The American Diabetes Association (ADA) 78th Scientific Sessions was held in Orlando, FL from 22–26 June 2018. While there were no major new ground-breaking results presented, the highlight of this year's conference was further details from oral semaglutide's (NVO) Phase III program, since the GLP-1 agonist appears more efficacious than other oral options and even than Novo's own injectable Victoza. While Novo had additional data from its trial of recently approved injectable semaglutide (Ozempic) showing advantages over Lilly's Trulicity, the latter had data from a higher-dose study.

A Januvia (MRK) study raised questions whether the DPP-IV inhibitor was more effective on glycemic control than SGLT-2 inhibitors in stage 2 renal impairment, though a KOL was skeptical. The results were particularly notable as soon after the conference, JNJ announced its SGLT-2 inhibitor Invokana met an interim analysis in its Phase III DN trial CREDENCE, though without details.

Meanwhile, long-acting insulins Toujeo (SNY) and Tresiba (NVO) continued to battle with competing data from a comparative trial and real- world studies.

A major highlight outside of study results was the ADA/EASD's release of a draft consensus treatment algorithm. While the algorithm retained metformin as the first-line choice, it subsequently focused on whether patients had atherosclerotic cardiovascular disease/heart failure, highlighting the GLP-1 agonists and SGLT-2 inhibitors with proven CV benefits. Other steps focused on whether there was a need for hypoglycemia minimization and weight loss.

On the device front, there was interesting data on Insulet's Omnipod Horizon hybrid closed-loop system, a three-way comparison of CGM devices (Senseonics' Eversense, Dexcom G5, and Abbott's Freestyle Libre Pro), and Medtronic's new Sugar.IQ app for its Guardian Connect CGM system.

We highlight these and other presentations below.

Like our report? Have any questions or feedback? Please let us know at askanalyst@sagientresearch.com.

GLP-1 Agonists

• New details from the primary ITT analyses required by regulators for oral semaglutide's Phase III trials were presented at the conference (PIONEER I) and in a company conference call (PIONEER 2, 4, 7), generally reinforcing prior top-line data, though with some differences. The top-line data had used a secondary analysis when patients were "on-treatment without rescue medication." Of note, those top-line results for PIONEER 2, 4, and 7 had only been released within weeks of the conference, building excitement for the drug, with advantages on both HbA1c and weight compared to SGLT-2 inhibitor Jardiance (Boehringer, LLY), injectable GLP-1 agonist Victoza, and DPP-IV inhibitor Januvia.
  • HbA1c differences were generally consistent between the two analytical methods, except in the PIONEER I trial against placebo, where it was still strong, but not as striking. The company attributed this to more placebo patients starting rescue medication.
  • Weight reduction against the active comparators in PIONEER 2, 4, and 7 was not as strong in the primary method, with little difference in weight compared to Jardiance.
    • However, on the latter point, Novo officials said oral semaglutide did have an advantage on waist circumference, and they are considering conducting a DEXA scan study to show that oral semaglutide does lead to more fat loss, with some of Jardiance's weight reduction due to diuresis of fluid.
    • Curiously, despite an advantage on weight compared to Victoza, oral semaglutide only had a modest weight advantage to Januvia, which does not usually cause weight loss, even in a fixed-dose study presented a couple days after the conference (PIONEER 3). Those studies allowed concomitant sulfonylureas/thiazolidenediones, raising a question whether these drugs can limit the benefit on weight seen with oral semaglutide.
  • While it is positive that oral semaglutide had rates of nausea fairly comparable to Victoza, which will help bolster its use before injectable GLP-1 agonists, given this is substantially higher than for other oral drugs, a variety of considerations are likely to come into play when deciding which oral drugs to use first.
  • Novo officials said if oral semaglutide's PIONEER 6 CV safety study is positive, they will approach regulators for a CV label indication for both oral and injectable (Ozempic) semaglutide, since the latter had also shown a benefit, though in a safety study too small for the FDA to approve a CV indication.
    • However, PIONEER 6 has even less statistical power to show such a benefit.
    • Novo also announced they are now planning a large CVOT in diabetics for oral semaglutide, rather than for Ozempic, though the latter will have a CVOT in obesity.
• New analyses from Ozempic's SUSTAIN-7 trial against LLY's Trulicity, as well as higher-dose data from the latter, illustrate the high stakes in competition among injectable GLP-1 agonists. Novo is hoping Ozempic will help it defend against Trulicity, which has been taking share from Novo's market-leading daily Victoza.
  • The analyses from SUSTAIN-7 were incremental, but reinforced prior published data, showing that for comparisons of both low and high doses, Ozempic was superior to Trulicity in terms of both HbA1c reduction and weight loss, across most HbA1c and weight categories, respectively. Still, even for the high dose of Ozempic, only about half the patients who started with a baseline HbA1c over 9% achieved the 7% HbA1c goal.
  • For its part, in a Phase II study, higher doses of Trulicity led to numerically greater reductions in HbA1c and body weight than the currently approved high dose, but while it is difficult to compare across trials, it seems possible Ozempic could retain a moderate advantage. Moreover, the highest dose of Trulicity tested appreciably increased the rate of nausea. A Phase III trial with the higher doses is under way and should complete in 2019.
  • Trulicity is expected to have results from its CVOT in early Q4, 2018, but as noted above, Novo announced that if oral semaglutide shows positive results on its CV safety study in Q4 2018, it will use that to submit to the FDA for a CV indication for Ozempic, the injectable formulation, as well, though the study is not strongly powered (please see above for more details).
• AZN's GLP-1/glucagon receptor dual agonist MEDI0382 led to a reasonably good reduction in fasting glucose relative to placebo in a 41-day Phase I/IIa study, as well as weight loss that was modest but could increase over time.
  • However, it was difficult to say whether it would have advantages over a single-agent GLP-1 agonist, though investigators felt its glucose-lowering could be better than semaglutide.
  • Rates of nausea/vomiting were fairly high, but the investigators implied these could potentially be reduced with titration.
  • A comparative study with Victoza could have data later in the year.

