The Retinal Vessel Density Can Reflect Cognitive Function in Patients with Alzheimer’s Disease: Evidence from Optical Coherence Tomography Angiography

Abstract

Background:

There is increasing evidence that Alzheimer’s disease (AD) patients may present decreased cerebral blood perfusion before pathological brain changes. Using the retina as a window to the brain, we can study disorders of the central nervous system through the eyes.

Objective:

This study aimed to investigate differences in retinal structure and vessel density (VD) between patients with mild AD and healthy controls (HCs). Furthermore, we explored the relationship between retinal VD and cognitive function.

Methods:

We enrolled 37 patients with AD and 29 age-matched HCs who underwent standard ophthalmic optical coherence tomography angiography (OCTA) for evaluation of the retinal layer thickness and VD parameters. Cognitive function was evaluated using a battery of neuropsychological assessments. Finally, the correlations among retinal layer thickness, VD parameters, and cognitive function were evaluated.

Results:

The retinal fiber layer thickness and retinal VD of patients with AD were significantly reduced compared with HCs. The retinal VD was significantly correlated with overall cognition, memory, executive, and visual-spatial perception functions. However, there was no significant between-group difference in the macular thickness.

Conclusion:

Our findings indicate a positive correlation between retinal VD and some, but not all, cognitive function domains. Most importantly, we demonstrated the role of OCTA in detecting early capillary changes, which could be a noninvasive biomarker for early AD.

Keywords

Alzheimer’s disease biomarker cognitive function optical coherence tomography angiography vessel density

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disease with an unclear onset and progressive development; it is a common cause of dementia [1]. AD is clinically characterized by impairment of memory, visual-spatial, and executive function, as well as personality and behavioral changes [1]. AD prevalence has been projected to triple by 2050 given the increasingly aging population [2]. Currently, the clinical diagnosis of chronic cognitive decline mainly relies on comprehensive examinations, including cognitive assessment, blood tests, magnetic resonance imaging (MRI), and positron emission tomography [3]. However, most are invasive and expensive procedures. Given the increasing incidence of AD and the lack of effective treatment options, there is a need for simple and non-invasive biomarkers for AD detection and diagnosis.

The cerebral pathology underlying AD may be reflected in the eye given the homology between the retina and the central nervous system [4, 5]. Brain lesions in patients with AD are characterized by Aβ deposition and tau-protein hyperphosphorylation in neurons, as well as existing cerebrovascular microcirculation disorder, which is directly associated with neuronal fiber tangles [6]. In addition, retinal micro-vasculature and cerebrovascular (including terminal arteries without anastomosis, barrier function, etc.) share similar characteristics [5]. Several studies have shown that characteristic pathological changes in patients with AD affect the retina [7 –11]. Therefore, the retina can be studied in the assessment of central neuron damage, brain volume, and the brain’s microcirculation system in patients with AD. This examination can be achieved using optical coherence tomography (OCT). The fundus parameters detected by non-invasive optical imaging tools could be promising biomarkers for early AD diagnosis.

OCT is a non-invasive retinal imaging technology that can be used to study retinal structural changes in patients with AD [12]. It has been shown that patients with AD have significantly reduced peripa-pillary retinal fiber layer (pRNFL) and macular thickness [13 –15]. Macular thickness is associated with reduced gray matter volume in the occipital and temporal lobe [16], and it has been shown that decreased RNFL could be associated with atrophy of the cingulate cortex and hippocampus in individuals without dementia [17, 18]. Moreover, a study reported an ass-ociation between changes in retinal structure and cognitive decline [19]. Further, patients with AD might present decreased cerebral blood perfusion before pathological brain changes; therefore, cerebral blood flow could be a useful marker that reflects an inc-reased risk of cognitive decline [20 –23]. There has been increasing attention on retinal microvasculature in patients with AD, and OCTA can provide a lot of information regarding these changes; it can allow co-mprehensive analysis of the retinal microvascular anatomy in patients with AD and provide a highly specific, sensitive, simple, replicable, and non-inva-sive method for early AD diagnosis [11 , 24].

There have been many studies on the retinal structure, but not retinal microvasculature, in patients with AD. Further, the relationship between retinal micro-vascular changes and multi-domain cognitive func-tion in patients with AD remains unclear. OCTA images obtained through continuous OCT cross-sec-tional scans allow three-dimensional reconstruction of the retinal and choroidal perfusion microvascular, which is an effective indicator for changes in fundus blood vessels in patients with AD [25]. Previous OCTA studies on patients with AD have reported changes in the optic disc and macular vascular density (VD); however, these reports were inconsistent and require further verification [7 , 26–28].

In addition, previous studies on patients with AD and healthy elderly individuals have indicated that RNFL and macular thickness are positively correlated with some cognitive function domains [19 , 30]. Only a few studies have confirmed the relationship between retinal microvasculature and overall cognitive function in patients with AD [31]. Most of these previous studies assessed cognitive function using the Mini-Mental State Examination (MMSE), which does not cover all cognitive domains affected in patients with AD. Consequently, this study aimed to determine the relationship between retinal VD in specific quadrants and multiple cognitive domains in patients with mild AD measured using a comprehensive neuropsychological assessment.

