On March 23, a number of NeuroPerspective subscribers and sources forwarded to us a ‘report’ prepared by Kerrisdale Capital on SAGE Therapeutics. Subheaded ‘Overhyped Lead Drug Headed for Failure,’ it gave the illusion of a detailed, seemingly well-researched critique of SAGE Therapeutics and SAGE-547 in Super-Refractory Status Epilepticus. However, even after a cursory review, as we noted to those correspondents, this report appeared to be slanted against SAGE, based on several false premises. With other obligations at hand, we did not assess the report more thoroughly until today: The more we have delved into it, the meticulous detail with which the report was prepared, far from clarifying the issues at hand, clearly appears to be in the service of camouflage, not clarity. A few points should be made:
1) Mechanism: Kerrisdale claims that SAGE-547’s mechanism of action is not dissimilar from that utilized by other agents deployed in the treatment of SSRE, citing binding-mechanism reports. Predicting the clinical impact of psychotropic drugs based on in vitro binding studies has often been a fool’s errand: Any number of companies have claimed that the addition or deletion of activity at some receptor subtype or another would be predictive of its clinical impact on certain clinical symptoms or side effects, and such predictions often fall short (e.g. vilazodone’s profile was predicted by its developer(s) to offer less impact upon sexual functioning in patients taking it for depression–it has not turned out that way). The clinical ‘truth’ of a putative mechanistic profile is ultimately only determined via human efficacy testing. Kerrisdale’s claim that SAGE-547 “is just not that special” because it relies upon GABA-A receptor potentiation, including an effect upon extrasynaptic receptors, comes close to being irrelevant. Receptor binding is not a binary, yes or no proposition–and the pattern and duration of inhibitory effects constitutes a complex domain. The relative value of a drug in this category can only be determined by its effect in the clinic, not by putative binding profiles.
One other note, regarding ‘duration:’ Kerrisdale makes reference to some cases of eventual improvement that appeared temporally incongruent with SAGE-547’s half-life, as if the former is incompatible with the latter, and the drug cannot have had an impact eventual weaning from anesthesia. The relationship between half-life and clinical effect is nowhere this clearcut, one recent example is the use of ketamine/esketamine in depression studies: Janssen’s recent Phase II showed that patients experienced antidepressant effects up to eight weeks following the last administration, though esketamine’s half-life is measured in hours.
2) The Treatment Population: Kerrisdale stakes the bulk of its case on the premise that SAGE-547 clinically does not do anything that other ‘third-line’ drugs cannot do. They provide a cascade of citations from the literature intended to make the case that drugs like midazolam and levetiracetam can also rescue patients from status epilepticus–citing meta-analyses that report a “64-68% base rate” of control provided by other GABAergic third-line therapies, albeit in populations that included both refractory and super-refractory cases. Kerrisdale claims that many of the ‘refractory’ patients had a long-enough period of ICU treatment that they should be considered super-refractory, and thus a suitable comparison group vis-a-vis SAGE’s treatment population. This is specious: Kerrisdale here is defining SSRE as based on duration of status epilepticus, whereas SAGE’s enrollment criteria in Phase II required that patients be treatment-refractory: “the presence of one or more breakthrough seizures>6 hours after initiation of the continuous IV AED/third-line agent (e.g. pentobarbitol, midazolam, propofol.” In other words, not responsive to the very drugs that Ferrisdale claims are just as efficacious as SAGE-547. The Phase III enrollment criteria includes these ‘third-line’ treatment-refractory patients, but adds patients who have been admitted to an ICU but have not yet been given a standard third-line treatment.
3) Placebo comparator: Kerrisdale predicts that the SAGE-547 Phase III trial will fail, based on the premise that it is not really different from other third-line drugs. But this trial does not involve an active comparator, SAGE-547 is being compared to a placebo, which even Kerrisdale would have to admit, has a different binding-profile and mechanism than does SAGE-547. Indeed, the Phase III inclusion of patients not yet tried on a third-line drug means that SAGE-547 has a wider avenue to success, not restricted to patients so ‘super-refractory’ that even the standard treatments have failed. This argument is not even logical.
4) Speaking of illogical, Kerrisdale complains that SAGE-547 does not correct the underlying pathology accounting for the SSRE, that it is a “band-aid.” No claim has ever been made that SAGE-547 is supposed to be curative. The endpoint of the trial is to provide at least 24 hours that are seizure-free after being weaned. This argument is the proverbial ‘red herring,’ completely irrelevant to the trial, and the goal of treatment.
5) Size of population: Kerrisdale claims that the incidence of SSRE is far lower than SAGE claims. Based on Kerrisdale’s performance on points 1-4, their credibility is not to be assumed. But in any event, this is something that will be assessed when SAGE-547 comes to market, and we would expect that if it is successful in patients refractory to other third-line agents, that it would become the standard of care. Kerrisdale will have exited its short position two or three years before we will have any sales data.
All in all, the more thoroughly we read Kerrisdale’s report, the less convincing it became. It reminds us that the 1980s rock band ‘Cheap Trick’ has recently made a comeback of sorts–Kerrisdale’s report exemplifies the kind of short-selling cheap trick that never went away, and it provides a vivid illustration of how a shower of informational chaff can be used to distort rather than illuminate. While we are optimistic about SAGE-547’s Phase III prospects in SSRE, extrapolating success in a small Phase II trial to a (relatively) large Phase III study is not to be automatically assumed–we look forward to seeing the data.
–NI Research does not have a position in SAGE Therapeutics. We have no relationship with SAGE, other than SAGE does subscribe to NeuroPerspective, as do most CNS companies. We did not contact SAGE Therapeutics in the preparation of this note, because it first of all was not necessary, and secondly, SAGE is in the awkward position of being a public company conducting a pivotal trial for the medication in question. Their ability to respond fully is thus somewhat restricted, an asymmetric position that short-sellers try to exploit.