Amylyx outlined the scope of its Phase 3 defeat, demonstrating the extent to which its now-withdrawn ALS drug Relyvrio failed to separate itself from placebo.
In the primary endpoint measuring patients’ declining levels of function, known as the ALSFRS-R, Relyvrio induced only a 0.343-point difference between the trial’s active and placebo arms after 48 weeks, according to Amylyx’s presentation on Tuesday at the American Academy of Neurology’s annual meeting.
Those taking Relyvrio saw their ALSFRS-R scores decline by 14.98 points, while the placebo group’s scores declined by 15.32 points. The difference resulted in the previously reported p-value of p-0.667.
Co-CEO Joshua Cohen told Endpoints News that Amylyx is still working to understand what specifically drove the failure.
Joshua Cohen“That’s what we hope to continue to learn,” Cohen said. “Might there be certain biomarkers, might there be certain biological differences, or otherwise? Or is it just statistical chance? I think all of those are possibilities and things we want to keep learning more about.”
Last month, Amylyx reported that Relyvrio failed the Phase 3 PHOENIX trial in one of the most anticipated biotech events of the year. The resounding and clear-cut defeat — with a lack of signal in any of the patient subgroups — prompted the company to voluntarily withdraw the drug in a move widely applauded by experts.
The AAN presentation gave further insight into the trial. For example, in Phase 3 participants who also met the criteria for the successful Phase 2 CENTAUR study (upon which Relyvrio received full FDA approval in September 2022), the drug coaxed a 2.01-point difference in ALSFRS-R and placebo.
But that group only made up about 25% of the participants in the Phase 3 trial. Amylyx enrolled 664 patients, the vast majority (83%) of whom lived in Europe, and randomized them at a 3-to-2 ratio between the active and placebo arms. Among patients who would not have qualified for the Phase 2 trial, the placebo group outperformed those on Relyvrio on the ALSFRS-R by a 0.101-point margin.
“When we designed CENTAUR, the idea was to select for a rapidly progressing homogeneous group,” Cohen said. “When we went into PHOENIX, we looked to broaden those slightly. Not dramatically, but slightly.”
Justin KleeThe 168 patients in the Phase 3 trial who could have qualified for Phase 2 also represented a sample size that wasn’t much larger than the 137 patients enrolled in the mid-stage trial itself. Co-CEO Justin Klee said the three biggest differences in inclusion/exclusion criteria for both studies involved time to the onset of first symptoms, presence of ALS symptoms in fewer parts of the body, and patients whose breathing hadn’t been as severely restricted by their disease.
In an effort to win approval in the EU, where regulators rejected the drug last year, Amylyx additionally looked at patients in that region. In this group, Relyvrio produced a 0.463-point advantage over placebo — slightly higher than the overall trial population, but still nowhere near the level needed for approval.
Patients who were not taking another approved ALS drug, called edaravone, also failed to see a benefit. Of the 664 patients in the Phase 3 study, 644 did not take edaravone, largely due to the drug not being approved in Europe, and Relyvrio generated a 0.431-point ALSFRS-R difference among them. Amylyx did not conduct a subgroup analysis for patients who took edaravone, given that it only consisted of 20 patients.
Amylyx said the Phase 3 study is continuing to collect overall survival data, which will not be available until 2025 or 2026. In order for the OS data to mature, a minimum of 70% of patients must have died or three years must have passed after February 2023 when the last patient was randomized, whichever comes first.
Amylyx also reported data from two secondary endpoints: a patient-reported questionnaire called the ALSAQ-40 and a measurement of breathing ability called slow vital capacity, or SVC.
In the ALSAQ-40, active arm patients saw their scores change by 39.8 points, compared to 38.4 from placebo-arm patients. For SVC, the average change from baseline for Relyvrio patients was -21.0, compared to -23.0 for placebo patients. The respective 1.41-point and 2.02-point differences were not statistically significant, though Amylyx didn’t report p-values.
There are few next steps available for Relyvrio, but Cohen and Klee say that they’ll continue to conduct biomarker analyses from the study. Researchers plan to conduct analyses for the Phase 2-like patient subgroup while continuing to collect Phase 3 survival data. Amylyx is also advancing the same compound in another rare disease called Wolfram syndrome.
“We think the drug has clear biological activity, and it’s really always a question of matching the mechanism of the drug with the mechanism of the disease,” Klee said. “And I think that’s where biomarkers can be informative.”