Specifically, this study aimed to investigate differences in the retinal structure and VD between patients with mild AD and healthy elderly individuals. Moreover, we aimed to further explore the relationship of the optic disc and macular VD with cognitive function in patients with AD.

MATERIALS AND METHODS

Participants

Between March 2019 and January 2020, partici-pants (n = 66) were enrolled from the Memory Disorders Clinic and the surrounding communities of the First Affiliated Hospital of Anhui Medical University, Hefei, Anhui, China. The participants were div-ided into the following two groups: the AD group (n =37), which included participants with mild AD; and the healthy control group (HC): (n = 29), which included healthy elderly individuals with normal cognitive function. All the participants underwent OCTA examination and neuropsychological assessments.

The inclusion criteria of participants with mild AD were as follows: 1) the core criteria for probable mild AD according to the National Institute of Neurological and Communicative Diseases and Str-oke/Alzheimer’s Disease and Related Disorders Association [32, 33]; 2) an MMSE score of 10–24; 3) a Clinical Dementia Rating (CDR) score between 0.5 and 1 [34]; 4) age between 50 and 85 years without limitations for sex and being right-handed; and 5) having been examined in the first affiliated hospital of Anhui Medical University by an eye specialist and lacking diabetic retinopathy, serious cataracts, glaucoma, age-related macular degeneration, optic neuropathy, and high refractive error (plus or minus 3D diopter). The inclusion criteria for the control group were a lack of AD clinical signs and being matched with the AD group according to age, sex, and education.

The exclusion criteria were as follows: 1) history of eye disease (e.g., eye trauma, eye surgery, and significant media opacity); 2) history of diabetes mellitus, uncontrolled hypertension, or other serious chronic medical conditions; 3) having other secon-dary dementia; 4) having organic brain defects on T1 or T2 images and a history of other neurologic disease; 5) any history or clinical signs of other severe psychiatric illnesses (such as major depression disorder); and 6) history of alcohol or drug abuse.

All the participants had normal hearing and normal vision or correction. This study was approved by the Anhui Medical Ethics Committee. All the experiments were conducted with the consent of the patient and their family, and written informed consent was obtained in accordance with the Declaration of Helsinki.

OCT angiography image acquisition

The True XR OCT equipment (Optovue, Inc., Fremont, CA, USA) was used to scan the peripapillary and macula of all the participants. The RTVue XR OCT Avanti with AngioVue Software (2016.2.0 version) was used to visualize vascular structures of the retina and choroid; it can provide quantitative VD and automatically locate each user-defined retinal layer. AngioVue scans of the optic disc and macular area comprise two back-to-back 3D scans that each have aiming, capture and quality review phases. It uses 840 nm diode laser sources at a scanning rate of 70,000 A-scans per second and provides noninvasive information regarding the vascular structure.

OCTA images of the peripapillary area and macula were obtained using a 4.5×4.5 mm² scan and a 3×3 mm² scan, respectively. After the scan was complete, the operator assessed the quality of the en face images and rescanned in case of excessive motion that cannot be corrected through software. In addition, the ophthalmologist checked the image quality and removed images with a quality index of <4/10 and images with motion artifacts or black lines caused by sudden eye movements [17].

The peripapillary area was divided into eight sectors based on the Garway-Heath map (superotem-poral, superonasal, inferotemporal, inferonasal, nasal superior, nasal inferior, temporal superior, and temporal inferior, [ST/SN/IT/IN/NS/NI/TS/TI]) in the radial peripapillary capillary (RPC) layer, which extends from the inner limiting membrane to the posterior margin of the RNFL [35]. Further, it was divided into two sectors based on Hemisphere Maps (Superior and Inferior Hemispheres, [S-Hemi/I-Hemi]).

Additionally, all macular vasculatures were ob-tained through the deep capillary plexus (DCP) and surface capillary plexus (SCP). Both superficial and deep plexus images were divided into two sectors (S-Hemi and I-Hemi) or four quadrants (superior, in-ferior, nasal, and temporal, [S/I/N/T]) based on Hemisphere and Quadrant Maps. Further, the full macula thickness in the above area was measured (Fig. 1).

Fig. 1

1) HD Angio Disc mode (optic disc) Report Layout Legend: a) Garway-Heath map; b) Hemisphere Maps; 2) Angio Retina mode (macular) Report Layout Legend: c, d) Hemispher and Quadrant Maps.

Neuropsychological assessments

Senior neuropsychological graduate students in the department of neurology performed all neuropsychological assessments immediately before or after OCT data acquisition. The following neuropsychological test battery was administered for comprehensive evaluation of cognitive function and clinical symptoms [36]: MMSE, Montreal Cognitive Assessment (MoCA), Lawton-Brody Activities of Daily Living scale, CDR, Hachinski Ischemic Scale (His), Hamilton anxiety scale (HAMA), and Hamilton Depression scale (HAMD). Individual cognition domains were assessed using the following tests: Chinese’s version of the Auditory Verbal Learning Test (CAVLT- immediate, delay, and recall, [CAVLT-I/D/R]), Digital Span Test (forward/backward, [DST-F/B]), Stroop Color-Word Test (SCWT-dot, word, and colored word, [SCWT-D/W/CW]), Trail Making Test A/B (TMT A/B), Clock-Drawing Test (CDT), and Verbal Fluency Test (VFT–letter and semantic, [VFT-L/S]).

