Skip to main content
PMC home page
Neurology logoLink to Neurology
. 2023 Feb 21;100(8):e874–e883. doi: 10.1212/WNL.0000000000201289

Longitudinal Changes in Blood Biomarkers of Clinical Alzheimer Disease in a Biracial Population Sample

Kumar B Rajan 1,✉, Elizabeth A McAninch 1, Neelum T Aggarwal 1, Lisa L Barnes 1, Robert S Wilson 1, Jennifer Weuve 1, Charles S DeCarli 1, Denis A Evans 1
PMCID: PMC9984218  PMID: 36446595

Abstract

Background and Objectives

Recent studies suggest the utility of blood biomarkers in detecting changes in neurodegenerative disorders. The objective of our research was to test the hypothesis that the longitudinal changes in total tau (t-tau), neurofilament light chain (Nf-L), and glial fibrillary acidic protein (GFAP) are associated with structural MRI and the development of clinical Alzheimer disease (AD) and cognitive decline.

Methods

Data came from a population-based sample with serum concentrations of t-tau, Nf-L, and GFAP and cognitive characteristics measured over 17 years. The inclusion criteria for this investigation were based on participants with blood samples, cognitive function testing, and clinical diagnosis for AD. The longitudinal changes in the serum biomarkers were examined using linear mixed models for log10-transformed concentrations.

Results

In 1,327 participants (60% Black participants and 60% women, the concentration of t-tau increased annually by 4.8% (95% CI = 4.0–5.6) and Nf-L by 5.9% (95% CI = 5.4–6.4). The longitudinal change in GFAP was higher among Black participants than among White participants (4.4% vs 3.5% per year, p = 0.028). Baseline MRI characteristics were associated with the longitudinal changes in serum biomarkers of clinical AD. Specifically, a higher baseline third ventricular volume was associated with a higher rate of increase in the concentration of t-tau, and white matter hyperintensities predicted a higher rate of increase in Nf-L. The rate of change in concentrations of t-tau, Nf-L, and GFAP was significantly higher among those who developed clinical AD than in those with no cognitive impairment. For each standard deviation unit decline in global cognition, longitudinal change in t-tau increased by 81% (95% CI = 76–86), Nf-L by 113% (95% CI = 105–120), and GFAP by 66% (95% CI = 62–70).

Discussion

Blood biomarkers showed significant longitudinal changes corresponding to cognitive decline, clinical AD, and structural MRI characteristics. Our findings show that longitudinal changes in serum biomarkers were associated with several cognitive endophenotypes.

Classification of Evidence

The study found Class II evidence that longitudinal changes in serum t-tau, Nf-L, and GFAP were associated with cognitive decline and the development of clinical AD in people older than 65 years.

Blood-based biomarkers can serve as secondary outcomes in therapeutic trials for Alzheimer disease (AD) and related dementias.1-3 The use of blood biomarkers in clinical trials requires understanding the longitudinal changes in these biomarkers and how they are connected to brain characteristics and cognitive decline, leading to clinical AD. Several studies report the relationship of plasma total tau with cognitive decline4-7 and changes in hippocampal volume and neurofibrillary tangles.8 Several studies also show the relationship of the neurofilament light chain (Nf-L) to the risk of neurodegeneration, leading to cognitive impairment and dementia.9-11 The glial fibrillary acidic protein (GFAP) that has a central role in the glial cell activity and their higher concentrations predict cognitive impairment,12,13 leading to AD.14 We found that serum concentrations of t-tau, Nf-L, and GFAP predicted the development of clinical AD and cognitive decline.15 However, this is the first study that looks at the 3 biomarkers in the same cohort and examines the longitudinal changes in these blood biomarkers that might help us better understand clinical AD pathogenesis and alterations in brain morphology.

To address this research area, we investigated the longitudinal changes in serum blood biomarkers of t-tau, Nf-L, and GFAP over 17 years in a biracial population sample. We tested whether these changes differed between Black and White participants. Second, we tested whether baseline structural MRI characteristics predicted the longitudinal changes in these blood biomarkers. We examined the longitudinal change in these biomarkers among those who progressed to clinical AD and in those with no cognitive impairment. Finally, we examined the rate of change in these biomarkers with the rate of change in global cognition, memory, and perceptual speed score tests.

Methods

Study Participants

The Chicago Health and Aging Project (CHAP) enrolled participants based on a door-to-door census in 4 Chicago neighborhoods with substantial proportions of Black and White residents.16 Inclusion criteria required participants to live in the geographical area, be older than 65 years, and consent to be enrolled in the population study. Approximately one-third of participants were selected for a clinical assessment for AD when they also provided blood samples stored below 80°C freezers. The selections were made evenly between Black and White participant groups. The CHAP study collected blood samples from 5,696 participants with 11,600 blood draws. Of them, 1,327 participants with 3,000 blood samples were selected for participants who provided multiple blood draws and underwent a clinical diagnosis for AD.

