Plasma Amyloid and in vivo Brain Amyloid in Late Middle-Aged Hispanics

Abstract

Background: Determining amyloid positivity is possible with cerebrospinal fluid and brain imaging of amyloid, but these methods are invasive and expensive.

Objective: To relate plasma amyloid-β (Aβ), measured using Single-molecule array (Simoatrademark) assays, to in vivo brain Aβ, measured using positron emission tomography (PET), examine the accuracy of plasma Aβ to predict brain Aβ positivity, and the relation of APOE ɛ4 with plasma Aβ.

Methods: We performed a cross-sectional analysis in a cohort of 345 late middle-aged Hispanic men and women (age 64 years, 72% women). Our primary plasma variable was Aβ42/Aβ40 ratio measured with Simoa. Brain Aβ burden was measured as global SUVR with 18F-Florbetaben PET examined continuously and categorically.

Results: Plasma Aβ42/Aβ40 ratio was inversely associated with global Aβ SUVR (β= -0.13, 95% Confidence Interval (CI): -0.23, -0.03; p = 0.013) and Aβ positivity (Odds Ratio: 0.59, 95% CI: 0.38, 0.91; p = 0.016), independent of demographics and APOE ɛ4. ROC curves (AUC = 0.73, 95% CI: 0.64, 0.82; p < 0.0001) showed that the optimal threshold for plasma Aβ42/Aβ40 ratio in relation to brain Aβ positivity was 0.060 with a sensitivity of 82.4% and specificity of 62.8%. APOE ɛ4 carriers had lower Aβ42/Aβ40 ratio and a higher Aβ positivity determined with the Aβ42/Aβ40 ratio threshold of 0.060.

Conclusion: Plasma Aβ42/Aβ40 ratio assayed using Simoa is weakly correlated with in vivo brain amyloid and has limited accuracy in screening for amyloid positivity and for studying risk factors of brain amyloid burden when in vivo imaging is not feasible.

Keywords: Alzheimer’s disease; amyloid; brain; hispanics; middle age; plasma; positron emission tomography.

Figure 1:

Pearson correlations for age, plasma amyloid β (Aβ) Aβ40, Aβ42, plasma Aβ42/Aβ40 ratio, and global Aβ standardized uptake value ratio (SUVR) measured with ¹⁸F-Florbetaben Positron Emission Tomography (PET) imaging in Hispanic baseline (BL) imaging visits (n = 345). Shaded cells indicate significant correlations (p < 0.05).

Figure 2:

Boxplots comparing the distribution of amyloid β (Aβ) between APOE-ε4 carriers and non-carriers. Aβ measured as global brain Aβ SUVR and the ratio of plasma Aβ42 to Aβ40. Panel A represents a comparison using global brain Aβ SUVR in the full sample (n = 345). Panel B represents a comparison using the ratio of plasma Aβ42 to Aβ40 in the full sample. Panel C represents a comparison using global brain Aβ SUVR removing one subject with an outlying plasma ratio value of 0.165 (n = 344). Panel D represents a comparison using the ratio of plasma Aβ42 to Aβ40 while removing the same outlier. Significance corresponds to a t-test comparing the mean of both groups.

Figure 3:

Receiver operative characteristic (ROC) curve for predicting Aβ positivity (k-means) by plasma Aβ42/Aβ40 ratios, with corresponding AUC value reported. The circle indicates the point at which sensitivity and specificity is maximized (i.e., the distance on the curve to the top left corner is minimized).