Biogen Inc. and Eisai Co. Ltd. reported that the highest dose of BAN2401 significantly cleared amyloid from the brain and improved cognition in patients with mild Alzheimer's disease, but questions remain about the exclusion of certain individuals from the high dose arm and whether that confounded the results from the large Phase II trial.
Reflecting some investor uncertainties over the data, Biogen's stock declined by more than 11% in after-hours trading after Eisai Chief Medical Officer Lynn Kramer reported the much-anticipated detailed results at 18 months from the 856-patient Study 201, presented on July 25 as a late-breaking oral presentation at the Alzheimer’s Association International Conference (AAIC) in Chicago.
In Tokyo, Eisai’s share price also was strongly down in mid-afternoon trading, falling by a similar percentage as investors digested the new data.
Results for the amyloid-beta protofibril-targeting antibody were largely in line with expectations following the top-line data released recently, which generated huge interest in the Alzheimer’s community and around the validity of the amyloid hypotheses, which is that accumulations of the protein contribute to the development of the disease.
One exception was the new disclosure that patients who carried the APOE4 gene – associated with early onset of Alzheimer's disease and rapid decline – were excluded from the high-dose BAN2401 group at the request of ex-US regulators.
Debate Over APOE4 Role
APOE4 carriers tend to experience amyloid-related imaging abnormalities (ARIA) at higher rates than non-carriers, which was the case in the BAN2401 study. However, there was some debate during Biogen's same-day call with analysts and investors about whether APOE4 carriers have a worse response to treatment.
If so, the higher dose of BAN2401 would perform better than placebo and the lower doses, because that arm of the trial had fewer APOE4 carriers.
Biogen Vice President of Clinical Development Samantha Budd Haeberlein argued during the company's call that APOE4 status has an impact on the age of Alzheimer's disease onset, but not on the efficacy of treatment.
Haeberlein noted that amyloid was cleared from patients' brains in a dose-dependent manner, according to PET scans, which would lead her to believe that APOE4 status did not play a role in BAN2401's effect on cognitive decline.
Biogen Executive Vice President and Chief Medical Officer Alfred Sandrock added that the APOE4 carrier versus non-carrier subgroup analysis probably will be the first subgroup analysis that Biogen and Eisai will conduct, and those data would be shared as soon as possible.
The second subgroup analysis, he said later in the call, probably will look at prodromal (pre-symptomatic) patients versus those with mild cognitive impairment.
Analysts seemed to agree that the APOE4 carrier versus non-carrier analysis is most important and that no conclusions can be drawn about the BAN2401 data until that is done.
Evercore ISI analyst Umer Raffat looked at other studies to see if APOE4 status impacted drug efficacy and had mixed findings, but he said the BAN2401 effect on cognition may be large enough that even with APOE4 patients in the mix the results could still be significant.
Amyloid Clearing's Link To Cognitive Benefits Depends On Analysis
Biogen and Eisai reported earlier in July in the top-line results that BAN2401 at its highest dose provided statistically significant reductions in amyloid plaques and slowed the rate of cognitive decline in the large Phase II trial, which enrolled patients with mild cognitive impairment (MCI) due to Alzheimer's disease or mild Alzheimer's with confirmed amyloid pathology in the brain.
BAN2401, which targets amyloid beta protofibrils by selectively binding to and neutralizing soluble, toxic amyloid beta aggregates, was originated by Swedish bioventure BioArctic and acquired globally by Eisai in late 2007, then brought under the 2014 multi-asset Alzheimer's drug development partnership with Biogen.
Eisai’s strategic alliance with Biogen in Alzheimer's dates back to 2014. As well as BAN2401 and the beta amyloid-targeting antibody aducanumab, this includes the Eisai BACE (beta-site amyloid precursor protein) inhibitor elenbecestat and a Biogen anti-tau antibody.
The firms will co-promote any marketed drugs under the alliance in major markets, and for BAN2401 (and elenbecestat), they are equally splitting overall costs including R&D expenses, with Eisai to book sales for both drugs and profits to be shared equally.
Biogen suggested on July 24, the day before the AAIC presentation, that the company and its partner may be able to pursue accelerated approval in the US based on the Phase II results, but noted such a decision depends on future discussions with the FDA. (Also see "Enthusiasm Builds For Biogen’s BAN2401 In Alzheimer's, But SMA Gene Therapy Hits A Snag" - Scrip, 24 Jul, 2018.)
