Optic Nerve Morphology Influences Structure-Function Relationship in Early Glaucoma With and Without High Myopia

Abstract

Purpose: The purpose of this study was to determine the influence of optic nerve head (ONH) rotation and tilt on the structure-function (S-F) relationship differences between early open-angle glaucoma (OAG) with and without high myopia (HM).

Methods: A total of 164 eyes, including 69 early highly myopic glaucoma (HMG) eyes, 60 early OAG eyes, and 35 healthy HM eyes, were included. All the eyes underwent spectral-domain optical coherence tomography, standard automated perimetry, and fundus photography simultaneously. S-F relationships were analyzed by comparing the peripapillary retinal nerve fiber layer thickness (pRNFLT) to the corresponding visual field sensitivity loss (VFSL) according to the Garway-Heath map. Linear regression analyses were used to evaluate the relationship between ONH rotation and tilt with pRNFLT and VFSL.

Results: Multivariate regression analysis demonstrated that the ovality index was associated with nasal (P = 0.014) pRNFLT in OAG eyes and with nasal (P < 0.001), temporal (P = 0.012), and superior nasal (P = 0.015) pRNFLT in healthy HM eyes, but not in HMG eyes. Moreover, ONH rotation was significantly associated with inferior nasal (P = 0.013) pRNFLT in HMG eyes and with inferior temporal (P = 0.029) and inferior nasal (P = 0.036) pRNFLT in healthy HM eyes, whereas no associations were observed in OAG eyes. The strongest relationship between pRNFLT and VFSL for HMG and OAG were found in the inferior temporal (R = 0.567) and superior temporal sector (R = 0.638), respectively. However, no significant relationships were observed in healthy HM eyes (all P > 0.05).

Conclusions: Optic disc rotation and tilt may affect the distribution of the pRNFL in HM eyes with or without glaucoma, resulting in abnormal correspondence between structure and function. The analysis of the S-F relationship in the inferior temporal sector could be a valuable factor in assessing HMG eyes.

Glaucoma is a progressive and irreversible optic neuropathy that leads to structural damage, such as morphological alterations in the optic nerve head (ONH), thinning of the retinal nerve fiber layer (RNFL), and functional impairment in visual field (VF) sensitivity.¹ A comprehensive analysis is essential for clinicians in the management of glaucoma progression, which enables effective assessment of both structural damage and functional impairment.^2–4 

Analysis of the structure-function (S-F) relationship has been widely used to improve glaucoma diagnosis and monitor glaucoma progression by evaluating the strength of the correlation.^3,5 Structural parameters, such as the RNFL thickness measured using optical coherence tomography (OCT) and vascular density assessed using OCT angiography (OCTA), are key components for analyzing the S-F relationship.^4,6 A moderate to strong correlation has been previously reported between structural damages, including Bruch's membrane opening-minimum rim width (BMO-MRW), peripapillary RNFL thickness (pRNFLT) and peripapillary vascular density (pVD), and VF defects in open-angle glaucoma (OAG) based on the Garway-Heath map.^7–11 Furthermore, the relationship varies across different types of glaucoma. In moderate glaucomatous eyes, the relationship between S-F is comparable to vasculature-function (VF), which is weaker than VF in early glaucoma.^7,10,12 Overall, previous research suggests that the primary structural changes may differ during glaucoma progression. However, only a few studies have investigated the reasons for this variability. 

Individual variations have been reported in the strength of correlation using the Garway-Heath map.¹³ Variations in ocular parameters, such as the ONH position and area, ONH morphology, and axial length (AL), influence the S-F relationship. Myopia axial elongation is closely associated with various changes in ONH morphology, such as optic disc rotation and tilt, and can lead to structural and functional defects that cannot be easily distinguished.^6,14 The influence of ONH morphology and refractive error on the distribution of pRNFLT and VF defects in myopia has been previously reported.^15–20 However, the reasons for the differences in the correspondence between structural damage and functional damage across various stages of glaucoma are not clear, especially in the early stages. Therefore, exploring the influence of myopia-related ONH changes on the S-F relationship is crucial. Furthermore, distinguishing between high-myopia (HM) eyes with or without glaucoma changes remains challenging for clinicians,⁶ which could be attributed to RNFL displacement or defects caused by axial elongation in HM eyes.^14,15 This study aimed to characterize optic disc rotation and tilt in eyes with early OAG and early highly myopic glaucoma (HMG), to evaluate their impact on the S-F relationship, and to determine the influence of myopia-related ONH morphology on glaucoma diagnosis. 

