Microstructure of Parapapillary Atrophy: Beta Zone and Gamma Zone

Abstract

Purpose.: To examine the morphologic features of parapapillary atrophy by using enhanced depth imaging optical coherence tomography (EDI-OCT) and color fundus photographs.

Methods.: The clinical observational comparative study included 80 normal eyes of 46 subjects and 80 eyes of 46 patients with primary open-angle glaucoma. Both groups did not vary significantly in axial length (P = 0.62) and refractive error (P = 0.30). Color fundus photographs and cross-sectional B-scan images obtained by EDI-OCT were examined. On the EDI-OCT images, we measured a gamma zone defined as the region between the temporal disc margin to the beginning of Bruch's membrane, and a beta zone defined as Bruch's membrane without retinal pigment epithelium.

Results.: The gamma zone (mean area: 1.13 ± 2.04 mm²) was significantly associated with longer axial length (P < 0.001; standardized coefficient beta: 0.48), longer vertical disc diameter (P < 0.001; beta: 0.43), older age (P = 0.008; beta: 0.22), and the absence of glaucoma (P = 0.03; beta: −0.19). The beta zone (mean area: 0.85 ± 0.60 mm²) was associated with longer axial length (P < 0.001; beta: 0.39) and the presence of glaucoma (P < 0.001; beta: 0.48).

Conclusions.: In addition to associations with older age, increasing myopia, and larger disc size, the EDI-OCT–defined gamma zone of parapapillary atrophy was associated with the absence of glaucoma, whereas the EDI-OCT–defined beta zone was associated with the presence of glaucoma. Differentiation between the beta zone and the gamma zone may be clinically useful.

Introduction

Parapapillary atrophy has so far been differentiated into an alpha zone and a beta zone. ¹ The alpha zone was defined as irregular hyperpigmentation and hypopigmentation and it was located in the periphery of the parapapillary atrophy. The beta zone was characterized by visible sclera and visible large choroidal vessels and location between the peripapillary scleral ring and alpha zone. Histologic studies in human eyes and monkey globes confirmed that alpha zone histologically corresponded to irregularities in the retinal pigment epithelium (RPE), and that the corresponding beta zone showed a complete loss of RPE cells and an almost complete loss of photoreceptors combined with a closure of the choriocapillaris. ^2–4 Enhanced depth imaging (EDI) of optical coherence tomography (OCT) is a new technique to visualize deeper structures of the macula and optic nerve head. ^5,6 Using the EDI-OCT technology, recent clinical studies have demonstrated now that in some medium myopic eyes, the end of Bruch's membrane did not touch the optic disc border. ^7–14 Since this region without Bruch's membrane did not fulfill the definition of the alpha zone nor of the beta zone, it may be called the gamma zone. It was the purpose of our study to assess whether beta zone and gamma zone as defined by EDI-OCT are associated with other parameters such as age, sex, axial length, refractive error, and glaucoma. 

Methods

The clinical observational study included ophthalmologically normal subjects and patients with primary open-angle glaucoma. The study protocol was approved by the Institutional Review Board of Eye and ENT Hospital, Fudan University and written informed consent was obtained from all participants. All investigations adhered to the tenets of the Declaration of Helsinki. All study participants underwent a complete ocular examination including refractometry and assessment of visual acuity, slit-lamp assisted biomicroscopy of the anterior and posterior segments of the eye, gonioscopy, applanation tonometry, biometry with measurement of the axial length (IOL Master; Carl Zeiss Meditec, Jena, Germany), color photography of the optic nerve head (CR-DGI; Canon, Inc., Tokyo, Japan), and perimetry (Humphrey Visual Field Analyzer; Zeiss, Inc., Oberkochen, Germany). Glaucoma was defined by the appearance of the optic nerve head (abnormal shape of the neuroretinal rim according to the Inferior–Superior–Nasal–Temporal (ISNT) rule, ¹⁵ defects in the retinal nerve fiber layer, and corresponding defects in the visual field. In the normal subjects, intraocular pressure was <21 mm Hg, the retinal nerve fiber layer thickness as measured by spectral-domain optical coherence tomography (OCT) (combined with the Heidelberg Retina Angiograph Spectralis HRA plus OCT; Heidelberg Engineering, Heidelberg, Germany) was within the normal range, and the visual field was unremarkable. Exclusion criteria were any retinal or optic nerve disease other than glaucoma, any neurological causes for visual field loss or optic nerve damage, previous retinal surgery, history of any ocular surgery in the past 3 months, or any other nonglaucomatous cause that might have affected the visual field or the status of the retinal nerve fiber layer. The inclusion criterion was a clear detectability of classic beta zone of parapapillary atrophy on the fundus photographs and on the enhanced depth imaging scans of OCT. 

