Asymmetry of Macular Vessel Density in Bilateral Early Open-angle Glaucoma With Unilateral Central 10-2 Visual Field Loss

Supplemental Digital Content is available in the text. Précis: Glaucomatous eyes without detectable 10-2 visual field loss showed significant macular vessel density loss, especially in inferior quadrant of perifoveal area. Macular vessel density loss spatially corresponded with structural and functional damage. Purpose: The purpose of this study was to investigate the characteristics, intereye and intraeye asymmetry of macular vessel density assessed by optical coherence tomography (OCT) angiography in bilateral early open-angle glaucoma with unilateral 10-2 visual field loss. Materials and Methods: Thirty-two eyes of 16 patients with bilateral early open-angle glaucoma and unilateral 10-2 visual field loss, and 13 eyes of 13 healthy participants were consecutively enrolled. All subjects underwent 30-2, 10-2 visual fields, OCT, and OCT angiography examinations. Intereye differences were compared between the perimetrically affected eye and the unaffected eye in the same patient. Intraeye differences were compared between the affected hemifields and the unaffected hemifields in the same eye with single-hemifield 10-2 visual field loss. Results: Macular whole image vessel density of the perimetrically unaffected eyes was lower than the healthy eyes (46.6% vs. 51.1%; P<0.001). Parafoveal vessel density parameters of the perimetrically affected eyes were comparable to the unaffected eyes (all P>0.05). Although inferior perifoveal vessel density of the perimetrically affected eyes was significantly lower than the unaffected eyes (42.2% vs. 46.2%; P=0.007), similar results were found in macular ganglion cell complex. In glaucomatous eyes with single-hemifield loss, perifoveal vessel density and macular ganglion cell complex of the affected hemifields were significantly worse than the unaffected hemifields (43.6% vs. 47.0%, 78.4 μm vs. 89.0 μm; P=0.023 and 0.005; respectively). Conclusions: Significant macular microvascular damage was present in glaucomatous eyes without detectable 10-2 visual field damage. The damage of inferior perifoveal vessel density was more significant in early glaucoma. Macular microvascular damage spatially corresponded with functional and structural damage.

O pen-angle glaucoma (OAG) is an optic neuropathy characterized by progressive damage of retinal ganglion cells (RGCs) and their axons. 1 Macula has the highest RGC density, with ∼50% of RGCs located in this region. 2,3 This anatomic structure makes macula an important position for early detection of structural and vascular changes in glaucoma. Previous studies have found that macula was involved in the early glaucomatous damage. [3][4][5][6] The pathogenesis of OAG remains unknown, but mounting studies have suggested that vascular factors played an important role in the development and progression of glaucoma. 7,8 Optical coherence tomography angiography (OCTA) based on the split-spectrum amplitude-decorrelation angiography algorithm can quantitatively and qualitatively visualize the decreased vessel density in glaucomatous eyes with high repeatability. [9][10][11] Few studies investigated the intraeye and intereye asymmetry of the macular vessel density in glaucomatous eyes. 12,13 However, these 2 studies used 24-2 visual field to assess the visual function of glaucoma. Although central 24-2 or 30-2 visual field test pattern is widely used in clinical setting for assessing glaucomatous function loss, De Moraes et al 14 found that commonly used glaucoma staging systems including Hodapp-Parrish-Anderson, visual field index, and Brusini staging systems, which are based on 24-2 or 30-2 visual field test results, might fail to detect the presence of glaucomatous macular damage and underestimate the disease severity. Previous study found that 24-2 visual filed could miss central visual field loss manifested in 10-2 visual field. 15 It is not surprising, because the test points of 10-2 visual field space every 2 degrees, which are denser than those of 24-2/30-2 visual field test pattern (grid 6 degrees). 15,16 More points of 10-2 visual field test pattern are tested within the macular region and could be more accurate to reflect the functional damage of macula.
OAG is generally a bilateral disease that manifests asymmetric glaucomatous optic nerve damage and visual field loss. 1 To understand the characteristics of early glaucomatous damage in macula, the current study investigated the differences of macular microvasculature between the perimetrically affected eyes and the unaffected eyes, and between the hemifields in glaucomatous eyes with singlehemifield visual field loss, based on central 10-2 visual field, which might provide insight into the pathophysiological mechanism of glaucoma.

MATERIALS AND METHODS
This prospective cross-sectional study was conducted at the Department of Glaucoma, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China, between January 2019 and May 2019. The study protocol was conducted in accordance with the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of Zhongshan Ophthalmic Center. Written informed consent was obtained from all participants.
Blood pressure including systolic blood pressure and diastolic blood pressure was measured in the left brachial artery at the height of heart by an automatic blood pressure instrument (Model: HM-7136; Omron Inc., Japan). Mean ocular perfusion pressure is two thirds of the mean artery pressure minus IOP, where mean artery pressure is calculated as two thirds of the diastolic blood pressure plus one third of the systolic blood pressure.