Methods

Participants

This retrospective, cross-sectional study was conducted at the Zhongshan Ophthalmic Center (ZOC), Sun Yat-Sen University, Guangzhou, China. The participants were enrolled and examined between January 2020 and December 2023. This study was approved by the Ethics Review Committee of the ZOC (approval number: 2023KYPJ105) and adhered to the principles outlined in the Declaration of Helsinki. Informed consent was obtained from all participants before enrollment. 

All participants underwent complete ophthalmologic examinations, which were performed by authors Yangjiani Li and Jinpeng Yang. These included assessments of refraction and best-corrected visual acuity (BCVA), intraocular pressure (IOP) determined by Goldmann applanation tonometry, AL (IOL Master 700; Carl Zeiss Meditec, Dublin, CA, USA), fundus photography (Clarus 500; Carl Zeiss Meditec, Jena, Germany), standard automated perimetry (Humphrey Field Analyzer 3; Carl Zeiss Meditec), and spectral-domain optical coherence tomography (SD-OCT; Spectralis, Heidelberg Engineering, Heidelberg, Germany). In addition, slit-lamp biomicroscopy and gonioscopy were performed. 

The inclusion criteria were as follows: participants aged ≥ 18 years, BCVA of 20/40 or better, diagnosis of early glaucoma (mean deviation [MD] ≥ –6 decibel [dB]), availability of good quality OCT scanning (signal quality ≥ 15), and reliable VF test results (< 20% fixation loss, < 15% false negative errors, and < 15% false positive errors). All patients with glaucoma had open angle on gonioscopy, characteristic glaucomatous optic neuropathy (GON) and glaucomatous visual field (VF) defect. GON was defined as cup to disc (C/D) ratio ≥ 0.7, C/D ratio asymmetry > 0.2, or focal neuroretinal rim thinning, especially at the superior temporal (ST) or inferior temporal (IT), etc. Corresponding Glaucomatous VF defect was defined as reliable early glaucomatous VF abnormality (MD ≥ –6 dB) without other ocular diseases that might cause VF abnormalities. Eyes with glaucoma that met the eligibility criteria were divided into two groups according to the AL: HMG (AL ≥ 26.0 mm) and OAG (AL < 26.0 mm).^1,21 

The exclusion criteria were as follows: history of intraocular surgery (e.g. cataract surgery and refractive surgery), secondary glaucoma, non-glaucomatous optic neuropathy, vascular or non-vascular retinopathies, and other ocular or systemic diseases (e.g. retinal detachment and Alzheimer's disease) that can impair the VF. Additionally, eyes with pathological myopia were excluded.²¹ 

Optical Coherence Tomography

SD-OCT imaging was performed using the Spectralis OCT (version 6.3.2) imaging platform. All participants had peripapillary retinal nerve fiber layer thickness (pRNFLT) measured using the circular scan protocol, which was composed of a 360 degree, 3.5-mm diameter peripapillary circle scan centered on the ONH. Only scans with clear fundus images and uninterrupted visibility of the pRNFLT were included in the analysis. 

The Spectralis OCT system divides the pRNFLT into the following six sectors: nasal (N), temporal (T), IT, inferior nasal (IN), ST, and superior nasal (SN). These measurements were automatically calculated by the Spectralis OCT software, as previously described elsewhere.¹² The ocular magnification effects of the AL can affect the true circle size, so the modified Littmann's by Bennett et al. was used for correcting pRNFLT.^22–24 Both global and sector-specific corrected pRNFLT values were analyzed. The procedures were performed by an experienced ophthalmologist (author S.Z.). 

These six sectors were used to assess the S-F relationship of pRNFLT parameters with their corresponding visual field sensitivity loss (VFSL) using the Garway-Heath map (Fig. 1)²⁵ and the Kanamori map.²⁶ 

Figure 1.

