Structure Function Relationships of Vessel Density and Retinal Nerve Fiber Layer Thickness in Early Glaucomatous Eyes With High Myopia

Abstract

Purpose: The purpose of this study was to evaluate the structure function relationship of circumpapillary vessel density (cpVD) with visual field sensitivity (VFS) and compare its characteristics with circumpapillary retinal nerve fiber layer thickness (cpRNFLT) in early glaucomatous (EG) and normal eyes with and without high myopia (HM).

Methods: Seventy-five EG (mean deviation > −6 dB) and 7 normal eyes with HM (axial length [AL] >26.5 mm) and 111 EG and 11 normal eyes without HM were enrolled in this retrospective cross-sectional study. All patients underwent circumpapillary optical coherence tomography (OCT) and OCT angiography (OCTA) scanning with the Cirrus HD-6000 with AngioPlex OCTA (Carl Zeiss Meditec, Dublin, CA, USA). Structure function correlations were determined by comparing global, superior, inferior, and Garway-Heath sectoral values for cpVD and cpRNFLT with its corresponding 24-2 and 10-2 VFS of Humphrey Visual Field (HFA) analyzer.

Results: CpVD showed no significant correlations with AL except for the nasal sector (P = 0.044), whereas cpRNFLT demonstrated significant positive association with AL in the global (P = 0.024), nasal (P = 0.020), and temporal (P < 0.001) sectors. In HM eyes, global and sectoral cpVD significantly correlated with corresponding VFS in all 24-2/10-2 VF sectors (all P < 0.05). CpVD-VFS correlation was significantly stronger than cpRNFLT-VFS in the nasal sector of HM eyes (P = 0.002) and temporal and inferior temporal sector of eyes without HM (P = 0.008 and P = 0.042, respectively).

Conclusions: In EG eyes with HM, cpVD was less affected by AL in comparison to cpRNFLT and cpVD-VFS correlation was significant in all 24-2/10-2 VF sectors. AL-associated advantages of cpVD-VFS over cpRNFLT-VFS were observed.

The global prevalence of high myopia (HM) is increasing dramatically and is estimated to reach 1 billion people by 2050.¹ Myopia is a well-known risk factor for open angle glaucoma (OAG)² and the risk of OAG is about six times higher in HM eyes.³ As glaucomatous damage is irreversible, early diagnosis of glaucoma in HM eyes is essential. However, diagnosis of glaucoma upon optic disc ophthalmoscopy and/or fundus photographs is challenging in HM eyes due to its characteristic disc morphology changes accompanied by axial length (AL) elongation (i.e. reduced contrast among the neuro-retinal rim and optic cup, optic disc tilt, and presence of peripapillary atrophy [PPA]).⁴ The role of optical coherence tomography (OCT) technology is highly important in objectively detecting early glaucomatous damage.^4–6 However, evaluation of HM eyes by conventional OCT parameters, such as circumpapillary retinal nerve fiber layer thickness (cpRNFLT), is also limited by confounding myopic anatomic changes, variation in retinal nerve fiber bundle distribution patterns, and magnification error induced inaccuracy, indicating the necessity for an alternative parameter to assess subtle glaucomatous changes that occur in early glaucomatous HM eyes.^6,7 

OCT angiography (OCTA) is one of the most recent technologies which enables noninvasive imaging to quantify layer specific ocular microcirculation of the retina and choroid. Although several studies have shown the usefulness of evaluating circumpapillary vessel density (cpVD) in detecting focal glaucomatous changes of early glaucomatous eyes,⁸ little is known about the potential of OCTA in HM eyes. There are reports of cpVD being reduced in HM eyes in comparison to eyes without HM,⁹ suggesting the influence of AL elongation on OCTA parameters. However, cpVD is a density parameter and is expected to be less sensitive to magnification effects. Exclusion of large vessels when calculating cpVD may also mitigate the impact of myopic nerve and vessel distribution pattern changes on its results. Two previous studies have shown significant vascular function relationship in moderate to advanced glaucomatous HM eyes and demonstrated the potential of cpVD as a parameter less affected by anatomic variation in comparison to cpRNFLT for evaluation of glaucomatous HM eyes.^10,11 However, to the best of our knowledge, structure function relationship of cpVD in early glaucomatous HM eyes, in which diagnosis is often difficult with the use of cpRNFLT, is yet to be explored. 

