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

INTRODUCTION

Decades of advances in Alzheimer’s disease (AD) research, particularly in cerebrospinal fluid (CSF) and brain imaging biomarkers [1, 2], have led to the dominance of brain amyloid as the main AD neuropathological construct, followed by tau and neurodegeneration [1]. These constructs feature prominently in the recent National Institute on Aging (NIA)/Alzheimer’s Association (AA) 2018 research framework [3]. This framework proposes to conduct research in which individuals are classified as having the AD continuum by amyloid positivity, defined by amyloid biomarkers. Determining amyloid positivity has been enabled by the widespread availability of accurate CSF and brain imaging markers of amyloid-β (Aβ). As compared with cognitively normal individuals, patients with AD dementia show lower CSF Aβ₄₂ concentration [4] and higher brain amyloid burden on amyloid positron emission tomography (PET) [5]. However, the use of CSF and PET biomarkers of amyloid status is limited by cost and by the relative burden of undergoing brain imaging with radiation exposure or lumbar puncture for the measurement of CSF biomarkers. Thus, inexpensive and safe blood-based biomarkers would be ideal for the ascertainment of Aβ-positivity status. More recently, ultrasensitive immunoassays coupled with mass spectrometry show greater promise [6]. For example, the commercially available Single-molecule array (Simoatrademark) is a novel method to measure Aβ₄₀ and Aβ₄₂ in plasma [7]. Whereas CSF and brain imaging biomarkers are relatively well-established and standardized in AD research, blood-based biomarkers need further development. Our primary objective was to examine the relation of plasma Aβ₄₂/Aβ₄₀ ratio, Aβ₄₀, and Aβ₄₂ measured using Simoa assays to in vivo brain Aβ measured using PET in a late middle-aged Hispanic cohort. We also examined the accuracy of plasma Aβ to predict brain amyloid positivity determined quantitatively using PET imaging. Lastly, we explored whether the association between APOE ɛ4 genotype and Aβ burden ascertained on PET could be replicated with plasma Aβ [8].

MATERIALS AND METHODS

Study design and population

This was a cross-sectional analysis of a community-based cohort of 345 participants of a study focusing on the relation of AD risk factors and brain amyloid in middle-aged Hispanics conducted at Columbia University Irving Medical Center (CUIMC) in New York City, recruited between 03/01/2016 and 07/31/2019 [9]. We targeted Hispanics because they are the most common ethnic group in the community surrounding CUIMC [10] and because there is a paucity of AD biomarkers studies in Non-Whites [1]. Participants were recruited from the community through various outreach activities, including presentations at churches and senior centers, posters at CUIMC and around the community, health fairs, and newspaper ads. Study research staff collaborated with the Community Engagement Core Resource of the Irving Institute for Clinical and Translational Research, the Clinical Translational Science Award at CUIMC, to conduct outreach in the community and promote the study. Recruitment was exclusively community-based and did not include recruitment from clinics or using electronic medical records. Inclusion criteria included men and women aged 55–69 years and who were able to undergo phlebotomy, clinical and neuropsychological assessments, 3T brain magnetic resonance imaging (MRI), and positron emission tomography (PET) with the Aβ radioligand ¹⁸F-Florbetaben. Exclusion criteria were a dementia diagnosis, cancer diagnosis other than non-melanoma skin cancer, and MRI contraindications. We screened 659 potential participants; 114 (17.3%) declined to participate, 178 (27.0%) were ineligible, and 16 (2.4%) did not complete study procedures (Supplementary Figure 1). One additional participant (0.2%) was excluded from the analyses due to incomplete data on APOE ɛ4 genotype, the most important genetic predictor of in vivo brain amyloid burden [5]. Five (0.8%) participants were also excluded due to incomplete plasma Aβ data. The mean interval between amyloid PET and MRI was 15.8 ± 33.4 days and the mean interval between PET scan and phlebotomy was 14.0 ± 24.2 days.

Aβ plasma biomarkers

Our primary plasma variable of interest was plasma Aβ₄₂/Aβ₄₀ ratio, as is usual in studies examining plasma Aβ [11 –13], but also examined plasma Aβ₄₀ and Aβ₄₂ individually. Plasma Aβ₄₀ and Aβ₄₂ were measured using the commercially available ultrasensitive Simoatrademark assays performed by Quanterixtrademark (Billerica, MA, USA) [14] using frozen plasma aliquots. Aβ₄₀ and Aβ₄₂ were assayed together using a multiplex assay (Human Neurology 3-Plex Total). The Simoa Aβ₄₀ and Aβ₄₂ assays target the N-terminus of beta amyloid and different C-terminus detection antibodies specific to Aβ₄₀ and Aβ₄₂. Aβ₄₂ (mean range: 0–400 pg/mL) has a lower limit of detection of 0.019–0.034 pg/mL, a reproducibility coefficient of variation (CV) = 7.5% and repeatability CV = 6.7% [14 –17]. Aβ₄₀ (mean range: 0–800 pg/mL) has a lower limit of detection of 0.16 pg/mL, a reproducibility CV = 5.1% and repeatability CV = 3.5% [17 –19]. All samples were diluted 4-fold for Aβ₄₂ and 8-fold for Aβ₄₀ using a proper sample diluent (PBS containing carrier protein and detergent) for measurement.

