It is a measure of the continuing fascination with Alzheimer’s disease that the tiniest glimmer of hope, in a failed study whose analysis has been tortured extensively, sent Biogen’s valuation up by $12bn this morning.
Though the data, from an adaptive trial of the beta-amyloid MAb BAN2401, should be treated with extreme caution they will renew hope that the beta-amyloid hypothesis does, after all, hold the key to tackling Alzheimer’s. As Biogen awaits next year’s all-important phase III readout for its later-stage project, aducanumab, here are 10 points that investors need to bear in mind.
1. What has happened?
Biogen/Eisai’s phase II Study 201 has yielded positive results for BAN2401 at 10mg/kg biweekly, the highest of five doses tested, at 18 months. No numerical data have been released, but the endpoints said to have been met are the ADCOMS disease score – a measure of efficacy – and brain amyloid reduction as measured by PET.
2. Wait, didn’t this study already fail?
Yes. In December Biogen said Study 201 had failed its primary endpoint: efficacy, according to ADCOMS, at 12 months. None of the dosing levels had at that point proved successful, and the trial was to remain blinded and proceed to a comprehensive, final 18-month analysis.
3. How was this trial designed?
Study 201 used adaptive randomisation to enrol 856 subjects with mild Alzheimer’s disease across five dose levels. It used a complex Bayesian statistical model to yield a faster result than traditional designs.
The 12-month analysis was relevant for the primary endpoint, while the 18-month “final” analysis was for secondary measures, in addition to 16 interim analyses for stopping for futility or lack of safety; none of the stopping criteria were met.
4. So how reliable are the 18-month data?
When a trial fails its primary endpoint, secondary measures become hypothesis-generating only. The effect size has not been disclosed, and neither have the number of patients on the highest dose, powering assumptions or statistical penalties. None of the other four doses has yielded promising results, though 10mg/kg biweekly is now said to have “begun to show a statistically significant” ADCOMS benefit at six and 12 months.
Bayesian analysis is complex and makes assumptions that sometimes seem at odds with a more traditional statistical approach. But if it is accepted that an adaptive trial merely aims to guide towards a dose and endpoint that must still be proved prospectively in phase III then there is little wrong with the data as they stand.
5. How reliable is the ADCOMS endpoint?
ADCOMS is a novel, non-validated measure developed by Eisai, and it is anyone’s guess as to how acceptable the FDA might see it as a pivotal endpoint. It is relevant to the extent that the more common CDR-SB score – suggested by the FDA as valid in early Alzheimer’s – makes up some 60% of this composite endpoint.
6. What is the trial’s most positive takeaway?
In addition to the hints of efficacy, safety is key. Vasogenic oedema (ARIA-E) has been a problem with other beta-amyloid MAbs, including aducanumab, leading to suggestions that these drugs lack a therapeutic window. BAN2401 previously caused virtually no ARIA-E, and the incidence in Study 201 with 10mg/kg biweekly is around 10%. This is just about acceptable; higher dosing might be possible.
7. So is the amyloid hypothesis alive again?
Biogen says the data provide “compelling evidence to support the amyloid hypothesis as a therapeutic target for Alzheimer’s disease”. Study 201 will be seen as lending the hypothesis fresh impetus, though many might argue that, even after multiple failures, the enthusiasm never waned.
8. Does Biogen now have a new phase III Alzheimer’s candidate?
Yes. The phase II data are very far from an unequivocal sign of efficacy, but in the current climate they are enough to warrant further work. Biogen aims to “continue to work towards ... delivering BAN2401 to patients ... as early as possible”.
9. What does this mean for aducanumab?
Aducanumab and BAN2401 are both MAbs based on the IgG1 isotype, and aducanumab’s pivotal trials are in early Alzheimer’s. This is probably where the similarities end. Aducanumab’s studies measure a sole primary endpoint, change in CDR-SB score at 18 months, and the project binds the solid form of amyloid beta in brain plaques; BAN2401 targets the more complex soluble protofibrils.
In a nutshell, beyond Biogen’s ability to design Alzheimer’s studies and its ability to torture data, there is very little in terms of read-across.
10. Is the exuberance warranted?
Biogen opened up 20% this morning, while Bioarctic, a virtually unknown Swedish company that originated BAN2401, traded up over 200%. Notably, Biogen’s stock now stands 10% above where it was trading just before Study 201 was said to have failed its primary analysis.
It is acceptable to reverse December’s write-off of BAN2401 as a viable asset, but Biogen’s stock movement implies more, seemingly pricing in a higher chance that aducanumab will succeed. This is a dangerous and unwarranted assumption.
The best that can be said is that Biogen is probably seeing something in the latest data. But what that is, and its relevance, still remain subject to full Study 201 disclosure – and then the trick needs to be repeated prospectively in a pivotal trial.
Given investors’ optimism about Alzheimer’s, however, such considerations probably matter little right now.
|BAN2401||Study 201||856 early Alzheimer's pts, adaptive trial, Bayesian analysis||Reported||NCT01767311|
|Aducanumab||Engage||1,605 early Alzheimer's pts, 18mth CDR-SB is primary endpoint||Nov 2019||NCT02477800|
|Aducanumab||Emerge||1,605 early Alzheimer's pts, 18mth CDR-SB is primary endpoint||Feb 2020||NCT02484547|