Furthermore, assessment of the relationship between glaucomatous structural change and loss of central visual field (VF) is very important in HM eyes because central VF damage (VFD) is observed even in the earlier stages of glaucomatous HM eyes¹² and has substantial impact on quality of vision.¹³ The relationship between cpVD and central 10-2 VF in glaucomatous eyes with HM has not been investigated despite the fact that VFD within the central 10-degree VF is often missed with 24-2 VF testing.¹⁴ 

We hypothesized that the advantages of cpVD in HM eyes will lead to a better structure function relationship in comparison to cpRNFLT in glaucomatous eyes with HM even in the earliest stages of the disease. We also hypothesized that differences in sectoral structure function relationships arising from myopic anatomic changes will be revealed and emphasize the strengths of OCTA in HM eyes. The purpose of this study is to evaluate global and sectoral structure function relationships of cpVD and 24-2/10-2 VF results in early glaucomatous eyes (including preperimetric glaucoma [PPG] and normal eyes) with and without HM and to compare them with the relationship between cpRNFLT and VF to explore the characteristics and potential of cpVD in HM eyes in the earliest stages of glaucoma. 

Methods

Participants

Protocols for this retrospective observational study were approved by the Research Ethics Committee of the Graduate School of Medicine and Faculty of Medicine at The University of Tokyo (Identifier: 2217) and adhered to the tenets of the Declaration of Helsinki. Patients gave informed consent for their information to be stored in the hospital database and used for retrospective research at their first visit. Study participants were notified of the protocol posted at the outpatient clinic and were provided with the opportunity to opt out of the study. 

Participants of this study included patients with early OAG or PPG (mean deviation [MD] > −6 dB) with and without HM and normal subjects who consulted for a routine eye examination or refractive error from the University of Tokyo Hospital (Tokyo, Japan), Yotsuya Shirato Eye Clinic (Tokyo, Japan) and Tajimi Iwase Eye Clinic (Gifu, Japan) between June 2020 and February 2023. HM was defined as AL >26.5 mm. 

All participants included in this study underwent the following ocular examinations: refraction and corneal curvature radius measurements, best corrected visual activity (BCVA), AL measurements, central corneal thickness (CCT) measurements, corneal curvature radius measurements, slit-lamp examination, intraocular pressure (IOP) measurements with Goldmann applanation tonometry, gonioscopy, fundus examination including optic nerve head (ONH) examination, optic disc stereophotography, OCT and OCTA imaging and Humphrey Field Analyzer (HFA; Carl Zeiss Meditec, Dublin, CA, USA) measurements with the 24-2 Swedish Interactive Threshold Algorithm standard strategy. Participants with early OAG and PPG also underwent HFA measurements with the 10-2 Swedish Interactive Threshold Algorithm standard strategy. 

Diagnosis of each eye was conducted independently by two glaucoma specialists (authors K.A. and H.S.) and disagreements were resolved by a third adjudicator (author M.A.). Inclusion criteria for this study were BCVA of 20/25 or better, AL <30 mm, attainment of good quality OCT/OCTA scanning (signal strength indices >7), and reliable VF results (<20% fixation loss, <15% false negative errors, and <15% false positive errors). Glaucomatous eyes had open angles on gonioscopy, glaucomatous ONH changes (i.e. neuro-retinal rim narrowing, notching, and the presence of retinal nerve fiber layer defects) upon fundus examination and stereophotographs, VF abnormality with MD > –6 dB in accordance with glaucomatous ONH changes. PPG eyes with glaucomatous ONH changes but without apparent VF abnormalities on the 24-2 HFA results were also included in this study. Normal eyes had normal optic disc appearances on fundus examination and fundus stereo photographs, IOP <21 mm Hg, no abnormal VF results, and no abnormal findings on slit-lamp examination and fundus examination. 

