Distribution of Choroidal Thickness and Choroidal Vessel Dilation in Healthy Japanese Individuals: The Nagahama Study.

Purpose: To report fundamental epidemiologic data for choroidal parameters such as choroidal thickness and index of choroidal vascularity in Japanese individuals and to evaluate their correlations with age, sex, systemic parameters, and other ocular parameters. Design: Population-based cohort study. Participants: A total of 9850 individuals participated in the first follow-up of the Nagahama Prospective Cohort for Comprehensive Human Bioscience (the Nagahama Study) conducted between 2013 and 2016.
Methods: All participants underwent standardized ophthalmic examinations, including OCT with enhanced depth imaging (EDI; RS-3000 Advance; Nidek). We manually segmented the choroidoscleral interface to measure subfoveal choroidal thickness (SFCT) and calculated the normalized choroidal intensity obtained with EDI (NCIEDI) and choroidal vascularity index (CVI). These are indices of choroidal brightness in OCT and reportedly represent the dilation of choroidal vessels. After summarizing the age-sex stratified distributions of SFCT, NCIEDI, and CVI, their associations with age, sex, axial length (AL), and spherical equivalent (SE) were evaluated using linear regression analysis with adjustments for possible confounders. Main Outcome Measures: Distribution of SFCT, NCIEDI, and CVI in the healthy Japanese population and their characteristics.
Results: Age-sex standardized SFCT, NCIEDI, and CVI were 291.2 μm, 0.653, and 66.88%, respectively. In both men and women, SFCT was associated negatively with age (P < 0.001) and NCIEDI was associated positively with age (P < 0.001). Although both SFCT and NCIEDI did not differ significantly between men and women overall (P = 0.87 and P = 0.21, respectively), among younger participants (35-50 years of age), men showed significantly greater SFCT than women (P < 0.001). Only in men was CVI associated positively with age (P < 0.001). In the multivariable analysis, SFCT was associated significantly with age, sex, AL, SE, and the interaction term of age and sex (P < 0.001). Independent of SFCT, NCIEDI and CVI were associated significantly with age (P < 0.001). Conclusions: We report normative Japanese SFCT, NCIEDI, and CVI data using a large general Japanese cohort. The association analysis of SFCT with NCIEDI and CVI suggested that younger individuals have a more lumen-rich choroid for their choroidal thickness than older individuals.

The choroid, a posterior part of the uvea, is a membranous tissue lying between the retina and sclera. It comprises abundant vessels and stromal tissue and is rich in melanocytes. Because of these anatomic and histologic characteristics, the choroid is multifunctional and involved in the nourishment of the outer retina, regulation of the temperature of the retina, regulation of intraocular pressure (IOP), and absorption of excess light. Thus, impairment of the choroid can cause numerous ocular diseases, including age-related macular degeneration (AMD), central serous chorioretinopathy (CSC), and high-myopia–related chorioretinal atrophy.2, 3, 4 Additionally, recent advancements in OCT technology have facilitated detailed noninvasive investigations of the choroidal structure, leading to the definition of the new disease concept, pachychoroid., The term pachychoroid indicates a thick choroid (“pachy-“ [prefix]: thick).

In this context, the clinical significance of parameters for the choroid, particularly choroidal thickness, is increasing. Nevertheless, despite the importance of the criterion for the normal thickness of the choroid, a uniform cutoff thickness value may be difficult to identify because the thickness can be influenced by a variety of factors, including age, axial length (AL), refractive error, and blood pressure.7, 8, 9, 10, 11, 12, 13 Thus, as we discussed in a previous report, these parameters should be considered together to understand the relationship between choroidal thickness and its pathogenesis. Furthermore, several methods have been proposed to quantify the variation of histologic choroidal structure, such as choroidal vascularity index (CVI) and choroidal intensity, which allow for understanding the choroid in more detail. To apply our wider understanding of the choroid to clinical practice, it is essential to accumulate epidemiologic data for the choroidal condition in healthy populations; however, the epidemiologic features of the choroidal condition have been investigated in only a few cohorts.7, 8, 9

In this article, we report fundamental epidemiologic data for the choroidal condition using a large general Japanese cohort, the Nagahama Prospective Cohort for Comprehensive Human Bioscience (the Nagahama Study; n = 9850). Although the choroidal thickness of normal adult individuals has been reported in multiple studies,7, 8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25 all these studies included a small number of individuals (n < 1000), except for the Beijing Eye Study (n = 3233) and Singapore Epidemiology of Eye Diseases Study (SEEDS; n = 1619). To the best of our knowledge, the current study provides the largest normative data of choroidal thickness to date. In addition, we also present the distribution of choroidal intensity and CVI, which are recently introduced indicators of the choroidal brightness in OCT and are thought to correlate negatively and positively, respectively, with the dilation of choroidal vessels.24, 25, 26, 27

Methods

This retrospective cross-sectional study followed the tenets of the Declaration of Helsinki. This study was also approved by the ethics committee of the Kyoto University Graduate School of Medicine and the Nagahama Municipal Review Board. Written informed consent was obtained from all participants.