The data presented on July 25 don't make the possibility of approval ahead of Phase III data any clearer given the APOE4-related concerns. Some analysts have already said that a further large randomized Phase III study with a pre-specified endpoint appears necessary, but this has yet to be confirmed by Eisai or Biogen.
Patients were treated with 2.5 mg/kg of BAN 2401 every two weeks, 5 mg/kg once-monthly, 5 mg/kg bi-weekly, 10 mg/kg monthly, 10 mg/kg bi-weekly or placebo. A Bayesian adaptive randomization trial design meant that patients could be randomized into treatment arms with a higher likelihood of efficacy after interim assessments of the data.
That meant that there were 247 patients in the placebo arm, 52 in the 2.5 mg/kg bi-weekly arm, 51 in the 5 mg/kg monthly arm, 92 treated with 5 mg/kg bi-weekly, 253 in the 10 mg/kg monthly arm and 161 in the 10 mg/kg bi-weekly arm.
The primary endpoint was a Bayesian assessment at 12 months of the likelihood of success in terms of the Alzheimer's Disease Composite Score (ADCOMS), which was not statistically significant at any dose. An 80% chance of providing a 25% slowing in ADCOMS was needed, but the highest dose of BAN2401 achieved only a 64% likelihood of meeting the 25% threshold.
However, at the 18 month stage, Biogen and Eisai reported at AAIC that the reduction of accumulated amyloid plaques in the brain was statistically significant at all doses. These were measured using standardized PET (positron emission tomography), and scans of patients treated with the highest dose of BAN2401 (10 mg/kg bi-weekly) and assessed by the Centiloid scale showed an observed mean reduction from 74.5 at baseline to 5.5 at 18 months (no p-value was reported).
A Mixed Effects Model with Repeated Measures (MMRM) assessment showed a mean reduction in amyloid load of 70 units (p<0.0001). PET scans also showed that 81% of patients treated with the highest BAN2401 dose converted from amyloid positive to amyloid negative at 18 months, which reached significance with a p value of <0.0001.
In addition, an analysis against predefined clinical endpoints at 18 months confirmed a slowing in cognitive decline from baseline based on ADCOMS, which was dose-dependent. At the highest dose this was 30% versus placebo (p=0.034), but a statistically significant slowing of decline was seen on ADCOMS as early as six months (p<0.05), and at 12 months (p<0.05).
Bayesian analysis of ADCOMS at 12 months gave an estimated probability that the highest dose slowed clinical decline more than placebo of 98%.
Cognition Effects Mixed
Biogen noted in its call that that, given BAN2401’s mode of action, there is a “complex relationship between plasma exposures and what's going on in the brain – Cmax is important to achieve.”
But in terms of effect on cognition, Biogen and Eisai reported dose-dependent effects via two out of three secondary endpoints. The highest BAN2401 dose showed a 30% slowing in clinical decline as assessed by ADCOMS compared with placebo at 18 months (p=0.034), and statistically significant slowing also was observed at six months (p<0.05) and 12 months (p<0.05).
Also, the highest BAN2401 dose slowed clinical decline by 47% compared to placebo at 18 months, according to the Alzheimer's Disease Assessment Scale-cognitive subscale (ADAS-Cog; p=0.017).
However, the slowing in cognitive decline from baseline as assessed by Clinical Dementia Rating Sum of Boxes (CDR-SB) was not statistically significant, though the companies said there was a dose-dependent slowing in cognitive decline from baseline, in excess of the pre-specified difference of 25% over the course of the study, and reaching 26% at 18 months for the highest dose versus placebo (the decline in the latter being in line with earlier research).
Dose-dependent increases in amyloid-beta levels were observed in cerebrospinal fluid (p<0.0001 at the highest dose at 18 months), and there was a statistically significant reduction in total tau levels against placebo in both of the 10mg/kg dose regimens (p<0.05).
ARIA, Infusion Reactions Most Common Adverse Events
Biogen and Eisai said their drug "demonstrated an acceptable tolerability profile through 18 months of study drug administration." The treatment-related adverse event rate was 26.5% for the placebo arm, 53.4% for the 10 mg/kg monthly treatment arm and 47.2% for the 10 mg/kg bi-weekly treatment arm.
Serious adverse event rates were 17.6% for the placebo arm, 12.3% for the 10 mg/kg monthly arm and 15.5% for the 10 mg/kg bi-weekly arm.
