Introduction

Glaucoma is a leading cause of irreversible blindness and myopia is highly prevalent in east Asia1. Myopia, especially high myopia, was one of the risk factors of glaucoma2. Diagnosing glaucoma in high myopic patients is challenging. A myopic optic disc change, like tilt and peripapillary atrophy may mimic the glaucomatous optic disc appearance. Optical coherence tomography (OCT) has been widely used in diagnosing and monitoring glaucoma. But the structural characteristics of high myopia lower the diagnostic accuracy of peripapillary retinal nerve fiber layer (pRNFL) thickness in detecting glaucoma3,4. Macular ganglion cell–inner plexiform layer (mGCIPL) thickness measurement was reported more effective than RNFL measurement for detecting glaucoma even in highly myopic eyes because the macular region has been less affected by optic disc variation5,6. Evaluation of disease progression is essential in the management of glaucoma. Not only diagnosing but also monitoring disease progression in highly myopic glaucoma is a challenge. Commercially available software and algorithms, Guided Progression Analysis (GPA) (Carl Zeiss Meditec, Dublin, CA), has provided progression analysis of RNFL and GCIPL thickness. However, the GPA does not include patients with high myopic patients in their database.

Vessel density parameters measured by optical coherence tomography angiography (OCTA) in the peripapillary and macular regions in the glaucomatous eyes were significantly lower than those in the healthy eyes7,8. Peripapillary/macular vessel density changes have been reported to be useful for the differentiating between glaucoma and healthy eyes, especially in high myopic patients because vessel density measurement is less affected by the low reflectance of the RNFL or optic disc deformation9,10. Studies have demonstrated that peripapillary vessel density had better topographic correlation with visual field comparing with the RNFL thickness in POAG eyes with high myopia11,12. However, there have been few reports determining longitudinal changes in vessel density in patients with high myopic glaucoma. Moreover, little is known about the change of pRNFL and mGCIPL thickness comparing to vessel density change in high myopic glaucoma in the longitudinal follow-up.

In the present study, we evaluated longitudinal changes of the macular and the peripapillary vessel density and the pRNFL and mGCIPL thickness in glaucoma patients with high myopia. We further analyzed the longitudinal changes of the structural thickness and the vessel density between the stable and the progression group.

Methods

Study design and participants

Highly myopic eyes (spherical equivalent <−6.00 diopters) with primary open angle glaucoma (POAG) examined at the glaucoma clinic of Far Eastern Memorial Hospital between September 2017 and September 2020 were consecutively enrolled on the basis of a retrospective medical-record review. The study was performed in accordance with the tenets of the Declaration of Helsinki. The study was approved by the institutional review board of the Research Ethics Review Committee at the Far Eastern Memorial Hospital, New Taipei City, Taiwan (approval no.:107135-E). All methods were performed in accordance with the relevant guidelines and regulations. Due to the retrospective nature of this study, the Research Ethics Review Committee of Far Eastern Memorial Hospital. (No. 107135-E) waived the need of obtaining informed consent.

The inclusion criteria for all groups were age more than 20 years, best corrected visual acuity 20/30, and open-angle confirmed by gonioscopy. Glaucomatous eyes were defined as glaucomatous damage to the optic disc (> 0.7 vertical cup/disc ratio, neuroretinal rim thinning, notching, or excavation) as accompanied by two corresponding and reliable abnormal visual field (VF) examinations, regardless of the intraocular pressure (IOP). Eyes with glaucomatous VF defects were defined as those with a glaucoma hemifield test result outside normal limits or a pattern standard deviation outside 95% of normal limits. Additionally, a cluster of three points with probabilities of 5% on the pattern deviation map in at least one hemifield, such as at least one point with a probability of 1%, or a cluster of two points with a probability of 1% was needed. A VF was defined as reliable when fixation losses were < 20%, and each of the false-positive and false-negative rates was < 15%. Subjects with evidence of retinal pathology, diabetes, hypertension, or non-glaucomatous optic nerve diseases were excluded, as well as eyes that had undergone previous laser therapy or ocular surgery (except cataract surgery), or the presence of any media opacities that prevented high-quality OCT scans.