Study Participants and Data Collection

We analyzed data from the second visit of the Nagahama Study obtained between 2013 and 2016. Although a detailed description of the Nagahama Study is provided elsewhere,, we summarized it briefly in the Supplemental Note (available https://www.ophthalmologyscience.org/). The present study analyzed a total of 9850 individuals 35 to 80 years of age.

All participants underwent a standardized ophthalmic examination, physical examinations, and blood tests. The standardized ophthalmic examination included measurement of the spherical equivalent (SE; ARK-530A; Nidek), IOP measurement by noncontact tonometry (full autotonometer TX-20P; Canon), central corneal thickness measurement (full autotonometer TX-20P), AL measurements by partial coherence interferometry (IOL Master; Carl Zeiss Meditec, Inc), color fundus photography (CR- DG10; Canon), and spectral-domain OCT (RS-3000 Advance; Nidek). Intraocular pressure and central corneal thickness were measured only from July 2013 through February 2016. Using OCT, a macular cross-scan of the right eye was performed in both normal and enhanced-depth modes. The scan width was 6 mm before February 2015 and 9 mm thereafter.

The following participants were excluded from this analysis: those without data for the right eye (n = 38), those for whom enhanced depth imaging (EDI) OCT images could not be obtained or the image quality was too poor to detect visually the entire layer of the choroid just under the fovea (n = 372), and those who had macular-involving disease (n = 115).

A questionnaire was administered to ascertain the history of cataract surgery, other ocular surgeries, and ocular laser treatment (including photocoagulation). The presence or absence of contact lens wear on the inspection day was recorded. Individuals who had undergone ocular surgery (excluding cataract surgery), ocular laser treatment, or both were excluded from all analyses. Additionally, individuals who had undergone cataract surgery were excluded from analyses of SE. Individuals who were unable to remove their contact lenses for special reasons also were excluded from analyses of SE, IOP, and central corneal thickness.

Measurement and Calculation of Choroidal Parameters

The inner limiting membrane, outer border of the nerve fiber layer (NFL), and Bruch’s membrane in each EDI OCT image were detected automatically by the built-in software. Using the same software, the segmentation of the Bruch’s membrane was corrected manually, and the outer border of the choroid was segmented manually as wide as visually detectable because the software was not equipped with an automated choroidal segmentation algorithm (Fig 1B). These manual procedures were performed by a retina specialist (Y.H.) and double-checked by a senior retina specialist (M.M.). In case of disagreement, the segmentation was corrected partially after discussion to determine the final grade. The choroidal thickness, defined as the distance between the Bruch’s membrane and the outer border of the choroid, was measured at all points in the area where the outer border of the choroid was segmented. The subfoveal choroidal thickness (SFCT) was defined as the thickness just under the fovea. The parafoveal choroidal thickness (PFCT) was defined as the thickness in a 6-mm-wide area centered at the fovea and analyzed only for participants with available choroidal thickness in the entirety of that area. For the analysis of SFCT, the mean values of horizontal and vertical scans were used. For the analysis of PFCT, the mean values calculated for each of the following 6 areas were used: 2-mm-wide areas centered at the fovea in horizontal and vertical scan, and those centered at the points 2 mm away from the fovea in the nasal, temporal, superior, and inferior directions.

Figure 1

Measurement of subfoveal choroidal thickness (SFCT) and normalized choroidal intensity obtained with enhanced depth imaging (NCIEDI). A, Raw unsegmented OCT image with enhanced depth imaging. B, Enhanced depth imaging OCT image after automatic and manual segmentation. The vitreous body was detected as the area over the inner limiting membrane (ILM). The nerve fiber layer (NFL) was detected as the area between the ILM and the outer border of the NFL. The choroid was detected as the area between the Bruch’s membrane (BM) and the outer border of the choroid. The SFCT was defined as the distance between the BM and the outer border of the choroid under the fovea. C, Pixels in a range of 6 mm around the fovea were used for measurement of mean vitreous intensity (VI), nerve fiber layer intensity (NFLI), and choroidal intensity (CI). In particular, for measurement of VI, the vitreous area, excluding the upper 40 and lower 20 pixels, was used. Normalized choroidal intensity obtained with enhanced depth imaging was calculated as follows: NCIEDI = (mean CI – mean VI) / mean NFLI.

