In Hindsight, Savara Wishes It Led With Patient-Reported Secondary Endpoint For Molgradex
Executive Summary
Orphan lung disease firm tries to sell Phase III setback in aPAP as a “successful failed study” but investors disagree, cutting Savara’s share price by three-quarters. Biotech hopes to talk with regulators after deeper dive into dataset.
It’s fairly standard for a biotech to frame a study that missed its primary endpoint as offering promise on other measures, but Savara Inc.’s attempt to depict its unsuccessful Phase III IMPALA study of Molgradex in autoimmune alveolar pulmonary proteinosis (aPAP) as a “successful failed study” is not resonating with investors.
The Austin, TX-based biotech reported after market closing on 12 June that Molgradex – an inhaled recombinant formulation of human granulocyte colony-stimulating factor (GM-CSF) – missed the primary endpoint in a 138-patient, placebo-controlled pivotal study. The trial did not hit a primary endpoint of improvement from baseline in alveolar-arterial oxygen gradient (A-aDO2), an outcome it attributed mainly to a strong placebo effect. A rare disorder affecting seven patients per million, according to Savara, aPAP is characterized by build-up of grainy material in the alveoli.
Despite Savara’s focus on a 12 June investor call on statistically significant improvement in the key secondary endpoint of St. George’s Respiratory Questionnaire (SGRQ), a patient-reported quality-of-life measure, investors showed immediate signs of bailing. Savara’s share price tumbled steeply in after-hours trading, continuing into 13 June and finished that day's trading at $2.62 per share, a 75.2% decrease from the closing price on 12 June.
In IMPALA, patients were randomized into three groups: a continuous treatment arm given 300mcg of Molgradex daily for 24 weeks; an intermittent treatment group that received either study drug or placebo, switched every seven days, for 24 weeks; and a placebo arm. For the primary endpoint, Savara – which says it is undertaking a deeper analysis of the trial data – did not report results from the intermittent treatment group, but revealed an average 12.1 millimeter of mercury (mmHg) improvement for the continuous treatment group and an 8.8mmHg average improvement in the placebo arm. This treatment difference of 4.6mmHg indicated the drug was active in improving lung function, Savara execs asserted, but did not meet statistical significance (p=0.17).
“We did observe a treatment effect in both of the active arms, which is, in terms of the magnitude of effect, very similar to what has been observed in prior uncontrolled studies over six months,” Savara president Taneli Jouhikainen told the call. “But what was very surprising, and certainly very disappointing, is that we saw what we would characterize, really, as a massive placebo effect for this type of disease.”
Demographic Differences In Arms Don’t Explain Trial Results
Patients were distributed fairly evenly amongst the three arms, the exec said, with 46 in the continuous treatment arm, 45 in the intermittent treatment arm and 47 in the placebo arm. An initial analysis showed some demographic differences between the arms, but nothing that would explain a material effect on the primary endpoint results, Jouhikainen noted.
As it conducts a deeper analysis, one question Savara will consider, he added, is whether the trial enrollment included too many patients with a mild state of aPAP disease. “We can speculate that perhaps despite our efforts to limit the patients who have too mild of a disease, it … may not have been quite adequate, and we will be analyzing this data very carefully to understand if perhaps the efficacy is best seen or better seen in more severe patients,” Jouhikainen said.
Jouhikainen called the SGRQ data “very impressive improvement” and explained that while Savara feared that a placebo effect would take a bigger toll on a patient-reported outcome than on an objective endpoint like A-aDO2, the exact opposite occurred. The trial showed an average 12.3 points of improvement on SGRQ score in the continuous treatment arm, compared with an average 4.7-point improvement in the placebo arm (p=0.01).
The 7.6 points of difference in SGRQ score is approximately twice what would be needed to show a clinically meaningful effect for a patient-reported outcome in lung disease, the exec pointed out. He also noted that the trial demonstrated statistically significant improvement in the diffusing capacity of the lungs for carbon dioxide (DLCO), suggesting the drug actively improves lung function.
“When we analyzed the effect to the DLCO, we could clearly see a dose-dependent improvement, which is statistically significant for the continuous dose and trending for the intermittent dose,” Jouhikainen said. “And so taken together, when we look at the DLCO on the one hand, we look at the A-aDO2 on the other hand and assess whether the drug is likely to have improved oxygen transport from the lungs to blood, our conclusion is that there's definitely a reason to believe that happened and the drug works in that fashion and that this is also associated with clinical benefit in form of quality-of-life improvement.”
He pointed out that Molgradex also showed a solid safety profile in the Phase III study, with similar numbers of adverse events across arms. Still, despite the safety profile and the indications seen with secondary endpoints, Jouhikainen conceded that “we have an uphill battle for approval with this dataset, but we do firmly believe that, based on this data, the drug works.”
Savara hopes to gain a further understanding of the IMPALA data and then discuss a potential path forward with both the US Food and Drug Administration and the European Medicines Agency, CEO Robert Neville told the call, which could mean filing for approval based on the data or running another study.
Analysts Expect Another Study Needed To Support Approval
JMP Securities called the data “a disappointing setback” for Molgradex and lowered its rating of Savara’s stock from “market outperform” to “market perform.” Analyst Liisa Bayko called SGRQ score “an important secondary endpoint” in a 13 June note. “We agree there is a signal of a drug effect; however, we believe another study will be necessary to support approval,” she wrote, pushing the drug’s expected launch date back from 2021 to 2024 and reducing its likelihood of approval from 60% to 40%.
Jefferies analyst David Steinberg offered a similar view in his note published on 13 June, saying he had pushed back Molgradex’s expected commercialization by two years, but had not removed the drug entirely from his Savara modeling. He added that SGRQ score might be a better indicator of efficacy in aPAP and asserted that Savara should have used that as its primary endpoint.
Savara, which also has Aerovanc in Phase III for methicillin-resistant Staphylococcus aureus (MRSA) infections in cystic fibrosis and two other mid-stage candidates for orphan lung indications, had $105m in cash on hand at the end of the first quarter, giving it runway into the second half of 2020, Steinberg added. This will give it a chance “to turn over at least two if not three additional cards in the next 12 months,” he said.
Savara obtained Molgradex in all-stock transaction in 2016 in which it acquired Danish company Serendex Pharmaceuticals AS. [See Deal] It went public less than a year later via a reverse merger with Mast Therapeutics Inc., a deal that gave Mast shareholders a 23% stake in the new company. [See Deal]