To be included in the analyses, all subjects were followed up for at least 2.5 years and a minimum of 4 OCT scanning sessions were included. All participants underwent medical therapy during the follow-up period. All patients had OCT imaging and VF testing at 4–6 months interval.

Ophthalmic examination and imaging

All subjects underwent a complete ophthalmic examination, including visual acuity assessment; auto-refraction (Auto Refractometer AR-610; Nidek Co, Ltd., Tokyo, Japan); slit-lamp biomicroscopy; gonioscopy; non-contact tonometry (CT-80, Topcon, Japan); and dilated stereoscopic examination of the optic disc. They also underwent digital color stereo disc photography; red free RNFL photography; optic nerve head (ONH) and macular imaging by HD-OCT (Cirrus; Carl Zeiss Meditec) and a central 30 − 2 threshold test of the Humphrey visual field (HVF, HFA II; Humphrey Instruments, Inc., Dublin, CA, USA).

Peripapillary (Optic Disc Cube 200 × 200 protocol) and macular (Macular Cube 512 × 128 protocol) scans (collectively referred to as ganglion cell analysis) were acquired using the Cirrus 5000 HD-OCT (Carl Zeiss Meditec, Inc.). Software released by the manufacturer was used to calculate RNFL and GCIPL thicknesses, as previously described13. OCTA imaging of the macula was performed using the Cirrus HD-OCT (version 10.0.0.14618). The procedure for OCTA imaging using the Cirrus HD-OCT has been detailed previously14. Angiographic images were generated through OCT-based microangiography (OMAG), and the macula was imaged using a 6 × 6 mm2 scan pattern. Angiometric software of the Cirrus HD-OCT automatically calculates vessel density (defined as the total area of perfused vasculature per unit area in the region of measurement) from the superficial retinal layer slab. This software calculates the vessel density parameters across four inner and four outer sectors of the Early Treatment Diabetic Retinopathy Study (ETDRS) grid over the macula. To avoid the myopic optic disc morphological change effect on the scan, we analyzed only the outer sector parameter both in the peripapillary and the macula area.

Image quality was assessed for all OCTA and OCT scans. Poor quality images, which were defined as those with a signal strength of < 7, poor centration, or motion artefacts and segmentation errors, were excluded from analysis.

Glaucoma progression

We adopted VF progression as the reference standard for glaucoma progression in the current study. The first VF test was excluded to minimize the learning effect. VF progression was determined by the Early Manifest Glaucoma Trial criteria, an event-based algorithm for detection of visual sensitivity loss at individual location. Visual field progression was defined as possible progression when at least 3 test points were flagged as having deteriorated significantly at the same test point locations in 2 consecutive fields15. These changes also had to have been observed in all the subsequent visits. If both eyes of a patient were classified into either the stable or the progressive group, the patient was included. If one eye was stable and the other was progressive, only the progressive eye was included, as there were few cases in the progression group.

Statistical analysis

Data analyses were conducted using IBM SPSS (version 22, IBM Corp, Armonk, NY). Descriptive statistics were used to present data distributions, including counts (percentages) for categorical variables and means ± standard deviations for continuous variables. To compare patient characteristics between the stable and progressive groups, the chi-squared test was used for categorical variables, while the Mann-Whitney U test was employed for continuous variables. Rates of change in each structural and vessel density parameter were estimated individually, as the regression slopes, using linear regression models. Mean differences in these rates of change between the two groups were compared using Generalized Linear Mixed Models, adjusted for age, sex, SE, mean IOP, and baseline MD, while accounting for repeated measures in patients with both eyes. Given the potential impact of glaucoma severity on the comparison, patients were further stratified into mild and moderate-to-severe groups based on whether their MD was greater or less than − 6 dB. Due to the small sample size in the subgroup analyses, the Mann-Whitney U test was used for further comparisons. The significance level was set at 0.05 for all statistical tests. The datasets used and/or analysed during the current study available from the corresponding author(Pei-Yao, Chang) on reasonable request.

