Relationship Between Alzheimer’s Disease and Retinal Choroidal Thickness: A Cross-Sectional Study

Abstract

Background:

The choroid is involved directly or indirectly in many pathological conditions such as Alzheimer’s disease (AD), Parkinson’s disease (PD), and multiple sclerosis (MS).

Objective:

The objective of this study was to investigate the association between retinal choroidal properties and the pathology of AD by determining choroidal thickness, hippocampus volume, cognitive functions, and plasma BACE1 activity.

Methods:

In this cross-sectional study, 37 patients with AD and 34 age-matched controls were included. Retinal choroidal thickness was measured via enhanced depth imaging optical coherence tomography. Hippocampal volume was measured via 3.0T MRI. Cognitive functions were evaluated using the Mini-Mental State Examination (MMSE) and Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-Cog). Plasma BACE1 activity was analyzed using a fluorescence substrate-based plasma assay, and regression model were to analyze the data.

Results:

Retinal choroidal thickness was significantly thinner in the AD group than in the control group [(114.81±81.30) μm versus (233.79±38.29) μm, p < 0.05]. Multivariable regression analysis indicated that the ADAS-cog scores (β=–0.772, p = 0.000) and age (β=–0.176, p = 0.015) were independently associated with choroidal thickness. The logistic regression model revealed that the subfoveal choroidal thickness was a significant predictor for AD (OR = 0.984, 95% CI: 0.972–0.997).

Conclusion:

There was a general tendency of choroid thinning as the cognitive function declined. Although choroidal thickness was not a potential indicator for early stage AD, it was valuable in monitoring AD progression.

Keywords

Alzheimer’s disease choroidal thickness cognitive performance hippocampus volume plasma BACE1 activity

INTRODUCTION

Alzheimer’s disease (AD) is one of the most common diseases of the aging population worldwide. According to the World Health Organization, the worldwide prevalence of AD among elderly people aged > 85 years may be as high as 20% –30%. In 2010, there were more than 35 million people with AD or other dementias in the world, and it is estimated that by 2050 the number will reach 115 million [1].

Although AD is thought to be a disease predo-minantly involving changes in cognitive function, patients with AD often have impaired visual function, including color perception, depth perception, contrast sensitivity, and visual field defects [2]. Reports have indicated that retinal-related degenerative diseases such as age-related macular degeneration and glaucoma may be associated with AD, and amyloid dep-osition leading to degeneration of the central nervous system is also considered to cause cellular damage to the optic nerve [3 –5]. It is well known that the retina and the central nervous system are homologous during embryonic development [6]. Similarly, the retinal and cerebral small blood vessels share similar anatomical and physiological properties such as blood–brain and blood–retina barrier functions, as well as nonanastomotic end arterial properties [7, 8]. Moreover, numerous studies suggest that vascular factors play a significant role in the pathogenesis of AD [9]. Thus, the retinal vascular system may act as a window for assessing how vascular pathology contributes to changes in AD.

The choroid, located between the retina and the sclera, is a film in the posterior two-thirds of the mid-ocular membrane and is composed of fibrous tissue, small blood vessels, and capillaries. The principal fu-nction of the choroid is to supply oxygen to the outer retina and vitreous chamber [10]. Choroidal blood flow is directly or indirectly involved in the pathophysiology of a variety of retina-related degenerative diseases. Changes in the choroidal vasculature may lead to significant changes in choroidal thickness. Therefore, analysis of choroidal thickness could have a potential role in the diagnosis, treatment, and prognosis of AD.

Optical coherence tomography (OCT) is a noncon-tact, noninvasive imaging technique that can measure choroidal thickness in vivo. Previous studies using OCT technology to observe structural alterations in the choroid of AD patients revealed a tendency toward choroidal thinning. Gharbiya et al. [11] reported on the correlation between AD and choroidal thickness in 2014: they found that choroid thickness was lower among patients with AD than among control subjects. In 2015, Bayhan et al. [12] published a similar study and concluded that the choroid was thinner in AD patients but was not related to Mini-Mental State Examination (MMSE) scores. In 2016, Bulut et al. [13] used enhanced depth imaging OCT (EDI-OCT) to observe choroidal thickness in 41 patients with AD, 38 patients with mild cognitive impairment (MCI), and 44 healthy controls. The results showed that choroidal thickness was lower in both the AD and MCI groups when compared to healthy controls, and there was a positive correlation between choroid thickness and MMSE scores. More recently, Cunha et al. [14] used EDI-OCT for 50 patients with mild AD (mean age 73.1 years), 152 patients without AD (mean age 71.03 years), and 50 elderly controls without AD (mean age 82.14 years). The authors found that among young patients with AD, the choroid was significantly thinner even when compared with elderly controls. The above studies indicate that the choroid in patients with AD is generally thinner than normal, but the relationship between cognitive impairment and choroidal thickness remains controversial.

Although choroidal thickness assessed by OCT is considered to be potentially useful for early diagnosis or as a follow-up tool [15], the role of the choro-idal vasculature in the pathogenesis of AD is unclear and there is a paucity of research regarding the relationship between changes in choroidal thickness and other diagnostic parameters of AD, such as cerebral structural changes, clinical characteristics, and blood biomarkers. To fill this gap in knowledge, we condu-cted this study to compare choroid thickness between patients with AD and healthy controls. We determ-ined hippocampal volume (HV), cognitive function, plasma volume, and retinal choroid thickness, and then analyzed the relationship between cerebral HV atrophy and retinal choroid thickness among AD patients. We further analyzed the specific correlation between cognitive impairment, plasma BACE1 activity, and retinal choroidal thickness. Finally, mul-tivariable linear regression models and logistic regressions were used to identify the relationship between choroidal thickness and AD.

