Lexicon reports positive topline results from second phase 3 inTandem2 study of SGLT-1/2 inhibitor sotagliflozin in type 1 diabetes – December 21, 2016

Executive Highlights

  • Lexicon reported topline results from the second pivotal inTandem2 trial of SGLT-1/2 dual inhibitor sotagliflozin in patients with type 1 diabetes.
  • The trial demonstrated placebo-adjusted A1c reductions of 0.36% and 0.35% with 200 mg and 400 mg doses of sotagliflozin, respectively, on top of insulin optimization (baseline A1c=7.7%-7.8% after insulin optimization, p<0.001).
  • DKA incidence was 0.4% and 1.1% in the 200 mg and 400 mg arms, compared to none in the placebo arm. Severe hypoglycemia was comparable in all three groups and was not significantly increased in the sotagliflozin arms.

This morning, Lexicon announced positive topline results from its phase 3 inTandem2 trial of SGLT-1/2 dual inhibitor sotagliflozin in patients with type 1 diabetes. This is the second pivotal trial of sotagliflozin to report results, following the positive topline results from the inTandem1 trial in September. In the inTandem2 trial (n=782 patients with type 1 diabetes in Europe and Israel), treatment with both the 200 mg and 400 mg doses of sotagliflozin demonstrated statistically significant and clinically meaningful placebo-adjusted reductions in A1c at 24 weeks, on top of optimized insulin therapy. Sotagliflozin produced placebo-adjusted A1c reductions of 0.36% and 0.35% with the 200 mg and 400 mg doses, respectively (p<0.001 for both doses). Impressively, these reductions occurred in the context of low baseline A1cs (7.7%-7.8%) following the six-week insulin optimization period.

Lexicon management generally characterized sotagliflozin as well-tolerated and did not express significant concerns about the safety and adverse event profile of the drug. Overall adverse event rates in the sotagliflozin arms were 56% in the 200 mg group and 54% in the 400 mg group, compared to 51% in the placebo group. Serious adverse event rate was 4.2% in both sotagliflozin arms, compared to 3.5% in the placebo arm. Discontinuations due to adverse events were 2% and 3% in the sotagliflozin arms, compared to 1.6% in the placebo arm. Management especially highlighted diarrhea and genital mycotic infection rates, which, while higher in the sotagliflozin arms, resulted in treatment discontinuation in less than 1% of participants across all three groups. Notably, there were two deaths in the trial – the first in Lexicon’s clinical program for sotagliflozin – though both were in the placebo arm of the trial.