For participants who underwent a 9-mm scan, we calculated the normalized choroidal intensity (NCI) for a 6-mm width around the fovea, which provided an estimate of the absolute brightness of the choroid in OCT and was thought to reflect the dilation of choroidal vessels. The concept of normalization involves correction of choroidal intensity using the brightness of the vitreous body and NFL as reference values for low and high brightness, respectively, to minimize the effects of several factors, such as media opacity and scan tilt, and to enable comparison of the choroidal intensity between individuals. We performed this procedure with EDI OCT, and the calculated value was referred to as NCI obtained with EDI (NCIEDI). Although the detailed calculation method has been described elsewhere, we briefly summarize it here. First, we calculated the mean values of the vitreous intensity, NFL intensity, and unnormalized choroidal intensity from the scan pixel measurements by ImageJ software (National Institutes of Health). Then, NCIEDI was calculated as follows: NCIEDI = (mean choroidal intensity – mean vitreous intensity) / mean NFL intensity. The resultant NCIEDI is unitless. In this procedure, to define the respective layers, the boundaries decided automatically and manually by built-in software were used, as described above. Especially for measuring vitreous intensity, the vitreous area excluding the upper 40 pixels (the upper edge of the image that can contain a lot of noise) and the lower 20 pixels (the vitreoretinal border region where the brightness can increase because of the posterior vitreous membrane) was used. The measurement procedure and calculation of NCIEDI are demonstrated in Figure 1. For the analysis of NCIEDI, the mean values of horizontal and vertical scans were used.

For all participants, CVI was calculated and analyzed. Although the detailed calculation method has been described elsewhere, we summarize it here briefly. First, the entire image of the horizontal B-scan was binarized using the Niblack method. Second, a region of the choroid 1.5-mm wide centered at the fovea center was extracted. Third, in the extracted area, the total choroidal area and the dark area (luminal area) were measured. The CVI was computed as the percentage of luminal area to total choroidal area. Similar to the NCIEDI measurements, we used the boundaries determined automatically and corrected manually by the built-in software to define the respective layers. The measurement procedure and calculation of CVI were performed using Fiji open source software.

Basic Clinical Parameters

Basic clinical parameters, including physical examination findings and plasma marker concentrations, were obtained from all participants. Age was calculated by dividing the number of days from birth to the date of examination by 365. Body mass index was calculated as follows: body mass index = weight / (height)2 (kg/m2). We measured central systolic blood pressure by using commercially available equipment (HEM-9000AI; Omron Healthcare), as reported previously., Briefly, the participants’ radial arterial waveform and brachial blood pressure after resting in the sitting position for 5 minutes were measured simultaneously. The measurements were measured twice, and the mean value was used for the analysis. The absolute pressure of the late systolic peak of the radial arterial waveform was defined as the central systolic blood pressure.

Statistical Analysis

Continuous data are shown as mean and standard deviation or standard error. Categorical data are shown as count and proportion. Because the age and sex distribution of the cohort deviated from those of the Japanese population, we calculated the age-sex standardized mean values of SFCT, NCIEDI, and CVI. We standardized the current data to the population statistics published by the Japanese government and calculated the population estimates of SFCT, NCIEDI, and CVI stratified by age and sex. We also analyzed the distribution of NCIEDI stratified by age and SFCT. The unpaired t test was performed to examine the differences in SFCT and NCIEDI between sexes in each age category. The distribution of PFCT was stratified by age, sex, and area and was analyzed in each scan direction. A paired t test with Bonferroni correction was performed to examine the difference in PFCT between the areas in each sex and age category. The Jonckheere-Terpstra test was performed to examine the trends in these stratified data.

To evaluate the contributions of age, sex, AL, and SE to SFCT, we conducted linear regression analysis with adjustments for known possible contributing factors. In addition, to evaluate the associations of age, sex, AL, and SE with NCIEDI and CVI, we also conducted linear regression analysis adjusted for known contributing factors, including choroidal thickness. Normalized choroidal intensity obtained with EDI was used after log10-transformation.

P values less than 0.05 were considered to be statistically significant. The statistical analyses were performed using R software version 4.1.0 (R Foundation for Statistical Computing).