Statistical analysis

Measurement data were expressed as mean± standard deviation $(\bar{x} \pm s)$ . Between-group comparisons for normally and non-normally distributed measurement data were performed using the independent sample t test and the Mann–Whitney U test, respectively. Between-group sex differences (categorical data) were analyzed using the χ ² test. The relationships of each of the averaged VD parameters in the optic disc and macula with the neuropsychological assessment scores of patients with AD were explored using partial correlation (age and sex were the calibration control variables), and multiple linear regression was used to further verify the results of partial correlation analysis. All statistical analyses were conducted using SPSS for Mac (23.0, IBM) and statistical significance was set at p < 0.05 (two-tailed).

RESULTS

We included 74 eyes of 37 patients with mild AD and 58 eyes of 29 HCs. After exclusion of poor quality OCTA scans (<4/10) and images with motion artifacts, 58 eyes of 34 patients with mild AD and 58 eyes of 29 HCs were included in the analysis of optic disc parameters, while 67 eyes of 37 patients with mild AD and 56 eyes of 29 HCs were included in the analysis of macula parameters. There was no significant between-group difference in the demographic variables (age, years of education, sex ratio), His, HAMA, and HAMD (all p > 0.05). The AD group had significantly worse scores for the remaining neuropsychological tests than those in the HC group (all p < 0.05) (Supplementary Table 1).

The RNFL thickness of the peripapillary area was significantly smaller in the AD group in the following sectors: Peripapillary (p = 0.002), S-Hemi (p = 0.012), I-Hemi (p = 0.003), SN (p = 0.049), NS (p = 0.002), NI (p = 0.002), IN (p < 0.001) (Table 1). There were no significant between-group differences in macular thickness (all p > 0.05) (Table 2). Compared with the HC group, the AD group had significantly decreased peripapillary RPC VD in the following sectors: Whole Image (p = 0.001), Peripapillary (p = 0.010), S-Hemi (p = 0.003), I-Hemi (p = 0.013), NS (p = 0.015), NI (p = 0.005), IN (p = 0.003), TI (p = 0.026), TS (p = 0.009), ST (p =0.044), and SN (p = 0.017) (Table 1). Further, when compared with the HC group the AD group had a significantly reduced macular SCP VD in the following sectors: Parafovea (p = 0.018), I-Hemi (p = 0.026), T (p = 0.007), N (p = 0.003) (Table 2). The AD group had a significantly reduced macular DCP VD only in the I-Hemi (p = 0.049) and temporal sectors (p = 0.013) (Table 2). There were no significant between-group differences in macular VD in the other sectors (p > 0.05) (Table 2).

Table 1

OCTA parameters of optic disc in subjects with AD and HC $(\bar{x} \pm s)$

OCTA parameters	AD (n = 34)	HC (n = 29)	T/Z value	p
Thickness (μm)
Peripapillary^b	105.03±14.367	112.46±10.080	– 3.186	0.002
GH S-Hemi^c	105.98±14.098	111.84±11.247	– 2.500	0.012
GH I-Hemi^c	102.08±19.236	113.08±11.675	– 2.962	0.003
GH-TS^c	77.14±29.93	75.79±10.370	–1.320	0.187
GH-ST^b	123.21±26.770	129.92±18.265	–1.558	0.122
GH-SN^b	131.21±23.205	140.21±24.782	– 1.993	0.049
GH-NS^c	97.90±18.38	107.66±15.688	– 3.061	0.002
GH-NI^b	82.87±19.293	94.80±20.693	– 3.157	0.002
GH-IN^b	124.22±30.052	143.47±22.205	– 3.924	P < 0.001
GH-IT^c	133.90±31.465	144.26±20.710	–1.485	0.137
GH-TI^b	69.78±14.811	68.66±11.912	0.446	0.656
Vessel density (%)
Whole Image_ RPC^b	48.15±4.327	50.37±2.745	– 3.294	0.001
Peripapillary_ RPC^c	50.77±4.779	53.22±2.752	– 2.562	0.010
GH S-Hemi_ RPC^c	51.03±4.561	53.34±3.001	– 2.984	0.003
GH I-Hemi_ RPC^c	49.54±7.538	53.08±3.111	– 2.490	0.013
GH-NS_RPC^c	47.17±6.500	49.91±3.667	– 2.421	0.015
GH-NI_RPC^b	46.47±6.739	49.52±4.130	– 2.888	0.005
GH-IN_RPC^b	49.06±9.324	53.16±4.333	– 3.040	0.003
GH-IT_RPC^c	54.24±10.160	58.15±4.388	–1.836	0.066
GH-TI_RPC^b	50.49±6.041	52.69±4.212	– 2.254	0.026
GH-TS_RPC^b	53.98±4.766	56.19±4.145	– 2.657	0.009
GH-ST_RPC^c	54.15±6.598	56.32±4.360	– 2.018	0.044
GH-SN_RPC^b	49.80±5.271	52.00±4.389	– 2.416	0.017