SGLT Inhibitors

• Additional data on SGLT inhibitors in type 1 diabetes did little to quell questions whether regulators will conclude the modest efficacy outweighs the risk of diabetic ketoacidosis (DKA), particularly once the drugs are released into broader real-world practice.
  • First conference details on 52-week data for Zynquista's (SNY, LXRX) inTandem1 and inTandem2 trials showed sustained efficacy versus placebo, albeit with a slightly smaller benefit on HbA1c than prior 24-week results.
    • On the other hand, an advantage on weight widened at 52 weeks in both studies.
    • Investigators concluded in the publications of the two studies that Zynquista lowered measures of hypoglycemia, indicating this could be due to the fact that it also targets SGLT-1. However, while there were signals of a benefit, it is difficult to draw conclusions, taking all of the data for the drug together. Still, this is likely a theme the company will want to promulgate, and in CGM sub-studies presented at the conference, Zynquista did appear to have a greater effect on post-prandial glucose than did SGLT-2 inhibitor Farxiga (AZN, BMY, Ono), offering more evidence for some differences in mechanism.
  • New 52-week data for Farxiga from DEPICT-1 and 24-week data from DEPICT-2 likewise reinforced its profile of modest efficacy in this segment, and unfortunately for the drug, rates of DKA were more comparable to Zynquista, unlike earlier more favorable 24-week data from DEPICT-1.

DPP-IV Inhibitors

• Januvia showed superior HbA1c lowering than SGLT-2 inhibitor Farxiga in patients with mild renal impairment, suggesting mild renal impairment has a stronger negative impact on Farxiga's efficacy than had been seen in pooled analyses of its prior trials, though consistent with early pharmacodynamic studies on urinary glucose excretion.
  • Some KOLs were surprised or felt the study was an outlier, but also that Farxiga's lower efficacy may not be as applicable for Invokana and Jardiance, which have shown stronger HbA1c reductions than Januvia in more general trials.
  • The diabetic nephropathy trials of SGLT-2 inhibitors that have released design details are testing patients with higher levels of albuminuria, so Merck may hope the data raise questions in more general patients with mild renal impairment, whether there is as compelling an argument for use of the class over Januvia.
    • Soon after the conference, JNJ announced that Invokana's CREDENCE DN trial was positive in an interim analysis, but details were not released yet, including whether there was an increase in amputations, a risk seen in its CANVAS program.

SGLT-2/DPP-IV Inhibitor Fixed-Dose Combination

• AZN’s Qtern (dapagliflozin/saxagliptin) achieved non-inferiority on HbA1c lowering compared to insulin glargine, with a trend for a slight advantage, though this could have been due to chance.
  • As expected, Qtern significantly reduced weight and hypoglycemic events compared to insulin glargine, as well.
  • While the findings offer additional reason to delay uptake of insulin until necessary, the increased cost/reimbursement issues associated with the FDC may continue to hinder uptake.

