Research

JAMA Otolaryngology–Head & Neck Surgery | Original Investigation

Association of Age-Related Hearing Loss With Cognitive

Function, Cognitive Impairment, and Dementia
A Systematic Review and Meta-analysis
David G. Loughrey, BA(Hons); Michelle E. Kelly, DPsychBAT; George A. Kelley, DA; Sabina Brennan, PhD;
Brian A. Lawlor, MD, FRCPI, FRCPsych

Invited Commentary page 127

IMPORTANCE Epidemiologic research on the possible link between age-related hearing loss Supplemental content
(ARHL) and cognitive decline and dementia has produced inconsistent results. Clarifying this
association is of interest because ARHL may be a risk factor for outcomes of clinical dementia.

OBJECTIVES To examine and estimate the association between ARHL and cognitive function,
cognitive impairment, and dementia through a systematic review and meta-analysis.

DATA SOURCES AND STUDY SELECTION A search of PubMed, the Cochrane Library, EMBASE,
and SCOPUS from inception to April 15, 2016, with cross-referencing of retrieved studies and
personal files for potentially eligible studies was performed. Keywords included hearing,
cognition, dementia, and Alzheimer disease. Cohort and cross-sectional studies published in
peer-reviewed literature and using objective outcome measures were included. Case-control
studies were excluded.

DATA EXTRACTION AND SYNTHESIS One reviewer extracted and another verified data. Both
reviewers independently assessed study quality. Estimates were pooled using random-effects
meta-analysis. Subgroup and meta-regression analyses of study-level characteristics were
performed.

MAIN OUTCOMES AND MEASURES Hearing loss measured by pure-tone audiometry only and
objective assessment measures of cognitive function, cognitive impairment, and dementia.
Cognitive function outcomes were converted to correlation coefficients (r value); cognitive
impairment and dementia outcomes, to odds ratios (ORs).

RESULTS Forty studies from 12 countries met our inclusion criteria. Of these, 36 unique
studies with an estimated 20 264 unique participants were included in the meta-analyses.
Author Affiliations: NEIL (Neuro
Based on the pooled maximally adjusted effect sizes using random-effects models, a small Enhancement for Independent Lives)
but significant association was found for ARHL within all domains of cognitive function. Programme, Trinity College Institute
Among cross-sectional studies, a significant association was found for cognitive impairment of Neuroscience, Trinity College
Dublin, Dublin, Ireland (Loughrey,
(OR, 2.00; 95% CI, 1.39-2.89) and dementia (OR, 2.42; 95% CI, 1.24-4.72). Among
Kelly, Brennan, Lawlor); School of
prospective cohort studies, a significant association was found for cognitive impairment (OR, Medicine, Trinity College Dublin,
1.22; 95% CI, 1.09-1.36) and dementia (OR, 1.28; 95% CI, 1.02-1.59) but not for Alzheimer Dublin, Ireland (Loughrey, Lawlor);
disease (OR, 1.69; 95% CI, 0.72-4.00). In further analyses, study, demographic, audiometric, Department of Psychology, National
University of Ireland Maynooth,
and analyses factors were associated with cognitive function. Vascular dysfunction and Kildare, Ireland (Kelly); Meta-Analytic
impaired verbal communication may contribute to the association between hearing loss and Research Group, School of Public
cognitive decline. Health, Department of Biostatistics,
Robert C. Byrd Health Sciences
Center, West Virginia University,
CONCLUSIONS AND RELEVANCE Age-related hearing loss is a possible biomarker and Morgantown (Kelley); Mercer’s
modifiable risk factor for cognitive decline, cognitive impairment, and dementia. Additional Institute for Successful Ageing,
research and randomized clinical trials are warranted to examine implications of treatment for St James Hospital, Dublin, Ireland
(Lawlor).
cognition and to explore possible causal mechanisms underlying this relationship.
Corresponding Author: David G.
Loughrey, BA(Hons), NEIL (Neuro
Enhancement for Independent Lives)
Programme, Trinity College Institute
of Neuroscience, Lloyd Building,
JAMA Otolaryngol Head Neck Surg. 2018;144(2):115-126. doi:10.1001/jamaoto.2017.2513 Trinity College Dublin, Room 3.10,
Published online December 7, 2017. Corrected on January 18, 2018. Dublin 2, Ireland ([email protected]).

(Reprinted) 115
© 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 04/18/2023

 Research Original Investigation Age-Related Hearing Loss, Cognitive Function, and Dementia

D
ementia affects an estimated 46.8 million persons
worldwide and is projected to affect approximately Key Points
131.5 million in 2050 with an estimated cost of US $818
Question Is age-related hearing loss associated with an increased
billion in 2015 and US $2 trillion by 2050.1 Current pharma- risk for cognitive decline, cognitive impairment, and dementia?
ceutical approaches targeting neuropathologic processes such
Findings In this systematic review and meta-analysis of 36
as Alzheimer disease (AD) offer limited benefit with symptom-
epidemiologic studies and 20 264 unique participants, age-related
modifying effects at best.2 Switching to a preventive strategy
hearing loss was significantly associated with decline in all main
through reduction of risk factors may be more beneficial than cognitive domains and with increased risk for cognitive
pharmacologic therapy after clinical expression of neuropatho- impairment and incident dementia. Increased risks for Alzheimer
logic changes 3 and may lead to significant reductions in disease and vascular dementia were nonsignificant.
medical costs.4
Meaning Age-related hearing loss is a possible biomarker and
Approximately one-third of adults older than 65 years ex- modifiable risk factor for cognitive decline, cognitive impairment,
periences a disabling hearing loss.5 Cohort studies indicate that and dementia.
age-related hearing loss (ARHL) precedes the onset of clinical
dementia by 5 to 10 years, is a possible noninvasive bio-
marker, and may offer a pathway to modify clinical outcomes.6 pure-tone audiometric assessment) as the main exposure vari-
As an emerging risk factor, a limited number of studies have able; (6) full inclusion of hearing loss sample (ie, no pure-tone
examined ARHL and cognitive decline. Epidemiologic find- audiometric cutoff); (7) assessment of cognitive function, cog-
ings have been inconsistent possibly owing to suboptimal nitive impairment,14 and/or dementia as outcome(s); and (8)
methods (eg, self-reported hearing loss or cognitive tests with exposure and outcome measurements obtained by health care
auditory stimuli).7 Prior reviews8-10 have not included a meta- professionals or trained investigators (ie, not based on self-
analysis or have included different measures of hearing im- reported data).
pairment and studies of different designs. Studies published on or before August 26, 2015, were re-
We conducted a systematic review and meta-analysis to trieved from the following 4 electronic databases by one of us
investigate and quantify the association between ARHL and (D.G.L.): (1) PubMed, (2) the Cochrane Library, (3) EMBASE, and
cognitive function, cognitive impairment, and dementia. We (4) SCOPUS. Keywords included hearing, cognition, demen-
reduced conceptual heterogeneity by including only observa- tia, and Alzheimer disease (eTable 1 in the Supplement). Re-
tional cross-sectional and cohort studies that assessed hear- sults were updated on April 15, 2016. Cross-referencing for po-
ing loss using pure-tone audiometry (the criterion standard). tentially eligible studies was conducted using retrieved studies
We conducted exploratory subgroup and meta-regression and personal files belonging to one of us (D.G.L.).
analyses to examine possible explanations for heterogeneity
owing to demographic, study, health, and analysis factors. Data Extraction and Quality Assessment
Two of us (D.G.L. and M.E.K.) independently screened for eli-
gible studies and conducted data extraction. If consensus could
not be reached, another of us (B.A.L.) acted as arbitrator for
Methods study inclusion, and another (G.E.K.) was consulted regard-
This systematic review was performed according to an a priori ing data extraction. Cognitive function was subdivided into 10
established protocol. It adhered to the Primary Reporting Items domains, including episodic memory (delayed recall and im-
for Systematic Reviews and Meta-analyses (PRISMA) mediate recall), executive functions (attention, fluency, rea-
Statement11 and met the Meta-analysis of Observational Stud- soning, and working memory), global cognition, processing
ies in Epidemiology (MOOSE) guidelines.12 All analyses were speed, semantic memory, and visuospatial ability.15 Among de-
conducted using Comprehensive Meta-Analysis software (ver- mentia studies, a secondary outcome of interest was any data
sion 3; Biostat). Institutional review board approval and in- that examined subgroups (eg, AD).
formed consent were not required for this systematic review Data from the most recently published study were se-
and meta-analysis. lected. Data from different studies that examined the same co-
hort were included if they were for different cognitive out-
Search Strategy and Selection Criteria comes and were treated as separate studies in analysis. Priority
Six a priori meta-analyses were planned across 2 levels of study was given to outcomes that were maximally adjusted for co-
design (cross-sectional and cohort) and 3 levels of outcome (cog- variates. Two of us (D.G.L. and M.E.K.) independently as-
nitive function, cognitive impairment, and dementia). The in- sessed the quality of reporting for each study using the
clusion criteria consisted of (1) cross-sectional and cohort stud- Strengthening the Reporting of Observational Studies in Epi-
ies, excluding case-control studies because of greater concern demiology (STROBE) instrument.16 With use of the Cohen κ
about sampling and retrospective analysis bias13 (all study de- coefficient,17 agreement was excellent (κ = 0.91) before cor-
signs have selected types of bias); (2) published studies (any lan- recting discrepant items.
guage); (3) study sample 18 years or older; (4) baseline sample
including the general, community-dwelling population rather Statistical Analysis
than special risk groups (eg, patients with coronary heart dis- We chose the Pearson r correlation coefficient as the effect size
ease); (5) individual’s peripheral hearing status (as assessed by of the linear association between hearing loss and cognitive

