Associations of Sensory and Motor Function with Blood-Based Biomarkers of Neurodegeneration and Alzheimer’s Disease in Midlife

Abstract

Pathological biomarkers of dementia and Alzheimer’s disease (AD) change decades before clinical symptoms. Common sensory and motor changes in aging adults may be early markers of neurodegeneration. We investigated if midlife sensory and motor functions in Beaver Dam Offspring Study (BOSS) participants (N=1529) were associated with longitudinal changes in blood-based biomarkers of neurodegeneration (neurofilament light chain (NfL); total tau (TTau)) and AD (amyloid beta (Aβ)). Mixed-effects models with baseline sensory and motor function as determinants and 10-year biomarker change as outcome were used. Participants with hearing impairment and worse motor function (among women) showed faster increases in NfL level over time (0.8%/year; 0.3%/year, respectively). There were no significant associations with TTau or Aβ.

We found consistent relationships between worse baseline hearing and motor function with a faster increase in neurodegeneration, specifically serum NfL level. Future studies with longer follow-up should determine if sensory and motor changes are more reflective of general neurodegeneration than AD-specific pathology and whether sensory and motor tests may be useful screening tools for neurodegeneration risk.

1. INTRODUCTION

From 1990 to 2016, the number of people world-wide living with all-cause dementia more than doubled from 20.2 million to 43.8 million, inflicting a huge burden on patients, caregivers, economies, and health-care systems, and is expected to further increase to 131.5 million by 2050 (GBD 2016; Prince et al, 2015). Age-related pathophysiological changes in the brain begin early in midlife and the preclinical phase of neurodegenerative conditions including Alzheimer’s disease (AD) may begin decades before clinical diagnosis (Fotenos et al, 2005; Sperling et al, 2011). The early detection of high-risk individuals in midlife has great potential for future targeted treatments to prevent or delay dementia and AD onset.

Epidemiological studies of sensory impairments in older adults have shown that hearing, vision, and olfaction impairments are associated with increased risk for the development of cognitive impairment, dementia, or AD, (Albers et al, 2015; Brenowitz et al, 2019; Fischer et al, 2016). Similarly, motor dysfunction has been associated with the risk of developing cognitive impairment or AD (Albers et al, 2015; Buchman et al, 2007; Camargo et al, 2016). The temporality of this association is not fully understood. Sensory and motor functions are highly integrated with the central nervous system (CNS) and share many risk factors of neurodegeneration with cognitive decline (Albers et al, 2015; Carson 2018; Schubert et al, 2019a; Whitson et al, 2018). Declines in sensory and motor function may indicate damage or degeneration within neural pathways or associated central processing areas and changes in these sensory and motor measures may be indicators of brain changes due to aging or disease (Albers et al, 2015; Carson 2018, Schubert et al, 2019; Whitson et al, 2018). Changes in sensory and motor function begin to develop in midlife with increasing incidence of impairments in sensory and motor function with increasing age (Dalton et al, 2020; Paulsen et al, 2018; Schubert et al, 2021; van der Willik et al, 2021). Since changes in sensory and motor function can be measured reliably, cost-effectively, and may be detectable at earlier timepoints, sensory and motor impairments could potentially serve as early markers for neurodegeneration.

The biomarker-based research framework developed by the National Institute on Aging and Alzheimer’s Association suggests classifying neuropathologic changes in the brain as AD pathologic change or non-AD pathologic change based on neuroimaging, or using biomarker levels of amyloid-β, phosphorylated tau, and total tau measured in cerebrospinal fluid (CSF) (Jack et al, 2018). Another emerging biomarker of general neurodegeneration is neurofilament light chain (NfL), a primary protein of the axonal cytoskeleton that increases in blood and CSF in conditions that cause axonal damage (Gaetani et al, 2019). Ultrasensitive assays using single molecule array (SIMOA) technology have been developed that can reliably measure concentrations of biomarkers related to neurodegeneration and AD, including NfL, amyloid-β₄₀ and amyloid-β₄₂ (Aβ₄₀, Aβ₄₂), and total tau (TTau), in blood (Blennow & Zetterberg, 2019). Data are limited on the distribution of these biomarkers in adults in the general population and it is not known if sensory and motor functions are associated with changes in midlife blood-based markers of neuronal injury, neurodegeneration, and AD.

The purpose of this study was to characterize the 10-year change in serum biomarkers of neurodegeneration and AD-related pathologic change in a general population cohort of middle-aged and older adults and to determine if sensory and motor function are associated with 10-year change. As changes in sensory or motor function may be detectable at earlier timepoints, are easily obtained, and relatively cost-effective to measure, this study aims to characterize the potential utility of these measures to identify those at higher risk for accumulation of biomarkers of neurodegeneration in the blood as indicators of CNS pathology. Further, we aimed to determine if these associations are independent of known risk factors for neurodegeneration, including behavioral factors, vascular disease, inflammation, metabolic dysregulation, and neurotoxin exposure.

2. METHODS

2.1. Study Design and Participants

The Beaver Dam Offspring Study (BOSS) is a longitudinal cohort study of the adult children of participants in the population-based Epidemiology of Hearing Loss Study (Cruickshanks et al, 1998). Baseline BOSS examinations took place in 2005–2008 with follow-up examinations at 5 (2010–2013) and 10 (2015–2017) years (Dalton et al, 2020; Nash et al, 2011; Paulsen et al, 2018; Schubert et al, 2021). Examinations included assessments of sensory and motor function, a blood draw, measures of vascular health, questionnaires on demographics and behavioral and medical history. The study was approved by the Health Sciences Institutional Review Board of the University of Wisconsin; written informed consent was obtained from all participants prior to each examination.

2.2. Measurement of NfL, Aβ₄₀, Aβ₄₂, and TTau

Blood collection and processing protocols were similar across phases. (Supplement 1).

Concentrations of NfL, Aβ₄₀, Aβ₄₂, and TTau, were measured in serum samples collected at the baseline, 5- and 10-year examinations and stored at −80° C until assayed at the Quanterix Simoa Accelerator Laboratory (Billerica, MA, USA) in 2021 (Schubert et al, 2022). The Simoa® NF-Light™ Advantage kit was used to assay samples for NfL concentration (Quanterix – NfL). The Simoa® Neurology 3-PlexA Kit was used to simultaneously assay the samples for Aβ₄₀, Aβ₄₂ and TTau concentrations (Quanterix 3-plex). Both assays were completed during a single thaw cycle. Participant and quality control samples were assayed in duplicate and the same lot of kits and reagents for each platform were used for all assays with 86 samples, 2 controls, and 8 calibrators per run. The average (range) intra-assay coefficients of variation (CV) for the controls were: NfL 4.7% (0.04–20.0%), Aβ₄₀: 5.1% (0.02–15.7%); Aβ₄₂: 4.2% (0.09–15.3%); TTau 4.7% (0.01–18.2%).

