Reproducibility of Alzheimer's Disease CSF-Biomarker Measurements under Clinical Routine Conditions

The differential diagnosis of dementias includes different diagnostic procedures. One of the key tools is the analysis of cerebrospinal fluid (CSF). Elevated CSF levels of total Tau (tTau) and phospho-181-Tau (pTau) in addition to decreased CSF Amyloid-βx-42 (Aβ42) are characteristic CSF hallmarks for Alzheimer's disease (AD). It has been reported previously that the ratio between Aβ42 and Amyloid-βx-40 (Aβ40) (Aβ42/40) can outperform CSF Aβ42 as a diagnostic AD biomarker and appears to be more robust to technical assay variations and pre-analytic procedures. To evaluate the reproducibility of these CSF biomarkers under clinical routine conditions, we compared the biomarker data measured in two different certified clinical laboratories. In this study we were able to show, that there is a disconcordant-rate between both laboratories of 29.6% for tTau and 25.0% for pTau. The proportion of disconcordant cases could be reduced from 31.5% (Aβ42) to 16.8% by applying Aβ42/40. To evaluate the inter-rater agreement, Cohen's kappa (k) was calculated. For tTau and pTau k was 0.449 and 0.528, respectively. The inter-rater agreement for Aβ42 alone appeared to be k = 0.170, indicating a “poor agreement” between the 2 different laboratories, whereas the diagnostic classifications of the subjects according to Aβ42/40 was k = 0.589, indicating a “moderate agreement”. These data point out that the measured CSF biomarker need to be interpreted with caution.