Article Text

Download PDFPDF

Morphological characteristics of parapapillary atrophy and subsequent visual field progression in primary open-angle glaucoma
  1. Ahnul Ha1,2,
  2. Yong Woo Kim1,3,
  3. Jinho Lee1,4,
  4. Eunoo Bak1,3,
  5. Young Soo Han1,3,
  6. Young Kook Kim1,3,
  7. Ki Ho Park1,3,
  8. Jin Wook Jeoung1,3
  1. 1 Department of Ophthalmology, Seoul National University College of Medicine, Seoul, South Korea
  2. 2 Department of Ophthalmology, Jeju National University Hospital, Jeju-si, South Korea
  3. 3 Department of Ophthalmology, Seoul National University Hospital, Seoul, South Korea
  4. 4 Department of Ophthalmology, Hallym University Chuncheon Sacred Heart Hospital, Chuncheon, South Korea
  1. Correspondence to Professor Jin Wook Jeoung, Department of Ophthalmology, Seoul National University Hospital, Jongno-gu, Seoul, Korea (the Republic of); neuroprotect{at}


Background/Aims To investigate the associations between the morphological characteristics of beta-zone parapapillary atrophy (β-zone PPA) and subsequent visual field (VF) progression in eyes with primary open-angle glaucoma (POAG).

Methods One hundred and twenty-one POAG eyes with β-zone PPA along with 48 normal eyes with β-zone PPA were included. β-zone PPA area was calculated based on the PPA pixel area/optic-disc pixel area ratio and the optical coherence tomography (OCT)-measured disc area. β-zone PPA margin irregularity was quantified as a function of both area (A) and perimeter (P, calculated as 1/(4πA/P²)). VF progression was defined using standard automated perimetry’s guided progression analysis software.

Results Of the 121 POAG eyes, 49 (40.5%) showed VF progression during the 10.1±1.9 years of follow-up. The baseline β-zone PPA area was similar among the three groups (Progressors, Non-progressors and Controls, p=0.995). However, the β-zone PPA irregularity index was significantly higher in the Progressors (p<0.001). The cumulative probability of VF progression was greater in the higher PPA irregularity index group (p<0.001, log-rank test). A Cox proportional hazards model showed the significant influences of optic disc haemorrhage (HR: 2.661, p=0.034) and higher baseline PPA irregularity index (HR: 1.455, p=0.007) on subsequent progression.

Conclusions In POAG eyes, baseline β-zone PPA margin regularity was significantly associated with subsequent VF progression. Irregular margin of β-zone PPA might be the mark of vulnerability in the parapapillary area to further glaucomatous damage.

  • glaucoma
  • imaging
  • optic nerve

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Glaucomatous optic neuropathy is associated with various morphologic changes to the optic nerve head (ONH). Such changes include both development and enlargement of beta-zone parapapillary atrophy (β-zone PPA), as characterised by atrophy of the retinal pigment epithelium, photoreceptors and choriocapillaris, and which results in distinctive sclera and choroidal vessel appearances.1 2

The spatial relationship between β-zone PPA and glaucomatous damage is well-established. Cho and Park showed a significant correlation in primary open-angle glaucoma (POAG) between angular location of localised retinal nerve fibre layer (RNFL) defect and β-zone PPA.3 Park et al reported a topographic correlation between PPA and visual field (VF) defect in patients with normal-tension glaucoma. Jonas and Naumann found a significant association among location of β-zone PPA and glaucomatous neuroretinal rim change and corresponding VF damage.2

Additionally, a growing body of evidence points to a correlation between β-zone PPA and subsequent glaucomatous progression. The presence of β-zone PPA was determined to increase the risk of VF deterioration in treated POAG eyes.4 Teng et al reported that largest β-zone PPA location correlates spatially with region of most rapid future VF progression in patients with glaucoma suffering both β-zone PPA and VF progression.5 By contrast, other studies have failed to identify any significant correlation between progression of β-zone PPA and neuroretinal rim change.6 7 Thus, the relationship between β-zone PPA and further glaucomatous deterioration remains unclear.

Accordingly, in this study, we compared progressive change of morphological features in cases of β-zone PPA between POAG eyes and normal healthy controls, which patients had been followed up for more than 10 years. In addition, we explored the relationship between β-zone PPA’s baseline morphological features and subsequent VF progression.


This study was approved by the Seoul National University Hospital (SNUH) Institutional Review Board and adhered to the tenets of the Declaration of Helsinki. The participants provided written informed consent.