Two sub-studies were conducted to provide support for the validity of the NfL, Aβ42/Aβ40 ratio, and TTau concentrations in serum and as potential markers of pathology. To determine the correlation between serum and plasma biomarker levels, NfL, Aβ₄₀, Aβ₄₂, and TTau were also measured in plasma samples for a subset (N=34) of the included BOSS participants (Supplement 1). We also sent samples for assay of these biomarkers from a subset of participants (N=122) in the Wisconsin Registry for Alzheimer’s Prevention (WRAP) and the Wisconsin Alzheimer’s Disease Research Center (ADRC) cohorts, two longitudinal studies on AD, who ranged from having normal cognition to those with dementia, and had corresponding biomarker measures in CSF and/or classification of amyloid pathology by Pittsburgh compound B carbon 11-labeled positron emission tomography (¹¹C-PiB PET) (Supplement 1) (Johnson et al, 2014, Racine et al, 2014). Additional imaging data, such as tau PET imaging, was not available for the current study.

2.3. Sensory and Motor Function

Sensory and motor function was measured by trained examiners using standardized protocols. Hearing was measured by pure tone audiometry and impairment was defined as a pure-tone average of the thresholds at 0.5, 1, 2, and 4 kHz greater than 25 decibels Hearing Level in either ear (Dalton et al, 2020). Either ear was chosen for this study to characterize the earliest onset of pathology rather than participants’ daily hearing function. Contrast sensitivity was measured using Pelli-Robson letter charts and visual impairment was defined as contrast sensitivity < 1.55 log units in the worse eye (Paulsen et al, 2018). We measured olfaction using the San Diego Odor Identification Test and impairment was defined as identifying fewer than 6 of 8 odorants correctly (Schubert et al, 2021).

Motor function measures included the time (seconds) to complete the Grooved Pegboard (GPB; Lafayette Instruments, Lafayette, IN, USA), a test of psychomotor function and manual dexterity (Strauss et al, 2006; Zhong et al, 2011); grip strength (kilograms), measured with a hand dynamometer (model 78010, Lafayette Instruments, Lafayette, IN, USA) (Strauss et al, 2006); and the physical function scale (PFS) of the Medical Outcomes Study Short Form Health Survey, a self- reported measure of difficulties in locomotion and endurance (Ware et al, 1993) (Supplement 1).

2.4. Other Variables

We assessed baseline covariates as potential confounders: age, sex, education, diabetes, hypertension, carotid artery plaque, current smoking status, exercise, weekly alcohol consumption in the past year, waist circumference, blood lead and cadmium, serum high sensitivity C-reactive protein (hsCRP), interleukin-6 (IL-6), soluble intercellular adhesion molecule-1 (sICAM-1), and soluble vascular cell adhesion molecule-1 (sVCAM-1) (Dalton et al, 2020) (Supplement 1).

2.5. Statistical Analyses

All statistical analyses were conducted with SAS 9.4 (SAS Institute, Inc., Cary, NC USA). Assays of samples for biomarker levels were completed in duplicate and the average of the duplicates was the biomarker concentration. Samples with a concentration below the limit of detection (LOD) were assigned a value halfway between zero and the LOD for that biomarker (Aβ₄₂ set to 0.09 pg/ml (N=9); TTau set to 0.04 pg/ml (N=109); no samples below LOD for NfL and Aβ₄₀). The Aβ₄₂/Aβ₄₀ ratio was calculated by dividing the concentration of Aβ₄₂ by the concentration of Aβ₄₀.

Descriptive statistics were used to summarize the baseline distribution and 10-year change of serum biomarker levels observed in our cohort. NfL and TTau concentration distributions were skewed and therefore natural log transformed in statistical models to comply with normality assumptions for linear models.

To determine the effect of baseline sensory impairment (hearing, vision, olfaction; impairment yes versus no) and motor function (GPB, grip strength, PFS; per standard deviation (SD)) on change in biomarker levels, we used linear mixed-effect models (LMEM) with baseline sensory and motor measures and repeated measures of biomarkers as the outcome (Supplement 2). The Aβ₄₂/Aβ₄₀ ratio was used, as the ratio is more indicative of neuropathology as compared to the individual measures of Aβ₄₀ and Aβ₄₂ (Schindler et al, 2019). The interaction between baseline sensory impairments and motor functions as determinants and time (per year) was the main variable of interest, indicating differences in rate of biomarker change by baseline sensory or motor function. We adjusted models for age, sex, random intercept, and random slope. When analysis resulted in a non-positive definite G matrix of the random effects, indicating poor model fit, we ran models with random participant-specific intercepts only (in models of Aβ₄₂/Aβ₄₀ ratio in men and TTau in women). Based on the literature, we repeated all models additionally adjusting for education, vascular and metabolic risk factors (diabetes, hypertension, carotid plaque, smoking, alcohol consumption, exercise, waist circumference), inflammation (hsCRP, IL-6, sICAM-1, sVCAM-1), and neurotoxin exposure (blood lead and cadmium) (Barnes & Yaffe, 2011; Schubert et al, 2019b; Wichmann et al, 2014; Zhong et al, 2011). Models were also explored with body mass index (BMI) included in place of waist circumference as body characteristics and central adiposity could have important and potentially differing effects on the measurement of blood-based biomarkers. We tested all models for sex interactions; due to a statistically significant sex by biomarker interaction, models for Aβ₄₂/Aβ₄₀ ratio were stratified by sex. Due to known differences in motor function measures by sex, and significant differences in measures of motor function by sex in the BOSS cohort, all models including motor function were also sex stratified (Ashendorf et al, 2009; Botoseneaunu et al, 2016; Syddall et al, 2009). Results from LMEM for the Aβ₄₂/Aβ₄₀ ratio were presented as mean difference in baseline ratio and mean change in the ratio per year, and results from models of log transformed NfL and TTau were exponentiated and are presented as percent difference in baseline concentration and percent change in concentration per year.

2.6. Sensitivity Analyses

Fully adjusted models were repeated excluding participants with biomarker levels reported above the upper limit of quantification to ensure they were not overly influencing findings this removed N=2 for analysis involving Aβ₄₂/Aβ₄₀ ratio and N=1 for TTau.