Insulins

• Long-acting basal insulin Toujeo (SNY) showed non-inferior glycemic control when compared to competitor Tresiba (NVO) in the open- label BRIGHT trial, but while it had a hypoglycemic benefit during the initial titration period, this did not persist in the maintenance phase. Nevertheless, it is somewhat positive for the drug that it was at least as good as Tresiba.
  • Novo's own comparative trial, with a longer follow-up, should have results in Q4 2018. A company official said it could take a longer maintenance period than in BRIGHT to show any difference in hypoglycemia.
  • Sanofi's real-world comparison studies presented at the conference showed comparable HbA1c reductions and hypoglycemic events, whereas Novo's showed a benefit on both for Tresiba. The reason for the difference is unclear.

Other Mechanisms

• Updated Phase II data from second-generation MetAP2 inhibitor ZGN-1061 were somewhat mixed. While the HbA1c reduction for the high dose was 0.7% already at 12 weeks, so it could become somewhat larger, values for the lower doses at the end of the study were close to baseline, raising questions how representative the high dose was. Weight loss was also modest, though it too could improve over time.
  • A higher dose is already being tested in the study, so those data should help better understand the profile of the drug.
  • The company will not take the drug into Phase III themselves, though they may do another Phase II study to enhance its profile.

Devices

• This year's meeting focused heavily on Bluetooth connectivity, mobile apps, and integrated solutions that are poised to transform diabetes care and improve outcomes.
• In a small, early stage trial, Insulet's Omnipod Horizon hybrid closed-loop system, which utilizes an untethered patch pump with no tubing, improved time-in-range (70-180mg/dl) compared to an earlier standard therapy phase in type 1 children, adolescents, and adults. The benefit came from lower time in both hyperglycemia and hypoglycemia, though patients were only in the hypoglycemia range for a small percentage of time in either phase.
  • One caveat is that there were different settings for the separate comparative phases of the study, limiting interpretation. The closed- loop system was tested in a supervised hotel setting with free-living conditions, while the open-loop standard therapy phase was at home.
  • Once full data from the trial are available and analyzed, the company will decide on whether to next proceed to a pre-pivotal trial, which it has said in the past would be in conjunction with the Dexcom G6 CGM.
  • The Omnipod Dash Insulin Management System, which is an open-loop system, was just approved several weeks prior to the conference. Interestingly, earlier in the year the pump gained Medicare Part D coverage, because it is disposable (under some plans, though, the Personal Diabetes Manager component apparently needs reimbursement through Part B).
• In a third-party comparison of three continuous glucose monitoring (CGM) devices in patients with type 1 diabetes, Senseonics' 90- day implantable Eversense, which was approved by the FDA the day before the conference began, performed modestly better overall than the seven-day Dexcom G5 and 14-day Abbott Freestyle Libre Pro (professional version). There were bigger differences among devices in the hypoglycemic range, though all of the CGM's performed worse there than at other glucose levels.
  • However, a potential source of bias in the study was that Eversense and the G5 were calibrated twice a day with the same device that was used to establish accuracy, as pointed out by an Abbott official, whereas the Freestyle Libre Pro does not require calibration.
• Medtronic announced the launch of its Sugar.IQ app, which is used in conjunction with its recently-approved Guardian Connect CGM system, along with new real-world data from a pilot program showing that the app improved time in range (mostly due to reduced hyperglycemia) and reduced hypoglycemia events, compared to baseline. Guardian Connect is the first CGM in the US to only display through a smartphone without an extra receiver.
• Reapplix released data from a European controlled study, showing improved wound healing with LeucoPatch (3CPatch in the US) versus standard-of-care in hard-to-heal diabetic foot ulcers.
  • With the positive data, the company also announced it was moving ahead to form a US company to commercialize the product in the US.

This report also has additional data/news on Victoza in type 1 diabetes, SGLT-2/GLP-1 agonist combinations, Farxiga in diabetics with heart failure, and Invokana (amputation risk from an observational study, CV effects by renal impairment status from CANVAS, and effects on diastolic dysfunction). For devices, there is additional news on several closed loop systems, including DBLG1 (Diabeloop, Cellnovo) and investigator-initiated trials on inpatient, non-critical-care use and inclusion of pramlintide.

For our disclosures, please read the Biomedtracker Research Standards.