Research

JAMA Ophthalmology | Original Investigation

Association Between Retinal Layer Thickness

and Cognitive Decline in Older Adults
Hyeong Min Kim, MD, MSc; Ji Won Han, MD, PhD; Young Joo Park, MD, MSc; Jong Bin Bae, MD, PhD;
Se Joon Woo, MD, PhD; Ki Woong Kim, MD, PhD

Invited Commentary page 691

IMPORTANCE Retinal layer thickness is hypothesized to be related to cognitive function Supplemental content
in patients with mild cognitive impairment (MCI) and Alzheimer disease (AD). However,
longitudinal cohort studies of the healthy older population are scarce.

OBJECTIVE To investigate the association between retinal layer thickness and cognitive
impairment and future cognitive decline in a community-based population cohort.

DESIGN, SETTING, AND PARTICIPANTS A total of 430 randomly sampled community-dwelling

Korean individuals 60 years or older participated in the baseline assessment (mean [SD],
76.3 [6.6] years) 215 of whom completed a mean (SD) of 5.4 (0.6) years (range, 4.1-6.2 years)
of follow-up. Using spectral-domain optical coherence tomography, the study team assessed
the thickness of 6 retinal layers in the macular region, the peripapillary retinal nerve fiber
layers (RNFLs), and the subfoveal choroid at baseline.

EXPOSURES Age, sex, education, diabetes, hypertension, and apolipoprotein E4 gene status.

MAIN OUTCOMES AND MEASURES Retinal layer thickness and cognitive function test scores
were analyzed.

RESULTS This study included 430 participants (female, 208 [48.6%]). Baseline macular RNFL
thickness was associated with baseline Consortium to Establish a Registry for Alzheimer’s
Disease (CERAD) score (coefficient [β] = 0.077; 95% CI, 0.054-0.100; P = .04 for total
macular area) and Mini-Mental State Examination (MMSE) score (coefficient [β] = 0.082;
95% CI, 0.063-0.101; P = .03 for total macular area). A thinner baseline total macular RNFL
thickness (lowest quartile, <231 μm) was associated with a larger decline in the CERAD and
MMSE scores during the follow-up period (P = .003 and P = .01, respectively). Furthermore,
participants with baseline total macular RNFL thickness below the lowest quartile cutoff
value presented a greater decline in cognitive scores and a higher prevalence of cognitive
impairment and Alzheimer disease than those with RNFL thickness above the lowest quartile
cutoff value.

CONCLUSIONS AND RELEVANCE In this study, macular RNFL thickness could be used as
a prognostic biomarker of long-term cognitive decline in adults 60 years or older. However,
to confirm these results, further large-scale population-based studies should be performed.

Author Affiliations: Author

affiliations are listed at the end of this
article.
Corresponding Author: Se Joon
Woo, MD, PhD, Department of
Ophthalmology (sejoon1@snu.ac.kr),
and Ki Woong Kim, MD, PhD,
Department of Neuropsychiatry
(kwkimmd@snu.ac.kr), Seoul
National University Bundang
Hospital, 173-82 Gumi-ro,
Bundang-gu, Seongnam-si,
JAMA Ophthalmol. 2022;140(7):683-690. doi:10.1001/jamaophthalmol.2022.1563 Gyeonggi-do, 13620,
Published online May 26, 2022. Republic of Korea.

(Reprinted) 683

Downloaded From: https://jamanetwork.com/ by a University of Florida User on 09/07/2023