2.7. Substudy Analyses

Pearson correlations were used to assess the strength of the linear association between biomarkers measured in serum and those in plasma and CSF samples. Mean differences in serum biomarker level by presence or absence of neuropathology, determined by CSF amyloid level and ¹¹C-PiB PET amyloid positivity, were tested with independent group t-tests.

3. RESULTS

3.1. Study population

The current study included 1529 BOSS participants with serum samples available from all three examinations (mean baseline age=49.2 years, SD=9.4, range = 22–84; 54% women, average 9.6 years of follow-up). Participants not eligible for the current study did not have available blood samples from all examinations because they: participated by questionnaire only in at least 1 examination phase (N=837), did not consent to a blood draw at one or more examination phases (N=263), did not participate in at least one follow-up examination (N=567), or died prior to completion of all follow-up examinations (N=102). The participants included in this study were similar in baseline characteristics to the entire baseline BOSS cohort (Table 1).

Table 1:

Baseline characteristics of the Beaver Dam Offspring Study participants with measured biomarkers of neurodegeneration and Alzheimer’s disease.

	Included (N=1529)	Full cohort (N=3298)
Sex, n(%)
Women	828 (54.2)	1802 (54.6)
Men	701 (45.8)	1496 (45.4)
Mean (SD) baseline age, years	49.2 (9.4)	49.2 (10.0)
Education, n(%)
0–12y	464 (30.5)	984 (30.1)
13–15y	534 (35.2)	1098 (33.6)
16+y	521 (34.3)	1188 (36.3)
Hearing Impairment (PTA>25dBHL), n(%)
No	1311 (85.8)	2443 (85.8)
Yes	217 (14.2)	404 (14.2)
Vision impairment (CS log triplets<1.55), n(%)
No	1277 (83.6)	2340 (82.3)
Yes	250 (16.4)	505 (17.7)
Olfactory impairment (SDOIT<6), n(%)
No	1474 (96.5)	2739 (96.2)
Yes	54 (3.5)	109 (3.8)
Mean (SD) Grooved Pegboard Time, s	71.7 (15.6)	72.7 (18.3)
Mean (SD) Grip Strength, kg	38.4 (12.4)	37.9 (12.2)
Mean (SD) SF36 - PFS	88.3 (16.0)	87.8 (17.0)

CS=contrast sensitivity; dBHL= decibels hearing level; PTA= Pure tone average; s=seconds; SD= standard deviation; SDOIT= San Diego Odor Identification Test; SF36-PFS= Short form 36-Physical Functioning Scale; y=years

3.2. Baseline biomarker levels

Mean (SD) baseline serum levels of NfL, Aβ₄₂/Aβ₄₀ ratio, and TTau are reported in Table 2. Adjusted for age, levels of NfL and TTau were similar among women and men (Table 2), while men had significantly lower baseline Aβ₄₂/Aβ₄₀ ratio (β=−0.004, p=0.0005). Adjusted for sex, NfL (β=0.32 pg/mL per year, p<0.0001) was higher with older baseline age, while Aβ₄₂/Aβ₄₀ ratio was lower (β= −0.0003 per year, p<0.0001) (Table 2). There were no sex or age effects for baseline serum TTau.

Table 2:

Baseline mean and 10-year mean change in serum biomarkers of neurodegeneration and Alzheimer’s disease

	NfL (pg/mL)		Aβ42/40 ratio		Total Tau (pg/mL)
	Baseline mean (SD)	p-value2	Baseline mean (SD)	p-value2	Baseline mean (SD)	p-value1
Overall (N=1529)	11.1 (7.6)		0.077 (0.023)		0.71 (7.0)
Sex
Women (N=828)	11.0 (5.6)	0.91	0.079 (0.026)	0.0005	0.54 (0.41)	0.27
Men (N=701)	11.2 (9.4)		0.074 (0.018)		0.93 (10.3)
Baseline age (years)
21–34 (N=76)	7.4 (4.5)	<0.0001	0.085 (0.027)	<0.0001	0.47 (0.24)	0.96
35–44 (N=413)	8.2 (3.5)		0.080 (0.027)		0.53 (0.41)
45–54 (N=613)	10.7 (5.7)		0.077 (0.022)		0.97 (11.0)
55–64 (N=340)	14.0 (11.3)		0.073 (0.018)		0.57 (0.39)
65–84 (N=87)	18.9 (7.9)		0.071 (0.021)		0.56 (0.37)
	10-year mean change (SD)	p-value2	10-year mean change (SD)	p-value2	10-year mean change (SD)	p-value2
Overall (N=1529)	4.6 (11.0)		−0.004 (0.022)		0.03 (0.79)
Sex
Women (N=828)	4.5 (7.8)	0.90	−0.005 (0.024)	0.01	0.01 (0.53)	0.27
Men (N=701)	4.7 (13.9)		−0.002 (0.018)		0.05 (1.02)
Baseline age (years)
21–34 (N=76)	1.5 (4.8)	<0.0001	−0.005 (0.025)	0.18	0.08 (0.40)	0.78
35–44 (N=413)	2.9 (8.1)		−0.005 (0.026		−0.01 (0.47)
45–54 (N=613)	3.9 (7.7)		−0.004 (0.022)		0.06 (1.11)
55–64 (N=340)	6.5 (16.3)		−0.002 (0.016)		0.01 (0.51)
65–84 (N=87)	12.9 (16.1)		−0.003 (0.018)		0.06 (0.37)

Aβ=amyloid beta; NfL=neurofilament light chain; SD=standard deviation

¹Sex effect: adjusted for age; Age effect: adjusted for sex

3.3. 10-year change in biomarker levels

Overall, serum NfL increased, the Aβ₄₂/Aβ₄₀ ratio declined, and TTau had a slight increase during the 10-year follow-up period (Table 2). In age- and sex-adjusted models, serum biomarker levels of NfL and TTau increased on average over the 10-year follow-up (per year: 3.3% p=0.0001; 1.3%, p<0.0001, respectively), and the Aβ₄₂/Aβ₄₀ ratio declined (per year: −0.0004, p<0.0001). Rates of change in serum biomarker levels were similar in women and men except women had a more rapid decline in the Aβ₄₂/Aβ₄₀ ratio over the 10-year period when adjusting for baseline age (−0.0005 per year in women, −0.0002 per year in men, p=0.01). Graphical trajectories of biomarker change by sex are presented in Supplement 3. Adjusting for sex, older age groups had greater increases in NfL (0.26 per year of age, p<0.0001). No significant age effects were observed for 10-year change in Aβ₄₂/Aβ₄₀ ratio or TTau (Table 2).