Brain Aβ

Brain Aβ burden was ascertained as global standardized uptake value ratio (SUVR) measured with ¹⁸F-Florbetaben PET. MRI was conducted for automatic delineation of the region of interest (ROI) in the PET images. MRI images were acquired in a General Electric Signa Premier 3T scanner and processed with FreeSurfer (v6.0 http://surfer.nmr.mgh.harvard.edu/). Participants underwent ¹⁸F-Florbetaben PET imaging in a Siemens Biograph64 mCT/PET scanner (target dose: 8.1 mCi; 4×5 min frames; iterative reconstruction algorithm; voxel size: 1.6 x 1.6 x 1 mm³. Images were acquired over 20 min starting 90 min post-injection. Dynamic PET frames (4 scans) were aligned to the first frame using rigid-body registration and a static PET image was obtained by averaging the four registered frames. The static PET image was then registered with the CT scan obtained for attenuation correction during PET imaging reconstruction by rigid-body registration with information theoretic cost function to generate a fused image with skull. The structural T1-weighted image in FreeSurfer space was registered to the CT/PET fused image using normalized mutual information and six degrees of freedom. A combination of the two transformation matrices obtained from the two rigid-body registrations was used to transfer all Freesurfer regional masks and the cerebellar gray matter from FreeSurfer space to static PET image space using nearest neighbor interpolation [20]. The standardized uptake value (SUV), defined as the decay-corrected brain radioactivity concentration normalized for injected dose and body weight, was calculated in all FreeSurfer regions. The SUV in each region as well as each voxel was normalized to the SUV in cerebellar gray matter to derive the regional and voxel-wise SUVR. Overall mean Aβ burden was calculated from voxel-based, individual ROI, including lateral temporal cortex, parietal cortex, cingulate cortex, and frontal cortex.

We categorized Aβ as positive using an SUVR threshold of 1.34 [9], determined using the K-means clustering method, which identifies the partition between the 2 peaks in the Aβ SUVR distribution and quantitatively determines Aβ positivity (Fig. 1A).

Fig. 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).

Covariates

We examined age, education, Hispanic subgroup, and APOE ɛ4 genotype, and cystatin C as covariates. Hispanic subgroup was classified following the format of the 2010 Census by country or region of origin (e.g., Mexican, Puerto Rican, Cuban, Dominican) [21]. APOE ɛ4 genotyping was conducted by LGC genomics (Beverly, MA) using single nucleotide polymorphisms rs429358 and rs7412. Cystatin C was measured in serum using a particle enhanced immune turbidimetric assay (Cobas Integra 400 plus; Roche Diagnostics, Indianapolis, IN).

The rationale for the covariates is as follows. Age, sex, and education are important predictors of dementia. APOE ɛ4 genotype is the strongest risk factor for sporadic dementia due to AD [22] and also the strongest determinant of in vivo amyloid burden [8, 23]. Thus, we also examined APOE ɛ4 genotype as an exposure in relation to Aβ burden ascertained on PET and in plasma. We used cystatin as a measure of renal function to explore whether renal function affected plasma Aβ levels.

Statistical analyses

Global Aβ SUVR (Supplementary Figure 2A) had a bimodal distribution while plasma Aβ₄₀ (Supplementary Figure 2C), Aβ₄₂ (Supplementary Figure 2D), and Aβ₄₂/Aβ₄₀ ratio (Supplementary Figure 2B) were normally distributed. Descriptive analysis used chi-squared and ANOVA tests to examine differences in demographic characteristics across brain Aβ positivity categories. Pearson correlation coefficients were used to explore the relationship between brain Aβ (both globally and regionally) and plasma Aβ. Correlations were reported as both unadjusted and age-adjusted partial correlations. The association of APOE ɛ4 carrier status with brain and plasma Aβ was examined using t-tests. The relationship between plasma Aβ and continuous brain Aβ SUVR was evaluated using multivariable linear regression. The association between plasma Aβ and brain Aβ positivity was examined with multivariable logistic regression. We also examined Cohen’s D effect size to evaluate the standardized difference in plasma Aβ means between brain Aβ positivity groups. Models were adjusted for age, sex, and APOE ɛ4 carrier status. We examined plasma Aβ₄₀, Aβ₄₂, and Aβ₄₂/Aβ₄₀ ratio by increments of 1 standard deviation in relation to the outcome. The justification for standardizing the exposures and outcomes was that examining unit increments yielded significant but very small coefficients. We conducted additional sensitivity analyses adjusting models for Cystatin C to control for renal function.