 Research Original Investigation
A
lzheimer disease (AD) is a severe neurodegenerative
disorder affecting millions of people globally.1 Subjec-
tive cognitive decline, also described as preclinical AD,
is likely to progress over time to mild cognitive impairment
(MCI) and AD.2 Therefore, clinicians and researchers have in-
vestigated several biomarkers of AD to diagnose the disease
as early as possible. Certain structural brain magnetic reso-
nance imaging findings, 3 and cerebrospinal fluid–based
biomarkers4 have been identified. The apolipoprotein E (APOE
ε4) genotype has been recognized as a notable biomarker for
AD. Moreover, many efforts have been made to discover non-
invasive ophthalmic biomarkers for the identification of pa-
tients with cognitive impairment and AD.
Since Parisi et al5 observed peripapillary retinal nerve
fiber layer (RNFL) thinning in patients with AD, other studies
have described optical coherence tomography (OCT) param-
eters in patients with dementia, MCI, and preclinical AD. Most
studies have reported peripapillary RNFLs thinning in pa-
tients who have AD and MCI compared with their age-
matched controls.6-14 Recent advances in the visualization
and segmentation of retinal layer structures by OCT have
provided better opportunities to analyze retinal layer mor-
phology in vivo.15,16 Prior investigations have reported corre-
lations between the peripapillary RNFLs, ganglion cell (GC)–
inner plexiform layer (IPL), and other retinal layers and
cognitive impairment in patients with MCI or AD who showed
retinal layer thinning. 5-14,17-20 Population-based studies
conducted by the UK Biobank suggested that thinner RNFLs
could indicate baseline cognitive dysfunction and future
cognitive decline over time, leading to an increased risk of
AD development.21
OCT angiography was developed and the association be-
tween cognitive performance and changes in the retinal mi-
crovasculature was investigated. Yoon et al22 suggested that
patients with AD had reduced macular vessel and perfusion
density in the superficial capillary plexus compared with
patients who had MCI and healthy controls. O’Bryhim et al23
proposed that cognitively healthy participants with preclini-
cal AD presented with foveal avascular zone (FAZ) enlarge-
ment, and their most recently published 3-year longitudinal
follow-up results showed that the FAZ enlargement re-
mained, although no differences were found in the other reti-
nal structural parameters.24
Similar to the UK Biobank studies, this community-based
longitudinal cohort study conducted in Korea aimed to enroll
a large number of participants to determine the association
between retinal layer thickness and cognitive function in
healthy individuals and individuals with cognitive impair-
ment. Both cross-sectional and longitudinal follow-up stud-
ies were conducted to determine relevant retinal layer thick-
ness parameters.
Methods
The institutional review board of Seoul National University
Bundang Hospital approved this population-based longitudi-
nal cohort study, which adhered to the tenets of the Declara-
684 JAMA Ophthalmology July 2022 Volume 140, Number 7 (Reprinted)
Downloaded From: https://jamanetwork.com/ by a University of Florida User on 09/07/2023
Association Between Retinal Layer Thickness and Cognitive Decline in Older Adults
Key Points
Question Is retinal layer thickness associated with cognitive
decline in an older population?
Findings In this cohort study including 430 community-dwelling
participants in Korea, baseline macular retinal nerve fiber layer
(RNFL) thickness was associated with baseline cognitive function
scores and follow-up cognitive decline.
Meaning These findings suggest that macular RNFL thickness
could be considered a predictive biomarker for evaluating
cognitive function in older individuals.
tion of Helsinki. Written informed consent was obtained from
all participants. For the cohort study, we used the Strength-
ening the Reporting of Observational Studies in Epidemiol-
ogy (STROBE) reporting guideline. The study participants did
not receive any compensation or incentives.
Participants
We enrolled Korean adults 60 years and older from 2 population-
based longitudinal cohort studies: the Korean Longitudinal
Study on Health and Aging (KLoSHA)25 and the Korean Longi-
tudinal Study on Cognitive Aging and Dementia (KLOSCAD).26
The specific details of these studies are described in eMethods
1 in the Supplement. A total of 500 participants were enrolled,
and 70 were excluded because of high myopia (axial length
>26 mm), high intraocular pressure (>21 mm Hg), self-
reported glaucoma history, and combined ocular pathologies
that might affect retinal layer thickness, such as age-related
macular degeneration, epiretinal membrane, diabetic macu-
lar edema, and diabetic retinopathy found on OCT infrared
imaging, leaving 430 participants for the final analysis. Among
these 430 participants, 215 completed follow-up assess-
ments, with a mean (SD) follow-up duration of 5.4 (0.6) years
(range, 4.1-6.2 years) (eFigure 1 in the Supplement).
Ophthalmic and Cognitive Function Assessments
We conducted a baseline assessment from September 2010 to
September 2011 and a follow-up assessment from September
2015 to September 2016. The baseline assessment comprised
comprehensive ophthalmic examinations, including spectral-
domain OCT (SD-OCT) to determine the best-corrected visual
acuity, intraocular pressure, auto kerato-refractometry, opti-
cal biometry axial length calculation (IOL Master; Carl-Zeiss
Meditec), and SD-OCT (Spectralis; Heidelberg Engineering). The
macula protocol, consisting of a raster scan composed of 31
horizontal lines centered on the fovea, was performed with 25
frames averaged for each OCT B-scan, along with 6-mm length
and automatic real-time processing27 to obtain the retinal layer
thickness of both the macula/fovea and peripapillary areas.
Nine macular fields were adopted in the macula/fovea area
based on the Early Treatment Diabetic Retinopathy Study
group, consisting of 3 concentric rings centered on the fovea
measuring 1 mm (center), 3 mm (inner), and 6 mm (outer) (eFig-
ure 2 in the Supplement). We measured 6 individual retinal lay-
ers (RNFL-GC layer, IPL, inner nuclear layer, outer plexiform
layer, and outer nuclear layer) using a built-in automated
jamaophthalmology.com
 Association Between Retinal Layer Thickness and Cognitive Decline in Older Adults Original Investigation Research

Table 1. Demographic and Clinical Characteristics of the Participants

No. (%), follow-up group (n = 215)

All Dropout group
Characteristic (N = 430) (n = 215) Baseline Follow-up
Age, mean (SD), y 76.3 (6.6) 76.2 (6.5) 76.5 (6.7) 80.9 (6.5)
Sex, No. (%)
Women 208 (48.6) 102 (47.9) 106 (49.3) NA
Men 222 (51.4) 113 (52.1) 50.7 (109) NA
Education, mean (SD), y 9.1 (5.6) 9.2 (5.9) 9.4 (5.7) NA
Axial length, mean (SD), mm 23.4 (1.2) 23.5 (1.4) 23.1 (1.3) NA
IOP, mean (SD), mm Hg 12.1 (3.3) 11.9 (3.5) 12.6 (3.1) NA
APOE ε4, No. (%) 70 (16.2) 30 (14.0) 40 (22.8) NA
Diabetes, No. (%) 90 (20.1) 48 (22.3) 42 (19.5) NA
Hypertension, No. (%) 170 (29.5) 80 (37.2) 90 (41.9) NA
CERAD total score, mean (SD) 66.7 (12.1) 66.0 (16.5) 66.6 (11.2) 54.1 (11.4) Abbreviations: APOE, apolipoprotein
MMSE score, mean (SD) 24.5 (4.6) 24.4 (5.1) 25.2 (3.2) 21.5 (3.1) E; CERAD, Consortium to Establish
a Registry for Alzheimer’s Disease
Cognitive disorders, No. (%)
Neuropsychological Battery;
MCI 63 (14.7) 38 (17.7) 25 (11.6) 38 (17.7) IOP, intraocular pressure;
Alzheimer disease 12 (2.8) 6 (2.8) 6 (2.8) 9 (4.2) MCI, mild cognitive impairment;
MMSE, Mini-Mental State
All 75 (17.5) 44 (20.4) 31 (14.4) 47 (21.9)
Examination.

