Decline in proper name retrieval during story recall is associated with Alzheimer’s disease CSF biomarkers in cognitively unimpaired individuals

Abstract

AbstractBackgroundRemembering names of people and places is a common problem in aging that is exacerbated in Alzheimer’s disease (AD). Because retrieval of proper names (PNs) has been associated with brain regions affected by AD (anterior and medial temporal lobe), PN retrieval may be an especially sensitive measure of early AD pathology. Extending results of a previous study from WRAP (Mueller et al., 2020), this study examined associations between cerebrospinal fluid (CSF) AD biomarkers and longitudinal delayed story recall PN retrieval.MethodParticipants were from the BIOCARD study, a cohort of late‐middle aged adults enriched for parental history of AD. Participants were cognitively unimpaired at baseline with longitudinal item‐level data from the Logical Memory (LM; WMS‐R) Story A recall (N = 218; M(SD) baseline age = 58(9); M(SD) follow‐up = 4.1(2.9) years). CSF amyloid beta (Ab42/40) and phosphorylated tau181 (p‐tau181) were measured using automated electrochemiluminescence assays; AD biomarker abnormality (+) was determined using the bottom quartile for Ab42/40 and the top quartile for p‐tau181. We calculated total Story A PNs (range 0‐4) and total score (0‐25). We performed longitudinal linear mixed effects (LME) models examining the interaction between time (LM age, centered) and last available Ab42/40 and/or p‐tau181 status as predictors of PNs and total recall (subject‐specific random intercepts).ResultOf the 218 participants included, forty‐nine (22.4%) were Ab42/40+, 60 (27.5%) were p‐tau181+, and the remainder were A‐/T‐ (Table 1). In LME models, CSF‐Ab42/40+*age was not a significant predictor of either outcome. The p‐tau181+*age interaction showed significantly greater annual decline in PNs in the p‐tau181+ group than the p‐tau181‐ group (interaction beta = ‐.01, p = .02; simple slope estimate p‐tau181‐ = .002, p‐tau181+ = ‐.01) (Table 2, Fig.1). Interaction patterns were weaker yet similar for total recall. In follow‐up analyses, 15 participants who converted to MCI or dementia were removed and results remained unchanged. In an exploratory model combining biomarker statuses, the Ab42/40+/p‐tau181+*age interaction was non‐significant.ConclusionAs in the WRAP study, these results from BIOCARD indicate that PN recall is more sensitive to preclinical change than the traditional story recall total score metric; results also indicate that decline is greatest in those with elevated CSF p‐tau. Future directions should include replication in other longitudinal cohorts.