Lykos Therapeutics just got publicly bodyslammed by the Food and Drug Administration.
The backstory: Lykos is a privately-held company spun out of a nonprofit called the Multidisciplinary Association for Psychedelic Studies (MAPS). MAPS was founded by its current president Rick Doblin, an advocate for the use of psychedelic drugs; Lykos is an attempt to slap some corporate respectability on top of that drug advocacy while trying to make big bucks by peddling MDMA as a PTSD treatment.
Lykos made a big media splash when the FDA advanced its proposed “medicine” through Phase III trials—only for the company’s request that the agency recognize its product as a drug to be denied: an advisory panel overwhelmingly had doubts about the drug’s efficacy and its safety. More bad news followed when papers promoting the efficacy of MDMA for PTSD were retracted for ethical violations. Those violations included alleged sexual assault committed against a participant by a therapist in the company’s Phase II series of trials.
Now the FDA has made public the full letter it sent rejecting Lykos’ application . . . and it’s horrifying.
Let’s start at the top (the “you” in all the below is Lykos’ chief scientific officer and by extension the whole company):
You did not collect important information on events that the participant, therapist, or study physician considered “positive” or “favorable.” This information is necessary for FDA to assess signals of abuse potential and patient impairment in the clinical trials in order to adequately describe the drug effects
In other words, Lykos avoided collecting data that might suggest the therapy they were trying to get approval for has risks related to abuse and side effects. Huh, wonder why?
It gets worse.
FDA inspections also identified several unreported adverse events for at least two sites, which increase our concerns about the reliability of the safety data.
So, per the FDA, not only did Lykos avoid collecting data on responses that indicate abuse potential—they also avoided mentioning “adverse events.” Those are negative reactions to the therapy, which in this case could be equivalent to something like a bad trip and all its attendant psychological suffering. Though, because Lykos seems simply not to have reported what happened, the public may well never know.
As you were specifically advised, regardless of whether a participant, therapist, or physician may find a drug experience to be neutral, positive, or favorable, events that are relevant to the abuse potential of a drug or potential for impairment should have been captured and reported in the safety and abuse potential summary in the NDA. . . . Overall, based on this failure to collect important information on “positive” or “favorable” events, and the unreported adverse events identified by FDA at the study sites, there are substantial concerns about the reliability of the safety data which limit our ability to adequately assess the safety of midomafetamine for the treatment of PTSD.
“Midomafetamine” is MDMA, aka ecstasy. That “specifically advised” cuts to the bone: The FDA is saying in effect: “We told you you needed to capture all this data; you didn’t. And now no-one has any reason to trust the larger set of data around safety you submitted.”
But the agency isn’t done:
You have not demonstrated that the effect observed with midomafetamine is durable beyond the 18-week end-of-study assessments (8 weeks after the last dose of midomafetamine) in MAPP1 and MAPP2. As PTSD is a chronic condition, any drug indicated for the treatment of PTSD is expected to provide a durable treatment effect. Given the proposed dosing regimen of midomafetamine of a limited duration of three treatment sessions, it is critical to understand whether the treatment benefit is durable beyond the Week 18 assessment to inform the appropriate use of midomafetamine for labeling of the drug.
This is a doozy: The whole point of this trial was to find out if MDMA helped people with Post-Traumatic Stress Disorder, a terrible mental-health condition that can trigger major behavioral changes—and it’s known precisely for its duration. That doesn’t seem to have occurred to Lykos here. Or worse, it did, and they used a shorter assessment period out of fears a longer one would blow up their claims.
Then there’s the below:
[A]pproximately 40% of the enrolled participants reported prior use of midomafetamine, which is much higher than the background use of midomafetamine in the proposed indicated population with PTSD, and higher than what would be expected based on available data. Notably, FDA analyses of data from the 2022 National Survey on Drug Use and Health found that, in individuals with past-year moderate or severe mental illness, the lifetime prevalence of MDMA use was 16.2% and 20.5%, respectively. Additionally, the inspections conducted as part of this review detected very high rates of failures during the “prescreening” process prior to the formal screening (i.e., screening procedures and informed consent). Together these two factors suggest the possibility of selection bias, with the enrolled population not being representative of the general PTSD population. Not only does this potentially limit generalizability, it also raises the possibility of expectation bias given that participants who have prior experience with midomafetamine are more likely to recognize the effects of midomafetamine (i.e., leading to functional unblinding) and anticipate a treatment benefit.
This point is a little technical but it’s key. First, if you are doing a study on a drug like MDMA and your participants have a much higher previous use rate of the drug than the population your drug will roughly target, that’s a problem. It could mean that people have dependence issues that might make them bad fits for a study like this. It could mean they like and enjoy taking the drug recreationally. Second, the fact that so many people flunked out of the pre-screen process means that there’s likely an even bigger mismatch between the people actually in the study and the broader population it would be targeted at. That makes the data way less valuable. And as the FDA points out, it raises the issue of what is called “functional unblinding”–i.e. when people figure out whether they have been given the active drug or are in the control group because they know what the drug does, which can lead to results that understate risks and overstate efficacy because people are expecting positive results from the treatment.
