Analytical characteristics of the updated elecsys Abeta42 Gen 2 assay, including comparisons with the Gen 1 assay Developing topics

• Published in: Alzheimer's & dementia Vol. 16; no. S3
• Main Authors: Blennow, Kaj, Zetterberg, Henrik, Rutz, Sandra, Wittig, Martina, Manuilova, Ekaterina, Logan, Chad, Bauer, Ekaterina, Stomrud, Erik
• Format: Journal Article
• Language: English
• Published: 01.12.2020
• ISSN: 1552-5260; 1552-5279
• DOI: 10.1002/alz.047517

Abstract Background Cerebrospinal fluid (CSF) biomarker amyloid‐β 1–42 (Abeta42), quantified using the Elecsys ® Abeta42 (Gen 1) immunoassay, supports diagnosis of Alzheimer’s disease. Compared with Gen1, the updated assay Elecsys Abeta42 CSF II (Gen2) has a higher threshold for biotin interference, extended measuring range, and is re‐standardised according to certified reference Abeta42 peptide ERM‐DA480/‐481/‐482. The new Routine‐Use (RU) pre‐analytical protocol will be introduced with the Gen2 assay (reported separately). Here, we present the analytical performance of the Gen2 assay, including comparisons with Gen1. Method Analytical performance experiments were conducted on cobas e 601 analysers. Precision was investigated according to Clinical and Laboratory Standards Institute EP05‐A3 (21‐day experiment) using seven native and two control samples. For lot and instrument comparisons (cobas e 801 vs e 601), 125 native CSF sample pools were measured in one run. Gen1 and Gen2 were compared in two method comparison (MC) studies: in‐house study: 103 frozen CSF samples were measured in one run; external study (Pre‐analytical study 2): CSF samples from 24 patients with suspected normal pressure hydrocephalus were prospectively collected according to the RU, Roche Trial and BioFINDER protocols. MC was based on mean concentrations obtained from 3–4 aliquots per patient. Aliquot‐aliquot variability was compared for each protocol. MC analyses included calculation of Pearson’s correlation (r) and Passing‐Bablok regression fit. Result Precision analysis provided within‐run variability of <6% and intermediate precision of <8% for all samples. Measurements of three lots were highly correlated (r=1.00) with bias <5.2%, within the clinically relevant range. Measurements on cobas e 601 and e 801 analysers were highly correlated (r=0.995) and regression analysis yielded a slope of 0.941. Gen1 and Gen2 were strongly correlated in both MC studies (in‐house: r=0.999; external: r=0.990). Slope estimates were 0.75 and 0.79 in the in‐house and external studies, respectively (Figure). Aliquot‐aliquot variability was similar for Gen1 and Gen2 assays within each protocol and lower in the RU versus the Roche Trial protocols (Table). Conclusion The Abeta42 Gen2 assay strongly correlated with the Gen1 assay. High precision and low variability between lots and analysers showed the Gen2 assay provided additional benefits over Gen1 without compromising technical performance.