CSF cut-offs for MCI due to AD depend on APOEε4 carrier status

Amyloid and tau pathological accumulation should be considered for AD definition and before subjects enrolment in disease-modifying trials. Although age, APOEe4 and sex influence CSF biomarker levels, none of these variables are considered by current normality/abnormality cut-offs. Using baseline CSF data from two independent cohorts (PharmaCOG/European ADNI and ADNI), we investigated the effect of age, APOEe4 status and sex on CSF Aβ42/P-tau distribution and cut-off extraction by applying mixture models with covariates. The Aβ42/P-tau distribution revealed the presence of 3 subgroups (AD-like, intermediate, control-like) and 2 cut-offs. The identification of the intermediate subgroup and of the higher cut-off were APOEe4-dependent in both cohorts. APOE-specific classification (higher cut-off for APOEe4+, lower cut-off for APOEe4-) showed higher diagnostic accuracy in identifying MCI due to AD compared to single Aβ42 and Aβ42/P-tau cut-offs. APOEe4 influences amyloid and tau CSF markers and AD progression in MCI patients supporting i) the use of APOE specific cut-offs to identify MCI due to AD and, ii) the utility of considering APOE genotype for early AD diagnosis.