Results

Finally, 8881 participants were included in the current study. Participants’ clinical characteristics are shown in Table 1. The number of women was approximately twice as large as the number of men. The mean age was 57.6 ± 12.4 years. The mean IOP, AL, and SE were 14.6 ± 3.0 mmHg, 24.11 ± 1.38 mm, and –1.47 ± 2.89 diopters, respectively. The age-sex standardized SFCT, NCIEDI, and CVI were 291.2 μm, 0.653, and 66.88%, respectively.

Table 1

Participants’ Demographic Characteristics and Standardized Values of Choroidal Parameters

Characteristics	Data
Sex (male/female)	2839 (32.0)/6042 (68.0)
Age (yrs)	57.6 ± 12.4
Body mass index (kg/m2)	22.2 ± 3.3
Central systolic blood pressure (mmHg)	130.1 ± 19.4
HbA1c (%)	5.55 ± 0.49
High-density lipoprotein (mg/dL)	67.8 ± 17.4
Low-density lipoprotein (mg/dL)	118.4 ± 29.2
Intraocular pressure (mmHg)	14.6 ± 3.0
Central corneal thickness (μm)	543.88 ± 28.70
Axial length (mm)	24.11 ± 1.38
Spherical equivalent (D)	–1.47 ± 2.89
Subfoveal choroidal thickness (μm)∗	291.2
NCIEDI (unitless)∗,†	0.653
Choroidal vascularity index (%)∗	66.88

D = diopter; HbA1c = glycated hemoglobin; NCIEDI = normalized choroidal intensity obtained with enhanced depth imaging.

Data are presented as mean±standard deviation, mean, or no. (%), unless otherwise indicated.

Age-sex standardized values are shown.

Number of eligible participants was 4561.

The age-sex stratified distribution of SFCT is shown in Figure 2, and the age-sex stratified population estimates of SFCT, NCIEDI, and CVI are shown in Table 2. Although mean SFCT did not differ significantly between men and women overall (P = 0.87), men showed a significantly thicker choroid than women among the younger participants (35–40 years of age: 346.9 ± 6.8 μm vs. 307.2 ± 3.8 μm [P < 0.001]; 40–50 years: 329.1 ± 4.4 μm vs. 302.3 ± 2.6 μm [P < 0.001]). In both men and women, SFCT was associated negatively with age (P < 0.001). Similarly, although mean NCIEDI did not significantly differ between men and women overall (P = 0.21), men tended to show lower intensity than women among the younger participants (35–40 years of age: 0.533 ± 0.021 vs. 0.572 ± 0.012 [P = 0.11]; 40–50 years of age: 0.548 ± 0.010 vs. 0.596 ± 0.007 [P < 0.001]). In both men and women, NCIEDI was associated positively with age (P < 0.001). Choroidal vascularity index was significantly higher in men than in women overall (66.97 ± 0.06% vs. 66.79 ± 0.04%; P = 0.015) and men tended to show significantly higher CVI than women among the younger participants (35–40 years of age: 67.65 ± 0.14% vs. 67.12 ± 0.11% [P = 0.004]; 40–50 years of age: 67.42 ± 0.10 vs. 66.86 ± 0.08 [P < 0.001]). Overall, CVI was associated negatively with age (P < 0.001).

Figure 2

Boxplot showing the distribution of subfoveal choroidal thickness (SFCT) stratified by age and sex. The smallest first quartile (181.0 μm in men ≥ 70 years of age) and the largest third quartile (427.1 μm in men < 40 years of age) are shown as dotted lines for reference. In both men and women, SFCT was associated negatively with age (P < 0.001). Note that men showed a significantly thicker choroid than women among younger participants younger than 50 years (P < 0.001), and no statistically significant difference in choroidal thickness was found between sexes among older participants 50 years of age or older. This suggested the interaction between age and sex on SFCT.