Of the 5,696 participants with blood draws, 1,534 underwent clinical evaluations, of whom we selected all 454 participants had 4 or more blood draws and 2 or more cognitive assessments. The remaining 843 participants were randomly selected from 3,148 participants with 3 or fewer blood draws and 2 or more population assessments of cognitive tests. We compared the demographic characteristics of 1,327 participants selected for blood evaluations and 5,696 participants with blood samples. We found they had similar age, race and ethnicity, sex, and the ApoE4 allele frequency.

Standard Protocol Approvals, Registrations, and Patient Consents

The Institutional Review Board of the Rush University Medical Center approved the study protocols, and all participants provided written consent for blood collection, population interviews, and clinical evaluations. Data that support study findings are available through our research data portal17

Quantification of Serum T-Tau, Nf-L, and GFAP

After the blood draws during in-home assessments, study personnel collected and transported serum samples on dry ice to Rush Alzheimer Disease Center Biorepository to be frozen in −80C ultra-low freezers. All serum samples for this analysis were collected and frozen between 1994 and 2012. In mid-2019, we extracted previously unthawed samples and shipped them to Quanterix Corporation (Billerica, MA) by FedEx on dry ice. We assayed 3 neuronal cytoskeletal biomarkers using an ultrasensitive immunoassay performed in duplicates using the bead-based HD platform and the Neurology 4Plex A kit. The coefficient of variation ranged from 7.3% for t-tau and 3.0% for Nf-L and GFAP. We used the mean concentration based on the average of duplicate measurements for all our analyses.

Covariates

Race and ethnicity were assessed using the 1990 US census questions. We asked the following question: “What do you feel is the racial category which best describes yourself?” with the possible responses, “White,” “Black,” “American Indian,” “Eskimo,” “Aleut,” “Asian/Pacific Islander,” or “Refused.” We also asked, “Do you consider yourself to be of Hispanic origin?” with possible responses “Yes,” “No,” or “Refused.” The study had a small fraction (<1%) of Hispanic origin, and participants described themselves as either “Black” or “White.” We also asked participants their date of birth to determine their age, the number of years of formal schooling completed, and sex during birth. We examined the APOE genotype at the Broad Institute using 2 allelic variants: rs7412 and rs429358 using the hME Sequenom MassARRAY platform. Genotyping success rate was 100% for variant rs7412 and 99.8% for variant rs429358. Both SNPs were in Hardy-Weinberg equilibrium, and we created an indicator variable for participants with 1 or more copies of the APOE ε4 allele.

Clinical Diagnosis of AD and Dementia

The clinical diagnosis of AD and dementia followed National Institute of Neurological and Communicative Diseases and Stroke/Alzheimer’s Disease and Related Disorders Association guidelines requiring a history of cognitive decline and impairment in at least 2 cognitive domains.18 To minimize random variability and improve diagnostic decisions across clinicians and time, we developed educationally adjusted cut points on 11 tests19 and used impairment ratings for 5 cognitive domains: orientation, attention, memory, language, and perception.20 A neuropsychologist, who had access to all cognitive data, education, occupation, and ratings of sensorimotor problems and motivation, agreed or disagreed with each cognitive domain rating and supplied a revised rating in the event of disagreement. A board-certified neurologist with access to all clinical data then agreed or disagreed with the diagnosis algorithm and provided a revised diagnosis in case of a dispute. Dementia requires a history of cognitive decline and evidence of impairment in at least 2 cognitive domains, one of which must be memory to meet the criteria for AD.18

Global and Individual Tests of Cognition

A brief battery of 4 cognitive tests was administered to participants during in-home population assessments. We derived a global measure of cognitive function by averaging 4 tests—2 tests of episodic memory, 1 test of executive function, and the mini-mental state examination after centering and scaling each to the baseline mean and standard deviation of the original cohort.21 Individual cognitive tests were based on standardized tests of memory scores, executive function-based speed scores, and the mini-mental state examination.