116 JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 (Reprinted) jamaotolaryngology.com

© 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 04/18/2023

 Age-Related Hearing Loss, Cognitive Function, and Dementia Original Investigation Research

function (continuous variables). Negative scores indicated that

Figure 1. PRISMA Flow Diagram
greater hearing loss was associated with poorer cognition. Odds
ratios (ORs) were chosen for cognitive impairment and de- 1824 Records identified through 57 Additional records identified
mentia (categorical variables). Influence of various audiom- database search through other sources
1273 SCOPUS
etric criteria (eg, worse vs better ear) and cognitive tests (vi- 504 PubMed
sual vs auditory stimuli) on outcome were examined in 32 Cochrane Library
subgroup analyses. If the required outcome metric was not re- 15 EMBASE

ported in the study, values were calculated using available data.

Random-effects, method-of-moments models that incorpo-
1185 Records screened after duplicates removed
rate heterogeneity into the overall estimate were used to pool
effect sizes from each study.18 All outcomes were converted 1030 Records excluded
to Fisher z values or logarithm ORs for analysis purposes and 888 Off topic
105 Inappropriate design,
then converted back to the original metric (ie, r correlation co- outcomes, or population
efficient and OR, respectively). For both meta-analyses of cog- 37 Review
nitive function, multiple tests of the same cognitive domain
from the same study were collapsed into a single effect size 155 Full-text articles assessed for eligibility
and within-study subgroups were analyzed independently as
separate effect sizes. 115 Full-text articles excluded
43 No audiometry
Heterogeneity was examined using the Q test, and P ≤ .10 31 Case-control
was considered to be statistically significant.19 Inconsistency 19 Inappropriate sample
12 Inappropriate audiometric
was examined using the I 2 statistic, and the following criteria
grades were applied: less than 25% indicated very low; 25% 7 No statistical analysis
2 Inappropriate design
to less than 50%, low; 50% to less than 75%, moderate; and
1 Audiometric assessment
75% or greater, large.19 Small-study effects were examined not specified
using funnel plots, and the regression-intercept approach of
Egger and colleagues20 provided at least 10 effect sizes were 40 Studies included in qualitative synthesis
present. To examine the influence of each result on the
overall findings, outcomes were analyzed by deleting each 36 Studies included in quantitative synthesis (meta-analysis)
26 Cross-sectional cognition
study from the model once. Cumulative meta-analysis 9 Cohort cognition
ranked by year was used to examine the accumulation of 5 Cross-sectional impairment
3 Cohort impairment
evidence over time.21 3 Cohort dementia
We conducted subgroup and meta-regression analyses to 2 Cross-sectional dementia
examine heterogeneity between studies. Planned variables in-
cluded (1) study characteristics, (2) participant characteristics, Study selection for the meta-analysis. Some studies were allocated to more
(3) audiometric factors, (4) cognitive measures, and (5) statis- than 1 category.
tical analysis (eTable 2 in the Supplement provides a list of each
planned variable). For continuous variables, we used random-
effects meta-regression22 where at least 4 effect sizes were sults and eFigures 28 to 74 in the Supplement. Thirty-five of
found. For categorical variables, we examined between-group the 40 included studies (88%) met the criteria for at least 16
differences (between-group Q value) in effect sizes using mixed of 22 STROBE items. Further details on the main analyses and
effects analysis of variance–like models for meta-analysis22 if descriptions of the small-study, influence, and cumulative
at least 3 effect sizes were available for each category. These analyses are found in eFigures 2 to 27 and eTables 3 to 9 in the
analyses were considered to be exploratory. Supplement. Diagnostic criteria for each clinical outcome are
shown in the Table 1 and Table 2.
Twenty-six studies with 15 620 participants were in-
cluded in the cross-sectional cognitive function analysis.23-48
Results Two studies were omitted because of duplicate data.49,50 Nine
Characteristics and Quality of Included Studies studies with 8233 participants7,29,36,40,46,48,51-53 were in-
The characteristics of included studies are shown in eTable 3 cluded in the cohort cognitive function analysis with a fol-
in the Supplement. Of the 1185 citations reviewed, 40 low-up ranging from 2 to 23 years (mean [SD], 10.4 [6.7] years).
studies7,23-61 met the inclusion criteria, representing 34 471 par- Five studies with 6582 participants (797 cases of 6553
ticipants from 12 countries (Figure 1). Of these, 36 unique stud- included participants [12.2%])30,55-58 were included in the
ies with an estimated 20 264 unique participants were in- cross-sectional cognitive impairment analysis. Two studies
cluded in the meta-analyses. Study quality results are shown were omitted because of duplicate data.54,59 Three studies with
in Table 1 and Table 2 and eFigure 1 in the Supplement. Fur- 7817 participants (1395 cases of 6825 included participants
ther details on the main analyses are found in eFigures 2 to 27 [20.4%]) were included in the cohort cognitive impairment
and eTables 3 to 9 in the Supplement; and further details on analysis with a follow-up ranging from 6 to 18 years (mean [SD],
the small-study, influence, and cumulative analyses, in the eRe- 11.7 [6.0] years).7,40,55

jamaotolaryngology.com (Reprinted) JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 117

© 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 04/18/2023

 118
Table 1. Characteristics of Cross-sectional Studies
No. of Clinical
Partici- Age, Mean (SD) Audiometric Outcomes STROBE
Source Country Study Name pants or Range, y Female, % Assessmenta Cognitive Domains Assessed (Criteria) Covariates Scoreb
Anstey,23 1999 Australia NA 180 70.56 (7.13) 100 2, 4, and 8 kHz/both Attention, processing speed None Age, grip strength, forced 14
ears expiratory volume,
vibration sense, and
vision
Anstey and Australia NA 180 70.56 (7.13) 100 2, 4, and 8 kHz/both Processing speed, reasoning, None Age 17
Smith,24 1999 ears semantic memory, visuospatial
Research Original Investigation

ability, working memory

Anstey et al,25 Australia ALSA 894 77.7 (5.6) 49 0.5, 1, 2, 3, and 4 Immediate recall, processing None None 17
2001 kHz/both ears speed, semantic memory
Baltes and Germany BASE and 315 64.9 (22) NA 0.25, 0.5, 1, 2, 3, 4, 6, Fluency, global cognition, None Age 17
Lindenberger,26 young adult and 8 kHz/both ears immediate recall, processing
1997 sample speed, reasoning, semantic

Downloaded From: https://jamanetwork.com/ on 04/18/2023

memory
Bucks et al,27 Australia BHAS 1969 56.2 (5.5) 53.8 0.5, 1, 2, and 4 Attention, delayed recall, None Age, sex, educational 21
2016 kHz/better ear fluency, processing speed, attainment, depression,
working memory and premorbid IQ
Clark28 1960 United States NA 102 20-70 Approximately 3 kHz/both ears Attention, immediate recall, None None 11
50 processing speed, reasoning,
visuospatial ability
Deal et al,48 2015 United States ARIC 253 56.6 (5.3) 60.9 >25 dB, >40 dB; 0.5, 1, Attention, delayed recall, None Age, sex, educational 22
2, and 4 kHz/better ear fluency, global cognition, attainment, smoking,
processing speed, semantic hypertension, diabetes,
memory premorbid IQ, and
depression
Deal et al,29 2017 United States HABC 1889 75.5 (3) 52.73 0.5, 1, 2, and 4 Immediate recall, processing Dementia Age, sex, race, 22
kHz/better ear speed (diagnosis, educational attainment,
medication use study site, smoking,
or hypertension, diabetes,
race-stratified and stroke
3MS decline

JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 (Reprinted)
more than 1.5
SDs from the
baseline mean)
Dupuis et al,30 Canada NA 301 71.13 (7.4) 64 >25 dB; 0.5, 1, and 2 Global cognition Cognitive None 20
2015 kHz/worse ear impairment
(MoCA)

© 2017 American Medical Association. All rights reserved.