Table 2

Age-Sex Stratified Population Estimates of Choroidal Parameters

Parameter	Age (yrs)	Men		Women		All		P Value∗
		Mean (Standard Error)	No. (%)	Mean (Standard Error)	No. (%)	Mean (Standard Error)	No. (%)
SFCT (μm)	35–39	346.9 (6.8)	244 (2.7)	307.2 (3.8)	637 (7.2)	318.2 (3.4)	881 (9.9)	< 0.001
	40–49	329.1 (4.4)	555 (6.2)	302.3 (2.6)	1372 (15.4)	310.0 (2.3)	1927 (21.7)	< 0.001
	50–59	297.9 (5.0)	436 (4.9)	292.9 (2.8)	1335 (15.0)	294.1 (2.4)	1771 (19.9)	0.37
	60–69	271.5 (3.4)	860 (9.7)	279.6 (2.3)	1753 (19.7)	277.0 (1.9)	2613 (29.4)	0.049
	70+	252.5 (3.5)	744 (8.4)	262.5 (3.4)	945 (10.6)	258.1 (2.5)	1689 (19.0)	0.042
	All	288.3 (2.0)	2839 (32.0)	287.9 (1.3)	6042 (68.0)	288.1 (1.1)	8881 (100.0)	0.87
P value†		< 0.001		< 0.001		<0 .001
NCIEDI (unitless)	35–39	0.533 (0.021)	86 (1.9)	0.572 (0.012)	212 (4.6)	0.560 (0.011)	298 (6.5)	0.11
	40–49	0.548 (0.010)	302 (6.6)	0.596 (0.007)	717 (15.7)	0.582 (0.006)	1019 (22.3)	< 0.001
	50–59	0.620 (0.012)	285 (6.2)	0.628 (0.007)	812 (17.8)	0.626 (0.006)	1097 (24.1)	0.55
	60–69	0.685 (0.011)	455 (10.0)	0.676 (0.008)	900 (19.7)	0.679 (0.007)	1355 (29.7)	0.53
	70+	0.796 (0.016)	364 (8.0)	0.791 (0.015)	428 (9.4)	0.794 (0.011)	792 (17.4)	0.81
	All	0.663 (0.007)	1492 (32.7)	0.653 (0.004)	3069 (67.3)	0.657 (0.004)	4561 (100.0)	0.21
P value†		< 0.001		< 0.001		< 0.001
CVI (%)	35–39	67.65 (0.14)	244 (2.7)	67.12 (0.11)	637 (7.2)	67.27 (0.09)	881 (9.9)	0.004
	40–49	67.42 (0.10)	555 (6.2)	66.86 (0.08)	1372 (15.4)	67.03 (0.06)	1927 (21.7)	< 0.001
	50–59	66.85 (0.15)	436 (4.9)	66.82 (0.08)	1335 (15.0)	66.82 (0.07)	1771 (19.9)	0.84
	60–69	66.93 (0.11)	860 (9.7)	66.81 (0.09)	1753 (19.7)	66.85 (0.07)	2613 (29.4)	0.39
	70+	66.52 (0.15)	744 (8.4)	66.38 (0.14)	945 (10.6)	66.44 (0.10)	1689 (19.0)	0.47
	All	66.97 (0.06)	2839 (32.0)	66.79 (0.04)	6042 (68.0)	66.85 (0.04)	8881 (100.0)	0.015
P value†		< 0.001		0.39		< 0.001

CVI = choroidal vascularity index; NCIEDI = normalized choroidal intensity obtained with enhanced depth imaging; SFCT = subfoveal choroidal thickness.

Difference between sexes was tested by unpaired t test.

Association with age strata was tested by Jonckheere-Terpstra test.

The results of linear regression for SFCT are shown in Table 3. As expected, age, sex, the interaction term of age and sex, AL, and SE were associated significantly with SFCT (P < 0.001). This model showed that SFCT decreased by 28.15 μm for a 1-mm elongation of AL and by 6.23 μm for a 1-diopter myopic shift. Considering the interaction term of age and sex, SFCT decreased by 4.05 μm for each year of age in men and 2.79 μm for each year of age in women.

Table 3

Multivariable Analysis of Associations with Subfoveal Choroidal Thickness

Parameters	β	95% Confidence Interval for β	Standardized β	P Value
Age	–4.05	–4.35 to –3.75	–0.403	< 0.001
Sex (female)	–20.16	–25.06 to –15.27	–0.093	< 0.001
Interaction term of age and sex (female)	1.26	0.91–1.60	0.074	< 0.001
Axial length	–28.15	–30.86 to –25.44	–0.375	< 0.001
Spherical equivalent	6.23	4.94–7.52	0.176	< 0.001
Intraocular pressure	–0.89	–1.65 to –0.12	–0.026	0.023
Body mass index	0.59	–0.12 to 1.29	0.019	0.10
Low-density lipoprotein	0.06	–0.02 to 0.13	0.016	0.13
High-density lipoprotein	0.07	–0.06 to 0.21	0.013	0.27
Central corneal thickness	0.03	–0.04 to 0.11	0.010	0.37
Central systolic blood pressure	0.03	–0.09 to 0.15	0.005	0.66
Glycated hemoglobin	–0.33	–4.74 to 4.09	–0.002	0.88

