P2‐143: Method Comparison of AB(1‐42) Measured in Human Cerebrospinal Fluid Samples by Liquid Chromatography‐Tandem Mass Spectrometry, the Inno‐Bia ALZBIO3 Assay, and the Elecsys^® B‐Amyloid(1‐42) Assay

Abstract

Validated biomarkers for diagnosis of Alzheimer’s disease (AD) are urgently needed. Aβ(1-42) has been shown to correlate with the neurodegeneration observed in AD. Available assays for quantitation of Aβ(1-42) in cerebrospinal fluid (CSF) demonstrate significant variability and lack standardization to reference measurement procedures (RMPs). The Elecsys® β-Amyloid(1-42) assay is a new electrochemiluminescence immunoassay standardized to an approved RMP based on liquid chromatography–tandem mass spectrometry (LC-MS/MS) (Leinenbach et al., 2014), and has proven reproducibility and lot-to-lot comparability (Bittner et al., 2015). The objective of the current study is to evaluate the comparability of the Elecsys® β-Amyloid(1-42) assay to the INNO-BIA AlzBio3 assay (Fujirebio) and an alternative LC-MS/MS candidate RMP (Korecka et al., 2014). 250 de-identified frozen human CSF samples were obtained from laboratory testing samples. Aβ(1-42) concentrations were compared between the Elecsys® β-Amyloid(1–42) and AlzBio3 assays, and LC-MS/MS. All samples underwent ≤2 freeze/thaw cycles. The correlation between different methods was assessed by correlation coefficients (Spearman and Pearson). Regression analysis (Passing–Bablok) was used to evaluate the systematic difference between the Elecsys® β-Amyloid(1-42) assay and LC-MS/MS method. A further 129 CSF samples from pathologically confirmed AD cases (n=62) and age-matched controls (n=67) were analyzed with Elecsys® β-Amyloid(1-42) assay and LC-MS/MS; AlzBio3 Aβ(1-42) concentrations were already available for these samples. Measured Aβ(1-42) concentrations from the Elecsys® β-Amyloid(1–42) assay and from the LC-MS/MS method showed a linear relationship (Figure), with a correlation coefficient (Pearson’s r) of 0.94. Passing–Bablok regression analysis revealed an intercept of 50.8 and slope of 0.97, indicating a percentage bias of <10% within the measuring range. The correlation between the Elecsys® β-Amyloid(1–42) and AlzBio3 assays showed a slightly non-linear relationship with a correlation coefficient (Spearman’s rho) of 0.88. For the 129 additional samples, the clinical performance of all three assays was equivalent in terms of ROC-AUC.