MATERIALS AND METHODS

Participants

This study was approved by the Ethics Committee of Beijing Geriatric Hospital (approval no. 2017–013). Risks and benefits were discussed with each participant or his/her guardian, and written informed consent was obtained before beginning data collection and examinations. The trial registration number is ChiCTR1800014839.

Participants with AD were recruited from the Memory Clinic and Center for Cognitive Disorders of Beijing Geriatric Hospital in Beijing, China. Age-matched, cognitive-normal participants were volunteers enrolled from the Medical Examination Center of Beijing Geriatric Hospital. For the AD group, the inclusion criteria were: 1) age ≥60 years; 2) met the National Institute on Aging and Alzheimer’s Association criteria for AD in 2011 [16]; 3) Hachinski Ischemia Scale (HIS) score of < 4; and 4) best-cor-rected visual acuity ≥0.6. The exclusion criteria were as follows: 1) those with glaucoma, high myopia (refraction > 6 diopters), uveitis, retinal optic neuro-pathy, macular degeneration, intraocular surgery wi-thin 6 months, post-laser photocoagulation, or other diseases affecting the optic nerve; 2) those with refractive media turbidity, such as cataract nuclear grade IV, affecting OCT imaging and choroidal thickness examination; 3) those with other diseases, such as diabetes, hypertension, or coronary heart disease, that may influence choroidal thickness; 4) those with other types of dementia or other central nervous system degeneration that may cause optic nerve deg-eneration (such as Parkinson’s disease or multiple sclerosis); and 5) long-term smokers. For the control group of age-matched, cognitively normal volunteers, the inclusion criteria were: 1) age ≥60 years; 2) no memory loss symptoms and MMSE score of ≥28 points; and 3) best-corrected visual acuity ≥0.6. The exclusion criteria for the control group were the same as for the AD group.

Sample size

To calculate the sample size [12, 13], we assumed an effect size of 55–60 for the between-group difference in mean choroidal thickness, with standard deviation of 4–4.5, power of 0.90, and a type 1 error of 0.05: the minimum sample size calculated using MedCalc software was 27 in each group. Considering that 10% losses might occur during the study, we decided to include a total of 60 subjects, with 30 patients in the AD group and 30 participants in the control group.

Collection of clinical data

Demographics and clinical information were recorded for all participants, including sex, age, educational level, race, dominant hand, weight, height, and blood pressure. Body mass index (BMI) was defined as body weight (kg)/body height² (m²). Mean arterial pressure (MAP) was defined as (1/3×systolic blood pressure) (2/3×diastolic blood pressure). History of drinking, smoking, and use of psychoactive substance was also recorded. Blood samples were acquired for measurement of plasma BACE1 activity.

OCT image acquisition and feature extraction

All participants underwent a complete ophthalmic examination, including diopter and intraocular pressure (IOP) measurements, slit-lamp and fundus examinations, and choroidal thickness measurements at the Ophthalmic Center of Beijing Geriatric Hospital. Measurements were performed at the same time of day (9:00–11 : 00 a.m.) to avoid diurnal fluctuations, and all participants were banned from smoking, drinking tea or coffee, or engaging in intensive activities for at least 12 h before the assessment. Systolic and diastolic blood pressures were measured before OCT using a sphygmomanometer with participants in the sitting position. All examinations were performed by two experienced ophthalmologists who were blinded to the study design and reviewed the final results to exclude participants with retinopathy or choroidal imaging failure.

Choroidal thickness was measured using a Spect-ralis OCT scanner (Heidelberg Engineering, Heidel-berg, Germany) with EDI mode. The parameters were as follows: maximum A scanning frequency, 40 kHz; scanning depth, 1.9 mm; scanning width, 9.0 mm; transverse optical resolution rate, 14μm; and vertical optical resolution, 7μm. Choroidal thickness was de-fined as the perpendicular distance from the retinal pigment epithelium to the sclera boundary [17]. The choroidal thickness was measured manually beneath the fovea (subfoveal choroidal thickness, SFCT) and at 12 more locations near the fovea: 500μm nasal to the fovea (N₁), 1,500μm nasal to the fovea (N₂), 3,000μm nasal to the fovea (N₃), 500μm temporal to the fovea (T₁), 1,500μm temporal to the fovea (T₂), 3,000μm temporal to the fovea (T₃), 500μm superior direction to the fovea (S₁), 1,500μm superior to the fovea (S₂), 3,000μm superior to the fovea (S₃), 500μm inferior to the fovea (I₁), 1,500μm inferior to the fovea (I₂), and 3,000μm inferior to the fovea (I₃). Three consecutive measurements of choroidal thickness at each location were performed over 3 days at the same time, and the average values were used for analysis. In addition, the above examination was carried out on both eyes for each participant, and one eye with clear choroidal images was selected for final analysis.

Magnetic resonance imaging acquisition and feature extraction

HV was measured using a MAGNETOM Skyra 3.0-T magnetic resonance imaging (MRI) instrument (Siemens, Munich, Germany). First, all participants underwent routine transectional sagittal spin echo T1-weighted imaging (T1WI) and fast spin-echo T2-weighted imaging to exclude other organic brain dis-eases. Next, on the oblique coronal plane for T1WI, the region of interest (ROI) of the hippocampus was delineated using a mouse and the hippocampal area of each layer was obtained. The volume was calculated according to the layer thickness, and the absolute HV was obtained by adding layer-by-layer volumes. The ROI of the hippocampus was defined using the method of Watson and Maller [18, 19]. In addition, to eliminate the influence of differences in cranial cavity size on HV, the absolute HV was standardized using the entire cranial cavity volume according to the Cendes method [20] and adjusted standardized right HV and left HV values were used for final analyses.

Neuropsychological tests

All participants underwent neuropsychological tests, including the MMSE [21], Alzheimer’s Disease Assessment Scale-cognitive subscale (ADAS-cog) [22], and Clinical Dementia Rating (CDR) tests [23]. In addition, the CDR score was used to measure the severity of dementia for the AD group. The tests were completed by two experienced neuropsychiatrists who were blinded to the study design, and the intraclass correlation coefficient (ICC) between the two neuropsychiatrists was 0.91.