Parafoveal choroidal thickness was analyzed in 7280 participants. The age- and sex-stratified PFCT distribution is shown in Figure 3. In the horizontal scan (Fig 3A), PFCT was thickest in the foveal area, followed by the temporal area and nasal area (289.7 ± 1.2 μm vs. 280.4 ± 1.0 μm vs. 222.8 ± 1.1 μm; P < 0.001). In the vertical scan (Fig 3B), PFCT showed no significant difference between the superior area and the foveal area (297.2 ± 1.1 μm vs. 296.3 ± 1.2 μm; P = 0.06), whereas PFCT was thinnest in the inferior area (276.9 ± 1.1 μm; P < 0.001). The population estimates of PFCT stratified by age, sex, and area are shown in Supplemental Table 1 (available at https://www.ophthalmologyscience.org/). In all areas, PFCT was significantly thicker in men than women among younger participants (35–50 years of age; P < 0.001), showed no significant difference between sexes in middle-aged participants (50–65 years of age; P > 0.05), and tended to be thinner in men than women in older participants (older than 65 years; P = 0.040 in the inferior area, P = 0.004 in the nasal area, and P < 0.001 in the other areas).

Figure 3

Age-sex stratified parafoveal choroidal thickness (PFCT). A, B, Line graphs and boxplots showing the choroidal thickness according to the distance from the fovea, stratified by age and sex: results from the (A) horizontal scan and (B) vertical scan. A, For any stratum, PFCT was thickest in the foveal area, followed by the temporal area and the nasal area. B, For any stratum, PFCT showed no significant difference between the superior area and the foveal area, whereas it was thinnest in the inferior area (P < 0.001). A, B, Regardless of sex and area, PFCT was associated negatively with age (P < 0.001). Note that men showed a greater decrease in PFCT with age than women in all areas.

Normalized choroidal intensity obtained with EDI was analyzed in 4561 participants. The distribution of NCIEDI stratified by age and SFCT is shown in Figure 4. In all age strata, NCIEDI was correlated negatively with SFCT (P < 0.001), that is, NCIEDI decreased as the choroid became thicker. In each SFCT category (except for < 100 μm), NCIEDI tended to increase with age (SFCT between 100 and 500 μm, P < 0.001; SFCT > 500 μm, P = 0.064). In the scatterplots shown in Supplemental Figures 1 and 2 (available at https://www.ophthalmologyscience.org/), the negative and positive correlation between NCIEDI or CVI and SFCT increased in strength as the patients became older. Table 4 summarizes the results of the linear regression analysis for NCIEDI and CVI. On the basis of Supplemental Figures 1 and 2, we included an interaction term for age and SFCT. Although AL and SE were not associated significantly with NCIEDI (P = 0.069 and P = 0.63, respectively), NCIEDI was associated significantly with age, SFCT, and their interaction term, as expected (β = 1.90 × 10–3, standardized β = 0.161 [P < 0.001]; β = –1.09 × 10–3, standardized β = –0.749 [P < 0.001]; and β = –0.01 × 10–3, standardized β = –0.062 [P < 0.001]; respectively). Similar to NCIEDI, CVI was associated significantly with age, SFCT, and their interaction term (β = –1.74 × 10–4, standardized β = 0.091 [P < 0.001]; β = 0.60 × 10–4, standardized β = 0.259 [P < 0.001]; and β = 0.02 × 10–4, standardized β = 0.130 [P < 0.001]; respectively) and also associated with sex, AL, and SE (β = –26.20 × 10–4, standardized β = –0.051 [P < 0.001]; β = –15.41 × 10–4, standardized β = –0.089 [P < 0.001]; and β = 9.31 × 10–4, standardized β = 0.113 [P < 0.001]; respectively).

Figure 4

Boxplot of the distribution of log10-transformed normalized choroidal intensity obtained with enhanced depth imaging (NCIEDI) stratified by age and subfoveal choroidal thickness (SFCT). In all age categories, NCIEDI was correlated negatively with SFCT (P < 0.001). In each SFCT category (except for < 100 μm), NCIEDI tended to increase with age (SFCT 100–500 μm, P < 0.001; SFCT > 500 μm, P = 0.064).

Table 4

Multivariable Analysis of Associations with Normalized Choroidal Intensity Obtained with Enhanced Depth Imaging and Choroidal Vascularity Index