program (HEYEX software; eFigure 2 in the Supplement). In
our study, the outer nuclear layer included the thick hypore-
flective band, the anatomical outer nuclear layer, and the Henle
fiber layer, in accord with the terminology before the 2014
International Nomenclature for Optical Coherence Tomogra-
phy Panel consensus.28 Two independent retina specialists
(H.M.K. and Y.J.P.) manually measured the subfoveal choroi-
dal thickness using the enhanced-depth imaging mode, and
the average of repeated measures was analyzed. We obtained
the average thickness of the peripapillary RNFLs measured in
6 different areas (eFigure 2 in the Supplement). We also col-
lected and analyzed retinal thickness data for the outer ring,
inner ring, and total macular area. The right eye of each par-
ticipant was selected for analysis of the retinal thickness data.
Research neuropsychologists or trained research nurses
performed comprehensive neuropsychological assessments,
including the Korean version of the Consortium to Establish
a Registry for Alzheimer’s Disease Assessment Packet (CERAD-K)
and the Mini-Mental State Examination (MMSE) (eMethods 2
in the Supplement). A panel of research neuropsychiatrists
diagnosed dementia according to the diagnostic criteria of the
Diagnostic and Statistical Manual of Mental Disorders (Fourth
Edition).29 MCI was diagnosed according to the consensus
criteria from the International Working Group on MCI.30 The
presence of objective cognitive impairment was ascertained
when the performance of the participants was −1.5 SD or be-
low the age-, sex-, and education-adjusted norms in any of
the neuropsychological tests. We classified participants into
cognitively normal and cognitively impaired (MCI or demen-
tia) groups.
Statistical Analysis
Continuous variables were compared between the groups using
t test and categorical variables were compared using Pearson
χ2 test. Continuous variables within participants were com-
pared using paired t tests. We examined the associations be-
tween baseline retinal layer thickness and baseline cognitive
test scores (CERAD total score and MMSE) using multiple lin-
ear regression analyses and changes in cognitive test scores
using repeated-measures analysis of variance. In addition, the
lowest quartile of baseline macular RNFL thickness was set to
compare the 2 groups below and above the cutoff value. In both
analyses, we adjusted for age, sex, level of education, pres-
ence of the APOE ε4 allele, diabetes, and hypertension as
covariates. Two-sided hypothesis testing was performed, and
the null hypothesis was rejected if P = .05; the P values were
not adjusted for multiple analyses. The 95% CI was also docu-
mented. All statistical analyses were performed using IBM
SPSS Statistics for Windows (version 25.0; SPSS Inc).
Results
This study included 430 participants (female, 206 [48.6%]). The
demographic and clinical characteristics of the study partici-
pants are summarized in Table 1. Cognitive disorders were less
common in participants who responded to the follow-up as-
sessment (follow-up group) than in those who did not (drop-
out group). However, all the other characteristics were compa-
rable. In the follow-up group, both CERAD total score and MMSE
scores decreased and the frequency of cognitive disorders in-
creased during the follow-up period. At the baseline assess-
ment, the follow-up group showed slightly thicker retinal lay-
ers than the dropout group in the macular RNFL inner layer
(mean [SD], 94.7 [14.0] vs 90.4 [12.4]; P < .001), all layers of the
macular RNFL mean ([SD], 257.2 [34.7] vs 250.2 [34.9]; P = .04),
ganglion cell outer layer mean ([SD], 131.3 [18.3] vs 135.6 [19.8];
P = .02), and inner nuclear inner layer mean ([SD], 160.1 [15.6]
vs 156.9 [15.7]; P = .04) (eTable 1 in the Supplement).
At baseline, the macular RNFL thickness was associated
with the baseline CERAD total score (baseline CERAD total
score for outer thickness: β = 0.071; 95% CI, 0.047-0.095;

jamaophthalmology.com (Reprinted) JAMA Ophthalmology July 2022 Volume 140, Number 7 685

Downloaded From: https://jamanetwork.com/ by a University of Florida User on 09/07/2023

 Research Original Investigation Association Between Retinal Layer Thickness and Cognitive Decline in Older Adults

Table 2. Association of Baseline Retinal Layer Thickness With Baseline Cognitive Performancea

CERAD total score MMSE score

Characteristic β P value β P value
Subfoveal choroid 0.021 .40 0.024 .42
Retinal nerve fiber layer
Average peripapillary 0.034 .29 0.015 .56
Macular
Outer 0.071 .04 0.078 .04
Inner 0.069 .047 0.074 .04
Total 0.077 .04 0.082 .03
Ganglion cell layer
Outer 0.004 .91 0.016 .65
Inner 0.001 .98 0.018 .62
Total 0.003 .94 0.004 .92
Inner plexiform layer
Outer 0.008 .82 0.006 .87
Inner 0.006 .87 0.049 .19
Total 0.008 .82 0.026 .47
Inner nuclear layer
Outer 0.012 .72 0.039 .29
Inner 0.032 .36 0.021 .57
Total 0.015 .67 0.033 .37
Outer plexiform layer
Outer 0.033 .34 0.025 .50 Abbreviations: CERAD, Consortium to
Establish a Registry for Alzheimer’s
Inner 0.039 .26 0.022 .55
Disease Neuropsychological Battery;
Total 0.042 .23 0.025 .50 MMSE, Mini-Mental State
Examination.
Outer nuclear layer
a
Multiple linear regression analyses
Outer 0.032 .37 0.028 .45
adjusted for age, sex, level of
Inner 0.025 .48 0.052 .15 education, diabetes, hypertension,
and the presence of the
Total 0.029 .41 0.044 .23
apolipoprotein ε4 allele.