Inclusion criteria

The current study included patients from the POAG cohort of an ongoing study at SNUH’s Glaucoma Clinic. Among the patients examined between January 2004 and June 2018 and included in the POAG cohort, those regularly followed up for at least 10 years were considered as subjects to be included. The glaucoma group’s inclusion criteria were as follows: (1) diagnosis of POAG; (2) at least two colour stereo disc photography (SDP) images of sufficient quality that had been taken at intervals of more than 10 years; and (3) a minimum of 10 reliable VF results (excluding the first test).

POAG diagnosis had been made according to the following criteria, irrespective of the level of untreated intraocular pressure (IOP): presence of glaucomatous optic disc changes (eg, diffuse and/or localised notching, thinning, RNFL defects on SDP and/or red-free fundus photography), and an open angle as confirmed by gonioscopic examination. β-zone PPA was considered not to be a criterion for classification of glaucomatous optic neuropathy.

For the control group, patients regularly examined due to a family history of glaucoma or of cataract were enrolled consecutively. For inclusion in this study, the following criteria had to be met: (1) baseline IOP less than 21 mm Hg, with no history of elevated IOP; (2) normal optic disc and retina according to SDP/RNFL photography; and (3) normal VF. The eyes followed as glaucoma-suspect (eg, cup-to-disc ratio >0.6) were not included in the control group.

All of the enrolled patients underwent a comprehensive ophthalmic examination that included best corrected visual acuity assessment, refraction, slit-lamp biomicroscopy, Goldmann applanation tonometry (Haag-Streit, Koniz, Switzerland), dilated fundus examination, digital colour SDP, red-free RNFL photography, central corneal thickness measurement (Orbscan 73II, Bausch&Lomb, Rochester, New York, USA), axial length (AXL) measurement (AxisIIPR; Quantel Medical, Inc, Bozeman, Montana, USA), Cirrus spectral domain-optical coherence tomography (SD-OCT; Carl Zeiss Meditec, Dublin, California, USA) and a central 24-2 threshold test of Humphrey Visual Field (HFA II; Humphrey Instruments Inc, Dublin, California, USA).

Exclusion criteria

PPA is one of the key morphologic features of myopic ONH.8 The presence of β-zone PPA can be determined by both glaucomatous damage and myopic change.9 We intended to exclude the effects of structural alterations to the ONH and/or parapapillary retina accompanying axial elongation of the eyeball. Thus, only eyes with AXL 24.0 mm or less were included in the final analysis.10 If both eyes were eligible according to the inclusion criteria, one eye was selected randomly for further study. See online supplementary eText for additional exclusion criteria.

Planimetric analysis of PPA

Optic disc tilt ratio, optic disc area and PPA area were measured on SDP/RNFL photographs using ImageJ software (V.1.48; Wayne Rasband, National Institute of Health, Bethesda, Maryland, USA). Images with poor quality owing to cataract or technical reasons that prevented reliably outlining the disc margin or PPA were excluded. All of the PPA parameters were measured independently by two glaucoma specialists (AH/EB), in a masked fashion, without knowledge of the pertinent clinical diagnosis or of any other clinical information. The structures for quantification were outlined on the inside edge in order that the trace thickness could be incorporated into the total delineated area.

Optic disc tilt was measured according to the ovality index. Thus, the tilt ratio was that between the longest and shortest ONH diameters. The optic disc area was measured based on the SD-OCT-camera-related and eye-related magnification factors by substitution of the subject's AXL, as noted in a previous study.11 The α-zone was defined as an irregular area of hypopigmentation and hyperpigmentation adjacent to the scleral ring or located on the outer side of the β-zone. The β-zone, which extended from the scleral ring, was characterised by the absence of the retinal pigment epithelium, which made visible either the sclera or the choroid and its large vessels. The β-zone PPA area and clinical disc margin were plotted using a mouse-driven cursor to trace the disc and β-zone PPA margins onto the SDP/RNFL image directly. The peripapillary scleral ring was incorporated into the β-zone measurements. Because the width of the scleral ring usually was very thin, any error introduced by adding its area to the β-zone was deemed inconsequential. Subsequently, the pixel areas of the β-zone PPA and clinical disc were obtained by ImageJ software. Then, the β-zone PPA area was calculated from the PPA pixel area/disc pixel area ratio as previously described.11 12

PPA irregularity index

Morphometric analysis of PPA has focused on area, angular extent or largest radial extent4 5 13 14; however, PPA shape varies widely among people.15 In the present study, therefore, we introduced an additional parameter, namely the PPA irregularity index, to reflect the detailed morphological variability of PPA.