VF results were diagnosed as abnormal if one of the following criteria was met: (1) the pattern deviation probability plot showed a cluster of 3 or more points with a probability of less than 5% and at least 1 point with a probability less than 1% in an expected location, (2) the pattern standard deviation had a probability of less than 5%, or (3) the glaucoma hemifield test indicated that the field is out of normal limits, according to the criteria described by Anderson and Patella.¹⁵ 

Eyes with a history of glaucoma, corneal and vitreous surgery, corneal opacity, clinically significant cataract, retinal disease, and non-glaucomatous optic neuropathy were excluded from this study. Eyes with pathological myopia or its suspects (diffuse chorioretinal atrophy, patchy chorioretinal atrophy, lacquer cracks, myopic choroidal neovascularization, and macular atrophy) were excluded to avoid confoundment between glaucomatous changes and pathological myopic changes.¹⁶ When both eyes of one subject were eligible for the study, one eye was randomly selected. 

Optical Coherence Tomography and Angiography Imaging

OCT and OCTA imaging were performed using the Cirrus HD-6000 with AngioPlex OCTA (Carl Zeiss Meditec) with a scan speed of 100,000 A-scans per second and eye tracking technology. All subjects underwent 4.5 × 4.5 mm ONH angiography scans after pupillary dilation. 

CpRNFLT parameters were calculated from B-scans obtained by the ONH angiography scans and not from separate ONH cube scans to ensure exact alignment of measurement areas between the two parameters. CpVD was calculated as the density of the superficial capillary plexus between the inner limiting membrane and RNFL inside a 4.5 mm diameter Bruch's membrane opening (BMO)-centered circle excluding the optic disc area (2 mm diameter BMO-centered circle). CpVD is defined as the total area of perfused vasculature per unit area in the region of measurement and large vessels were excluded from the analysis. Magnification errors for cpVD measurement areas and cpRNFLT measurement circles were corrected using a modified Littmann's formula¹⁷ provided by the manufacturer and corrected cpVD/cpRNFLT values were calculated from exported raw OCT data upon magnification-error corrected measurement areas/circles. All OCT/OCTA scans were checked by two experienced examiners (authors K.A. and H.S.) and cases with poor quality OCT/OCTA scans due to (1) vitreous opacity, (2) motion artifacts and blinks, (3) defocusing, (4) segmentation errors (erroneous identification of the borders of RNFL or incomplete segmentation by the automated algorithm and PPA associated artifacts) in OCT/OCTA measurement areas,¹⁸ (5) missing data due to excessive disc tilt, and 6) displacement of the measurement center were excluded.¹⁹ 

Three consecutive ONH angiography scans were taken on the same visit from 15 HM eyes and 15 eyes without HM to assess intra-visit reproducibility. 

Mapping of Structure Parameters to Visual Field

Global and 6 Garway-Heath sectoral²⁰ cpVD/cpRNFLT values calculated from the OCT raw data were used to assess structure function relationships of cpVD/cpRNFLT parameters with their corresponding 24-2 mean VF sensitivity (VFS). 

The superior and inferior quadrants of cpVD/cpRNFLT were corresponded to the inferior and superior 10-2 mean VFS, respectively. In addition, the inferior temporal and temporal Garway-Heath disc sectors were corresponded to the mean VFS of the superior and inferior 10-2 sectors according to a mapping advocated by Hood et al.²¹ (Fig. 1). 

Figure 1.

 

The structure function mapping of optic disc (A) and 10-2 visual field (B) according to the report by Hood et al. Sixteen testing points of the superior visual field correspond to the inferior temporal Garway-Heath sector and 29 testing points of the inferior visual field to the temporal sector.

Statistical Analysis

All data are reported as the mean ± standard deviation unless otherwise specified. Background parameters between eyes with and without HM were compared with unpaired t-tests. Analysis of covariance (ANCOVA) adjusting for difference in age and MD of 24-2 were performed to cpVD/cpRNFLT parameters between eyes with and without HM. Categorical variables were compared using chi-square tests. Multiple linear regression analysis was used to examine the relationship of age, AL, corneal curvature radius and MD of 24-2 with cpVD/cpRNFLT. Bonferroni's correction by the number of included explanatory variables were conducted to account for multiple comparisons. Intra-visit reproducibility was evaluated by intra-class correlation coefficients. Structure function relationships of the global and sectoral cpVD/cpRNFLT and mean VFS of the corresponding regions (cpVD-VFS and cpRNFLT-VFS, respectively) were determined using Pearson's correlation analysis. Steiger's test was used to evaluate differences in correlation between cpVD-VFS and cpRNFLT-VFS relationships. Statistical analyses were performed with commercially available software (SPSS version 27.0; SPSS, Inc., Chicago, IL, USA). 