Parameters	Normalized Choroidal Intensity Obtained with Enhanced Depth Imaging				Choroidal Vascularity Index
	β (× 10–3)	95% Confidence Interval for β (× 10–3)	Standardized β	P Value	β (× 10–4)	95% Confidence Interval for β (× 10–4)	Standardized β	P Value
Age	1.90	1.54–2.26	0.161	< 0.001	–1.74	–2.23 to –1.25	–0.091	< 0.001
Sex (female)	7.12	–0.62 to 14.85	0.023	0.072	–26.20	–37.23 to –15.17	–0.051	< 0.001
Axial length	–4.15	–8.61 to 0.32	–0.038	0.069	–15.41	–21.67 to –9.15	–0.089	< 0.001
Spherical equivalent	–0.51	–2.58 to 1.56	–0.010	0.63	9.31	6.39–12.23	0.113	< 0.001
SFCT	–1.09	–1.13 to –1.05	–0.749	<0.001	0.60	0.55–0.66	0.259	< 0.001
Interaction term of age and SFCT	–0.01	–0.01 to 0.00	–0.062	< 0.001	0.02	0.02–0.03	0.130	< 0.001
Low-density lipoprotein	–0.08	–0.19 to 0.03	–0.016	0.15	0.09	–0.07 to 0.25	0.011	0.28
High-density lipoprotein	–0.13	–0.34 to 0.08	–0.016	0.22	–0.01	–0.31 to 0.28	–0.001	0.93
Body mass index	0.58	–0.53 to 1.69	0.013	0.31	0.81	–0.76 to 2.38	0.011	0.31
Interocular pressure	0.51	–0.69 to 1.72	0.010	0.41	0.24	–1.47 to 1.95	0.003	0.78
Glycated hemoglobin	2.86	–4.07 to 9.80	0.010	0.42	–10.63	–20.52 to –0.74	–0.022	0.035
Central corneal thickness	–0.04	–0.16 to 0.08	–0.007	0.54	–0.03	–0.20 to 0.14	–0.004	0.71
Central Systolic blood pressure	–0.02	–0.22 to 0.17	–0.003	0.80	–0.05	–0.32 to 0.23	–0.004	0.73

SFCT = subfoveal choroidal thickness.

The data for normalized choroidal intensity obtained with enhanced depth imaging and choroidal vascularity index were log10-transformed before analysis.

Discussion

With the development of the EDI technique for OCT, the choroid increasingly is being recognized as a clinically important structure, because the pathologic features of the choroid have been revealed to be associated with various ocular diseases such as AMD, high myopia, Vogt-Koyanagi-Harada disease, and CSC. In particular, the concept of pachychoroid disease has gained attention because pachychoroid pathologic features, such as a thickened choroid and dilated choroidal vessels, can cause retinal pigment epithelium damage and choroidal neovascularization, sometimes resulting in irreversible visual loss. These conditions are termed pachychoroid pigment epitheliopathy and pachychoroid neovasculopathy, respectively, and together with CSC, they constitute the core elements of pachychoroid-related diseases.,

In previous studies,39, 40 we investigated the molecular biological mechanisms underlying pachychoroid through genome-wide association studies of choroidal thickness and CSC. We found that CFH and VIPR2 were the important genetic determinants of choroidal thickness and that these genes together with TNFRSF10A and GATA5 are the susceptibility genes for CSC. We also reported the clinical importance of defining pachychoroid for differentiating pachychoroid neovasculopathy from AMD, because pachychoroid neovasculopathy often seems to be AMD and is misdiagnosed as a result, especially in Asia, where pachychoroid diseases are common., Subsequent studies also support this idea, and the criteria for a thick choroid currently are the subject of debate.,, However, despite the increasing clinical importance of the choroid, only a few large-scale cohort studies have provided normative data for choroidal thickness., In this study, we reported the age-sex stratified normative data for choroidal parameters, that is, choroidal thickness and choroidal intensity, using a relatively large general Japanese cohort from the Nagahama Study (n = 9850).

The age-sex standardized SFCT (291.2 μm) in the present study was thicker than that reported in earlier population-based cohort studies: 253.8 ± 107.4 μm in the Beijing Eye Study, which investigated 3233 eyes of 3233 Chinese individuals 50 to 93 years of age (mean, 64.3 ± 9.6 years), and 255.2 ± 102.6 μm in the SEEDS, which investigated 2794 eyes of 1619 Chinese, Indian, and Malay individuals 50 to 90 years of age (mean, 60.9 ± 7.7 years). This difference may be attributed to the fact that the current study included younger individuals, who tend to have a thicker choroid, than the previous studies. The effect sizes in the Beijing Eye Study and SEEDS were –44.7 μm and –32.1 μm per 1-mm axial elongation, respectively, and –4.12 μm and –4.33 μm per 1 year of age, respectively, which are almost compatible with the data obtained in the current study (Table 3). Parafoveal choroidal thickness also showed a tendency compatible with the result of SEEDS, which analyzed macular volume scans; PFCT was thickest in the superior area, followed by the foveal area, the temporal area, the inferior area, and the nasal area. However, the significant interaction between age and sex on SFCT is a novel finding of the current study (Fig 2; Tables 2 and 3).