Study Subjects
Consecutive patients diagnosed as bilateral early OAG with 10-2 visual field loss in 1 eye and intact 10-2 visual field result in the contralateral eye, were included in this study. OAG patients of this study consist of juvenile-onset openangle glaucoma (JOAG) and primary open-angle glaucoma (POAG, also known as adult-onset OAG) patients. The common inclusion criteria for OAG in the current study were the presence of glaucomatous optic disc changes (neuroretinal rim thinning, excavation, notching, or retinal nerve fiber layer defect) confirmed by 2 glaucoma experts (J. H. and H.X.) on both dilated fundus examination and stereoscopic optic disc photographs, a history of untreated IOP > 21 mm Hg, above 18 years old. The major difference between JOAG and POAG is the age of glaucoma onset: the age of JOAG patient is younger than 35 years old, and the age of POAG patient is 35 years or older. 17,18 Visual field results were not used for the definition of OAG. Although the definition of early stage of OAG was based exclusively upon the 30-2 visual field test with mean deviation better than −6.0 dB in at least 2 consecutive and repeatable visual field tests. Central 10-2 visual field loss was defined as a cluster of 3 continuous points (5%, 5%, 1%; 5%, 2%, 2%; or worse) within 1 hemifield on the pattern deviation or total deviation plot based on cluster criteria, and the points of the cluster were allowed to lie on the edge of the visual field plot. 15,19 Healthy participants were recruited from hospital staff, who were required to have no family history of glaucoma, IOP < 21 mm Hg, normal optic nerve appearance, and normal 10-2 and 30-2 visual fields which were defined as pattern standard deviation within the 95% confidence limits and glaucoma hemifield test result within normal limits, in both eyes. Only 1 eye in 1 healthy participant was randomly selected.
Common inclusion criteria for all participants were age above 18 years, open-angle on gonioscopy, best-corrected visual acuity ≥ 20/40, refractive error between +3.0 and −6.0 D. Exclusion criteria were coexisting ocular diseases, such as retinal disease, nonglaucomatous optic neuropathy, uveitis; a medical history of systemic disease, such as diabetes mellitus, hypertension; an intraocular surgery except for uncomplicated cataract surgery 6 months before enrollment; and presence of significant media opacities which prevented from obtaining high-quality images.

Visual Field Tests
All participants underwent visual field examinations using central 30-2 and central 10-2 pattern with Swedish Interactive Threshold Algorithm in a standard condition (Humphrey Field Analyzer II; Carl Zeiss Meditec Inc., Dublin, CA). 30-2 visual field test and 10-2 visual field test were performed on the same day. 10-2 visual field test was performed after 30-2 visual field test with a rest period of at least 15 minutes. Only reliable visual field results, defined as fixation losses <20%, false-positive rates <15%, false-negative rates <30%, were included in the current study. We further categorized 10-2 visual filed defects as single-hemifield defect (superior or inferior hemifield) and bilateral hemifields defect (Fig. 1). The eyes with single-hemifield defect were selected for further intraeye analyses.

SD-OCT and OCTA Imaging
Imaging of fundus ultrastructure and microvasculature were acquired with a RTVue-XR Avanti SD-OCT system with Angiovue (Optovue Inc.). Pupil was dilated before imaging. The Optic Nerve Head scan protocol was used to measure circumpapillary retinal nerve fiber layer (cpRNFL) thickness and rim area, which centered on the optic disc along a 3.45 mm diameter circle. The ganglion cell complex scan protocol, which centered 1 mm temporal to the fovea and covered a 7 mm diameter circular area, was used to measure macular ganglion cell complex (mGCC) thickness.
High-density Angio Retina 6×6-mm protocol was used to cover the macula. We analyzed the superficial vessel density, which extended from inner limiting membrane to inner plexiform layer 9 μm below. Vessel density measurement was automatically obtained with built-in software (Optovue Inc.; software version: 2017.1.0.151). Macular whole image vessel density was measured in a 6×6-mm 2 area. Parafoveal vessel density was measured within an annulus with inner diameter of 1 mm and outer diameter of 3 mm. Perifoveal vessel density was measured within an annulus with inner diameter of 3 mm and outer diameter of 6 mm. Vessel density was defined as the total area occupied by the vessels.
Poor quality images showing scan quality <7/10, vitreous floater artifacts or segmental errors were excluded. 5,20 Statistical Analysis Statistical analyses were performed using SPSS software (version: 20.0; SPSS Inc., Chicago, IL). Continuous variables were presented as means (95% confident interval). χ 2 Test was used to compare categorical variables such as sex. The sample size of this study was relatively small, therefore, nonparametric tests was used. Mann-Whitney test was used to compare the differences between healthy eyes and 10-2 visual field affected eyes, and between healthy eyes and 10-2 visual field unaffected eyes. Wilcoxon signed ranks test was used to compare the intereye differences between 10-2 visual field affected eyes and 10-2 visual field unaffected fellow eyes, and the intraeye differences between affected hemifields and unaffected hemifields in perimetrically affected eyes with single-hemifield. P-value <0.05 was considered statistically significant. Bonferroni correction was applied for multiple comparisons. P-value <0.017 (0.05/3) was considered statistically significant in multiple comparisons.

