The API Generation program: Umibecestat treatment and discontinuation effects on hippocampal and whole brain volumes in the overall population and amyloid‐negative APOE4 homozygotes: Alzheimer's Prevention Initiative (API) Generation program:...

Background: In addition to early mild non‐progressive cognitive worsening, BACE inhibitors have been associated with early non‐progressive reductions in hippocampal and whole brain volume in trial participants with elevated amyloid (A+, Egan et al, 2019). We have begun to evaluate the effects of umibecestat treatment and discontinuation on cognitively unimpaired amyloid‐positive (A+) and A‐ APOE4 homozygotes and A+ heterozygotes, testing the hypotheses that BACE inhibitor‐related hippocampal and whole brain shrinkage is apparent soon after treatment initiation, non‐progressive, limited to A+ participants, and reversible following discontinuation. Method: The API Generation Program had begun to evaluate umibecestat in 1,623 cognitively unimpaired 60‐75 year‐old participants including 921 A+ APOE4 heterozygotes and 702 homozygotes (A+ or A‐).Following randomization, volumetric brain MRIs were scheduled at month 6, month 12 and then annually. After discontinuation of study medication, participants continued to be assessed under double‐blind conditions, and MRIs were acquired within 3 months after the last dose. We investigated 6‐month hippocampal and whole brain volume changes and their relationship to RBANS immediate, delayed and total memory score changes; and we plan to investigate the persistence or reversibility of these changes within 3 months of discontinuation after study completion in Q2 2020. Result: Umibecestat was associated with significant 6‐month hippocampal and whole brain shrinkage in the aggregate APOE4 homozygote and A+ heterozygote group, but not in the A‐ homozygotes. While an early non‐progressive worsening in RBANS immediate, delayed, and total memory scores (Graf et al, AAIC 2020 submitted) was seen in the overall population as well as in A‐ HMs, we failed to demonstrate any significant association between brain shrinkage and cognitive worsening in the overall group (R2 coefficient <0.1) in the active treatment group for any of the visits (on‐ and off‐ treatment). Conclusion: BACE inhibitor‐related brain shrinkage appears to be early, non‐progressive, related to pre‐existing amyloid plaque burden, and unrelated to its cognitive worsening. Pending findings will help to clarify the extent to which it is also reversible. Together, our cognitive and MRI findings could help to assess potential benefits versus risks of BACE inhibitors in future prevention trials. [ABSTRACT FROM AUTHOR]