the robustness of the type 1 data, though it is possible those patients may indeed benefit more from the technology.

Additional data or news involve PCSK9 inhibitor Praluent (SNY, REGN), with lipid results from trials of diabetics; oral GLP-1 agonist semaglutide (NVO), though just Phase I studies; PLX039 (ProLynx), a hydrogel-microsphere drug-delivery system for potential monthly dosing of a GLP-1 agonist; INO352 (Ajinomoto), an allosteric modulator of GLP-1 in preclinical development; BioChaperone technology (Adocia) for the development of glargine-liraglutide and -dulaglutide combinations; oral insulin from NVO and Oramed; type 1 diabetes immune intervention trials (BCG, Gleevec [NVS], oral insulin, Diamyd [DMYDY]); glucagon receptor antagonist REMD-477 (REMD); oxyntomodulin analogs (including TT401 [OPK, LLY]); glucokinase activators (TMG-123 [Teijin], HMS5552 [Hua Medicine]; dual GLP-1/glucagon agonist SP-1373 (Velocity/Spitfire) and standardization of CGM.