3.4. Sensory and motor function and 10-year change in serum NfL

Baseline effects and rates of change in NfL by sensory and motor function from linear mixed effect models are displayed in Table 3. Participants with baseline hearing impairment had a more rapid increase in serum NfL over the 10-year follow-up period in both minimally and fully adjusted models than those without hearing impairment (~0.8% faster increase per year, p=0.004; 3.2% versus 4.0% per year in those without and with hearing impairment, respectively) (Figures 1a–1b). No significant relationships were observed between baseline visual or olfactory impairment and change in serum NfL over the 10-year follow-up period.

Table 3:

Differences in baseline levels and rate of change in serum neurofilament light chain (NfL) over 10-years by baseline sensory and motor function.

	NfL, % (95% CI)
	Overall (N=1529)
Hearing Impairment	Model 1		Model 2
Baseline Effect	−0.1 (−5.3, 5.3)		−0.9 (−6.0, 4.6)
No hearing impairment, change per year	3.2 (2.9, 3.4)†		3.2 (2.9, 3.4)†
Hearing impairment, change per year	4.1 (3.3, 4.9)†		4.0 (3.2, 6.4)†
Visual Impairment	Model 1		Model 2
Baseline Effect	−1.7 (−6.4, 3.3)		−2.0 (−6.8, 3.0)
No vision impairment, change per year	3.2 (3.0, 3.4)		3.2 (3.0, 3.4)
Vision impairment, change per year	3.6 (2.9, 4.4)		3.7 (2.9, 4.5)
Olfactory Impairment	Model 1		Model 2
Baseline Effect	−4.3 (−13.2, 5.5)		−6.0 (−15.0, 3.9)
No olfactory impairment, change per year	3.3 (3.1, 3.5)		3.3 (3.1, 3.5)
Olfactory impairment, change per year	4.0 (2.7, 5.3)		4.0 (2.7, 5.4)
	Women (N=828)		Men (N=701)
Grooved Pegboard Test (GPB) Time	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD slower1	0.6 (−2.0, 3.2)	2.1 (−0.5, 4.9)	0.4 (−2.4, 3.3)	0.9 (−2.0, 3.9)
Average time, change per year	3.2 (3.0, 3.5)†	3.2 (3.0, 3.5)†	3.4 (3.1, 3.7)	3.4 (3.0, 3.7)
1 SD above average time, change per year	3.7 (3.2, 4.1)†	3.6 (3.1, 4.1)†	3.7 (3.1, 4.3)	3.6 (3.0, 4.3)
Grip Strength	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD lower2	5.3 (2.5, 8.1)†	3.5 (0.8, 6.4)†	2.7 (−0.1, 5.6)	0.7 (−2.1, 3.6)
Average grip, change per year	3.2 (3.0, 3.5)†	3.3 (3.0, 3.5)†	3.4 (3.1, 3.7)	3.4 (3.0, 3.7)
1 SD below average grip, change per year	3.6 (3.1, 4.1)†	3.6 (3.1, 4.1)†	3.7 (3.0, 4.3)	3.6 (3.0, 4.3)
Physical Functioning Scale (PFS)	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD lower3	−1.3 (−3.7, 1.2)	2.6 (−0.2, 5.5)	1.6 (−1.1, 4.4)	2.3 (−0.6, 5.2)
Average PFS score, change per year	3.2 (3.0, 3.5)†	3.2 (3.0, 3.5)†	3.4 (3.1, 3.7)	3.4 (3.0, 3.7)
1 SD below average PFS score, change per year	3.8 (3.3, 4.3)†	3.9 (3.4, 4.4)†	3.4 (2.8, 4.1)	3.4 (2.7, 4.0)

Model 1: Linear mixed-effects model adjusted for baseline age, sex (in sensory models only), random slope, and random intercept. Change per year by sensory/motor function calculated from regression estimates for time and time*sensory/motor interaction terms.

Model 2: Linear mixed-effects model 1, additionally adjusted for baseline education, diabetes, hypertension, presence of carotid artery plaque, smoking status, weekly alcohol consumption, regular exercise, waist circumference, blood lead level, blood cadmium level, serum high-sensitivity C-reactive protein, serum interleukin-6, natural log of serum soluble intercellular adhesion molecule-1, natural log of serum soluble vascular cell adhesion molecule-1. Change per year by sensory/motor function calculated from regression estimates for time and time*sensory/motor interaction terms.

¹Centered at mean GPB time: Women= 68s, Men=76s; Standard deviation (SD): Women=13.6s, Men=16.6s

²Centered at mean grip strength: Women=29kg, Men=49kg; SD for grip strength; Women=5.9kg, Men=9.3kg

³Centered at mean PFS score: Women=87, Men=90; SD for PFS: Women=16.9, Men=14.7

^†p<0.05 for difference in baseline biomarker level or rate of change in biomarker

Figure 1a:

Serum neurofilament light chain (NfL) trajectory by hearing impairment status (Women)¹

Figure 1b:

Serum neurofilament light chain (NfL) trajectory by hearing impairment status (Men)¹

For all three measures of motor function, individuals with worse motor function had a faster rate of increase in serum NfL concentration but effects were only statistically significant in women. Women with slower performance on the Grooved Pegboard Test, had more rapid increases in serum NfL over time (0.4% faster per year per SD (13.6s) slower test time, p=0.002; 3.2% versus 3.6% per year in those with average GPT time and those 1 SD above average GPT time, respectively) (Figure 2a). The estimate for men was similar (0.2% per year per SD slower time, p=0.07) (Figure 2b). Among women, lower baseline grip strength was associated with higher baseline NfL (3.5% higher NfL per SD (5.9kg) weaker grip; 95%CI=0.8,6.4), and a faster rate of increase in NfL over time (0.3% per year faster per SD (5.9kg) weaker grip strength, p=0.006; 3.3% versus 3.6% per year in those with average grip strength and 1 SD below average grip strength, respectively) (Figure 3a). A similar difference in rate of change by grip strength was observed among men, (0.2% per year per SD (9.3kg) weaker, p=0.07) (Figure 3b). Lower scores on the SF36-PFS were associated with faster rates of increase in NfL over time in women (0.6% per year faster per SD lower PFS score; p=0.006; 3.2% versus 3.9% per year in those with average PFS score and 1 SD below average score, respectively) (Figure 4a); there was no relationship among men (Figure 4b).