Given the serious nature of the concerns raised about the reliability of the safety data, and also acknowledging comments made during the open public hearing of the advisory committee and in the public docket raising concerns about study conduct and the adequacy of collection of adverse event information, we recommend that you consider an independent third-party data audit of all study records and reports, including the recordings of the treatment sessions, to identify unreported or under-reported adverse events.
The FDA isn’t saying that Lykos is lying here, but the suggestion of a third-party audit is damning evidence that the agency simply does not trust any data coming out of this trial.
And with what looks like good reason:
Your application did not include laboratory data (e.g., liver analytes, electrolytes) from the phase 3 studies and the cardiac safety assessment was incomplete.
That . . . seems like a massive oversight. We hope it wasn’t intentional!
Although the following are not considered approvability issues, we recommend that they be addressed in your clinical trial:
1. Characterize the extent to which psychotherapy in the treatment contributes to the treatment benefit and if psychotherapy is necessary for a potential treatment benefit of midomafetamine to inform labeling . . . Consider a factorial study design that includes an arm with no psychotherapy.
2. Improve the diversity of your study population. For proposed recommendations regarding study diversity, refer to our recently published draft guidance for industry, Diversity Action Plans to Improve Enrollment of Participants from Underrepresented Populations in Clinical Studies.
I.e., The FDA seems to be saying that because of the way Lykos set up the trial it’s hard to tell if the benefits of the treatment come from the MDMA or the therapy they administer with the drug (we’re going to guess it’s the therapy). And it turns out that despite all the hippy-dippy good vibes energy of the pro-psychedelics movement, in this case they failed to have good enrolled population diversity—another thing that can degrade the generalizability of any gathered data.
This one is just a whole laundry list of issues the FDA says need improvement:
1. Drug-drug interaction study: We recommend that you consider assessing the potential for an interaction between midomafetamine and SSRI drugs to determine if a taper of SSRIs is necessary prior to initiation of treatment midomafetamine.
2. PK study in patients with impaired hepatic function: Midomafetamine is primarily metabolized in the liver. The impaired liver function is anticipated to increase the systemic exposures of midomafetamine. Therefore, we recommend that you conduct a PK study in patients with hepatic impairment. We note that you plan to conduct a PK study in participants with moderate hepatic impairment as a post-marketing phase 4 study (MPKH).
3. PK study in patients with renal impairment: Based on the available data, midomafetamine appears to be minimally excreted through renal pathway. Therefore, the PK of midomafetamine is not likely to be altered in patients with mild to moderate renal impairment (estimated glomerular filtration rate (eGFR <60 mL/min). However, given that uremic toxin could affect the hepatic disposition of midomafetamine in patients with severe renal impairment (eGFR: <30 mL/min), the review team recommends a reduced design renal impairment study for midomafetamine in patients with severe renal impairment.
4. Clinical lactation study: When lactating women with PTSD use midomafetamine, there is a potential for midomafetamine to get excreted into the breast milk. Therefore, it is important to conduct a milk-only clinical lactation study to understand the extent of midomafetamine that is excreted into breast milk.
5. Pre- and post-natal development study: After reviewing the data available from your conducted reproductive toxicology studies and data found in the literature, we have determined that there is a gap in the characterization of the risk of midomafetamine treatment during pregnancy on the development of the CNS and other developing systems.
These are all worrying, but given the way psychedelics keep getting pushed as “wellness” treatments, numbers 4 and 5 seem like whoppers. Maybe Lykos should have, you know, thought of that before?
The most alarming part of the larger story here, though, does not concern only Lykos. The broader field of psychedelic therapy rests on just such shaky ground. SAM EVP Luke Niforatos delineated this in a piece around psilocybin for the Colorado Gazette:
Consider a 2024 publication from researchers in Ireland and Canada, a review of nearly 200 published psilocybin studies, and 80 ongoing ones. Their findings should have put the brakes on efforts at the local, state and federal level to greenlight legal psilocybin.
Almost half the studies in which participants ingested psilocybin did not use a control group; i.e., there was no meaningful baseline against which to compare the results of the “treatment.”
The median number of study participants was 22; within that, the median number of those who had taken the drug was 18.
Almost 40% of the studies relied for their primary outcome assessment — the tool by which results of taking psilocybin are observed and categorized — on self-reporting and questionnaires as opposed to more objective clinical measures.
The similarities to the advocacy-first, science-second approach the FDA detected with Lykos is uncanny. The release of this document is a point-by-point case study in just how shaky the science here really is, and how overblown the claims of help. In that sense, it’s a massive public service.