In the field of dermatology, ‘border irregularity index’ has been used as a parameter for quantification of one of the features of malignant skin lesions.16 17 This is predicated on the fact that as the border becomes more irregular, the perimeter length is further increased relative to the area.18 Correspondingly, the PPA margin irregularity index was quantified as a function of area (A) and perimeter (P, calculated as 1/(4πA/P²)), a higher value indicating greater border irregularity. The margin of β-zone PPA was plotted directly onto the SDP/RNFL image as described above. Representative examples of β-zone PPA and their respective irregularity indices are shown in figure 1.

Figure 1

Measurement of PPA irregularity index. The figure’s first row includes left-eye optic disc photographs showing the delineation of the β-zone PPA outer margin in the white dotted line. The second row shows the β-zone PPA configuration in the black background. (A) POAG eye showing irregular PPA margin and, accordingly, higher irregularity index. (B) Example of POAG eye with relatively regular margin of β-zone PPA and, accordingly, lower irregularity index. The PPA irregularity index was quantified as a function of area (A) and perimeter (P), and was calculated as 1/ (4πA/P²), a higher value indicating greater border irregularity. β-zone PPA, beta-zone parapapillary atrophy; POAG, primary open-angle glaucoma.

Assessment of PPA progression

Both qualitative and quantitative approaches were applied to determine β-zone PPA progression. For the qualitative assessment, two glaucoma specialists (AH/EB) evaluated each subject’s SDP/RNFL photographs at the baseline and final visit. Conspicuously increased vessel visibility and atrophy in the β-zone area were considered to represent PPA progression. PPA progression was confirmed by agreement of the two experts. If they could not agree, a third examiner (JWJ) made the determination.

As a quantitative approach, the measurements of β-zone PPA area at the baseline and at the final visit were compared. The average difference of the two ophthalmologists’ measurements of PPA area ranged from 1.7% to 9.7%. Thus, if the PPA area increase exceeded 10% of the baseline value, the eye was classified as ‘progression of β-zone PPA.’

Assessment of VF progression

VF progression was determined using commercial software (Humphrey Field Analyzer Guided Progression Analysis; Carl Zeiss Meditec), and detailed criteria are provided in online supplementary eText.

Data analysis

The normally distributed data were compared by independent t-test, and the categorical data were analysed by χ2 test. Longitudinal changes of PPA parameters were evaluated within each group by paired t-test. The PPA characteristics were compared among the three groups by one-way analysis of variance with the Bonferroni post hoc test. The intraobserver and interobserver reproducibilities were evaluated by the intraclass correlation coefficients and their CIs.

We calculated Youden’s index to identify the optimised cut-off value of PPA irregularity index for prediction of VF progression. Kaplan-Meier survival analysis and the log-rank test were then used to compare, between the groups, VF progression’s cumulative risk ratio. The first time at which progression was identified was regarded as the survival analysis endpoint. The end of follow-up was the point at which patients without progression were censored. The HRs of the associations between the potential risk factors and glaucoma progression were determined by Cox proportional hazards modelling. For each factor, a univariate analysis was performed, and cases having a p value <0.1 were included in multivariate model. The final model was developed by backward elimination, and adjusted HRs with 95% CIs were calculated. Statistical analysis was performed using the MedCalc software (V., Mariakerke, Belgium). Two-sided p value <0.05 was considered to represent statistical significance.


Two hundred and eleven POAG and 218 normal eyes meeting the baseline eligibility criteria were included in this study. Among them, β-zone PPA was present in 121 (57.3%) and 48 (22.0%) eyes, respectively. Among those 121 POAG eyes with β-zone PPA, 49 showed VF progression. The demographic characteristics of both the included and the excluded patients are shown in online supplementary eTables 1–3.

Comparison of PPA characteristics among study subjects

Table 1 shows the PPA parameters at the baseline and final visit. The baseline PPA irregularity index did not show any correlation with optic disc tilt ratio (p=0.200) or β-zone PPA area (p=0.390). There were no significant differences in baseline optic disc tilt ratio or β-zone PPA area among the three groups (p=0.425 and p=0.995). The baseline PPA irregularity index, however, was higher in progressors than in non-progressors or normal subjects, and the differences were statistically significant (p<0.001). At the final visit, the optic disc tilt ratio was similar in all three groups (p=0.442). Both the β-zone PPA area and the PPA irregularity index were significantly greater and higher, respectively, in the progressor group (p=0.013 and p<0.001).