Era et al,31 1986 Finland NA 547 31-35, 51-55, 0 0.5, 1, and 2 kHz; PTT Fluency, reasoning, visuospatial None None 17
71-75 4 kHz/better ear ability, working memory
Gussekloo The Netherlands Leiden 85+ 459 85 (0) 66 1, 2, and 4 kHz/both Attention, delayed recall, global None Sex and educational 17
et al,32 2005 study ears cognition, immediate recall, attainment
processing speed
Harrison Bush United States SKILL 894 73.47 (6) 57.8 0.5, 1, and 2 Attention, global cognition, None Age, sex, educational 21
et al,33 2015 kHz/better ear immediate recall, processing attainment, race,
speed, working memory diabetes, heart disease,
hypertension, stroke, and
depression

(continued)
Age-Related Hearing Loss, Cognitive Function, and Dementia

jamaotolaryngology.com
 Table 1. Characteristics of Cross-sectional Studies (continued)
No. of Clinical
Partici- Age, Mean (SD) Audiometric Outcomes STROBE
Source Country Study Name pants or Range, y Female, % Assessmenta Cognitive Domains Assessed (Criteria) Covariates Scoreb
Helzner et al,49 United States HABC 2052 77.5 (2.8) 52.7 >25 dB; 0.5, 1, and 2 Global cognition None Age, sex, educational 19
2005 kHz/worse ear attainment, household
income, study site, blood

jamaotolaryngology.com
pressure, diabetes, CVD,
cerebrovascular disease,
hip bone mineral density,
history of ear surgery,
alcohol use, smoking,
walking calorie
expenditure, ototoxic
medication use, and
occupational noise

Downloaded From: https://jamanetwork.com/ on 04/18/2023

exposure
Herbst and United Kingdom NA 253 ≥70 64 ≥35 dB; 1, 2, and 4 None Dementia None 14
Humphrey,60 kHz/better ear (CARE)
1980
Age-Related Hearing Loss, Cognitive Function, and Dementia

Heron and United Kingdom NA 540 20-79 44.44 1 kHz/both ears Attention, immediate recall, None None 22
Chown,34 1967 processing speed, reasoning,
semantic memory
Hofer et al,35 Denmark, Finland, NORA 1041 75 (0) 57.26 0.25 kHz, 0.5, 1, and 2 Fluency, immediate recall, None None 18
2003 and Sweden kHz; 4 and 8 kHz/both processing speed, reasoning,
ears working memory
Hong et al,36 Australia BMES 2334 >49 NA >40 dB; 0.5, 1, 2, and Global cognition None None 20
2016 4 kHz /worse and
better ear
Karpa et al,54 Australia BMHS 2815 66.6 (9.3) 56.7 >25 dB; 0.5, 1, 2, and None Cognitive None 20
2010 4 kHz/better ear impairment
(MMSE)
Kiely et al,55 2012 Australia ALSA and 4221 73.6 (8.9) 53.7 0.5, 1, 2, and 4 None Cognitive Age, years in study, sex, 20
BMES kHz/better ear impairment educational attainment,
(MMSE) diabetes, stroke,
hypertension, workplace
noise exposure, and
high-frequency
audiometric noise
notches

© 2017 American Medical Association. All rights reserved.

Kurniawan et al,56 The Netherlands Leiden 85+ 435 85 (0) 66.7 >35 dB; 1, 2, and 4 None Cognitive None 19
2012 study kHz/better ear impairment
(MMSE)
Li et al,37 1998 Germany NA 179 30-51 51.96 0.25, 0.5, 1, 2, 3, 4, 6, Fluency, immediate recall, global None Age 17
and 8 kHz/both ears cognition, processing speed,
reasoning, semantic memory
Lin,38 2011 United States NHANES 605 64.1 (2.9) 52.9 0.5, 1, and 2, 4 Processing speed None Age, sex, hearing aid, 20
kHz/better ear income, educational
attainment, race, and
CVD risk factors
(diabetes, hypertension,
smoking, and stroke)

(continued)
Original Investigation Research

(Reprinted) JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2
119
 120
Table 1. Characteristics of Cross-sectional Studies (continued)
No. of Clinical
Partici- Age, Mean (SD) Audiometric Outcomes STROBE
Source Country Study Name pants or Range, y Female, % Assessmenta Cognitive Domains Assessed (Criteria) Covariates Scoreb
39
Lin et al, 2011 United States BLSA 347 71 (7.2) 35.2 0.5, 1, 2, and 4 Attention, fluency, global None Age, sex, race, 19
kHz/better ear cognition, immediate recall, educational attainment,
processing speed, semantic diabetes, smoking, and
memory hypertension
Lin et al,40 2013 United States HABC 1984 77.4 (2.76) 52.1 >25 dB; 0.5, 1, 2, and Global cognition, processing Cognitive Age, sex, educational 20
4 kHz/better ear speed impairment attainment,
Research Original Investigation

(3MS score <80 race/ethnicity, study site,

or decline >5 hypertension, diabetes,
from baseline) smoking, and stroke
Lindenberger and Germany BASE 156 84.9 (9) 50 0.25, 0.5, 1, 2, 3, 4, 6, Global cognition None Age and vision 18
Baltes,41 1994 and 8 kHz/both ears
Lindenberger and Germany BASE 516 84.9 (8.7) 50 0.25, 0.5, 1, 2, 3, 4, 6, Global cognition None Age 17

Downloaded From: https://jamanetwork.com/ on 04/18/2023

Baltes,50 1997 and 8 kHz/both ears
López-Torres Spain NA 1161 73.3 (5.9) 55.9 ≥40 dB; 1 and 2 None Cognitive None 20
Hidalgo et al,57 kHz/either ear; ≥40 dB; impairment
2009 1 or 2 kHz/both ears (SPMSQ)
MacDonald et Australia VLS 125 78.9 (3.12) 61.6 0.5, 1, and 2 kHz/both Immediate recall, processing None None 18
al,42 2004 ears speed, reasoning, semantic
memory, working memory
Quaranta et al,58 Italy GA 488 72.8 (6.2) 39.3 >35 dB; 0.5, 1, and 2 None Cognitive Age, sex, and educational 17
2014 kHz/both ears impairment attainment
(Neuropsycho-
logical
assessment14
Dementia
(DSM-5)
Schaie et al,43 United States NA 47 76.4 (NA) 48.9 0.128, 0.256, 0.512, Global cognition None Age 15
1964 1.024, 2.048, 4.096,
and 8.192 kHz/both
ears

JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 (Reprinted)
Sugawara et al,44 Japan NA 846 63.9 (8.3) 63.4 >25 dB; 0.5, 1, and 2 Global cognition None Age, sex, and educational 18
2011 kHz/better ear attainment
Tay et al,59 2006 Australia BMES 3509 66.7 (NA) 57 >40 dB; 0.5, 1, 2, and None Cognitive Age, sex, educational 22
4 kHz/better ear impairment attainment, and history
(MMSE) of stroke

© 2017 American Medical Association. All rights reserved.

