Association of cerebrospinal fluid and plasma biomarkers with longitudinal tau accumulation

Abstract

AbstractBackgroundDue to its tight relation with neurodegeneration and cognitive decline, tau pathology is increasingly recognized as an attractive therapeutic target in Alzheimer’s disease. To be more efficient, clinical trials testing anti‐tau drugs need to target subjects at risk for longitudinal tau accumulation. So far, only elevated signal in an amyloid‐β (Aβ) PET scan has been consistently identified as predictor of future tau burden. Here, we studied whether more accessible cerebrospinal fluid (CSF) and plasma biomarkers were associated with PET‐measured longitudinal tau accumulation.MethodWe included two partially overlapping cohorts of participants from the Alzheimer’s Disease Neuroimaging Initiative that underwent baseline and follow‐up [18F]flortaucipir (FTP) scans, as well as baseline Aβ PET scans. In cohort 1 (93 cognitively unimpaired (CU) and 70 impaired (CI)), all participants had available CSF biomarkers (Aβ1‐42, p‐tau181, and p‐tau181/Aβ1‐42 ratio) at baseline; in cohort 2 (37 CU and 37 CI), participants had a plasma p‐tau181 measurement within 2 years prior to the baseline FTP scan. Annual standardized uptake value ratio (SUVR) change on FTP scans was estimated at the voxel level and in a previously defined meta‐region of interest (ROI) using linear mixed effects models. In cohort 1, correlations of CSF markers with FTP SUVR annual change were determined using Pearson correlation coefficients, separately for CU and CI. In cohort 2, we instead used clinical diagnosis‐adjusted partial correlations in the combined cohort (CU+CI) due to lower sample size.ResultAmong CU individuals in cohort 1, all CSF biomarkers showed moderately strong correlations with FTP SUVR change; CSF p‐tau181/Aβ1‐42 ratio showed the strongest association with FTP SUVR change, slightly outperforming Aβ PET (Fig. 1). Similarly, CSF p‐tau181/Aβ1‐42 ratio was the strongest correlate of FTP SUVR change in CI subjects (Fig. 2). In cohort 2, plasma p‐tau181 was moderately correlated with FTP SUVR change, though not stronger than Aβ PET (Fig. 3).ConclusionCSF biomarkers might be as effective as Aβ PET for the prediction of longitudinal tau accumulation. Plasma p‐tau181, though not as strongly correlated as Aβ PET, might represent a cost‐effective solution for the prediction of tau accumulation. These findings may have implications for clinical trials with anti‐tau drugs.