Plasma BACE1 assay

Plasma BACE1 activity was measured as previously described [27, 28]. In brief, a synthetic peptide substrate containing a β-cleavage site [MCA-Glu Val Lys Val Asp Ala Glu Phe-(Lys DNP)-OH; GL Biochem, Shanghai, China] was placed in reaction buffer (50 mmol/L acetic acid buffer and 100 mmol/L sodium chloride) at a concentration of 10 mmol/L for the BACE1 activity assay. First, a 10-μL sample of plasma was mixed with 100μL of buffer, resulting in a pH of 4.5, which is the optimal pH for the BACE1 activity assay. Then the fluorescence intensity was measured at 420 nm after excitation at 320 nm using a microplate reader (Synergy H1; Bio Tek, Winooski, VT, USA). Total protein was measured using a bicinchoninic acid kit (Thermo Fisher Scientific, Waltham, MA, USA; #23225), and BACE1 activity was corrected using the plasma total protein content. Finally, BACE1 activity was calculated using V_max and V_mean values and expressed in fluorescence units/time.

Statistical analysis

SPSS statistical software for Windows, version 20 (SPSS, Chicago, IL, USA) was used for statistical analysis. Data are expressed as the mean±standard deviation (SD), or percentage. The normality of the values was analyzed using the Shapiro-Wilk test. The independent t-test was used for comparisons between continuous variables satisfying the normal distribution, and the Mann-Whitney U test was used for comparisons between continuous variables not satisfying a normal distribution. The chi-squared test was used for comparisons between non-continuous variables. Correlations between clinical parameters (HV, cognitive level, and plasma BACE1 activity) and choroidal thickness were determined based on the Pearson’s or Spearman’s correlation coefficient. The Kendall rank correlation coefficient was used to examine correlations between choroid thickness and dementia severity. Multivariable linear regression analysis was used to identify independent risk factors for SFCT, and a logistic regression model was conducted to detect the prediction accuracy of SFCT for AD. p < 0.05 was considered statistically significant.

RESULTS

Demographics, clinical characteristics, and cognitive performance

Participants, including healthy volunteers and pa-tients with AD, were recruited from February 1, 2018 to September 16, 2020 at Beijing Geriatric Hospital in China. Following initial screening, a total of 364 individuals were considered, and a diagnosis of AD was made by two experienced neuropsychiatrists; 252 subjects who did not meet the inclusion criteria were excluded. This left 114 participants who met the inclusion and exclusion criteria for enrollment. Of these, 43 cases were excluded for reasons that included spontaneous withdrawal from the study, suboptimal MRI quality, poor OCT image quality, unqualified blood specimens, or missing data. Therefore, data were collected for 71 participants [37 patients with AD (52.1%) and 34 age-matched controls (47.9%); mean (SD); age: 80.8 (6.63) years; age range: 65–96 years]. The trial flowchart is shown in Fig. 1, and the demographics and clinical characteristics of all participants are shown in Table 1.

Fig. 1

Trial flowchart.

Table 1

Demographics of subjects and clinical characteristics

	AD group	Control group	p
	(n = 37)	(n = 34)
Sex	10/27	16/18	0.080^a
(Male/Female, n)
Age (y)	83.13±5.68	79.73±7.53	0.158^b
Education duration (y)	5.79±4.21	6.85±4.91	0.331^b
Race (n)
Han nationality (%)	34 (97.3%)	33 (97.1%)	1.000^a
Manchu (%)	1 (2.7%)	1 (2.9%)
Dominant hand
Right-hander (%)	35 (94.6%)	32 (96.2%)	0.931^a
Left-hander (%)	2 (5.4%)2	2 (3.8%)
HIS (mean±SD)	0.70±0.74	0.70±0.67	0.707^b
BMI (kg/m²)	21.84±4.11	22.29±4.83	0.959^b
MAP (mmHg)	80.19±9.69	78.15±10.36	0.394^b
Diopter (D)	–1.45±1.99	–0.22±5.98	0.287^b
IOP (mmHg)	18.15±1.72	17.74±3.09	0.112^b
Axial length (mm)	23.39±0.59	23.43±1.51	0.258^b
Laterality
Left eye (n)	17 (45.9%)	15 (44.1%)	0.877^a
Right eye (n)	20 (54.1%)	19 (55.9%)
MMSE mean	7.86±6.86	29.41±1.01	0.000^b
ADAS-Cog mean	48.40±17.49	3.00±1.80	0.000^b
CDR mean	2.54±0.60	0	0.000^b
HV_left (mm³)	2767.35±359.76	3480.20±349.49	0.000^b
HV_right (mm³)	2901.72±357.70	3542.79±281.94	0.000^b
HV_total (mm³)	5650.16±732.55	7023.00±583.67	0.000^b
TICV (mm³)	337108.83±34533.13	334265.20±36362.57	0.420^c
Adjusted HV_left (mm³)	2744.64±446.31	3462.63±434.76	0.000^c
Adjusted HV_right (mm³)	2866.72±440.67	3528.76±423.79	0.000^c
Adjusted HV_total (mm³)	5611.36±881.60	6991.39±821.11	0.000^c
Plasma BACE1 activity	757.93±388.21	556.35±142.13	0.005^b
(ng/mL)

^aχ² test for categorical variables. ^bMann-Whitney U test. ^cUnpaired t test with Levene’s test for equality of variances. AD, Alzheimer’s disease; HIS, Hachinski Ischemic Scale; IOP, intraocular pressure; BMI, body mass index; MAP, mean arterial pressure (1/3×systolic blood pressure) (2/3×diastolic blood pressure); MMSE, Mini-Mental State Examination; ADAS-Cog, The Alzheimer’s Disease Assessment Scale-Cognitive section; CDR, Clinical Dementia Rating; HV, hippocampal volume; TICV, total intracranial volume; SD, standard deviation.

