• Biogen and partner Eisai announced Wednesday that a once-failed experimental drug for Alzheimer’s disease slowed cognitive decline in patients with early forms of the memory-robbing condition, a finding that offers equal measures of optimism and caution in a field that has yet to hit on an effective treatment.
• After 18 months of study, patients who received the highest dose of the drug, called BAN2401, did 30% better than those given placebo as measured by a clinical test. Treatment with the drug also led to dose-dependent reductions in the sticky amyloid plaques characteristic of Alzheimer’s.
• Nearly every compound tested in Alzheimer’s has eventually failed, making Wednesday’s results a rare success. Questions and caveats remain, however, compounded by the negative results seen at one year of treatment during the same study.

Even after dozens of clinical setbacks in Alzheimer’s, expectations for BAN2401 were high after Biogen and Eisai unexpectedly announced earlier this month that a Phase 2 study had shown a cognitive benefit to the drug.

Detailed data from that trial were released for the first time Wednesday at the Alzheimer’s Association International Conference, and the results simultaneously exceeded predictions, yet fell short of hopes for a home run.

“We’re cautiously optimistic,” said Keith Fargo, director of scientific programs at the Alzheimer’s Association, in an interview. “If they pan out to be true in Phase 3 clinical trials, we are talking about disease modification and that is something the community desperately needs.”

Wall Street, on the other hand, was less nuanced in its reaction. Biogen stock fell by 11% at market open Thursday, while shares in AC Immune, another Alzheimer’s drugmaker, slumped by a similar percentage.

The trial enrolled 856 patients with early Alzheimer’s disease into six groups: five different dose regimens of BAN2401 and a placebo group. Treatment with BAN2401 cleared amyloid plaques in all dose groups, with 81% of those in the highest dose cohort converting from amyloid positive to amyloid negative at 18 months.

Participants who received that high dose — 10 mg per kg biweekly — saw a statistically significant 30% slower decline from baseline than those given placebo as measured by a clinical endpoint Eisai scientists developed to test cognition and function. Known as ADCOMS, the measure combines elements of some of the better-known disease scales like ADAS-cog and CDR-SB.

On ADAS-cog alone, the high dose of BAN2401 slowed decline by 47% versus placebo, while treatment led to a positive but not statically significant 26% difference on CDR-SB.

Prior to Wednesday, a survey of investors by Mizuho suggested a 20% reduction on the ADCOMS measure would be considered positive.

Yet the results also came with substantial caveats. The study was designed to primarily test the efficacy of BAN2401 at 12 months using statistical methods designed to more quickly indicate whether the trial would succeed. On that measure, the study failed, making the results at 18 months secondary and only hypothesis generating.

And while the highest dose led to a statistically significant reduction on the ADCOMS measure, that was not the case for lower doses of the drug.

This is important because enrollment of carriers of a genetic risk variant into the highest dose was restricted following a 2014 request from a non-U.S. regulator. This resulted in fewer patients with this variant, known as APOE4, in the high-dose group and more in the second-highest treatment group.

“It is a reasonable hypothesis that the reduction of APOE4 positive individuals in the highest dose arm could potentially explain the differences in the cognitive outcomes,” said Fargo. “That’s a scientific question that we don’t know the answer to yet but certainly that is a legitimate criticism.”

In an interview Wednesday, Samantha Budd Haeberlein, VP of late-stage clinical development for Alzheimer’s at Biogen, said the regulator’s request to limit allocation of APOE4 carriers was related to concerns over a side effect known as ARIA. Individuals with the APOE4 variant are thought to be at greater risk of developing ARIA, or amyloid-related imaging abnormalities.

No more than 10% of people treated with BAN2401 at any dose level experienced ARIA with edema, or brain swelling.

Biogen and Eisai’s data still mark a notable achievement in the Alzheimer’s field. A spate of recent trial failures and dead-ends has shaken confidence in the theory that wiping out amyloid plaques could slow or halt cognitive decline in Alzheimer’s patients.

BAN2401 offers a counterpoint to those concerns and boosts optimism in further study of the drug.

“The results look real,” said Howard Fillit, chief science officer at the Alzheimer’s Drug Discovery Foundation, in an interview. “There is a clinical signal here that is optimistic and exciting.”

Biogen and Eisai plan to discuss the data with major global regulators on a path forward for BAN2401.

“I think we believe some further work is needed but what that looks like is something we need to work out,” Haeberlein said.

While BAN2401’s next steps are yet to be decided, attention now turns to Biogen and Eisai’s aducanumab, another anti-amyloid drug currently in two large Phase 3 trials. Results are expected in late 2019 or early 2020.