O1‐05‐03: Upward Drift in CSf AB42 Values over 10 Years
Suzanne Elizabeth Schindler, Courtney L. Sutphen, John C. Morris, Anne M. Fagan · Alzheimer s & Dementia · 2016
The best established CSF biomarkers for Alzheimer’s disease are levels of amyloid β-peptide 42 (Aβ42), total tau (tau) and ptau-181 (ptau). One limitation of commercial ELISA kits is a typical shelf life of <1 year, so biomarker research programs that collect samples over many years generate data from many different kit lots. We examined whether a widely used commercial assay for CSF Aβ42, tau and ptau provided stable analyte measurements over 10 years. CSF samples (n=341) were collected from research participants enrolled in studies of aging and Alzheimer’s disease at the Knight Alzheimer’s Disease Research Center at Washington University in St. Louis from 2003-2013. Samples were initially assayed a mean of 6 months following collection using the INNOTEST ELISA (Fujirebio, formerly Innogenetics, Ghent, Belgium) with kits that were current at the time of assay. We subsequently re-assayed banked aliquots of the same samples using a single lot of the “Improved INNOTEST” ELISA. We found that Aβ42 values as measured by the single lot of the Improved INNOTEST were significantly higher than the Aβ42 values obtained using many prior lots of the INNOTEST kits (the initial INNOTEST value/single lot Improved INNOTEST was <1 in nearly all cases, see Fig. 1). Further, when compared with values from the single lot of the Improved INNOTEST, Aβ42 values obtained using multiple lots of the INNOTEST kits have been drifting upwards over the last decade by 2.6% per year (p<0.0001, Fig. 1). In contrast, tau levels have been drifting slightly downwards by 1.0% per year (p<0.01) and ptau levels have been drifting slightly upwards by 1.7% per year (p<0.05). In our dataset there is a significant drift in CSF analyte values over one decade as measured by INNOTEST. This drift effect is largest in magnitude for Aβ42. We suggest that investigators 1) evaluate the stability of CSF analyte values in their own cohort, especially if the values were obtained over multiple years using multiple assay lots, and/or 2) include lot number as a covariate in their analyses. Aβ42 values by single run Improved INNOTEST versus initial values measured after CSF collection