Associations between Inflammation and Multisensory Impairment among Older Adults

Abstract

Background:

Impairment in multiple senses (multisensory impairment) is common in older adults but the underlying mechanisms are unclear. We evaluated whether common blood-based markers of inflammation (e.g., Interleukin 6 (IL-6), C-Reactive Protein (CRP), and Tumor Necrosis Factor Alpha (TNF-α)), were associated with multisensory impairment.

Methods:

We analyzed data from 1674 participants in the Health, Aging, and Body Composition Study, a prospective cohort study of Black and White older adults who were aged 70–79 at enrollment. IL-6, CRP, and TNF-α were assayed from blood samples at Year 1. Sensory function in 4 domains was assessed in Years 3–5; impairment was defined with clinical cut-points. Vision was measured by visual acuity and contrast sensitivity; hearing by pure tone audiometry (500, 1000, 2000, and 4000 Hz); smell by the 12-item Cross Cultural Smell Identification Test; and touch by vibration detection threshold and monofilament of the big toe. A previously developed multisensory impairment score (0–12) was calculated based on sample quartiles and summed across sensory domains. Regression models evaluated the associations of inflammation markers with individual and multiple sensory impairments (as separate outcomes) with adjustment for demographics, health conditions, and health behaviors.

Results:

Higher CRP (ß=0.07; 95%CI: 0.01–0.12; p=0.01) and IL-6 (ß=0.11; 95%CI: 0.04–0.18; p=0.003) levels were associated with number of sensory impairments. Participants with highest quartile of IL-6 (OR=1.45; 95%CI: 1.09–1.92; p=.01) and TNF-α (OR=1.46; 95%CI: 1.12–1.91; p=.005) had higher odds of a poor multisensory impairment score. High CRP was associated with impaired vision (OR=1.45; 95%CI:1.08–1.93; p=.01) and high TNF-α was associated with touch impairment (OR=1.63; 95%CI:1.15–2.30; p=.006). Having multiple high markers was also associated with multisensory (OR: 1.76;95% CI: 1.20–2.58;p= 0.004) and vision impairment (OR: 1.55; 95% CI: 1.13–2.13;p= 0.004).

Conclusions:

Markers of inflammation were associated with multisensory impairment, but there were fewer associations with individual sensory impairments.

INTRODUCTION

Sensory impairments are associated with poor health outcomes among older adults, including increased risk of mortality and dementia.^1,2 Sensory function declines with age, and impairments in multiple senses (multisensory impairment) are common.³ Compared to individual sensory impairments, multisensory impairment is more strongly associated with poor health outcomes.^1,2,4 However, the etiology of multisensory impairment remains unclear.

Sensory impairments may be affected by systemic disease processes such as chronic inflammation, which is also common in aging.⁵ Inflammation may play a role in the pathogenesis of certain eye diseases,⁶ poor smell,⁷ and inner ear degradation in age-related hearing loss.⁸ Inflammatory markers such as interleukin-6 (IL-6) or C-reactive protein (CRP) have been associated with age-related macular degeneration (AMD),^9,10 as well as impairments in hearing^11,12 and olfaction^13,14, but not all studies found associations.^15–17 Tumor Necrosis Factor-α (TNFα), another inflammation marker, has been less well studied with sensory loss but has been linked to sudden hearing loss¹⁸ and diabetic-retinopathy.¹⁹ However, there remains limited data evaluating the associations in larger community-based (not clinically based) studies of older adults with objective measures of sensory function. Few studies have examined whether inflammation is associated with measures of multisensory impairment. Such work could improve our understanding of the etiology of age-related sensory loss and avenues for prevention.

Our objective was to examine the associations between several markers of inflammation (CRP, IL-6, and TNFα) and multisensory impairment in older adults from the Health, Aging, and Body Composition (Health ABC) Study. We also examined associations with individual sensory impairments.

