Eli Lilly has shared full data from a phase 2 clinical trial it hailed as showing a “significant slowing of decline” in Alzheimer’s disease patients. The donanemab data are a mixed bag, with a slight win on one disease scale undermined by a failure on a more widely used measure of Alzheimer’s.
In January, Lilly revealed donanemab, formerly known as LY3002813, improved a composite measure of cognition and daily function in patients with early symptomatic Alzheimer’s. The finding, in a field defined by repeated failures, was a rare glimmer of hope and offered encouragement for people who take a positive view of Biogen’s similar drug candidate aducanumab.
Full data from the 257-subject clinical trial, published in The New England Journal of Medicine, show patients on donanemab experienced a 6.86 decline in scores on the Integrated Alzheimer’s Disease Rating Scale (iADRS). Scores in the placebo group fell 10.06. Lower iADRS scores indicate greater cognitive and functional impairment.
The difference between the iADRS scores was enough for the trial to hit its primary endpoint with a p-value of 0.04. Analysts at Jefferies hailed the result as a positive for Lilly and the broader effort to treat Alzheimer’s by targeting amyloid.
“LLY's positive Phase II donanemab study adds another (+) datapoint for Abeta antibodies (and BIIB) that directly target and remove plaque and wisely enrich for milder pts,” the analysts wrote in a note to investors.
However, more skeptical readings of the data are possible, in part because it is unclear exactly what the narrow statistical improvement in iADRS scores means in terms of clinical outcomes.
Lilly has used iADRS, which combines scores from the better established ADAS-Cog and ADCS-iADL, across multiple trials. In the solanezumab phase 3 clinical trials, iADRS was the only measure that consistently found the drug performed better than placebo. Two of those trials failed the Clinical Dementia Rating Scale–Sum of Boxes (CDR-SB) endpoint but hit on iADRS.
The donanemab data reveal a similar divergence between iADRS and CDR-SB results. In the phase 2, Lilly found donanemab had no statistically significant effect on CDR-SB over placebo. The success on iADRS, which is little used outside of Lilly, and failure on the well-understood CDR-SB complicates efforts to understand what the apparent effects of donanemab mean in practical terms for patients. Lilly achieved the results in a carefully selected population, with 257 of the 1,955 patients assessed meeting the eligibility criteria and joining the study.
Lilly is running a second, 500-subject phase 2 clinical trial that uses change in CDR-SB as the primary goal, with iADRS serving as a secondary endpoint. The Jefferies analysts see potential for Lilly to frame the phase 2 program as a success on the strength of two hits against the iADRS endpoints.
“While FDA considers CDR-sb a regulatory endpoint, we don't get any sense investors would see a major issue if LLY had two positive studies on iADRS and presumably LLY remains in close dialog with FDA,” the analysts wrote.
Lilly's shares opened down 7% in early trading Monday morning.
Missing the CDR-SB endpoint prevented Lilly from forming definite conclusions about its other secondary objectives. Patients on donanemab performed better numerically on multiple measures, but it is unclear whether the results reflect benefits delivered by the drug.
Lilly more conclusively showed that donanemab, which zeroes in on amyloid plaques, reduces levels of the aggregates of misfolded proteins in the brain. On the safety and tolerability front, incidence of ARIA-E—amyloid-related imaging abnormalities—affected 26.7% of patients on donanemab and caused symptoms in 6.1% of people on the drug. The rate of ARIA-E in the placebo arm was 0.8%.
Patients on aducanumab have also developed ARIA-E. That was one of many data points analysts compared between the aducanumab and donanemab trials. The Jefferies analysts, who think aducanumab has a good shot at winning approval, presented a positive take on the Lilly data for Biogen.
“We could argue the data was promising and adds to the hypothesis for approving - but not too good such that it would obviously be better than Adu on efficacy and/or for potential competitor market share in the future,” the analysts wrote.