ARIA and infusion-related reactions were the most common treatment-related adverse events, and there was a 9.9% rate of ARIA with edema (ARIA-E) at the highest BAN2401 dose, but ARIA-E did not occur in more than 10% of patients in any of the treatment arms.
However, the rate of ARIA-E in APOE4 carriers was 14.6% at the highest dose; all patients with MRI-confirmed ARIA-E were discontinued from the study. Concerns around safety related to ARIA meant that regulators wanted to limit number of patients in the high-dose group who were carriers.
APOE4 Status May Or May Not Be An Efficacy Predictor
"Confounding the data, the proportion of subjects in the 10mg/kg [bi-weekly] arm that were APOE4+ was substantially lower than the placebo arm and other dose groups, which may have impacted the baseline rates of cognitive decline between the various arms and contributed to the appearance of benefit compared to placebo," Leerink analyst Geoffrey Porges wrote in a July 25 note.
The percentage of patients who carried the APOE4 gene was just 30% in the highest dose arm and 70% to 91% in all other arms in the Phase II study.
"The reason behind this difference in patient enrollment was due to a request from health authorities outside of the US that APOE4+ subjects be removed from the highest dose arm and randomized solely to the other lower doses of BAN2401 or placebo," Porges said.
Haeberlein said during Biogen's call that data to date for the alliance's Phase III study with aducanumab had not shown a connection between APOE4 status and treatment efficacy.
Evercore ISI's Raffat noted that aducanumab results show that APOE4 carriers may do slightly worse on amyloid-reducing therapy than non-carriers, but the difference in cognitive decline may not be enough to make an anti-amyloid drug ineffective.
The analyst also looked at data from a 1,901-patient Phase III study for Merck's BACE inhibitor verubecestat, which appeared to show that non-carriers had worse efficacy than APOE4 carriers.
The forthcoming analysis of APOE4 carriers versus non-carriers may or may not settle the debate related to BAN2401, however, without a large Phase III dataset.
“Is APOE4 confounding results? We'll find that out soon,” Biogen’s Sandrock said.
Amyloid Doubts Remain Regardless Of APOE4 Subanalysis
Datamonitor Healthcare analyst Zara Fulton told Scrip that the BAN2401 data were unconvincing in terms of the amyloid hypothesis without Phase III results showing that reducing amyloid plaques in the brain will result in a slowing of Alzheimer's disease progression.
The amyloid hypothesis has yet to be proven in a Phase III trial, despite many efforts, but Biogen and Eisai remain confident in this approach via BAN2401, the Phase III BACE inhibitor elenbecestat for which Phase II results were presented at AAIC, the anti-amyloid agent aducanumab with Phase III results expected in 2020 and other programs. (Also see "Eisai/Biogen Remain In BACE Race As Alzheimer's Contenders Dwindle" - Scrip, 6 Jun, 2018.)
"I find these Phase II [BAN2401] data unconvincing as patients on lower doses did numerically worse than those on placebo, and only the highest dose, tested in 161 patients, met the primary endpoint. Unfortunately, promising Phase II data often do not play out in Phase III trials," Fulton said.
"I understand why Biogen/Eisai would want to race for FDA approval based on these results as replicating the statistically significant benefit for the highest tested dose in a large Phase III trial is far from a sure bet," she continued.
"The companies are likely banking on the FDA being swayed by the extremely high unmet need in AD, acute public awareness and lack of disease-modifying therapies."
Biogen's CMO Sandrock stood by BAN2401 as proof of the amyloid hypothesis, however, noting that it's important to attack "more aggregated forms [of amyloid plaques] with the right antibody and the right approach."
He noted that the very next step would be to have discussions with regulators, because there is a “range of possibilities here” on how to proceed. “Colleagues at Eisai are talking about additional trials,” he added.
Leerink's Porges wrote that BAN2401 largely lived up to expectations, but more data are needed to confirm efficacy observed to date.
"The results presented today have exceeded the Street’s expectations on ADCOMS, but do not provide confirmatory evidence that BAN2401 had a meaningful benefit on standard cognitive measures due to the differences in baseline patient characteristics across the arms," the analyst wrote.
"We do believe the data today is supportive of the amyloid hypothesis, but continue to believe that aducanumab remains Biogen’s most valuable asset in this indication."
With contributions from Ian Haydock in Tokyo.