METHODS

Population

Health ABC is a cohort study of 3075 Black and White men and women aged 70–79 years and recruited in 1997–1998 (Year 1) from a random subset of Medicare-eligible adults living in Pittsburgh, PA and Memphis, TN.²⁰ To be eligible for Health ABC, participants were required to report no difficulties in daily activities, climb 10 stairs and walk ¼ mile without difficulty, be free from life-threatening cancers, and plan to remain in the study area for at least 3 years. Those enrolled were followed with yearly clinical examinations, 6-month phone calls, and medical record review for major health events until death, drop-out, or study close (2013–2014). All participants provided informed consent, and the study was approved by the institutional review boards at the study coordinating center and each clinical site. The analysis presented here includes 1674 participants with complete measures of inflammation and sensory test data (from Years 3–5). Participants were excluded if they died, dropped-out, or had dementia through study Year 5 (n=575); or if they had one or more missing sensory assessment (n=600), inflammation marker (n=167), or covariate measure (n=59). Excluded participants were on average older and were more likely to be Black, have not completed high school, have a lower income, have smoked, and have comorbidities compared to the analytic sample (Supplementary Table S1).

Inflammation Markers

Biological specimens were processed according to standardized protocols. Blood samples were drawn in Year 1 after an 8-hour fast and stored at −70°C at the Health ABC core laboratory at the University of Vermont.²¹ Serum CRP level was measured by enzyme-linked immunosorbent assay (ELISA) in duplicate based on purified protein and polyclonal anti-CRP antibodies (Calbiochem, EMD Biosciences Inc, Darmstadt, Germany). The CRP assay was standardized according to the World Health Organization First International Reference Standard with a sensitivity of 0.08 μg/mL. Plasma IL-6 level was measured by ELISA with a detectable limit of 0.10 pg/mL (R&D Systems, Minneapolis, MN). Plasma TNF-α concentration was measured in duplicate using the ELISA High Sensitivity HSTA50 kit (R&D Systems., Minneapolis, MN), with a lower detection limit of 0.18 pg/ml, and the detection range was 0.5–32 pg/ml. We defined high levels of each marker based on the highest sample quartile (CRP > 3.14 μg/mL, IL-6 > 2.79 pg/ml, TNF-α > 4.11 pg/ml).

Multisensory Impairment

Binocular vision (with corrective lenses) was assessed in Year 3 using the Bailey-Lovie distance visual acuity test²² (scores were converted to Snellen equivalents) and the Pelli-Robson contrast sensitivity test.²³ Hearing (without hearing aids) was measured during Year 5 with audiometric assessments conducted in a sound-treated booth and audiometric thresholds were measured at 0.5, 1, 2, and 4 kHz. Pure tone average (PTA) was calculated as an average of hearing thresholds for the better hearing ear. Touch was assessed during study Year 4 by measuring lower extremity sensory nerve function in two ways: (1) vibration detection threshold (in μm) at the bottom of the big toe, and (2) monofilament testing with both standard (10g) and light (1.4g) monofilaments; participants were classified as either able or unable to detect 3 of 4 touches with each monofilament. Olfaction was assessed using the Cross Cultural Smell Identification Test²⁴ in study Year 3, a 12-item culturally sensitive measure.

While there are no established scores for quantifying multisensory impairment, we used two measures, developed in our prior work.^4,5 First, we defined impairment for each of the four sensory functions listed above using clinical or previously published cut-points, described in detail elsewhere.^4,5 Visual impairment was defined as current visual acuity of 20/40 or worse or a log contrast sensitivity of 1.55 or less. Impaired smell was defined as the lowest tertile of olfactory scores based on previously calculated race- and sex-stratified tertiles of function (good, moderate, and poor) from Yaffe et al (2017).²⁵ We defined touch impairment as inability to detect the standard 10g monofilament or a vibration threshold limit >130μm, based on prior work.²⁶ A PTA >40 decibels was defined as moderate to severe hearing loss. We then summed the number of impairments across the four senses (0–4), and participants were categorized as having none, one, or multiple impairments.

Second, we created a summary score measure of multisensory function based on contrast sensitivity for vision, olfactory scores for smell, vibration detection for touch, and PTA for hearing. For each sensory domain, participants were assigned a score of 0–3 based on sample quartiles of the original score, with higher scores indicating worse sensory function. A summary multisensory score (0–12) was calculated by summing across senses.⁴ We defined a poor multisensory score as 6 or higher.