Results

The study included 94 eyes of 75 patients with POAG. Of the 94 eyes, 9 eyes had a poor-quality optic disc OCTA scan, 7 eyes had a poor-quality macular OCTA scan, and 7 eyes had poor quality both on disc and macular OCTA scans during follow up. OCTA data of these eyes were excluded, while the rest of the data from these eyes were used for the analysis. Finally, 62 glaucoma patients (71 eyes) with high myopia were included in the analyses. 47 eyes had stable visual field measurement, while 24 eyes had visual field progression during the follow up period. Table 1 summarizes the baseline clinical characteristics, VF results, OCT, and OCTA measurements of the included patients and presents the baseline and final OCT and OCTA parameters between the stable and the progressive group. Mean follow-up durations were 2.89 ± 0.62 years in the stable group and 2.86 ± 0.88 years in the progressive group, respectively. The mean number of OCT examinations performed during the follow-up was significantly more in the progressive group (5.50 ± 1.18) than that in the stable group (4.53 ± 0.86) (range, 3–7 examinations). The baseline mean visual field deviation and visual field index was significantly less in the progressive group than that in the stable group.

Table 1 Baseline characteristics of the glaucoma groups.
Full size table

Table 2 showed vessel vessel density parameters and structural parameters. The measurement in macular vessel density was not significantly different between the progressive and the stable groups, except in the inferior quadrant. The measurement in the optic disc vessel density was not significantly different between the two groups in every quadrant.

Table 2 Vessel perfusion density parameters measured by optical coherence tomography angiography (OCTA) and structural parameters measured by optical coherence tomography (OCT).
Full size table

Rate of change in the structural and vessel density parameter in the stable and the progressive groups was displayed in Table 3. The mean difference in slopes between the two groups was compared using generalized linear mixed model with adjustment for age, gender, SE, mean IOP, and baseline mean visual field deviation. Mean difference between the stable and progressive groups in the slope of macular PD and peripapillary PD was significantly different (all p < 0.05), while the slope of macular ganglion cell-inner plexiform layer (mGCIPL) thickness and peripapillary retinal nerve fiber layer (pRNFL) thicknesses showed no significant difference between the two groups (all p > 0.05). We further stratified our patients by glaucoma severity into mild group (MD>−6 dB) and moderate to severe group (MD<−6 dB). The result suggested that the decrease of inferior pRNFL thickness in the mild group was more in the progressive group but not reach the significance level. However, the slope of macular PD and peripapillary PD in the progressive group decreased significantly more than that in the stable group (all p < 0.05) (Table 4). Figure 1 demonstrated the vascular and structural image measured by OCTA and OCT at the baseline and 3 years later in a 58-year-old patient with progressive glaucoma.

Table 3 Multivariable analysis for rate of change in the structural and vessel density parameters between the stable and the progression groups.
Full size table
Table 4 Comparisons of rate of change in the structural and vessel density parameters between the stable and the progressive groups, stratified by severity.
Full size table
Fig. 1
figure 1

The vascular and structural image measured by OCTA and OCT at the baseline and 3 years later in a 58-year-old patient with progressive glaucoma. OCT angiography image of a 6 × 6 mm scan exhibited the superficial vessels at the peripapillary (a) and the macular area (b) at baseline and peripapillary (d) and the macular area (e) 3 years later. The grids represented the sectors across which the vessel length densities (/mm) were calculated. The area between the inner two circles in the 6 × 6 mm scan represented the inner sectors, and the area between the outer two circles represented the outer sectors. Combined RNFL and GCIPL deviation maps indicated structural glaucomatous damage at baseline (c, left) and 3 years later (f, left). OCT scan showed the GCIPL thickness (µm) provided by the ganglion cell analysis report (c, right bottom) and RNFL thickness (µm) (f, right top) in each sector at baseline and 3 years later (f, left). GCIPL, ganglion cell-inner plexiform layer; OCT, optical coherence tomography.

