Automatic assessment of classical Alzheimer’s disease biomarkers in paired cerebrospinal fluid and plasma samples
Giovanni Bellomo, Alfredo Megaro, Andrea Toja, Giovanna Nardi, Anna Wojdala, Marta Filidei, Federico Paolini Paoletti, Lorenzo Gaetani, Lucilla Parnetti- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
A/T/(N) cerebrospinal fluid (CSF) biomarkers, i.e., amyloid‐β 1‐42 and 1‐40 ratio (Aβ42/Aβ40, A), tau phosphorylated at threonine 181 (p‐T181‐tau, T), and total tau (N), define molecular diagnosis of Alzheimer’s disease (AD). Automated platforms have brought a significant achievement in reducing the pre‐analytical variability. Measuring such biomarkers in blood might allow longitudinal sampling for disease monitoring, in view of disease‐modifying therapies. Along with this premise, to test the agreement between CSF and plasma measurements, we assessed Aβ42, Aβ40 and p‐T181‐tau with Lumipulse® G600II automated platform in paired CSF and plasma samples from 224 patients.
Method
We selected 224 plasma samples collected from subjects with different clinical diagnosis (see Table 1), for whom CSF measurements were performed between 2018 and 2021. Both plasma and CSF were collected at the time of lumbar puncture following international guidelines. ATN CSF profiles have been defined according to cutoff values measured in our lab. One‐way analysis of variance with post‐hoc Tukey test was applied to determine statistically significant differences among groups, linear regression and correlation analysis were applied to determine degree of agreement between CSF and plasma biomarkers, and ROC analysis was applied to test the ability to differentiate AD‐like CSF profiles (A+/T+) from normal CSF profiles (A‐/T‐).
Result
Distributions of Aβ42/Aβ40 ratio and p‐T181‐tau for the different A/T groups are shown in Figure 1. Plasma Aβ42/Aβ40 ratio was found to be significantly decreased in A+/T‐ and A+/T+ subjects as compared with A‐/T‐ and A‐/T+; p‐T181‐tau was found to be significantly increased in A+/T+ subjects as compared with all the other A/T profiles. Plasma Aβ42/Aβ40 ratio and p‐T181‐tau showed moderate correlation with their CSF equivalents (r = 0.62 and 0.60, p << 0.001, respectively). Both biomarkers could detect AD‐like CSF profiles (A+/T+) with an area under the ROC curve > 0.8 (see Figure 2).
Conclusion
The automated assessment of Aβ42/Aβ40 and p‐T181‐tau in plasma can satisfactorily reflect the CSF A+T+ status. Larger cohorts are needed to verify the concordance of plasma with CSF biomarkers