Other Clinical Characteristics

At baseline, participants reported age, sex, race, educational attainment, income, smoking status (never or ever), heavy alcohol consumption (1+ drink per day); and sedentary lifestyle (no walking during the week). Height and weight were measured in-person and body mass index (BMI) was calculated based on kg/m². History of diabetes, heart disease, hypertension, and stroke were based on medical record review, reported medications, and/or laboratory values. History of depression was defined as a score of 10 or higher on the Center for Epidemiologic Study Depression scale short form.²⁷

Statistical Analyses

Descriptive statistics characterized participants overall and by inflammation levels. Linear regression models evaluated the associations of baseline inflammation markers with multisensory impairment measures (outcomes). We ran separate models for each inflammation marker (continuous, log-transformed). Next, we used logistic regression models to examine associations of high levels of inflammation markers with sensory impairments. Inflammation markers were included simultaneously in models and outcomes were individual impairments (hearing, vision, smell, and touch). We examined poor multisensory score (>=6) as an outcome and to determine if associations were independent of individual senses, we adjusted for individual sensory impairments. We also evaluated re-ran similar models using number of high inflammatory markers (none, one, multiple) as the primary predictor. All models adjusted for age, sex, race, education, BMI, heart disease, diabetes, sedentary lifestyle, depression, smoking, and alcohol use. We conducted analyses in R, version 4.4.1; tests were two-sided with α=0.05.

RESULTS

Of the 1674 participants included in analyses, 861 (51.4%) were female, 571 (34.1%) were Black, and the mean age at baseline was 73.9 years (SD=2.8). Participants with high inflammation on at least one marker (n=797; 47.6%) were more likely to be Black, have a history of smoking, report being sedentary, have a BMI of 30 or higher, and have hypertension, diabetes, or stroke (Table 1). Correlations were weak for CRP and IL-6 = 0.36 but lower for CRP and TNF-α = 0.09 and IL-6 and TNF-α = 0.16. 217 (13.0%) participants had high levels on at least 2 markers and only 53 (3.2%) had high levels on all 3 markers; the most common combination was high CRP and IL-6 (Supplementary Figure S1)

Table 1.

Participant Characteristics at Baseline, Overall and by Inflammation Status

Characteristic*	All Participants n=1674	Low Inflammation n=877	High Inflammation n=797
	N (%) unless specified otherwise
Age (years), mean (SD)	73.9 (2.8)	73.9 (2.8)	73.9 (2.8)
Female	861 (51.4)	432 (49.3)	429 (53.8)
Black	571 (34.1)	273 (31.1)	298 (37.4)
Highschool Education	1335 (79.7)	711 (81.1)	624 (78.3)
Income above 50k/year	271 (18.3)	156 (20.3)	115 (16.2)
Ever Smoker	896 (53.5)	445 (50.7)	451 (56.6)
Alcohol use (> 1 drink/day)	129 (7.7)	60 (6.8)	69 (8.7)
Sedentary (0 walking/week)	672 (40.1)	328 (37.4)	344 (43.2)
Body Mass Index >= 30	407 (24.3)	159 (18.1)	248 (31.1)
Hypertension	877 (52.4)	401 (45.7)	476 (59.7)
Cardiovascular disease	338 (20.2)	174 (19.8)	164 (20.6)
Diabetes	278 (16.6)	110 (12.5)	168 (21.1)
Stroke	126 (7.5)	51 (5.8)	75 (9.4)
Depression	63 (3.8)	28 (3.2)	35 (4.4)
Inflammatory Markers
C-Reactive Protein, μg/mL	2.8 (4.3)	1.4 (0.7)	4.4 (5.8)
Interleukin-6, pg/mL	2.3 (1.8)	1.4 (0.6)	3.2 (2.2)
Tumor Necrosis Factor-α, pg/mL	3.4 (1.6)	2.8 (0.7)	4.1 (2.0)
Sensory impairment
One	650 (38.8)	321 (36.6)	329 (41.3)
Multiple	433 (25.9)	206 (23.5)	227 (28.5)
Multisensory impairment score, mean (SD)	5.4 (2.5)	5.2 (2.5)	5.6 (2.5)

Abbreviations: pg/mL, picogram per milliliter; SD, standard deviation; μg/mL, microgram per milliliter

^*Data collected at baseline except for sensory impairment measures (collected in years 3–5).