Full size image

Discussion

Although growing body of evidence has suggested that OCTA imaging is essential for patients with high myopic glaucoma10,11,12, there is limited empirical evidence for macular and optic disc vessel density assessment in the follow-up of high myopic glaucoma. Lin et al.16 reported the rate of macular capillary density loss in the deep capillary plexus in a longitudinal study in high myopic patients without glaucoma. Miguel A et al.17did a systemic review using OCTA to detect longitudinal microvascular changes in glaucoma. They included studies evaluating vessel density change following trabeculectomy18and peripapillary choroidal microvascular dropout19. There was only one study20in the systemic review comparing the longitudinal change of OCT and OCTA directly. However, the study20 investigated only the macular area without the peripapillary area. Our study demonstrated the evidence that the validity of the longitudinal follow-up of both the macular and optic vessel density to expedite the detection of glaucoma progression in the highly myopic eyes.

The retinal vessel density was reduced in high myopia, indicating that the microvessel network may be stretched due to axial elongation21. Suwan et al. demonstrated peripapillary microvasculature was reduced both in myopia and glaucoma, however, microvasculature attenuated to a greater extent in glaucoma than in myopia22. Studies have demonstrated that peripapillary vessel density had better topographic correlation with visual field comparing with the pRNFL thickness in POAG eyes with high myopia11,12. In our study, the baseline VF defect was significantly worse in the progression group than that in the stable group. The baseline pRNFL(superior, inferior and average), mGCIPL (every section) along with the macular prefusion density (inferior and mean) was significantly lower in the progression group than that in the stable group. However, the optic disc vessel density was lower in the progression group, but did not reach to a significantly level. Two reasons might account for this. First, we excluded patients with the segmentation errors in our study. The potential advantage of OCTA comparing to OCT is that it may be less affected by high myopic anatomical variation of disc because segmentation errors occurred frequently in the high myopic eyes when using OCT to measure the pRNFL thickness. Lee et al.12 did not find that retinal vessel density was superior to the RNFL thickness in the correlation with the visual field sensitivity loss in the high myopic patients without segmentation errors. Second, the number of patients in our progression group might not be enough to be significant different between the two groups.

Meanwhile, we found that the change of the macular vessel density and peripapillary VD was significant more in the progression group, while the RNFL and GCIPL thickness was not. It was reported that highly myopic eyes even without glaucoma showed significantly faster reduction in the macular capillary density but only in the deep capillary plexus instead of the superficial capillary plexus16. In this study, we evaluated only the macular and peripapillary superficial capillary plexus, therefore, high myopia itself associated changes did not influence our results. Hou et al.23 found macular vessel density thinning was faster than GCC thinning and was associated with severity of glaucoma. Their conclusion supported our findings; however, they included only refraction with ± 5.0 diopters(D), which is different from the high myopic population in our study. Shang et al.24 reported the rates of macular vessel density loss were significantly faster in more advanced group than in milder group, which also supported our findings that OCTA measurements could detect vascular deterioration over time at different stages. Besides the macular vessel density change, our study further demonstrated the peripapillary vessel density change was also significant more in the progression group.