 

Heat maps depicted variations of S-F relationship in the Pearson correlation coefficients (r). Heat map of Garway-Heath sectoral peripapillary retinal nerve fiber layer thickness (pRNFLT) and corresponding VFSL in open angle glaucoma (OAG) eyes. The T, IT, ST, and SN sectors demonstrated significant association (A). Heat map of Garway-Heath sectoral pRNFLT and corresponding VFSL in highly myopic glaucoma (HMG) eyes. The T, IT, and ST sectors demonstrated significant association (B). Heat map of Garway-Heath sectoral pRNFLT and corresponding VFSL in healthy high myopia (HM) eyes. All six sectors had no significant correlation (C). IN, inferior nasal; IT, inferior temporal; N, nasal; pRNFLT, peripapillary retinal nerve fiber layer thickness; SN, superior nasal; ST, superior temporal; T, temporal. Significant values are in bold type.

Standard Automated Perimetry

All participants underwent VF testing using the Swedish interactive threshold algorithm Fast 30-2 Fast pattern on a Humphrey Field Analyzer (Carl Zeiss Meditec, Dublin, CA, USA), and OCT measurements were completed simultaneously. To minimize the effects of age, the VF total deviation (TD) map was used to evaluate the S-F relationship.³ First, the TD map was divided into six sectors according to the correspondence map described by Garway-Heath et al.²⁵ Second, the visual sensitivity in each Garway-Heath sector was converted from the dB scale to a linear scale using the formula: 1/Lambert = (10)^0.1*dB.^3,5 Third, the average 1/Lambert values were logged and converted back to dB. Finally, the mean VFSL in these six sectors was calculated. 

ONH Measurement

Fundus photographs centered on the optic disc and macular region were reviewed using the ImageJ software (ImageJ version .1.53t; available at http://imagej.nih.gov/ij; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD, USA). Initially, fundus photography was conducted with the optic disc and macula centered, utilizing standardized settings. Subsequently, images of poor quality, which due to low lighting, haziness, or obscured features of the optic disc, were excluded. Two observers (authors Q.Z. and S.Z.) measured the ONH parameters using ImageJ software. The average values were used for further analysis in this study. 

First, the optic disc margin was manually identified, and the DrawEllipse macro was used to automatically draw the best-fit ellipse. In the ellipse, the longest and shortest axes identified by ImageJ were exported to.csv files to calculate the ovality index (OI). Next, a vertical meridian was drawn to measure the degree of rotation (Fig. 2). The details have been provided in the other section.²⁷ 

Figure 2.

 

Color fundus photography (A), visual field (VF) pattern deviation map (B) and optical coherence tomography (OCT) (C, D) from a 29-year-old female patient with highly myopic glaucoma. The orange line indicates (A) inferior rotation of the optic disc. The red region (B) shows VF defect in inferior nasal sector. The red line (C) and the red region (D) indicate the yellow color code at the superior nasal sector. Note that OCT shows a retinal nerve fiber layer (RNFL) thickness (C, D), which is not corresponding to the VF impairment (B). RNFL shifts from the superior nasal to the superior temporal (red arrow, C) due to the rotation of the optic disc, resulting in the inaccurate RNFL thickness measurement (red region, D). INF, inferior; MD, mean deviation; N, nasal; NAS, nasal; NI, nasal inferior; NS, nasal superior; PSD, pattern standard deviation; SUP, superior; T, temporal; TI, temporal inferior; TMP, temporal; TS, temporal superior.

Optic disc tilt was assessed using the OI, defined as the ratio of the longest to the shortest disc diameter.²⁸ The ONH rotation angle was defined as the deviation of the long axis of the optic disc from the vertical meridian.²⁹ The vertical meridian was considered a vertical line 90 degrees from a horizontal line connecting the fovea to the center of the optic disc. The inter-observer reproducibility of OI and optic disc rotation angle measurements were evaluated by calculating intraclass correlation coefficients (ICCs) for paired measurements. 

Statistical Analysis

Numerical variables are presented as the mean ± standard deviation (SD). Analysis of variance (ANOVA) with a post hoc Tukey-Kramer test or Kruskal–Wallis test was performed to compare the numerical variables. We conducted the Shapiro-Wilk normality test to assess deviations from normal distributions. A χ² test was used to compare frequencies of categorical variables. The S-F relationship between the pRNFLT and the corresponding mean VFSL was evaluated using Pearson's correlation analysis, and Fisher's z-transformation was used to evaluate the differences in the strength of the correlation coefficients between the OAG and HMG. Multivariate linear regression analysis was performed to investigate the association among age, AL, MD, OI, ONH rotation angle, and pRNFLT. 