Choroidal thickness parameters

EDI-OCT was conducted to assess choroidal thickness in the study participants. Typical images are sho-wn in Fig. 2. The results show that in the control group, the thickest choroidal area was in the subf-ovea region (subfovea choroidal thickness, SFCT), followed by the temporal side, the superior side, and the inferior side; the nasal side had the thinnest choroidal area. Specifically, at regions 500μm from the fovea, the order for choroidal thickness was T₁ >S₁ > N₁ > I₁; at regions 1,500μm from the fovea, the order of was T₂ > I₂ > S₂ > N₂; and at regions 3,000μm from the fovea, the order was S₃ > I₃ > T₃ > N₃. However, for the AD group, SFCT was thinner than the choroidal thickness at T₁, S₁, N₁, and I₁. In areas surrounding the fovea, the order for choroidal thickness was I₁ > S₁ > T₁ > N₁; S₂ > T₂ > I₂ > N₂; and S₃ > T₃ > I₃ > N₃.

Fig. 2

Choroidal thickness measurements in both healthy individuals (A) and patients with AD (B).

Choroidal thickness measured at the 13 locations were compared between the AD and control groups. The results show that the choroidal thickness at the subfovea and at T₁, T₂, N₁, N_2, S₁, I₁, I₂, and I₃ were significantly thinner in the AD group than in the control group (Table 2, Fig. 3). Furthermore, the AD participants were divided by disease severity into subgroups with mild to moderate AD (CDR = 1 or 2) and severe AD (CDR = 3). Mann-Whitney U tests were used to compare the choroidal thickness differences in 13 locations. In severe AD, the choroidal thickness measurements at SFCT, T₁, T₂, N₁, and I₃ were significantly thinner in mild to moderate AD (Table 3).

Table 2

Comparison of choroidal thickness between control and AD groups at different measurement locations

OCT parameter	AD group	Control group	p
	(n = 37)	(n = 34)
SFCT (μm)	114.81±81.30	233.79±38.29	0.000^a
T₁ (μm)	121.24±72.70	234.14±3.17	0.000^a
N₁ (μm)	124.29±87.19	214.11±41.88	0.000^b
S₁ (μm)	155.56±94.34	216.67±49.11	0.008^a
I₁ (μm)	161.72±97.68	211.64±46.35	0.016^b
T₂ (μm)	152.91±74.84	216.67±49.11	0.000^b
N₂ (μm)	111.10±58.35	156.55±51.86	0.000^b
S₂ (μm)	176.08±108.13	207.91±60.46	0.108^b
I₂ (μm)	150.54±85.85	211.64±46.36	0.006^b
T₃ (μm)	161.78±80.90	194.82±84.08	0.140^b
N₃ (μm)	105.05±71.10	97.97±29.50	0.530^b
S₃ (μm)	193.35±124.04	204.94±69.98	0.289^b
I₃ (μm)	144.18±83.52	199.32±59.29	0.002^b

^at test. ^bMann-Whitney U test. AD, Alzheimer’s disease; SFCT, subfoveal choroidal thickness; T, N, S, I are the choroidal thicknesses at different sectorial points (temporal, nasal, superior, inferior) to the fovea.

Fig. 3

Choroidal thickness differences between the groups.

Table 3

Comparison of choroidal thickness between mild to moderate AD group and severe AD group at different measurement locations

OCT parameter	Mild to moderate AD	Severe AD	p
	(n = 15)	(n = 22)
SFCT (μm)	165.73±69.60	80.09±70.62	0.000
T₁ (μm)	150.86±72.90	101.04±66.80	0.020
N₁ (μm)	149.40±66.66	107.18±95.51	0.007
S₁ (μm)	173.00±72.11	143.68±106.88	0.178
I₁ (μm)	180.53±71.54	148.90±111.89	0.101
T₂ (μm)	184.86±67.63	131.13±72.99	0.048
N₂ (μm)	128.06±72.01	99.54±45.11	0.199
S₂ (μm)	182.33±71.59	171.81±128.76	0.353
I₂ (μm)	165.73±68.80	140.18±95.92	0.159
T₃ (μm)	174.66±67.83	153.00±89.19	0.654
N₃ (μm)	106.66±75.67	103.95±69.60	0.957
S₃ (μm)	202.73±76.43	186.95±149.59	0.353
I₃ (μm)	176.46±76.83	122.18±82.28	0.048

Mann-Whitney U test. AD, Alzheimer’s disease; SFCT, subfoveal choroidal thickness; T, N, S, I are the choroidal thicknesses at different sectorial points (temporal, nasal, superior, inferior) to the fovea.

Correlation between hippocampal volume and SFCT in patients with AD

Pearson’s correlation analyses for the AD group revealed a significant positive correlation between total adjusted hippocampus volume and SFCT (r = 0.441, p = 0.006). Adjusted right and left hippocampal volumes were also analyzed; SFCT was directly correlated with the adjusted left (r = 0.455, p = 0.005) and adjusted right (r = 0.437, p = 0.007) hippocampal volume (Table 4).

Table 4

Correlations between clinical parameters (cognitive function, hippocampal volume, and plasma BACE1 activity) and subfoveal choroidal thickness in patients with AD

Parameter	SFCT
	r	p
MMSE	0.772	0.000^a
ADAS-cog	–0.695	0.000^a
Adjusted HV_left (mm³)	0.455	0.005^b
Adjusted HV_right (mm³)	0.437	0.007^b
Adjusted HV_total (mm³)	0.441	0.006^b
Plasma BACE1 activity	–0.156	0.328^a
(ng/mL)

^aSpearman’s correlation test. ^bPearson’s correlation test. SFCT, subfoveal choroidal thickness; MMSE, Mini-Mental State Examination; ADAS-Cog, The Alzheimer’s Disease Assessment Scale-Cognitive section.