Tomographic images of the parapapillary fundus were taken by using the enhanced depth imaging mode of the optical coherent tomograph (Spectralis HRA-OCT; a combination of the Heidelberg Retina Angiograph HRA and Spectralis; Heidelberg Engineering). For the EDI-OCT imaging of the parapapillary region, the OCT device was set to image a 15° × 15° rectangle for horizontal scans and vertical scans including the whole parapapillary region atrophy and the optic disc. This rectangle was scanned with 97 sections, and each section had 20 OCT frames averaged. The distance between each of the B-scans was 61 μm. The color fundus photographs were imported from a desk computer by the built-in software of the Spectralis-HRA device. They were overlaid onto the near-infrared reflectance image obtained by the OCT. The images were not automatically aligned. After a double-click, at least three reference points, such as blood vessels, were set on each image. After confirmation, the overlay was copied to the other image. The border of the optic disc (inner margin of the peripapillary ring) was marked on the near-infrared reflectance image so that the corresponding line could be seen on the B-scans. Measurements were made using the measurement tools built into the software of the Spectralis-HRA, including linear and area measurements, corrected for the magnification by the optic media of the eye. According to the cross-sectional OCT-EDI scans, we marked the borders of the gamma zone (region between optic disc border and end of Bruch's membrane) (Figs. 1, 2) and the beta zone (region between end of Bruch's membrane and beginning of the retinal pigment epithelium with underlying Bruch's membrane). The areas of gamma zone and beta zone were then measured on the corresponding near-infrared reflectance images obtained by the OCT, after overlaying the color fundus photographs onto the near-infrared reflectance images. The points corresponding to the widest parts of the gamma and beta zones were then marked. On the corresponding near-infrared reflectance images obtained by the OCT, we then measured the maximal width of the gamma zone and the beta zone as the distance between the marked points and the optic disc border in radial centripetal direction to the center of the optic disc. Additionally, the area and shape (ovality index) of the optic disc were measured on the near-infrared reflectance OCT images. The ovality index was calculated by dividing the shortest optic disc diameter by the longest disc diameter. The shape of the optic disc was classified as “round” when its ovality index was ≥0.8 and classified as “oval” when its ovality index was <0.8. 

Figure 1. 

 

Clinical photograph and optical coherence tomogram of the parapapillary region; Gamma zone (no Bruch's membrane): between black arrow (optic disc border) and blue arrow; beta zone (Bruch's membrane without retinal pigment epithelium): between blue arrow and red arrow (upper half of figure); and between both blue arrows (lower half of figure).

Figure 2. 

 

Clinical photograph and optical coherence tomogram of the parapapillary region; the gamma zone (no Bruch's membrane): between black arrow and blue arrow; the beta zone (Bruch's membrane without retinal pigment epithelium): between blue arrow and red arrow; the alpha zone (Bruch's membrane with irregular retinal pigment epithelium): between red arrow and white arrow.

To assess the reproducibility of the technique, two examiners (YD, HH) reexamined 10 images 10 times. The coefficient of reproducibility was calculated as the mean of the SD values divided by the mean of the means. 

The statistical analysis was performed using a commercially available statistical software package (SPSS for Windows, version 20.0, SPSS, Inc., Chicago, IL). Only one randomly chosen eye per subject was taken for the statistical analysis. In a first step of the analysis, we calculated the means and SDs as well as medians and ranges of the main outcome parameters. Using the Kolmogorov–Smirnov test, we assessed the Gaussian distribution of the outcome parameters. In a second step, we compared the sizes of the parapapillary zones between the glaucoma group and the control group using tests for unpaired samples (nonparametric test for the gamma zone parameters; parametric test for the beta zone parameters). In a third step of the analysis, we performed a univariate analysis to assess associations between the parapapillary zones and other ocular and systemic parameters. In a fourth and final step, we carried out a multivariate analysis, the list of independent parameters of which primarily included all variables that were significantly associated with the parapapillary region in univariate analysis. We then dropped step by step those independent parameters that were no longer significantly associated with the parapapillary zone. 95% confidence intervals (95% CI) were presented. All P values were two-sided and considered statistically significant when they were <0.05. 