Results

Two hundred ninety-eight eyes of 215 glaucoma subjects and 36 eyes of 18 normal subjects met the inclusion criteria. After excluding 47 eyes (14 eyes for history of glaucoma surgery or retinal disease, 4 eyes for pathological myopia or its suspects, and 34 eyes for insufficient quality and/or segmentation errors of OCT/OCTA scans) and randomly selecting one eye from subjects whose both eyes were eligible, 204 eyes of 204 subjects (82 HM eyes and 122 eyes without HM) were enrolled in the current study. Table 1 presents background data of the study subjects. There was no significant difference in age, sex, visual acuity, IOP, CCT, 24-2/10-2 VF MD, and pattern standard deviation between the 2 groups. Corneal curvature radius was significantly larger in HM eyes, as previously reported.²² Intra-visit reproducibilities of the angiography scans were 0.908 and 0.964 for global cpVD of eyes with and without HM, respectively. 

Table 1.

 

Demographic and Clinical Characteristics of Early Glaucomatous and Normal Eyes With and Without High Myopia

 

Global and Garway-Heath sectoral values of cpVD and cpRNFLT are presented in Table 2. There was no significant difference in cpVD between eyes with and without HM in all sectors except for the nasal sector. Global, nasal, and temporal cpRNFLT were significantly thicker in HM eyes. There was no significant association between cpVD and AL except for in the nasal sector, whereas global, nasal, and temporal cpRNFLT demonstrated significant positive association with AL in multiple regression analysis after correction for age, AL, corneal curvature radius, and MD of 24-2 (Table 3). 

Table 2.

 

Global and Garway-Heath Sectoral cpVD and cpRNFLT Measurements of Early Glaucomatous and Normal Eyes With and Without High Myopia

 

Table 3.

 

Relationship of Age, Axial Length, Corneal Curvature, and Mean Deviation of 24-2 With Global and Garway-Heath Sectoral cpVD and cpRNFLT

 

Table 4 presents the structure function relationships of cpVD/cpRNFLT with its corresponding 24-2 VFS. In HM eyes, sectoral cpVD-VFS correlation was significantly stronger than that of cpRNFLT-VFS in the nasal region (P = 0.002), whereas global and other sectors demonstrated similar significant structure function relationships between cpVD-VFS and cpRNFLT-VFS (Fig. 2). In eyes without HM, both cpVD-VFS and cpRNFLT-VFS structure-function relationships were significant in all sectors except for the superior nasal and nasal sector with a significantly stronger association between cpVD-VFS in the inferior temporal region (P = 0.042; Fig. 3). Table 5 presents structure function relationships of cpVD/cpRNFLT with its corresponding 10-2 VFS. All sectoral parameters demonstrated significant cpVD-VFS and cpRNFLT-VFS structure-function relationships in both eyes with and without HM. In the temporal sector of eyes without HM, cpVD-VFS demonstrated a significantly stronger structure-function relationship in comparison to cpRNFLT-VFS (P = 0.008). 

Table 4.

 

Global and Garway-Heath Sectoral Structure-Function Relationship of cpVD and cpRNFLT With 24-2 Visual Field Sensitivity in Early Glaucomatous and Normal Eyes With and Without High Myopia

 

Figure 2.

 

Scatter plots of global and Garway-Heath sectoral circumpapillary vessel density (cpVD), circumpapillary retinal nerve fiber layer thickness (cpRNFLT), and corresponding visual field sensitivity (VFS) in eyes with high myopia. Only cpVD demonstrated significant correlation with VFS in the nasal sector (R = 0.358, P < 0.001).

Figure 3.

 

Scatter plots of global and Garway-Heath sectoral circumpapillary vessel density (cpVD), circumpapillary retinal nerve fiber layer thickness (cpRNFLT) and corresponding visual field sensitivity (VFS) in eyes without high myopia. CpVD demonstrated significantly stronger correlation with VFS in comparison with cpRNFLT in the inferior temporal sector (P = 0.042).

Table 5.