We also assessed the accuracy with which plasma Aβ could predict brain Aβ positivity through the use of receiver operating characteristic (ROC) curves, plotting the true positive rate of these predictions against the false positive rate. The area under these curves (AUC) is a measure of each model’s ability to distinguish between brain Aβ positivity and negativity, where a higher AUC represents a more accurate prediction. These results were compared to the accuracy of APOE ɛ4 carrier status in similarly predicting brain Aβ positivity. We calculated 95% confidence intervals for AUC values using stratified bootstrap samples (n = 1000). The general significance of AUC curves was determined using a Wilcoxon test [24]. AUC values were compared using a nonparametric approach outlined in DeLong et al. [25]. We determined the optimal value for plasma Aβ₄₂/Aβ₄₀ ratio corresponding with amyloid positivity by identifying which plasma Aβ₄₂/Aβ₄₀ ratio value maximized sensitivity and specificity of the corresponding ROC curve. We calculated the positive predictive value (PPV) and negative predictive value (NPV) for each of these values. Statistical significance was considered at p < 0.05. Analyses were conducted using SAS version 9.4 m5 and R version 3.6.0.

RESULTS

The mean age of the sample was 64.1 ± 3.3 years, 72% were women, and the APOE ɛ4 carrier prevalence was 35.7% (Table 1). There were no differences in demographic variables between Aβ-negative and Aβ-positive participants (ascertained with PET), but as expected, the prevalence of APOE ɛ4 carriers among Aβ-positive participants was more than twice that among Aβ-negative participants. Plasma Aβ₄₀ levels were similar across Aβ positive and Aβ negative participants. Plasma Aβ₄₂ and Plasma Aβ₄₂/Aβ₄₀ ratio were approximately 10% lower among Aβ positive participants compared with Aβ negative participants. Similarly, there were small effect sizes between Aβ positive and Aβ negative participants for Plasma Aβ₄₀ (d = 0.01) levels, while the same differences for Plasma Aβ₄₂ (d = 0.37) and Plasma Aβ₄₂/Aβ₄₀ ratio (d = 0.55) were much larger.

Table 1

Descriptive statistics for entire sample and by amyloid β (Aβ) positivity determined from Aβ positron emission tomography. Aβ positivity was by k-means clustering (SUVR = 1.34)

Characteristic	Entire Sample (n = 345)	Aβ Negative (n = 317)	Aβ Positive (n = 28)	p
Age, mean (SD), y	64.13 (3.32)	64.06 (3.33)	64.88 (3.16)	0.20
Women, No. (%)	248 (71.88)	227 (71.61)	21 (75.00)	0.87
Education, mean (SD), y	10.55 (3.96)	10.45 (3.97)	11.71 (3.68)	0.092
Ethnicity, No. (%)
Dominican	295 (85.51)	273 (86.12)	22 (78.57)	0.42
Other Caribbean Hispanic	21 (6.09)	18 (5.68)	3 (10.71)
South American	18 (5.22)	16 (5.05)	2 (7.14)
Unspecified Hispanic	7 (2.03)	6 (1.89)	1 (3.57)
Central American	4 (1.16)	4 (1.26)	0 (0)
APOE ɛ4, No. (%)	123 (35.65)	102 (32.18)	21 (75.00)	< 0.0001
Cystatin C, mean (SD), mg/L	0.97 (0.43)	0.98 (0.44)	0.94 (0.15)	0.78
SRT: Total Recall, mean (SD), words	38.94 (8.82)	39.00 (8.69)	38.33 (10.40)	0.75
Plasma Aβ₄₀, mean (SD), pg/mL	222.98 (61.15)	223.05 (61.14)	222.18 (62.33)	0.94
Plasma Aβ₄₂, mean (SD), pg/mL	14.27 (3.98)	14.39 (3.98)	12.91 (3.77)	0.057
Aβ₄₂/Aβ₄₀ ratio, mean (SD)	0.065 (0.013)	0.065 (0.013)	0.058 (0.009)	0.0005

Plasma Aβ₄₂/Aβ₄₀ ratio was weakly but significantly inversely correlated with age, plasma Aβ₄₀, global Aβ SUVR, and Aβ SUVR in frontal, cingulate, parietal, and temporal ROIs, and moderately correlated with plasma Aβ₄₂ (Fig. 1). One individual had an outlying plasma Aβ₄₂/Aβ₄₀ ratio value (0.17), but the magnitudes of our correlations were remained the same even after removing this value. The magnitudes of our correlations also remained the same after adjusting for age.