In agreement with previous studies, the worse glaucoma severity was a risk factor for glaucoma progression25. Therefore, in our retrospective design, it was reasonable to find that the glaucoma severity was worse in the progression group than that in the stable group. Macular GCIPL and pRNFL was reported useful for monitoring glaucoma progression. Progressive macular GCIPL thinning and parapapillary RNFL thinning are mutually predictive of VF progression13and even development of VF defect in glaucoma suspect26. Progressive mGCIPL is less likely to reach the measurement floor than RNFL thickness in advanced glaucoma27. Monitoring mGCIPL thickness was reported effective for predicting glaucoma progression even in high myopia28. The rate of change in the thickness of mGCIPL and pRNFL was different according to the glaucoma severity, with more amount reduction in milder cases due to their thicker baseline GCIPL and RNFL thickness. In our study, the glaucoma severity was milder in the stable group. That may explain when we analyzed all our patients, the mGCIPL and pRNFL thickness change had no significant difference between the stable and the progression group because the stable group had thicker thickness at baseline. After we further stratified our patients by glaucoma severity into mild group (MD>−6 dB) and moderate to severe group (MD<−6dB), we found in the mild group the decrease of mGCIPL and pRNFL thickness was more in the progression group but did not reach to a significant level. Small sample size might explain for this. Meanwhile, the same in the mild group, we found the decrease of vessel density whether in the peripapillary or in the macular area showed significant more in the progression group than that in the stable group. The significantly more decrease in both the macular and peripapillary vessel density change than the structural change in the progression group may be owing to better correlation with visual filed sensitivity11,12. In other words, we found in the mild high myopic glaucoma, observing vessel density change could be helpful to find out the progressive patients, while the structural change whether mGCIPL or pRNFL thickness change could not. Vessel density change may be a more effective measurement than structural change in monitoring glaucoma progression in the high myopic patients. Macular vessel density was reported to reach a lower floor than GCIPL thickness20. When we looked at the moderate to severe group, it is interesting to find that the change of the superior quadrant peripapillary vessel density instead of macular vessel density was significant more in the progression group. The inferior quadrant peripapillary vessel density was also more in the progression group but did not reach to a significant level. However, there was only 10 patients in the stable and 11 patients in the progression group, the macular vessel density should be further evaluated when more patients are included. Wu et al.29 reported that OCT and OCTA showed limited agreement on event-based progression detection, but both OCT and OCTA detected more progressors than VF. However, they included patients with refraction within ± 5.0 diopters. In this study, we found in the mild highly myopic glaucoma, observing peripapillary or macular vessel density change could be helpful to find out the progressive patients; while the structural OCT change could not. OCT and OCTA may provide different but complementary information about highly myopic glaucoma progression. Our findings may provide insights for future development of built-in OCTA progression analysis software or artificial intelligence assisted detection of glaucoma progression. Further studies are needed to validate if observing serial change of vessel density in high myopic glaucoma patients is helpful to identify progression in advance of the time-consuming VF examination.

The major strengths of the present study include its longitudinal study design and the comprehensive investigation of the rates of macular and peripapillary vessel density changes in glaucomatous eyes with high myopia. The limitations of this study should be noted as well. First, only good-quality OCTA images without segmentation errors were included in this study; 24.5% patients were excluded due to bad image qualities. This may have introduced selection bias and affected the generalizability of the results. Second, because of the retrospective design, the severity was not the same between the stable and the progression groups. Therefore, we adjusted the visual field defect in the mixed model analysis. Moreover, we did the sub-group analysis according to glaucoma severity. Third, the sample size was relatively small in this study because collecting progressive high myopic glaucoma patients with good OCTA image is not easy. Future prospective studies with more study participants are needed to compare the change of OCT and OCTA findings in patients with high myopic glaucoma. Last, because our longitudinal study design, the new version of the AngioPlex(Cirrus, Version 11) analysing the peripapillary vessel density was not installed when we started to collect the baseline examination of our patients. The sector definitions of radial peripapillary capillaries in this study were automatically demarcated by the commercial software of AngioPlex (Cirrus, Version 10). However, the ETDRS grid used in Version 10 was primarily developed for diabetic retinopathy macular evaluation rather than glaucoma evaluation.

In conclusion, the performance of the vessel density changes whether in the macular or disc area was remarkable in differentiating stable or progressive glaucoma with high myopia. These findings offer new insights into the application of OCTA in evaluation glaucoma progression in patients with high myopia.