P = .04; baseline CERAD total score for inner thickness,
β = 0.069; 95% CI, 0.051-0.087; P = .047; baseline CERAD total
score for total thickness: β = 0.077; 95% CI, 0.054-0.100;
P = .04) and the baseline MMSE score (MMSE score for outer
thickness: β = 0.078; 95% CI, 0.059-0.09; P = .04; MMSE score
for inner thickness: β = 0.074; 95% CI, 0.058-0.090; P = .04;
MMSE score for total thickness: β = 0.082; 95% CI, 0.063-
0.101; P = .03) (Table 2). In the longitudinal study, macular
RNFL thickness was not associated with decreases in the
CERAD total score and MMSE scores during the follow-up
period: outer RNFL CERAD total score: F208 = 1.170; 95% CI,
0.845-1.495; P = .21; MMSE: F208 = 0.956; 95% CI, 0.637-
1.275; P = .59; inner RNFL CERAD total score: F208 = 1.229; 95%
CI, 0.896-1.52; P = .16; MMSE: F208 = 1.358; 95% CI, 0.958-
1.758; P = .01; total RNFL CERAD total score: F208 = 1.420; 95%
CI, 0.971-1.869; P = .07; and MMSE: F208 = 0.879; 95% CI,
0.582-1.176; P = .74 (Table 3).
However, after setting the cutoff value as the lowest quar-
tile (outer, 135 μm; inner, 86 μm; total, 231 μm), the baseline
macular RNFL thickness was associated with a decrease in both
CERAD total score and MMSE scores during the follow-up
period. A thinner baseline outer RNFL was associated with a
decline in both CERAD total score (F208 = 6.912; 95% CI, 5.478-
8.346; P = .009) and MMSE scores (F208 = 4.265; 95% CI, 3.189-
5.341; P = .04); the inner RNFL in both CERAD total score
(F208 = 7.600; 95% CI, 5.913-9.287; P = .006); MMSE scores
(F208 = 6.626; 95% CI, 5.237-8.015; P = .01); and the total RNFL
in both CERAD total score (F 208 = 8.980; 95% CI, 6.792-
11.168; P = .003); and MMSE scores (F208 = 6.645; 95% CI,
4.968-8.322; P = .01) (Table 3). The decline in annual cogni-
tive performance scores according to baseline macular RNFL
is shown in Figure 1 and eTable 2 in the Supplement. In addi-
tion, the baseline and follow-up CERAD total scores and MMSE
scores divided by the lowest quartile of total macular RNFL
(<231 μm) are shown in Figure 2 and eTable 2 in the Supple-
ment. Interestingly, the cognitive performance scores were not
different at baseline but showed differences at follow-up.
Figures 1 and 2 show that participants with a thinner baseline
macular RNFL were likely to experience a greater degree of cog-
nitive decline, leading to cognitive impairment.
The baseline and follow-up prevalence of MCI and AD ac-
cording to the lowest quartile are shown in eTable 2 in the
Supplement. In the lower-quartile group (n = 55), the preva-
lence of MCI increased from baseline from 27.2% (15 of 55) to

686 JAMA Ophthalmology July 2022 Volume 140, Number 7 (Reprinted) jamaophthalmology.com

Downloaded From: https://jamanetwork.com/ by a University of Florida User on 09/07/2023

 Association Between Retinal Layer Thickness and Cognitive Decline in Older Adults Original Investigation Research

Table 3. Association of Baseline Macular Retinal Nerve Fiber Layer Thickness With Changes
in Cognitive Performance During the Follow-up Perioda

CERAD total score MMSE

Characteristic F208 P value F208 P value
Outer
Thickness 1.170 .21 0.956 .56
Lowest quartile (135 μm) 6.912 .01 4.265 .04
Inner
Thickness 1.229 .16 1.358 .09 Abbreviations: CERAD, Consortium to
Establish a Registry for Alzheimer’s
Lowest quartile (86 μm) 7.600 .006 6.626 .01
Disease Neuropsychological Battery;
Total MMSE, Mini-Mental State
Thickness 1.420 .07 0.879 .74 Examination.
a
Repeated-measures analysis
Lowest quartile (231 μm) 8.980 .003 6.645 .01
of variance.

Figure 1. Decline in Annual Cognitive Performance Scores According to the Baseline Total Macular Retinal Nerve Fiber Layer (RNFL) Quartile Groups

Below lowest quartile (<231 μm) Above lowest quartile

A Outer macular RNFL B Inner macular RNFL C Total macular RNFL

5 5 5

a a
a
Annual cognitive decline, score

Annual cognitive decline, score

Annual cognitive decline, score

4 4 4

3 3 3

2 2 2

a a a

1 1 1

0 0 0
CERAD-TS MMSE CERAD-TS MMSE CERAD-TS MMSE

Outer, inner, and total macular RNFL thickness data were subdivided into to Establish a Registry for Alzheimer’s Disease Neuropsychological Battery total
2 groups, ie, below the lowest quartile and above the lowest quartile. The score; MMSE, Mini-Mental State Examination.
degree of cognitive function decline was higher in the below-lowest-quartile a
P < .05.
group (thinner macular RNFL). Abbreviations: CERAD-TS indicates Consortium