We compared the association of APOE ɛ4 carrier status with brain Aβ and plasma Aβ (Fig. 2). APOE ɛ4 carrier status was associated with higher brain Aβ SUVR (β= 0.08; 95% Confidence Interval (CI): 0.05, 0.11; p < 0.0001). APOE ɛ4 carrier status was not associated with plasma Aβ₄₀ nor plasma Aβ₄₂, but APOE ɛ4 carriers had lower plasma Aβ₄₂/Aβ₄₀ ratios (β= –0.003; 95% CI: –0.006, 0.0002; p = 0.067). However, the outlying plasma Aβ₄₂/Aβ₄₀ ratio value imparted an effect on the association between plasma Aβ₄₂/Aβ₄₀ ratio and APOE ɛ4 carrier status, and removing this value resulted in a significant association (β= –0.004; 95% CI: –0.006, –0.001; p = 0.0039). APOE ɛ4 carriers also had a higher odds of amyloid positivity compared to non-carriers (OR = 6.32, CI: 2.72, 16.49; p < 0.0001). These results were robust to adjustments for Cystatin C.

Fig. 2

Boxplots comparing the distribution of amyloid-β (Aβ) between APOE ɛ4 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.

Plasma Aβ₄₂/Aβ₄₀ ratio was inversely associated with global Aβ SUVR and Aβ positivity (Table 2), independent of demographics and APOE ɛ4 carrier status. The associations for plasma Aβ₄₂ were weaker but were in the same direction as for Aβ₄₂/Aβ₄₀ ratio. Plasma Aβ₄₀ was not associated with global Aβ SUVR or Aβ positivity. These results were robust to additional adjustments for Cystatin C.

Table 2

Regression analysis relating 1 standard deviation increment in plasma amyloid-β (Aβ) Aβ₄₀, Aβ₄₂, and Aβ₄₂/Aβ₄₀ ratio in relation to global Aβ derived from PET imaging. Global brain Aβ was examined continuously by linear regression and categorically by logistic regression. Global Aβ positivity was determined by k-means (SUVR = 1.34). Standardized betas and odds ratios are reported. Model 1 is unadjusted. Model 2 adjusts for age, and sex. Model 3 adjusts for age, sex, and APOE ɛ4 carrier status

		Model 1		Model 2		Model 3
	No.	Aβ Outcome	β (95% CI), OR (95% CI)	p	β (95% CI), OR (95% CI)	p	β (95% CI), OR (95% CI)	p
β (95% CI) Global Aβ SUVR
Aβ₄₀			0.03 (–0.08, 0.13)	0.60	0.02 (–0.08, 0.13)	0.67	0.01 (–0.09, 0.11)	0.86
Aβ₄₂	345		–0.09 (–0.20, 0.02)	0.095	–0.09 (–0.20, 0.01)	0.080	–0.08 (–0.18, 0.02)	0.14
Aβ₄₂/Aβ₄₀			–0.16 (–0.27, –0.06)	0.0028	–0.16 (–0.26, –0.05)	0.0031	–0.13 (–0.23, –0.03)	0.013
OR (95% CI) Aβ Positivity by k-means (SUVR = 1.34)
Aβ₄₀			0.99 (0.64, 1.40)	0.94	0.96 (0.62, 1.37)	0.84	0.94 (0.62, 1.33)	0.74
Aβ₄₂	345	Positive	0.68 (0.45, 1.01)	0.060	0.68 (0.45, 1.01)	0.060	0.69 (0.44, 1.06)	0.093
Aβ₄₂/Aβ₄₀			0.59 (0.41, 0.86)	0.0047	0.61 (0.42, 0.88)	0.0081	0.59 (0.38, 0.91)	0.016

