Abstract
INTRODUCTION: We aimed to determine the clinical efficacy of treating apolipoprotein E (ApoE) ε4 homozygotes with recently approved anti-amyloid antibodies. METHODS: Data were derived from supplementary analyses in the regulatory studies Clarity (lecanemab) and TRAILBLAZER-ALZ2 (donanemab). We used Bayesian reanalysis with an independent t-statistic to determine evidence for or against an effect of antibody treatment on Clinical Dementia Rating scale Sum of Boxes (CDR-SB) in ApoE ε4 homozygotes, and a Bayesian random-effect meta-analysis to determine the effect size. RESULTS: The Bayesian reanalysis showed moderate evidence of no effect for both antibodies. For donanemab and lecanemab, the odds of no difference in treatment effect were nearly three times greater than the odds of a difference. The meta-analysis revealed a small effect of -0.06 CDR-SB points in favor of treatment with a moderate heterogeneity estimate. The Bayes factor was 0.26, indicating that the absence of an effect was almost four times more likely than the presence of an effect. DISCUSSION: The most likely explanation for our results is the lack of a treatment effect for lecanemab and donanemab in ApoE ε4 homozygotes. This could reflect inadequate exposure to the antibody due to more severe side effects, subsequent treatment interruptions, and lower dosing or a biologically driven lack of efficacy in a genetically determined disease. Our results support the view that excluding these cases from treatment is justifiable because of the higher risk of side effects and the lack of clinical efficacy. HIGHLIGHTS: Lecanemab and donanemab were clinically ineffective in ApoE ε4 homozygotes.ApoE ε4 homozygotes should not receive these treatments due to inefficacy.Future trials should adopt Bayesian analysis strategies.Bayesian analysis provides evidence for or against treatment effects.Bayesian inference provides clinically interpretable results.