All statistical analyses were performed using SPSS software (version 26.0; SPSS Inc., Chicago, IL, USA). All comparisons between correlations were conducted using the cocor package in R software.³⁰ A P value of ˂ 0.05 indicated statistical significance. 

Results

This study initially included 254 participants (254 eyes), and 90 eyes of 90 participants were excluded based on the predefined criteria because of unreliable SD-OCT, VF examinations, or fundus photographs. A total of 164 eyes comprising 69 eyes with HMG, 60 eyes with OAG, and 35 healthy eyes with HM were enrolled in the current study. The demographic characteristics of the participants are listed in Table 1. The glaucoma groups were comparable in age (45.27 ± 14.92 for patients with OAG vs. 40.81 ± 11.80 of patients with HMG, P > 0.05), VF MD (−2.93 ± 1.54 vs. –3.13 ± 1.60, P > 0.05), and VF pattern standard deviation (PSD) (3.05 ± 1.90 vs. 3.29 ± 1.75, P > 0.05). Statistically significant differences in age, AL, OI, ONH rotation, and VF PSD were observed among the three groups (P < 0.05). However, there were no significant differences in sex, eye laterality, or 30-2 VF MD (P > 0.05). The interobserver ICC for OI and optic disc rotation were 0.970 and 0.993, respectively. 

Table 1.

 

Demographics and Ocular Characteristics of Study Population

 

Table 1 presents the mean pRNFLT and VFSL values for the Garway-Heath sector. The VFSL values were comparable between the OAG and HMG, except in the IT sector. No significant differences in pRNFLT were observed between the two groups, except in the T sector. 

The S-F relationship of pRNFLT thinning with its corresponding VF defects according to the Garway-Heath map and modified Kanamori map are summarized in Table 2 and Figure 1. In healthy HM eyes, no significant relationships were observed in any of the six sectors. The global, T, IT, and ST sectors showed significantly S-F relationship between with pRNFLT and VFSL in the HMG eyes, with the strongest correlation in the IT sector. In OAG eyes, pRNFLT-VFSL relationship was significant in all sectors except for N and IN. Notably, the ST sector (R = 0.638 vs. 0.379, P = 0.025) and SN sector (R = 0.340 vs. –0.011, P = 0.022) demonstrated a significantly stronger S-F relationship in OAG eyes compared with HMG eyes (see Table 2; Supplementary Fig. S1). 

Table 2.

 

Comparison of Structure-Function Relationship Between VFSL and pRNFLT

 

To analyze the difference in the S-F relationship between the OAG and HMG, linear regression analyses were used to determine the factors associated with regional structure and function parameters. In Table 3 and Supplementary Table S1, multivariate linear regression analysis revealed that OI and ONH rotation were weakly associated with pRNFLT and VF defects in OAG eyes after adjusting for age, AL, VF MD, OI, and rotation, but a statistically significant difference was observed in HM eyes with or without glaucoma. Specifically, OI was associated with N (P = 0.014), pRNFLT and T (P = 0.030) VFSL in OAG eyes, and with N (P = 0.002) and SN (P = 0.041) VFSL in HMG eyes. In healthy HM eyes, OI was associated with N (P < 0.001), T (P = 0.012), and SN (P = 0.015) pRNFLT. Additionally, ONH rotation was significantly associated with IN (P = 0.013) pRNFLT and N (P = 0.011) VFSL in HMG eyes and with IT (P = 0.029) and IN (P = 0.036) pRNFLT and T (P = 0.033) VFSL in healthy HM eyes. However, no significant correlation was observed in OAG eyes. 

Table 3.

 

Multivariate Linear Regression Analysis to Determine the Factors Associated With Garway-Heath Sectoral pRNFLT

 

Discussion

Our findings revealed differences in the S-F relationship between OAG and HMG eyes in the early stages of glaucoma (see Fig. 1). Additionally, compared with OAG, the HMG exhibited weaker correlations, whereas the HM without glaucoma exhibited no correlation across the six regions. Furthermore, based on the Spectralis OCT pRNFLT, the Garway-Heath map was superior to the other S-F maps. Our study is unique in that it included only patients with early glaucoma and explained that the structure of the ONH influences changes in RNFL distribution, potentially contributing to the observed differences in the S-F relationship (see Fig. 2). 

