Abstract
Background:
Middle-age has been identified as a critical time-period for health later in life. Identifying factors associated with worse brain function in middle-aged adults may help identify ways to preserve brain function with aging.
Objective:
To evaluate factors associated with a novel measure of brain aging in middle-aged and older adults.
Design:
Longitudinal cohort study
Setting:
Beaver Dam Offspring Study (BOSS) baseline (2005–2008), 5- (2010–2013), and 10-year examinations (2015–2017).
Participants:
2285 adults (22–84 years) with complete sensorineural and neurocognitive data at the 5-year examination.
Measurements:
Principal components analysis (PCA) was performed combining 5-year sensorineural (hearing, vision, olfaction) and cognitive (Trail Making Test A and B, Digit Symbol Substitution Test, Verbal Fluency Test, Auditory Verbal Learning Test) test data. Participants with a standardized PCA score < −1 were classified as having brain aging. Incident brain aging was defined as a PCA score < −1 at 10-years among participants who had a PCA score ≥ −1 at 5-years. Logistic regression and Poisson models were used to estimate associations between baseline factors and prevalent or incident brain aging, respectively.
Results:
Older age, being male, current smoking, larger waist circumference, not consuming alcohol, cardiovascular disease, and Interleukin-6 were associated with greater odds, while more education and exercise were associated with decreased odds, of prevalent brain aging. In addition to age and sex, less than a college education, higher levels of soluble intercellular adhesion molecule-1, diabetes, depressive symptoms, and history of head injury were associated with an increased 5-year risk of incident brain aging.
Conclusion:
In the current study vascular and inflammatory factors were associated with a new brain aging marker in middle-aged and older adults. Many of these factors are modifiable highlighting the importance of addressing health and lifestyle factors in midlife to potentially preserve function for better brain health later in life.
Keywords: Hearing, Olfactory, Inflammation, Cardiovascular, Brain Aging
INTRODUCTION
Middle-age has been identified as a critical time-period for health later in life. Especially for brain health, where it has been established that age-related pathophysiological changes and cognitive decline may begin in midlife and the prodromal period for neurodegenerative diseases may be 20 years or more. 1,2 As aging processes begin in midlife, baselines and trajectories for physical and cognitive health in older age may be established and exposures during this period, both endogenous (blood pressure, atherosclerosis) and exogenous factors (smoking, exercise), may have more significant impacts on health than later in life.3–6 Therefore, identifying factors associated with brain health in middle-age may help identify ways to preserve brain function with aging.
The search for modifiable risk factors for cognitive decline, Alzheimer’s disease (AD), and other dementias, has been the focus of numerous epidemiological studies.3–10 In addition to vascular-related risk factors such as hypertension and inflammation, age-related changes in hearing, vision, and olfaction may be early signals of increased risk for these neurological changes.11–17 Several studies have reported associations between sensory impairments and development of cognitive impairment or decline,11–15 and we have previously shown impairments in hearing, vision or olfaction are associated with worse cognitive function in middle-aged adults 16 and development of cognitive impairment in older adults.17 It has been hypothesized that shared etiologic pathways may explain these associations between sensorineural aging changes and cognitive decline (“common cause” hypothesis) whilst others have suggested that sensory loss causes cognitive decline due to sensory deprivation, social isolation or cognitive overload. 18–20 (Figure 1A) Alternatively, these associations may reflect the close integration of neural processing in sensory systems and cognition. (Figure 1B) It is well-recognized that we see, hear, and smell not with the eyes, ears, and nose, but with the brain; the peripheral sensory organs are responsible for signal transduction and transmission while translation, de-coding and interpretation are higher order functions performed by the brain. Therefore, age-related sensorineural losses in hearing, vision, and olfaction are measured by functional tests that evaluate the integrity of the entire sensorineural system, including aspects of brain function, and are not simple measures of organ specific pathology. 18
Figure 1. Proposed Hypotheses for Relationship between Sensory Impairments and Cognitive Function.
Panel A: 1) Sensory impairments cause cognitive decline due to sensory deprivation, social isolation or cognitive overload. 2) Other factor(s) cause both sensory impairments and cognitive impairment. Adapted from Whitson et al. J Am Geriatr Soc. 2018 Sep 24. doi: 10.1111/jgs.15506.20
Panel B: Associations between sensorineural and cognitive functions due to close neural integration.
Measures of brain function generally focus on neurocognitive function tests but sensorineural and cognitive functions are highly integrated.19,20 The Oxford Dictionary defines cognition as “The mental action or process of acquiring knowledge and understanding through thought, experience and the senses.” 21 Neurocognitive function test batteries rely on processing auditory and/or visual information and sensorineural function tests rely on attention, memory, processing speed and executive function. 18,19 Consequently, both sensorineural and cognitive function measures are capturing aspects of neuronal function to some degree.
Baltes and Lindenberger concluded more than 20 years ago that “…..the increase in the age-associated link between sensory and intellectual functioning may reflect brain aging and that the search for explanations of cognitive aging phenomena would benefit from attending to factors that are shared between the 2 domains.” 22 With this in mind we sought to explore the concept of an expanded measure of brain function that combined sensorineural and cognitive function measures as an indicator of brain function and identify factors associated with this common measure of neurological function in middle-aged and older adults.