Correlation between cognitive function (MMSE and ADAS-Cog scores) and SFCT in patients with AD

According to Spearman’s correlation analyses, SFCT was directly correlated with MMSE scores (r = 0.772, p = 0.000) and inversely correlated with ADAS-Cog scores (r= –0.695, p = 0.000) (Table 4). Kendall rank correlation analysis was used to assess differences between the AD subgroups, and the results showed that for the mild to moderate AD, the correlation between SFCT and MMSE and ADAS-Cog was not statistically significant. By contrast, SFCT was significantly positively correlated with MMSE (r = 0.618, p = 0.000) and negatively correlated with ADAS-Cog (r=–0.626, p = 0.000) for the severe AD.

Correlation between plasma BACE1 activity and SFCT in patients with AD

For plasma BACE1 activity, there was no significant correlation with SFCT (r = –0.156, p = 0.328) among patients with AD (Table 4).

Multivariable linear regression analysis

A multivariable linear regression model was used to determine the relationship between choroidal thi-ckness and groups, age, sex, years of education, BMI, MAP, diopters, IOP, axial length, laterality, cognitive function, adjusted HV_total, and plasma BACE1 activity. We used SFCT as the dependent variable and the above-mentioned factors as the independent variables. The multicollinearity of the independent variables was assessed by calculating the variance inflation factor (VIF) using linear regression. The confounders were excluded if they were statistically associated with another confounder in the model. There was an obvious collinearity between ADAS-cog and MMSE (VIF = 26.245 and 26.081), and ADAS-cog reflected cognitive function more preci-sely, thus MMSE was excluded. After the initial screening of independent variables, the groups, age, sex, years of education, BMI, MAP, diopters, IOP, axial length, laterality, ADAS-cog, adjusted HV_total, and plasma BACE1 activity were added to multiv-ariable linear regression models as independent variables, and backward removal of variables was used to exclude those that did not explain a significant pro-portion of variance. The regression equation was Y (SFCT) = 428.708–2.593×ADAS-Cog–2.324× age. The results show that the SFCT was significantly negatively correlated with ADAS-Cog (β=–0.772, p = 0.000) and age (β=–0.176, p = 0.015) (Table 5).

Table 5

Multiple linear regression analysis of choroidal thickness with other related factors^a

Related factors	B	β	SE	t	p
Constant	428.708		74.891	5.724	0.000
Age	–2.324	–0.176	0.932	–2.493	0.015
ADAS-Cog	–2.593	–0.772	0.237	–10.933	0.000

B, regression coefficient; β, Standardized Coefficients Beta; SE, standard error; ADAS-Cog, The Alzheimer’s Disease Assessment Scale-Cognitive section. ^aThe model was adjusted by group, age, sex, education duration, BMI, MAP, diopter, IOP, axial length, laterality, ADAS-Cog, adjusted HV_total, and plasma BACE1 activity. Not significant (p > 0.05) were not reported.

Logistic regression analyses

Logistic regression was used to detect the prediction accuracy of SFCT for AD, groups (whether or not it was AD) as the dependent variable. In the univariate model, SFCT was a significant predictor for AD, with an odds ratio (OR) of 0.975 [95% confidence interval (CI): 0.964–0.986]. The area under the ROC curve (AUC) of SFCT for AD was 0.865 (95% CI: 0.761–0.968). Furthermore, multivariable logistic regression was performed by adjusting sex, education duration, adjusted HVtotal, and plasma BACE1 activity. After controlling for all the other variables, we found an association between AD and SFCT in the model; the logistic regression analyses showed that SFCT was a significant predictor for AD, with the OR of 0.984 [95% CI: 0.972–0.997], and the AUC was 0.937 (95% CI: 0.761–0.968) (Table 6).

Table 6

Multivariable logistic regression analysis to detect the association between SFCT and AD^a

Related factors	B	SE	Wald	p	OR	95% CI
SFCT	–0.016	0.006	5.877	0.015	0.984	0.972–0.997
Adjusted HV_total	–0.002	0.001	7.213	0.007	0.998	0.997–1.000
Constant	10.123	3.870	6.844	0.009	24901.881

B, regression coefficient; SE, standard error; OR, odds ratio; CI, confidence interval; SFCT, subfoveal choroidal thickness; AD, Alzheimer’s disease ^aThe model was adjusted by sex, education duration, SFCT, adjusted HV_total, and plasma BACE1 activity. Not significant (p > 0.05) were not reported.

DISCUSSION

The choroid is a highly vascularized tissue that provides metabolic support for the retina, and changes in the choroidal vasculature may lead to significant changes in choroidal thickness. In our study, thickness measurements in control participants showed that the thickest choroidal areas were in the subfovea, followed by the temporal side, superior side, and inf-erior side, and the nasal side had the thinnest chor-oidal area, consistent with previous reports [29, 30]. However, for AD patients compared to the control group, we found that the choroidal thickness was significantly lower in the subfovea and temporal and superior parafoveal areas; moreover, the SFCT was thinner than in other parafoveal regions. The results indicate that the normal distribution of choroidal thickness had changed in patients with AD. These results indicating choroidal thinning in AD patients compared with control subjects are in accordance with results from previous studies [7 , 11–13]. Nevertheless, we found that the choroidal thickness values observed in different studies were slightly different. It is generally accepted that choroidal thickness is a highly variable parameter that is affected by variables such as age, axial length, diopter, and diurnal variations [31, 32]. Differences in results by study are therefore considered to be related to different control of these variables. Many previous studies of patients with AD have described optic nerve degeneration, including thinning of the retinal nerve fiber layer and loss of retinal ganglion cells [33]. Degeneration of ganglion cells might be accompanied by pathological changes in the vascular system, manifest as a decrease in choroidal tissue structure per unit area, changes in the choroidal microcirculation system, and finally morphological alterations involving thinning of the choroid, reflecting the important role of vascular factors in the pathogenesis of AD.