^bIndependent sample t test; ^cMann–Whitney U test. AD, Alzheimer’s disease; GH, Garway-Heath map; RPC, radial peripapillary capillary; S-Hemi, Superior Hemispheres; I-Hemi, Inferior Hemispheres; ST, Superotemporal; SN, Superonasal; IT, Inferotemporal; IN, Inferonasal; NS, Nasal Superior; NI, Nasal Inferior; TS, Temporal Superior; TI, Temporal Inferior.

Table 2

OCTA parameters of macular in subjects with AD and HC $(\bar{x} \pm s)$

OCTA parameters	AD (n = 37)	HC (n = 29)	T /Z value	p
Thickness (μm)
Foveal (1)^c	245.90±26.284	246.95±20.121	–1.590	0.112
T (1–3)^c	303.14±22.031	307.94±24.757	–1.783	0.075
S (1–3)^c	314.05±23.579	319.39±26.919	–1.743	0.081
N (1–3)^c	313.31±27.295	319.05±24.389	–1.901	0.057
I (1–3)^c	312.68±27.279	317.69±24.792	–1.755	0.079
S Hemi (1 –3)^c	310.94±23.597	316.34±25.727	–1.874	0.061
I Hemi (1–3)^c	310.50±25.095	315.66±24.841	–1.854	0.064
Vessel density (%)
Surface Capillary Plexus
L1-Fovea^b	15.80±6.975	15.94±6.264	–0.118	0.907
L1-ParaFovea^c	46.62±5.124	48.61±3.790	– 2.358	0.018
L1-Para-S-Hemi^c	46.64±5.452	48.40±4.057	–1.962	0.050
L1-Para-I-Hemi^c	46.69±5.446	48.83±3.777	– 2.225	0.026
L1-Para-T^c	45.58±5.428	48.01±3.243	– 2.689	0.007
L1-Para-S^c	48.37±5.979	49.49±4.404	–0.890	0.373
L1-Para-N^c	45.79±4.860	48.21±3.797	– 2.944	0.003
L1-Para-I^c	46.88±6.464	48.78±4.783	–1.582	0.114
Deep Capillary Plexus
L2-Fovea^c	28.80±8.147	28.90±8.298	–0.152	0.879
L2-ParaFovea^b	51.57±3.680	52.63±3.863	–1.551	0.124
L2-Para-S-Hemi^c	51.54±4.394	52.46±4.270	–1.007	0.314
L2-Para-I-Hemi^b	51.49±3.623	52.80±3.602	– 1.990	0.049
L2-Para-T^b	51.52±4.242	53.28±3.171	– 2.536	0.013
L2-Para-S^c	51.74±4.564	52.49±4.544	–0.825	0.409
L2-Para-N^b	52.52±3.887	53.54±3.512	–1.505	0.135
L2-Para-I^c	50.38±4.916	51.25±5.126	–1.109	0.267

^bIndependent sample t test; ^cMann–Whitney U test. AD, Alzheimer’s disease; L1, Surface Capillary Plexus; L2, Deep Capillary Plexus; Para, Parafovea; S-Hemi, Superior Hemispheres; I-Hemi, Inferior Hemispheres; S, Superior; I, Inferior; N, Nasal; T, Temporal.

Fig. 2

Example of fundus angiograms OCTA measurements taken from the left eye of: a) a healthy control with all quadrants with a thickness and vasculature normal for their age; and b) an AD patient with decreased RFNL thickness in the temporal quadrant (from left to right: optic disc thickness, peripapillary vascular density, and the macular superficial and deep vascular density).

We analyzed the correlations between optic disc OCTA parameters and cognitive function test scores using partial correlation coefficients. After calibra-tion control for age and sex, the S-Hemi sector of peripapillary RPC VD was significantly correlated with MMSE (r = 0.418, p = 0.019), MoCA (r = 0.382, p = 0.037), CAVLT-I (r = 0.370, p = 0.048), TMT-A (r = –0.501, p = 0.034), and TMT-B (r = –0.518, p =0.048) scores, whereas the NS sector of peripapillary RPC VD was significantly correlated with CAVLT-I (r = 0.480, p = 0.007) and TMT-B (r = –0.517, p =0.041) scores. The SN sector of the peripapillary RPC VD was significantly correlated with MMSE (r = 0.458, p = 0.010), MoCA (r = 0.529, p = 0.003), CAVLT-I (r = 0.428, p = 0.021), TMT-A (r = –0.665, p = 0.003), TMT-B (r = –0.603, p = 0.017), and CDT (r = 0.560, p = 0.002) scores among all patients with mild AD (Supplementary Table 2). There were no significant correlations between the other sectors’ VD parameters and cognitive function test scores (all p > 0.05, Supplementary Table 2).