METHODS
Study Population
The Beaver Dam Offspring Study (BOSS) is a longitudinal study of the adult children of participants in the Epidemiology of Hearing Loss Study, a population-based study of aging (1993-present). 23, 24 The baseline BOSS examination (BOSS1) occurred in 2005–2008 and follow-up examinations were conducted 5 (BOSS2; 2010–2013) and 10 years (BOSS3; 2015–2017) later. 23, 25–28 Participants included in the analyses had to have BOSS1covariate data and complete sensory and cognitive data from BOSS2. (See detailed methods in Supplementary Text) For the 5-year incidence analyses, participants also had to have complete sensory and cognitive data from BOSS3. This research was approved by the Health Sciences Institutional Review Board of the University of Wisconsin and informed consent was obtained from participants at each examination.
Sensory and Cognitive Measures
Pure-tone air and bone-conduction was used to measure hearing ability and hearing function was defined as the pure-tone average (PTA) of the thresholds at 0.5, 1, 2, and 4 kHz in the worse ear. 23 Vision was measured by contrast sensitivity as log contrast sensitivity units (CS) based on triplet scores in the better eye.28 Olfactory function was measured using the San Diego Odor Identification Test (SDOIT) and the score was the number of correctly identified odors.25 (See Supplementary Methods)
Participants completed a battery of 5 neurocognitive tests at BOSS2 and BOSS3 related to the cognitive domains of attention, speed, executive function, memory and verbal fluency(Trail Making Tests A and B (TMTA, TMTB), modified Rey Auditory Verbal Learning Test (RAVLT), Digit Symbol Substitution Test (DSST), Verbal Fluency Test (VFT)).29, 30 (See Supplementary Text)
Covariates
Vascular health measures, biomarkers, and participant interview for demographic, behavioral and medical history were obtained at baseline using standardized protocols. Blood pressure, height, weight and waist circumference were measured and carotid artery ultrasound scans were obtained for intima-media thickness (IMT) and plaque assessment.26 Arterial stiffness was measured by femoral and radial pulse wave velocity. Non-fasting blood samples obtained at baseline were tested for Hemoglobin A1C (A1C), total and high-density lipoprotein (HDL) cholesterol, interleukin-6 (IL-6), soluble intercellular adhesion molecule-1 (sICAM-1), high sensitivity C-reactive protein (hsCRP), tumor necrosis factor alpha (TNF-α), and soluble vascular cell adhesion molecule-1 (sVCAM-1). (See Supplementary Text) Inflammatory markers were log transformed (natural log (ln)) and analyzed as continuous variables.
Cardiovascular disease (CVD) was considered present if the participant reported a history of any one of eleven physician-diagnosed conditions or procedures. (Supplementary Text) Other demographic and behavioral history included years of education, weekly alcohol consumption in the past year, history of heavy alcohol use (4 or more drinks/day), smoking history (current versus not current), usual exercise (at least once a week long enough to work up a sweat) and presence of depressive symptoms (Centers for Epidemiological Studies Depression Scale (CES-D) score >15). 31
Statistical Methods
All analyses were completed using the SAS version 9.4 software (SAS Institute, Inc. Cary, NC). Composite scores of sensorineural (PTA, CS, SDOIT) and neurocognitive (TMTA, TMTB, DSST, AVLT, DSST) function test data were created using Principal Component Analysis (PCA) with the Factor procedure (Method=Principal with Score option); separate PCA scores were constructed for BOSS2 and BOSS3. Additive inverses of PTA, TMTA and TMTB were used so that lower scores represented worse function and test data were standardized. The detailed methods and coefficient loadings are shown in the Supplementary Text and Supplementary Table 1, respectively. Since the scores were standardized, the PCA score mean was zero and the standard deviation (SD) was 1. Prevalent brain aging was defined as a PCA score more than 1 SD below the mean (< −1) at BOSS2. Incident brain aging was defined as a PCA score more than 1 SD below the mean (<−1) at BOSS3 among participants who had a PCA score ≥−1 at BOSS2.
Logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals (CI) for baseline factors and prevalent brain aging at BOSS2. Poisson models fit to a binary response via Proc Genmod were used to estimate the relative risk (RR) for baseline factors associated with the 5-year incidence of brain aging at BOSS3. Factors significantly associated with brain aging in age-sex-adjusted models were evaluated in multivariable models. Manual stepwise selection was used to construct the multivariable models and significant factors were retained in a base multivariable model. To reduce the potential effects of collinearity between the inflammatory markers, they were tested individually in the multivariable base models. To ameliorate any potential collinearity between the comprehensive CVD variable and CVD-related factors, sensitivity analyses were run for hypertension, diabetes and HDL cholesterol without CVD in the multivariable model.
RESULTS
There were 2285 baseline participants (46% men) with complete sensory and cognitive data at BOSS2. The mean age of participants was 49 years (22–84 years) and most (93.7%) were less than 65 years of age. There were 326 (14.3%) participants classified as having prevalent brain aging at BOSS2 (PCA score <−1); the distribution by age group is shown in Supplementary Figure 1. On average, participants with a PCA score <−1 had a mean PTA 17 dB higher, identified one less odor correctly, had CS 0.08 log triplets lower, TMTA and TMTB times 17 and 57 seconds longer, respectively, produced 3 and 12 fewer words in the AVLT and VFT, respectively, and converted 18 fewer symbols on the DSST as compared to those with PCA scores ≥ −1. (Table 1)
Table 1.
