Abstract
Apathy is a common neuropsychiatric symptom (NPS) in Alzheimer's disease (AD) but can emerge earlier in prodromal and even preclinical stages as part of mild behavioural impairment (MBI-apathy), a syndrome defined by emergent and persistent NPS. In dementia, apathy is associated with higher morbidity, mortality and caregiver distress. However, the significance of MBI-apathy in dementia-free persons, including its associations with AD biomarkers, remains unclear. This study aimed to determine whether MBI-apathy is associated with biomarker evidence of amyloid-beta (Aβ) and tau [phosphorylated (p-tau) and total (t-tau)] in CSF. Because MBI predicts incident dementia better than NPS without MBI, we aimed to determine the association between apathy and AD biomarkers when it occurred as part of the MBI syndrome and when it did not. Dementia-free participants with mild cognitive impairment or normal cognition from the Alzheimer's Disease Neuroimaging Initiative were stratified by NPS status (MBI-apathy, non-apathy MBI, non-MBI NPS and no-NPS) based on the Neuropsychiatric Inventory (NPI) or NPI-Questionnaire (NPI-Q). Linear regressions modelled cross-sectional associations between NPS status (predictor) and CSF biomarker ratios (Aβ42/Aβ40, p-tau181/Aβ42 and t-tau/Aβ42; primary outcomes) and levels (Aβ40, Aβ42, p-tau181 and t-tau; exploratory outcomes), adjusting for age, sex, apolipoprotein E4, education, Mini-Mental State Examination and NPI version. Hierarchical linear mixed-effects (LME) models assessed longitudinal associations over 2 years, incorporating random intercepts and slopes to account for repeated measures. Fixed effects included NPS status, all covariates from the linear regression model, and an interaction term between NPS status and time. Among 477 participants (176 cognitively normal), 52 had MBI-apathy. Primary cross-sectional analyses showed that, compared with the no-NPS group, MBI-apathy was associated with higher CSF p-tau181/Aβ42 [11.25% (2.56%-20.68%); adjusted P = 0.018] and t-tau181/Aβ42 [10.26% (2.42%-18.70%); adjusted P = 0.018]. Exploratory analyses revealed that MBI-apathy was associated with higher CSF p-tau181 [5.98% (0.50%-11.77%); P = 0.032]. Primary LME models showed that MBI-apathy was associated with higher CSF p-tau181/Aβ42 [11.34% (2.55%-20.88%); adjusted P = 0.022] and t-tau181/Aβ42 [10.34% (2.41-18.88%); adjusted P = 0.022] over 2 years. Exploratory LME models revealed that MBI-apathy was associated with higher CSF p-tau181 [6.03% (0.56%-11.81%); P = 0.032] and t-tau [4.96% (0.07%-10.09%); P = 0.049] over 2 years. MBI-apathy was significantly associated with core AD biomarkers cross-sectionally and longitudinally, over 2 years, underscoring its relevance as a marker of AD pathological burden. An overall MBI composite score might reflect a broader spectrum of pathology and warrants further investigation.