LAUSR

Objective To assess the ability of a combination of synaptic CSF biomarkers to separate Alzheimer disease (AD) and non-AD disorders and to help in the differential diagnosis between neurocognitive diseases. Methods This was a retrospective cross-sectional monocentric study. All participants explored with CSF assessments for neurocognitive decline were invited to participate. After complete clinical and imaging evaluations, 243 patients were included. CSF synaptic (GAP-43, neurogranin, SNAP-25 total, SNAP-25aa40, synaptotagmin-1) and AD biomarkers were blindly quantified with ELISA or mass spectrometry. Statistical analysis compared CSF levels between the various groups of AD dementias (n = 81), mild cognitive impairment (MCI)–AD (n = 30), other MCI (n = 49), other dementias (OD) (n = 49), and neurologic controls (n = 35) and their discriminatory powers. Results All synaptic biomarkers were significantly increased in patients with MCI-AD and AD-dementia compared to the other groups. All synaptic biomarkers could efficiently discriminate AD dementias from OD (AUC ≥0.80). All but synaptotagmin were also able to discriminate patients with MCI-AD from controls (area under the curve [AUC] ≥0.85) and those with AD dementias from controls (AUC ≥0.80). Overall, CSF SNAP-25aa40 had the highest discriminative power (AUC 0.93 between patients with AD dementias and controls or OD, AUC 0.90 between those with MCI-AD and controls). Higher levels were associated with 2 alleles of APOE ε4. Conclusion All synaptic biomarkers tested had a good discriminatory power to distinguish patients with AD abnormal CSF from those with non-AD disorders. SNAP25aa40 demonstrated the highest power to discriminate AD CSF-positive patients from patients without AD and neurologic controls in this cohort. Classification of evidence This retrospective study provides Class II evidence that CSF synaptic biomarkers discriminate patients with AD from those without AD.