Mean Sensorineural and Neurocognitive Function Test Measures at BOSS 5 & 10-year Follow-ups by Prevalent and Incident Brain Aging
| Prevalent Brain Aging at 5 years | Incident Brain Aging at 10 years | |||
|---|---|---|---|---|
| Yes (n=326) | No (n=1959) | Yes (n=78) | No (n=1423) | |
| Function Measure | Mean (Range) | Mean (Range) | Mean (Range) | Mean (Range) |
| PTA at 0.5, 1, 2, 4 kHz (decibels) worse ear | 32.2 (−1.25–125) | 15.0 (−3.75–125) | 29.5 (6.3–101.3) | 18.3 (−3.75–125) |
| San Diego Odor Identification (# odors) | 6.8 (1–8) | 7.8 (1–8) | 6.6 (1–8) | 7.7 (1–8) |
| Contrast Sensitivity (log triplets) better eye | 1.55 (1.05–1.65) | 1.63 (1.05–2.25) | 1.61 (1.35–1.65) | 1.65 (1.20–1.95) |
| Trail Making Test, Part A (seconds) | 43.2 (20–147) | 26.0 (9–71) | 42.2 (19–88) | 26.0 (7–57) |
| Trail Making Test, Part B (seconds) | 118.5 (48–301) | 61.3 (14–156) | 116.5 (57–300) | 60.3 (11–148) |
| Auditory Verbal Learning Test (# words) | 4.8 (0–11) | 7.7 (0–15) | 5.5 (0–13) | 7.8 (0–15) |
| DSST (# symbols) | 41.1 (13–61) | 59.4 (29–93) | 43.7 (21–58) | 59.5 (35–93) |
| Verbal Fluency Test (# words) | 29.5 (7–55) | 41.5 (9–82) | 34.6 (13–64) | 43.6 (10–92) |
Note. Participants with brain aging have PCA standardized score > 1 SD below the mean PCA (<−1). DSST= Digit Symbol Substitution Test; PTA=Pure-tone Average
Men (OR=3.27, 95% CI=2.48, 4.32, versus women, adjusted for age) and older participants (OR=1.94, 95% CI=1.79, 2.10, per 5 year increase, adjusted for sex) were more likely to have prevalent brain aging at BOSS2. (Table 2) Factors found to be significantly related to an elevated likelihood of brain aging in age-sex-adjusted models were smoking, waist circumference, CVD, hypertension, diabetes, ln hsCRP, ln IL-6, ln sICAM-1, being a non-drinker, and depressive symptoms. Having 16 or more years of education, exercising at least once per week, and higher HDL cholesterol were associated with a decreased likelihood of brain aging. (Table 2)
Table 2.
BOSS1 Participant Characteristics by Brain Aging at BOSS2 and Age- and Sex-Adjusted Associations of BOSS1 factors with Brain Aging at BOSS2 and Incident Brain Aging at BOSS3
| Baseline Participant Characteristic | Brain Aging at BOSS2 | Incident Brain Aging BOSS3 | |||
|---|---|---|---|---|---|
| N | Yes N (%) | No N (%) | Age-Sex Adjusted OR (95% CI) | Age-Sex Adjusted RR (95% CI) | |
| Sex (Male) | 2285 | 217 (66.6) | 829 (42.3) | 3.27 (2.48, 4.32) | 2.50 (1.62, 3.87) |
| Education (16+ years) | 2285 | 56 (17.2) | 719 (36.7) | 0.32 (0.23, 0.45) | 0.50 (0.31, 0.80) |
| Current Smoker | 2281 | 62 (19.0) | 315 (16.1) | 1.99 (1.41, 2.81) | 1.83 (1.03, 3.26) |
| Alcohol consumption past year, g/week | 2280 | ||||
| None | 75 (23.0) | 166 (8.5) | 2.97 (2.02, 4.38) | 1.22 (0.57, 2.59) | |
| 0–14 | 116 (35.6) | 819 (41.9) | 1.00 | 1.00 | |
| 15–74 | 56 (17.2) | 501 (25.6) | 0.77 (0.52, 1.12) | 0.67 (0.37, 1.19) | |
| 75–140 | 35 (10.7) | 232 (11.9) | 0.92 (0.58, 1.47) | 0.77 (0.35, 1.69) | |
| >140 | 44 (13.5) | 236 (12.1) | 0.88 (0.57, 1.36) | 1.46 (0.86, 2.50) | |
| History Heavy Alcohol Use | 2278 | 71 (21.8) | 334 (17.1) | 1.12 (0.81, 1.57) | 1.60 (0.99, 2.58) |
| Exercise, ≥ 1/week | 2279 | 160 (49.2) | 1240 (63.5) | 0.60 (0.46, 0.78) | 0.94 (0.62, 1.43) |
| BMI > 30 kg/m2 | 2214 | 177 (55.8) | 826 (43.5) | 1.16 (0.88, 1.52) | 1.02 (0.68, 1.54) |
| Cardiovascular Disease | 2268 | 51 (15.9) | 95 (4.9) | 1.83 (1.20, 2.80) | 1.60 (0.94, 2.72) |
| Hypertension | 2245 | 190 (58.8) | 624 (32.5) | 1.36 (1.03, 1.79) | 1.54 (1.01, 2.36) |
| Diabetes | 2275 | 48 (14.8) | 78 (4.0) | 2.27 (1.46, 3.53) | 2.68 (1.66, 4.34) |
| Hemoglobin A1C ≥ 6.5% | 2174 | 20 (6.5) | 56 (3.0) | 1.38 (0.77, 2.46) | 2.77 (1.59, 4.82) |
