Cerebrospinal fluid biomarkers distinguish postmortem-confirmed Alzheimer's disease from other dementias and healthy controls in the OPTIMA cohort.

Cerebrospinal fluid (CSF) amyloid-beta (A beta) and tau have been studied as markers of Alzheimer's disease (AD). Combined A beta 42 and t-tau distinguishes AD from healthy controls with a sensitivity and specificity (sens/spec) near 89% across studies. This study examined these markers in the homogeneous OPTIMA cohort, using extensive longitudinal follow up and postmortem evaluation to confirm clinicopathological status. Baseline CSF was analyzed from 227 participants with AD (97% autopsy-confirmed), mild cognitive impairment (MCI; 73% confirmed), other dementia syndrome (ODS; 100% confirmed), and controls (CTL; 27% confirmed, follow up approximately 9-13 years). Biomarker concentrations were analyzed using validated ELISAs. AD patients had lower CSF A beta 42 and higher t-tau, p-tau, t-tau/A beta 42, and t-tau/A beta 40 compared to CTLs, with MCI intermediate. CTL and MCI participants who progressed to AD demonstrated more AD-like profiles. A beta 40, sA beta PP beta, and sA beta PP beta were lower in AD compared to CTL. High-level discriminators of AD from CTL were t-tau/A beta 40 (AUROC 0.986, sens/spec of 92%/94%), p-tau/A beta 42 (AUROC 0.972, sens/spec of 94%/90%), and A beta 42 (AUROC 0.941,sens/spec of 88%). For discriminating AD from ODS, p-tau/A beta 42 demonstrated sens/spec of 88%/100% (95%/86% at the AD versus CTL cutoff) and A beta 42 demonstrated sens/spec of 84%/100%(88%/100% at the AD versus CTL cutoff). In a well-characterized, homogeneous population, a single cutoff for baseline CSF A beta and tau markers can distinguish AD with a high level of sens/spec compared to other studies. It may be important to characterize sources of demographic and biological variability to support the effective use of CSF diagnostic assays in the broader AD population.