Item

Abstract

Background:
A reduced amyloid-β (Aβ)42/40 peptide ratio in blood plasma represents a peripheral biomarker of the cerebral amyloid pathology observed in Alzheimer’s disease brains. The magnitude of the measurable effect in plasma is smaller than in cerebrospinal fluid, presumably due to dilution by Aβ peptides originating from peripheral sources. We hypothesized that the observable effect in plasma can be accentuated to some extent by specifically measuring Aβ1–42 and Aβ1–40 instead of AβX–42 and AβX–40.

Methods:
We assessed the plasma AβX–42/X–40 and Aβ1–42/1–40 ratios in an idealized clinical sample by semi-automated Aβ immunoprecipitation followed by closely related sandwich immunoassays. The amyloid-positive and amyloid-negative groups (dichotomized according to Aβ42/40 in cerebrospinal fluid) were compared regarding the median difference, mean difference, standardized effect size (Cohen’s d) and receiver operating characteristic curves. For statistical evaluation, we applied bootstrapping.

Results:
The median Aβ1–42/1–40 ratio was 20.86% lower in amyloid-positive subjects than in the amyloid-negative group, while the median AβX–42/X–40 ratio was only 15.56% lower. The relative mean difference between amyloid-positive and amyloid-negative subjects was −18.34% for plasma Aβ1–42/1–40 compared to −15.50% for AβX–42/X–40. Cohen’s d was 1.73 for Aβ1–42/1–40 and 1.48 for plasma AβX–42/X–40. Unadjusted p-values < 0.05 were obtained after .632 bootstrapping for all three parameters. Receiver operating characteristic analysis indicated very similar areas under the curves for plasma Aβ1–42/1–40 and AβX–42/X–40.

Conclusions:
Our findings support the hypothesis that the relatively small difference in the plasma Aβ42/40 ratio between subjects with and without evidence of brain amyloidosis can be accentuated by specifically measuring Aβ1–42/1–40 instead of AβX–42/X–40. A simplified theoretical model explaining this observation is presented.