We used partial correlation coefficients to analyze the correlations between the macular OCTA parameters and cognitive function test scores of the AD group. After calibration control for age and sex, only the I-Hemi (r = –0.353, p = 0.047) sector of macular SCP VD was significantly correlated with the SCWT-W score in all patients with mild AD (Supplementary Table 3). There were no significant correlations between other sectors of the macular SCP VD and cognitive function test scores (all p > 0.05, Supplementary Table 3). The Parafovea (r = –0.376, p = 0.031), S-Hemi (r = –0.415, p = 0.016), T (r =–0.350, p = 0.046), and N (r = –0.360, p = 0.039) sectors of macular DCP VD were significantly correl-ated with SCWT-D scores of patients with mild AD. The Parafovea (r = –0.437, p = 0.012), S-Hemi (r =–0.424, p = 0.016), I-Hemi (r = –0.377, p = 0.033), T (r = –0.381, p = 0.031), S (r = –0.399, p = 0.024), and N (r = –0.407, p = 0.021) sectors of macular DCP VD were significantly correlated with SCWT-W scores of patients with mild AD. The Parafovea (r = –0.366, p = 0.043), T (r = –0.369, p = 0.041), and N (r = –0.369, p = 0.041) sectors of macular DCP VD were significantly correlated with SCWT-CW scores of patients with mild AD. The S-Hemi of macular DCP VD was significantly correlated with the TMT-B scores (r = –0.535, p = 0.027) of patients with mild AD (Supplementary Table 4). There were no significant correlations between macular DCP VD in other sectors and cognitive function test scores (all p > 0.05, Supplementary Table 4).

We verified the results of partial correlation analysis by using multiple linear regression, and found that although the results were slightly different, the correlation between retinal vessel density and overall cognition, execution, and visual space remained. We found that the SN sector of peripapillary RPC VD can predict overall cognitive, executive, and visuospatial functions, whereas the NS and TS sectors of peripapillary RPC VD can predict memory function, and the NI sector of peripapillary RPC VD can predict visuospatial function. Further, the T and S-Hemi sectors of macular SCP VD can predict executive function, and the T sectors of macular DCP VD can predict executive function (see the Supplementary Material for details).

DISCUSSION

This study showed that the AD group had signifi-cantly reduced pRNFL thickness and VD of peripapillary and macula compared with the HC group. Decreased pRNFL thickness and VD in patients with mild AD is suggestive of retinal neuron and microvascular system damage, which corresponds to pathological brain damage [4 , 37]. This suggests a novel method for assessing the effect of neuronal and microvasculature system damage on mild patients with AD.

The retina comprises specialized neuronal layers connected by synapses, including retinal ganglion cells (RGCs). Light that enters the eye is captured by photoreceptor cells in the outermost retinal layer, initiating a cascade of neuronal signals that eventually reach the RGCs, the axons of which form the optic nerve [4, 37]. According to previous reports, the optic nerves of many patients with AD exhibit predomin-ant loss of RGCs, which contribute to large caliber fibers to the optic nerve [8, 38]. Therefore, decreased pRNFL and macular thickness can be present in pa-tients with mild AD. Consistent with our findings, previous studies have reported significant thinning of pRNFL and macula in patients with mild AD [14 , 39–41]. Although we observed a numerical, but not significant, reduction of the full macular thickness in the AD group, most previous studies have shown that patients with AD have significantly thinner macular GC-IPL [42]. These inconsistent findings could be attributed to measurement errors or mean differences.

Moreover, we found that VD in the peripapillary area (except IT), macular SCP (I-Hemi, T, and N), and DCP (I-Hemi and T) was significantly lower in patients with mild AD than in HCs. Our findings indicate that VD analysis of the retinal using OCTA images could help identify retinal VD loss before the manifestation of apparent clinical retinopathy in patients with mild AD. The brain and retina microvasculature are homologous and their blood supplies are anatomically closely associated, with both vascular networks sharing similar vascular regulatory processes; therefore, changes in cerebral blood vessels could be reflected by the retinal vessels [5]. Moreover, age-related vascular changes have been shown to accompany or even precede AD development and vascular pathology may be a pathogenic factor for AD [20, 43]. Thus, as indicated by previous reports, reduced retinal VD can be found in the early stages of AD [28 , 44–46]. However, there have been inconsistent reports regarding the areas in which retinal VD is significantly reduced, which could be attributed to differences in the parameters used in image acquisition. Our findings suggest that the peripapillary area (except the IT), macular SCP (I-Hemi, T, and N), and DCP (I-Hemi, T) could be good biomarkers to detect and monitor the early cognitive changes of AD. However, the cross-sectional design of this study does not allow positive conclusions.