We examined a ROC curve to evaluate the accuracy of the plasma Aβ₄₂/Aβ₄₀ ratio in estimating brain Aβ positivity, and the AUC was 0.73 (95% CI: 0.64, 0.82; p < 0.0001). We also considered alternative methods of predicting estimating brain Aβ positivity, evaluating the accuracy of APOE ɛ4 carrier status as a predictor (AUC = 0.71, 95% CI: 0.62, 0.80; p < 0.0001) as well as APOE ɛ4 carrier status and plasma Aβ₄₂/Aβ₄₀ ratio used jointly as predictors (AUC = 0.81, 95% CI: 0.73, 0.87; p < 0.0001). There was no statistical difference in predictive strength between the accuracy of the plasma Aβ₄₂/Aβ₄₀ ratio and APOE ɛ4 carrier status (p = 0.82) or APOE ɛ4 carrier status and plasma Aβ₄₂/Aβ₄₀ ratio modelled simultaneously (p = 0.18) by DeLong’s test. We further explored the point at which the sensitivity and specificity of Aβ positivity predictions were optimized. This was done by finding the point on the ROC curve closest to 100% sensitivity and 100% specificity. This was a threshold of 82.4% sensitivity and 62.8% specificity. This corresponded to a plasma Aβ₄₂/Aβ₄₀ ratio value of 0.06 (Fig. 3). We examined the association of APOE ɛ4 carrier status with amyloid positivity defined by this plasma Aβ₄₂/Aβ₄₀ ratio threshold (0.06) and found that individuals carrying an ɛ4 allele had higher odds of amyloid positivity by plasma Aβ₄₂/Aβ₄₀ ratio threshold compared to APOE ɛ4 non-carriers (OR = 2.73, 95% CI: 1.69, 4.43; p < 0.0001).

Fig. 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).

Finally, we explored the effect of setting sensitivity and specificity levels at 90%, depending on whether we intend to use plasma Aβ₄₂/Aβ₄₀ ratio to screen for persons who are likely to be brain Aβ-positive versus screen out those who are Aβ-negative. Using a plasma Aβ₄₂/Aβ₄0 ratio cutoff with a sensitivity of 90% (1 in 10 false negative) with the goal of capturing as many truly Aβ-positive participants as possible, the specificity is 39.7%, meaning that nearly 6 in 10 individuals who screen positive will be Aβ-negative. Using this cutoff there was a PPV of 11.6% and an NPV of 97.8% for predicting brain amyloid positivity, implying that only 1 out of every 50 negative test result will correspond to an amyloid positive individual. If we set the specificity at 90% (1 in 10 false positive) with the intent of excluding as many Aβ-negative individuals as possible, the sensitivity would be 25%, meaning that only a quarter of Aβ-positive individuals screened would be included. This cutoff had a PPV of 18.1% and an NPV of 93.1% for predicting brain amyloid positivity. Thus, it is of potentially greater value to use plasma Aβ₄₂/Aβ₄₀ ratio in screening for Aβ-positive individuals as compared to Aβ-negative individuals.

DISCUSSION

We found that in a community-based late middle age cohort that plasma Aβ₄₂/Aβ₄₀ ratio is weakly but significantly correlated with global brain amyloid SUVR, with the ability to predict amyloid positivity determined quantitatively with amyloid PET at a limited level (AUC within a 0.70–0.80 range) [26]. Although the plasma Aβ₄₂/Aβ₄₀ ratio generally had weak PPV due to the low prevalence of brain amyloid positivity, it had a high NPV and rarely classified amyloid positive cases as negative. Plasma Aβ₄₂/Aβ₄₀ ratio measured using Simoa may have limited use as a surrogate marker of brain amyloid levels in epidemiologic, prevention, or treatment studies given the weak correlation. Plasma Aβ₄₂/Aβ₄₀ ratio measured using Simoa seems to have limited accuracy for screening for Aβ-positive individuals in late middle-age, an important period of the lifespan for interventions targeting AD. Our finding showing that plasma Aβ₄₂/Aβ₄₀ ratio is lower among APOE ɛ4 carriers as compared with non-carriers, in the same direction as for Aβ SUVR, suggests that plasma Aβ₄₂/Aβ₄₀ ratio may be used, albeit with limited accuracy, for the examination of the relation between risk factors and Aβ burden when brain imaging or lumbar puncture is not possible. Moreover, APOE ɛ4 and plasma Aβ₄₂/Aβ₄₀ ratio had similar strength in identifying brain Aβ positivity, but a higher magnitude AUC in models considering plasma Aβ and APOE ɛ4 together may suggest that plasma Aβ is able to capture information about brain Aβ positivity that was not observed through the association between APOE ɛ4 and Aβ positivity.

The development of accurate blood-based biomarkers of AD has lagged behind brain imaging and CSF biomarkers, but recent advances enable the use of blood-based biomarkers in AD research. The blood-brain barrier is altered in aging and AD [27]. The increased permeability between the brain and the periphery makes it possible for blood-based biomarkers to be representative of pre-clinical changes in AD [28]. Extant proteomic methods to measure blood-based biomarkers for AD include immunocapture, and aptamer-based techniques. However, issues around lower limit of detection, depletion of lower molecular weight proteins, and antibody availability have limited the use of these methods in particular [29]. More recently, ultrasensitive immunoassays and mass spectrometry show greater promise [6, 30]. The commercially available Simoatrademark is a novel method to measure Aβ₄₀ and Aβ₄₂ in plasma [7], that is increasingly used in research. Whereas CSF and brain imaging biomarkers are relatively well established and standardized in AD research, research using blood-based biomarkers has advanced more recently.