| Head Injury | 2283 | 91 (27.9) | 538 (27.5) | 0.96 (0.71, 1.29) | 1.65 (1.07, 2.54) |
| Depressive Symptoms | 2210 | 63 (20.3) | 243 (12.8) | 2.15 (1.51, 3.06) | 2.04 (1.16, 3.58) |
| Carotid Artery Plaque | 2198 | 139 (44.6) | 372 (19.7) | 1.19 (0.89, 1.60) | 1.19 (0.77, 1.83) |
| Mean (SD) | Mean (SD) | ||||
| Age (per +5 years) | 2285 | 58.2 (9.8) | 47.5 (8.9) | 1.94 (1.79, 2.10) | 1.78 (1.59, 1.99) |
| Waist (per +5 cm) | 2212 | 107.1 (15.7) | 98.7 (16.3) | 1.08 (1.04, 1.13) | 1.03 (0.97, 1.11) |
| Mean IMT (per +0.1 mm) | 2201 | 0.74 (0.19) | 0.64 (0.12) | 1.03 (0.94, 1.13) | 1.08 (0.96, 1.22) |
| Non-HDL Cholesterol (per +30 mg/dL) | 2184 | 152.9 (37.6) | 153.7 (38.9) | 0.98 (0.88, 1.09) | 0.90 (0.72, 1.12) |
| HDL Cholesterol (per +10 mg/dL) | 2184 | 45.9 (14.6) | 50.3 (14.5) | 0.88 (0.79, 0.98) | 1.12 (0.95, 1.32) |
| hsCRP (mg/L) | 2139 | 3.47 (6.5) | 2.65 (4.5) | 1.21 (1.06, 1.38)b | 1.15 (0.90, 1.47)b |
| IL-6 (pg/mL) | 2123 | 3.72 (8.8) | 2.29 (4.5) | 1.64 (1.35, 2.01)b | 1.29 (0.89, 1.87)b |
| TNF-α (pg/mL) | 2139 | 0.70 (0.74) | 0.67 (1.8) | 1.20 (0.99, 1.46)b | 1.01 (0.69, 1.46)b |
| sICAM-1 (ng/mL) (per 0.25 log units) | 2129 | 244.0 (97.3) | 220.7 (66.7) | 1.23 (1.10, 1.38)b | 1.31 (1.09, 1.57)b |
| sVCAM-1 (ng/mL) (per 0.25 log units) | 2139 | 657.8 (259.5) | 582.3 (186.2) | 1.12 (0.99, 1.26)b | 1.08 (0.87, 1.32)b |
| Femoral PWV (per +1 m/s)a | 1255 | 8.6 (3.9) | 8.0 (5.2) | 0.98 (0.94, 1.02) | 0.98 (0.94, 1.04) |
| Radial PWV (per +1 m/s)a | 1412 | 10.2 (4.4) | 10.4 (5.4) | 0.97 (0.94, 1.01) | 1.02 (0.98, 1.05) |
Note. Participants with brain aging have PCA score <−1. Prevalent Brain aging at 5-years: n=326; No Brain aging: n= 1959; Incident brain aging at 10-years n=78. BOSS= Beaver Dam Offspring Study; BMI: Body Mass Index; OR=odds ratio; RR= relative risk; CI=confidence interval; HDL= high density lipoprotein; hsCrP=high sensitivity C-reactive protein; IL-6=interleukin-6; IMT: carotid artery intima media thickness; sICAM-1=soluble intercellular adhesion molecule-1, sVCAM-1=soluble vascular cell adhesion molecule-1; TNF-α =tumor necrosis factor alpha; PWV= Pulse-wave velocity.
Pulse-wave velocity data were only collected on a subset of participants.
Log transformed (natural log (ln))
In multivariable modeling older age (OR=1.92, 95% CI=1.76, 2.10, per 5 years), being male (OR=3.55, 95% CI=2.57, 4.90), current smoking (OR = 1.57, 95% CI=1.08, 2.27, versus non-smokers), waist circumference (OR=1.05, 95% CI=1.004, 1.10, per 5 cm), CVD (OR=1.63, 95% CI = 1.04, 2.56) and being a non-drinker (OR=2.54, 95% CI=1.70, 3.81, vs 0–14 g/week) were associated with an increased likelihood, while 16 or more years of education (OR = 0.37, 95% CI = 0.26, 0.53, vs < 16 years) and exercise (OR=0.74, 95% CI = 0.56, 0.99, ≥1/week vs <1/week) were associated with a decreased likelihood, of brain aging. (Table 3, Model 1) Adding (ln) IL-6 (OR=1.35, 95% CI= 1.07, 1.70) to the model slightly attenuated the estimates for smoking, exercise, and waist circumference and they were no longer statistically significant suggesting some of these factors may be working through inflammatory pathways. (Table 3, Model 2) Results were similar after controlling for depressive symptoms which was not significant in the final model (OR=1.46, 95% CI=0.98, 2.17; CES-D >15) (model not shown). Hypertension, diabetes, hsCRP, sICAM-1 and HDL cholesterol were not significant in the multivariable models or sensitivity analyses. In age-adjusted sex-stratified analyses, estimates were similar to those in the overall multivariable models; in women alcohol, and in men alcohol, education, smoking, and IL-6 remained statistically significant.
Table 3.