Furthermore, our findings indicate that decreased retinal VD is associated with cognitive function in-volved in the prefrontal, parietal, and temporal cortex. We observed a significant positive correlation between retinal VD and specific cognitive functions in patients with AD that had not been previously reported. We found that peripapillary RPC VD was significantly correlated with the MMSE, MoCA, CAVLT, TMT, and CDT; macular SCP VD was significantly correlated with the SCWT; and macular DCP VD was significantly correlated with the SCWT and TMT. The MMSE and MoCA tests are used to evaluate overall cognition; the CAVLT test is used to evaluate memory and re-extraction function; the SCWT and TMT are used to evaluate executive function; and the CDT is used to evaluate visual-spatial function [36]. The frontal, parietal, and temporal lobe form the brain’s attention, memory, and execution network. Damage to these brain structures can induce abnormal cognitive behavior that has been reported in patients with mild AD [47]. Further, regionally reduced cerebral blood flow in patients with AD is observed mainly in frontal, temporal, parietal, and medial-temporal regions; these regions overlap with those showing consistent atrophy in later AD stages [21]. Therefore, these studies further confirm our hypothesis.

Several studies have analyzed the correlation between retinal thickness and cognitive function in patients with AD and healthy elderly people. They found that retinal thickness could predict cognitive domains and help identify individuals at risk or in preclinical stages of AD [19 , 30]. Moreover, other studies have reported an association between the retinal structure and brain volume [16–18 , 45]. Recent reports have suggested that changes in cerebral blood flow in patients with AD occur before pathological brain changes; further, cerebral vascular dysfunctions are early and pivotal contributors to AD development and cerebral amyloid angiopathy is a reliable predictor of cognitive decline [20 , 48]. The retina is homologous with the central nervous system and can be used from brain assessment [4]. Therefore, changes in retinal microvasculature reflect changes in cerebral microvasculature in patients with AD. Decreased retinal VD in patients with mild AD could be an imaging biomarker for screening individuals with AD symptoms.

Notably, using the same area division method, there were more areas with significant differences in peripapillary and macular VD than in pRNFL and macular thickness. This indicates that peripapillary and macular VD are more sensitive than pRNFL and macular thickness as biomarkers for early diagnosis of mild AD. However, this should be verified by a receiver operating characteristic curve analysis.

This cross-sectional study had a small sample size and lacked further verification via MRI imaging analysis. Therefore, we could not assess the va-lue of retinal VD measurements in the assessment of AD progression. However, our findings confirm the decreased pRNFL thickness and retinal microvascular density in patients with mild AD, as well as the positive correlation between retinal VD and some cognitive function domains (e.g., overall cognition, memory, executive, and visual-spatial perception function). There is a need for further longitudinal studies with larger sample sizes to allow early diagnosis of patients with AD. Finally, OCTA allows rapid, noninvasive, and quantitative evaluation of the retinal morphology and microvascular, which helps translate multimodal eye data into brain function indicators and provides a more sensitive biomarker for early AD diagnosis.

CONCLUSION

In conclusion, our study showed that compared with the normal age-matched HCs, patients with mild AD had significantly reduced pRNFL thickness, peripapillary RPC VD, macular SCP, and DCP VD, which might represent a potential noninvasive biomarker for early AD. Furthermore, retinal blood flow density was associated with neuropsychological performance in multiple cognitive domains.

Larger cohort and longitudinal studies are required in the future to explore the temporal relationship of retinal microvascular damage and pathological RNFL loss with cognitive decline.

AVAILABILITY OF DATA AND MATERIAL

All datasets generated for this study are included in the manuscript and/or the supplementary files. The data that support the findings of this study are available from the corresponding authors upon request.

Footnotes

ACKNOWLEDGMENTS

We thank the participants for their cooperation during this study. This work was supported by the National Key R&D Program of China (nos. 2016YFC1306400, 2018YFC1314504, 2016YFC1305904, and 2018YFC1314200); the Natural Science Foundation of China (nos. 31970979).

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Lane

, Hardy

, Schott

(2018) Alzheimer’s disease. Eur J Neurol 25, 59–70.

Hebert

, Weuve

, Scherr

, Evans

(2013) Alzheimer disease in the United States (2010–2050) estimated using the 2010 census. Am Acad Neurol 80, 1778–1783.

Reitz

, Mayeux

(2014) Alzheimer disease: Epidemiology, diagnostic criteria, risk factors and biomarkers. Biochem Pharmacol 88, 640–651.

London

, Benhar

, Schwartz

(2013) The retina as a window to the brain-from eye research to CNS disorders. Nat Rev Neurol 9, 44–53.

Patton

, Aslam

, Macgillivray

, Pattie

, Deary

, Dhillon

(2005) Retinal vascular image analysis as a potential screening tool for cerebrovascular disease: A rationale based on homology between cerebral and retinal microvasculatures. J Anat 206, 319–348.

Kalaria

(2000) The role of cerebral ischemia in Alzheimer’s disease. Neurobiol Aging 21, 321–330.

Golzan

, Goozee

, Georgevsky

, Avolio

, Chatterjee

, Shen

, Gupta

, Chung

, Savage

, Orr

, Martins

, Graham

(2017) Retinal vascular and structural changes are associated with amyloid burden in the elderly: Ophthalmic biomarkers of preclinical Alzheimer’s disease. Alzheimers Res Ther 9, 13–22.