Using Simoatrademark technology in 248 participants aged 61 ± 9 years with subjective cognitive decline from the SCIENCe project and Amsterdam Dementia Cohort, where amyloid PET was available for 69 participants, the predictive accuracy of plasma amyloid to discriminate participants with an abnormal amyloid PET scan from those with a normal amyloid PET scan yielded an AUC of 66% (95% CI = 53–79%) for plasma Aβ₄₂ alone and 68% (95% CI = 55–82%) for the plasma Aβ₄₂/Aβ₄₀ ratio. The authors reported that the number of PET scans would be reduced by 54% when applying the plasma ratio as prescreener suggesting that plasma Aβ₄₂/Aβ₄₀ has the potential to be used as a screening measure to identify AD related neuropathological changes in cognitively normal individuals with subjective cognitive decline [19]. Plasma Aβ₄₂/Aβ₄₀, measured with Simoatrademark, was also predictive of cerebral amyloidosis in a sample of 276 cognitively intact individuals aged approximately 77 years with subjective memory complaints from the INSIGHT-preAD study, a French academic university-based cohort that is part of the Alzheimer Precision Medicine Initiative Cohort Program, with an AUC of 0.77 [13]. Compared to these studies, our study examined the relation of plasma Aβ with in vivo brain Aβ in a community-based cohort recruited irrespective of cognitive status and found a modest correlation between plasma Aβ₄₂/Aβ₄₀ ratio and brain Aβ SUVR, and limited accuracy of plasma Aβ₄₂/Aβ₄₀ ratio in predicting amyloid positivity, with an AUC of 0.73, lower than in recent studies. It is also possible that the association between plasma amyloid and brain Aβ SUVR is stronger in older adults than it is in middle-aged or late-middle-aged individuals. A recent longitudinal study from the Pittsburgh center of the Ginkgo Evaluation of Memory Study (n = 194, mean age = 85 years), reported that plasma Aβ₄₂/Aβ₄₀ ratio at baseline was not associated with amyloid PET positivity. However, the association was significant at 8 years follow-up in cognitively normal participants [31].

Our study has some limitations to consider. It is possible that other techniques of measuring plasma amyloid such as mass spectrometry is superior to Simoa assays [30]. We could not however compare the accuracy of the Simoa assay to other plasma biomarkers of amyloid positivity, including other plasma Aβ assays [32, 33] and plasma assays of phosphorylated forms of tau [34]. We cannot address the ability of plasma Aβ₄₂/Aβ₄₀ ratio in predicting brain amyloid changes given the cross-sectional nature of our study. Despite these limitations, our study also has strengths, including the use of state-of-the-art amyloid imaging to ascertain amyloid status quantitatively. This study also addresses the lack of biomarker studies in non-white individuals. Moreover, most our findings were similar to European studies mentioned above, suggesting that our findings are in fact generalizable.

The main implications of our study are that plasma Aβ₄₂/Aβ₄₀ ratio measured using Simoa has limited accuracy to screen for amyloid positivity and for studying risk factors of brain amyloid burden when n-vivo amyloid imaging is not feasible.

Footnotes

ACKNOWLEDGMENTS

This work was funded by National Institutes of Health grants R01AG050440, RF1AG051556. Partial support was provided by grants K24AG045334, P30AG059303, and ULT1TR001873. H Zetterberg is a Wallenberg Scholar supported by grants from the Swedish Research Council (#2018-02532), the European Research Council (#681712), Swedish State Support for Clinical Research (#ALFGBG-720931), the Alzheimer Drug Discovery Foundation (ADDF), USA (#201809-2016862), and the UK Dementia Research Institute at UCL.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Jack

Jr , Knopman

, Jagust

, Petersen

, Weiner

, Aisen

, Shaw

, Vemuri

, Wiste

, Weigand

, Lesnick

, Pankratz

, Donohue

, Trojanowski

(2013) Tracking pathophysiological processes in Alzheimer’s disease: An updated hypothetical model of dynamic biomarkers. Lancet Neurol 12, 207–216.

Jack

Jr , Knopman

, Jagust

, Shaw

, Aisen

, Weiner

, Petersen

, Trojanowski

(2010) Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol 9, 119–128.

Jack

Jr , Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

, Contributors (2018) NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement 14, 535–562.

Olsson

, Lautner

, Andreasson

, Öhrfelt

, Portelius

, Bjerke

, Hölttä

, Rosén

, Olsson

, Strobel

, Wu

, Dakin

, Petzold

, Blennow

, Zetterberg

(2016) CSF and blood biomarkers for the diagnosis of Alzheimer’s disease: A systematic review and meta-analysis. Lancet Neurol 15, 673–684.