Many studies have reported that the cerebral hippocampus is closely related to cognitive processing, memory, and other cognitive functions. In the present study, we calculated right HV, left HV, and total HV from structural MRI scans for patients with AD and control participants. The results show that the left HV, right HV, and total HV were significantly lower in patients with AD than in control participants, which is consistent with previous results [34 –36]. To further understand the role of choroidal thickness in the pathogenesis of AD, we investigated the relationship between SFCT and HV in the AD group. The results revealed a significant positive correlation between HV and choroidal thickness among patients with AD. This indicates that as HV decreased, the choroid gra-dually became thinner. Notably, after adjusting for age, sex, MAP, and other confounding factors, the association was not significant on multivariable linear regression. A possible explanation might be the small sample size. The lack of statistical significance also suggests that the relationship between HV and choroidal thickness is complex, so the role of the choroidal vasculature in the pathogenesis of AD remains to be further characterized. However, this is the first report on a correlation analysis of choroidal thickness and HV among patients with AD. In the future, a large-scale longitudinal comparative study would be helpful to assess the relationship between choroidal thickness and cerebral structural changes among patients with AD, and to further clarify the underlying pathological mechanisms.

One purpose of this study was to determine the relationship between choroidal thickness and cognitive impairment among AD patients. MMSE is the assessment tool most commonly used to detect dementia, but it has some disadvantages: the items are not comprehensive enough, some items are too simple, and it is susceptible to false-positive or false-negative results due to the educational level of the subjects being tested. Compared with the MMSE, the ADAS-Cog gives more accurate reflection of a patient’s cognitive deficit; its reliability and efficiency have been confirmed as acceptable in many clinical studies [24 –26], and some items and subscales of the ADAS-Cog may be better measures of cognitive functioning for different stages of AD [24 –26]. Therefore, we used MMSE and ADAS-Cog together to evaluate cognitive function. The re-sults show that SFCT was positively correlated with the MMSE score and negatively correlated with the ADAS-Cog score. Our findings are in agreement with Bulut et al. [13], who observed a positive correlation between choroidal thickness and MMSE scores among patients with AD. To the best of our knowledge, ours is the first study to analyze ADAS-Cog scores in relation to choroidal thickness. Compared with MMSE, ADAS-Cog more accurately reflects cognitive function in AD, and its reliability and eff-ectiveness have been confirmed by many clinical studies [24–26 , 37]. Multivariable correlation analyses after adjustment for the above factors revealed that the correlation of SFCT with ADAS-Cog was still significant. Both of the scales together demonstrated that SFCT thinning in AD patients was related to a decline in cognitive function. Subgroup analysis results showed that for the group with mild to moderate AD, there was no correlation between SFCT and cognitive function. However, for the severe AD group, SFCT was significantly positively correlated with the MMSE score and negatively correlated with the ADAS-Cog score, suggesting that choroidal thinning as a morphological change in the vascular sys-tem may occur during the late rather than the early stages of AD.

BACE1 is considered a candidate biomarker for ea-rly diagnosis of AD [38, 39]. We therefore analyzed the relationship between plasma BACE1 and cho-roidal thickness but found no significant correlation. BACE1 is a major factor in the production of amyloid from amyloid precursor protein and may be a factor in the relationship between choroidal thickness and cognition [40 –43]. Previous studies reported that BACE1 knockout mice have lower production of Aβ and that cognitive decline is significantly reduced in the absence of BACE1 [44]. In addition, many studies have found that BACE1 is a stress response protease, so AD-related oxidative stress, inflammation, and hypoxia can increase the level of BACE1 expression in both cerebrospinal fluid and plasma [45, 46]. Therefore, BACE1 activity is considered an important marker in the early diagnosis, prevention, and treatment of AD [38, 39]. To the best of our knowledge, this is the first study to compare the relationship between choroidal thickness and plasma BACE1 activity, but we found no significant correlation between them. A possible explanation might be that changes in choroidal thickness do not occur in parallel with changes in BACE1 activity at the same stage. Recent data suggest that a change in BACE1 activity is an early event in the AD pathological process. Our logistic regression model revealed that the SFCT was a significant predictor for AD, and the SFCT in the severe AD was thinner than those of mild to moderate AD. In addition, the subgroup correlations analysis revealed the significant correlation between choroidal thickness and cognitive function only in the severe AD, but not in the mild to moderate AD. These findings suggest that a significant decrease in choroidal thickness is a morphological change that is most likely to occur during later stages of AD pathophysiology. Therefore, we concluded that choroidal thickness is not a potential marker for early AD diagnosis, but might be valuable in monitoring the pathological progression of AD.

Strengths and limitations

Previous studies relied on use of the MMSE as a neurocognitive testing scale. The MMSE is widely used, but it lacks sufficient sensitivity, suffers from a ceiling effect, and has a relatively high false-positive rate in detecting early dementia in AD. A strength of our study is the use of a standardized battery of cognitive tests, including MMSE and ADAS-Cog, for more precise evaluation of cognitive function and more accurate assessment of the relationship between cognitive function and choroidal thickness.

Cerebral hippocampal volume atrophy has been significantly associated with disease progression in AD and has been validated in longitudinal studies as a characteristic of AD. In addition, BACE1 plays a crucial role in AD development or progression, and BACE1 activity is a candidate biomarker of AD. Thus, we analyzed the relationship between hippocampal volume, plasma BACE1 activity, and choroidal thickness to gain a better understanding of the relationship between changes in choroidal thickness and AD. To the best of our knowledge, this is the first study to compare hippocampal volume and plasma BACE1 activity with choroidal thickness.