Sadun

, Bassi

(1990) Optic nerve damage in Alzheimer’s disease. Ophthalmology 97, 9–17.

, Vince

(2015) Hyperspectral imaging signatures detect amyloidopathy in Alzheimer’s mouse retina well before onset of cognitive decline. ACS Chem Neurosci 6, 306–315.

10.

, Leung

, Tsang

, So

, Chiu

, Chang

(2012) Review: Tauopathy in the retina and optic nerve: Does it shadow pathological changes in the brain? Mol Vis 18, 2700–2710.

11.

Cabrera DeBuc

, Somfai

, Koller

(2017) Retinal microvascular network alterations: Potential biomarkers of cerebrovascular and neural diseases. Am J Physiol Heart Circ Physiol 312, 201–212.

12.

Husvogt

, Ploner

, Maier

(2018) Optical coherence tomography. In Medical Imaging Systems: An Introductory Guide, Maier A, Steidl S, Christlein V, Hornegger J, eds., Cham (CH), pp. 251-261.

13.

Doustar

, Torbati

, Black

, Koronyo

, Koronyo-Hamaoui

(2017) Optical coherence tomography in Alzheimer’s disease and other neurodegenerative diseases. Front Neurol 8, 701–714.

14.

Gao

, Liu

, Li

, Bai

, Liu

(2015) Abnormal retinal nerve fiber layer thickness and macula lutea in patients with mild cognitive impairment and Alzheimer’s disease. Arch Gerontol Geriatr 60, 162–167.

15.

Ascaso

, Cruz

, Modrego

, Lopez-Anton

, Santabarbara

, Pascual

, Lobo

, Cristobal

(2014) Retinal alterations in mild cognitive impairment and Alzheimer’s disease: An optical coherence tomography study. J Neurol 261, 1522–1530.

16.

Liu

, Ong

, Hilal

, Loke

, Wong

, Chen

, Cheung

, Zhou

(2016) The association between retinal neuronal layer and brain structure is disrupted in patients with cognitive impairment and Alzheimer’s disease. J Alzheimers Dis 54, 585–595.

17.

Shi

, Zheng

, Hu

, Jiang

, Cao

, Chen

, Mei

, Li

, Shen

(2019) Retinal nerve fiber layer thinning is associated with brain atrophy: A longitudinal study in nondemented older adults. Front Aging Neurosci 11, 69–79.

18.

Shi

, Cao

, Hu

, Jiang

, Mei

, Zheng

, Chen

, Wang

, Cao

, Li

, Shen

(2019) Retinal nerve fiber layer thickness is associated with hippocampus and lingual gyrus volumes in nondemented older adults. Prog Neuropsychopharmacol Biol Psychiatry 99, 109824.

19.

Almeida

ALM

, Pires

, Figueiredo

, Costa-Cunha

LVF

, Zacharias

, Preti

, Monteiro

MLR

, Cunha

(2019) Correlation between cognitive impairment and retinal neural loss assessed by swept-source optical coherence tomography in patients with mild cognitive impairment. Alzheimers Dement (Amst) 11, 659–669.

20.

Cortes-Canteli

, Iadecola

(2020) Alzheimer’s disease and vascular aging: JACC Focus Seminar. J Am Coll Cardiol 75, 942–951.

21.

Kim

, Alvarado

, Stephens

, Wey

, Wang

DJJ

, Leritz

, Salat

(2020) Associations between cerebral blood flow and structural and functional brain imaging measures in individuals with neuropsychologically defined mild cognitive impairment. Neurobiol Aging 86, 64–74.

22.

Ruitenberg

, den Heijer

, Bakker

, van Swieten

, Koudstaal

, Hofman

, Breteler

(2005) Cerebral hypoperfusion and clinical onset of dementia: The Rotterdam Study. Ann Neurol 57, 789–794.

23.

Brown

, Thore

(2011) Review: Cerebral microvascular pathology in ageing and neurodegeneration. Neuropathol Appl Neurobiol 37, 56–74.

24.

Yang

, Yuan

, Wang

, Chen

(2019) Comparison of the repeatability of macular vascular density measurements using four optical coherence tomography angiography systems. J Ophthalmol 2019, 4372580.

25.

Puliafito

(2014) OCT angiography: The next era of OCT technology emerges. Ophthalmic Surg Lasers Imaging Retina 45, 360.

26.

Cheung

, Ikram

, Chen

, Wong

(2017) Imaging retina to study dementia and stroke. Prog Retin Eye Res 57, 89–107.

27.

O’Bryhim

, Apte

, Kung

, Coble

, Van Stavern

(2018) Association of preclinical Alzheimer disease with optical coherence tomographic angiography findings. JAMA Ophthalmol 136, 1242–1248.

28.

Yoon

, Grewal

, Thompson

, Polascik

, Dunn

, Burke

, Fekrat

(2019) Retinal microvascular and neurodegenerative changes in Alzheimer’s disease and mild cognitive impairment compared with control participants. Ophthalmol Retina 3, 489–499.