Ossenkoppele

, Jansen

, Rabinovici

, Knol

, van der Flier

, van Berckel

BNM

, Scheltens

, Visser

, Amyloid

PETSG

, Verfaillie

SCJ

, Zwan

, Adriaanse

, Lammertsma

, Barkhof

, Jagust

, Miller

, Rosen

, Landau

, Villemagne

, Rowe

, Lee

, Na

, Seo

, Sarazin

, Roe

, Sabri

, Barthel

, Koglin

, Hodges

, Leyton

, Vandenberghe

, van Laere

, Drzezga

, Forster

, Grimmer

, Sánchez-Juan

, Carril

, Mok

, Camus

, Klunk

, Cohen

, Meyer

, Hellwig

, Newberg

, Frederiksen

, Fleisher

, Mintun

, Wolk

, Nordberg

, Rinne

, Chételat

, Lleo

, Blesa

, Fortea

, Madsen

, Rodrigue

, Brooks

(2015) Prevalence of amyloid PET positivity in dementia syndromes: A meta-analysis. JAMA 313, 1939–1949.

Andreasson

, Blennow

, Zetterberg

(2016) Update on ultrasensitive technologies to facilitate research on blood biomarkers for central nervous system disorders. Alzheimers Dementia (Amst) 3, 98–102.

Zetterberg

, Mörtberg

, Song

, Chang

, Provuncher

, Patel

, Ferrell

, Fournier

, Kan

, Campbell

, Meyer

, Rivnak

, Pink

, Minnehan

, Piech

, Rissin

, Duffy

, Rubertsson

, Wilson

, Blennow

(2011) Hypoxia due to cardiac arrest induces a time-dependent increase in serum amyloid β levels in humans. PLoS One 6, e28263.

Palta

, Rippon

, Reitz

, He

HA-O

, Sherwood

, Ceballos

, Teresi

, Razlighi

, Moreno

, Brickman

AA-O

, Luchsinger

(2020) Apolipoprotein E genotype and in-vivo amyloid burden in middle-aged hispanics. Neurology 95, e2086–e2094.

Luchsinger

, Palta

, Rippon

, Sherwood

, Soto

, Ceballos

, Laing

, Igwe

, He

, Razlighi

, Teresi

, Moreno

, Brickman

(2020) Pre-diabetes, but not type 2 diabetes, is related to brain amyloid in late middle-age. J Alzheimers Dis 75, 1241–1252.

10.

Luchsinger

, Cabral

, Eimicke

, Manly

, Teresi

(2015) Glycemia, diabetes status, and cognition in hispanic adults aged 55-64 years. Psychosom Med 77, 653–663.

11.

Risacher

, Fandos

, Romero

, Sherriff

, Pesini

, Saykin

, Apostolova

(2019) Plasma amyloid beta levels are associated with cerebral amyloid and tau deposition. Alzheimers Dement (Amst) 11, 510–519.

12.

Schindler

, Bollinger

, Ovod

, Mawuenyega

, Li

, Gordon

, Holtzman

, Morris

, Benzinger

TLS

, Xiong

, Fagan

, Bateman

(2019) High-precision plasma beta-amyloid 42/40 predicts current and future brain amyloidosis. Neurology 93, e1647–e1659.

13.

Vergallo

, Megret

, Lista

, Cavedo

, Zetterberg

, Blennow

, Vanmechelen

, De Vos

, Habert

, Potier

, Dubois

, Neri

, Hampel

, INSIGHT-preAD study group; Alzheimer Precision Medicine Initiative (APMI) (2019) Plasma amyloid beta 40/42 ratio predicts cerebral amyloidosis in cognitively normal individuals at risk for Alzheimer’s disease. Alzheimers Dement 15, 764–775.

14.

Wilson

, Rissin

, Kan

, Fournier

, Piech

, Campbell

, Meyer

, Fishburn

, Cabrera

, Patel

, Frew

, Chen

, Chang

, Ferrell

, von Einem

, McGuigan

, Reinhardt

, Sayer

, Vielsack

, Duffy

(2015) The Simoa HD-1 Analyzer: A novel fully automated digital immunoassay analyzer with single-molecule sensitivity and multiplexing. J Lab Autom 21, 533–547.

15.

Zetterberg

, Wilson

, Andreasson

, Minthon

, Blennow

, Randall

, Hansson

(2013) Plasma tau levels in Alzheimer’s disease. Alzheimers Res Ther 5, 9.

16.

Blennow

, Zetterberg

(2018) The past and the future of Alzheimer’s disease fluid biomarkers. J Alzheimers Dis 62, 1125–1140.