41.8% (23 of 55) and the prevalence of AD increased from 7.2% ated with baseline cognitive performance scores. Instead, af-
(4 of 55) to 10.9% (6 of 55) at follow-up, whereas in the above- ter setting the cutoff value as the lowest quartile, the baseline
lowest-quartile group (n = 160), the prevalence of MCI in- outer, inner, and total macular RNFL thicknesses were asso-
creased from 6.3% (10 of 160) to 9.4% (15 of 160) for MCI and ciated with changes in both CERAD total scores and MMSE
from 1.3% (2 of 160) to 1.9% (3 of 160) for AD at follow-up. The scores (Table 3). Moreover, there were no differences in the cog-
below-lowest-quartile group showed a much higher rate of MCI nitive score and RNFL thickness according to APOE ε4 status
and AD prevalence than the above-lowest-quartile group. (eTable 3 in the Supplement).
Previously, researchers found that retinal layer thickness
in the RNFL, GC-IPL, and other areas was related to cognitive
impairment; thus, retinal layer assessment with OCT could be
Discussion a convenient method of cognitive function assesment.8-10,12,17,20
This longitudinal cohort study enrolled older community- However, recent studies on patients with preclinical AD have re-
based participants to attempt to reveal the association be- ported some conflicting results. Snyder et al31 observed a thick
tween OCT-measured retinal layer thickness and cognitive IPL in patients with preclinical AD, whereas López-Cuenca et al32
function and decline. The results of the cross-sectional study observed a thin macular RNFL, IPL, inner nuclear layer, and outer
showed that the change in thicknesses of the outer and inner plexiform layer. In contrast, van de Kreeke et al33,34 showed no
macular RNFLs were associated with the baseline CERAD total differences in any of the measured retinal layers, including the
scores and MMSE scores (Table 2). In the longitudinal study, peripapillary and macular RNFL, ganglion cell layer, and IPL.
the overall baseline macular RNFL thickness was not associ- Therefore, an exploratory cross-sectional analysis was performed

jamaophthalmology.com (Reprinted) JAMA Ophthalmology July 2022 Volume 140, Number 7 687

Downloaded From: https://jamanetwork.com/ by a University of Florida User on 09/07/2023

 Research Original Investigation
Figure 2. Cognitive Performance Scores According to the Baseline Total Macular Retinal Nerve Fiber Layer (RNFL) Quartile Groups
A CERAD-TS
75
70
Above lowest quartile
65
60
Points
55
Below lowest quartile (<231 μm)
50 a
45
40
Baseline (n = 215) Follow-up (n = 215)
Both Consortium to Establish a Registry for Alzheimer’s Disease
Neuropsychological Battery total score and Mini-Mental State Examination
scores were not different at baseline; however, they differed at the follow-up
evaluation. The below-lowest-quartile group showed a greater cognitive decline
to discover which retinal layer is relevant for cognitive perfor-
mance. We found that only the macular RFNL was associated
with neuropsychiatric test scores and that it could be a candi-
date layer for predicting cognitive decline.
Unlike previous investigations focusing on patients with
MCI and AD, these longitudinal population-based studies
included older participants from the general population in
local communities to detect possible associations between
retinal layer thickness and early changes in cognitive func-
tion. Méndez-Gómez et al 35 from the Three-City-Alienor
cohort (427 participants) proposed that RNFL thickness was
not associated with initial cognitive scores using the MMSE,
but it was associated with episodic memory loss over 2 years
of follow-up. A UK Biobank prospective, multicenter,
community-based study21 reported that a thinner RNFL was
associated with worse cognitive function at baseline and dur-
ing follow-up and presented an increased risk of dementia.
The UK Biobank study examined large cohorts of more than
30 000 individuals at baseline and 1251 individuals at the
3-year follow-up using cognitive tests including prospective
memory, pairs matching, numeric and verbal reasoning, and
reaction time with touch screens. The Rotterdam study36
analyzed more than 3000 participants and revealed that a
thinner GC-IPL was associated with dementia, whereas
a thinner RNFL at baseline was associated with an increased
risk of developing dementia. Along with these population-
based longitudinal cohort studies, we assessed the cognitive
performance of the enrolled participants at baseline and final
follow-up at 4 to 6 years and matched it with the baseline
retinal layer thickness data.
Recent OCT angiography studies that investigated the as-
sociation between OCT angiography parameters and cogni-
tive decline showed that the FAZ area is enlarged in patients
who had preclinical AD with positive biomarkers.22-24 Our study
is in line with the OCT angiography studies as macular micro-
vasculature changes, including in FAZ areas, are known to be
correlated with macular RNFL thickness.37 However, it is still
688 JAMA Ophthalmology July 2022 Volume 140, Number 7 (Reprinted)
Downloaded From: https://jamanetwork.com/ by a University of Florida User on 09/07/2023
Association Between Retinal Layer Thickness and Cognitive Decline in Older Adults
B MMSE
28
26
Above lowest quartile
24
Points
22
Below lowest quartile (<231 μm) a
20
18
Baseline (n = 215) Follow-up (n = 215)
than the above-lowest-quartile group.
a
P< .05.
unknown whether enlargement of the FAZ precedes macular
RNFL thinning, which is the main mechanism of the associa-
tion between cognitive decline and macular retinal changes
in the microvasculature and neural structures.
The participants of the current study were community-
based people 60 years and older; thus, unlike in prior inves-
tigations, most participants presented with normal cognitive
function or MCI, with only 12 patients (2.8%) having definite
AD. Therefore, we considered our cohort population suitable
for assessing the association between retinal layer thickness
and cognitive function in a real-world clinical setting. Our re-
sults showed that baseline macular RNFL thickness was asso-
ciated with baseline neuropsychiatric test scores. However,
contrary to other studies, the average peripapillary RNFL thick-
ness showed no differences. We presumed that macular RNFL
thinning might precede peripapillary RNFL thinning in the
early phase of cortical degeneration.
Another characteristic of our longitudinal study was that
we enrolled a larger number of participants than previous case-
control studies. Moreover, because we adjusted for param-
eters including age, sex, education level, diabetes, hyperten-
sion, and APOE ε4 status—that is, variables that might affect
cognitive function assessment—we conducted more compre-
hensive statistical analyses compared with prior investiga-
tions. In addition, we investigated the possibility of predict-
ing cognitive function decline based on baseline retinal layer
thickness in a longitudinal study and found that baseline macu-
lar RNFL thickness may be associated with the prognosis of
cognitive function.
Another strength of our study was that we analyzed both
cognitive function assessment scores (CERAD and MMSE). Al-
though the MMSE has certain disadvantages, such as limited
examination of visuospatial cognitive ability, bias against
poorly educated participants, and poor sensitivity in detect-
ing patients with very mild dementia, it is a quick and conve-
nient test to diagnose cognitive impairment in a clinical set-
ting. Hence, our results could be useful for clinicians, both
jamaophthalmology.com
 Association Between Retinal Layer Thickness and Cognitive Decline in Older Adults Original Investigation Research

