Short-term variability of Alzheimer's disease plasma biomarkers in a mixed memory clinic cohort

For clinical implementation of Alzheimer's disease (AD) blood-based biomarkers (BBMs), knowledge of short-term variability, is crucial to ensure safe and correct biomarker interpretation, i.e., to capture changes or treatment effects that lie beyond that of expected short-term variability and considered clinically relevant. In this study we investigated short-term intra- and inter-individual variability of AD biomarkers in the intended use population, memory clinic patients. In a consecutive sample of memory clinic patients (AD n = 27, non-AD n = 20), blood samples were collected on three separate days within a period of 36 days and analysed for plasma Aβ40, Aβ42, p-tau181, p-tau217, p-tau231, T-tau, neurofilament light (NfL), and glial fibrillary acidic protein (GFAP). We measured intra- and inter-individual variability and explored if the variability could be affected by confounding factors. Secondly, we established the minimum change required to detect a difference between two given blood samples that exceeds intra-individual variability and analytical variation (reference change value, RCV). Finally, we tested if classification accuracy varied across the three visits. Intra-individual variability ranged from ~ 3% (Aβ42/40) to ~ 12% (T-tau). Inter-individual variability ranged from ~ 7% (Aβ40) to ~ 39% (NfL). Adjusting the models for time, eGFR, Hba1c, and BMI did not affect the variation. RCV was lowest for Aβ42/Aβ40 (- ~ 15%/ + ~ 17%) and highest in p-tau181 (- ~ 30/ + ~ 42%). No variation in classification accuracies was found across visits. We found low intra-individual variability, robust to various factors, appropriate to capture individual changes in AD BBMs, while moderate inter-individual variability may give rise to caution in diagnostic contexts. High RCVs may pose challenges for AD BBMs with low fold changes and consequently, short-term variability is important to take into consideration when, e.g., estimating intervention effect and longitudinal changes of AD BBM levels. Clinicaltrials.gov (NCT05175664), date of registration 2021-12-01.