Multivariable Models of BOSS Baseline Participant Characteristics and Brain Aging at 5-Year Follow-up
| Model 1 | Model 2 | |
|---|---|---|
| Baseline Factors | OR (95% CI) | OR (95% CI) |
| Age (per +5 years) | 1.92 (1.76, 2.10) | 1.92 (1.76, 2.11) |
| Sex (Male) | 3.55 (2.57, 4.90) | 3.74 (2.67, 5.26) |
| Education (≥16 years) | 0.37 (0.26, 0.53) | 0.39 (0.27, 0.57) |
| Current Smoker | 1.57 (1.08, 2.27) | 1.44 (0.98, 2.12) |
| Exercise (≥ 1/week) | 0.74 (0.56, 0.99) | 0.77 (0.57, 1.04) |
| Waist (per +5 cm) | 1.05 (1.004, 1.10) | 1.02 (0.97, 1.07) |
| Cardiovascular disease | 1.63 (1.04, 2.56) | 1.61 (1.001, 2.58) |
| Alcohol consumption past year, g/week (vs 0–14) | ||
| None | 2.54 (1.70, 3.81) | 2.85 (1.87, 4.33) |
| 15–74 | 0.85 (0.57, 1.27) | 0.90 (0.60, 1.36) |
| 75–140 | 1.04 (0.64, 1.70) | 1.01 (0.61, 1.68) |
| ≥141 | 0.83 (0.52, 1.31) | 0.90 (0.56, 1.45) |
| Ln IL-6 (pg/mL) | 1.35 (1.07, 1.70) |
Note. Model 1 n=2200. Model 2 n=2089. Participants with brain aging have PCA score <−1. BOSS= Beaver Dam Offspring Study; CI=confidence interval; Ln IL-6= natural log interleukin-6; OR=odds ratio.
5-Year Incidence of Brain Aging
Of the 1501 participants with a PCA score ≥ −1 at BOSS2 and with follow-up at BOSS3 (at risk group); there were 78 participants with incident brain aging (PCA score < −1) at BOSS3. (Supplementary Figure 2) Most mean differences in individual sensorineural and neurocognitive test performance by incident brain aging were similar in magnitude to those seen with prevalent brain aging although PTA and CS differences were slightly smaller. (Table 1) Age, sex, smoking, hypertension, a history of head injury, depressive symptoms, diabetes, a hemoglobin A1C ≥6.5% and Ln sICAM-1 at baseline were associated with an increased risk, and more years of education with a decreased risk, of incident brain aging in age-sex adjusted models.(Table 2) In a multivariable model, older age (Relative risk (RR)=1.80, 95% CI=1.59, 2.04, per 5 years), being male (RR=2.71, 95% CI=1.74, 4.23), 16 or more years of education (RR = 0.57, 95% CI = 0.38, 0.94, vs < 16 years), a history of head injury (RR=1.77, 95% CI=1.14, 2.75), depressive symptoms (RR=1.98, 95% CI= 1.15, 3.42, CES-D>15), having diabetes (RR=2.29, 95% CI=1.35, 3.88) and higher levels of sICAM-1 (RR=1.24, 95% CI=1.04, 1.49, per 0.25 log units) were associated with the risk of developing brain aging over 5 years. (Table 4) Sex-specific analyses were not done due to the small number of incident cases.
Table 4.
Multivariable Models of BOSS Baseline Factors and Incident Brain Aging at 10-Years
| Baseline Risk Factors | RR (95% CI) |
|---|---|
| Age (per +5 years) | 1.80 (1.59, 2.04) |
| Sex (Male) | 2.71 (1.74, 4.23) |
| Education (≥16 years) | 0.57 (0.38, 0.94) |
| History of Head Injury | 1.77 (1.14, 2.75) |
| Diabetes | 2.29 (1.35, 3.88) |
| Ln sICAM-1 (ng/mL) (per 0.25 log units) | 1.24 (1.04, 1.49) |
| Depressive Symptoms | 1.98 (1.15, 3.42) |
Note. n=1402. Participants with brain aging have PCA score <−1 BOSS3. BOSS: Beaver Dam Offspring Study; CI=Confidence; Ln sICAM-1: = natural log soluble Intercellular Adhesion Molecule-1; RR=Relative Risk; Interval.
DISCUSSION
In the current study we used a combined measure of sensorineural and neurocognitive functions as a novel marker of brain function to assess factors associated with brain aging in middle-aged and older adults. Participants with brain aging had overall worse performance on sensorineural and neurocognitive function measures as compared to the cohort as a whole, but on average, deficits in function were mild (Table 1). Nevertheless, we detected several risk factors associated with this marker and these results add to the evidence that pathways of vascular, inflammatory and metabolic dysregulation are important contributors to brain aging. (Figure 1, panel B) As many of the identified factors are modifiable, these findings suggest that some brain aging may be delayed or prevented. Similar to early interventions in midlife to prevent CVD, addressing these modifiable factors to prevent early changes in neurological health may reduce or delay brain aging.3–6, 9,10
The innovative component of our study was the use of a composite measure of sensorineural and neurocognitive function as a marker of brain aging. Longitudinal studies have found sensorineural impairments in hearing, vision and olfaction predict cognitive decline or impairment 11–14, 17 suggesting that sensorineural functions may be sensitive to early effects of pathophysiological aging processes or, aging changes may occur early in brain regions associated with sensorineural function. As sensation/perception and cognition are not independent, dysfunction in one neural system will lead to effects in other neural systems. 11,19 For example, the eye is considered an extension of the brain and changes in the eye may reflect changes in the central nervous system 33; olfactory receptor neurons project directly from the olfactory epithelium in the nose into the brain and AD-like pathology has been found to begin early in areas of the brain related to olfactory processing34; and presbycusis is likely the result of aging and pathological changes to the auditory system and the brain (auditory cortex).35 Therefore, we combined these functions, which have close neural integration, into one outcome, versus as an exposure (sensorineural) predicting an outcome (neurocognitive), to evaluate risk factors related to neurological (brain) function. This approach appears to be useful for identifying important opportunities for intervention to preserve multiple aspects of brain function in later life.