neuropsychiatrists and ophthalmologists, in matching OCT reti-
nal thickness data and MMSE scores in the real world.
Limitations
This study has limitations. First, this longitudinal study did not
perform β-amyloid (Aβ) positron emission tomography and mag-
netic resonance imaging , as cerebral Aβ deposition starts ap-
proximately 15 to 20 years before the onset of AD, which is cru-
cial for detecting future candidates for AD who are still cognitively
healthy.38 van de Kreeke et al33,34 studied cognitively normal in-
dividuals with amyloid pathology on positron emission tomog-
raphy (preclinical AD) and observed no differences in retinal layer
thickness in the macular and peripapillary RNFL in both cross-
sectional and longitudinal studies. In contrast, Byun et al39 re-
cently reported that Aβ-positive patients with preclinical AD
showed reduced inner nasal macular thickness and RNFL thick-
ness compared with Aβ-negative patients.39 Second, numer-
ous parameters besides the selected confounding variables of
age, sex, education level, diabetes, hypertension, and APOE ε4
status could also influence cognitive function; thus, strict ad-
justments might not be feasible. Third, in the follow-up period,
baseline macular RNFL thickness as a score changes variable was
not associated with either CERAD or MMSE; instead, setting the
lowest quartile as the cutoff value revealed a correlation be-
tween these 2 parameters. Further longitudinal studies with
larger sample sizes are warranted. Fourth, among the baseline
430 participants, only half participated in the follow-up study.
This may have affected the longitudinal analysis in our cohort
study. Fifth, our longitudinal study follow-up lasted for a mean
(SD) of 5.4 (0.6) years (range, 4.1-6.2 years), which is a consid-
erably small period of time in the development of neurodegen-
erative diseases. Subsequent follow-up at an additional time
point of 5 years, approximately 10 years from baseline, was
planned in this cohort study. Additionally, we could not per-
form the final follow-up SD-OCT evaluation; therefore, in the lon-
gitudinal study, we could not correlate changes in cognitive func-
tion score with changes in retinal layer thickness. Further
investigations are needed to evaluate OCT retinal thickness
changes throughout regular follow-up intervals.
Conclusions
In conclusion, this longitudinal cohort study revealed that
OCT-measured baseline macular RNFL thickness was associ-
ated with future cognitive decline. We propose that a thinner
macular RNFL may predict a decline in cognitive perfor-
mance. Overall, macular RNFL thickness may be considered
a noninvasive ocular biomarker for assessing changes in cog-
nitive function in patients. However, regarding the limita-
tions mentioned above, further population-based investiga-
tions with a long-term follow-up are necessary.

ARTICLE INFORMATION
Accepted for Publication: April 7, 2022.
Published Online: May 26, 2022.
doi:10.1001/jamaophthalmol.2022.1563
Open Access: This is an open access article
distributed under the terms of the CC-BY License.
© 2022 Kim HM et al. JAMA Ophthalmology.
Author Affiliations: Department of
Ophthalmology, Seoul National University College
of Medicine, Seoul National University Bundang
Hospital, Seongnam, Republic of Korea (H. M. Kim,
Woo); Department of Neuropsychiatry, Seoul
National University Bundang Hospital, Seongnam,
Republic of Korea (Han, Bae, K. W. Kim);
Department of Ophthalmology, Kangwon National
University School of Medicine, Kangwon National
University Hospital, Chuncheon, Republic of Korea
(Park); Department of Psychiatry and Behavioral
Science, Seoul National University College of
Medicine, Seoul, Republic of Korea (K. W. Kim);
Department of Brain and Cognitive Science, Seoul
National University College of Natural Sciences,
Seoul, Republic of Korea (K. W. Kim).
Author Contributions: Drs Se Joon Woo and
Ki Woong Kim had full access to all of the data in
the study and take responsibility for the integrity
of the data and the accuracy of the data analysis.
Drs Hyeong Min Kim and Ji Won Han contributed
equally to this work.
Concept and design: H. Kim, Park, Woo, K. Kim.
Acquisition, analysis, or interpretation of data:
H. Kim, Han, Park, Woo, K. Kim.
Drafting of the manuscript: H. Kim, Han, Park, Woo.
Critical revision of the manuscript for important
intellectual content: H. Kim, Han, Woo, K. Kim.
Statistical analysis: H. Kim, Han, K. Kim.
Obtained funding: Woo, K. Kim.
Administrative, technical, or material support:
H. Kim, Han, Park, Woo, K. Kim.
Supervision: Han, Woo, K. Kim.
Conflict of Interest Disclosures: Dr Woo reported
grants from Seoul National University Bundang
Hospital (11-2010-037 and 18-2018-024) and
grants from National Research Foundation
(2020R1F1A1072795) during the conduct of the
study. No other disclosures were reported.
Funding/Support: This research was supported
by a research grant from Seoul National University
Bundang Hospital (11-2010-037, 18-2018-024),
a National Research Foundation grant funded by
the Korean government (Ministry of Science and
information and communication technology)
(grant 2020R1F1A1072795), and a grant from
the Korean Health Technology R&D Project,
Ministry of Health and Welfare, Republic of Korea
(grant HI09C1379 [A092077]).
Role of the Funder/Sponsor: The funders had no
role in the design and conduct of the study;
collection, management, analysis, and
interpretation of the data; preparation, review,
or approval of the manuscript; and decision to
submit the manuscript for publication.
Additional Contributions: We thank Min Seok Kim,
MD, MSc, Retina Center, Moon's Eye Clinic, Suwon,
Korea, and Jun Seong Bae, Hyun Seok Seo, and
Hee Seo Choi, College of Medicine, Seoul National
University, for their assistance in data cleansing
and analysis. The contributors did not receive
compensation.
REFERENCES
1. GBD 2016 Neurology Collaborators. Global,
regional, and national burden of neurological
disorders, 1990-2016: a systematic analysis for the
Global Burden of Disease Study 2016. Lancet Neurol.
2019;18(5):459-480. doi:10.1016/S1474-4422(18)
30499-X
2. Rabin LA, Smart CM, Amariglio RE. Subjective
cognitive decline in preclinical Alzheimer’s disease.
Annu Rev Clin Psychol. 2017;13:369-396.
doi:10.1146/annurev-clinpsy-032816-045136
3. McConathy J, Sheline YI. Imaging biomarkers
associated with cognitive decline: a review. Biol
Psychiatry. 2015;77(8):685-692. doi:10.1016/j.
biopsych.2014.08.024
4. Kulic L, Unschuld PG. Recent advances in
cerebrospinal fluid biomarkers for the detection of
preclinical Alzheimer’s disease. Curr Opin Neurol.
2016;29(6):749-755. doi:10.1097/WCO.
0000000000000399
5. Parisi V, Restuccia R, Fattapposta F, Mina C,
Bucci MG, Pierelli F. Morphological and functional
retinal impairment in Alzheimer’s disease patients.
Clin Neurophysiol. 2001;112(10):1860-1867.
doi:10.1016/S1388-2457(01)00620-4
6. Ferrari L, Huang SC, Magnani G, Ambrosi A,
Comi G, Leocani L. Optical coherence tomography
reveals retinal neuroaxonal thinning in
frontotemporal dementia as in Alzheimer’s disease.
J Alzheimers Dis. 2017;56(3):1101-1107.
doi:10.3233/JAD-160886
7. Garcia-Martin E, Bambo MP, Marques ML, et al.
Ganglion cell layer measurements correlate with
disease severity in patients with Alzheimer’s
disease. Acta Ophthalmol. 2016;94(6):e454-e459.
doi:10.1111/aos.12977
8. Iseri PK, Altinaş O, Tokay T, Yüksel N.
Relationship between cognitive impairment
and retinal morphological and visual functional
abnormalities in Alzheimer disease.