Results

The study included 80 normal eyes of 45 subjects and 80 eyes of 45 patients with primary open-angle glaucoma. Mean age was significantly (P = 0.03) higher in the glaucoma group (45.5 ± 14.5 years; median: 46 years; range, 17–74 years) than that in the control group (38.4 ± 13.3 years; median: 39 years; range, 12–67 years). Both groups did not vary significantly in axial length (25.5 ± 2.0 mm; median: 25.30 mm; range, 22.7–30.0 mm) versus 25.6 ± 2.3 mm (median: 25.89 mm; range, 21.9–33.6 mm; P = 0.62) and in refractive error (−5.2 ± 4.0 diopters; median: −5.0 diopters; range, range −15.0–0 diopters) versus −6.0 ± 3.9 diopters (median: −5.50 diopters; range, −21.0–0 diopters; P = 0.30). All eyes were phakic. 

Area Measurements

Mean area of the gamma zone was 1.13 ± 2.04 mm² (median: 0.65; range, 0–15.04 mm²). Area of the gamma zone was not normally distributed (P < 0.001). Gamma area zone was significantly (P = 0.03; Mann–Whitney U test) smaller in the glaucoma group (1.00 ± 2.33 mm²) than that in the control group (1.25 ± 1.71 mm²). In univariate analysis, the gamma zone area was significantly related with myopic refractive error (P < 0.001), axial length (P < 0.001) (Fig. 3), vertical optic disc diameter (P < 0.001), optic disc ovality (P < 0.001), and area of the beta zone (P = 0.04) (Table). The gamma zone area was not significantly associated with age (P = 0.35), and horizontal disc diameter (P = 0.75), nor with sex (P = 0.70; Mann–Whitney U test). 

Figure 3. 

 

Scatterplot showing the correlation between area of gamma zone of parapapillary atrophy and axial length.

Table

 

Associations between Gamma Zone Area (mm²) or Beta Zone Area (mm²) and Ocular Parameters (Univariate Analysis)

Parameter	P Value	Standardized Coefficient Beta	Regression Coefficient	95% Confidence Interval
Gamma zone area
Refractive error, diopter	<0.001	−0.66	−0.34	−0.42, −0.26
Axial length, mm	<0.001	0.61	0.60	0.42, 0.77
Vertical disc diameter, mm	<0.001	0.53	0.004	0.003, 0.006
Optic disc ovality index	<0.001	−0.50	−10.2	−13.9, −6.4
Beta zone area, mm2	0.04	0.22	0.73	0.03, 1.44
Age, y	0.35
Horizontal disc diameter	0.75
Beta zone area
Refractive error, diopter	0.007	−0.28	−0.04	−0.07, −0.01
Axial length, mm	0.001	0.35	0.10	0.04, 0.16
Vertical disc diameter, mm	0.01	0.26	0.001	0.000, 0.001
Optic disc area, mm2	0.03	0.23	0.23	0.02, 0.44
Gamma zone area, mm2	0.04	0.22	0.06	0.002, 0.12
Age, y	0.03	0.23	0.010	0.001, 0.018
Horizontal disc diameter	0.05	0.21	0.000	0.000, 0.001

Model building for multivariate analysis began with the list of independent parameters including age, axial length, the presence of glaucoma, beta zone area, vertical disc diameter, and disc ovality. From this full model, nonsignificant terms were removed step by step beginning with the parameter with the highest P value (beta zone area; P = 0.77). With the reduced list of independent parameters, the multivariate analysis was repeated, leading to exclusion of the optic disc ovality parameter (P = 0.36), until eventually all remaining independent parameters showed a significant association with gamma zone area: the gamma zone area was associated with longer axial length (P < 0.001; beta: 0.48; B: 0.47 [95% CI: 0.31, 0.63]), longer vertical disc diameter (P < 0.001; beta: 0.43; B: 0.003 [95% CI: 0.002, 0.005]), older age (P = 0.008; beta: 0.22; B: 0.03 [95% CI: 0.01, 0.06]), and the absence of glaucoma (P = 0.03; beta: −0.19; B: −0.77 [95% CI: −1.49, −0.06]). 

Mean area of the beta zone was 0.85 ± 0.60 mm² (median: 0.74 mm²; range, 0–2.85 mm²). Area (P = 0.15) and width (P = 0.23) of the beta zone were normally distributed. Beta zone area was significantly (P < 0.001) larger in the glaucoma group (1.14 ± 0.59 mm²) than that in the control group (0.56 ± 0.45 mm²) (Fig. 4). In univariate analysis, the beta zone area was significantly associated with older age (P = 0.03), increasing axial length (P = 0.001), myopic refractive error (P = 0.001), longer vertical disc diameter (P = 0.01), larger disc area (P = 0.03), and larger gamma zone area (P = 0.04) (Table). It was not significantly associated with sex (P = 0.70), optic disc ovality (P = 0.71), and horizontal disc diameter (P = 0.05). Model building for multivariate analysis again began with the list of independent parameters including age, axial length, the presence of glaucoma, gamma zone area, and vertical disc diameter. In a first step, the gamma zone area (P = 0.93) was removed, followed by the vertical disc diameter (P = 0.29) and age (P = 0.21). Then, eventually all remaining independent parameters showed a significant association with beta zone area: the beta zone area was associated with longer axial length (P < 0.001; beta: 0.39; B: 0.11 [95% CI: 0.06, 0.16]) and the presence of glaucoma (P < 0.001; beta: 0.48; B: 0.56 [95% CI: 0.35, 0.78]). 

Figure 4. 

 

Boxplots showing the distribution of the area of beta zone of parapapillary atrophy in the control group and in the glaucoma group.

Width Measurements

Mean width of the gamma zone was 0.41 ± 0.46 mm (median: 0.31; range, 0–2.26 mm). In univariate analysis, gamma zone width was significantly related with refractive error (P < 0.001; correlation coefficient r = 0.80), axial length (P < 0.001; r = 0.75; equation of the regression line: gamma zone width [mm] = 0.16 × axial length [mm] − 3.69), smaller horizontal disc diameter (P = 0.03; r = −0.23), longer vertical disc diameter (P < 0.001; r = 0.51), disc ovality (P < 0.001; r = −0.71), and width of beta zone (P = 0.03; r = 0.23). The gamma zone width was not significantly associated with age (P = 0.98) and sex (P = 0.52). In the following multivariate analysis, we removed step by step nonsignificant terms from the list of independent variables. It started with beta zone diameter (P = 0.59), followed by disc ovality (P = 0.27), and age (P = 0.06). Finally, the gamma zone width was significantly associated with longer axial length (P < 0.001; beta: 0.64; B: 0.14 [95% CI: 0.11, 0.17]), longer vertical disc diameter (P < 0.001; beta: 0.35; B: 0.61 [95% CI: 0.35, 0.86]), and the absence of glaucoma (P = 0.02; beta: −0.18; B: −0.16 [95% CI: −0.29, −0.03]). 

Mean width of the beta zone was 0.25 ± 0.17 mm (median: 0.22; range, 0–0.97 mm). In univariate analysis, beta zone width was significantly associated with longer axial length (P < 0.001, r = 0.40), myopic refractive error (P = 0.001; r = −0.33), and longer vertical disc diameter (P = 0.008; r = 0.28). It was significantly larger in the glaucomatous group than that in the control group (0.30 ± 0.17 mm versus 0.20 ± 0.17 mm; P = 0.002). It was not significantly associated with age (P = 0.05), sex (P = 0.34), disc ovality (P = 0.0245), and horizontal disc diameter (P = 0.24). In the multivariate analysis, beta zone width was significantly associated with axial length (P < 0.001; beta: 0.41; B: 35 [95% CI: 18, 52]) and the presence of glaucoma (P = 0.01; beta: 0.27; B: 96 [95% CI: 24, 168]), whereas vertical disc diameter (P = 0.11) was no longer significantly associated. 

The coefficients of reproducibility for examiner 1 and examiner 2, respectively, were 0.7% and 0.9%, 2.3% and 2.5%, and 3.2% and 3.5% for the repeated assessments of the area of the optic disc, beta area, and gamma area, respectively. 

Discussion

In our clinical study on parapapillary atrophy, we differentiated on OCT images between a peripheral beta zone defined as Bruch's membrane devoid of retinal pigment epithelium cells, and a more centrally located gamma zone defined as parapapillary sclera without overlying choroid, Bruch's membrane, and deep retinal layers. The results showed that beta zone was significantly associated with the presence of glaucoma, whereas the gamma zone was associated with the absence of glaucoma. Both zones increased with age, myopia (axial length), and disc size. 

The results of our study on gamma zone as parapapillary region between the border of the optic nerve and the beginning of Bruch's membrane confirm other previous clinical studies. ^7–14 Using spectral-domain optical coherence tomography, Hayashi and colleagues ¹³ examined the parapapillary region in 100 patients with primary open-angle glaucoma and in 100 normal subjects. They found that the parapapillary beta zone according to their definition was composed of straight or downward-curved Bruch's membrane in 68 eyes or of a downward-bending slope lacking Bruch's membrane in 79 eyes. This latter region without Bruch's membrane was termed “gamma” zone in our study. In the study reported by Hayashi et al., ¹³ the presence of glaucoma and less myopic refractive error were associated with the curved-type of Bruch's membrane, and the region without Bruch's membrane (called gamma zone in our study) was associated with myopic refractive error. These clinical findings agree with our results in that the gamma zone as compared with the beta zone was more strongly associated with axial length, and in that gamma zone was correlated with the absence of glaucoma, whereas beta zone was associated with the presence of glaucoma. Park and colleagues ⁹ assessed the microstructural anatomy of clinical beta zone parapapillary atrophy by using Fourier-domain optical coherence tomography. They found that the edge of Bruch's membrane did not extend to the optic disc margin in all eyes, what would be the equivalent of gamma zone in our previous histologic study. ¹⁶ Lee and colleagues ⁸ evaluated the cross-sectional configurations of peripapillary atrophy alpha zone and beta zone in normal subjects using spectral-domain optical coherence tomography. Among other findings, they reported on slope and step configurations of the scleral bed and hump- and wedge-shaped appearances of Bruch's membrane in the peripapillary region, and that the presence of the step configuration was associated with myopia and longer axial length. This step configuration resembled the gamma zone in our previous histologic study. ¹⁶  

Interestingly, the gamma zone in our study was strongly associated with axial length with a steep increase starting at an axial length of approximately 26.5 mm (Fig. 3). The cutoff value of an axial length of 26.5 mm is similar to the cutoff values of approximately −8 diopters for the differentiation between medium myopia and high myopia as suggested in clinical studies. ^17,18 The gamma zone was not significantly (P = 0.41) associated with the size of the beta zone in the multivariate analysis, suggesting that both zones were not directly depending on each other. 

In recent studies by Reis and colleagues, ¹⁴ it has been discussed to use the opening of Bruch's membrane (as determined by spectral-domain OCT) as the margin of the optic nerve head to measure the neuroretinal rim area. One of the reasons for this procedure was that Bruch's membrane ending can overhang into the nasal region of the scleral optic nerve canal. If in that situation the margin of the scleral optic nerve canal is taken for the assessment of the neuroretinal rim area, the rim area measurements are falsely large, since the rim is measured perpendicularly (as it should be) where the nerve fibers dive into the canal, and additionally en face in the region of the overhanging Bruch's membrane. If now Bruch's membrane opening is taken as optic disc margin and if the gamma zone is present, the advantage for a valid rim area measurement one has on the nasal side with the overhanging Bruch's membrane would be counteracted by a disadvantage on the temporal side, if the nerve fiber tissue in the gamma zone would all be counted as the neuroretinal rim area. If, however, the rim area is defined as the thickness of the retinal nerve fiber layer at the end of Bruch's membrane and is measured perpendicular to the retinal nerve fiber layer surface, one would have included the advantage of taking care of an overhanging Bruch's membrane end in the nasal side without having the disadvantage in gamma zone on the temporal side. This procedure would be a retinal nerve fiber layer measurement at the Bruch's membrane opening. Compared with the circular retinal nerve fiber layer thickness measurement in the peripapillary region, the potential advantage would be that it may be easier to outline the deep margin of the retinal nerve fiber layer at the end of Bruch's membrane than within the retinal tissue. 