RESULTS
Thirty-two eyes of 16 patients with bilateral early OAG and unilateral 10-2 visual field loss and 13 eyes of 13 healthy participants met the inclusion criteria and were included in the current study. There were 8 patients aged 20 to 35 years old, diagnosed as JOAG; and 8 patients older than 35 years old, diagnosed as POAG. The demographic and ocular characteristics of healthy eyes and OAG eyes are described in Table S1 (Supplemental Digital Content 1, http://links.lww.com/IJG/ A398). There were no significant differences between OAG patients and healthy participants concerning age, sex, and mean ocular perfusion pressure (all P > 0.05). The axial length of 10-2 perimetrically affected eyes was longer than 10-2 perimetrically unaffected eyes, while the difference was borderline significant [24.2 (23.5, 24.8) mm vs. 24.1 (23.5, 24.8) mm; P = 0.017]. As expected, all results of cpRNFL thicknesses of the perimetrically affected eyes were thinner compared with unaffected eyes and healthy eyes (all P < 0.017), while only average and inferior mGCC thicknesses of the affected eyes were thinner compared with the unaffected eyes (P = 0.001 and 0.001, respectively). No optic disc hemorrhage was found in any patients of this study.
Superficial macular vessel density measurements in healthy, perimetrically affected and unaffected eyes are shown in Table S2 (Supplemental Digital Content 1, http://links.lww.com/IJG/ A398). All results of the macular vessel densities of the affected eyes, except for nasal parafoveal vessel density (P = 0.028), were significantly decreased compared with healthy eyes (all P < 0.017). All results of macular vessel densities of the unaffected eyes, except for superior and nasal parafoveal vessel density, nasal perifoveal vessel density (all P > 0.017), were significantly lower compared with healthy eyes (all P < 0.017). The intereye analyses showed that macular whole image vessel density of the affected eyes was comparable to the unaffected eyes [44.7% (42.8%, 46.6%) vs. 46.6% (45.1%, 48.0%); P = 0.028]. In regard to the vessel density in different areas of the macula, we found that average and all quadrants' parafoveal vessel densities of the affected eyes were comparable to the unaffected eyes (all P > 0.017). However, inferior perifoveal vessel densities of the affected eyes were significantly lower than the unaffected eyes (P = 0.007).
Twelve eyes with single-hemifield loss from 16 eyes with 10-2 visual field loss were selected for further intraeye analyses. Comparisons of visual field, structural, microvascular results in the hemispheres corresponding to the affected hemifields and the unaffected hemifields in glaucomatous eyes with single-hemifield 10-2 visual field loss are summarized in Table S3 (Supplemental Digital Content 1, http://links.lww.com/IJG/A398). As expected, mean sensitivity of the unaffected hemifields was significantly Representative cases of a healthy eye, and 2 eyes of a patient with bilateral early OAG and unilateral 10-2 visual field loss were presented in Figure 2. Macular vessel densities of the affected eye and the unaffected eye were significantly decreased compared with the healthy eye. The affected eye showed a well-defined inferior arcuate macular microvascular defect that spatially corresponded well with inferior mGCC thinning and superior 10-2 visual field defect.