jamaophthalmology.com (Reprinted) JAMA Ophthalmology July 2022 Volume 140, Number 7 689

Downloaded From: https://jamanetwork.com/ by a University of Florida User on 09/07/2023

 Research Original Investigation
J Neuroophthalmol. 2006;26(1):18-24.
doi:10.1097/01.wno.0000204645.56873.26
9. Kesler A, Vakhapova V, Korczyn AD, Naftaliev E,
Neudorfer M. Retinal thickness in patients with
mild cognitive impairment and Alzheimer’s disease.
Clin Neurol Neurosurg. 2011;113(7):523-526.
doi:10.1016/j.clineuro.2011.02.014
10. Kirbas S, Turkyilmaz K, Anlar O, Tufekci A,
Durmus M. Retinal nerve fiber layer thickness in
patients with Alzheimer disease. J Neuroophthalmol.
2013;33(1):58-61. doi:10.1097/WNO.
0b013e318267fd5f
11. Larrosa JM, Garcia-Martin E, Bambo MP, et al.
Potential new diagnostic tool for Alzheimer’s
disease using a linear discriminant function for
Fourier domain optical coherence tomography.
Invest Ophthalmol Vis Sci. 2014;55(5):3043-3051.
doi:10.1167/iovs.13-13629
12. Liu D, Zhang L, Li Z, et al. Thinner changes of
the retinal nerve fiber layer in patients with mild
cognitive impairment and Alzheimer’s disease. BMC
Neurol. 2015;15:14. doi:10.1186/s12883-015-0268-6
13. Moschos MM, Markopoulos I, Chatziralli I, et al.
Structural and functional impairment of the retina
and optic nerve in Alzheimer’s disease. Curr
Alzheimer Res. 2012;9(7):782-788. doi:10.2174/
156720512802455340
14. Paquet C, Boissonnot M, Roger F, Dighiero P,
Gil R, Hugon J. Abnormal retinal thickness in
patients with mild cognitive impairment and
Alzheimer’s disease. Neurosci Lett. 2007;420(2):
97-99. doi:10.1016/j.neulet.2007.02.090
15. Hee MR, Izatt JA, Swanson EA, et al. Optical
coherence tomography of the human retina. Arch
Ophthalmol. 1995;113(3):325-332. doi:10.1001/
archopht.1995.01100030081025
16. Huang D, Swanson EA, Lin CP, et al. Optical
coherence tomography. Science. 1991;254(5035):
1178-1181. doi:10.1126/science.1957169
17. Ascaso FJ, Cruz N, Modrego PJ, et al.
Retinal alterations in mild cognitive impairment
and Alzheimer’s disease: an optical coherence
tomography study. J Neurol. 2014;261(8):1522-1530.
doi:10.1007/s00415-014-7374-z
18. Berisha F, Feke GT, Trempe CL, McMeel JW,
Schepens CL. Retinal abnormalities in early
Alzheimer’s disease. Invest Ophthalmol Vis Sci.
2007;48(5):2285-2289. doi:10.1167/iovs.06-1029
19. Cunha JP, Proença R, Dias-Santos A, et al.
OCT in Alzheimer’s disease: thinning of the RNFL
690 JAMA Ophthalmology July 2022 Volume 140, Number 7 (Reprinted)
Downloaded From: https://jamanetwork.com/ by a University of Florida User on 09/07/2023
Association Between Retinal Layer Thickness and Cognitive Decline in Older Adults
and superior hemiretina. Graefes Arch Clin Exp
Ophthalmol. 2017;255(9):1827-1835. doi:10.1007/
s00417-017-3715-9
20. Marziani E, Pomati S, Ramolfo P, et al.
Evaluation of retinal nerve fiber layer and ganglion
cell layer thickness in Alzheimer’s disease using
spectral-domain optical coherence tomography.
Invest Ophthalmol Vis Sci. 2013;54(9):5953-5958.
doi:10.1167/iovs.13-12046
21. Ko F, Muthy ZA, Gallacher J, et al; UK Biobank
Eye & Vision Consortium. Association of retinal
nerve fiber layer thinning with current and future
cognitive decline: a study using optical coherence
tomography. JAMA Neurol. 2018;75(10):1198-1205.
doi:10.1001/jamaneurol.2018.1578
22. Yoon SP, Grewal DS, Thompson AC, et al.
Retinal microvascular and neurodegenerative
changes in Alzheimer’s disease and mild cognitive
impairment compared with control participants.