Table 1

Comparison of PPA's characteristics between patients with POAG and normal subjects (mean±SD)

Among the 48 normal eyes, 13 (27.1%) showed β-zone PPA progression by either qualitative or quantitative assessment. Of the 121 POAG eyes, 94 (77.7%) manifested β-zone PPA progression. In the majority of the POAG eyes with VF progression, β-zone PPA progression was apparent (48/49, 98.0%). The detailed results on the repeatability and reproducibility of measurements and on the longitudinal changes of the PPA parameters are provided in the online supplementary eText.19

Comparison of VF progression according to PPA irregularity index

We calculated Youden’s index to find the optimised PPA irregularity index value discriminating the VF progression group, and identified the cut-off point of 1.23. We dichotomised the data based on this cut-off value (ie, PPA irregularity index >1.23) and plotted Kaplan-Meier survival curves (online supplementary efigure 1). The cumulative probability of VF progression was greater in the higher PPA irregularity index group (p<0.001); 5-year survival rates for VF progression were 0.42±0.08 and 0.86±0.04 in the higher and the lower PPA irregularity index group, respectively.

Factors associated with VF progression in patients with POAG and β-zone PPA

According to the multivariate model, two factors were associated with progression: presence of disc haemorrhage (HR: 2.661, p=0.034) and higher baseline PPA irregularity index (HR: 1.455, p=0.007). The full statistical results are summarised in table 2.

Table 2

Univariate and multivariate risk factors for VF deterioration in patients with POAG and PPA

Representative cases

Figure 2 highlights representative cases of POAG eyes in the progressor and non-progressor groups along with a normal eye. The first row provides the baseline examination findings, and the second row includes the results at the final visit. Figure 3 shows the serial examination results in figure 2 cases. Note that the baseline PPA irregularity index was higher in progressors and that both PPA area and PPA irregularity index increased during the 10 years of follow-up.

Figure 2

Representative cases of POAG eyes in progressor and non-progressor groups and in normal eye. The first row in the figure shows the baseline examination findings on baseline PPA area, PPA irregularity index, colour SDP/RNFL photography and VF test. The second row provides the examination results at the final visit. Note that the baseline PPA irregularity index was higher in progressors, and that both PPA area and PPA irregularity index were significantly increased over the course of the 10 year follow-up period. POAG, primary open-angle glaucoma; PPA, parapapillary atrophy; RNFL, retinal nerve fibre layer; SDP, stereo disc photography; VF, visual field.

Figure 3

Serial examination results for representative figure 2 cases: (A) progressor; (B) non-progressor; (C) normal eye. The first rows in each figure show serial SDPs over the course of the 10 year follow-up period. The second rows show schematics of optic disc and β-zone PPA, as well as PPA area and PPA irregularity index value. The third rows in parts (A) and (B) depict the serial VF results. OAG, open-angle glaucoma; PPA, parapapillary atrophy; SDP, stereo disc photography; VF, visual field.


This study demonstrated the association of greater β-zone PPA margin irregularity with higher risk of subsequent VF progression in non-myopic POAG eyes. Additionally, we introduced an analytic method for quantification and comparison of the degree of β-zone PPA margin irregularity (ie, a PPA irregularity index).

Enlargement of PPA, particularly of the β-zone, has been correlated with progressive glaucomatous damage in a number of studies employing various methodologies. In a retrospective SDP review done by Rockwood and Anderson, a noticeable PPA increase was observed in 21% of eyes with progressive glaucoma versus only 4% of eyes with non-progressive glaucoma over the course of a 12 month follow-up period.20 In another retrospective study, 93% of 28 eyes in which progression of PPA had been observed showed either progressive optic disc damage or VF loss.6 Contrastingly, Quigley et al found no difference in PPA prevalence between patients with ocular hypertension who progressed to overt glaucoma after a 5-year follow-up and those who did not.21 Savatovsky et al, correspondingly, showed equal enlargement of PPA area between patients with ocular hypertension who had progressed to glaucoma over a mean 12.3 years of follow-up and those who had not.22

In the current study also, we found that β-zone PPA enlargement had occurred in both normal subjects and POAG eyes, regardless of further VF progression, though the proportion of PPA progression cases was much higher in the eyes with VF deterioration. Based on the results of previous studies and the findings on our own series, β-zone PPA enlargement might be associated with not only glaucomatous damage but also with age-related change20 23; this suggests that PPA size alone is not a sufficient parameter for assessment of the risk of further glaucomatous change.