In recent years, the Garway-Heath map has made substantial contributions to the field of glaucoma by exploring the structural and functional correlations in the disease, and it has been widely adopted in studies to assess the correlation between RNFL thickness and VF defects.⁵ Most previous studies have demonstrated moderate S-F relationships in glaucoma patients using the Garway-Heath map.^7–11 The Garway-Heath map aligns the VF test locations with specific pRNFL distributions within the ONH using color fundus images, outlining the corresponding S-F zones. However, Kanamori et al. proposed an alternative map by grouping the VF test points with RNFL thickness sectors measured by Stratus OCT and found a moderate S-F relationship in glaucomatous eyes.²⁶ They hypothesized that RNFL thickness measured by OCT at a distance from the ONH margin and for non-radially projected retinal nerve fibers might not accurately reflect the true S-F relationship in glaucoma. Wu et al. reported stronger correlations using the modified Kanamori map than did the Garway-Heath map and verified the corresponding relationship between RNFL changes and VF defects in each sector using the Hood model.³¹ Compared with other S-F maps in early stage glaucoma, our findings demonstrated superior correlations based on the Garway-Heath map, with higher correlation coefficients associated with the ST sector (R = 0.638; see Table 2 and Fig. 1). A possible explanation could be the projection of retinal nerve fibers approximately 1.73 mm from the optic disc center, corresponding with the 6 ONH sectors of the Garway-Heath. The Garway-Heath map may be an effective method for S-F analysis in glaucoma. 

The S-F relationship in glaucoma is an important topic for clinical and research purposes, as it offers valuable insights into the pathophysiology of the disease, aids in monitoring its progression, and evaluates the effectiveness of treatment. This S-F relationship may be exclusive to glaucoma, and the severity of glaucoma could affect sectoral correlation coefficients.⁸ Most previous studies have reported a moderate-high correlation between OCT pRNFLT and Humphrey VF test, especially in the early to moderate stages of the disease.^7–10,16 Stronger correlations are observed in the ST or IT sectors of the pRNFL.^7,9,13,29 However, studies focusing exclusively on the early stages of glaucoma are scarce. In the present S-F study, we observed the strongest correlation between pRNFLT and VFSL for HMG in the IT sector (R = 0.567), and for OAG in the ST sector (R = 0.638), respectively. Conversely, no statistically significant correlations were found in all sectors in healthy eyes with HM (see Table 2). The difference in the S-F relationship could be attributed to refractive errors. Moreover, there were few correlations between structural pRNFL thinning and functional VF defects in healthy HM eyes. These findings were consistent with those of previous studies demonstrating that this association exists only in glaucomatous eyes and not in those with refractive errors.^32,33 These findings highlight that, analyzing the S-F relationship using the Garway-Heath map may be critical for differentiating between early glaucomatous eyes and healthy HM eyes. In addition, our findings suggest that the IT and ST sectors may serve as important prognostic factors for patients with HMG because the structural and functional dysfunction in these sectors was not associated with the OI (P > 0.05) or ONH rotation (P > 0.05; see Table 3). 

The influence of ocular parameters on the mapping of RNFL has been previously documented, with promising results for the use of individualized maps in detecting glaucoma and monitoring glaucomatous progression.^{13,17–20,34} In myopia studies, Lamparter et al. identified optic disc rotation and optic disc tilt as significant factors influencing RNFL distribution.¹³ In addition, Sung et al. reported that optic disc rotation had a significant correlation with RNFL thickness.³⁵ Furthermore, some studies on glaucoma have found that the abnormal areas of the ONH parameters align with the distribution of the RNFL and correspond to the location of VF defects.¹⁵ Elze et al. and Baniasadi et al. revealed retinal areas of risk false positive and false negative diagnoses due to ONH morphology.^17,19,20 Han et al. reported that deep ONH parameters, such as optic disc tilt, are consistent with the location of RNFL defects.³⁶ Additionally, Lan et al. found that the optic disc rotation resulted in wedge-shaped RNFL defects in OAG.¹⁶ Sung et al. indicated that the direction of optic disc rotation predicts the progression of VF defects.²⁹ Interestingly, our findings highlighted that optic disc rotation and tilt were associated with the reginal pRNFLT in HM eyes with and without glaucoma using the Garway-Heath map, but not in OAG. Moreover, optic disc rotation and tilt had a minimal effect on regional pRNFLT and VF in OAG (see Table 3, Supplementary Table S1). This may explain the differences in the S-F relationship between OAG and HMG (see Fig. 2). 