Thomas et al,45 United States NA 259 72 (NA) 54 0.5, 1, and 2 Delayed recall, global cognition, None None 13
1983 kHz/better ear reasoning, working memory
Valentijn et al,46 The Netherlands MAAS 391 65.1 (6.6) 48.6 1, 2, and 4 kHz/better Attention, delayed recall, None None (cross-sectional) 20
2005 ear fluency, immediate recall, Age, sex, educational
processing speed attainment, and baseline
hearing and cognitive
function
van Boxtel et al,47 The Netherlands MAAS 453 51.4 (16.5) 50.8 1, 2, and 4 KHz/better Delayed recall, immediate recall, None Age, sex, and educational 18
2000 ear processing speed attainment; processing
speed (delayed and
immediate recall only)
Abbreviations: ALSA, Australian Longitudinal Study of Aging; ARIC, Atherosclerosis Risk in Communities neurocognitive study; BASE, Berlin Aging Study; BHAS, Busselton Healthy Aging Study; BLSA, Baltimore Longitudinal Study of
Aging; BMES, Blue Mountains Eye Study; BMHS, Blue Mountains Hearing Study; CARE, Comprehensive Assessment and Referral Evaluation; CVD, cardiovascular disease; DSM-5, Diagnostic and Statistical Manual of Mental Disorders
(Fifth Edition); GA, Great Age study; HABC, Health, Aging and Body Composition study; MAAS, Maastricht Aging Study; MoCA, Montreal Cognitive Assessment; MMSE, Mini-Mental State Examination; NA, not applicable; NHANES,
National Health and Nutritional Examination Survey; NORA, Nordic Research on Aging study; PTT, pure-tone threshold; SKILL, Staying Keen in Later Life study; SPMSQ, Short Portable Mental Status Questionnaire; STROBE,
Age-Related Hearing Loss, Cognitive Function, and Dementia

jamaotolaryngology.com
Strengthening the Reporting of Observational Studies in Epidemiology; VLS, Victoria Longitudinal Study; 3MS, Modified Mini-Mental State Examination.
a
Expressed as frequencies and ears assessed in pure-tone audiometry. Cutoff decibel level is included where applied to determine case of hearing loss.
b
Scores range from 0 to 22, with higher scores indicating better study quality.
 Table 2. Characteristics of Cohort Studies
No. of
Source (Length Partici- Age, Mean (SD) or Audiometric STROBE
of Study, y) Country Study Name pants Range, y Female, % Assessmenta Cognitive Domains Assessed Clinical Outcomes (Criteria) Covariates Scoreb
Anstey et al,51 Australia ALSA 2087 ≥65 49.4 0.5, 1, and 2 kHz Immediate recall, None Age 15
2001 (2) or 3 and 4 kHz or 6 processing speed, semantic
and 8 kHz/both memory

jamaotolaryngology.com
ears
Anstey et al,52 Australia ALSA 1823 77.77 (6.56) 48.8 PTA of lesser PTT Immediate recall, None Age, sex, educational attainment, 20
2003 (8) at 2, 3, and 4 kHz processing speed, semantic depression, self-rated health, and
in either ear memory number of medical conditions
Deal et al,48 United States ARIC 253 56.6 (5.3) 60.9 >25 dB, >40 dB; Attention, delayed recall, None Age, sex, educational attainment, 22
2015 (23) 0.5, 1, 2, and 4 fluency, global cognition, smoking, hypertension, diabetes,
kHz/better ear processing speed, semantic premorbid IQ, and depression
memory
Deal et al,29 United States HABC 1889 75.5 (3) 52.73 0.5, 1, 2, and 4 Immediate recall, Dementia (diagnosis, Age, sex, race, educational 22

Downloaded From: https://jamanetwork.com/ on 04/18/2023

2017 (9) kHz/better ear processing speed medication use or attainment, study site, smoking,
race-stratified 3MS decline hypertension, diabetes, and
more than 1.5 SDs from the stroke
baseline mean)
Age-Related Hearing Loss, Cognitive Function, and Dementia

Gallacher et al,7 United Kingdom CaPS 1057 56.1 (4.4) 0 0.5, 1, 2, and 4 Delayed recall, global Cognitive impairment Age, social class, anxiety, 21
2012 (17) kHz/both ears cognition, immediate recall, (NINCDS-AIREN, DSM-IV, baseline cognitive function
processing speed, reasoning and no functional (cognitive function only), and
impairment); dememtia premorbid IQ (clinical outcomes
(DSM-IV or NINCDS-AIREN); only)
AD (DSM-IV, most met
criteria for
NINCDS-ADRDA); vascular
dementia (NINCDS-AIREN)
Hong et al,36 Australia BMES 2334 >49 NA >40 dB; 0.5, 1, 2, Global cognition None Age and sex 20
2016 (10) and 4 kHz /worse
and better ear
Kiely et al,55 Australia ALSA and 4221 73.6 (8.9) 53.7 0.5, 1, 2, and 4 None Cognitive impairment Age, years in study, sex, 20
2012 (11) BMES kHz/better ear (MMSE) educational attainment, diabetes,
stroke, hypertension, workplace
noise exposure, and
high-frequency audiometric
noise notches
Lin et al,61 United States BLSA 639 36-90 43.7 0.5, 1, 2, and 4 None Dementia (DSM-III), AD Age, sex, race, educational 21
2011 (18) kHz/better ear (NINCDS-ADRDA) attainment, diabetes, smoking,
hypertension, and baseline

© 2017 American Medical Association. All rights reserved.

cognitive function
Lin et al,40 United States HABC 1984 77.4 (2.76) 52.1 >25 dB; 0.5, 1, 2, Global cognition, processing Cognitive impairment (3MS Age, sex, educational attainment, 20
2013 (6) and 4 kHz/better speed score <80 or decline >5 race/ethnicity, study site,
ear from baseline) hypertension, diabetes, smoking,
and stroke
Lindenberger Germany BASE 516 84.9 (8.7) 50 2, 3, 4, and 6 Fluency, immediate recall, None Age, time to death, and risk for 18
and Ghisletta,53 kHz/both ears processing speed dementia
2009 (13)
Valentijn et al,46 The MAAS 391 65.1 (6.6) 48.6 1, 2, and 4 Attention, delayed recall, None Age, sex, educational attainment, 20
2005 Netherlands kHz/better ear fluency, immediate recall, and baseline hearing and
(6) processing speed cognitive function
Abbreviations: ALSA, Australian Longitudinal Study of Aging; ARIC, Atherosclerosis Risk in Communities neurocognitive study; BASE, Berlin Aging Study; BLSA, Baltimore Longitudinal Study of Aging; BMES, Blue Mountains Eye Study;
CaPS, Caerphilly Prospective Study; DSM-III, Diagnostic and Statistical Manual of Mental Disorders (Third Edition); DSM-IV, Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition): HABC, Health, Aging and Body
Composition study; MAAS, Maastricht Aging Study; MMSE, Mini-Mental State Examination; NA, not applicable; NINCDS-ADRDA, National Institute of Neurological and Communicative Disorders and Stroke–Alzheimer Disease and
Related Disorders Association; NINCDS-AIREN, NINCDS–Association Internationale pour la Recherché et l’Enseignement en Neurosciences; PTA, pure-tone average; PTT, pure-tone threshold; STROBE, Strengthening the Reporting of
Original Investigation Research

(Reprinted) JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2
Observational Studies in Epidemiology; 3MS, Modified Mini-Mental State Examination.
a
Expressed as frequencies and ears assessed in pure-tone audiometry. Cutoff decibel level is included where applied to determine case of hearing loss.
b
Scores range from 0 to 22, with higher scores indicating better study quality.

121
 Research Original Investigation Age-Related Hearing Loss, Cognitive Function, and Dementia

Figure 2. Forest Plot of Correlations for Cognition Cross-sectional Outcomes

No. of Indicates Indicates

Outcome Participants/Events r Value (95% CI) Decline Improvement
Attention 5159/11 −0.16 (−0.24 to −0.07)
Delayed recall 3808/7 −0.10 (−0.16 to −0.04)
Fluency 4629/9 −0.08 (−0.12 to −0.04)
Global cognition 7702/15 −0.15 (−0.18 to −0.11)
Immediate recall 6747/15 −0.14 (−0.20 to −0.09)
Processing speed 10 660/20 −0.13 (−0.18 to −0.08)
Reasoning 3128/12 −0.18 (−0.25 to −0.10)
Semantic memory 2906/10 −0.14 (−0.20 to −0.08)
Visuospatial ability 669/5 −0.11 (−0.19 to −0.03)
Working memory 4855/9 −0.10 (−0.15 to −0.05)
Twenty-six studies were included in
Summary 15 620/113 −0.12 (−0.14 to −0.10)
the analysis.23-48 Squares represent
−0.50 −0.25 0 0.25 0.50 correlation (r value); different sizes of
r Value (95% CI) markers, weight; diamond, overall
correlation; and error bars, 95% CIs.

