Longitudinal change in plasma biomarkers by amyloid PET status

Abstract

AbstractBackgroundIt is important to better understand the longitudinal trajectories of plasma measures of proteins implicated in Alzheimer’s disease among cognitively normal older adults.MethodWe investigated whether baseline levels and longitudinal rates of change of three Quanterix SIMOA plasma measures (amyloid‐beta 42 to 40 ratio [Aβ42/Aβ40], glial fibrillary acidic protein [GFAP], and neurofilament light chain [NfL]) differ by amyloid group as determined using Pittsburgh compound B (PiB) PET in the Baltimore Longitudinal Study of Aging (N=147 PiB– and 57 PiB+; 640 longitudinal measures). All participants were cognitively normal at baseline. We used linear mixed effects models with longitudinal plasma measures as the outcome. We conducted an unadjusted analysis that included binary PiB group, time from baseline, and their interaction. We also conducted an adjusted analysis that additionally included the following explanatory variables: baseline age and its interaction with time, sex, APOE ε4 carrier vs noncarrier, race, and estimated glomerular filtration rate (eGFR) from serum creatinine concurrent with plasma measurements. All variables except time were mean‐centered.ResultParticipant characteristics are presented in Table 1. Figures 1 and 2 illustrate plasma biomarkers versus age and longitudinal visit schedule per participant, respectively. PiB+ individuals had lower Aβ42/Aβ40, higher GFAP, and higher NfL at baseline (Table 2). Both PiB groups exhibited longitudinal increases in GFAP and NfL (Table 3). Aβ42/Aβ40 declined longitudinally in the PiB– group, but annual change in the PiB+ group was not statistically significant. PiB group difference in rates of change in plasma measures was statistically significant for Aβ42/Aβ40 but not for NfL. Difference in rate of change in GFAP was statistically significant in the unadjusted model only. Adjusted model fixed effect estimates are depicted in Figure 3.ConclusionSteeper decline in plasma Aβ42/Aβ40 in the PiB– group suggests that individuals may exhibit decline in this measure prior to the onset of amyloid PET positivity. Despite exhibiting baseline PiB group differences, plasma GFAP and NfL have similar rates of increase in both PiB groups. Reliable detection of such changes at the individual level, prior to the onset of cognitive impairment, may enable better personalized prediction of future outcomes.