Understanding the pathogenesis of glaucoma is crucial for elucidating its pathogenesis.³ A well-established relationship has been reported between structural damage and functional impairment in glaucoma.^2,3 Therefore, the structure that has a strong correlation with VF defects may contribute to glaucoma progression. Compared to OAG, retinal microvasculature damage may be an early pathogenic mechanism of HMG.³⁷ Fan et al. used OCTA to reveal a significant impairment in retinal vasoreactivity of peripapillary capillaries in early-to-moderate HMG than in glaucomatous eyes without HM.³⁸ Lee et al. reported that the decrease in peripapillary retinal microvasculature is consistent with the location of pRNFL defects in glaucomatous eyes with myopia.³⁹ Yarmohammadi et al. demonstrated that retinal vessel density is associated with VF independent of RNFL thickness.⁴⁰ In a prospective study, Rao et al. reported that lower vascular density and perfusion density can lead to pRNFL loss.⁴¹ Additionally, compared with OAG, the S-F relationship is weaker than the vasculature-function relationship in early HMG, whereas the correlations are similar in the moderate stage of disease.^7,10,12 These studies imply that changes in blood flow parameters may precede changes in cpRNFLT in HMG. However, our results revealed significant S-F relationships only in the IT and ST sectors, but not in the N sector, in HMG. These discrepant findings may indicate that vascular mechanisms may play a crucial role in early HMG. 

This study had several limitations that must be acknowledged. Our results show no correlation in several sectors, which differs from recent research findings. We speculate that this may be attributed to the inclusion of only patients with early-stage glaucoma (MD ≥ –6 dB) in our study. The severity of glaucoma influences the S-F relationship, with weaker correlations observed in early-stage glaucomatous eyes. Moreover, another definition of optic disc tilt in 3-D was the angle between reference plane and ONH plane. By definition, 3-D optic disc tilt offers a more intuitive approach to determining the angle and direction of the tilt. However, it should be noted that ONH analysis based on fundus photography is both convenient and cost-effective in primary clinical settings. In addition, the regional RNFL thickness was not adjusted for the optic disc rotation angle, limiting the accuracy. Future prospective and longitudinal studies are required to address these issues and improve S-F relationships in early HMG. Another limitation is the lack of microvasculature parameters assessed by OCTA. However, pRNFLT measured using OCT is an objective indicator commonly used for monitoring glaucoma progression. The primary objective of this study was to evaluate the impact of optic disc rotation and tilt on the correlations between pRNFLT and VF defects and to provide evidence for S-F analysis. Looking ahead, the discovery of more specific parameters could enable the use of a single indicator to diagnose glaucoma, thereby propelling the evolution of studies on the S-F relationship. 

In conclusion, the S-F relationship analysis, based on the Garway-Heath map, can be used to differentiate OAG, HMG, and HM. Optic disc rotation and tilt may influence the distribution of the pRNFL to some extent in HMG, resulting in abnormal correspondence between structure and function. The S-F relationship analysis in the IT and ST sectors is more stable and useful for managing patients with HMG. Therefore, clinical practitioners should consider the impact of optic disc rotation on pRNFL when evaluating glaucoma progression. 

Acknowledgments

The authors thank all research assistants and nursing staff who contributed to the practical organization and execution of this study. We are grateful to Hui Xiao for her valuable support and guidance throughout the entire work, especially in the methodology. 

Supported by the National Natural Science Foundation of China (82471067) and the Guangdong Basic and Applied Basic Research Foundation (2022A1515012168 and 2024A1515013296). 

Disclosure: J. Yang, None; Y. Li, None; Q. Zhang, None; S. Zeng, None; H. Huang, None; C. Wu, None; Z. Liu, None; J. Tang, None; S. Wu, None; Y. Chen, None; Y. Zhuo, None; Y. Yang, None; Y. Li, None