Our study has some limitations. First, this was a cross-sectional study and there was no longitudinal follow-up to assess dynamic changes in choroidal thi-ckness with disease progression or regression. Second, the study was observational and was limited to measuring morphological choroidal changes, and not to studying changes in choroidal microstructure, such as changes in the structure of choroidal blood vessels or the number of cells in choroidal tissues, or the relationship between choroidal microstructural changes, so AD progression and different stages need further study. In addition, many small vessel vascular changes can be found in AD, but we only choose the hippocampal volume as the main observation index in MRI, which is a deficiency of this study, so the white matter hyperintensity should be observed in future studies. Finally, another major limitation of our study is the sample size, so changes in choroidal thickness as a predictor of later stages of AD should be further explored using a larger sample size.

CONCLUSIONS

We found that thinner choroid was correlated with cognitive function declines among patients with AD; however, the correlation was only apparent in cases of severe AD, and not in cases of mild to moderate AD. Besides aging, multivariable linear regression models with choroidal thickness as the dependent variable supported the importance of AD as a risk factor for choroidal atrophy. Studies with larger size samples are required to confirm our findings, but the possible association between choroid thinning and AD suggests that noninvasive techniques could potentially be used to follow the disease course and to monitor the efficacy of treatments.

DATA ACCESSIBITY

Information about the experimental methods and data described in this article is available to scientific and medical communities for review, verification, and research studies. For data inquiry, please contact Dr. Xin Ma at maxinanding@ccmu.edu.cn or Mo Li at limo333@163.com

Footnotes

ACKNOWLEDGMENTS

We thank our colleagues in the Beijing Geriat-ric Hospital for their efforts in this study. In addi-tion, we would like to thank all the participants for their involvement and cooperation. This study was funded by Beijing Municipal Administration of Hospitals Clinical Medicine Development of Special Funding Support (No. ZYXL201834). We thank International Science Editing () for editing this manuscript.

Authors’ disclosures available online ().

References

Lim

, Li

, He

, Vingrys

, Wong

, Currier

, Mullen

, Bui

, Nguyen

(2016) The eye as a biomarker for Alzheimer’s disease. Front Neurosci 10, 536.

Tzekov

, Mullan

(2014) Vision function abnormalities in Alzheimer disease. Surv Ophthalmol 59, 414–433.

Sivak

(2013) The aging eye: Common degenerative mechanisms between the Alzheimer’s brain and retinal disease. Invest Ophthalmol Vis Sci 54, 871–880.

Ghiso

, Doudevski

, Ritch

, Rostagno

(2013) Alzheimer’s disease and glaucoma: Mechanistic similarities and differences. J Glaucoma 22(Suppl 5), S36–S38.

Kaarniranta

, Salminen

, Haapasalo

, Soininen

, Hiltunen

(2011) Age-related macular degeneration (AMD): Alzheimer’s disease in the eye? J Alzheimers Dis 24, 615–631.

Wang

, Zhu

, Shi

, Li

, Shen

(2015) Meta-analysis of the relationship of peripheral retinal nerve fiber layer thickness to Alzheimer’s disease and mild cognitive impairment. Shanghai Arch Psychiatry 27, 263–279.

Lee

, Abraham

, Swenor

, Sharrett

, Ramulu

(2018) Application of optical coherence tomography in the detection and classification of cognitive decline. J Curr Glaucoma Pract 12, 10–18.

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

, Akinyemi

, Ihara

(2012) Does vascular pathology contribute to Alzheimer changes? J Neurol Sci 322, 141–147.

10.

Laviers

, Zambarakji

(2014) Enhanced depth imaging-OCT of the choroid: A review of the current literature. Graefes Arch Clin Exp Ophthalmol 252, 1871–1883.

11.

Gharbiya

, Trebbastoni

, Parisi

, Manganiello

, Cruciani

, D’Antonio

, De Vico

, Imbriano

, Campanelli

, De Lena

(2014) Choroidal thinning as a new finding in Alzheimer’s disease: Evidence from enhanced depth imaging spectral domain optical coherence tomography. J Alzheimers Dis 40, 907–917.

12.

Bayhan

, Aslan

, Celikbilek

, Tanik

, Gurdal

(2015) Evaluation of the chorioretinal thickness changes in Alzheimer’s disease using spectral-domain optical coherence tomography. Clin Exp Ophthalmol 43, 145–151.

13.

Bulut

, Yaman

, Erol

, Kurtulus

, Toslak

, Dogan

, Turgut

, Kaya

(2016) Choroidal thickness in patients with mild cognitive impairment and Alzheimer’s type dementia. J Ophthalmol 2016, 7291257.

14.

Cunha

, Proenca

, Dias-Santos

, Melancia

, Almeida

, Aguas

, Santos

, Alves

, Ferreira

, Papoila

, Louro

, Castanheira-Dinis

(2017) Choroidal thinning: Alzheimer’s disease and aging. Alzheimers Dement (Amst) 8, 11–17.

15.

Di Staso

, Ciancaglini

, Abdolrahimzadeh

, D’Apolito

, Scuderi

(2019) Optical coherence tomography of choroid in common neurological diseases. In Vivo 33, 1403–1409.

16.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

, 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.

17.

Wallman

, Nickla

(2010) The multifunctional choroid. Prog Retin Eye Res 29, 144–168.

18.

Réglade-Meslin

, Anstey

, Maller

(2006) Sex and symmetry differences in hippocampal volumetrics: Before and beyond the opening of the crus of the fornix. Hippocampus 16, 80–90.

19.

Xiong

, Lancaster

, Szabo

(2001) Amygdalar and hippocampal volumetry in control participants: Differences regarding handedness. Am J Neuroradiol 22, 1342–1345.

20.

Leproux

, Cendes

, Melanson

, Ethier

, Evans

, Peters

, Andermann

(1993) MRI of amygdala and hippocampus in temporal lobe epilepsy. J Comput Assist Tomogr 17, 206–210.

