Plasma Aβ42/40 and PET amyloid associations among late‐middle‐aged African Americans: Preliminary results from the AA‐FAIM study

Abstract Background Alzheimer’s disease (AD) models of disease progression have been built largely on non‐Hispanic White samples. Availability of AD blood‐based biomarkers may improve screening for AD risk and facilitate greater diversity in research studies; little is known about these biomarkers within African Americans (AA). Here, we characterized associations between plasma Ab 42/40 and Positron Emission Tomography (PET) amyloid measures in an AA sample. Method Participants in the African Americans Fighting Alzheimer’s in Mid‐Life (AA‐FAIM) with >=1 PET scan and plasma Ab 42/40 assay were included (n=34 participants; n=117 EDTA plasma samples). The ratio Ab 42/40 was calculated from plasma Ab 40 and Ab 42 quantified by C2N Diagnostics (PrecivityAD™). PET 11 C‐Pittsburgh compound B (PIB)‐derived amyloid measures included a mean cortical DVR (global PiB DVR), PET_A+ (DVR>1.19) and estimated PET_A+ duration at each plasma sample. In separate models where “time” was age or PET_A+ duration at each plasma sample (outcome=plasma Ab 42/40 ), we used mixed effects models to estimate simple slopes from time * PET_A+ status interactions. Sensitivity, specificity and related statistics were calculated for two plasma_A+ cutoffs (0.0975, West et al., 2021; 0.089, Hu et al., 2022). Result Baseline plasma sample and baseline PET scan average(sd) ages were 61(8.6) and 65(9), respectively. Closest plasma Ab 42/40 explained 47% of variability in last global PiB DVR (mean(sd)=‐0.85(0.91) years plasma‐to‐PET). Seven(20.6%) participants were PET_A+. In longitudinal analyses, simple age‐slopes for plasma Ab 42/40 change were ‐0.00061(0.00024) for PET_A+ and ‐0.00057(0.00018) for PET_A‐ (age * PET_A status interaction NS; model AICc=‐792.6). In a parallel mixed effect model, simple slopes were ‐0.00093(0.00026) for PET_A+ change/year and ‐0.00039(0.00012) for PET_A‐ change/year (interaction p∼0.06; AICc=‐800.6). Figures 1 and 2 depict relationships among plasma and PET amyloid measures. Positive and negative predictive values were 50.0% and 95.5% (0.0975 cut‐off) and 68% and 83% (0.089 cut‐off; see Figure 1 for additional statistics). Conclusion Preliminary results from this small African American sample provide encouraging data regarding plasma Ab 42/40 ’s potential to screen for PET A+ in AAs. PET A+ duration explained more variability than age in plasma Ab 42/40 trajectories. The on‐going AA‐FAIM study continues the critical effort to increase representation of AA in AD research, ensuring findings are broadly applicable.