ARTICLE IN BRIEF
Esai and Biogen are seeking accelerated approval for an antibody to BAN2401, which reduced cognitive decline by two measures in early Alzheimer's disease after 18 months. Independent experts offered cautious optimism about the findings but also expressed skepticism about whether it was ready for clinical practice.
CHICAGO—The antibody BAN2401 used at the highest dose reduced cognitive decline by two measures in early Alzheimer's disease after 18 months, while also reducing amyloid burden, according to results presented by the drug's manufacturer here at the Alzheimer's Association International Conference (AAIC).
“This is the first large clinical trial to support the amyloid hypothesis,” said Lynn D. Kramer, MD, FAAN, chief clinical officer of the neurology business group at Eisai.
The results were eagerly awaited, but several experts not associated with the trial were mostly restrained in their response to the data. They noted that questions remain about the study methodology, which involved response-adaptive randomization, and that further assessment of the drug is needed in a larger phase 3 trial.
The findings come just several months after Eisai, developing the drug with Biogen, announced that the drug hadn't met its more rigorous, 12-month endpoint in its interim analysis. That endpoint involved a much higher threshold, while the new 18-month endpoints were based on a standard statistical analysis, Dr. Kramer said. A standard analysis was not done at the 12-month mark in order to preserve the integrity of the trial as it continued on, he said.
In the 856-person trial, all patients had amyloid-beta at baseline and were randomized to placebo or to one of five active treatment groups, from 2.5 mg/kg bi-weekly to 10 mg/kg bi-weekly. After the first 196 patients were randomized in a traditional fashion, a computer algorithm was used to randomize the non-placebo patients to the treatment dose it predicted would have the best results. After that, the algorithm was used to assess the results every three months and determined the dosing group that would most likely benefit patients, and they were re-assigned accordingly. The untraditional randomization made the results a bit more difficult to assess, independent experts said.
The most effective dose was ultimately the highest dose. But many more patients were randomized to the second-highest dose (253 compared to 161), because the regulatory agency, the European Medicines Agency, insisted that no patients who were positive for the high-risk genetic marker, apolipoprotein E4 (APOE4), be randomized to the highest dose, unless they'd already received six months of treatment; they shared concerns about amyloid-related imaging abnormalities thought to represent edema (ARIA-E), Dr. Kramer said. That resulted in 26 patients dropping out of the highest-dose group, which left 48 in the highest dose group.
Also unusual was the use of a measurement tool developed in-house by Biogen and Eisai. Called the Alzheimer's Disease Composite Score (ADCOMS), it is meant to assess cognition and function in early Alzheimer's by drawing on information from four mild cognitive impairment study datasets and four mild Alzheimer's disease study datasets and uses certain components geared toward early disease that were pulled from other scales.
The ADCOMS used items from the Alzheimer's Disease Assessment Scale-Cognition (ADAS-Cog), the Mini mental Status Examination (MMSE), the Clinical Dementia Rating-Sum of Boxes (CDR-SOB), and the Neuropsychological Test Battery (NTB), and incorporated activities of daily living (ADL).
ADCOMS items included delayed word recall, word recognition, personal care, judgment and problem-solving, ability to do home tasks and hobbies, and drawing.
Researchers found that the 10 mg/kg dose given biweekly produced the only significant slowing of cognitive decline. Patients receiving this dose had a 30 percent slower decline than patients on placebo on the ADCOMS scale (p=.034). The worst possible score on the ADCOMS is 1.97, and the baseline score of participants in the trial was 0.33.
“So the changes in terms of the score are very small and the patients starting in the trial are very mild compared to the traditional endpoints,” Dr. Kramer said.
The highest dose was also found to slow decline by 47 percent compared to placebo on the ADAS-Cog scale (p=.017). Patients in the placebo group fell by just shy of five points over the 18 months, compared to about 2.5 points for those on the highest dose. The other doses produced slowing of decline but not significantly compared to placebo.
BAN2401 did not produce a significant slowing of decline on CDR-SOB, even at the highest dose.
Researchers also found that the highest dose significantly reduced amyloid burden on positron emission tomography (PET) by 70 units on the centiloid scale (p<.0001) over the placebo group, in which patients saw no drop. Dr. Kramer said that 81 percent of patients converted from being amyloid-positive at the start of the trial to amyloid-negative by 18 months on a visual read of PET imaging.