Figure 2a:

Serum neurofilament light chain (NfL) trajectory by Grooved Pegboard Test time (Women)¹

Figure 2b:

Serum neurofilament light chain (NfL) trajectory by Grooved Pegboard Test time (Men)¹

Figure 3a:

Serum neurofilament light chain (NfL) trajectory by grip strength (Women)¹

Figure 3b:

Serum neurofilament light chain (NfL) trajectory by grip strength (Men)¹

Figure 4a:

Serum neurofilament light chain (NfL) trajectory by Short form 36-Physical Functioning Scale (Women)¹

Figure 4b:

Serum neurofilament light chain (NfL) trajectory by Short form 36-Physical Functioning Scale (Men)¹

3.5. Sensory and motor function and 10-year change in serum Aβ₄₂/Aβ₄₀ ratio

Baseline effects and rates of change in Aβ₄₂/Aβ₄₀ ratio by sensory and motor function from linear mixed effect models are displayed in Table 4. Women with hearing impairment had no change in the Aβ₄₂/Aβ₄₀ ratio over time (0.0000 per year, 95%CI=−0.0008, 0.0008) while those without impairment declined (−0.0006 per year, 95%CI=−0.0008, −0.0004) in fully adjusted models, although this interaction did not reach statistical significance (p=0.07). A significant difference in rate of decline in the Aβ₄₂/Aβ₄₀ ratio was found by baseline visual impairment among women. Those with visual impairment declined more slowly than those without (0.0005 per year slower, p=0.048; −0.0002 versus −0.0006 in those with and without visual impairment, respectively), although mean baseline Aβ₄₂/Aβ₄₀ ratio was lower among those with visual impairment (−0.004; 95%CI=−0.008, 0.0003), albeit a non-significant difference. Among men, baseline visual impairment was not associated with change in Aβ₄₂/Aβ₄₀ ratio. There was no significant relationship between olfactory impairment and Aβ₄₂/Aβ₄₀ ratio in this study.

Table 4:

Differences in baseline levels and rate of change in serum amyloid beta (Aβ) 42/20 ratio over 10-years by baseline sensory and motor function.

	Aβ42/40 ratio, (95% CI)
	Women (N=828)		Men (N=701)
Hearing Impairment	Model 1	Model 2	Model 1	Model 2
Baseline Effect	−0.003 (−0.009, 0.002)	−0.004 (−0.010, 0.002)	−0.003 (−0.007, 0.0001)	−0.003 (−0.006, 0.0010)
No hearing impairment, change per year	−0.0006 (−0.0008, −0.0004)	−0.0006 (−0.0008, −0.0004)	−0.0003 (−0.0005, −0.0001)	−0.0003 (−0.0005, −0.0001)
Hearing impairment, change per year	−0.0001 (−0.0009, 0.0007)	0.0000 (−0.0008, 0.0008)	−0.0000 (−0.0007, 0.0006)	−0.0001 (−0.0007, 0.0006)
Visual Impairment	Model 1	Model 2	Model 1	Model 2
Baseline Effect	−0.004 (−0.008, 0.0004)	−0.004 (−0.008, 0.0003)	0.002 (−0.003, 0.006)	0.002 (−0.002, 0.006)
No vision impairment, change per year	−0.0006 (−0.0008, −0.0004)	−0.0006 (−0.0008, −0.0004)†	−0.0002 (−0.0004, 0.0000)	−0.0002 (−0.0004, 0.0000)
Vision impairment, change per year	−0.0003 (−0.0009, 0.0004)	−0.0002 (−0.0008, 0.0005)†	−0.0005 (−0.001, 0.0002)	−0.0005 (−0.001, 0.0002)
Olfactory Impairment	Model 1	Model 2	Model 1	Model 2
Baseline Effect	−0.005 (−0.016, 0.006)	−0.006 (−0.019, 0.006)	−0.004 (−0.010, 0.002)	−0.003 (−0.009, 0.003)
No olfactory impairment, change per year	−0.0006 (−0.0007, −0.0004)	−0.0005 (−0.0007, −0.0004)	−0.0003 (−0.0004, −0.0001)	−0.0003 (−0.0004, −0.0001)
Olfactory impairment, change per year	0.0000 (−0.001, 0.001)	−0.0001 (−0.002, 0.001)	0.0.003 (−0.0007, 0.001)	0.0002 (−0.0007, 0.001)
Grooved Pegboard Test (GPB) Time	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD slower1	0.0000 (−002, 0.002)	0.0005 (−0.001, 0.002)	−0.001 (−0.003, −0.0001)†	−0.001 (−0.003, 0.0004)
Average time, change per year	−0.0005 (−0.0007, −0.0004)	−0.0005 (−0.0007, −0.0004)	−0.0002 (−0.0004, −0.0001)	−0.0002 (−0.0004, −0.0001)
1 SD above average time, change per year	−0.0004 (−0.0008, −0.0001)	−0.0004 (−0.0008, 0.0000)	−0.0002 (−0.0005, 0.0002)	−0.0002 (−0.0005, 0.0002)
Grip Strength	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD lower2	0.0002 (−0.002, 0.002)	0.0003 (−0.001, 0.002)	−0.0001 (−0.002, 0.001)	0.0000 (−0.001, 0.001)
Average grip, change per year	0.0005 (0.0004, 0.0007)	0.0005 (0.0004, 0.0007)	0.0002 (0.0001, 0.0004)	0.0002 (0.0001, 0.0004)
1 SD below average grip, change per year	0.0005 (0.0002, 0.0009)	0.0005 (0.0001, 0.0009)	0.0003 (−0.0001, 0.0006)	0.0002 (−0.0001, 0.0006)
Physical Functioning Scale (PFS)	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD lower3	−0.002 (−004, −0.0004)†	−0.001 (−0.003, 0.0004)	−0.0009 (−0.002, 0.0005)	−0.0006 (−0.002, 0.0009)
Average PFS score, change per year	0.0005 (0.0004, 0.0007)	0.0005 (0.0004, 0.0007)	0.0002 (0.0001, 0.0004)	0.0002 (0.0001, 0.0004)
1 SD below average PFS score, change per year	0.0007 (0.0003, 0.001)	0.0007 (0.0003, 0.001)	0.0003 (−0.0001, 0.0006)	0.0003 (−0.0001, 0.0006)

Model 1: Linear mixed-effects model adjusted for baseline age, sex (in sensory models only), random slope (in women only), and random intercept. Change per year by sensory/motor function calculated from regression estimates for time and time*sensory/motor interaction terms.