The significant interaction between age and sex on SFCT (P < 0.001) indicates that the effect of age on SFCT differs significantly between sexes. In fact, Figure 2 and Table 2 show that men harbor a significantly thicker choroid than women among participants younger than 50 years, whereas no statistically significant differences were found among older participants after Bonferroni correction, with men tending to show a thinner choroid. Because this significant interaction was noted even after adjusting for possible confounders, including refractive status and AL, and a similar trend also was observed in PFCT (Fig 3; Supplemental Table 1), we believe that the results are robust. The most important pachychoroid spectrum disease, CSC, is known to occur predominantly in men (male-to-female ratio, 7:1) in their 40s. The current distribution of choroidal thickness may suggest the presence of the underlying choroidal pathologic features of CSC among younger men. To date, although the choroidal thickness in men has been reported to be more than that in women,7, 8, 9 age-sex stratified analysis has not been conducted, probably because of the limited sample sizes of previous reports. The novel finding of the current study should be replicated in other cohorts and large epidemiologic consortiums, such as the European Eye Epidemiology Consortium and the Asian Eye Epidemiology Consortium.

Pachychoroid is characterized not only by choroidal thickness, but also by dilation of choroidal vessels., Several methods for quantifying choroidal vessel dilation have been proposed to improve our understanding of the choroidal structure in more detail. For example, Branchini et al introduced the light-to-dark ratio, which is an indicator of the ratio of the choroidal stromal area to the choroidal vessel lumen area. Sonoda et al, using the freely available ImageJ software, evaluated the proportion of black (luminal) and white (interstitial) areas on binarized EDI choroidal images, and this evaluation method was later modified into CVI by Agrawal et al. Similarly, Balasubramanian et al calculated the NCI, which is the absolute brightness of the choroid with reference to the dark structure vitreous body and the bright structure NFL. The age-sex standardized CVI (66.88%; n = 8881) in the present study was consistent with that reported in an earlier population-based study (mean ± standard deviation, 65.61 ± 2.33%; n = 345).

These parameters, that is, indicators of choroidal vessel dilation, are known to correlate strongly with choroidal thickness,, which also has been observed in the current study (Fig 4; Supplemental Figs 1 and 2). In addition, our age-stratified analysis (Supplemental Figs 1 and 2) and multivariable analysis (Table 4) were the first to reveal that the strength of the correlation between SFCT and NCIEDI or CVI varied with age (P < 0.001); younger individuals showed a more lumen-rich choroid for their choroidal thickness than older individuals. This is consistent with the findings of previous histologic reports in which blood vessels (low-intensity areas) decrease with age more predominantly than stromal tissue (high-intensity areas).47, 48, 49 Although CVI was reported previously to be associated with only SFCT in a multivariable analysis, we were able to find additional controlling factors because of a larger sample size, notably age, sex, AL, SE, and the interaction term of age and SFCT (Table 4). Age, SFCT, and their interaction term were associated significantly with both NCIEDI and CVI, suggesting that these parameters represent similar aspects of the choroidal status, although not exactly the same. Further research is needed to clarify the clinical relevance and differences between these parameters. Considering choroidal thickness and brightness characterized different aspects of the choroidal structure, a combination of these parameters may allow differentiation of the pachychoroid spectrum.

The potential limitations of our study need to be mentioned. First, because the current study was a community-based one, the overall distributions of age and sex were different from those in the normal Japanese population. However, by using age-sex based stratification and standardization to present summary data, we believe we minimized its influence. Second, although the current study presented normative NCIEDI data for the first time, these values need to be interpreted carefully. Because the current study used EDI OCT images, which increase the intensity of the choroid, unlike the findings in the original report proposing NCI, NCIEDI is higher than NCI obtained with normal OCT images. However, because EDI is currently essential for evaluating the choroid, we believe that normative NCIEDI data are rather important. Finally, the generalizability of these findings is limited to the Japanese population undergoing assessments with the SD-OCT RS-3000 Advance device. Further replication in other ethnicities and with various OCT methods is expected.

In summary, we present normative data for choroidal thickness, choroidal intensity, and CVI in Japanese individuals by using a large general Japanese population from the Nagahama Study. The current data indicated the presence of a significantly thicker choroid in younger men, which might have some relationship to pachychoroid pathogenesis. The association analysis of SFCT with NCIEDI and CVI suggested that younger individuals showed a more lumen-rich choroid for their choroidal thickness than older individuals. Further replication of these findings in larger cohorts such as an epidemiologic consortium is expected.