Ophthalmol Retina. 2019;3(6):489-499.
doi:10.1016/j.oret.2019.02.002
23. O’Bryhim BE, Apte RS, Kung N, Coble D,
Van Stavern GP. Association of preclinical Alzheimer
disease with optical coherence tomographic
angiography findings. JAMA Ophthalmol. 2018;136
(11):1242-1248. doi:10.1001/jamaophthalmol.2018.
3556
24. O’Bryhim BE, Lin JB, Van Stavern GP, Apte RS.
OCT angiography findings in preclinical Alzheimer’s
disease: 3-year follow-up. Ophthalmology. 2021;128
(10):1489-1491. doi:10.1016/j.ophtha.2021.02.016
25. Park JH, Lim S, Lim J-Y, et al. An overview
of the Korean Longitudinal Study on Health and
Aging. Psychiatry Investig. 2007;4(2):84-95.
26. Han JW, Kim TH, Kwak KP, et al. Overview of
the Korean Longitudinal Study on Cognitive Aging
and Dementia. Psychiatry Investig. 2018;15(8):
767-774. doi:10.30773/pi.2018.06.02
27. Ryoo NK, Ahn SJ, Park KH, et al. Thickness of
retina and choroid in the elderly population and
its association with Complement Factor H
polymorphism: KLoSHA Eye study. PLoS One. 2018;
13(12):e0209276. doi:10.1371/journal.pone.0209276
28. Staurenghi G, Sadda S, Chakravarthy U,
Spaide RF; International Nomenclature for Optical
Coherence Tomography (IN–OCT) Panel. Proposed
lexicon for anatomic landmarks in normal posterior
segment spectral-domain optical coherence
tomography: the IN•OCT consensus. Ophthalmology.
2014;121(8):1572-1578. doi:10.1016/j.ophtha.2014.
02.023
29. American Psychiatric Association. Diagnostic
and Statistical Manual of Mental Disorders. 4th ed.
American Psychiatric Association; 1994.
30. Petersen RC. Mild cognitive impairment as
a diagnostic entity. J Intern Med. 2004;256(3):183-
194. doi:10.1111/j.1365-2796.2004.01388.x
31. Snyder PJ, Johnson LN, Lim YY, et al.
Nonvascular retinal imaging markers of preclinical
Alzheimer’s disease. Alzheimers Dement (Amst).
2016;4:169-178. doi:10.1016/j.dadm.2016.09.001
32. López-Cuenca I, de Hoz R, Salobrar-García E,
et al. Macular thickness decrease in asymptomatic
subjects at high genetic risk of developing
Alzheimer’s disease: an OCT study. J Clin Med.
2020;9(6):1728. doi:10.3390/jcm9061728
33. van de Kreeke JA, Nguyen HT, den Haan J, et al.
Retinal layer thickness in preclinical Alzheimer’s
disease. Acta Ophthalmol. 2019;97(8):798-804.
doi:10.1111/aos.14121
34. van de Kreeke JA, Nguyen HT, Konijnenberg E,
et al. Longitudinal retinal layer changes in preclinical
Alzheimer’s disease. Acta Ophthalmol. 2021:99(5):
538-544.
35. Méndez-Gómez JL, Rougier MB, Tellouck L,
et al. Peripapillary retinal nerve fiber layer thickness
and the evolution of cognitive performance in
an elderly population. Front Neurol. 2017;8:93.
doi:10.3389/fneur.2017.00093
36. Mutlu U, Colijn JM, Ikram MA, et al. Association
of retinal neurodegeneration on optical coherence
tomography with dementia: a population-based
study. JAMA Neurol. 2018;75(10):1256-1263.
doi:10.1001/jamaneurol.2018.1563
37. Zhou Y, Zhou M, Gao M, Liu H, Sun X. Factors
affecting the foveal avascular zone area in healthy
eyes among young Chinese adults. Biomed Res Int.
2020;7361492. doi:10.1155/2020/7361492
38. Jack CR Jr, Knopman DS, Jagust WJ, et al.
Hypothetical model of dynamic biomarkers of the
Alzheimer’s pathological cascade. Lancet Neurol.
2010;9(1):119-128. doi:10.1016/S1474-4422(09)
70299-6
39. Byun MS, Park SW, Lee JH, et al; KBASE
Research Group. Association of retinal changes
with Alzheimer disease neuroimaging biomarkers
in cognitively normal individuals. JAMA Ophthalmol.
2021;139(5):548-556. doi:10.1001/jamaophthalmol.
2021.0320
jamaophthalmology.com