Abstract INTRODUCTION Carriers of the ε4 allele of the apolipoprotein E ( APOE ) gene have an increased risk for Alzheimer's disease (AD) and amyloid‐related imaging abnormalities (ARIAs) upon anti‐amyloid beta (Αβ) immunotherapy. Measuring apoE4 and pan‐apoE proteins in blood plasma for apoE4 proteotyping may offer an alternative to APOE genotyping. METHODS We assessed apoE4 proteotyping accuracy in 479 participants of the prospective FINd DElirium RIsk factors (FINDERI) study in patients undergoing cardiac surgery and compared results to quantitative polymerase chain reaction (qPCR) genotyping. RESULTS Proteotype–genotype discordance occurred in 8 of 479 participants (1.67%). Five of 17 proteotype homozygotes were genotypically heterozygous. Replacing manufacturer provided cut points with custom data‐driven thresholds substantially improved classification performance. DISCUSSION We confirm the reported overall high classification performance of apoE4 proteotyping but underscore the need to re‐evaluate the generalizability of the cut points provided with the assay kits. Misclassification of heterozygous APOE ε4 carriers as homozygous could erroneously exclude eligible patients from anti‐amyloid therapies.
Klafki et al. (2026) studied this question.