Model 2: Linear mixed-effects model 1, additionally adjusted for baseline education, diabetes, hypertension, presence of carotid artery plaque, smoking status, weekly alcohol consumption, regular exercise, waist circumference, blood lead level, blood cadmium level, serum high-sensitivity C-reactive protein, serum interleukin-6, natural log of serum soluble intercellular adhesion molecule-1, natural log of serum soluble vascular cell adhesion molecule-1. Change per year by sensory/motor function calculated from regression estimates for time and time*sensory/motor interaction terms.

¹Centered at mean GPB time: Women= 68s, Men=76s; Standard deviation (SD): Women=13.6s, Men=16.6s

²Centered at mean grip strength: Women=29kg, Men=49kg; SD for grip strength; Women=5.9kg, Men=9.3kg

³Centered at mean PFS score: Women=87, Men=90; SD for PFS: Women=16.9, Men=14.7

^†p<0.05 for difference in baseline biomarker level or rate of change in biomarker

In minimally adjusted models, slower baseline peg time was associated with lower baseline Aβ₄₂/Aβ₄₀ ratio among men, and lower PFS was associated with lower baseline Aβ₄₂/Aβ₄₀ ratio among women. These associations were attenuated and no longer significant after further adjustment. There were no significant differences in rate of change in Aβ₄₂/Aβ₄₀ ratio over time by motor functions.

3.6. Sensory and motor function and 10-year change in serum TT

Baseline effects and rates of change in TTau by sensory and motor function from linear mixed effect models are displayed in Table 5. Rates of change in TTau over the 10-year period were relatively stable across all models in both minimally and fully adjusted analyses. No significant relationships were observed between sensory or motor functions and 10-year change in TT.

Table 5:

Differences in baseline levels and rate of change in serum total tau (TTau) over 10 years by baseline sensory and motor function.

	TTau, % (95% CI)
	Overall (N=1529)
Hearing Impairment	Model 1		Model 2
Baseline Effect	2.5 (−7.8, 14.0)		0.5 (−9.9, 12.1)
No hearing impairment, change per year	1.3 (0.8, 1.8)		1.4 (0.9, 1.8)
Hearing impairment, change per year	1.2 (−0.5, 3.0)		1.3 (−0.5, 3.1)
Visual Impairment	Model 1		Model 2
Baseline Effect	5.1 (−4.8, 16.2)		6.3 (−4.0, 17.7)
No vision impairment, change per year	1.3 (0.8, 1.8)		1.4 (0.9, 1.9)
Vision impairment, change per year	1.5 (−0.2, 3.3)		1.2 (−0.5, 3.0)
Olfactory Impairment	Model 1		Model 2
Baseline Effect	16.1 (−4.8, 41.6)		9.4 (−11.1, 34.6)
No olfactory impairment, change per year	1.3 (0.9, 1.8)		1.4 (0.9, 1.8)
Olfactory impairment, change per year	0.2 (−2.6, 3.1)		0.7 (−2.3, 3.8)
	Women (N=828)		Men (N=701)
Grooved Pegboard Test (GPB) Time	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD slower1	−0.8 (−5.2, 3.8)	−3.2 (−7.7, 1.5)	1.3 (−3.9, 6.9)	−0.5 (−6.0, 5.2)
Average peg time, change per year	1.5 (0.9, 2.1)	1.6 (0.9, 2.2)	1.1 (0.4, 1.7)	1.1 (0.5, 1.7)
1 SD above average peg time, change per year	1.9 (0.7, 3.1)	2.0 (0.7, 3.2)	0.9 (−0.4, 2.1)	1.0 (−0.3, 2.2)
Grip Strength	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD lower2	−2.8 (−7.2, 1.9)	−3.2 (−7.8, 1.7)	3.9 (−1.5, 9.6)	3.9 (−1.6, 9.8)
Average grip, change per year	1.6 (0.9, 2.2)	1.6 (1.0, 2.2)	1.1 (0.4, 1.7)	1.1 (0.5, 1.7)
1 SD below average grip, change per year	1.8 (0.6, 3.0)	1.9 (0.6, 3.1)	0.7 (−0.5, 1.9)	0.7 (−0.6, 2.0)
Physical Functioning Scale (PFS)	Model 1	Model 2	Model 1	Model 2
Baseline effect, per SD lower3	0.4 (−3.9, 5.0)	−1.2 (−5.9, 3.8)	2.9 (−2.4, 8.4)	1.5 (−3.9, 7.3)
Average PFS score, change per year	1.5 (0.9, 2.1)	1.6 (0.9, 2.2)	1.1 (0.4, 1.7)	1.1 (0.5, 1.7)
1 SD below average PFS score, change per year	1.9 (0.7, 3.1)	1.9 (0.7, 3.2)	1.3 (0.0, 2.5)	1.3 (0.0, 2.6)

Model 1: Linear mixed-effects model adjusted for baseline age, sex (in sensory models only), random slope (in men only), and random intercept. Change per year by sensory/motor function calculated from regression estimates for time and time*sensory/motor interaction terms.

Model 2: Linear mixed-effects model 1, additionally adjusted for baseline education, diabetes, hypertension, presence of carotid artery plaque, smoking status, weekly alcohol consumption, regular exercise, waist circumference, blood lead level, blood cadmium level, serum high-sensitivity C-reactive protein, serum interleukin-6, natural log of serum soluble intercellular adhesion molecule-1, natural log of serum soluble vascular cell adhesion molecule-1. Change per year by sensory/motor function calculated from regression estimates for time and time*sensory/motor interaction terms.

¹Centered at mean GPB time: Women= 68s, Men=76s; Standard deviation (SD): Women=13.6s, Men=16.6s

²Centered at mean grip strength: Women=29kg, Men=49kg; SD for grip strength; Women=5.9kg, Men=9.3kg

³Centered at mean PFS score: Women=87, Men=90; SD for PFS: Women=16.9, Men=14.7

3.7. Sensitivity analyses

All reported results were unchanged in sensitivity analyses excluding participants with measured serum biomarker level greater than the upper limit of quantification. Results from models including BMI instead of waist circumference had nearly identical results.

3.8. Substudy analyses

Among 34 BOSS participants included in a comparison of serum and plasma biomarkers, NfL concentrations were on average similar. Concentrations of TTau were lower in serum and the Aβ₄₂/Aβ₄₀ ratio higher (Supplement 4). We found a very strong correlation between serum and plasma NfL concentrations (correlation coefficient=0.96, p<0.0001), a moderate correlation for the Aβ₄₂/Aβ₄₀ ratio (corr. coeff.=0.68, p<0.0001), and a weak correlation for TTau (corr. coeff.=0.35, p=0.04) (Supplement 4).