Structural MRI Evaluations

All participants were scanned at the High-Tech Imaging Center, Palos Heights, IL using the same General Electric 1.5T scanner. Three imaging sequences used fluid-attenuated inversion recovery, spoiled gradient recalled acquisition with an echo-time minimum, and double-spin echo with a repetition time of 2,100 ms.22 The scans were digitally transferred to UC Davis for processing and analysis. In brief, nonbrain tissue was removed using an atlas-based method and nonlinearly registered by a cubic B-spline deformation template adapted for those older than 60 years.23 The B-spline deformation was modeled using a spatially smooth thin-plate spline interpolation. Structural MRI images were processed to remove the skull using an atlas-based method.24,25 Gray, white, and CSF tissue segmentation algorithm was designed to enhance accuracy at likely tissue boundaries after correcting image intensity inhomogeneities. A multiatlas hippocampal segmentation algorithm computed 25 hippocampal and ventricular volumes.25 A registration-based method used in the study conducted by Das et al.27 consisted of an initial probabilistic segmentation of gray matter, white matter, and CSF after intensity inhomogeneity correction using segmentation methods.28 For each boundary voxel on the GM/WM boundary, the thickness was calculated as the distance moved under the registration transformation and propagated across the GM mask.

Statistical Analysis

The mean values, SDs, and frequencies describe the baseline demographic and cognitive characteristics, such as age, the number of formal years of education completed, self-reported race and ethnicity, sex of the participant, and 4 neuropsychological test scores. The distribution of t-tau, Nf-L, and GFAP were positively skewed, and these biomarkers were log10 transformed for longitudinal regression analysis. All regression models adjusted for age at first blood assay (centered at 80 years), education (centered at 12 years), female sex, Black race, chronic health conditions, and the APOE ε4 alleles.

The first analyses modeled the longitudinal person-specific changes in the blood serum biomarkers using a linear mixed-effects regression model with person-specific intercept and slope.29 The time since baseline blood assessment, in years, captured the annual rate of change in log10-transformed blood biomarkers over time. We used interactions of the time since baseline measurement with an indicator for Black race, female sex, the APOE ε4 alleles, and education to test whether the annual change in log10-transformed blood biomarkers was different by these characteristics. We also included each baseline MRI characteristic—total brain volume, white matter hyperintensities, hippocampal, lateral, and third ventricular volumes, and the interaction with the time variable used to test the association of these characteristics with the change in serum biomarkers. In the third set of analyses, we stratified participants by clinical AD status, participants who developed clinical AD, and participants with no cognitive impairment. Using a linear mixed-effects model, we examined the rate of change in log10-transformed blood biomarkers over time. Finally, we estimated the demographic-adjusted person-specific rate of change in cognitive function and the rate of change in the log10-transformed regression model. A regression model with a person-specific rate of change as the predictor and the person-specific rate of change in blood biomarkers was used to examine the relationship of cognitive decline with change in blood biomarkers. All linear mixed-effects regression models were performed using the nlme library, and graphical representations were performed with the R program.30

Data Availability

Data used for publication within the article are available upon request for replication from the study data portal riha.rush.edu/dataportal.html.

Results

The study sample consisted of 1,327 participants with 3,000 blood draws for a median of 3 draws and an average follow-up of 6.0 years (maximum 17 years). The average age of the participants at the first blood draw was 78.7 (SD = 6.3) years, with average education of 12.5 (SD = 3.6) years (Table 1). The study sample consisted of 61% Black participants and 62% women. At the first blood draw, the average total brain volume was 853 mL. The average white matter hyperintensity (WMH) was 7.29 cc and did not differ between Black and White participants (p = 0.25). However, after adjusting for total brain volume and demographic characteristics, the hippocampal volume was 0.23 cc (95% CI = 0.10, 0.35) lower among Black participants than White participants.

Table 1.

Baseline Characteristics of 1,327 Participants by Race and Ethnicity

graphic file with name WNL-2022-201188t1.jpg

Open in a new tab

The baseline demographic characteristics of participants selected for blood biomarkers (N = 1,327) did not differ significantly from the entire cohort (N = 10,802) and those who provided blood samples (N = 5,696). The global cognitive function test scores were significantly higher among White participants than among Black participants (Table 1). However, we found no significant differences in the rate of cognitive decline between White and Black participants. In our sample, Black participants have a higher prevalence of clinical AD. However, the incidence rates were similar between White and Black participants. The follow-up time was longer in White participants by approximately 1 year compared with that in Black participants, and the number of blood draws was similar between White and Black participants.

The log10-transformed geometric mean (95% CI) and raw mean concentrations of serum biomarkers are also summarized in Table 1. The baseline concentration of t-tau did not differ between Black and White participants. However, Nf-L was significantly lower by 13.7% (95% CI = 7.5–19.5) among Black participants than among White participants, which reduced to 6.8% (95% CI = 0.4–12.8) after adjusting for age. The concentration of GFAP among Black participants was also lower by 9.3% (95% CI = 3.5–14.7) compared with that among White participants. However, this difference reduced to 1.2% (95% CI = −3.8 to 7.3) after adjusting for age and was no longer statistically different between Black and White participants (p = 0.51).