Our exploration of this brain aging marker found inflammatory and vascular related factors (age, sex, education, smoking, exercise, waist, CVD, alcohol) associated with brain aging in midlife. These factors are consistent with those reported in previous studies of sensorineural or neurocognitive function. 3–5, 7, 9- 11, 26, 36–41 Notably, IL-6 and sICAM-1 were associated with prevalent and incident brain aging, respectively. Higher levels of systemic inflammation may both directly, and through its deleterious effects on vascular health, promote or accelerate pathophysiological aging changes.42–44 IL-6 attenuated the estimates for smoking, adiposity and exercise suggesting these factors may also have effects on brain aging or function through inflammation. sICAM-1 has been associated with early atherosclerotic processes, 44 advanced plaque in arteries and future cardiac events,45 and insulin resistance. 46 In the current study diabetes and sICAM-1 were both independent predictors of incident brain aging suggesting metabolic dysregulation plays a role in brain aging. (Figure1) As rates of cognitive impairment/AD are declining 47 we may be seeing the cumulative impact of interventions emphasizing healthy lifestyles for early control of vascular disease and diabetes.
Consistent with the recognition that even subtle brain injury can have long-term effects on brain function, 48 participants who reported a history of head injury had an increased risk of developing brain aging. Depressive symptoms were also associated with an increased risk of brain aging; it has been hypothesized that underlying subcortical pathology may be associated with both depression and cognitive function. 49 Although significant in age-sex-adjusted models, hypertension and HDL cholesterol were not significant in multivariable models or sensitivity models excluding CVD. The BOSS cohort is generally healthy and factors associated with this measure of worse brain function may be those associated with early subtle changes and therefore may be different than factors associated with more advanced outcomes such as dementia. Likewise, those with brain aging may represent a mixed group of physiologic aging and pre-clinical AD (pathology) which may have limited our ability to detect some factors if the early risk factors for those conditions differ.
Residual confounding may have limited our ability to detect some factors associated with brain aging and the BOSS cohort is primarily non-Hispanic white which may limit the generalizability of findings to other populations. As functional measures evaluated whole sensorineural systems, it was not possible to separate out potential contributions of peripheral organ (ear, nose, eye) dysfunction to the brain aging marker. Strengths of this study include the large, well-defined cohort, prospective design, standardized measures of sensorineural and neurocognitive function, measured inflammatory and vascular markers and detailed demographic, behavioral and medical histories. The brain aging measure used as the outcome was based on performance on sensorineural and neurocognitive function tests and classification was based on a composite PCA score relative to the whole cohort which allowed us to evaluate factors associated with worse brain function in this younger cohort with low levels of impairment and overall good function. Future studies are needed to validate this novel marker of brain aging and determine if this combined measure is more sensitive than neurocognitive measures alone.
Conclusion
We identified several vascular and inflammatory factors associated with a novel measure of brain aging in middle-aged and older adults. Many of these factors are modifiable highlighting the importance of addressing health and lifestyle factors in midlife to potentially preserve brain function and health later in life. We believe the findings of this study support the concept of using an integrated metric of brain function in studies of brain aging rather than domain specific measures (hearing, memory, etc.). The brain has a multitude of functions which may be affected by aging or pathology at different rates and time points; including sensorineural function together with cognitive function as part of the outcome, not the exposure, provides additional information about neuronal function. This measure may be especially useful in studies of brain aging in younger populations where sensorineural and neurocognitive impairments are less common. Identifying modifiable risk factors associated with worse brain function in middle-aged adults will provide more opportunity for intervention and prevention.
Supplementary Material
Supplementary Figure 1: Prevalent Brain Aging by Age Group
Supplementary Figure 2: Incident Brain Aging by Age Group
Supplementary Table 1: Coefficient Loadings for Principal Component Analysis
ACKNOWLEDGMENTS
FUNDING: This work was supported by R01AG021917 (KJC) from the National Institute on Aging; National Eye Institute; and National Institute on Deafness and Other Communication Disorders and an unrestricted grant from Research to Prevent Blindness. The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Institute on Aging or the National Institutes of Health.
Sponsor’s Role: The sponsor had no role in the design, methods, subject recruitment, data collections, analysis and preparation of paper.