Interestingly, POAG eyes with an irregular margin of β-zone PPA showed a higher risk of subsequent VF progression. It remains unclear why PPA margin irregularity is associated with glaucoma progression. However, it is possible that PPA shape in glaucomatous eyes reflects vascular and mechanical vulnerability around the ONH. Due to the fact that the ONH’s prelaminar portion receives its blood supply mainly from the peripapillary choroid via branches of short posterior ciliary arteries of a characteristic sectoral distribution,24–27 the absence or dysfunction of the centripetal branch in the PPA sector is possibly associated with ischaemic ONH damage in that focal segment.28 Additionally, irregular margin of PPA can be attributed to uneven mechanical forces on the focal tissues around the ONH. In these processes, greater β-zone PPA margin irregularity might reflect a more pathologic mechanism associated with glaucomatous damage rather than with physiologic change.

Jonas et al histologically subdivided classic β-zone PPA into newly defined β-zone PPA existing with Bruch’s membrane (BM) and newly defined Ɣ-zone PPA containing no overlying BM.29 Recent studies have demonstrated that such microstructures of the PPA might have different associations with glaucomatous change. That is, β-zone PPA (β-zone with BM) is associated with glaucoma but not with myopia, whereas Ɣ-zone PPA (β-zone without BM) is unrelated to glaucoma but related to myopia.30 31In the present study, although we excluded eyes with myopia (AXL >24 mm) in order to minimise the influence of Ɣ-zone PPA, its potential effects on measurements of PPA parameters and on the association between PPA parameters and glaucoma progression should be considered. Future studies using objective methods (eg, OCT) to evaluate PPA microstructures should test whether our result holds true.

Our findings must be interpreted in the light of a number of limitations. First, although we found good intra-reader and inter-reader agreements for the PPA parameters, determination of PPA borders was somewhat subjective. In an actual clinical setting, it would be difficult to correctly define the PPA margin in some cases; thus, this factor should be taken carefully into account when interpreting PPA parameters. In addition, in eyes with regional β-zone PPA, focal PPA irregularity can be diluted due to the relative smooth disc border of the area without PPA. In our opinion, however, the PPA irregularity index can reflect the overall morphological features of the PPA around the entire ONH. That is, since peripapillary areas absent of PPA are likely not to be associated with increased risk of glaucoma, dilution of overall irregularity may also reflect a lower risk of glaucomatous change in that ONH. Nonetheless, future studies with novel parameters to analyse regional changes around the ONH should be continued. Second, we studied mostly low-baseline IOP POAG eyes (92.8% of the subjects had a baseline IOP ≤21 mm Hg), and all of the subjects were Korean. Thus, caution needs to be exercised in any attempts to generalise our results. Third, since we included only eyes without myopia in the final analysis, our results might not be directly applicable to myopic populations. Although our analysis was confined to subjects with non-myopic POAG, by doing so, we intended to exclude the effects of structural alterations to the ONH and/or parapapillary retina accompanying axial elongation of the eyeball. We carefully enrolled subjects based on AXL, considering the fact that refractive error can be affected by the density of nuclear sclerosis.32 Further investigation on morphological characteristics of PPA and their associations with glaucomatous progression in POAG eyes with myopia is underway. Finally, the control group included subjects with a family history of glaucoma. Even though they did not develop glaucoma over the average 10.1 year follow-up, they still had an increased risk of developing the disease relative to those without any such history.

In conclusion, this study found a significant correlation between greater β-zone PPA margin irregularity with subsequent VF progression in POAG. This suggests that eyes with irregular margin of β-zone PPA, parapapillary tissues are increasingly vulnerable to further glaucomatous injury. Eyes with irregular margin of β-zone PPA should be more carefully monitored while keeping the possibility of VF deterioration in mind.



  • Presented at This paper was presented at the E-poster Discussion session, American Academy of Ophthalmology Annual Meeting, San Francisco, 2019.

  • Contributors AH, YKK, KHP and JWJ performed the study design. AH and JWJ helped in writing the article. Data collection was done by AH, JL, EB and YSH. Analysis and interpretation of the data were done by AH and YWK. AH, JL, EB and YSH helped in literature search. AH, YKK and KHP helped in critical revision of the article. JWJ gave the final approval of the article.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data are available upon reasonable request. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost B Jonas