29.

Mammadova

, Neppl

, Denburg

, West Greenlee

(2020) Reduced retinal thickness predicts age-related changes in cognitive function. Front Aging Neurosci 12, 81–87.

30.

, Muthy

, Gallacher

, Sudlow

, Rees

, Yang

, Keane

, Petzold

, Khaw

, Reisman

, Strouthidis

, Foster

, Patel

, UK Biobank Eye & Vision Consortium (2018) Association of retinal nerve fiber layer thinning with current and future cognitive decline: A study using optical coherence tomography. JAMA Neurol 75, 1198–1205.

31.

Zhang

, Zhou

, Knoll

, Samra

, Ward

, Weintraub

, Fawzi

(2019) Parafoveal vessel loss and correlation between peripapillary vessel density and cognitive performance in amnestic mild cognitive impairment and early Alzheimer’s disease on optical coherence tomography angiography. PLoS One 14, e0214685.

32.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr. , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

33.

McKhann

, Drachman

, Folstein

, Katzman

, Price

, Stadlan

(1984) Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group* under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 34, 939–944.

34.

Morris

(1993) The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 43, 2412–2412.

35.

Garway-Heath

, Poinoosawmy

, Fitzke

, Hitchings

(2000) Mapping the visual field to the optic disc in normal tension glaucoma eyes. Ophthalmology 107, 1809–1815.

36.

Kessels

RPC

, Henriks

MPH

(2016) Neuropsychological assessment. Encyclopedia of Mental Health (2nd ed.), Friedman HW, ed. Elsevier, Oxford, pp. 197–201.

37.

Wässle

(2004) Parallel processing in the mammalian retina. Nat Rev Neurosci 5, 747–757.

38.

Parisi

, Restuccia

, Fattapposta

, Mina

, Bucci

, Pierelli

(2001) Morphological and functional retinal impairment in Alzheimer’s disease patients. Clin Neurophysiol 112, 1860–1867.

39.

Chan

VTT

, Sun

, Tang

, Chen

, Wong

, Tham

, Wong

, Chen

, Ikram

, Whitson

, Lad

, Mok

VCT

, Cheung

(2019) Spectral-domain OCT measurements in Alzheimer’s disease: A systematic review and meta-analysis. Ophthalmology 126, 497–510.

40.

Yap

, Balendra

, Almonte

, Cordeiro

(2019) Retinal correlates of neurological disorders. Ther Adv Chronic Dis 10, 2040622319882205.

41.

Alber

, Goldfarb

, Thompson

, Arthur

, Hernandez

, Cheng

, DeBuc

, Cordeiro

, Provetti-Cunha

, den Haan

, Van Stavern

, Salloway

, Sinoff

, Snyder

(2020) Developing retinal biomarkers for the earliest stages of Alzheimer’s disease: What we know, what we don’t, and how to move forward. Alzheimers Dement 16, 229–243.

42.

Cheung

, Ong

, Hilal

, Ikram

, Low

, Ong

, Venketasubramanian

, Yap

, Seow

, Chen

, Wong

(2015) Retinal ganglion cell analysis using high-definition optical coherence tomography in patients with mild cognitive impairment and Alzheimer’s disease. J Alzheimers Dis 45, 45–56.

43.

Shi

, Koronyo

, Rentsendorj

, Regis

, Sheyn

, Fuchs

, Kramerov

, Ljubimov

, Dumitrascu

, Rodriguez

, Barron

, Hinton

, Black

, Miller

, Mirzaei

, Koronyo-Hamaoui

(2020) Identification of early pericyte loss and vascular amyloidosis in Alzheimer’s disease retina. Acta Neuropathol 139, 813–836.

44.

, Zhang

, Azhati

, Li

, Xu

, Liu

(2020) Retinal microvascular attenuation in mental cognitive impairment and Alzheimer’s disease by optical coherence tomography angiography. Acta Ophthalmol 98, e781–e787.

45.

Yoon

, Thompson

, Polascik

, Calixte

, Burke

, Petrella

, Grewal

, Fekrat

(2019) Correlation of OCTA and volumetric MRI in mild cognitive impairment and Alzheimer’s disease. Ophthalmic Surg Lasers Imaging Retina 50, 709–718.

46.

Lahme

, Esser

, Mihailovic

, Schubert

, Lauermann

, Johnen

, Eter

, Duning

, Alnawaiseh

(2018) Evaluation of ocular perfusion in Alzheimer’s disease using optical coherence tomography angiography. J Alzheimers Dis 66, 1745–1752.

47.

Bero

, Yan

, Roh

, Cirrito

, Stewart

, Raichle

, Lee

, Holtzman

(2011) Neuronal activity regulates the regional vulnerability to amyloid-beta deposition. Nat Neurosci 14, 750–756.

48.

Boyle

PA YL

, Nag

, Leurgans

, Wilson

, Bennett

(2015) Cerebral amyloid angiopathy and cognitive outcomes in community-based older persons. Neurology 85, 1930–1936.