Adverse Events as a Cause of Unblinding of Allocated Arms in Anti-Amyloid Therapy Trials: A Meta-Analysis of the Predictive Value.

Abstract

Anti-amyloid drugs for early Alzheimer's disease, including lecanemab, are associated with adverse events (AEs), such as amyloid-related imaging abnormalities (ARIA)-edema/effusion (E), ARIA-hemorrhage, and infusion-related reactions, which can indicate allocated arms in clinical trials. Herein, we evaluated the predictive value of AEs using a meta-analysis to estimate their incidence and simulated positive predictive value (PPV). The PPV for ARIA-E was high (0.915), but that for ARIA hemorrhage was low (0.630). Infusion-related reactions had a high PPV of 0.910, but with a wide confidence interval. Our results suggest the need to ameliorate the unblinding effects of AEs, particularly ARIA-E in trials.