Core blood biomarkers of Alzheimer's disease: A single-center real-world performance study

Abstract

Background: The new criteria for Alzheimer's disease pave the way for the introduction of core blood biomarkers of Alzheimer's disease (BBAD) into clinical practice. However, this depends on the demonstration of sufficient accuracy and robustness of BBADs in the intended population.

Objectives: To assess the diagnostic performance of core BBADs in our memory clinic, comparing them with cerebrospinal fluid (CSF) analysis.

Design: Real-world cross-sectional observational study.

Setting: Memory Clinic of Fondazione IRCCS "San Gerardo dei Tintori," Monza, Italy.

Participants: n = 102 consecutive outpatients (mean age: 71.0 ± 7.6 years) with cognitive impairment undergoing routine lumbar puncture.

Measurements: CSF Aβ40, Aβ42, tTau, and pTau181 levels were measured. Plasma biomarkers were evaluated using Lumipulse® G600II. Logistic regression and Receiver Operating Characteristic (ROC) analysis were used to assess biomarker performance. The diagnosis of Alzheimer's disease was based on CSF Aβ42/40 ratio.

Results: Plasma pTau217 demonstrated the highest diagnostic accuracy (AUC=0.91), followed by pTau181 (AUC=0.88) and Aβ42/40 (AUC=0.83). In robustness analyses, only pTau217 and pTau181 performance remained consistent, while that of Aβ42/40 ratio declined with added random variability. pTau217 significantly outperformed other BBAD, with the exception of pTau181. pTau BBAD were significant predictors of baseline Mini-Mental State Examination scores.

Conclusions: Plasma pTau217, measured with Lumipulse®, is a robust and reliable BBAD for detecting amyloid pathology in a memory clinic setting, offering a practical and less invasive alternative to traditional CSF testing.

Keywords: Alzheimer's disease; Amyloid; BBAD; Dementia; Plasma biomarkers; Tau.

Fig. 1

Boxplots for plasma Aβ42, pTau181, pTau217 and Aβ42/40 ratio, according to CSF amyloid status. Outliers (above 95th and under 5th percentiles) have been excluded. The horizontal line indicates the optimal cut-off based on maximization of the Youden index.

Fig. 2

AUC with 95% confidence interval for the different plasma biomarkers in discriminating between amyloid positive and negative subjects.

Fig. 3

Spearman correlation coefficients between plasma biomarkers (rows) and all plasma and CSF biomarkers (columns). Non-significant correlations have been suppressed with an X.

Fig. 4

Results of a two cut-offs approach for plasma pTau181 and pTau217.