Serum biomarker levels were validated against CSF-based and ¹¹C-PiB PET-based amyloid positivity with samples from the WRAP/ADRC cohort. A strong correlation was found between serum and CSF NfL concentration (corr. coeff.=0.71, p<0.0001), and weak correlations for the Aβ₄₂/Aβ₄₀ ratio and TTau (corr. coeff.=0.39, p<0.0001; corr. coeff.=0.21, p=0.03, respectively) (Supplement 4). However, in those with CSF amyloid positivity, serum concentrations of NfL and TTau were significantly higher, and the Aβ₄₂/Aβ₄₀ ratio was significantly lower, compared to those without CSF pathology. In those with amyloid pathology identified by ¹¹C-PiB PET, serum NfL concentrations were significantly higher, and the Aβ₄₂/Aβ₄₀ ratio was significantly lower, compared to those without brain amyloid pathology. There was no significant difference in serum TTau between those with ¹¹C-PiB PET amyloid positive status and those with negative status. (Supplement 4).

4. DISCUSSION

In the present study, we evaluated associations of sensory and motor function with serum NfL, Aβ₄₂/Aβ₄₀ ratio_, and TTau, biomarkers of neurodegeneration and AD pathology in a sample of community-living adults. We found that baseline measures of hearing impairment and motor function were associated with baseline and 10-year rate of change in NfL, a non-specific marker of neuronal injury and neurodegeneration, independent of known risk factors of neurodegeneration. Importantly, the observed associations between hearing and motor impairments and biomarkers of neurodegeneration were significant in models controlling for many covariates, including behavioral, metabolic, vascular, and inflammatory factors. This indicates that these factors do not fully explain the relationships between sensory function and these biomarkers. Additionally, we characterized the distributions of serum NfL, Aβ₄₂/Aβ₄₀ ratio, and TTau by age, sex and 10-year change, in a predominantly middle-aged community living population_. To our knowledge this is the largest study of serum biomarkers in a general population cohort with an age range from younger to older adulthood. The results of this study extend previous research on the relationship between sensory and motor function and neurodegeneration and cognitive dysfunction, to include the study of midlife blood biomarkers of neurodegeneration and AD. This study also adds important information about the distribution and change over time of these biomarkers in the adult population.

4.1. Serum biomarkers by age and sex, and the 10-year change in biomarker concentrations

Previous studies of blood-based neurodegenerative biomarkers have focused largely on plasma measures, populations with existing pathology or disease, cohorts at high-risk for AD or neurodegenerative disorders, or older adults (de Wolf et al, 2020; Mielke et al, 2017; Preische et al, 2019; Schindler et al, 2019; Sullivan et al, 2021; Verberk et al, 2018). Data on age and sex patterns of these biomarkers and particularly on their change over time in the general population across the adult age spectrum have been limited, especially serum-based markers (Hviid et al, 2020). In the current study of a primarily middle-aged cohort, NfL was higher with older baseline age, Aβ₄₂/Aβ₄₀ ratio was lower, and TTau concentrations were on average not different across baseline age groups. These patterns are consistent with previous studies using plasma, except for TTau, which has been previously reported to weakly correlated with age, although in populations considerably older than the BOSS cohort (de Wolfe et al, 2020; Mielke et al, 2017; Schindler et al, 2019). Over the 10 years, mean NfL and TTau concentrations increased and the mean Aβ₄₂/Aβ₄₀ ratio decreased as the population aged. Our results extend existing research to younger age ranges. Few longitudinal studies of these blood biomarkers have included participants in their 30’s and 40’s and it is notable that a decrease in average Aβ₄₂/Aβ₄₀ ratio trajectory can be seen as early as the fourth decade. Additional studies should further characterize serum-based biomarker levels in order to establish normative data.

Men had a significantly lower Aβ₄₂/Aβ₄₀ ratio than women. Sex differences were also seen in the change in Aβ₄₂/Aβ₄₀ ratio over time as women had a more rapid decline in the ratio than men. Sex differences in Aβ₄₂/Aβ₄₀ ratio have not been widely reported, although two other studies have recently found lower plasma Aβ₄₂/Aβ₄₀ ratios in men versus women (Schindler et al, 2019, Sullivan et al, 2021). Women have higher incidence and prevalence rates of dementia and AD, especially at older ages but, the reason(s) for these differences are still under debate. Some studies have attributed the sex-difference in rates to selective survival among men, while other studies have suggested that women may be more vulnerable to dementia and AD due to biological differences related to hormonal differences, menopausal changes, or differences in brain structure (Beam et al, 2018; Chene et al, 2015; Snyder et al, 2016). Therefore, different pathological mechanisms or trajectories of neurodegeneration in men and women are plausible and might be reflected in blood biomarker levels. There were no differences by sex in baseline NfL and TTau.

Concentrations of Aβ₄₀, Aβ_42, and TTau have been reported to be lower in serum than plasma when measured by SIMOA and serum-based measures have been used less frequently in research (Ashton et al, 2021; Verberk et al, 2021). In our comparison of biomarker concentrations in serum and plasma, NfL and Aβ₄₂/Aβ₄₀ ratio had strong and moderate correlations, respectively, while TTau concentrations, which were on average low in this primarily middle-aged cohort, were weakly correlated. These findings are consistent with those of another small study (Ashton et al, 2021). Additionally, in the sub-study of serum samples from the WRAP/ADRC, two cohorts enriched for risk for and history of AD, concentrations of serum NfL and the Aβ₄₂/Aβ₄₀ ratio were significantly different between those with and without CSF-based and ¹¹C-PiB PET confirmed amyloid pathology. These results indicate that serum NfL and Aβ₄₂/Aβ₄₀ ratio may be useful in identifying neurodegenerative pathology and provides support for the use of stored serum samples in longitudinal cohort studies (Schubert et al, 2022).