Those with high inflammation were more likely to have multiple sensory impairments (28.5% vs. 23.5%) (Table 1). In linear regression models higher levels of CRP and IL-6 were associated with number of sensory impairments and IL-6 was also associated with worse multisensory function scores. (Table 2); TNF-α level was borderline associated with worse multisensory function. Having one or multiple high levels of the inflammatory markers was associated with worse multisensory function in a graded fashion (Table 2).

Table 2.

Linear regression model results for the associations of baseline inflammation markers (predictor) with multisensory impairment (outcome)^*

Inflammation Marker**	Number of Impairments (0–4) ß (95% CI)	p	Multisensory Function Score (0–12) ß (95% CI)	p
CRP, log μg/mL	0.07 (0.01, 0.12)	0.01	0.10 (-0.04, 0.24)	0.16
IL-6, log pg/mL	0.11 (0.04, 0.18)	0.003	0.44 (0.26, 0.63)	<0.001
TNF-α, log pg/mL	0.03 (-0.07, 0.14)	0.54	0.25 (-0.03, 0.54)	0.08
High Inflammation
None	Ref		Ref
One Marker	0.12 (0.02, 0.21)	0.02	0.35 (0.09, 0.61)	0.009
Multiple Markers	0.22 (0.10, 0.34)	<0.001	0.58 (0.26, 0.90)	<0.001

Abbreviations: CRP, C-reactive protein; IL-6, interleukin-6; pg/mL, picogram per milliliter; SD, standard deviation; TNF-α, Tumor Necrosis Factor-α; μg/mL, microgram per milliliter

^*All models adjusted for age, sex, race, education, body mass index, heart disease, diabetes, sedentary lifestyle, depression, smoking, and alcohol use.

^**CRP, IL-6, and TNF-α values were log transformed and included as the primary predictor in separate linear regression models; number of high inflammation markers was included as a categorial predictor in separate models; high levels were based on the highest quartile of each inflammation marker. Estimates (ßs) note the estimated difference in the number of impairments or multisensory score for 1 unit of the predictor.

In models including inflammatory markers simultaneously (Figure 1a), high IL-6 (OR=1.51; 95% CI: 1.08–2.12; p=.016) and high TNF-α (OR=1.48; 95% CI: 1.07–2.04; p=.018) were associated with higher odds of a poor multisensory score even with adjusting for individual sensory impairments. There were no associations with IL-6 and any individual sensory impairments. High CRP was associated with a higher odds of impaired vision (OR=1.45; 95% CI: 1.08–1.93; p=.013) and high TNF-α was associated with a higher odds of touch impairment (OR=1.63; 95% CI: 1.15–2.30; p=.006). Having multiple high levels of inflammation (Figure 1b) was associated with 1.76 times higher odds of poor multisensory score (95% CI: 1.20–2.58;p= 0.004) adjusting for individual sensory impairments, and higher odds of vision impairment (OR: 1.55; 95% CI: 1.13–2.13;p= 0.004). Having one high marker was associated with vision impairment (OR 1.33; 95% CI: 1.03–1.74;p=0.03). There were no significant interactions by sex or race (p>0.1).

Figure 1. Logistic regression results for the associations of high levels of each inflammatory marker (A) and number of high markers (B) with sensory impairments (outcomes).

In panel A high levels Interleukin 6 (IL-6), Tumor Necrosis Factor Alpha (TNF-α) and C-Reactive Protein (CRP) based on the highest sample quartile were included simultaneously as predictors in models with each sensory impairment as the outcome (multisensory impairment was defined as a poor multisensory score (6 or higher, range 0–12)). In Panel B a categorical variable for number of high markers was included as primary predictor in predictors in models with each sensory impairment as the outcome. All models adjusted for age, sex, race, education, body mass index, heart disease, diabetes, sedentary lifestyle, depression, smoking, and alcohol use. Multisensory impairment models also additionally adjusted for impairment in vision, hearing, smell, or touch.

DISCUSSION

We evaluated the associations between levels of CRP, IL-6, and TNF-α with both individual and multisensory impairments. All markers of inflammation were associated with multisensory impairment, especially IL-6. CRP was also associated with vision impairment, and TNF-α was associated with touch impairment. Having multiple markers at high levels was associated almost 1.8 times odds of multisensory impairment. This work suggests that inflammation is associated with multisensory impairment in older adults. Associations were less consistent with individual sensory impairments and did not fully account for associations with multisensory impairment.

