Timing of changes in Alzheimer's disease plasma biomarkers as assessed by amyloid and tau PET clocks.

Plasma biomarkers for Alzheimer's disease (AD) are increasingly being used to assist in making an etiological diagnosis for cognitively impaired (CI) individuals or to identify cognitively unimpaired (CU) individuals with AD pathology who may be eligible for prevention trials. However, a better understanding of the timing of plasma biomarker changes is needed to optimize their use in clinical and research settings. The aim of this study was to evaluate the timing of change of key AD plasma biomarkers (Aβ42/Aβ40, p-tau217, p-tau181, GFAP and NfL) from six different companies, along with established AD biomarkers, using AD progression timelines based on amyloid and tau PET. We used data from the Alzheimer's Disease Neuroimaging Initiative (ADNI), including 784 individuals with longitudinal F-florbetapir amyloid PET and 359 individuals with longitudinal F-flortaucipir tau PET, to estimate age at amyloid and tau positivity, defined as the age at the first positive PET scan. Of these, longitudinal plasma biomarker measures were available from 190 individuals with an estimated age at amyloid positivity and 70 individuals with an estimated age at tau positivity. Age at tau positivity was a stronger predictor of symptom onset than age at amyloid positivity in 17 individuals who progressed from CU to CI during their participation in the ADNI study (Adj R = 0.86 Adj R = 0.38), and therefore was used to estimate symptom onset age for all individuals with an estimated age at tau positivity. Generalized additive mixed models (GAMMs) were used to model biomarker trajectories across years since amyloid positivity, tau positivity, and symptom onset, and to identify the earliest timepoint of biomarker abnormality when compared to a reference group of amyloid- and tau-negative CU individuals, as well as time periods of significant change in biomarkers. All plasma biomarkers except NfL became abnormal prior to amyloid and tau positivity. Plasma Aβ42/Aβ40 was the first biomarker to reach abnormality consistently across timelines and plasma GFAP became abnormal early in the tau timeline. Plasma Aβ42/Aβ40 levels reached a plateau, while plasma p-tau217, p-tau181, GFAP and NfL increased throughout disease progression. Some differences in the timing of change were observed across biomarker assays. The primary utility of plasma Aβ42/Aβ40 may lie in early identification of individuals at high risk of AD. In contrast, p-tau217, p-tau181, GFAP and NfL increase throughout the estimated timelines, supporting their potential as biomarkers for staging and monitoring disease progression.