An antibody that targets clumps of soluble amyloid-beta before they aggregate into plaques has passed its phase 2 challenge, posting statistically significant results on a combined measure of cognition and function while decreasing amyloid deposition in the brain.
Full data won’t be released until a late-breaker session on July 25 at the Alzheimer’s Association International Conference in Chicago, so it’s still impossible to fully dissect the 18-month, dose exploration study. But according to codevelopers Eisai and Biogen, the significant clinical improvements accrued to the highest dose tested (10 mg/kg intravenously, twice a month) and were evident as early as 6 months. Amyloid cleared in a dose-dependent manner as well.
Dr. Lynn Kramer
Although there is no available information on P values or effect sizes, the trial design designated success as at least a 25% reduction in the rate of decline over 1 year, relative to placebo.
In an interview, Lynn Kramer, MD, of Eisai declined to give further details. “I will say, however, that we believe these changes are clinically meaningful,” said Dr. Kramer, chief medical officer of the company’s neurology division.
These results make BAN2401 the first antiamyloid antibody to score significant results in both cognition and amyloid brain imaging a clinical trial, and warrant tempered optimism, according to Keith Fargo, PhD, director of scientific programs for the Alzheimer’s Association.
“We can’t know whether this is a breakthrough until we get past phase 3,” Dr. Fargo said in an interview. “In drug development, you’re never done until you’re done.”
Richard J. Caselli, MD, agreed with the tempered enthusiasm.
Dr. Richard J. Caselli
“This is very encouraging, but we await the full data and ultimately the phase 3 trial,” said Dr. Caselli, professor of neurology the Mayo Clinic Arizona in Scottsdale and is also associate director and clinical core director of the Arizona Alzheimer’s Disease Center. “I would also say that 25% slowing in disease progression is nice and unprecedented but clinically not exactly earth shattering. If it holds up in phase 3, the cost-benefit will need to be carefully weighed.”
Nevertheless, the publicly available data seem to lay a firm foundation for future studies, he said. “It’s unusual to see this kind of clarity in both clinical and biomarker results in a phase 2. There’s a history of large drug companies taking things from phase 2 to phase 3 on data that have been less clear than this, teasing out results from subgroups or using biomarker but not clinical data to make a decision on moving forward.”
The BAN2401 research team staked this commitment in the 2016 paper describing the study methodology. “They wanted to have clear evidence on both biomarker and clinical efficacy to allow them to make a decision [to move into phase 3],” Dr. Fargo said. “I don’t know if they are or not, but I’d be surprised if they don’t, and I think they would be on solid footing.”
BAN2401 selectively binds to amyloid-beta protofibrils – large AB oligomers that are still soluble – and targets them for clearance. It was originally developed by Swedish biopharma company BioArctic; Eisai acquired the molecule in 2007 and entered the Biogen deal in 2014.
The study randomized 856 patients with mild cognitive impairment or early Alzheimer’s dementia to six treatment arms: BAN2401 2.5 mg/kg biweekly, 5 mg/kg monthly, 5 mg/kg biweekly, 10 mg/kg monthly, and 10 mg/kg biweekly, or placebo. All patients had PET-confirmed amyloid brain pathology, a key baseline requirement; only one other antiamyloid agent (Biogen’s antiamyloid antibody aducanumab) has completed a phase 2 study in a purely amyloid-positive cohort. Before PET imaging, patients with non-Alzheimer’s pathology comprised up to 30% of Alzheimer’s drug studies, which researchers say confounded results and likely contributed to the long string of antiamyloid failures.
The coprimary endpoints were reduction of brain amyloid on PET scan and slowing of progression as measured by the Alzheimer’s Disease Composite Score (ADCOMS), a new tool developed and promoted by Eisai. ADCOMS combines measures from the Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-Cog), Clinical Dementia Rating Sum of Boxes (CDR-sb), and the Mini-Mental State Examination (MMSE). In devising ADCOMS, researchers chose individual components of each tool that are most likely to change in very-early-stage disease. The scale was validated retrospectively in 1,160 patients who were included in four datasets. This is the first time it’s been used in an Alzheimer’s study.
The BAN2401 trial employed a Bayesian adaptive randomization, a computer-driven algorithm designed to drive patients to the two most effective doses. Often seen in trials of faster-progressing diseases, like cancer, it is unusual in Alzheimer’s studies, Dr. Fargo said.
“I would say it’s unorthodox, but it doesn’t give me heartburn. It’s an adaptive trial design that’s been considered hard to do in the AD field, because the changes between placebo and treatment groups tend to be very slow. In studies that have a very clear biomarker endpoint, like tumor size for example, that everyone can agree on, you can quickly see which treatment arm is most effective. In general, we haven’t been able to do this in AD because the signal changes so slowly.”
Dr. Keith Fargo
The BAN2401 algorithm assessed outcomes monthly, and reallocated incoming subjects according to the most recent efficacy data. But the 12-month primary endpoint assessment, a widely publicized failure, was “unfortunate,” Dr. Fargo said.
“This study had a built-in 12-month endpoint, at which time the sponsor could declare either futility or early success, and stop it for either one.” If neither was achieved, the study was to continue for 18 months and reassess outcomes. This was the situation Eisai and Biogen encountered in December when they announced that BAN2401 had failed to achieve its clinical and imaging endpoints but that the study would continue as planned. That news generated a large amount of negative press for the companies – which saw stock prices plunge – and contributed to media and consumer confusion about the trial design’s validity, Dr. Fargo said.Does herpesvirus join the list of infective degenerative brain diseases?
The 18-month data released on July 5 also included some information on adverse events. Amyloid-related imaging abnormalities, both edematous and hemorrhagic (ARIA-E and ARIA-H) occurred in an undisclosed number of patients. Dr. Kramer said ARIA occurred in “not more than 10% in any of the treatment arms” and in less than 15% of patients who carried an apolipoprotein E e4 allele. Only a portion of these patients were included in the final analysis, and this may have influenced the results somewhat negatively, Dr. Kramer said. In the first years of the study, before ARIA was understood to be largely subclinical and self-resolving, any patient who developed it was dropped.
“We initially dropped every patient who had ARIA, and this is one of the issues that actually make the observed effect [of BAN2401] probably a little bit conservative. Patients with ARIA are typically responding to amyloid removal, and they could be expected to do better.”