Figure 3. Forest Plot of Correlations for Cognition Cohort Outcomes

No. of Participants/ Indicates Indicates

Outcome Events r Value (95% CI) Decline Improvement
Attention 391/1 −0.10 (−0.20 to 0.00)
Delayed recall 1774/4 −0.10 (−0.15 to −0.05)
Fluency 1233/4 −0.07 (−0.14 to 0.01)
Global cognition 4227/6 −0.14 (−0.19 to −0.09)
Immediate recall 4225/6 −0.06 (−0.10 to −0.02)
Processing speed 6462/10 −0.08 (−0.14 to −0.03)
Reasoning 1057/1 −0.06 (−0.12 to 0.00)
Nine studies were included in the
Semantic memory 707/1 −0.14 (−0.23 to −0.05)
analysis.7,29,36,40,46,48,51-53 Squares
Summary 8233/33 −0.09 (−0.11 to −0.07) represent correlation (r value);
−0.50 −0.25 0 0.25 0.50 different sizes of markers, weight;
r Value (95% CI) diamond, overall correlation; and
error bars, 95% CIs.

Two studies with 741 participants (59 cases of 679 in- CI, –0.14 to 0.01). These results included global cognition
cluded participants [8.7%])58,60 were included in the cross- (r = −0.14; 95% CI, –0.19 to –0.09), executive functions (range,
sectional dementia analysis. One study assessed dementia (39 r = −0.06 [95% CI, –0.12 to –0.004] to r = −0.10 [95% CI, –0.20
cases of 245 included participants [15.9%]),60 and the other as- to –0.001]), episodic memory (range, r = −0.06 [95% CI, –0.10
sessed AD (20 cases of 434 included participants [4.6%]).58 to –0.02] to r = −0.10 [95% CI, –0.15 to –0.05), processing speed
Three studies with 3585 participants7,29,61 (10.4%) were in- (r = −0.08; 95% CI, –0.14 to –0.03), and semantic memory
cluded in the cohort dementia analysis with a follow-up length (r = −0.14; 95% CI, –0.23 to –0.05) (Figure 2, Figure 3, and eFig-
ranging from 9 to 18 years (mean [SD], 15.0 [5.2] years). All 3 ures 10-27 and eTables 1-9 in the Supplement). No cohort data
studies reported incident dementia outcomes (366 cases of were available for visuospatial ability or working memory.
3439 included participants [10.6%]), 2 examined an AD sub- Heterogeneity was significant in most domains (Q range, 0.0-
set (78 cases of 1491 included participants [5.2%]),7,61 and 1 ex- 79.9). Inconsistency ranged from very low to high.
amined a vascular dementia subset (38 cases of 870 included
participants [4.4%]).7 Hearing Loss and Cognitive Impairment
We found a statistically significant association between ARHL
Hearing Loss and Cognitive Function and cognitive impairment across cross-sectional (OR, 2.00; 95%
We found a small but statistically significant association be- CI, 1.39-2.89) and cohort studies (OR, 1.22; 95% CI, 1.09-1.36)
tween ARHL and all 10 cognitive domains of interest in cross- (eFigures 2 and 3 and eTable 7 in the Supplement). Statisti-
sectional studies, including global cognition (r = −0.15; 95% CI, cally significant heterogeneity (Q range, 0.1-23.7) and a large
−0.18 to −0.11), executive functions (range, r = −0.08 [95% CI, amount of inconsistency were observed in cross-sectional but
−0.12 to −0.04] to r = −0.18 [95% CI, −0.25 to −0.10), episodic not in cohort studies.
memory (range, r = −0.10 [95% CI, −0.16 to −0.04] to r = −0.14
[95% CI, –0.20 to –0.09]), processing speed (r = −0.13; 95% CI, Hearing Loss and Dementia
–0.18 to 0.08), semantic memory (r = −0.14; 95% CI, –0.20 to We found a significant association between ARHL and demen-
–0.08), and visuospatial ability (r = −0.11; 95% CI, –0.19 to –0.03). tia in cross-sectional (OR, 2.42; 95% CI, 1.24-4.72) and cohort
Similar results were observed in 7 of 8 domains in cohort stud- (OR, 1.28; 95% CI, 1.02-1.59) studies (eFigures 4-9 and eTable
ies, excluding fluency, which was not significant (r = −0.07; 95% 7 in the Supplement). Statistically significant heterogeneity (Q

122 JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 (Reprinted) jamaotolaryngology.com

© 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 04/18/2023

 Age-Related Hearing Loss, Cognitive Function, and Dementia Original Investigation Research

range, 0.4-6.6) and a moderate amount of inconsistency were cross-sectional attention and immediate recall. The propor-
observed in cohort but not cross-sectional studies. No statis- tion of individuals diagnosed with hearing loss by study au-
tically significant association was found between ARHL and thors weakened the association with immediate recall. Re-
AD for cross-sectional (OR, 1.80; 95% CI, 0.58-5.60) or cohort sults were otherwise mixed and nonsignificant.
(OR, 1.69; 95% CI, 0.72-4.00) studies. In addition, the asso-
ciation between ARHL and vascular dementia was not signifi- Cognitive Measures
cant (OR, 2.40; 95% CI, 0.99-5.82). Results were mostly minor and inconsistent with respect to
whether the cognitive test was accessible to a sample with hear-
Subgroup Analyses and Meta-regression ing loss. The only significant result found a stronger associa-
The results of the subgroup and meta-regression analyses for tion for nonbiased tests.
cognitive outcomes are summarized below (eTables 8 and 9
in the Supplement). The respective Fisher z values (modera- Statistical Analysis
tor analysis), slope (meta-regression), SEs, and 95% CIs for each A stronger association was generally found for studies that used
variable are available in eTables 10 to 36 in the Supplement. correlation as the statistical model (compared with linear re-
gression or linear mixed models) and those that reported re-
Study Characteristics sults as significant. Studies that used age, sex, race, educa-
Studies conducted in the United States reported weaker asso- tional attainment, and vascular factors as covariates in their
ciations between ARHL and cognition compared with Austra- analysis typically reported weaker (sometimes significantly
lian and European studies, possibly owing to differences in weaker) associations. This same trend was observed for stud-
prevalence of ARHL or cognitive decline and dementia. Asso- ies that controlled for stroke, hypertension, diabetes, and cur-
ciations generally became weaker with later publication dates rent or previous smoking. Controlling for depression signifi-
(possibly owing to increased adjustment for covariates) and, cantly weakened the association with cross-sectional attention.
in some cases, with higher STROBE score. Results for journal Results for premorbid IQ were mixed and nonsignificant ex-
impact factor were mixed. Among cohort studies, results for cept for cohort global cognition.
length of follow-up were mostly insignificant. Because of a lack of data, no other a priori variables were
examined. Other variables were reviewed ad hoc. A signifi-
Participant Characteristics cantly weaker association was generally found for analyses that
Cross-sectional associations were weaker when studies ex- controlled for study site. These analyses were not conducted
cluded participants with cognitive impairment and dementia for cognitive impairment and dementia outcomes owing to lack
and included participants with cardiovascular risk. Associa- of studies, with the exception of cross-sectional cognitive im-
tions with cohort processing speed were mixed with regard to pairment studies. Year of publication, age (mean and mini-
whether participants with cognitive impairment were re- mum), sex, sample degree of hearing loss, proportion with
moved at baseline or in analysis. The age and sex of the sample hearing loss and cognitive impairment, impact factor, and
generally had mixed results. Associations were weaker for stud- STROBE were assessed (eTable 37 in the Supplement). No as-
ies with mixed-race participants compared with studies in sociation was statistically significant.
which the breakdown by race was not declared. Associations
were typically stronger with an increased proportion of white
participants but weaker with black participants and nonsig-
nificant for those of other races, possibly owing to selective
Discussion
survival. Associations were also typically stronger with an in- In this meta-analysis, ARHL had significant associations with ac-
creased proportion of primary educational attainment, weaker celerated multidomain cognitive decline, cognitive impairment,
with tertiary educational attainment, and mixed with second- and dementia, thus supporting further consideration of ARHL
ary educational attainment and mean years of education. as a risk factor for these outcomes.3,6 The associations, although
Smoking (current and previous) had a significant association. small, were comparable in size and significance with other more
commonly researched risk factors using meta-analysis.3
Audiometric Factors The result for AD indicated increased risk with ARHL but
Stronger associations were usually found for lower- was nonsignificant, most likely owing to small sample sizes or
frequency hearing loss (<4 kHz) and when auditory function to causal factors other than AD etiology underpinning the
was assessed with both ears (compared with only the better association.6 Age-related hearing loss has been associated with
ear). No significant difference was found for hearing loss ex- increased global and regional gray matter atrophy and white
amined as a categorical (>25 dB) vs a continuous variable. matter hyperintensities, whereas AD substrate has been found
Weaker associations were generally found when studies used in the auditory neural regions but not in the peripheral audi-
a sound-treated room or booth or followed the World Health tory structures.63
Organization criteria.62 Declared inclusion of hearing aid us- Study quality assessment showed that reporting was gen-
ers weakened the association for immediate recall and seman- erally of very good quality. Poor reporting of attrition rates may
tic memory. However, the proportion of hearing aid users in- conceal a greater decline in cognition and risk for dementia in
cluded in the study had no significant result. The sample degree older cohorts owing to higher numbers of dropouts among
of hearing loss significantly weakened the association with those with poorer health. Subgroup analysis found no bias for

jamaotolaryngology.com (Reprinted) JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 123