Footnotes
Conflict of Interest: None
Supporting Information
Supplementary Methods: Additional detail on study population, measures, statistical methods
REFERENCES
- 1.Singh-Manoux A, Kivimaki M, Glymour MM, et al. Timing of onset of cognitive decline: results from Whitehall II prospective cohort study. BMJ 2012; 5: 344:d7622. doi: 10.1136bmj/d7622. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Sperling R, Aisen PS, Beckett LA et al. Toward defining the preclinical stages of Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimer Dement. 2011; 7: 280–292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bendlin BB, Carlsson CM, Gleason CE, et al. Midlife predictors of Alzheimer’s disease. Maturitas. 2010; 65: 131–137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Debette S, Seshadri S, Beiser A et al. Midlife vascular risk factor exposure accelerates structural brain aging and cognitive decline. Neurol. 2011; 77: 461–468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fitzpatrick AL, Kuller LH, Lopez OL, et al. Midlife and late-life obesity and the risk of dementia. Arch Neurol. 2009; 66(3): 336–342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Qiu C, Winblad B, Fratiglioni L. The age-dependent relation of blood pressure to cognitive function and dementia. Lancet Neurol. 2005; 4: 487–499. [DOI] [PubMed] [Google Scholar]
- 7.Podewils LJ, Guallar E, Kuller LH et al. Physical activity, APOE Genotype and dementia risk: Findings from the Cardiovascular Health Cognition Study. Am J Epidemiol 2005; 161(7):639–651. [DOI] [PubMed] [Google Scholar]
- 8.Plassman BL, Williams JW, Burke JR, Holsinger T, Benjamin S. Systematic Review: Factors associated with risk for and possible prevention of cognitive decline in later life. Ann Intern Med. 2010; 153: 182–193. [DOI] [PubMed] [Google Scholar]
- 9.Gottesman RF, Albert MS, Alonso A et al. Associations between midlife vascular risk factors and 25-year incident dementia in the Atherosclerosis Risk in Communities Study. JAMA Neurol. 2017;74(10):1246–1254. doi: 10.1001/jamaneurol.2017.1658 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Whitmer RA, Sidney S, Selby J, Johnston C, Yaffe K. Midlife cardiovascular risk factors and risk of dementia in late life. Neurol 2005; 64: 277–281. [DOI] [PubMed] [Google Scholar]
- 11.Albers MW, Gilmore GC, Kaye J, et al. At the interface of sensory and motor dysfunctions and Alzheimer’s disease. Alzheimers Dement. 2015;11(1): 70–98. doi: 10.1016/j.jalz.2014.04.514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Schubert CR, Carmichael LL, Murphy C et al. Olfaction and the 5-year incidence of cognitive impairment in an epidemiological study of older adults. J Am Geriatr Soc 2008; 56:1517–1521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Valentijn SA, van Boxtel MP, van Hooren SA et al. Change in sensory functioning predicts change in cognitive functioning: results from a 6-year follow-up in the Maastricht aging study. J Am Geriatr Soc 2005; 53:374–380. [DOI] [PubMed] [Google Scholar]
- 14.Roberts RO, Christianson TJH, Kremers WK et al. Association between olfactory dysfunction and amnestic mild cognitive impairment and Alzheimer’s disease dementia. JAMA Neurol. 2016; 73(1): 93–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Humes LE. Age-related changes in cognitive and sensory processing: Focus on middle-aged adults. Am J Audiol 2015;24: 94–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Schubert CR, Cruickshanks KJ, Fischer ME, et al. Sensory impairments and cognitive function in middle-aged adults. J Gerontol A Biol Sci Med Sci. 2017; 72(8):1087–1090. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Fischer ME, Cruickshanks KJ, Schubert CR et al. Age-related sensory impairments and risk of cognitive impairment. J Am Geriatr Soc 2016; 64(10):1981–1987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lindenberger U and Baltes PB. Sensory functioning and intelligence in old age: A strong connection. Psychol Aging. 1994;9(3): 339–355. [DOI] [PubMed] [Google Scholar]
- 19.Wayne RV, Johnsrude IS. A review of causal mechanisms underlying the link between age-related hearing loss and cognitive decline. Ageing Res Rev. 2015; 23: 154–166. [DOI] [PubMed] [Google Scholar]
- 20.Whitson HE, Cronin-golomb A, Cruickshanks KJ, et al. American Geriatrics Society and National Institute on Aging Bench to Bedside Conference: Sensory impairment and cognitive declin in older adults. J Am Geriatr Soc. 2018. September 24. doi: 10.1111/jgs.15506. [Epub ahead of print]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Oxford Dictionary https://en.oxforddictionaries.com/definition/cognition accessed 2/6/2018.
