Abstract
The cerebrospinal fluid (CSF) biomarkers amyloid-β (Aβ)1-42, T-tau, and P-tau have good diagnostic accuracy for clinically diagnosed Alzheimer's disease (AD). However, in multi-center studies, the predictive values of the CSF biomarkers have been lower, possibly due to differences in procedures for lumbar puncture and CSF handling and storage, and to differences in patient populations, clinical evaluations, and diagnostic procedures. Here we investigate the diagnostic accuracy of CSF biomarkers in a well defined homogeneous mono-center population. We also evaluate an extended panel of amyloid related biomarkers. Sixty consecutive patients admitted for cognitive impairment to a memory clinic were recruited. The participants included patients with AD or mild cognitive impairment (MCI) diagnosed with AD upon follow-up (n = 32), patients with stable MCI (n = 13), patients with other dementias diagnosed at primary evaluation or upon follow-up (n = 15), and healthy controls (n = 20). CSF was analyzed for Aβ1-42, T-tau, P-tau, AβX-38, AβX-40, AβX-42, sAβPPα, and sAβPPβ. In multivariate analysis, the core biomarkers Aβ1-42, T-tau, and P-tau demonstrated a high ability to diagnose AD versus the combined groups of controls and stable MCI, with an area under the receiver operating characteristic curve (AUROC) of 0.97 (95% CI 0.93–1.00, p < 0.0001). The additional biomarkers only marginally increased AUROC to 0.98 (95% CI 0.95–1.00, p < 0.0001), this increase mainly mediated by AβX-42. In conclusion, CSF biomarkers Aβ1-42, T-tau, and P-tau have very high diagnostic accuracy in a well defined cohort of untreated patients, demonstrating the excellent potency of CSF biomarkers to identify pathological processes in AD when a stringent analytical protocol is used.