Plasma Alzheimer″s assays agree on level but much less on individual rates of change

INTRODUCTION: Serial plasma testing is proposed for monitoring Alzheimer′s disease and for trials, but its reliability for a person′s rate of change is unknown. METHODS: In 393 ADNI participants with three plasma samples over a median 5.0 years, 15 measures of five biomarkers were assayed on aliquots of the same samples, separating shared change from assay-specific variation at each sampling occasion. Results were tested in 1,093 A4/LEARN participants. RESULTS: Assays agreed well on cross-sectional levels (intraclass correlation 0.75—0.95 for pTau217, pTau181, GFAP, NfL) but less on individual rates of change (0.44, 95% CI 0.37—0.50, for pTau217). Assay-specific slope variance was approximately zero, so disagreement was predominantly random. Three-sample single-assay slope reliability was at most 0.40 (0.34—0.46) for pTau217, 0.42 (0.33—0.51) for pTau181, 0.47 (0.37—0.60) for GFAP, 0.52 (0.40—0.64) for NfL, and 0.10 (0.03—0.20) for Aβ42/40. Two pTau217 measurements had to differ by 65%—131%, depending on the assay, to exceed measurement noise. Detecting 25% slowing of pTau217 change required 924 (598—1,468) to 4,030 (2,091—9,256) participants per arm, or 532 (357—778) to 965 (655—1,431) with averaged measures. In A4/LEARN, single-assay pTau217 reliability was 0.54 (0.46—0.61), and a 7.3- rather than 4.8-year span raised it by 0.10 (0.00-0.22); a fourth sample within it added little. DISCUSSION: With three single-assay samples, much of an individual′s measured change is noise. Longer spans, endpoint replication, and assay choice, rather than additional intermediate samples, are needed for reliable serial testing and trials.

Authors

Institutions

Publication Details

Journal
medRxiv
Published
2026-11-18
DOI
https://doi.org/10.64898/2026.11.16.26368193
Primary Topic
Dementia and Cognitive Impairment Research
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Plasma Alzheimer″s assays agree on level but much less on individual rates of change

Ahmet Begde, Ivan G. Koychev, James Benedict Rowe, Vanessa Raymont et al.
8 links
medRxiv
Dementia and Cognitive Impairment Research
preprint

Plasma Alzheimer″s assays agree on level but much less on individual rates of change

Ahmet Begde, Ivan G. Koychev, James Benedict Rowe, Vanessa Raymont, Kellen K. Petersen, Alzheimer’s Disease Neuroimaging Initiative, Rachel Brohier, Rosanna McEwen-Smith, Yi Yang
preprint en

Abstract

INTRODUCTION: Serial plasma testing is proposed for monitoring Alzheimer′s disease and for trials, but its reliability for a person′s rate of change is unknown. METHODS: In 393 ADNI participants with three plasma samples over a median 5.0 years, 15 measures of five biomarkers were assayed on aliquots of the same samples, separating shared change from assay-specific variation at each sampling occasion. Results were tested in 1,093 A4/LEARN participants. RESULTS: Assays agreed well on cross-sectional levels (intraclass correlation 0.75—0.95 for pTau217, pTau181, GFAP, NfL) but less on individual rates of change (0.44, 95% CI 0.37—0.50, for pTau217). Assay-specific slope variance was approximately zero, so disagreement was predominantly random. Three-sample single-assay slope reliability was at most 0.40 (0.34—0.46) for pTau217, 0.42 (0.33—0.51) for pTau181, 0.47 (0.37—0.60) for GFAP, 0.52 (0.40—0.64) for NfL, and 0.10 (0.03—0.20) for Aβ42/40. Two pTau217 measurements had to differ by 65%—131%, depending on the assay, to exceed measurement noise. Detecting 25% slowing of pTau217 change required 924 (598—1,468) to 4,030 (2,091—9,256) participants per arm, or 532 (357—778) to 965 (655—1,431) with averaged measures. In A4/LEARN, single-assay pTau217 reliability was 0.54 (0.46—0.61), and a 7.3- rather than 4.8-year span raised it by 0.10 (0.00-0.22); a fourth sample within it added little. DISCUSSION: With three single-assay samples, much of an individual′s measured change is noise. Longer spans, endpoint replication, and assay choice, rather than additional intermediate samples, are needed for reliable serial testing and trials.

medRxiv
Washington University in St. Louis (US), University of Cambridge (GB), Cambridge University Hospitals NHS Foundation Trust (GB), University of Oxford (GB), Imperial College London (GB)
Alzheimer's Society, National Institute for Health and Care Research
Good health and well-being
Dementia and Cognitive Impairment Research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon