Intereye Difference in the Microstructure of Parapapillary Atrophy in Unilateral Primary Open-Angle Glaucoma

Abstract

Purpose: To investigate intereye differences in the parapapillary atrophy (PPA) microstructure in unilateral primary open-angle glaucoma (POAG).

Methods: Bilateral optic nerves from 172 unilateral POAG patients having clinical PPA shown in infrared fundus imaging were scanned using spectral-domain optical coherence tomography (SD-OCT). Based on the extent of Bruch's membrane (BM) within the PPA area, the PPA was divided into β-zone (area with BM) and γ-zone (area devoid of BM). The width of each type of PPA and disc tilt angle were measured in radial B-scan images. Disc tilt ratio and torsion degree were determined using the radial B-scan and infrared fundus images. Measurements were compared between the glaucomatous and contralateral healthy eyes, and the factors associated with the presence of glaucoma were determined.

Results: The width of β-zone was larger in glaucomatous eyes compared with the contralateral healthy eyes (283.67 ± 157.45 vs. 214.42 ± 145.39 μm, P < 0.001), whereas there were no significant intereye differences in the γ-zone width (P = 0.392). Eyes with POAG had a larger β-zone width (P < 0.001), and larger tilt angle (P = 0.032) compared with the contralateral healthy eyes. In multivariate regression analysis, larger β-zone width (P < 0.001), and lower IOP (P = 0.001) were revealed as significant factors associated with the presence of glaucoma.

Conclusions: The β-zone width was revealed as a significant lateralizing factor for unilateral POAG, whereas the γ-zone was not associated with the presence of glaucoma.

Glaucoma is a multifactorial disease.^1,2 Although elevated intraocular pressure (IOP) is considered the most important factor for the development and progression of glaucoma,^3–5 glaucomatous damage also occurs when the IOP is within the statistically normal range. Specifically, it is intriguing that glaucomatous damage can occur unilaterally in some patients despite a similar IOP in both eyes. 

Parapapillary atrophy (PPA) has long been suggested as one of the structural characteristics associated with glaucoma.^6,7 It has been considered a significant risk factor for both the presence of glaucoma⁸ and its progression.⁹ Recently, it has been demonstrated that the PPA could be morphologically categorized based on the location of the termination of Bruch's membrane (BM) within the PPA area.^10–12 Although the PPA with an intact BM was related to the presence of glaucoma¹⁰ and a faster rate of retinal nerve fiber layer (RNFL) thinning,⁹ the PPA devoid of BM was associated with axial elongation^9,12 but was independent of the presence of glaucoma.¹⁰ It has recently been shown that the area devoid of BM appears as the reflection of the combination of externally oblique border tissue of Elschnig and sclera, rather than being a tissue related to atrophy.¹³ Based on the results, we hypothesized that the mechanism of development differs between areas within the conventional PPA region, and thus each type of PPA may have a different significance for glaucoma pathogenesis. 

Because both eyes are under the same systemic environment, investigating the structural differences between the eyes of patients with unilateral glaucoma may provide a clue to understanding the pathogenic mechanisms of primary open-angle glaucoma (POAG). The purpose of this study was to investigate whether the structural characteristics of the PPA differ between eyes in patients with unilateral POAG. 

Methods

Study Subjects

The participants in this study included patients from the Investigating Glaucoma Progression Study (IGPS), which is an ongoing prospective study that has been under way since August 2011 at the Seoul National University Bundang Hospital Glaucoma Clinic. The study protocol was reviewed and approved by the Seoul National University Bundang Hospital (SNUBH) Institutional Review Board. This study was carried out in accordance with the recommendations of the Declaration of Helsinki for biomedical research involving human subjects. 

The database of patients included in the IGPS between January 2012 and May 2015 was reviewed. The subjects underwent a complete ophthalmic examination, including a visual acuity assessment, refraction, slit-lamp biomicroscopy, gonioscopy, Goldmann applanation tonometry, and dilated stereoscopic examination of the optic disc. They also underwent measurements of central corneal thickness (Orbscan II; Bausch & Lomb Surgical, Rochester, NY, USA) and axial length (AXL) (IOL Master version 5; Carl Zeiss Meditec, Dublin, CA, USA), corneal curvature (KR-1800; Topcon, Tokyo, Japan), stereo disc photography (EOS D60 digital camera; Canon, Utsunomiya-shi, Tochigi-ken, Japan), scanning of the optic disc and circumpapillary RNFL thickness measurements using spectral-domain optical coherence tomography (SD-OCT) (Spectralis OCT, Heidelberg Engineering, Heidelberg, Germany), and standard automated perimetry (24-2 Swedish interactive thresholding algorithm and Humphrey Field Analyzer II 750; Carl Zeiss Meditec). 

The subjects included in the IPGS were required to have a best-corrected visual acuity of at least 20/40, spherical refraction of −6.0 to +3.0 diopters (D), and a cylinder correction of −3.0 to +3.0 D. Those with a history of ocular surgery other than cataract or glaucoma surgery, of intraocular diseases (e.g., diabetic retinopathy or retinal vessel occlusion), or of neurologic diseases (e.g., pituitary tumor) that could cause visual field deficits were excluded. 

To be included in the present study, patients were required to be diagnosed with unilateral POAG and to have a clinically apparent PPA in at least one eye that was visible on an infrared fundus image with a width of ≥200 μm on at least one radial B-scan image as measured by the built-in caliper tool of the Spectralis OCT.¹² Eyes with poor image quality in which the clinical PPA was not delineated clearly on SD-OCT images were excluded. Subjects with a large PPA that exceeded the area covered by each radial scan were also excluded from this study. 

Primary open-angle glaucoma was diagnosed according to the following criteria: an open angle on gonioscopy, glaucomatous optic nerve damage (e.g., the presence of focal thinning or notching of the neuroretinal rim, or diffuse thinning of the neuroretinal rim), and associated visual field defect without ocular disease or conditions that may cause visual field abnormalities. A glaucomatous visual field change was defined as being present when two or more of the following criteria were fulfilled: (1) outside the normal limits on the Glaucoma Hemifield Test; (2) three abnormal points with a probability of being normal of P < 5%, and one with P < 1% by pattern deviation; or (3) a pattern standard deviation of P < 5%. Those visual field defects were confirmed on two consecutive reliable tests (fixation loss rate ≤20%, false-positive and false-negative error rates ≤25%). The contralateral eye had to have an open angle on gonioscopy, a normal-appearing optic disc, and a normal visual field. 

To exclude the influence of IOP, only patients with a diurnal IOP of consistently ≤21 mm Hg in both eyes were included in the study. Diurnal measurements of IOP were obtained every 2 hours from 9:00 AM to 5:00 PM, and the average of the five measurements was defined as the untreated IOP. The IOP at examination was defined as the IOP at the time of the SD-OCT imaging. 

Spectral-Domain OCT Imaging of the Parapapillary Area

The optic nerve head, including the PPA, was imaged by the Spectralis OCT using the enhanced depth imaging (EDI) technique as reported by Spaide and Jonas.¹⁴ The imaging was performed using a 48-radial scan centered on the optic disc. The scan angle spanned 20 degrees, and the distance between each scan was 3.75 degrees. A potential magnification error was removed by entering the corneal curvature of each eye into the Spectralis OCT system before performing SD-OCT scanning. Only eyes having acceptable scans with a good-quality image (i.e., quality score ≥ 15) obtained for all of the 48 scans were included for analysis. 

Measurement of the Width of the PPA

The PPA width was measured on the EDI SD-OCT images, as described previously.¹² The conventional PPA was divided into the β-zone (the area with intact BM) and γ-zone (the area devoid of BM) according to the presence/absence of BM in the conventional PPA area. Based on the recent notion that the γ-zone reflects the combination of the externally oblique border tissue of Elschnig and sclera, rather than being a tissue-related to atrophy,¹³ the terms β-zone and γ-zone are used here rather than PPA_+BM and PPA_−BM, respectively, as was used in our previous studies.^12,15 The distance from the termination of BM to the clinical PPA margin and that from the disc margin to the boundary of the termination of BM were defined as the β-zone and γ-zone widths, respectively.^9,12 Measurements were obtained at the 12 clock-hour meridians, using 6 of the 48 radial scan images, and the means of the measured values were determined as the β-zone width and the γ-zone width of the eye (Fig. 1). The maximum values among the 12 measurements were defined as the maximum β-zone width and maximum γ-zone width. Measurements were performed twice using the built-in caliper tool of the Spectralis OCT system on the display window of the Spectralis viewer (Heidelberg Eye Explorer software version 1.7.0.0; Heidelberg Engineering) by two experienced ophthalmologists (YJY and EJL) who were masked to the subjects' clinical information, and the mean of the values from the two observers was used for data analysis. Disagreements were resolved by consensus between the two ophthalmologists or a third adjudicator (T-WK). Excellent interobserver reproducibility for the measurement of the β-zone and γ-zone widths has been reported by our previous study (intraclass correlation coefficients = 0.962 and 0.977, respectively).¹² 

Figure 1

 

Measurement of the PPA width and optic disc tilt angle on the EDI SD-OCT images. (A) Color-converted infrared fundus image of a glaucomatous eye. Green solid lines indicate the location where the six radial B-scans were selected. Measurements were performed at the 12 clock-hour meridians in the peripapillary area. (B) B-scan image obtained along the 3 to 9 o'clock meridian. The PPA was divided into β-zone (area with BM) and γ-zone (area devoid of BM). The distance from the termination point of BM (red arrowhead) to the PPA margin (white arrowhead) and that from the optic disc margin (black arrowhead) to the boundary of the BM termination (red arrowhead) were defined as the β-zone and γ-zone widths, respectively. The optic disc tilt angle was measured in the radial B-scan image obtained along the shortest optic disc diameter. The angle between the two lines (white solid lines), one of which connected the nasal (open red arrowhead) and temporal points of BM termination (closed red arrowhead) and the other, which connected the nasal BM termination point (open red arrowhead) and the temporal optic disc margin (black arrowhead), was defined as the optic disc tilt angle (α). The measurements were performed at 12 meridians in the six radial B-scan images, and the average of the 12 values were considered for analysis.

Determination of the Optic Disc Tilt Angle, Tilt Ratio, and Torsion Degree

On the basis of previous reports that showed that both the γ-zone and the optic disc tilt angle were associated with myopic axial elongation,^12,15,16 and that the optic disc torsion degree was larger in the visual field–affected eyes of glaucoma patients with unilateral visual field defects,¹⁷ the optic disc tilt angle, tilt ratio, and torsion degree were also determined and compared between the glaucomatous eyes and contralateral healthy eyes. 

The optic disc tilt angle was measured along the shortest optic disc diameter using one of the radial B-scan images, following the method described by Hosseini et al.¹⁶ In brief, the radial B-scan image passing through the short axis of the optic disc was selected from the 48 radial images. The angle between the two lines, one of which connected the nasal and temporal points of Bruch's membrane opening (BMO) and the other, which connected the nasal BMO point and the temporal optic disc margin, was defined as the optic disc tilt angle (Fig. 1).¹⁶ 

The optic disc tilt ratio and torsion degree were determined based on the BMO defined using radial B-scan images, and measured on infrared fundus images.¹⁸ The optic disc tilt ratio was determined by calculating the ovality index, which was defined as the ratio of the longest and shortest diameter of the BMO.¹⁹ The optic disc torsion degree was defined as the angular deviation of the long axis of the BMO from the vertical line, which is perpendicular to a horizontal line connecting the fovea and the center of the BMO, based on the method described by Chauhan et al.¹⁸ Positive and negative values indicated supranasal rotation and inferotemporal rotation of the torsion degree, respectively.^18,20 

Statistical Analysis

The comparison between groups was performed using a χ² test for categorical variables and a paired t-test for continuous variables. Factors associated with the presence of glaucoma were assessed using logistic regression analysis. Variables with statistical significance of P < 0.20 in the univariate analysis were included in the multivariate model. Analyses were performed using SPSS software version 22.0 (SPSS, Inc., Chicago, IL, USA). The data are presented as mean ± SD values unless stated otherwise, and the cutoff for statistical significance was set at P < 0.05. 

Results

Baseline Characteristics

This study initially involved 233 patients diagnosed with unilateral POAG, of which 61 were excluded because the PPA was not visible with infrared fundus imaging for either eye (n = 48) or poor image quality did not allow clear visualization of the PPA (n = 13). The remaining 172 patients included 93 men (54.1%) and 79 women (45.9%). The average patient age was 53.8 ± 14.6 years. Table 1 compares the clinical characteristics between glaucomatous eyes and the contralateral healthy eyes. There were no intereye differences in terms of the refractive error, AXL, central corneal thickness, and untreated IOP (all P > 0.05). Glaucomatous eyes had a lower IOP at SD-OCT examination (P < 0.001) and a worse visual field mean deviation (MD) and pattern standard deviation (PSD) (all P < 0.001). Both the mean and maximum widths of the β-zone were larger in glaucomatous eyes (P < 0.001) than in healthy eyes, whereas there was no significant intereye difference in the mean and maximum γ-zone widths (P = 0.392 and 0.147, respectively). The optic disc tilt angle was larger in glaucomatous eyes than in contralateral healthy eyes (P = 0.032). However, the optic disc tilt ratio and torsion degree, which were based on the BMO, did not show a significant intereye difference (P = 0.078, 0.145, and 0.083 for the tilt ratio, torsion degree, and absolute torsion degree, respectively) (Table 1). 

Table 1

 

Comparison of Clinical Characteristics of Unilateral POAG Patients Between Glaucomatous Eyes and Contralateral Healthy Eyes

 

Factors Associated With the Presence of Glaucoma

Table 2 gives the results of the logistic regression analysis to assess the factors associated with the presence of glaucoma. In the univariate analysis, a larger β-zone width (P < 0.001), and a lower IOP at OCT examination (P = 0.004), were significantly associated with the presence of glaucoma. Multivariate analysis also revealed β-zone width (P < 0.001) and IOP at examination (P = 0.001) as significant factors. 

Table 2

 

Factors Associated With the Presence of Glaucoma in Unilateral POAG Patients

 

Factors Associated With the Presence of Glaucoma in Myopic and Nonmyopic Subgroups

Because the analyses of factors associated with glaucoma in overall subjects revealed a significant influence of only β-zone width but not γ-zone width, which is associated with myopic axial elongation,^10,11,21 the analyses were performed again for myopic and nonmyopic subgroups. Patients with a spherical refraction of less than −2.00 D or AXL of ≥24.0 mm in the glaucomatous eyes were classified as the myopic subgroup,¹⁷ and the others were classified as the nonmyopic subgroup. Table 3 compares the clinical characteristics of the glaucomatous eyes and the contralateral healthy eyes in the myopic and nonmyopic subgroups. In both groups, the mean and maximum widths of the β-zone were larger in the glaucomatous eyes than in the contralateral healthy eyes (all P < 0.001). In the myopic subgroup, there was no intereye difference for either spherical error (P = 0.248) or AXL (P = 0.475), whereas the glaucomatous eyes had more tilted optic discs with a larger tilt angle (P = 0.003) and a larger tilt ratio (P = 0.023). Meanwhile, none of the optic disc tilt-related parameters showed significant intereye differences in the nonmyopic subgroup. 

Table 3

 

Intereye Comparison of the Baseline Characteristics of Unilateral POAG in Myopic and Nonmyopic Subgroups

 

Table 4 gives the factors associated with the glaucomatous eyes in the myopic and nonmyopic subgroups. In the myopic subgroup, a larger β-zone width (P = 0.002) and a lower IOP at examination (P = 0.004) were significantly associated with the presence of glaucoma in the univariate analysis. The multivariate analysis revealed that a larger disc tilt angle (P = 0.001), a larger β-zone width (P < 0.001), and a lower IOP at examination (P = 0.004) were associated with the presence of glaucoma. In the nonmyopic subgroup, only a larger β-zone width was significantly associated with the presence of glaucoma in both the univariate and multivariate analyses (P = 0.007 and 0.005, respectively). 

Table 4

 

Factors Associated With the Presence of a Glaucoma in Unilateral POAG Patients in Myopic and Nonmyopic Subgroups

 

Representative Cases

Figure 2 shows a myopic patient with unilateral POAG where both eyes had γ-zone, but the β-zone was only present in the glaucomatous eye. 

Figure 2

 

A case of a myopic patient with unilateral POAG in the right eye. Both circumpapillary RNFL thickness profiles (C, G) and visual field pattern deviation plots (D, H) confirm the presence of glaucomatous damage in the right eye, but not in the left eye. (A, B) Note that the glaucomatous eye has both β-zone (area between the white and red arrowheads) and γ-zone (area between the red and black arrowheads). (E, F) Meanwhile, in the contralateral healthy eye, the entire clinical PPA area consists of γ-zone (area between the black and white arrowheads).

Discussion

The present study showed that the β-zone, but not γ-zone, was a significant lateralizing factor for unilateral POAG with a similar IOP in both eyes. The result suggests that the presumed features of glaucomatous damage associated with the PPA may be mainly related to the β-zone rather than to the γ-zone. 

Previous studies have suggested a different pathogenesis and a different influence on glaucoma depending on the microstructure of the clinical PPA. It has been suggested that the PPA having BM tissue is associated with age-related atrophy of the BM-retinal pigment epithelium complex, whereas the PPA without BM results from axial elongation in young subjects with myopic progression.¹² Jonas et al.¹⁰ showed that the β-zone was associated with glaucoma, while the γ-zone was independent of glaucoma. Later, Kim et al.⁹ reported that glaucomatous eyes with β-zone had a faster rate of RNFL thinning than those with γ-zone or without any type of clinical PPA. Our result is in line with the previous findings. We speculate that a collapse of choriocapillaries in the parapapillary region and reduced perfusion may be associated with the development of the β-zone,²² which may also induce ischemic optic nerve damage by reduction of nutrition to the β-zone and thereby increase vulnerability to glaucomatous damage.^23,24 However, the result of this study does not explain the causal relationship between the development of the β-zone, decreased perfusion, and glaucomatous axonal damage. A longitudinal follow-up observation is required to reveal the causal relationship between the β-zone and the development of glaucoma. 

Myopia is considered one of the risk factors for the development of glaucoma,^25,26 whereas its influence on glaucoma progression is not yet clarified.^27–29 In a study by Kim et al.,⁹ patients with γ-zone, which is associated with myopia, had a slower rate of glaucomatous RNFL thinning. The study suggested that the development of glaucoma in myopic eyes may be associated with increased shearing force on the lamina cribrosa due to increased scleral tension during axial elongation, while such changes may stop with aging, relieving the stress on the lamina cribrosa. In the present study, the analysis performed within the myopic subgroup showed a significant influence of β-zone with a similar size of γ-zone and a similar refractive error and AXL. The finding suggests that with a similar degree of myopia, the presence of β-zone may play an additional role in the development of glaucoma. 

A larger optic disc tilt angle was significantly associated with the presence of glaucoma in the myopic subgroup. This is in agreement with previous reports, and it supports the hypothesis that optic disc tilt may increase tensile stress on the axonal fibers, leading to damage to the axons.^16,30 Meanwhile, neither the tilt ratio nor the torsion degree had a significant relationship with glaucoma, either in the subjects overall or in the myopic subgroup, which is not consistent with previous reports.^17,30 The discrepancy is mainly due to the definitions of optic disc tilt and torsion differing between the studies: measurements were based on the BMO in the present study, whereas previous studies used parameters based on the clinical optic disc margin. It has been suggested that the BMO is a more consistent parameter than the clinical optic disc margin, because the latter does not comprise a single anatomic structure.^13,31 The conventional tilt and torsion calculated based on the clinical optic disc martin represented the ovality of clinical optic disc and the angular deviation of the ovality, respectively, and thus did not account for the actual tilt or torsion. We believe that using the OCT-determined BMO may yield more reliable measurements of the degree of optic disc tilt or torsion. 

On the other hand, it also should be noted that although there were significant intereye differences in the tilt angle, the absolute magnitude of the difference was small. On the other hand, none of the tilt-related parameters were associated with the presence of glaucoma in the nonmyopic subgroup. Although this is a natural finding based on the negligible degree of optic disc tilt in both eyes in the nonmyopic subgroup, this finding further supports the notion that the pathogenic mechanism associated with optic disc tilt is mainly linked to axial elongation during myopia progression, and the mechanism is likely to be confined to myopic patients. 

A lower IOP at OCT examination was another significant factor for glaucoma. This may be associated with glaucomatous eyes being under IOP-lowering treatment at the time of the OCT examination. 

As had been described previously, this study is limited by its cross-sectional design. Therefore, the causal relationship between the glaucomatous axonal damage and the development of β- or γ-zone cannot be explained by the current study. In addition, the possibility of differentiating between peripapillary tissues in ways other than dividing them into areas with/without BM should be considered. Last, the presence of PPA was determined based on only infrared images, which could have missed eyes that had PPA that was not visible in infrared images. 

In conclusion, the width of the β-zone, but not the width of the γ-zone, was a significant lateralizing factor for unilateral POAG, which suggests that the presumed features of glaucomatous damage associated with the conventional PPA may be mainly related to the β-zone. In addition, a significant influence of both the β-zone width and optic disc tilt angle on the presence of glaucoma was found with similar degrees of myopia in both of the eyes of myopic patients. This may indicate that, in the development of unilateral POAG, atrophic change of the parapapillary tissue may play a role in addition to mechanical stretching by optic disc tilt. 

Acknowledgments

Supported by the National Research Foundation of Korea Grant, and the Korean Government (Grant 2013R1A1A1A05004781), Seoul, Korea. The funding organization had no role in the design or conduct of this research. The authors have no proprietary or commercial interest in any of the materials discussed in this paper. 

Disclosure: Y.J. Yoo, None; E.J. Lee, None; T.-W. Kim, None