The physiologic existence of the gamma zone, particularly in myopic eyes, may call attention to the anatomic fact that the optic nerve head is composed of a two-layered hole: one hole in the sclera forming the scleral optic nerve head canal, and one hole in Bruch's membrane, the so-called Bruch's membrane opening. The margin of the scleral optic nerve head serves for the ophthalmoscopic definition of the border of the optic disc. The opening of Bruch's membrane can be seen only by OCT. Since Bruch's membrane is not firmly attached to the sclera but separated by the spongy choroid from the sclera, a shifting of Bruch's membrane during the myopic elongation of the globe could occur, resulting in a slight temporal transposition of Bruch's membrane opening in relation to the optic nerve scleral canal. This would lead to an overhanging of Bruch's membrane ending into the nasal region of the scleral optic nerve head canal, and a scleral region bared of Bruch's membrane on the temporal margin of the scleral optic nerve head canal. This bared area in the temporal parapapillary region has been called the gamma zone, which represents the difference in area between the clinical disc border as defined ophthalmoscopically as the scleral optic nerve canal and Bruch's membrane opening. 

The beta zone as defined on the basis of EDI-OCT images was significantly associated with glaucoma. It shows that the histologic changes observed in the histologic beta zone, that is, loss of retinal pigment epithelium cells and photoreceptors and a closure of the choriocapillaris, may be related to the glaucomatous optic neuropathy. ^3,4 The conventional ophthalmoscopic beta zone has so far been defined as visible sclera and visible large choroidal vessel upon ophthalmoscopy. It thus included the (new) beta zone and the (new) gamma zone. Since the new beta zone was associated with glaucoma, whereas the new gamma zone was significantly associated with the absence of glaucoma, one may infer that the clinical differentiation between gamma zone and beta zone (both of which have so far been summarized into the formerly “beta zone” as defined by simple ophthalmoscopy) may increase the diagnostic precision of beta zone for glaucoma. 

Potential limitations of our study should be mentioned. First, it was a hospital-based study with inclusion criteria and exclusion criteria. As for any hospital-based investigation, there was the possibility of a bias by the inclusion and exclusion criteria and by the referral of the patients by the ophthalmologists. Second, the differentiation between gamma zone and beta zone needed the help of the EDI-OCT, so that clinically without an EDI-OCT device, the gamma zone cannot reliably be differentiated from the beta zone. Third, interestingly, beta zone and gamma zone were associated with the vertical disc diameter, whereas the associations with the horizontal disc diameter were unclear and not significant. The reason for this observation might have been the vertical tilting of the optic nerve head in myopic eyes, so that the transpupillary view onto the optic disc occurred in an oblique angle. Due to this perspective factor, the horizontal disc diameter was measured falsely low in myopic eyes, whereas the determination of the vertical disc diameter was mostly unaffected by the perspective factor. Fourth, our study can only be regarded as a pilot study, clearly showing the need for further and larger investigations. These studies may correlate clinical photographs with EDI-OCT images of the parapapillary region to search for clinical markers to better differentiate between both zones without the help of an EDI-OCT. 

In conclusion, the EDI-OCT–defined gamma zone of parapapillary atrophy was associated with the absence of glaucoma, whereas the EDI-OCT–defined beta zone was associated with the presence of glaucoma. Both zones were additionally correlated with older age, increasing myopia, and larger disc size. Clinical differentiation between beta zone and gamma zone may be useful. 

Footnotes

 Supported by the Key Clinical Program of the Chinese Ministry of Health Grant 2011‐36, National Science Foundation of China Grants 81170817 and 81170838, and Shanghai Science and Technology Commission Grant 11PJ1402100. The authors alone are responsible for the content and writing of the paper.

Footnotes

 Disclosure: Y. Dai, None; J.B. Jonas, None; H. Huang, None; M. Wang, None; X. Sun, None