Automated high-throughput quantification of plasma p-tau217 and APOE-ε4 for Alzheimer’s disease diagnosis and cognitive decline in a memory cohort

Abstract Background Blood-based biomarkers could improve the precision of Alzheimer’s disease (AD) clinical diagnosis and expand access to targeted treatments. Therefore, we evaluated the diagnostic accuracy of plasma Elecsys p-tau217 (Roche) and compared it with Elecsys p-tau181 (Roche) and Lumipulse p-tau217 (Fujirebio). We also assessed the added value of APOE-ε4 carrier status, plasma Aβ42 and Aβ42/40 in a memory-clinic cohort and evaluated associations with longitudinal cognition. Materials and methods A total of 187 patients from the Cognitive Centre Ghent University (CCUG) biobank, classified as AD (n = 103) or non-AD cognitive disorders (n = 84) based on CSF biomarkers (CSF Aβ42/40 ratio, total tau and p-tau181), were included. Plasma Elecsys p-tau181, p-tau217, and APOE-ε4 were measured on the Roche cobas® pro e801, and p-tau217, Aβ42, and Aβ40 on the Fujirebio LUMIPULSE G1200. ROC analyses and single- and two-cut-off strategies (95% sensitivity/specificity) were applied. Subgroup analyses examined APOE-ε4 status, renal function, cerebral amyloid angiopathy (CAA) features, Fazekas score, and longitudinal cognition. Results Elecsys plasma p-tau217 showed high discriminative performance for AD versus non-AD (AUC 0.939), comparable to Lumipulse p-tau217 (AUC 0.950; p = 0.485). Elecsys p-tau181 performed lower than Elecsys p-tau217 (AUC 0.903; p = 0.043). Using a two-cut-off strategy, the intermediate proportion was 19.9% for Elecsys p-tau217, 11.9% for Lumipulse p-tau217, and 33.2% for Elecsys p-tau181. Adding APOE-ε4 to Elecsys p-tau217 improved discriminative performance (AUC 0.970, p = 0.02) and reduced intermediates to 11.0%. Adjustment for Aβ42 on the Fujirebio platform did not significantly increase the AUC (0.950 vs. 0.957; p = 0.322) and modestly reduced intermediate classifications (11.9% to 10.0%). Higher baseline Elecsys p-tau217 was associated with lower baseline MoCA and a trend towards faster MoCA decline (p = 0.07). Age, sex, renal function, Fazekas score, and CAA were not significantly associated with Elecsys p-tau217 concentrations. Conclusion Plasma Elecsys p-tau217 measured on an automated high-throughput platform shows excellent diagnostic accuracy for AD. Incorporating APOE-ε4 further improves classification, while Aβ42 adjustment had only limited additional impact. Baseline p-tau217 also reflects cognitive severity and may relate to subsequent cognitive decline in the memory-clinic setting.