The revelation that a mid-stage trial of Biogen/Eisai’s beta-amyloid MAb BAN2401 had shown glimmers of hope has fired up the Alzheimer’s world and, along with it, the stock markets. Now comes the tricky bit: providing hard data to convince investors that BAN2401 could be clinically relevant.
There is not long to wait, as full data from this trial, Study 201, are to be unveiled at the AAIC conference in Chicago on July 25, making this meeting unmissable. Could BAN2401 be the first drug to halt the course of Alzheimer’s and, if so, when could it be filed for approval?
The most optimistic investors, having already convinced themselves that the answer to the first question is yes, now want the second answered. Many will look to the AAIC for evidence of a strong clinical effect to back up the bull case, which has already been bolstered by the US FDA earlier saying it might accept a single study to back registration of an Alzheimer's drug.
That said, however much hope has been generated by Eisai and Biogen saying the results would be discussed with regulatory agencies, investors could be disappointed.
Regulators must always be consulted on a path forward, so there is no surprise here. And the FDA might indeed accept a single trial – but this must be well controlled and backed by confirmatory evidence. Study 201 used Bayesian statistics and adaptive randomisation, and at the end of the day it failed; any effect shown is thus at best hypothesis-generating.
Many at AAIC might struggle to come to terms with this, especially given the complexity of the Bayesian statistics used to analyse Study 201’s primary 12-month endpoint. Only the final, 18-month analysis used a traditional analysis, but even though this was claimed to yield statistical significance for a 10mg/kg biweekly dose this result is exploratory because of the primary fail.
Key data points
Beyond this investors should look for several data points when detailed results from Study 201 are unveiled at the AAIC’s oral session on July 25, with a concurrent presentation on Eisai’s website.
Clearly they will want to know the 10mg/day dose’s effect size in terms of the ADCOMS endpoint. Study 201’s design reckoned on a 25% reduction in rate of decline, so this is a likely minimum, but the absolute numbers will depend on the baseline score, which will have been influenced by the fact that 64% of enrollees had prodromal disease and the rest mild Alzheimer’s (all were beta-amyloid positive).
What effect was seen in each of these two populations will also determine market assumptions and phase III design. Furthermore, it is not clear whether the FDA will accept the as-yet unvalidated composite ADCOMS measure as a registrational endpoint (Biogen rekindles the eternal beta-amyloid flame, July 6, 2018).
Given Study 201’s adaptive design it is not clear how many of its 856 subjects received the active dose – Evercore ISI’s Umer Raffat reckons on 169-217. This number could help determine how strongly any clinical effects are viewed.
At least as importantly, investors will want more information about the brain swelling side-efffect ARIA-E, said to have been experienced by 10% of 10mg/day subjects. It is not clear why, when BAN2401 targets soluble protofibrils rather than solid beta-amyloid, it caused ARIA-E, especially when none was seen in phase I.
That said, 10% is still relatively low. BAN2401 could yet become Biogen’s Alzheimer’s lead, even though aducanumab is further advanced, if it can show efficacy with reduced toxicity.
And how does all this relate to the stock markets? The combined valuations of Biogen, Eisai and Bioarctic, BAN2401’s originator, surged $17bn on July 6 and have yet to dip, illustrating another problem: reading too much into Study 201's results.
Study 201 is a hypothesis-generator, but it cannot show hard evidence of BAN2401 slowing Alzheimer’s; all the full data can do is point a way to the design of a pivotal trial that must still prospectively show a clinical effect.