Head-to-Head Comparison of 8 Plasma Amyloid-β 42/40 Assays in Alzheimer Disease

Abstract

Key Points Question How well does plasma amyloid-β 42/40 (Aβ42/40), measured using 8 different assays, detect brain Aβ pathology in the early stages of Alzheimer disease? Findings In this study, including 408 participants from 2 independent cohorts (BioFINDER and Alzheimer Disease Neuroimaging Initiative), plasma Aβ42/40 quantified using certain mass spectrometry–based methods showed better discriminative accuracy than immunoassays when identifying individuals with abnormal intracerebral Aβ status according to cerebrospinal fluid Aβ42/40 levels and Aβ positron emission tomography. Meaning Certain mass spectrometry–based plasma tests might have sufficient performance to detect brain Aβ pathology in Alzheimer disease.

Performance of plasma amyloid, tau, and astrocyte biomarkers to identify cerebral AD pathophysiology

“…These results add to the conflicting literature in which some studies suggest that plasma Aβ42 or Aβ42/40 ratio strongly correlated with brain Aβ (AUC = 88 to 97%) [21,25], while others suggest only a moderate association between plasma and PET assessments of Aβ [43,44]. Moreover, previous research revealed that IP-MS performed better than Simoa in terms of Aβ estimates [23], whereas we found similar discriminative accuracy using both techniques. Several biological and analytical factors may contribute to the divergence of our results.…”

“…For example, biological factors such as peripheral Aβ expression accounts to >50% of the global plasma Aβ [45] and the modest fold-change between CU and CI in plasma (10–20%) compared to CSF Aβ (40– 60%) may lead to the high susceptibility of plasma Aβ measures to small variabilities in pre-analytical approaches and cohorts characteristics [20, 21, 44]. In addition, differences in the analytical performance of distinct techniques used to measure plasma Aβ can also play a role in the difference in performance found across studies [23]. We have shown lack of robust correlation between plasma Aβ values and the other AD-related plasma markers.…”

“…Plasma Aβ can be quantified by immunoprecipitation coupled to mass spectrometry (IP-MS) and by novel ultrasensitive immunoassay techniques [20][21][22][23], including fully automated clinical chemistry instruments [24]. Both methods can detect with relatively high precision individuals with abnormal brain Aβ burden or those at high risk of future conversion to Aβ-PET positivity [11,20,21,23,25].…”

“…Plasma Aβ can be quantified by immunoprecipitation coupled to mass spectrometry (IP-MS) and by novel ultrasensitive immunoassay techniques [20][21][22][23], including fully automated clinical chemistry instruments [24]. Both methods can detect with relatively high precision individuals with abnormal brain Aβ burden or those at high risk of future conversion to Aβ-PET positivity [11,20,21,23,25]. More recently, plasma levels of tau protein phosphorylated at threonine 181 (p-tau181), 217 (p-tau217), and 231 (p-tau231) have shown high accuracy to detect both tau and Aβ pathologies [8,11,13].…”

Blood amyloid and tau biomarkers as predictors of cerebrospinal fluid profiles

“…These plasma biomarkers are also those identified as the best predictors of AD (Palmqvist et al 2021 ), but here, the context of use is different, and their performance are event higher than for discriminating AD from non-AD. Other differences are noted such as the fact that Aβ 1–42 detection by IP-MS outperformed other Aβ 1–42 detections, but this was not the case when considering the Aβ 1–42 /Aβ 1–40 ratio, thus differently than when AD is the performance criterion (Janelidze et al 2021 ). As mentioned above, this could be partly explained by the fact that diseases other than AD showed pathological amyloid or tau profiles.…”