 

Global and Sectoral Structure-Function Relationship of cpVD and cpRNFLT With 10-2 Visual Field Sensitivity in Early Glaucomatous Eyes With and Without High Myopia

 

Discussions

OCT technology is widely used for objective detection of glaucomatous damage. However, the use of conventional parameters, such as cpRNFLT, in HM eyes has been challenged by several issues and OCTA is drawing attention as a novel evaluation method that is less affected by myopic ONH changes. Previous studies have reported that mechanical stretching of the retina due to AL elongation causes temporal shifting of the superior and inferior cpRNFL bundles. This RNFL distribution shift, which results in thicker cpRNFLT in the temporal sector,²³ is often a cause for false-positive diagnosis in the cpRNFLT deviation maps generally based on normative databases that do not include HM eyes. Large vessels, which are also susceptible to this myopic shift, can be excluded in the analysis of cpVD and thus creates an advantage for OCTA in assessing HM eyes. However, the characteristics and usefulness of cpVD in glaucomatous HM eyes, especially in the earlier stages, is yet to be explored. 

Furthermore, cpRNFLT in HM eyes need to be handled with caution because previous studies have reported the underestimation of cpRNFLT measurements in eyes with long AL due to magnification errors.^24,25 In addition, the commonly used magnification correction method of simply multiplying magnification factors on cpRNFLT values measured on magnification uncorrected measurement circles assume a linear relationship between the measurement circle diameter and cpRNFLT values.²⁶ However, this assumption is not necessarily true and the distribution of cpRNFLT differs among sectors.²⁶ Although the impact magnification error will be diluted with cpVD values due to the density nature of the parameter, macular VD is known to be affected by magnification error,²⁷ and similar correction should also be applied on cpVD values. Therefore, in our study, both cpRNFLT and cpVD values were calculated from exported raw OCT data upon a magnification-error corrected measurement circle/area, which is considered to be the most appropriate among the currently available correction methods. 

Wang et al. and Yaprak et al. reported a small but significant decrease (2-5%) in global and sectoral cpVD in non-glaucomatous HM eyes suggesting that AL-related retinal stretching may have an impact on the density of superficial retinal vessels.^9,28 On the other hand, Shin et al. examined moderate to advanced glaucomatous eyes and reported no significant difference in cpVD between eyes with and without HM. They speculated that glaucomatous vascular dropouts were overwhelmingly larger than the decrease of cpVD caused by myopic retinal stretching in eyes with substantial glaucomatous damage.¹⁰ In our study of early glaucomatous eyes, cpVD demonstrated no significant difference between eyes with and without HM and no significant correlations with AL in all sectors except for a slight difference in the nasal sector. Our results indicate that the effect of myopic stretching on cpVD was subtle in the presence of glaucomatous change, even at the earliest stage of glaucoma, whereas cpRNFLT demonstrated significant increases with longer AL in global, nasal, and temporal values. 

Some previous studies reported the structure function relationship of cpVD with 24-2 VF in moderate to advanced glaucomatous eyes with HM to find promising results of the use of OCTA in HM eyes. Shin et al. reported that cpVD-VFS was stronger than cpRNFLT-VFS in global, superior nasal, nasal, and temporal sectors.¹⁰ Lee et al. focused on segmentation errors and reported that correlation was comparable between cpVD-VFS and cpRNFLT-VFS when HM eyes with segmentation errors were excluded.¹¹ Because the relationship of OCT/OCTA parameters with VF varies with disease severity,²⁹ we focused on early glaucomatous HM eyes without segmentation errors and cpVD and cpRNFL were both calculated from magnification corrected measurement areas upon one ONH angiography scan enabling a precisely registered measurement area between cpVD and cpRNFLT. Our results demonstrated that cpVD showed significant relationship with its corresponding 24-2 and 10-2 VFS in all global and sectoral areas of HM eyes, indicating its usefulness in evaluating glaucomatous damage even in the earliest stage of glaucoma. Furthermore, varying structure function relationship characteristics were observed between sectoral cpVD-VFS and cpRNFLT-VFS depending on the presence of HM. CpVD-VFS association was significantly stronger in comparison to cpRNFLT-VFS association in the nasal sector in HM eyes, whereas stronger cpVD-VFS association was observed in the temporal and inferior temporal sector of eyes without HM. 