4.2. Sensory impairments and rates of biomarker change

We found that those with baseline hearing impairment had a more rapid increase in NfL over the 10-year follow-up period, although effect sizes were small. This is consistent with studies showing associations of hearing function with cognitive decline which may be due to neurodegenerative processes (Loughrey et al, 2018; Merten et al, 2020; Shen et al, 2018). This suggests that hearing loss could be indicative of early neurodegeneration in the brain or auditory system, and that a hearing screening test could potentially be useful to identify those at higher risk of neurodegeneration. Interestingly, in the current study, we did not observe a significant relationship between hearing impairment and amyloid beta ratio, a marker of AD specific pathology. Previous studies have found associations of hearing loss with all-cause dementia and AD (Gates et al, 2011; Loughrey et al, 2017; Yuan et al, 2018). This could mean that the current study indicates that hearing might be more closely related to non-specific neurodegeneration or that longer follow-up may be needed to detect an association between hearing and AD pathology, given the limited change in AD-specific biomarkers in this younger cohort. We found an unexpected association of a faster decline of the Aβ₄₂/Aβ₄₀ ratio over time in women without vision impairment compared to those with vision impairment. However, the mean baseline Aβ₄₂/Aβ₄₀ ratio was considerably lower among those with vision impairment. Thus, the unexpected direction of effect for change in Aβ₄₂/Aβ₄₀ ratio by vision impairment could be due to a floor effect, i.e., women with vision impairment starting lower leaves less room for decline over time. Similarly, this could suggest that women with vision impairment may have ‘aged faster’ and decline in their Aβ₄₂/Aβ₄₀ ratio had plateaued. This finding may also be due to chance.

4.3. Motor Function and NfL, the Aβ₄₂/Aβ₄₀ ratio, and TTau

In the current study, we found consistent relationships between baseline motor function measures and change in NfL concentrations, while there were no associations with changes in the Aβ₄₂/Aβ₄₀ ratio or TTau. Weaker grip strength,slower grooved pegboard test performance, and lower PFSwere associated with more rapid increases in NfL concentrations over 10 years. There was a similar rate of change in levels for both men and women, while estimates were not statistically significant in men.

The faster increase in serum NfL concentrations among those with worse performance on motor function tests may reflect ongoing neurodegenerative processes, including peripheral nerve degeneration which occurs with aging, that affect both grip strength and manual dexterity. Grip strength is not only a measure of muscle mass, but it has also been associated with cognitive decline and AD, and lower brain volume (Buchman et al, 2007; Camargo et al, 2016; Carson 2018). Likewise, performance on the grooved pegboard test is associated with connectivity in supplementary motor areas in the brain (Lammers et al, 2020) and with cognitive function (Ashendorf et al, 2009). Motor tasks are highly dependent on central mechanisms of neuromuscular control and therefore functional motor measures may be sensitive markers of CNS changes (Carson 2018; Seidler et al, 2010). We also found that self-reported poorer physical functioning, based on a lower PFS, was associated with a faster increase in NfL over ten years among women. These findings suggest that the PFS may be useful for predicting risk of neurodegeneration among women. In recent years, NfL has shown great potential as a biomarker of neuronal damage in many neurological conditions including multiple sclerosis and Parkinson’s disease (Gaetani et al, 2019). In the current study, the association between measures of motor function and serum NfL suggest that these motor measures may be useful for identifying individuals with an increased risk for neurodegeneration, particularly for women. It is unclear why the observed differences in rate of change in NfL by motor function was only significant among women, but there are known sex differences in physical functioning as people age (Gordon et al, 2017). Previous studies have found that women are more likely to have physical disability and frailty than men (Botoseneanu et al, 2016; Gordon et al, 2017; Merrill et al, 1997), which would be consistent with faster increases in NfL and neurodegeneration. Moreover, studies have found sex differences in self-reported disability, with women tending to overreport and men to underreport disability when compared to objectively measured physical function (Botoseneanu et al, 2016; Merrill et al, 1997). Further studies are warranted to determine the role of these sex differences in motor function and biomarkers of neurodegeneration.

4.4. Strengths and Limitations

The strengths of our study include the large, well-characterized cohort with 10 years of follow-up with standardized measures of multiple sensory and motor functions and stored blood samples from three time points spanning 10 years. The objective measurement of sensory function is a strength of the current study, while self-reported physical function and use of grip strength and the GPB test as proxies for motor function are potential weaknesses. Pre-analytical handling and storage of blood samples across all phases was in concordance with currently recommended protocols for measuring NfL, Aβ₄₂, Aβ₄₀, and TTau in blood (Supplement 1) (Ashton et al, 2021; Schubert et al, 2022; Verberk et al, 2021). The biomarker assays were conducted using the kits and reagents from the same lots. We were able to determine the correlation of biomarker concentrations in serum and plasma and to evaluate the ability of these serum biomarkers to differentiate between people with and without neuropathology based on two accepted measures of amyloid pathology, CSF biomarkers and ¹¹C-PiB PET confirmed amyloid.

The current study included only participants with available samples and data from all three BOSS phases which could lead to selection bias, although the baseline characteristics of the study sample were similar to the full baseline BOSS cohort. The BOSS cohort is predominantly non-Hispanic White, which may limit the generalizability of our findings to other populations. While we aimed to determine changes in neurodegenerative biomarkers in midlife, the general good health and relatively young age of the cohort, and the small levels of biomarker change over time may make it difficult to detect effects. While analyses controlled for a wide range of potential confounders, other factors which could be related to sensory and motor function and blood-based biomarker levels, such as kidney and liver function, were not available. Serum TTau levels were low in this younger population with many participants having concentrations below the LOD. This may indicate that serum TTau may not be as good of a biomarker of neurodegeneration in younger populations, and we may not have had the ability to detect associations with sensory and motor functions. Serum TTau may be more useful in older populations with higher TTau levels or with more neurodegenerative disease, especially AD. Additionally, recent studies have shown that phosphorylated-tau (PTau) may be a more specific indicator of pathological tau and AD and future studies should evaluate the relationship between sensory and motor function and PTau (Jack et al, 2018; Janelidze et al, 2020).

5. CONCLUSIONS

We found relationships between midlife sensory (hearing) and motor functions and long-term changes in serum NfL concentration, a non-specific marker of neurodegeneration. These relationships remained when taking other traditional risk factors for neurodegeneration into account. No consistent relationships were found between sensory and motor function and changes in biomarkers specific to AD. Future studies with longer follow-up should determine if sensory and motor changes are more reflective of general neurodegenerative versus AD-specific pathology. If confirmed, hearing or motor function tests could potentially become relevant screening tools to identify those at higher risk for neurodegeneration to target use of future prevention and intervention strategies.

Highlights.

• Serum NfL and Tau levels increased, and Aβ₄₂/Aβ₄₀ ratio decreased over 10 years

• Serum NfL levels correlated well with NfL levels in plasma and CSF

• Serum NfL was higher and Aβ₄₂/Aβ₄₀ ratio lower in those with CSF or PET amyloid

• Midlife hearing impairment associated with faster increase in NfL level over time

• Worse midlife motor function associated with faster increase in NfL over time