RESUMO
BACKGROUND: Phosphorylated tau (p-tau) is a specific blood biomarker for Alzheimer's disease (AD) pathology. Multiple p-tau biomarkers on several analytical platforms are poised for clinical use. The Alzheimer's Association Global Biomarker Standardisation Consortium plasma phospho-tau Round Robin study engaged assay developers in a blinded case-control study on plasma p-tau, aiming to learn which assays provide the largest fold-changes in AD compared to non-AD, have the strongest relationship between plasma and cerebrospinal fluid (CSF), and show the most consistent relationships between methods (commutability) in measuring both patient samples and candidate reference materials (CRM). METHODS: Thirty-three different p-tau biomarker assays, built on eight different analytical platforms, were used to quantify paired plasma and CSF samples from 40 participants. AD biomarker status was categorised as "AD pathology" (n=25) and "non-AD pathology" (n=15) by CSF Aß42/Aß40 (US-FDA; CE-IVDR) and p-tau181 (CE-IVDR) methods. The commutability of four CRM, at three concentrations, was assessed across assays. FINDINGS: Plasma p-tau217 consistently demonstrated higher fold-changes between AD and non-AD pathology groups, compared to other p-tau epitopes. Fujirebio LUMIPULSE G, UGOT IPMS, and Lilly MSD p-tau217 assays provided the highest median fold-changes. In CSF, p-tau217 assays also performed best, and exhibited substantially larger fold-changes than their plasma counterparts, despite similar diagnostic performance. P-tau217 showed the strongest correlations between plasma assays (rho=0.81 to 0.97). Plasma p-tau levels were weakly-to-moderately correlated with CSF p-tau, and correlations were non-significant within the AD group alone. The evaluated CRM were not commutable across assays. INTERPRETATION: Plasma p-tau217 measures had larger fold-changes and discriminative accuracies for detecting AD pathology, and better agreement across platforms than other plasma p-tau variants. Plasma and CSF markers of p-tau, measured by immunoassays, are not substantially correlated, questioning the interchangeability of their continuous relationship. Further work is warranted to understand the pathophysiology underlying this dissociation, and to develop suitable reference materials facilitating cross-assay standardisation. FUNDING: Alzheimer's Association (#ADSF-24-1284328-C).