- 22.Baltes PR and Lindenberger U. Emergence of a powerful connection between sensory and cognitive functions across the adult life span: A new window to the study of cognitive aging. Psychol Aging. 1997; 12(1): 12–21. [DOI] [PubMed] [Google Scholar]
- 23.Nash SD, Cruickshanks KJ, Klein R et al. The prevalence of hearing impairment and associated risk factors: the Beaver Dam Offspring Study. Arc Otolaryngol Head Neck Surg 2011;137(5):432–439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cruickshanks KJ, Wiley TL, Tweed TS et al. Prevalence of hearing loss in older adults in Beaver Dam, Wisconsin: the Epidemiology of Hearing Loss Study. Am J Epidemiol 1998;148:879–886. [DOI] [PubMed] [Google Scholar]
- 25.Schubert CR, Cruickshanks KJ, Fischer ME, et al. Olfactory impairment in an adult population: The Beaver Dam Offspring Study. Chem Senses. 2012; 37(4): 325–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Schubert CR, Cruickshanks KJ, Fischer ME et al. Carotid intima media thickness, atherosclerosis and 5-year decline in odor identification: The Beaver Dam Offspring Study. J Gerontol A Biol Sci Med Sci. 2015; 70(7):879–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Fischer ME, Schubert CR, Nondahl DM et al. Subclinical atherosclerosis and increased risk of hearing impairment. Atherosclerosis 2015; 238(2):344–349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Paulsen AJ, Schubert CR, Johnson LJ, et al. Cadmium, lead and the 10-year incidence of contrast sensitivity impairment: The Beaver Dam Offspring Study. JAMA Ophthalmol. 2018. epub ahead of print, doi: 10.1001/jamaophthalmol.2018.3931. [DOI] [Google Scholar]
- 29.Reitan RM (1992). Trail Making Test Manual for Administration and Scoring. Reitan Neuropsychology Laboratory, Tucson, AZ. [Google Scholar]
- 30.Strauss E, Sherman E, Spreen O. A compendium of neuropsychological tests: administration, norms, and commentary. 3rd Ed. USA: Oxford University Press, 2009. [Google Scholar]
- 31.Radloff LS (1977). The CES-D scale: A self-report depression scale for research in the general population. Appl Psychol Measure. 1977; 1(3): 385–401. [Google Scholar]
- 32. Jolliffe IT. (2002). Principal Component Analysis 2nd Ed. Series: Springer Series in Statistics. New York: Springer. [Google Scholar]
- 33.Pathai S, shiels PG, Lawn SD, Cook D, Gilbert C. The eye as a model of ageing in translational research – Moleclar, epigenetic and clinical aspects. Age Res Rev. 2013; 12: 490–508. [DOI] [PubMed] [Google Scholar]
- 34.Braak H and Braak E. Diagnostic criteria for neuropathologic assessment of Alzheimer’s disease. Neurobiol Aging. 1997; 18(S4):S85–S88. [DOI] [PubMed] [Google Scholar]
- 35.Humes LE, Dubno JR, Gordon-Salant S. et al. Central Presbycusis: A review and evaluation of the evidence. J Am Acad Audiol 2012; 23(8): 635–666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Cruickshanks KJ, Nondahl DM, Dalton DS et al. Smoking, central adiposity, and poor glycemic control increase risk of hearing impairment. J Am Geriatr Soc 2015;63(5):918–924. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Schubert CR, Cruickshanks KJ, Nondahl DM, Klein BE, Klein R, Fischer ME. Association of exercise with lower long-term risk of olfactory impairment in older adults. JAMA Otolaryngol HeadNeck Surg. 2013; 139:1061–1066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Klein R, Lee KE, Gangnon RE et al. Relation of smoking, drinking, and physical activity to changes in vision over a 20-year period: the Beaver Dam Eye Study. Ophthalmology 2014; 121(6):1220–1228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Wichmann MA, Cruickshanks KJ, Carlsson CM et al. Long-term systemic inflammation and cognitive impairment in a population-based cohort. J Am Geriatr Soc 2014;62(9):1683–1691. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Ngandu T, Helkala EL, Soininen H, et al. Alcohol drinking and cognitive function: Findings from the Cardiovascular risk Factors and Dementia (CAIDE) Study. Dement Geriatr Cogn Disord. 2007; 23: 140–149. [DOI] [PubMed] [Google Scholar]
- 41.Nash SD, Cruickshanks KJ, Zhan W et al. Long-term assessment of systemic inflammation and the cumulative incidence of age-related hearing impairment in the epidemiology of hearing loss study. J Gerontol A Biol Sci Med Sci 2014;69(2):207–214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Singh T, Newman AB. Inflammatory markers in population studies of aging. Ageing Res Rev. 2011; 10: 319–329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Libby P, Ridker PM, Maseri A. Inflammation and atherosclerosis. Circulation. 2002;105:1135–1143. [DOI] [PubMed] [Google Scholar]
- 44.Gross MD, Bielinski SJ, Suarez-Lopez JR, et al. Circulating soluble intercellular adhesion molecule a and subclinical atherosclerosis: the Coronary Artery Risk Development in Young Adults Study. Clin Chem. 2012; 58(2): 411–420. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Stoner L, Lucero AA, Palmer BR. Jones LM, Young JM, Faulkner J. Inflammatory biomarkers for predicting cardiovascular disease. Clin Biochem. 2013; 46: 1353–1371. [DOI] [PubMed] [Google Scholar]
- 46.Hak AE, Pols HAP, Stehouwer CDA. Markers of inflammation and cellular adhesion molecules in relation to insulin resistance in nondiabetic elderly: The Rotterdam Study. J Clin Endocrinol Metab. 2001; 86: 4398–4405. [DOI] [PubMed] [Google Scholar]
- 47.Langa KM, Larson EB, Crimmins EM, et al. A comparison of the prevalence of dementia in the United States in 1000 and 2012. JAMA Intern Med. 2017;177(1): 51–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.McInnes K, Friesen CL, MacKenzie DE, Westwood DA, Boe SG. Mild traumatic brain injury (mTBI) and chronic cognitive impairment: A scoping review. PLoS ONE. 2017. 12(4): e017487 10.1371/journal.pone.0174847. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Comijs HC, van Tilburg T, Geerlings SW, et al. Do severity and duration of depressive symptoms prdict cognitive decline in older persons? Results of the Longitudinal Aging Study Amsterdam. Aging Clin Exp Res. 2003; 16(3): 226–232. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure 1: Prevalent Brain Aging by Age Group
Supplementary Figure 2: Incident Brain Aging by Age Group
Supplementary Table 1: Coefficient Loadings for Principal Component Analysis