Our strongest findings were that IL-6 and having multiple high markers were associated with multisensory impairment. Although some studies have found individual sensory associations with IL-6,^9,11,13 the evidence is mixed.¹⁵ CRP and TNF-α were associated with multisensory impairment in several models as well. Together, this suggests multisensory impairment is linked to chronic inflammatory processes. These associations remained after adjustment for individual sensory impairments, perhaps indicating that there is a global disease processes underlying multisensory impairment. This adds to prior work markers of inflammation are associated with other aging outcomes including lower muscle mass²¹ and mortality.²⁸ Aging and other chronic diseases facilitate increased release of cytokines (including IL-6 and TNF-α) which creates a chronic pro-inflammatory state (aka “inflammaging”). CRP is an acute phase protein that increases in response to an increase in cytokines, especially IL-6 levels.²⁹ Chronic inflammation that impacts cellular repair, immune system response, endothelial and cardiometabolic dysfunction which may lead to degeneration of tissues in sensory structures.^6,8,29 IL-6 in particular has been highlighted as a marker of inflammaging shared across age-related diseases.²⁹ CRP levels were only slightly correlated with IL-6, however their biologic link and that findings that CRP was not associated with multisensory impairment in models also including IL-6 further suggests IL-6 as the stronger predictor. TNF-α however, also was independently associated with multisensory impairment, suggesting multiple cytokines may be important. However, future studies will be needed to clarify and replicate differences between specific markers. Overall, we found all markers, especially multiple high levels, were associated with multisensory impairment.

We found more limited associations with individual sensory impairments. CRP and number of high markers was associated with vision impairment. CRP is associated with AMD in prior work;¹⁰ however, there are relatively few participants reporting AMD in Health ABC (~100) suggesting the association may be more general with vision impairment. TNF-α was associated with touch impairment. TNF-α may play an important role in insulin resistance and diabetes,³⁰ and is associated with diabetes complications including diabetic-retinopathy.¹⁹ Although we adjusted for diabetes, diabetes is more common for those with touch and multisensory impairment.² CRP and TNF-α may represent distinct disease mechanisms that impact specific sensory impairments, however since these were also associated with multisensory impairment mechanisms may overlap. We found no associations between any markers with hearing or smell. While this is in contrast to studies that have found associations between inflammation and both hearing^11,12 and smell,^13,14 it aligns with other research showing no associations between inflammatory markers and individual sensory impairments.^15–17

Strengths of this work include multiple objective sensory measures, comparisons between individual and multiple sensory impairments, several markers of inflammation, a biracial cohort, and robust covariate adjustment. This study also has some limitations; sensory measures were assessed across multiple years without longitudinal assessments, and participants were Black or White and healthy compared to the general population. Health ABC participants in our study differed in demographics and were healthier than those excluded. As such, results could be impacted by selection biased and may not be representative of all older adults. Future work will be needed in more diverse and representative populations.

We found associations between inflammation markers and multisensory impairment in a cohort of Black and White older adults. While our results contribute to an understudied area, future work is needed to determine whether chronic inflammation is a risk factor for the development of sensory impairment or whether it explains the association between sensory loss and other poor health outcomes.

Supplementary Material

Supplement

Supplementary Figure S1. Counts and overlap of the highest quartile of C-Reactive Protein (CRP), Interleukin 6 (IL6), and Tumor Necrosis Factor Alpha (TNFa).

Supplementary Table S1. Participant characteristics of the analytic sample compared to those excluded from the analytic sample (due to death, drop-out, or missing data)

Key Points.

• Serum and plasma markers of inflammation (Interleukin 6 (IL-6), C-Reactive Protein (CRP), and Tumor Necrosis Factor Alpha (TNF-α) were associated with sensory impairments in multiple domains (multisensory impairment).

• High CRP was also associated with vision impairment and high TNF-α was associated with touch impairment.

• Associations of inflammatory markers with multisensory impairment were more consistent than associations with individual sensory impairments.

Why does this paper matter?

Chronic inflammation may increase risk for sensory impairments; however, the associations between markers of inflammation and sensory function in individual and multiple domains have been less studied. This work adds to our understanding of inflammation as a potential risk factor for sensory impairments among older adults.