DISCUSSION
To the best of our knowledge, this is the first study investigating the intraeye and intereye asymmetry of macular vessel density in bilateral early OAG eyes with unilateral 10-2 visual field loss. In the current study, we found that macular vessel density was significantly decreased even in 10-2 perimetrically unaffected glaucoma eyes. The intereye and intraeye analyses showed that significant damage of macular vessel density was present in perifoveal but not parafoveal area in early OAG. In addition, the damage of macular vessel density spatially corresponded to mGCC damage and visual field damage.
Macula has the highest density of RGC and is crucial for daily visual function such as reading and driving. 3 The vessels supplying nutrition for the macula to meet their high metabolism derive from the central retinal artery and contain exclusively the capillaries and small vessels (arterioles and venules). 21 Previous studies found that central 10-2 visual field could be more accurate to reflect the function and the glaucomatous damage of macula when compared with 24-2/30-2 visual field. 15,16 Therefore, the current study adopted 10-2 rather than 24-2/30-2 visual field to investigate the differences between the perimetrically affected eyes and the unaffected eyes in the same patient, and the differences between the hemifields in the same glaucomatous eye with single-hemifield visual field loss.
In regard to the structural changes in OAG eyes with unilateral 24-2 visual field loss, previous studies found that perimetrically unaffected eyes had smaller rim area, thinner cpRNFL and mGCC when compared with healthy eyes. 12,22,23 Similarly, we found that rim area, cpRNFL, and mGCC of glaucomatous eyes with unaffected 10-2 visual field were significantly decreased compared with healthy eyes. Our results indicated that significant structural damage occurred even in glaucomatous eyes with unaffected 10-2 visual field.
With the application of OCTA, the vessel density in whole image area and in various areas of macula could be noninvasively quantified with high repeatability. 24 Previous study found that parafoveal vessel density of 24-2 visual field unaffected eyes was comparable to the contralateral affected eyes, and was significantly lower than healthy eyes. 12 Similarly, average parafoveal vessel density of the 10-2 visual field unaffected eyes was comparable to the affected eyes and was significantly lower than healthy eyes in the current study, which indicated that significant microvascular damage of macula occurred in eyes without detectable 10-2 visual field loss. Previous studies qualitatively observed significant vessel dropout in peripheral area but not in parafoveal area. 25,26 Furthermore, we conducted quantitative research and found that inferior perifoveal vessel density of the affected eyes was significantly lower than the unaffected eyes. Moreover, we found that inferior mGCC thicknesses of the affected eyes was also thinner than the unaffected eyes. These findings suggested that the damage of perifoveal vessel density, especially in inferior quadrant, was more significant and spatially corresponded to the structural damage in early OAG.
Glaucomatous eyes with single-hemifield loss are good research objects for investigating the association between damage of retinal microvasculature and structural damage of macula, and between damage of retinal microvasculature and visual field loss. A previous study compared the hemifield differences of the macular vessel density in early to moderate OAG eyes with single-hemifield defect of 24-2 visual field, and found that parafoveal vessel density of the unaffected hemifields was significantly higher than the affected hemifields. 13 Although we found comparable parafoveal vessel density between the unaffected hemifields and the affected hemifields in early OAG. The probable explanation accounting for this discrepancy may due to different inclusion criteria between the 2 studies. The current study, which was based on 10-2 visual field loss, included exclusively the early glaucoma and focused on the glaucomatous damage on macula. This finding suggested that parafoveal vessel density might not be involved in the very early damage of glaucoma. However, it was interesting to find that perifoveal vessel density and mGCC thickness of the affected hemifields were lower and thinner, respectively, compared with the unaffected hemifields. This finding further suggested that the damage of perifoveal vessel density was more prominent in early OAG, and macular microvasculature defect spatially corresponded to structural damage and visual field loss.
The average age of glaucoma patients of this study is relatively young when compared with previous studies. 12,13 Eight of 16 early OAG patients were younger than 35 years old. These OAG patients can be also diagnosed as JOAG which is characterized by the presence of OAG in patients younger than 35 years old. 17,18 Overall, the current study comprises both the JOAG and adult-onset OAG (POAG), which represent a relative full spectrum of OAG. Therefore, the results and conclusions of this study may not be generalizable to the exclusive adult-onset OAG (POAG) patients.
A previous study found that eyes with visual field defect within 10 degree had a higher incidence of optic disc hemorrhage. 27 Similarly, a recent study by Shukla et al 28 found that optic disc hemorrhage was significantly correlated with the presence and progression of both 24-2 and 10-2 visual field defects. Although we found no optic disc hemorrhage in any patients of this study, we recommend the careful scrutinization of the central visual field using 10-2 visual field test in glaucoma patients with optic disc hemorrhage.
There are several limitations to the current study. First, sample size was relatively small in this study. However, the differences of variable parameters were considerably compared within intereyes and intraeyes. This approach is advantageous in improving the statistical power because it avoids the effect of confounding factors on vessel density measurement such as age, sex, and anatomic variabilities. 12 Second, we cannot determine the cause-effect relationship between microvascular damage and structural damage because of the nature of cross-section of this study. Longitudinal study is needed to resolve this issue.
In conclusion, significant macular microvascular damage was present even in glaucomatous eyes without detectable 10-2 visual field loss. The damage of perifoveal vessel density, especially in inferior quadrant, was more prominent in early glaucoma. Macular microvascular damage spatially corresponded to structural and visual field damages. Further study is needed to investigate the cause-effect relationship between microvascular damage and structural damage in early OAG patients.