© 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 04/18/2023

 Research Original Investigation Age-Related Hearing Loss, Cognitive Function, and Dementia

verbal or audio cognitive tests. However, some potential bias and vascular dementia.75 In patients with ARHL, these do-
may have existed because a stronger effect size was found with mains may benefit from improved verbal communication
substandard audiometric assessment. through use of hearing aids. Additional randomized clinical trials
exploring the cognitive benefits of hearing loss treatment are
Causal Mechanisms for ARHL and Cognitive Decline required, as is more research as to whether treatment, alone or
The association between ARHL and cognitive decline re- as part of a wider approach to risk factors, modifies dementia
mains unclear.61 One hypothesis is a common etiology, such outcomes. Neuroimaging studies could examine modification
as decline in the vascular system or a broader physiological de- of cortical changes and neurocognitive compensation with hear-
cline. Age-related hearing loss has been linked with multiple ing aid use in speech tasks. Future epidemiologic research might
indicators of functional decline and is a biomarker for frailty assess whether ARHL is associated with cognitive decline in-
syndrome, which has been causally linked to dementia.64 Other dependently of neuropathologic hallmarks of dementia and
hypotheses suggest that the association may be mechanistic, whether a mediator of this association exists (eg, loneliness).
for example, ARHL causing cognitive decline through im- Also of interest would be whether cognitive reserve moderates
paired speech perception.61 cognitive decline in the population with ARHL. Our results in-
Vascular risk factors contributed significantly to decline dicated a moderator effect of educational attainment, which is
in global cognition and processing speed. However, the pooled often used as a proxy for cognitive reserve.76
effect size of studies controlling for vascular risk factors in these Increasing evidence suggests that ARHL is associated with
outcomes remained significant, suggesting other contribut- a wide range of health issues, higher disease burden, and in-
ing factors, for example, depression, which significantly mod- creased risk for hospitalization,64,77 leading to greater aware-
erated the association with attention. ness of this condition as a critical public health concern.70,77
Of interest, the pattern of decline observed in this study was In the United States, only 1 in 5 adults with hearing loss wears
consistent with estimated cognitive outcomes based on behav- hearing aids, possibly owing to cost, lack of insurance cover-
ioral and neuroimaging research.65 This research reports in- age, or lack of knowledge of health care options, particularly
creased recruitment of short-term memory and executive func- for milder loss.70 The National Academies of Sciences, Engi-
tions to aid speech perception after acquired hearing loss and neering, and Medicine recently outlined several recommen-
concomitant decline in auditory cortex regions.66 This situa- dations to address this issue, with implications for public health
tion is estimated to lead to less decline in these functions but services and policy.70,78 Initiatives to expand access to treat-
greater decline in episodic and semantic long-term memory ow- ment through screening programs, expand delivery of hear-
ing to reallocation of cognitive resources.65 Consistent with this ing services, and provide coverage for assistive hearing de-
research, we observed that hearing loss was less associated with vices would be beneficial.70 In addition, primary health care
decline in executive functions and immediate recall compared clinicians would benefit from standard guidelines for screen-
with delayed and semantic memory and was increasingly less ing and referring patients with hearing loss.70
predictive of decline in attention and immediate recall among
those with greater hearing loss. In addition, semantic memory, Strengths and Limitations
usually maintained in older age compared with episodic To the our knowledge, this study is the first systematic re-
memory,67 demonstrated a decline similar to that of episodic view and meta-analysis of ARHL and cognitive decline that
memory. Furthermore, the results indicated that hearing aids used only pure-tone thresholds as the audiometric criteria. Our
may benefit short-term and semantic memory. strict inclusion criteria in study design and measurement al-
The stronger association for low- to middle-frequency hear- lowed us to reduce conceptual heterogeneity and thus pro-
ing loss with immediate recall and processing speed may be vide the most accurate quantitative measure of this associa-
attributable to advanced aging as ARHL progresses from high tion. Considerable heterogeneity remained across most
to low frequencies.64 Of interest, vascular dysfunction has been outcomes. However, in any adjusted estimate of effect size for
associated with lower-frequency hearing loss and white mat- risk factors derived from aging studies, residual confounding
ter hyperintensities.68 Alternatively, reallocation of execu- will exist. Extensive subgroup and meta-regression analyses
tive functions to support accuracy in speech perception may investigating this heterogeneity provided insight into how fu-
be associated with decline in performance speed, as also ob- ture studies may reduce bias and explore the potential basis
served in older adults with visual processing deficits.69 of this association in experimental and clinical trials.
This study has several limitations. We could not examine
Future Directions whether studies controlled for etiology of hearing loss (eg, con-
Cognitive decline is influenced by multiple modifiable health genital or prelingual deafness). However, because of the low
factors.3 Hearing loss may be another serviceable risk factor, be- prevalence (<2%) of hearing loss in patients younger than 40
cause it is easily diagnosed and can be treated.70 Although as- years, this prevalence was considered to be insignificant.77
sociations were small, treatment may cumulatively benefit cog- Some of the meta-analyses had a low number of effect sizes.
nition as observed in intervention studies in older adults without We could not examine other planned moderators and covar-
cognitive impairment.71-73 This benefit was not observed in pa- iates, such as attrition, owing to lack of data. For meta-
tients with dementia, but treatment may still reduce disability.74 analyses of dementia subgroups, the number of cases was
Decline in lexical or semantic, episodic memory, and execu- small. Furthermore, because the meta-analyses were of ob-
tive functions is used by clinicians as a marker for probable AD servational studies, support for any inferences regarding the

124 JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 (Reprinted) jamaotolaryngology.com

© 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 04/18/2023

 Age-Related Hearing Loss, Cognitive Function, and Dementia Original Investigation Research

causal nature of the association is limited and cannot provide

direct evidence for policy recommendations. However, our Conclusions
analyses of prospective studies give an indication of the tem-
poral order of the association consistent with a causal effect. Age-related hearing loss is a potential risk factor for cognitive
Further research is required to determine whether a causal re- decline, cognitive impairment, and dementia. The effect sizes
lationship exists. Owing to the large number of statistical tests for all 3 main outcomes were small, but they compared with
conducted, some of our findings could have been the result meta-analytic estimates for other risk factors more com-
of chance. However, we did not want to risk missing poten- monly investigated in this population. Additional research, par-
tially important findings that could be tested in future origi- ticularly randomized clinical trials, is warranted to examine
nal studies. Finally, as is the case with any aggregate data meta- cognitive implications of treatment and to explore the pos-
analysis, the potential for ecological fallacy exists. sible causal mechanisms underlying this relationship.

