Comparison of two analytical platforms for CSF biomarkers of Alzheimer's disease

Cerebrospinal fluid (CSF) biomarkers of Alzheimer's disease (AD) are currently being assessed with two different assays. Our objective was to study if there is a correlation between values obtained by both techniques, to compare their validity and search for conversion factor between values obtained for every protein. We compared the performances of two commonly used platforms, an enzyme-linked immunosorbent assay (ELISA) and a multiplex (xMAP) technology for measurement of CSF Aβ 1-42, total tau (T-tau), and phosphorylated tau 181 (P-tau 181p) proteins, in 30 AD patients and 28 control subjects. The relations between the variables of both techniques were evaluated using the Spearman p correlation coefficient (α = 0.05). Receiver operating characteristic and area under the curve (AUC) analyses were calculated for the variables of both techniques. The two assays platforms yielded different absolute values for the various analytes, always higher in ELISA. We found some correction factor between values: 2,1- to 3-fold for Aβ 1-42; 4,1- to 4,6-fold for T-tau; and 1,4- to 1,6-fold for P-tau 181p. In addition, those values were highly correlated (Aβ 1-42: r = 0.70, P < 0.01; T-tau: r = 0.90, P < 0.01; P-tau 181p: r = 0.85, P < 0.01) and the AUC for the variables showed very similar values. In conclusion, the results obtained with ELISA and xMAP platforms were highly correlated and its validity is very similar. Differences in absolute values point to the need for a clear description of the technique used. Moreover, we found some conversion factor between values of every protein that may be useful for transformation between both techniques.