Longitudinal Change in Blood Biomarkers Over Time

After adjusting for age, sex, race and ethnicity, education, and the presence of the APOE ε4 alleles, the longitudinal change in concentration of serum t-tau was 0.020 pg/mL (95% CI = 0.017, 0.023) per year or approximately 4.8% (95% CI = 4.0–5.6) increase per year (Table 2). The longitudinal change in concentration of serum Nf-L was 0.025 pg/mL (95% CI = 0.022–0.027) per year or approximately 5.9% (95% CI = 5.4–6.4) increase per year. Finally, the longitudinal change in concentration of GFAP was 0.017 pg/mL (95% CI = 0.015–0.019) per year or approximately 4.0% (95% CI = 3.6–4.4) increase per year. These findings show that the 3 blood biomarkers exhibit varying annual increase in serum concentrations.

Table 2.

Annual Rate of Percentage Change in Serum Concentration of Total Tau, Neurofilament Light Chain, and Glial Fibrillary Acidic Protein in 1,327 Participants From a Biracial Population Sample

graphic file with name WNL-2022-201188t2.jpg

Open in a new tab

Regarding racial differences, the longitudinal change in the concentrations of t-tau (p = 0.49) and Nf-L (p = 0.22) was not different between Black and White participants (Table 2). However, the annual increase in GFAP (4.4% vs 3.5%, p = 0.028) was higher among Black participants than among White participants. These findings suggest that the changes in astroglial cell neurons might longitudinally vary between Black and White participants. We found no significant association of longitudinal changes in the 3 biomarkers with demographic and genetic risk factors, such as age, sex, education, and the APOE ε4 alleles.

Baseline Structural MRI Characteristics and Change in Blood Biomarkers

In a sample of 435 participants, several baseline structural MRI characteristics predicted longitudinal changes in blood biomarkers (Table 3). Participants with higher third ventricular volume had a higher rate of longitudinal change in the concentration of t-tau. Specifically, for each mL with higher third ventricular volume, the annual increase in t-tau was higher by 0.015 pg/mL per year or 3.6% (95% CI = 0.8–6.5) increase in the concentration of t-tau per year. We also found an association of baseline WMH with a higher increase in Nf-L concentration. For every 5 mL higher WMH, the rate of longitudinal change in NF-L was higher by 0.010 pg/mL per year, 2.4% (95% CI = 1.2–3.6) increase in longitudinal change in Nf-L per year. Our analysis did not find a significant association of structural MRI characteristics with the change in GFAP over time.

Table 3.

Annual Rate of Percentage Change in Blood Concentrations of Total Tau (T-Tau), Neurofilament Light Chain (Nf-L), and Glial Fibrillary Acidic Protein (GFAP) and Neuroimaging Biomarkers in 742 Participants

graphic file with name WNL-2022-201188t3.jpg

Open in a new tab

Examining change in Nf-L further by quartiles of WMH, participants with WMH above 19 mL (fourth quartile) had their annual Nf-L increase by 7.3% (95% CI = 5.8–8.9) (Figure 1). By contrast, those below 3 mL (first quartile) increased 3.8% (95% CI = 2.4–5.2) per year. Participants with WMH between 3.0 and 7.8 mL (second quartile) had their Nf-L increase by 6.6% (p = 0.003) and those with WMH between 7.8 and 19 mL (third quartile) had their Nf-L increase by 5.0% (p = 0.019), both of which were significantly higher than the longitudinal change in participants with WMH below 3 mL and combined into one group, as shown in Figure 1.

Figure 1. Association of Quartiles of Baseline White Matter Hyperintensities (WMH) With Longitudinal Change in Neurofilament Light Chain (Nf-L).

Figure 1

Open in a new tab

Association of WMH with longitudinal changes in NfL (A), total tau (t‐tau) (B), and glial fibrillary acidic protein (GFAP) (C). The second and third quartiles of WMH were combined into one because their associations were similar and primarily overlapped. The association of baseline WMH with the annual rate of change in serum concentration of biomarker was adjusted for age, female sex, Black race, education, and the ApoE4 allele.

