BackgroundAdvances in analytical methods have recently paved the way to Alzheimer’s disease (AD) biomarkers testing in blood along with the more established CSF testing. To ensure a forthcoming application of this low-invasive diagnostic that might allow to recognize early onset of dementia, appropriate pathological cut-points need to be defined. MethodsIn this cross-sectional study we measured blood and CSF neurofilament light chain (NFL), phosphorylated tau (pTau 181), Amyloid-β1-42 (AB 1–42) and Amyloid-β1-40 (AB 1–40) on a fully automated chemiluminescent platform (Lumipulse, Fujirebio) in 80 cognitively impaired patients and 55 cognitively unimpaired subjects. Clinical cut points were calculated with receiver-operator characteristic (ROC) curve analysis and a head-to-head comparison of blood and CSF testing was performed. ResultsBlood NFL best discriminant thresholds to distinguish neurodegenerative diseases from controls varied age-dependently, being 19 and 33 pg/mL in subjects 50–65 years and > 65 years respectively. AD was best framed by AB 1–42/1–40 ratio < 0.079 and ptau181 > 1 pg/mL. Though a strong correlation for all biomarkers, only blood AB ratio was equal to CSF testing for AD diagnosis. ConclusionsThe specific context of use might be considered to define the cut-offs of blood biomarkers of neurodegenerative diseases. Future efforts towards reference materials for each AD blood biomarker will improve clinical cut-offs.