  • Sotagliflozin treatment was associated with higher DKA incidence – 1 patient experienced DKA in the 200 mg arm and 3 patients experienced DKA in the 400 mg arm, compared to none in the placebo arm. This translates into a 0.4% event rate in the 200 mg arm and a 1.1% event rate in the 400 mg arm. Lexicon management characterized these results as low DKA incidence and emphasized that the company is pleased with the safety profile of sotagliflozin. These rates are lower than the 1% and 3% rates reported in the inTandem1 phase 3 trial. For comparison, the phase 2 trial of J&J’s SGLT-2 inhibitor Invokana (canagliflozin) in type 1 diabetes found a significant increase in ketone-related adverse events (11 events and 6 events with the two respective doses vs. none with placebo), including serious DKA (6 events and 4 events with the two respective doses vs. none with placebo). While it’s of course difficult to compare between trials, we find it notable that canagliflozin was associated with 2x-4x more DKA than sotagliflozin, despite a trial less than half the size (n=351 in canagliflozin trial, compared to n=782 for inTandem2). Notably, all events in the canagliflozin trial were associated with precipitating factors such as pump failure, missed insulin doses, or concurrent illness. In a webcast accompanying the inTandem2 results announcement, Lexicon management shared that the DKA events were predominantly in patients on insulin pumps, as opposed to MDI – we’re curious if pump malfunctions may have played a role in the events, though the company emphasized that it does not yet have details on the circumstances surrounding the DKA incidences. We continue to believe that the DKA risk associated with this class can be mitigated with good patient (and provider!) education. That said, DKA is a serious and major concern for many and we imagine that this risk will be raised as a potential concern by regulatory authorities. DKA concerns have derailed development of adjunctive therapies for type 1 diabetes before – Novo Nordisk declined to pursue a type 1 diabetes indication for its GLP-1 agonist Vicotza (liraglutide) despite modest but statistically significant A1c reductions due to its safety profile that, along with worrisome increases in hypoglycemia, included eight cases of DKA in the ADJUNCT ONE trial (n=1,398) and a nearly four-fold increase in hyperglycemia with ketosis in the ADJUNCT TWO trial (n=835). In our view, it will be important for Lexicon to take this risk seriously and carefully consider how to present this risk in the context of sotagliflozin’s overall benefit-risk profile.
    • In a separate call, Lexicon management expressed optimism that the DKA risk can be managed, especially with increasing patient and provider familiarity with sotagliflozin. Management pointed out that there were fewer incidences of DKA in inTandem2 compared to inTandem1 (which observed three events in the 200 mg group and eight events in the 400 mg group) and suggested that the company is making progress in managing DKA risk at least in clinical trials. Management particularly pointed to increased experience with sotagliflozin treatment and protocols as a major driver of lower DKA rates. We certainly hope that some of the education strategies employed by Lexicon can be translated into real-world clinical practice in order to produce as positive of a benefit-risk ratio as possible. In the vein, we’re extremely interested in the “net benefit” composite endpoint that Lexicon has introduced in its phase 3 program (“net benefit” is the primary endpoint of inTandem3 and a secondary endpoint in inTandem1 and inTandem2). The composite endpoint consists of proportion of participants that are able to achieve an A1c target <7% with no DKA and no severe hypoglycemia, and we hope that this composite can help clearly illustrate the benefit-risk profile of sotagliflozin for regulatory authorities and the public.
  • Notably, severe hypoglycemia was neither increased nor decreased with sotagliflozin treatment compared to placebo. Severe hypoglycemia rate was 3.8% in the 200 mg arm and 2.3% in the 400 mg arm, compared to 2.7% in the placebo arm. In the earlier inTandem1 trial, sotagliflozin was associated with a numerically lower hypoglycemia rate, though the reduction was not statistically significant. In a separate call with Lexicon management, we learned that, in the inTandem2 trial, documented hypoglycemia was defined as a blood glucose reading of less than 70 mg/dl and severe hypoglycemia was defined as hypoglycemia resulting in loss of consciousness or seizure, ore requiring the assistance of another person. That said, management shared that pre-specified secondary endpoints in the CGM sub-study will evaluate the incidence of severe hypoglycemia as defined as a blood glucose value of 54 mg/dl or lower. Results from the phase 2 inTandem4 dose-ranging study and the phase 2 JDRF-partnered study of sotagliflozin in younger, less well-managed patients with type 1 diabetes found similarly low rates of severe hypoglycemia in the sotagliflozin and placebo groups. While severe hypoglycemia is a notoriously difficult endpoint to measure and demonstrate improvement in clinical trials (we hope that greater standardization and consensus on its definition should help matters), we would’ve loved to see a benefit on this front given the promising data we’ve seen to date from phase 2 studies on sotagliflozin’s benefits on time-in-range and postprandial glucose – we hope that the CGM study with a more quantitative definition of severe hypoglycemia may be able to help demonstrate a benefit. Time-in-range and postprandial glucose have a major impact on self-management for many patients with type 1 diabetes – fewer and more modest postprandial spikes could reduce bolus insulin need and reduce the risk of severe hypoglycemia later on. Indeed, while the company does not yet have formal qualitative patient feedback data, management shared in Q&A a patient quote that illustrates sotagliflozin’s impact of glycemic variability: “Before I took sotagliflozin, my glucose profile was like mountains and canyons. Now, it’s like hills and valleys.” Lexicon has not shared any secondary endpoint data from this study or from other trials in the phase 3 development program. However, inTandem2 included a CGM sub-study and we are especially eager to see the results on time-in-range from this study. In our call with Lexicon, management also suggested that it will be particularly informative to examine hypoglycemia rates during the pre-randomization insulin optimization period compared to the post-randomization active treatment period – we continue to eagerly await the full results!
  • Lexicon shared that it expects to have “some presentations with data” at ADA 2017 in San Diego. We hope this will include full 24 week results, including secondary endpoints and results from the CGM and body mass composition sub-studies, from the inTandem1 and inTandem2 phase 3 trials. Both inTandem1 and inTandem2 include 52-week long-term extension studies that Lexicon suggested will not be ready in time for ADA 2017. Even so, we’re extremely curious to learn more about sotagliflozin’s impact on glycemic variability, postprandial glucose, body weight, blood pressure, the net benefit composite, and quality of life.
  • Management continued to express optimism that the company will be able to file sotagliflozin for regulatory approval in type 1 diabetes ahead of type 2 diabetes. The Sanofi-run type 2 diabetes phase 3 program just got underway in 4Q16 – the 240-patient study in drug-naïve patients is expected to complete in March 2018 and the 500-patient study in metformin-treated patients is expected to complete in March 2019. This represents a significant lag behind the Lexicon-managed type 1 diabetes program – full 52-week extension results from inTandem1 and inTandem2 should be available mid-2017 and the third and final inTandem3 study is expected to report results in mid-2017 as well. With topline results from two pivotal trials in hand, Lexicon management shared that it is initiating conversations with Sanofi regarding next steps and that the two companies would engage in discussions with the FDA in 2017. Assuming that the FDA responds positively to the idea of a standalone type 1 diabetes submission (a big if, considering that the FDA had previously expressed a preference for a joint type 1 and type 2 diabetes phase 3 program), management estimated that the submission could occur in the early 2018 time frame. In either case, management suggested that Lexicon will be able to offer more clarity on next steps after the first quarter of 2017.

Close Concerns Questions

Q: How was severe hypoglycemia defined in the trial?

Q: Would a more rigorous hypoglycemia trial – similar perhaps to the SWITCH 1 and 2 trial design for Tresiba (insulin degludec) – demonstrate an improvement in severe hypoglycemia?

Q: To what degree might the DKA findings be related to an imperfect understanding of how to optimize insulin dosage in the context of an SGLT-1/2 dual inhibitor on the part of the trial investigators?

Q: What kind of patient and provider education tactics might mitigate the incidence of DKA?

Q: What were the results for the secondary endpoints? How do the time-in-range, postprandial glucose, weight reduction, etc. compare amongst the two phase 3 trials and compare to the phase 2 trials?

Q: Will we see more drastic A1c and secondary endpoint results in the inTandem3 study, considering that inTandem3 does not include a pre-randomization insulin optimization component?

Q: How will sotagliflozin fare in the “real world” in which likely many patients won’t achieve optimal insulin therapy prior to initiating sotagliflozin therapy?

Q: What are the long-term cardiovascular and renal effects of sotagliflozin? We assume there will not be a large-scale randomized, controlled outcomes trial for sotagliflozin in type 1 diabetes specifically, but could these outcomes be assessed with big data and registries? Could the Helmsley Charitable Trust or the JDRF fund a long-term follow-up study to assess this?

-- by Helen Gao and Kelly Close