Longitudinal Change in Blood Biomarkers and Clinical AD and Dementia

Figure 2 shows the longitudinal changes in blood biomarkers for those who developed clinical AD and those with no cognitive impairment. The baseline t-tau was higher by 32%, and the longitudinal change was higher by 8% per year among those with clinical AD compared with those with no cognitive impairment. The differences and changes in Nf-L and GFAP for participants who developed clinical AD were strikingly higher than the differences for those with no cognitive impairment. The baseline Nf-L was higher by 62%, and the longitudinal change was higher by 88% per year among those who developed clinical AD compared with those with no cognitive impairment. Regarding GFAP, the baseline concentration was higher by 45%. The longitudinal change was higher by 80% per year among those who developed clinical AD than among those with no cognitive impairment. The average increases in Nf-L and GFAP were much higher than the average increases seen in t-tau. The rate of change in longitudinal blood biomarkers among those who developed clinical AD did not change when adjusting for baseline WMH characteristics.

Figure 2. Predicted 10-Year Change in Concentrations of Total Tau (T‐Tau), Neurofilament Light Chain (NfL), and Glial Fibrillary Acidic Protein (GFAP) Among Those Diagnosed With Clinical AD and No Cognitive Impairment in a Biracial Population Sample.

Figure 2

Open in a new tab

Longitudinal change in t‐tau (A), NfL (B), and GFAP (C) among those who developed clinical AD and those who did not develop cognitive impairment. The models adjusted for age, female sex, Black race, education, and the ApoE4 allele with log10-transformed serum biomarkers concentrations as outcome measures.

We found similar increases in the baseline differences and longitudinal changes in serum biomarkers with dementia. The baseline concentrations were higher among those who developed dementia—35% for t-tau, 64% for Nf-L, and 48% for GFAP. The longitudinal change was higher by 9% per year for t-tau, 90% per year for Nf-L, and 83% per year for GFAP (data not shown).

Longitudinal Change in Blood Biomarkers and Cognitive Decline

Longitudinal changes in blood biomarkers were significantly associated with the longitudinal cognitive decline after adjusting for demographic characteristics and the APOE ε4 alleles (Figure 3). For each SD unit decline in global cognition, longitudinal change in t-tau increased by 81% (95% CI = 76, 86), Nf-L by 113% (95% CI = 105–120), and GFAP by 66% (95% CI = 62–70) (Table 4). The association of longitudinal change in blood biomarkers and corresponding longitudinal decline in memory test scores was also substantial, ranging from 96% (95% CI = 92–99) for GFAP to 173% (95% CI = 159–189) for Nf-L. We also found similar associations between longitudinal changes in blood biomarkers and longitudinal decline in speed test scores. These findings show that the longitudinal change in blood biomarkers corresponds to the longitudinal change in cognitive function. We observed that changes in Nf-L and GFAP showed a higher association with cognitive decline than t-tau.

Figure 3. Association of Percentage Increase in Concentration of Total Tau (T‐Tau), Neurofilament Light Chain (NfL), and Glial Fibrillary Acidic Protein (GFAP) With Change in Standardized Units of Global Cognition, Episodic Memory, and Executive Functioning in a Biracial Population Sample.

Figure 3

Open in a new tab

Percent change in the neurological biomarkers for change in SD unit for global cognition (A), episodic memory (B) and executive function (C). The relationship of change in serum concentrations of total tau, neurofilament light chain, and glial fibrillary acidic protein with cognitive function was adjusted for age, female sex, Black race, education, and the ApoE4 allele.

Table 4.

Percentage Increase in Serum Concentration of Total Tau (T‐Tau), Neurofilament Light Chain (NfL), and Glial Fibrillary Acidic Protein (GFAP) Corresponding to 1 SD Unit Decline in Global and Individual Tests of Cognition in 1,327 Participants

graphic file with name WNL-2022-201188t4.jpg

Open in a new tab

Classification of Evidence

The primary research question examines the longitudinal change in serum biomarkers of t-tau, Nf-L, and GFAP and their association with cognitive decline and the development of clinical AD in people older than 65 years. The study found Class II evidence that longitudinal changes in serum t-tau, Nf-L, and GFAP were associated with cognitive decline and the development of clinical AD in people older than 65 years.

Discussion

Our study builds on previous findings that serum t-tau, Nf-L, and GFAP predict the development of clinical AD and dementia and cognitive decline in the population sample.15 Our study findings were significant and showed that the longitudinal change in blood biomarkers was substantial. They were associated with baseline structural MRI, development of clinical AD and dementia, and longitudinal changes in global and individual cognition tests, typical during preclinical and prodromal phases of development of clinical AD. These findings are significant because the longitudinal changes in blood biomarkers in biracial studies are yet to be published. A recent study on the relationship of change in serum Nf-L with familial AD9 suggests changes in blood biomarkers can be substantial and provide valuable insights into disease development. Additional findings suggest that the rate of change in Nf-L in CSF may not be static, and these associations may attenuate with worse disease severity.31