ARTICLE INFORMATION 2. Thies W, Bleiler L; Alzheimer’s Association. 2013 16. Vandenbroucke JP, von Elm E, Altman DG, et al;
Accepted for Publication: October 5, 2017. Alzheimer’s disease facts and figures. Alzheimers STROBE Initiative. Strengthening the Reporting of
Dement. 2013;9(2):208-245. Observational Studies in Epidemiology (STROBE):
Published Online: December 7, 2017. explanation and elaboration. Int J Surg. 2014;12(12):
doi:10.1001/jamaoto.2017.2513 3. Barnes DE, Yaffe K. The projected effect of risk
factor reduction on Alzheimer’s disease prevalence. 1500-1524.
Correction: This article was corrected on January Lancet Neurol. 2011;10(9):819-828. 17. Cohen J. Weighted kappa: nominal scale
18, 2018, to correct an error in Figure 1, a rounding agreement with provision for scaled disagreement
error in Figure 2, and errors in the numbers of 4. Lin PJ, Yang Z, Fillit HM, Cohen JT, Neumann PJ.
Unintended benefits: the potential economic or partial credit. Psychol Bull. 1968;70(4):213-220.
participants and events in Figure 3.
impact of addressing risk factors to prevent 18. DerSimonian R, Laird N. Meta-analysis in clinical
Author Contributions: Mr Loughrey had full access Alzheimer’s disease. Health Aff (Millwood). 2014;33 trials. Control Clin Trials. 1986;7(3):177-188.
to all the data in the study and takes responsibility (4):547-554.
for the integrity of the data and the accuracy of the 19. Higgins JP, Thompson SG, Deeks JJ, Altman DG.
data analysis. 5. World Health Organisation. Deafness and Measuring inconsistency in meta-analyses. BMJ.
Study concept and design: All authors. hearing loss. http://www.who.int/mediacentre 2003;327(7414):557-560.
Acquisition, analysis, or interpretation of data: /factsheets/fs300/en/. Updated February 2017. 20. Egger M, Davey Smith G, Schneider M, Minder
Loughrey, Kelly, Kelley. Accessed August 27, 2016. C. Bias in meta-analysis detected by a simple,
Drafting of the manuscript: Loughrey, Kelly, Kelley, 6. Albers MW, Gilmore GC, Kaye J, et al. At the graphical test. BMJ. 1997;315(7109):629-634.
Brennan. interface of sensory and motor dysfunctions and 21. Lau J, Schmid CH, Chalmers TC. Cumulative
Critical revision of the manuscript for important Alzheimer’s disease. Alzheimers Dement. 2015;11(1): meta-analysis of clinical trials builds evidence for
intellectual content: All authors. 70-98. exemplary medical care. J Clin Epidemiol. 1995;48
Statistical analysis: Loughrey, Kelly, Kelley. 7. Gallacher J, Ilubaera V, Ben-Shlomo Y, et al. (1):45-57.
Obtained funding: Loughrey. Auditory threshold, phonologic demand, and
Administrative, technical, or material support: 22. Borenstein M, Hedges LV, Higgins JPT,
incident dementia. Neurology. 2012;79(15):1583- Rothstein HR. Introduction to Meta-Analysis. London,
Loughrey. 1590.
Study supervision: Loughrey, Brennan, Lawlor. England: Wiley; 2011.
8. Gennis V, Garry PJ, Haaland KY, Yeo RA, 23. Anstey KJ. Sensorimotor variables and forced
Conflict of Interest Disclosures: The authors have Goodwin JS. Hearing and cognition in the elderly:
completed and submitted the ICMJE Form for expiratory volume as correlates of speed, accuracy,
new findings and a review of the literature. Arch and variability in reaction time performance in late
Disclosure of Potential Conflicts of Interest, and Intern Med. 1991;151(11):2259-2264.
none were reported. adulthood. Neuropsychol Dev Cogn B Aging
9. Taljaard DS, Olaithe M, Brennan-Jones CG, Neuropsychol Cogn. 1999;6(2):84-95.
Funding/Support: This study was supported by Eikelboom RH, Bucks RS. The relationship between
DeafHear (Mr Loughrey), the Irish Research Council 24. Anstey KJ, Smith GA. Interrelationships among
hearing impairment and cognitive function: biological markers of aging, health, activity,
(Mr Loughrey), the Central Remedial Clinic (Mr a meta-analysis in adults. Clin Otolaryngol. 2016;41
Loughrey), in part by award U54GM104942 from acculturation, and cognitive performance in late
(6):718-729. adulthood. Psychol Aging. 1999;14(4):605-618.
the National Institute of General Medical Sciences
of the National Institutes of Health (Dr Kelley), and 10. Cherko M, Hickson L, Bhutta M. Auditory 25. Anstey KJ, Luszcz MA, Sanchez L.
in part by Atlantic Philanthropies (Dr Brennan). deprivation and health in the elderly. Maturitas. A reevaluation of the common factor theory of
2016;88:52-57. shared variance among age, sensory function, and
Role of the Funder/Sponsor: The sponsors had no
role in the design and conduct of the study; 11. Liberati A, Altman DG, Tetzlaff J, et al. The cognitive function in older adults. J Gerontol B
collection, management, analysis, and PRISMA statement for reporting systematic reviews Psychol Sci Soc Sci. 2001;56(1):3-11.
interpretation of the data; preparation, review, or and meta-analyses of studies that evaluate health 26. Baltes PB, Lindenberger U. Emergence of a
approval of the manuscript; and decision to submit care interventions: explanation and elaboration. powerful connection between sensory and
the manuscript for publication. J Clin Epidemiol. 2009;62(10):e1-e34. cognitive functions across the adult life span: a new
Disclaimer: The content is solely the responsibility 12. Stroup DF, Berlin JA, Morton SC, et al; window to the study of cognitive aging? Psychol
of the authors and does not necessarily represent Meta-analysis of Observational Studies in Aging. 1997;12(1):12-21.
the official views of the National Institutes of Epidemiology (MOOSE) Group. Meta-analysis of 27. Bucks RS, Dunlop PD, Taljaard DS, et al. Hearing
Health. Observational Studies in Epidemiology: a proposal loss and cognition in the Busselton Baby Boomer
for reporting. JAMA. 2000;283(15):2008-2012. cohort: an epidemiological study. Laryngoscope.
REFERENCES 13. Mann CJ. Observational research methods: 2016;126(10):2367-2375.
1. Prince M, Wimo A, Guerchet M, Ali G, Wu Y, Prina research design II: cohort, cross sectional, and 28. Clark JW. The aging dimension: a factorial
M. The global impact of dementia: an analysis of case-control studies. Emerg Med J. 2003;20(1):54- analysis of individual differences with age on
prevalence, incidence, cost and trends. World 60. psychological and physiological measurements.
Alzheimer Report 2015. https://www.alz.co.uk 14. Petersen RC. Mild cognitive impairment as a J Gerontol. 1960;15:183-187.
/research/WorldAlzheimerReport2015.pdf. 2015. diagnostic entity. J Intern Med. 2004;256(3):183-194. 29. Deal JA, Betz J, Yaffe K, et al. Hearing
Accessed August 27, 2016. 15. Lezak MD. Neuropsychological Assessment. New impairment and incident dementia and cognitive
York, NY: Oxford University Press; 2004.

jamaotolaryngology.com (Reprinted) JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 125

© 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 04/18/2023

 Research Original Investigation Age-Related Hearing Loss, Cognitive Function, and Dementia