One possible reason for the difference in nasal cpVD-VFS and cpRNFLT-VFS in HM eyes could be the effect of blind spot enlargements often observed in HM eyes. Nasal VFD is not common in early glaucoma, but blind spot enlargements are more likely to occur due to PPA and disc tilt in HM eyes,³⁰ which may have resulted in reduced sensitivity of the 4 testing points corresponding to the nasal sector. Sakaguchi et al. also reported the weak but significant nasal cpVD-VFS association in moderate glaucomatous eyes with mild myopia and reduced cpVD in the nasal and temporal sector of eyes with PPA.³¹ However, nasal cpRNFLT did not show significant structure function relationship both in eyes with and without HM. Nasal cpRNFLT of eyes without HM has been reported to demonstrate lower diagnostic ability than other optic disc sectors³² due to the lesser relative loss of cpRNFLT in the nasal region compared to that of superior and inferior regions. The nasal cpRNFLT values of HM eyes becomes even more unreliable due to myopic optic disc tilt and temporal shifting of the retina.³³ Our results showed thicker nasal cpRNFLT in HM eyes which was consistent with the report by Kang et al.²⁵ However, some studies report that nasal cpRNFLT becomes thinner with AL elongation³⁴ and others report that nasal cpRNFLT does not change with AL³⁵ suggesting that there is no consensus on the AL-related change of nasal cpRNFLT to date. Our results were able to demonstrate the stability of cpVD in assessing the nasal structure of early glaucomatous HM eyes. 

In eyes without HM, cpVD-VFS correlation was significantly stronger than cpRNFLT-VFS in the inferior temporal sector with its corresponding 24-2 VFS and the temporal sector with its corresponding 10-2 VFS. These sectors are regions known to be directly related to central VFD.²¹ Park et al. reported that patients with normal tension glaucoma (NTG) with autonomic dysfunction and abnormal peripheral microcirculation presented central VFD and suggested the involvement of vascular factors.³⁶ Our results also suggested greater vascular involvement in the central VFD, indicating the importance of evaluating cpVD for early detection of central VFD. We predicted similar or larger association of cpVD-VFS in these central VFD related regions of HM eyes. However, although cpVD-VFS correlation was stronger than cpRNFLT-VFS in the temporal and inferior temporal sectors of the HM eyes, statistically significant difference was not reached. CpVD-VFS correlation in these regions of HM eyes may have been confounded by other myopic structural changes, such as PPA and optic disc tilt, that also affect central 10-2 VFS and temporal cpVD^31,37 and further investigation is warranted on the elucidation of the role of microcirculation on central VFD in HM eyes. 

There are some limitations to our study. First, both NTG and high tension glaucoma (HTG) were included in this study. Varying results have been reported on the differences of OCTA parameters between NTG and HTG.³⁸ However, the majority of the eyes included in this study were NTG eyes (approximately 80%) and the impact is not projected to be large. Second, effect of disc tilt and PPA, which may affect cpVD/cpRNFLT parameters, was not considered. Future studies are needed to explore the effect of disc tilt and PPA on cpVD/cpRNFLT and VF. Third, mapping of ONH and 10-2 VF has not been established. Although some studies have attempted to reveal the relationship between 10-2 VFD and the ONH,^21,39 there is still no consensus on methods of mapping 10-2 VFS with specific regions of the ONH. Although our results showed significant correlations in all considered sectoral analysis, future research on 10-2 mapping is desired for more precise evaluation of structure function relationship using 10-2 VF. Finally, the use of antiglaucoma eye drops was not an exclusion criterion for our study. Some topical ocular hypotensive medications are reported to affect retinal blood flow^40–42 and may have had some effect in our study results. 

In conclusion, OCTA measured cpVD was less affected by AL and demonstrated significant structure function relationship in all 24-2/10-2 VF sectors in our cohort of early glaucomatous HM eyes. CpVD-VFS association was significantly stronger than cpRNFLT-VFS in the nasal sector of HM eyes and temporal sectors of eyes without HM revealing an HM-associated advantage in cpVD-VFS characteristics. These discoveries suggest the potential of the use of OCTA in early glaucomatous HM eyes. 

Acknowledgments

Disclosure: K. Akiyama, None; H. Saito, Carl Zeiss, Meditec (R); S. Shirato, None; A. Iwase, Carl Zeiss, Meditec (F, R); S. Aoki, None; K. Sugimoto, None; T. Fujishiro, None; H. Murata, None; R. Sakata, None; M. Honjo, None; M. Aihara, Carl Zeiss, Meditec (R)