The lack of association of the ApoE4 allele and demographic characteristics with longitudinal blood biomarkers changes was noteworthy. This finding suggests that longitudinal changes might not be directly associated with demographic or genetic risk factors as hypothesized but more tightly with the preclinical and prodromal changes of the disease as manifested by brain structure and cognitive functioning changes. The lack of racial differences in longitudinal changes in t-tau and NF-L suggest that the disease progression might not differ between Black and White patients. However, the underlying pathologies of brain structural change and cognitive decline may differ.32,33 Of significance were the findings that ventricular volumes increase the longitudinal change of t-tau. By contrast, higher WMH increases the longitudinal change in Nf-L, suggesting the role of mixed and vascular pathology on the longitudinal change in Nf-L.

The higher rate of longitudinal change in GFAP among Black participants was concerning because these proteins provide astroglia cells that support and nourish cells in the brain and spinal cord and play a significant role in learning and cognition.34 While pathologic studies show a relationship between astroglia expression and degenerative or vascular diseases,35,36 the relationship of these glial proteins measured in the blood to differences between Black and White participants was novel, perhaps suggesting a higher amount of progressive atrophy in Black patients related to the astroglia cells. These findings provide additional secondary outcomes to monitor disease activity in treatment trials or may be the source of disease modification.

Serum Nf-L also was higher in individuals with silent cerebral vascular lesions and WMH severity.37,38 We showed a similar relationship between cerebral vascular lesions characterized by white matter injuries and a significantly faster longitudinal rate of change in Nf-L concentration. However, WMH did not influence t-tau or GFAP, suggesting that the pathobiological changes in serum Nf-L may be more specific to cerebrovascular disease. Alternatively, Nf-L may be a sensitive marker of structural brain changes detected with even subtle brain changes due to WMH. Enlargement of the third ventricular volume, which reflects an actual loss of neuronal tissue, results in higher concentrations of t-tau over time. Differences in these structural MRI characteristics to longitudinal change in blood biomarkers were potentially significant because they suggest differences in pathologic mechanisms. These blood biomarkers may become elevated over time.

Our study limitations include restriction to 3 cytoskeletal biomarkers and a lack of Hispanic, Asian, and American Indian participants and neuropathologic material. We assayed the blood biomarkers before the availability of various phospho-tau assays. However, we are in the process of assessing p-tau biomarkers in our sample. Therefore, they were not specific to the AD process and might be less informative on the disease pathogenesis. The inclusion of ultrasensitive assays of other neurodegeneration markers, especially ones representing noncytoskeletal elements, may have provided further clues to support or deny their potential as markers of AD pathogenesis. Autopsy material availability may have provided more pathogenetic relevance to the longitudinal change in blood biomarkers.

Nonetheless, the inability to look at whether the factors that lead to the systematic disadvantage of Black patients contribute to or account for differences in disease biomarkers over time is also a potential limitation. Although having more blood draws was a strength, it also means that we did not include participants who died earlier or lost to follow-up. In our sample, 309 (23%) had 1 blood evaluation, 583 (44%) had 2 blood evaluations, 264 (20%) had 3 blood evaluations, and 171 (13%) had 4 or more blood evaluations. Hence, the sample has a combination of low and high blood counts.

The preclinical course of AD may be longer than has been demonstrated,39 and a middle life or earlier study may provide a better understanding of the pathogenesis of the disease, especially in those with MCI. Another fluid biomarker development perspective limitation is the lack of a replication cohort. The external validity is limited to heterogenous populations with cognitive impairment not supported by AD biomarkers, becoming less relevant for clinical trials. The Nf-L and tau could be differentially sensitive to the same neurodegenerative process.40 The region of interest-driven method to assess the relationships between the fluid biomarker change and volume change has limitations. Interrogating this relationship through parametric statistical mapping (e.g., voxel‐based morphometry analysis) could yield a more complete and possibly different picture. Another limitation is that it is hard to interpret these differences due to the lack of brain autopsies when the population is potentially heterogeneous. The study population with cognitive impairment defined by the National Institute of Neurological Disorders and Stroke/ADRDA criteria can be called amnestic multidomain dementia rather than clinical AD.

This study addresses several questions related to the neurodegenerative process of late-life cognitive decline and clinical AD. Specifically, we show that convenient, cost-efficient measures of blood biomarkers showed increases in their concentrations over time and were associated with longitudinal changes in cognition observed during the same observational period. We also showed that the neuroimaging biomarkers of structural MRI predicted the longitudinal change in these blood biomarkers, adding further evidence to the clinical significance of longitudinal assessment of blood biomarkers. Our findings suggest that serum biomarkers provide relational associations with the cognitive and brain health characteristics during the development of clinical AD in older at-risk populations.