17.

Janelidze

, Stomrud

, Palmqvist

, Zetterberg

, van Westen

, Jeromin

, Song

, Hanlon

, Tan Hehir

, Baker

, Blennow

, Hansson

(2016) Plasma beta-amyloid in Alzheimer’s disease and vascular disease. Sci Rep 6, 26801.

18.

Lue

, Guerra

, Walker

(2017) Amyloid beta and tau as Alzheimer’s disease blood biomarkers: Promise from new technologies. Neurol Ther 6, 25–36.

19.

Verberk

IMW

, Slot

, Verfaillie

SCJ

, Heijst

, Prins

, van Berckel

BNM

, Scheltens

, Teunissen

, van der Flier

(2018) Plasma amyloid as prescreener for the earliest Alzheimer pathological changes. Ann Neurol 84, 648–658.

20.

Tahmi

, Bou-Zeid

, Razlighi

(2019) A fully automatic technique for precise localization and quantification of Amyloid-β PET scans. J Nucl Med 60, 1771–1779.

21.

Humes

, Jones

, Ramirez

(2011) Overview of Race and Hispanic Origin: 2010. 2010 Census Briefs, U.S. Department of Commerce, U.S. Census Bureau, Washington, DC.

22.

(2019) 2019 Alzheimer’s disease facts and figures. Alzheimers Dement 15, 321–387.

23.

Jansen

, Ossenkoppele

, Knol

, Tijms

, Scheltens

, Verhey

FRJ

, Visser

, and the Amyloid Biomarker Study Group (2015) Prevalence of cerebral amyloid pathology in persons without dementia: A meta-analysisprevalence of cerebral amyloid pathology in persons without dementiaprevalence of cerebral amyloid pathology in persons without dementia. JAMA 313, 1924–1938.

24.

Hanley

, McNeil

(1983) A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 148, 839–843.

25.

DeLong

, DeLong

, Clarke-Pearson

(1988) Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 44, 837–845.

26.

Mandrekar

(2010) Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol 5, 1315–1316.

27.

Viggars

, Wharton

, Simpson

, Matthews

, Brayne

, Savva

, Garwood

, Drew

, Shaw

, Ince

(2011) Alterations in the blood brain barrier in ageing cerebral cortex in relationship to Alzheimer-type pathology: A study in the MRC-CFAS population neuropathology cohort. Neurosci Lett 505, 25–30.

28.

Montagne

, Barnes

, Sweeney

, Halliday

, Sagare

, Zhao

, Toga

, Jacobs

, Liu

, Amezcua

, Harrington

, Chui

, Law

, Zlokovic

(2015) Blood-brain barrier breakdown in the aging human hippocampus. Neuron 85, 296–302.

29.

Baird

, Westwood

, Lovestone

(2015) Blood-based proteomic biomarkers of Alzheimer’s disease pathology. Front Neurol 6, 236.

30.

, Schindler

, Bollinger

, Ovod

, Mawuenyega

, Weiner

, Leslie

, Masters

, Fowler

, Trojanowski

, Korecka

, Martins

, Janelidze

, Hansson

, Bateman

(2022) Validation of plasma amyloid-β 42/40 for detecting Alzheimer disease amyloid plaques. Neurology 98, e688–e699.

31.

Lopez

, Klunk

, Mathis

, Snitz

, Chang

, Tracy

, Kuller

(2019) Relationship of amyloid-β1–42 in blood and brain amyloid: Ginkgo Evaluation of Memory Study. Brain Commun 2, fcz038.

32.

Janelidze

, Teunissen

, Zetterberg

, Allué

, Sarasa

, Eichenlaub

, Bittner

, Ovod

, Verberk

IMW

, Toba

, Nakamura

, Bateman

, Blennow

, Hansson

(2021) Head-to-head comparison of 8 plasma amyloid-β 42/40 assays in Alzheimer disease. JAMA Neurol 78, 1375–1382.

33.

Palmqvist

, Janelidze

, Stomrud

, Zetterberg

, Karl

, Zink

, Bittner

, Mattsson

, Eichenlaub

, Blennow

, Hansson

(2019) Performance of fully automated plasma assays as screening tests for Alzheimer disease-related beta-amyloid status. JAMA Neurol 76, 1060–1069.

34.

Janelidze

, Mattsson

, Palmqvist

, Smith

, Beach

, Serrano

, Chai

, Proctor

, Eichenlaub

, Zetterberg

, Blennow

, Reiman

, Stomrud

, Dage

, Hansson

(2020) Plasma P-tau181 in Alzheimer’s disease: Relationship to other biomarkers, differential diagnosis, neuropathology and longitudinal progression to Alzheimer’s dementia. Nat Med 26, 379–386.