Overall, there were no differences in serious treatment-related adverse events across the groups. But amyloid-related imaging abnormalities thought to represent edema (ARIA-E) was seen in 9.9 percent of patients in the highest dose group, compared to just 0.8 percent in the placebo group. Researchers said they were mild to moderate findings and that they typically resolved within four to 12 weeks.
Dr. Kramer said the company will be talking to regulatory authorities around the world to seek accelerated approval and to get guidance on how to go about future studies of the drug.
“In our opinion, we have fairly conclusive results,” he said. “We view this as certainly robust enough results to approach regulatory authorities to discuss next steps.”
After his talk at AAIC, Dr. Kramer was asked, “How much of the cognitive effects could be due to the removal of the people with APOE4 from the high-dose group?” [Those with APOE4 were not removed from the placebo group.] Dr. Kramer did not answer directly, but said those patients were instead moved to the second-highest-dose group. He also said he could not comment yet on results of APOE4-positive patients compared to those who were APOE4-negative.
He said the investigators “argued vehemently” against having the APOE4 patients removed from the highest-dose group. “And as you can see, the results suggest we might have been right,” he added.
Independent experts said they were cautiously optimistic that the findings suggested that reducing amyloid can improve cognitive symptoms in AD, but their enthusiasm was tempered by questions about the study design — whether the different populations in the study arms reflected true randomization — and whether the decision to exclude patients with APOE may have skewed the results. They told Neurology Today that the impressive findings in this phase 2 trial need to bear out in larger phase 3 trials.
Paul Aisen, MD, director of the Alzheimer's Therapeutic Research Institute at the University of Southern California, said the patient imbalance “does create a possible bias” but he did not see it as a major concern.
“We expect E4 carriers to progress a bit more rapidly, even among biomarker-confirmed participants,” he said. “But I do not think this could account for the apparent cognitive/clinical benefit in the high-dose group. The results are exciting, and consistent with aducanumab findings (in which) dramatic amyloid removal is associated with cognitive/clinical benefit at the stage of ‘early’ AD. Of course, a more standard-design confirmatory trial is necessary. I expect that the cognitive benefit would be even greater if the antibody were administered at an earlier, asymptomatic stage of disease.”
David S. Knopman, MD, FAAN, professor of neurology at Mayo Clinic in Rochester, MN, noted that the initial gross point difference on the ADAS-Cog for BAN2401 over placebo — the drug was about 2.5 points better — was not much different from what has been seen initially for prior Alzheimer's therapies, such as donepezil, that in the long run have proven rather disappointing.
“The question is: What does that effect size mean?” he said. “Over a year, who knows?”
Longer and larger phase 3 trials are needed to truly gauge how the drug is likely to affect patients' lives, he said. “This was intended to find a dose, it wasn't intended to be a definitive study.”
Julie A. Schneider, MD, MS, professor of neurology at Rush University Medical Center, said the unusual methodology means it will take a bit of time to fully grasp the results. “We're really going to have to study how this was done in order to feel more confident about the findings,” she said. “We have to see what this looks like clinically.”
She said that she is hoping for a steeper decline in the treatment group compared to the placebo group. “We really want to see a change in the slope of decline, rather than just a bounce up,” she said.
Maria Carrillo, PhD, chief science officer at the Alzheimer's Association, said these findings indicate that the drug could be a part of a broader attack on the disease.
“The scientific community is in agreement that combination therapy is the future and that there won't be a silver bullet,” she said. “So being able to delay the disease even for a year can have a huge impact on lives.”
She said a multitude of treatment approaches — treatment of vascular dementia causes, and other amyloid approaches, such as tau — “can give us hope for a higher magnitude of effect that can lead us to treating the disease chronically instead of how we're just waiting [with] symptomatic treatments and ultimately [watching] people continue to die of this disease.”
Dr. Kramer is an employee of Eisai Co, the manufacturer of BAN2401. Dr. Aisen is an advisor and receives consulting fees from Biogen regarding aducanumab but is not involved with the current trial. Dr. Knopman serves on the data and safety monitoring board and receives personal compensation for the DIAN-TU study.