Amylyx: Crucial Advisory Committee Meet Tomorrow (NASDAQ:AMLX)

Newly IPO-ed Amylyx Pharma (NASDAQ:AMLX) has had a rough few weeks. I covered the stock quickly in January, when it was just launching, and by February the stock had nearly doubled ahead of an upcoming PDUFA on June 30 for lead candidate AMX0035 targeting amyotrophic lateral sclerosis (ALS). On Feb. 16, the company announced that the FDA will hold an advisory committee meeting on March 30. The stock went up on that news.

However, on Feb. 28, exactly one month before the FDA released its damaging briefing documents for the adcomm, the stock started falling. It fell steadily for a month, then saw a spike briefly before the actual release of the briefing docs, after which it again fell sharply. It is now trading lower than its early-January prices.

The briefing docs begin by stating the FDA’s position that two well-designed large trials are generally needed for approval, barring cases of serious diseases where “the trial has demonstrated a clinically meaningful and statistically very persuasive effect on mortality, severe or irreversible morbidity, or prevention of a disease with potentially serious outcome, and confirmation of the result in a second trial would be impracticable or unethical.”

Now, AMLX produced two data-points for the FDA to consider. One was from the CENTAUR study, a single double-blind, placebo-controlled Phase 2 Study in 137 patients with ALS. First, as a brief background:

In 2019, FDA published a Guidance for Industry “Amyotrophic Lateral Sclerosis – Developing Drugs for Treatment,” which noted that survival time in ALS varies greatly, and that functional endpoints can be confounded by loss of data because of patient deaths. To address this, FDA recommends sponsors use an analysis method that combines survival and function in a single overall measure, such as the joint rank test

Now, the data that AMLX produced was not about such a joint rank test, but from a purely functional endpoint that the FDA does not prefer. Here, the company reported a positive, but not too persuasive, outcome, with a p-value of 0.034. The endpoint, however, had interpretive difficulties because, among other things, it did not account for deaths in the trial. So the FDA asked the company to begin a confirmatory phase 3 trial.

While the company has begun a 600-patient global trial, it also published data from an open-label extension study which seemed to show a survival benefit in patients who initially received AMX0035 compared to those patients who originally received placebo in the CENTAUR study. This data also, according to the FDA, has interpretability issues given the large number of trial dropouts, among other things. However, the FDA still invited the company to submit an NDA while the phase 3 study, to be completed in 2024, was still ongoing.

In a Type C Meeting on Feb. 4, 2021, which occurred after the OLE data presentation, the FDA said that while the new data was encouraging, they still insisted on another study.

The briefing documents are detailed and technical. One point the FDA makes multiple times is that if linearity of the primary endpoint of ALSFRS-R score is not assumed, then it leads to a failure of stat sig. As the FDA states – “Based on this analysis, the Week 24 treatment difference is estimated as 1.68 (S.E.=1.06) with a p-value of 0.1134.” A p-value less than 0.05 is considered statistically significant.

Besides, there was an amount of lost data in the trial, and possibility of bias because a much higher proportion of placebo patients took edaravone or riluzole even during the trial. The FDA concludes its analysis by saying this:

The potentially incorrect assumptions about linearity of the functional rating scale, the ignoring of deaths in the primary analysis, and the higher percentage of patients starting intercurrent treatment for ALS, in addition to the randomization error during the study, make it challenging to interpret the positive p-value in this small sample size.

The company performed a post hoc joint rank analysis, which, using their method, gave a p-value of 0.037. However, the FDA pointed out certain errors in the way the company handled missing data and ignored two deaths in the ITT population. If the joint rank analysis is done accounting for these errors, a p-value of 0.079 is obtained, which is not statistically significant.

On the positive side, the FDA agreed with the company that AMX0035 appeared to be safe and generally well-tolerated, with no unnatural safety signals seen.

Given what we discussed so far, I see two outcomes here. One, the FDA could give a sort of accelerated approval to the drug, with the phase 3 trial made into a post-approval confirmatory trial. This assumption takes into consideration the unmet medical burden of ALS, and the molecule’s broadly benign safety profile. I would give this a 40% chance. The second option could be that the FDA will delay potential approval until 2024, when the phase 3 confirmatory trial will be completed. Given the general direction of the FDA briefing docs, the “mild” benefit observed in the trial, problems with the study conduct and robustness, etc., I think this has a 60% chance of happening. There could be a third option – approval if the 24-week analysis is satisfactory – but I do not put much stock in that option.

Then there’s always the unknown factor. According to Evaluate, the FDA’s change of heart for the NDA came about because of lobbying from the ALS Foundation and other patient advocacy groups. One wonders how much pressure these groups can bring to bear on the regulatory agency and whether the advisory committee is the FDA’s way of publicly passing the buck for a drug they would rather wait two more years.