Thinning of the Eye’s Retinal Layer a Marker of Cognitive Decline?

Thinning of the eye’s macular retinal nerve fiber layer (RNFL) is linked to cognitive decline in older adults, new research suggests.

Results from a study of more than 400 adults showed an association between baseline macular RNFL thickness and scores on two cognitive tests — with thinner baseline macular RNFL associated with greater cognitive decline.

A steeper decline in cognitive scores and a higher prevalence of cognitive impairment and Alzheimer’s disease (AD) were also shown in participants with baseline total macular RNFL thickness below the lowest quartile cutoff value vs with RNFL thickness above that threshold.

“In this study, macular RNFL thickness could be used as a prognostic biomarker of long-term cognitive decline in adults 60 years or older,” investigators led by Hyeong Min Kim, MD, Seoul National University College of Medicine, Seoul National University Bundang Hospital , Seongnam, Korea, write.

The findings were published online May 26 in JAMA Ophthalmology.

Early Diagnosis

Subjective cognitive decline, or preclinical AD, precedes mild cognitive impairment (MCI) and AD. Because of this, clinicians and researchers “have investigated several biomarkers of AD to diagnose the disease as early as possible,” the investigators note.

These include the discovery of noninvasive ophthalmic biomarkers that might identify patients with cognitive impairment and AD. Since publication of a 2001 study reporting peripapillary RNFL thinning in patients with AD, other studies have reported similar findings and advances in visualization.

In addition, the segmentation of retinal layer structures by optical coherence tomography (OCT) has “provided better opportunities to analyze retinal layer morphology in vivo,” the researchers write.

In the current study, they looked both cross-sectionally and longitudinally at retinal layer thickness parameters and their potential association with future recognition decline and AD.

The investigators assessed 430 randomly selected community-dwelling adults (mean age, 76.3 years; 48.6% women) from two population-based longitudinal cohort studies: the Korean Longitudinal Study on Health and Aging and the Longitudinal Study on Cognitive Aging and Dementia. Of these participants, 215 completed a mean of 5.4 years of follow-up.

Spectral-domain OCT was used to assess the thickness of six retinal layers of the macular region, RNFLs, and subfoveal choroid. In addition, participants completed two neuropsychological assessments at baseline and at follow-up: 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).

Covariates included age, sex, level of education, presence of the APOE e4 allele, diabetes, and hypertension.

Ocular Biomarker

Baseline total macular RNFL thickness was associated with scores on both cognitive assessments at baseline.

Assessment Tool Coefficient[β](95% CI) P Value
CEAD-K .077 (.054 – .100) .04
MMSE .082 (.063 – .101) .03
CI = confidence interval

The same baseline associations between RNFL thickness and both assessment tools held true for outer as well as inner macular RNFL thickness.

However, in the longitudinal study, macular RNFL thickness was not associated with decreases in CERAD or MMSE scores during the follow-up period.

Interval, when the researchers set the cutoff value as the lowest quartile (less than 231 mm total RNFL thickness), they found that participants with the thinnest baseline total macular RNFL thickness had the largest decline in CERAD and MMSE scores during the follow-up period (P = .003 and P = .01, respectively) compared with those above that threshold. The same finding held true with outer as well as inner macular RNFL thickness.

Changes in the prevalence of MCI and AD from baseline to follow-up in the below-lowest and above-lowest quartile groups (n = 55 and n = 160, respectively) are shown in the table below.

Quartile MCI Prevalence AD Prevalence

baseline: 27.2%

follow-up: 41.8%

baseline: 7.2%

follow-up: 10.9%


baseline: 6.3%

follow-up: 9.4%

baseline: 1.3%

follow-up: 1.9%

No differences were found in the association of cognitive score and RNFL thickness according to APOE e4 status.

The investigators note several study limitations. For example, only half of the baseline participants continued to the follow-up phase, which might have affected the findings of the longitudinal analysis. Also, the follow-up period lasted for only about 5 years, “which is a significantly small period of time in the development of neurodegenerative diseases,” they write.

In addition, the study did reveal that “OCT-measured baseline macular RNFL thickness was associated with future cognitive decline,” they add.

The researchers propose that “a thinner macular RNFL thickness may predict a decline in cognitive performance” and “a thinner macular RNFL thickness may be considered a noninvasive ocular biomarker for assessing changes in cognitive function in patients.”

However, further population-based investigations with long-term follow-up are necessary, they write.

Inexpensive, Noninvasive

Commenting for Medscape Medical NewsHoward Fillit, MD, cofounder and chief science officer at the Alzheimer’s Drug Discovery Foundation and clinical professor of geriatric medicine and palliative care, medicine, and neuroscience at The Icahn School of Medicine at Mount Sinai in New York, said there is “a lot of interest now in retinal imaging as a technology for detecting early-stage AD.”

The current researchers assessed the retinal layer of the eye, which is particularly populated by neurons and by peripapillary nerve fibers, which come from the brain, he noted.

“So, it’s logical, plausible, and reasonable that what’s happening in the retinal layer of the eye is reflective of neurodegeneration in the brain,” and this was demonstrated in the study, said Fillit, who was not involved with the research.

OCT technology is commonly used in ophthalmological practice, but not for the purpose described in this study.

There are other, more advanced technologies under investigation that may be able to detect amyloid and be used as add-ons to OCT devices, such as hyperspectral camera imaging, that might “find its place in the algorithm” for AD diagnoses, Fillit said.

“Imagine a world where older people going for their annual eye checkup could have a 5- to 10-minute examination to screen for AD. Or, if they already have memory problems, a diagnosis of AD could be confirmed based on the presence of amyloid plaques at the back of the eye,” Fillit said. Although the technology is not currently available in mainstream clinical practice, when it is, it will be “relatively inexpensive and noninvasive,” he added.

In an accompanying editorial, Rajendra Apte, MD, PhD, professor of ophthalmology and visual sciences and vice chair for innovation and translation, Washington University School of Medicine, St. Louis, Missouri, writes that the study “contributes to our evolving knowledge of the ability of retinal imaging to identify biomarkers that predict development of cognitive impairment and dementia.”

JAMA Ophthalmol. Published online May 26, 2022. Abstract, Editorial

This research was supported by a research grant from Seoul National University Bundang Hospital, a National Research Foundation grant funded by the Korean government, and a grant from the Korean Health Technology R&D Project, Ministry of Health and Welfare, Republic of Korea. The investigators and Fillit have reported no relevant financial relationships. The Alzheimer’s Drug Discovery Foundation is invested in retinal programs but was not involved with this study. Apte reported having received research funding and Department of Ophthalmology funding from the Jeffrey Fort Innovation Fund, the Starr Foundation, and an unrestricted grant from Research to Prevent Blindness.

For more Medscape Psychiatry news, join us on Twitter and Facebook.


Leave a Comment