Glossary

AD

Alzheimer disease

GFAP

glial fibrillary acidic protein

Nf-L

neurofilament light chain

t-tau

total tau

WMH

white matter hyperintensities

Appendix. Authors

Appendix.

Open in a new tab

Footnotes

Class of Evidence: NPub.org/coe

Study Funding

This study was funded by NIH grants R01AG051635, RF1AG057532, R01AG058679, and R01AG073627.

Disclosure

The authors report no relevant disclosures. Go to Neurology.org/N for full disclosures.

References

  • 1.Zetterberg H, Burnham SC. Blood-based molecular biomarkers for Alzheimer's disease. Mol Brain. 2019;12(1):26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Hampel H, Vergallo A, Afshar M, et al. Blood-based systems biology biomarkers for next-generation clinical trials in Alzheimer's disease. Dialogues Clin Neurosci. 2019;21(2):177-191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Li D, Mielke MM, Bell WR, et al. Blood biomarkers as surrogate endpoints of treatment responses to aerobic exercise and cognitive training (ACT) in amnestic mild cognitive impairment: the blood biomarkers study protocol of a randomized controlled trial (the ACT Trial). Trials. 2020;21(1):19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Peltz CB, Kenney K, Gill J, Diaz-Arrastia R, Gardner RC, Yaffe K. Blood biomarkers of traumatic brain injury and cognitive impairment in older veterans. Neurology. 2020;95(9):e1126-e1133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Mattsson N, Zetterberg H, Janelidze S, et al. Plasma tau in Alzheimer disease. Neurology. 2016;87(17):1827-1835. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Zetterberg H. Blood-based biomarkers for Alzheimer's disease—an update. J Neurosci Methods. 2019;319:2-6. [DOI] [PubMed] [Google Scholar]
  • 7.Mielke MM, Hagen CE, Wennberg AMV, et al. Association of plasma total tau level with cognitive decline and risk of mild cognitive impairment or dementia in the mayo clinic study on aging. JAMA Neurol. 2017;74(9):1073-1080. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Pase MP, Beiser AS, Himali JJ, et al. Assessment of plasma total tau level as a predictive biomarker for dementia and related endophenotypes. JAMA Neurol. 2019;76(5):598-606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Preische O, Schultz SA, Apel A, et al. Serum neurofilament dynamics predicts neurodegeneration and clinical progression in presymptomatic Alzheimer's disease. Nat Med. 2019;25(2):277-283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.de Wolf F, Ghanbari M, Licher S, et al. Plasma tau, neurofilament light chain and amyloid-β levels and risk of dementia; a population-based cohort study. Brain. 2020;143(4):1220-1232. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Ma W, Zhang J, Xu J, Feng D, Wang X, Zhang F. Elevated levels of serum neurofilament light chain associated with cognitive impairment in vascular dementia. Dis Markers. 2020:6612871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Oeckl P, Halbgebauer S, Anderl-Straub S, et al. Glial fibrillary acidic protein in serum is increased in Alzheimer's disease and correlates with cognitive impairment. J Alzheimers Dis. 2019;67(2):481-488. [DOI] [PubMed] [Google Scholar]
  • 13.Elahi FM, Casaletto KB, La Joie R, et al. Plasma biomarkers of astrocytic and neuronal dysfunction in early- and late-onset Alzheimer's disease. Alzheimers Dement. 2020;16(4):681-695. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Asken BM, Elahi FM, La Joie R, et al. Plasma glial fibrillary acidic protein levels differ along the spectra of amyloid burden and clinical disease stage. J Alzheimers Dis. 2020;78(1):265-276. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Rajan KB, Aggarwal NT, McAninch EA, et al. Remote blood biomarkers of longitudinal cognitive outcomes in a population study. Ann Neurol. 2020;88(6):1065-1076. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Rajan KB, Wilson RS, Weuve J, et al. Cognitive impairment 18 years prior to clinical diagnosis of Alzheimer's disease dementia. Neurology. 2015;85(10):898-904. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Data Portal. Rush Institute for Healthy Aging. https://www.riha.rush.edu/dataportal.html
  • 18.McKhann G, Drachman D, Folstein M, et al. Clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group Under the Auspices of Department of Health and Human Services Task Force on Alzheimer's disease. Neurology. 1984;34(7):939-944. [DOI] [PubMed] [Google Scholar]
  • 19.Bennett DA, Wilson RS, Schneider JA, et al. Natural history of mild cognitive impairment in older persons. Neurology. 2002;59(2):198-205. [DOI] [PubMed] [Google Scholar]
