SAN DIEGO – Positive findings from a post hoc subanalysis of two unsuccessful studies represent “a major step forward in Alzheimer’s disease research” and could set the antiamyloid antibody up as a “foothold” in slowing disease progression, study investigators said at the Clinical Trials on Alzheimer’s Disease conference.
Michele G. Sullivan/MDedge News
Dr. Samantha Budd Haeberlein of Biogen (second from right) presented the data. Panelists included Dr. Ronald Petersen (at podium), and from right, Dr. Paul Aisen, Dr. Sharon Cohen, and Dr. Stephen Salloway.
After full follow-up of 78 weeks, patients with mild Alzheimer’s disease (AD) who took the highest 10-mg/kg dose for a full 14 doses experienced up to a 53% slowing of functional decline on the Clinical Dementia Rating–Sum of Boxes (CDR-SB) in one study and a 48% slowing in the other study – relative to placebo – a result that might give them “an extra year or 2” of independence; they might perhaps retain the ability to drive and even stay employed, said Sharon Cohen, MD, a panelist at the meeting’s aducanumab presentation session and a clinical investigator in EMERGE, one of two phase 3 studies from which the data were derived.
Samantha Budd Haeberlein, PhD, Biogen’s vice president and head of late-stage clinical development in Alzheimer’s disease, presented the new data. They “are complex” and require much more study before investigators, clinicians, and federal regulators can fully embrace them, said the panelists who discussed the results. Nevertheless, Biogen, which is codeveloping the antibody with partner Eisai, said in October it will put aducanumab forward to the Food and Drug Administration in a new drug application for the first-ever AD disease-modifying agent. FDA regulators have said they will review the data.
The new subanalysis comprised 570 of 3,285 patients in two identical studies with negative primary endpoint results. One, ENGAGE, failed to reach both its primary and secondary endpoints; the other, EMERGE, was halted last spring after a futility analysis determined that aducanumab was unlikely to confer significant benefit. The post hoc subanalysis looked at a combined subset of those who received the highest 10-mg/kg dose for the full 78 weeks of each trial. The statistically significant functional endpoints occurred in this group, comprised largely of apolipoprotein E epsilon-4 (APOE4) allele carriers.
“The futility analysis of EMERGE was highly unfortunate,” said panelist Paul Aisen, MD, founding director of the Alzheimer’s Therapeutic Research Institute at the University of Southern California, Los Angeles. “Clearly in the final analysis, EMERGE was positive in the primary endpoints, and now the secondary analysis of both studies is positive and consistent.” The diverging trajectory of placebo and treatment groups continued to the end of follow-up in both studies, a finding that at least suggests continuing improvement, he added.
Biogen undertook the pooled analysis after ENGAGE’s futility analysis. Early in the development program, concern about amyloid-related imaging abnormalities (ARIA) in APOE4 carriers led Biogen to stratify doses in that group.
“When we started [creating aducanumab trials], we stratified the dose so that e4 carriers had the lowest dose, but in PRIME [the phase 1b study], we saw the best result from the 10-mg/kg dose, so we believed that was important for efficacy. However, we didn’t have sufficient evidence to believe that it was safe to put carriers on that dose. In EMERGE, we saw that carriers could safely take it until the end of the study.”
Since the trials were running almost synchronously, a new version of randomization ensued. This allowed more e4 carriers to go forward on the 10-mg/kg dose.
“I would not normally recommend changing dose in the middle of a phase 3 trial, but it did have a real impact in the high-dose group,” Dr. Haeberlein said. Additionally, by the time of data lock after the futility analysis, more patients had completed the entire 78 weeks at the 10-mg/kg dose. Cumulative dosing ended up being quite different in the APOE4 carriers after this new version ensued. Before, the median cumulative dose for both carriers and noncarriers was 116 mg/kg. After the change, the median cumulative dose was 153 mg/kg. And before the alteration, 21% in EMERGE and 15% in ENGAGE received the full 14 possible 10-mg/kg doses. After the change, 51% in EMERGE and 47% in ENGAGE received the full 14 doses of 10 mg/kg.
The pooled analysis comprised this combined group, which was then largely composed of APOE4 carriers.
Imaging confirmed such dose-driven reductions in both brain amyloid plaques and phosphorylated tau. Although amyloid reduction has never been tied to cognitive or functional benefits, tau reduction has been associated with nonsignificant cognitive benefits in prior studies.
In the primary analysis of ENGAGE, aducanumab conferred no cognitive or functional benefit. In EMERGE, there were significant cognitive improvements on both the Mini Mental State Exam score (an 18% slowing of decline relative to placebo) and the Alzheimer’s Disease Assessment Scale cognitive portion (a 27% slowing).
However, the functional improvements seen in the pooled post hoc data “are a big deal,” and probably more meaningful to patients and families than the memory improvements, Dr. Cohen said.
“Those of us who know this disease well know what it means to lose yourself slice by slice, and anything you can hang onto is a triumph,” said Dr. Cohen, medical director and principal investigator of the Toronto Memory Program, an independent medical facility for dementia care and research. “I am pleased with a 27% slowing of cognitive decline, but a 40% slowing of functional decline is what will be really meaningful to patients. This is a long, slow disease, and if we can slow it at all, we’re winning out.”
Safety endpoints, especially ARIA, were not unexpected considering past studies. ARIA occurred in 41% of patients treated with the high aducanumab dose in EMERGE and in 40% in ENGAGE. It was largely asymptomatic (80% in EMERGE and 71% in ENGAGE). Headache was the next most common adverse event, followed by dizziness, visual disturbance, and nausea and vomiting. ARIA generally resolved within 4-6 weeks, and most patients continued their 10-mg/kg dose.
Dr. Laurie Ryan
Biogen intends to begin a new study, an open-label nonrandomized trial that will offer the 10-mg/kg dose to all patients in both trials, including those who took placebo. This may provide interesting data regarding redosing patients who were off their successful 10-mg/kg dose for an extended period of time, said Laurie Ryan, PhD, chief of the Dementias of Aging Branch in the Division of Neuroscience at the National Institute on Aging.
“If those in the high-dose group had a regression of their improvements and then improved again when restarted, that would certainly tell us something,” she said in an interview. Likewise, researchers will be carefully looking at any placebo group response. “But we have to remember that this will not be a randomized study,” and will bring with it all the issues that such a study typically carries.
“I agree it’s unfortunate that they had to stop the EMERGE trial,” she said. “It really did complicate the results, even though they are certainly trending in the right way. But we have had a number of post hoc analyses that show APOE4-positive benefiting, or e4-negative benefiting, and these haven’t panned out.”