21.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

22.

Rosen

, Mohs

, Davis

(1984) A new rating scale for Alzheimer’s disease. Am J Psychiatry 141, 1356–1364.

23.

Morris

(1993) The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 43, 2412–2414.

24.

Loveman

, Green

, Kirby

, Takeda

, Picot

, Payne

, Clegg

(2006) The clinical and cost-effectiveness of donepezil, rivastigmine, galantamine and memantine for Alzheimer’s disease. Health Technol Assess 10, 1–160.

25.

, Wu

, Zhou

, Liu

, Dong

(2008) Huperzine A for Alzheimer’s disease. Cochrane Database Syst Rev 2, D5592.

26.

Birks

, Grimley

(2015) Rivastigmine for Alzheimer’s disease. Cochrane Database Syst Rev 4, D1191.

27.

Zhang

, Cui

, Gao

, Li

, Zhang

, Liu

, Yan

, Shen

, Li

(2019) Elevated cleavage of neuregulin-1 by beta-secretase 1 in plasma of schizophrenia patients. Prog Neuropsychopharmacol Biol Psychiatry 90, 161–168.

28.

Shen

, Wang

, Sun

, Yao

, Keegan

, Mullan

, Wilson

, Lista

, Leyhe

, Laske

, Rujescu

, Levey

, Wallin

, Blennow

, Li

, Hampel

(2018) Increased plasma beta-secretase 1 may predict conversion to Alzheimer’s disease dementia in individuals with mild cognitive impairment. Biol Psychiatry 83, 447–455.

29.

Margolis

, Spaide

(2009) A pilot study of enhanced depth imaging optical coherence tomography of the choroid in normal eyes. Am J Ophthalmol 147, 811–815.

30.

Ikuno

, Kawaguchi

, Nouchi

, Yasuno

(2010) Choroidal thickness in healthy Japanese subjects. Invest Ophthalmol Vis Sci 51, 2173–2176.

31.

Barteselli

, Chhablani

, El-Emam

, Wang

, Chuang

, Kozak

, Cheng

, Bartsch

, Freeman

(2012) Choroidal volume variations with age, axial length, and sex in healthy subjects: A three-dimensional analysis. Ophthalmology 119, 2572–2578.

32.

Usui

, Ikuno

, Akiba

, Maruko

, Sekiryu

, Nishida

, Iida

(2012) Circadian changes in subfoveal choroidal thickness and the relationship with circulatory factors in healthy subjects. Invest Ophthalmol Vis Sci 53, 2300–2307.

33.

Blanks

, Torigoe

, Hinton

, Blanks

(1996) Retinal pathology in Alzheimer’s disease. I. Ganglion cell loss in foveal/parafoveal retina. Neurobiol Aging 17, 377–384.

34.

Davatzikos

, Resnick

, Wu

, Parmpi

, Clark

(2008) Individual patient diagnosis of AD and FTD via high-dimensional pattern classification of MRI. Neuroimage 41, 1220–1227.

35.

Gerischer

, Fehlner

, Kobe

, Prehn

, Antonenko

, Grittner

, Braun

, Sack

, Floel

(2018) Combining viscoelasticity, diffusivity and volume of the hippocampus for the diagnosis of Alzheimer’s disease based on magnetic resonance imaging. Neuroimage Clin 18, 485–493.

36.

Dawe

, Yu

, Arfanakis

, Schneider

, Bennett

, Boyle

(2020) Late-life cognitive decline is associated with hippocampal volume, above and beyond its associations with traditional neuropathologic indices. Alzheimers Dement 16, 209–218.

37.

Andrieu

, Sampaio

, Vellas

, Gordon

, European Task Force Group (2007) Disease-modifying trials in Alzheimer’s disease: A European task force consensus. Lancet Neurol 6, 56–62.

38.

Kandalepas

, Vassar

(2011) The β-secretase enzyme BACE1 as a therapeutic target for Alzheimer’s disease. Alzheimers Res Ther 3, 20.

39.

Moechars

, Lichtenthaler

, Barão

(2016) BACE1 Physiological functions may limit its use as therapeutic target for Alzheimer’s disease. Trends Neurosci 39, 158–169.

40.

Cheng

, He

, Lee

, Yao

, Li

, Shen

(2014) High activities of BACE1 in brains with mild cognitive impairment. Am J Pathol 184, 141–147.

41.

Koelsch

(2017) BACE1 Function and inhibition: Implications of Intervention in the amyloid pathway of Alzheimer’s disease pathology. Molecules 22, 1723.

42.

Yan

, Vassar

(2014) Targeting the beta secretase BACE1 for Alzheimer’s disease therapy. Lancet Neurol 13, 319–329.

43.

Yan

, Ma

, Gai

, Cai

, Luo

(2014) Can BACE1 inhibition mitigate early axonal pathology in neurological diseases? J Alzheimers Dis 38, 705–718.

44.

Kandalepas

, Vassar

(2014) The normal and pathologic roles of the Alzheimer’s beta-secretase, BACE1. Curr Alzheimer Res 11, 441–449.

45.

Bai

, Mao

, Tian

, Yu

, Zeng

, Ouyang

, Yu

, Li

, Wang

, Deng

, Wei

, Luo

(2015) Calcium sensing receptor mediated the excessive generation of beta-amyloid peptide induced by hypoxia in vivo and in vitro. Biochem Biophys Res Commun 459, 568–573.

46.

Yang

, Mahmood

, Kim

, Choi

, Lee

, Cho

, Hyun

, Kim

, Chang

, Kim

(2018) Phloroglucinol ameliorates cognitive impairments by reducing the amyloid beta peptide burden and pro-inflammatory cytokines in the hippocampus of 5XFAD mice. Free Radic Biol Med 126, 221–234.