  • 20.Bennett DA, Schneider JA, Aggarwal NT, et al. Decision rules guiding the clinical diagnosis of Alzheimer's disease in two community-based cohort studies compared to standard practice in a clinic-based cohort study. Neuroepidemiology. 2006;27(3):169-176. [DOI] [PubMed] [Google Scholar]
  • 21.Wilson RS, Bennett DA, Beckett LA, et al. Cognitive activity in older persons from a geographically defined population. J Gerontol B Psychol Sci Soc Sci. 1999;54(3):155-160. [DOI] [PubMed] [Google Scholar]
  • 22.Aggarwal NT, Wilson RS, Bienias JL, et al. The association of magnetic resonance imaging measures with cognitive function in a biracial population sample. Arch Neurol. 2010;67(4):475-482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Rajapakse JC, Giedd JN, DeCarli C, et al. A technique for single-channel MR brain tissue segmentation: application to a pediatric sample. Magn Reson Imaging. 1996;14(9):1053-1065. [DOI] [PubMed] [Google Scholar]
  • 24.Lee DY, Fletcher E, Martinez O, et al. Vascular and degenerative processes differentially affect regional interhemispheric connections in normal aging, mild cognitive impairment, and Alzheimer disease. Stroke. 2010;41:1791-1797. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Aljabar P, Heckemann RA, Hammers A, et al. Multi-atlas based segmentation of brain images: atlas selection and its effect on accuracy. Neuroimage. 2009;46(3):726-738. [DOI] [PubMed] [Google Scholar]
  • 26.Fletcher E, Carmichael O, DeCarli C. MRI non-uniformity correction through interleaved bias estimation and B-spline deformation with a template. Conf Proc Annu Int Conf IEEE Eng Med Biol Soc Conf. 2012;2012:106-109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Das SR, Avants BB, Grossman M, Gee JC. Registration based cortical thickness measurement. Neuroimage. 2009;45(3):867-879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Fletcher E, Singh B, Harvey D, Carmichael O, Decarli C. Adaptive image segmentation for robust measurement of longitudinal brain tissue change. Conf Proc Annu Int Conf IEEE Eng Med Biol Soc IEEE Eng Med Biol Soc Conf. 2012;2012:5319-5322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Diggle PJ, Heagerty P, Liang KY, Zeger SL. Analysis of Longitudinal Data. : Oxford University Press; 2002. [Google Scholar]
  • 30.R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, , 2012. [Google Scholar]
  • 31.Lleó A, Alcolea D, Martínez-Lage P, et al. Longitudinal cerebrospinal fluid biomarker trajectories along the Alzheimer's disease continuum in the BIOMARKAPD study. Alzheimers Dement. 2019;15:742-753. [DOI] [PubMed] [Google Scholar]
  • 32.Barnes LL, Leurgans S, Aggarwal NT, et al. Mixed pathology is more likely in black than white decedents with Alzheimer dementia. Neurology. 2015;85:528-534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.DeCarli C, Reed BR, Jagust W, Martinez O, Ortega M, Mungas D. Brain behavior relationships among African Americans, whites, and Hispanics. Alzheimer Dis Assoc Disord. 2008;22(4):382-391. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Fields RD, Araque A, Johansen-Berg H, et al. Glial biology in learning and cognition. Neuroscientist. 2014;20(5):426-431. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Fakhoury M. Microglia and Astrocytes in Alzheimer's disease: implications for therapy. Curr Neuropharmacol. 2018;16(5):508-518. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Liu Z, Chopp M. Astrocytes, therapeutic targets for neuroprotection and neurorestoration in ischemic stroke. Prog Neurobiol. 2016;144:103-120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Gattringer T, Pinter D, Enzinger C, et al. Serum neurofilament light is sensitive to active cerebral vascular small vessel disease. Neurology. 2017;89(20):2108-2114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Moore EE, Hohman TJ, Badami FS, et al. Neurofilament relates to white matter microstructure in older adults. Neurobiol Aging. 2018;70:233-241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Jack CR Jr, Knopman DS, Jagust WJ, et al. Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade. Lancet Neurol. 2010;9(1):119-128. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Illán-Gala I, Lleo A, Karydas A, et al. Plasma tau and neurofilament light in frontotemporal lobar degeneration and Alzheimer disease. Neurology. 2021;96:e671-e683. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

Data used for publication within the article are available upon request for replication from the study data portal riha.rush.edu/dataportal.html.

Articles from Neurology are provided here courtesy of American Academy of Neurology

RESOURCES