decline in older adults: the Health ABC Study. 6-year follow-up in the Maastricht Aging Study. 62. World Health Organization. Grades of hearing
J Gerontol A Biol Sci Med Sci. 2017;72(5):703-709. J Am Geriatr Soc. 2005;53(3):374-380. impairment. 2016. http://www.who.int/pbd
30. Dupuis K, Pichora-Fuller MK, Chasteen AL, 47. van Boxtel MP, van Beijsterveldt CE, Houx PJ, /deafness/hearing_impairment_grades/en/.
Marchuk V, Singh G, Smith SL. Effects of hearing Anteunis LJ, Metsemakers JF, Jolles J. Mild hearing Accessed August 27, 2016.
and vision impairments on the Montreal Cognitive impairment can reduce verbal memory 63. Lin FR, Ferrucci L, An Y, et al. Association of
Assessment. Neuropsychol Dev Cogn B Aging performance in a healthy adult population. J Clin hearing impairment with brain volume changes in
Neuropsychol Cogn. 2015;22(4):413-437. Exp Neuropsychol. 2000;22(1):147-154. older adults. Neuroimage. 2014;90:84-92.
31. Era P, Jokela J, Qvarnberg Y, Heikkinen E. 48. Deal JA, Sharrett AR, Albert MS, et al. Hearing 64. Panza F, Solfrizzi V, Logroscino G. Age-related
Pure-tone thresholds, speech understanding, and impairment and cognitive decline: a pilot study hearing impairment—a risk factor and frailty marker
their correlates in samples of men of different ages. conducted within the Atherosclerosis Risk in for dementia and AD. Nat Rev Neurol. 2015;11(3):
Audiology. 1986;25(6):338-352. Communities neurocognitive study. Am J Epidemiol. 166-175.
32. Gussekloo J, de Craen AJM, Oduber C, 2015;181(9):680-690. 65. Rönnberg J, Lunner T, Zekveld A, et al. The
van Boxtel MPJ, Westendorp RGJ. Sensory 49. Helzner EP, Cauley JA, Pratt SR, et al. Race and Ease of Language Understanding (ELU) model:
impairment and cognitive functioning in oldest-old sex differences in age-related hearing loss: the theoretical, empirical, and clinical advances. Front
subjects: the Leiden 85+ Study. Am J Geriatr Health, Aging and Body Composition Study. J Am Syst Neurosci. 2013;7:31.
Psychiatry. 2005;13(9):781-786. Geriatr Soc. 2005;53(12):2119-2127. 66. Campbell J, Sharma A. Compensatory changes
33. Harrison Bush AL, Lister JJ, Lin FR, Betz J, 50. Lindenberger U, Baltes PB. Intellectual in cortical resource allocation in adults with hearing
Edwards JD. Peripheral hearing and cognition: functioning in old and very old age: cross-sectional loss. Front Syst Neurosci. 2013;7:71.
evidence from the Staying Keen in Later Life (SKILL) results from the Berlin Aging Study. Psychol Aging. 67. Salthouse TA. Selective review of cognitive
study. Ear Hear. 2015;36(4):395-407. 1997;12(3):410-432. aging. J Int Neuropsychol Soc. 2010;16(5):754-760.
34. Heron AIC, Chown SM. Age and Function. 51. Anstey KJ, Luszcz MA, Sanchez L. Two-year 68. Eckert MA, Kuchinsky SE, Vaden KI, Cute SL,
London, England: Churchill Press; 1967. decline in vision but not hearing is associated with Spampinato MV, Dubno JR. White matter
35. Hofer SM, Berg S, Era P. Evaluating the memory decline in very old adults in a hyperintensities predict low frequency hearing in
interdependence of aging-related changes in visual population-based sample. Gerontology. 2001;47(5): older adults. J Assoc Res Otolaryngol. 2013;14(3):
and auditory acuity, balance, and cognitive 289-293. 425-433.
functioning. Psychol Aging. 2003;18(2):285-305. 52. Anstey KJ, Hofer SM, Luszcz MA. A latent 69. Grady CL, Maisog JM, Horwitz B, et al.
36. Hong T, Mitchell P, Burlutsky G, Liew G, Wang growth curve analysis of late-life sensory and Age-related changes in cortical blood flow
JJ. Visual impairment, hearing loss and cognitive cognitive function over 8 years: evidence for activation during visual processing of faces and
function in an older population: longitudinal specific and common factors underlying change. location. J Neurosci. 1994;14(3, pt 2):1450-1462.
findings from the Blue Mountains Eye Study. PLoS Psychol Aging. 2003;18(4):714-726.
70. Lin FR, Hazzard WR, Blazer DG. Priorities for
One. 2016;11(1):e0147646. 53. Lindenberger U, Ghisletta P. Cognitive and improving hearing health care for adults: a report
37. Li S-C, Jordanova M, Lindenberger U. From sensory declines in old age: gauging the evidence from the National Academies of Sciences,
good senses to good sense: a link between tactile for a common cause. Psychol Aging. 2009;24(1):1- Engineering, and Medicine. JAMA. 2016;316(8):
information processing and intelligence. Intelligence. 16. 819-820.
1998;26(2):99-122. 54. Karpa MJ, Gopinath B, Beath K, et al. 71. Mosnier I, Bebear JP, Marx M, et al.
38. Lin FR. Hearing loss and cognition among older Associations between hearing impairment and Improvement of cognitive function after cochlear
adults in the United States. J Gerontol A Biol Sci Med mortality risk in older persons: the Blue Mountains implantation in elderly patients. JAMA Otolaryngol
Sci. 2011;66(10):1131-1136. Hearing Study. Ann Epidemiol. 2010;20(6):452-459. Head Neck Surg. 2015;141(5):442-450.
39. Lin FR, Ferrucci L, Metter EJ, An Y, Zonderman 55. Kiely KM, Gopinath B, Mitchell P, Luszcz M, 72. Mulrow CD, Aguilar C, Endicott JE, et al.
AB, Resnick SM. Hearing loss and cognition in the Anstey KJ. Cognitive, health, and sociodemographic Quality-of-life changes and hearing impairment:
Baltimore Longitudinal Study of Aging. predictors of longitudinal decline in hearing acuity a randomized trial. Ann Intern Med. 1990;113(3):
Neuropsychology. 2011;25(6):763-770. among older adults. J Gerontol A Biol Sci Med Sci. 188-194.
2012;67(9):997-1003.
40. Lin FR, Yaffe K, Xia J, et al; Health ABC Study 73. Acar B, Yurekli MF, Babademez MA, Karabulut
Group. Hearing loss and cognitive decline in older 56. Kurniawan C, Westendorp RG, de Craen AJ, H, Karasen RM. Effects of hearing aids on cognitive
adults. JAMA Intern Med. 2013;173(4):293-299. Gussekloo J, de Laat J, van Exel E. Gene dose of functions and depressive signs in elderly people.
apolipoprotein E and age-related hearing loss. Arch Gerontol Geriatr. 2011;52(3):250-252.
41. Lindenberger U, Baltes PB. Sensory functioning Neurobiol Aging. 2012;33(9):2230.e7-2230.e12.
and intelligence in old age: a strong connection. 74. Allen NH, Burns A, Newton V, et al. The effects
Psychol Aging. 1994;9(3):339-355. 57. López-Torres Hidalgo J, Boix Gras C, Téllez of improving hearing in dementia. Age Ageing.
Lapeira J, López Verdejo MA, del Campo del Campo 2003;32(2):189-193.
42. MacDonald SW, Dixon RA, Cohen AL, Hazlitt JM, Escobar Rabadán F. Functional status of elderly
JE. Biological age and 12-year cognitive change in people with hearing loss. Arch Gerontol Geriatr. 75. Salmon D. Neuropsychological features of mild
older adults: findings from the Victoria Longitudinal 2009;49(1):88-92. cognitive impairment and preclinical Alzheimer’s
Study. Gerontology. 2004;50(2):64-81. disease. In: Pardon MC, Bondi MW, eds. Behavioral
58. Quaranta N, Coppola F, Casulli M, et al. The Neurobiology of Aging. Heidelberg, Germany:
43. Schaie KW, Baltes P, Strother CR. A study of prevalence of peripheral and central hearing
auditory sensitivity in advanced age. J Gerontol. Springer; 2012.
impairment and its relation to cognition in older
1964;19:453-457. adults [published correction appears in Audiol 76. Stern Y. Cognitive reserve in ageing and
44. Sugawara N, Sasaki A, Yasui-Furukori N, et al. Neurootol. 2015;20(2):135]. Audiol Neurootol. 2014; Alzheimer’s disease. Lancet Neurol. 2012;11(11):
Hearing impairment and cognitive function among 19(suppl 1):10-14. 1006-1012.
a community-dwelling population in Japan. Ann 59. Tay T, Wang JJ, Kifley A, Lindley R, Newall P, 77. Wilson BS, Tucci DL, Merson MH, O’Donoghue
Gen Psychiatry. 2011;10(1):27. Mitchell P. Sensory and cognitive association in GM. Global hearing health care: new findings and
45. Thomas PD, Hunt WC, Garry PJ, Hood RB, older persons: findings from an older Australian perspectives. Lancet. 2017;390(10098):934.
Goodwin JM, Goodwin JS. Hearing acuity in a population. Gerontology. 2006;52(6):386-394. 78. National Academies of Sciences, Engineering,
healthy elderly population: effects on emotional, 60. Herbst KG, Humphrey C. Hearing impairment and Medicine. Hearing Health Care for Adults:
cognitive, and social status. J Gerontol. 1983;38(3): and mental state in the elderly living at home. BMJ. Priorities for Improving Access and Affordability.
321-325. 1980;281(6245):903-905. Washington, DC: The National Academies Press;
46. Valentijn SA, van Boxtel MP, van Hooren SA, 2016.
61. Lin FR, Metter EJ, O’Brien RJ, Resnick SM,
et al. Change in sensory functioning predicts Zonderman AB, Ferrucci L. Hearing loss and
change in cognitive functioning: results from a incident dementia. Arch Neurol. 2011;68(2):214-220.

126 JAMA Otolaryngology–Head & Neck Surgery February 2018 Volume 144, Number 2 (Reprinted) jamaotolaryngology.com

© 2017 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 04/18/2023