A Decrease in Bruch’s Membrane Opening-Minimum Rim Area Precedes Decreased Retinal Nerve Fiber Layer Thickness and Visual Field Loss in Glaucoma

Précis: A decrease in Bruch’s membrane opening-minimum rim area, which represents the optic nerve head (ONH), preceded a decrease in the peripapillary retinal nerve fiber layer thickness (RNFLT) and the visual field index (VFI). Purpose: This study aimed to investigate the relative comparison between a decrease in BMO-MRA, the peripapillary RNFLT, and the VFI, according to the severity of glaucoma. Materials and Methods: This retrospective cross-sectional study included 121 eyes (73 with open-angle glaucoma and 48 normal eyes). The ONH and retinal nerve fiber layer were analyzed using spectral domain optical coherence tomography, and VFI was obtained using the Humphrey Field Analyzer. The tipping points of RNFLT for VFI and BMO-MRA were estimated using broken-stick regression models. Polynomial regression analysis was performed, and the changes in the 3 parameters were expressed as a graph. Results: The tipping point of the RNFLT for the VFI was 88.62 μm [95% confidence interval (CI): 79.59-97.65; P=0.001]. The tipping point of the RNFLT for BMO-MRA was 60.00 μm (95% CI: 48.28-71.72; P=0.220). Above the tipping point, BMO-MRA decreased with a decrease in the RNFLT (slope=0.0135; 95% CI: 0.0115-0.0155; P<0.001); below the tipping point, BMO-MRA did not decrease significantly (slope=0.0002; 95% CI: −0.0177 to 0.0181; P=0.983). Polynomial regression analysis showed that with the progression of glaucoma, BMO-MRA decreased more rapidly, and this preceded a decrease in the RNFLT followed by a decrease in the VFI. Conclusion: The ONH parameter, BMO-MRA, showed a faster decrease than RNFLT and VFI in early glaucoma. BMO-MRA may help detect early glaucomatous damage and its progression.

G laucoma is a degenerative optic neuropathy characterized by retinal ganglion cell loss that leads to characteristic changes in the optic nerve head (ONH) and retinal nerve fiber layer (RNFL). Since these structural changes cause irreversible functional deterioration of vision, an accurate understanding of the relationship between ONH, RNFL, and visual field (VF) is crucial for diagnosing early glaucoma and detecting its progression.
It has been reported that structural changes in the ONH and RNFL may precede loss of visual function and can be used to detect disease progression and preserve vision in glaucoma patients. [1][2][3][4][5][6] In addition, it also important to understand the differences between ONH and RNFL changes in glaucoma progression for early glaucoma diagnosis and in understanding the pathophysiology of glaucoma.
Bruch's membrane opening (BMO)-based parameters in spectral domain optical coherence tomography (SD-OCT) have been introduced as new structural measurements for analyzing ONH. 7,8 Among them, BMO-minimum rim area (BMO-MRA), a 2-dimensional parameter, has been shown to provide benefits compared with 1-dimensional parameters for analyzing the neuroretinal rim tissue by SD-OCT. 7,[9][10][11] In this study, we analyzed the BMO-MRA, peripapillary retinal nerve fiber layer thickness (RNFLT), and the visual field index (VFI) in healthy subjects and patients with various severities of glaucoma to evaluate the relative comparison between ONH, RNFL, and VF according to glaucoma severity.

MATERIALS AND METHODS
In this cross-sectional study, we retrospectively reviewed the medical records of patients who visited the Department of Ophthalmology, Kangdong Sacred Heart Hospital, between March 2017 and February 2018. Subjects were enrolled if they were either healthy or open-angle glaucoma patients aged 18 years or older. This study was approved by the Institutional Review Board of Kangdong Sacred Heart Hospital and adhered to the tenets of the Declaration of Helsinki.
Healthy subjects were defined as those with a bestcorrected visual acuity 20/40 or better, IOP < 21 mm Hg; open angles on gonioscopy; absence of glaucomatous optic disc appearance, disc hemorrhage, and RNFL defects on RNFL photography; normal RNFLT on SD-OCT; and normal VF. Subjects with glaucoma suspect were excluded.
Patients with open angles on gonioscopy, glaucomatous optic neuropathy, and VF defects were diagnosed as open-angle glaucoma. 12 Glaucomatous optic neuropathy was defined as neuroretinal rim thinning, notching, or generalized loss of the neuroretinal rim on color disc photography. Glaucomatous VF defects were considered to be present by confirmation of the following criteria by at least 2 reliable examinations: (1) a cluster of 3 points with < 5% probability on the pattern deviation map in at least 1 hemifield, with at least 1 point with <1% probability; (2) glaucoma hemifield test results outside normal limits; and (3) a pattern SD outside 95% of the normal limits.
The exclusion criteria were the presence of corneal, lens, or vitreous opacity that could affect fundus photography or VF test, SD-OCT image quality scores < 20, SD-OCT segmentation error in RNFL and ONH analysis, history of retinal disease including diabetic or hypertensive retinopathy, patients with VF defects caused by nonglaucomatous optic neuropathy or neurological disease, history of previous trauma or ocular surgery except uncomplicated cataract surgery, and high myopia with > −6.0 diopter or axial length > 26 mm. For those who met the inclusion criteria for both eyes, only 1 eye was selected for inclusion in the study.
Standard automated perimetry was performed using the 24-2 Swedish interactive threshold algorithm standard strategy, and all participants had been subjected to nearrefractive correction. VF tests were performed in all patients within 6 months after the SD-OCT examination. The evaluations of the VF test were considered reliable if the falsepositive and false-negative errors were <15% and if the fixation loss was < 20%.
The ONH and RNFL were analyzed using SD-OCT, and the global RNFLT was obtained automatically. A scan pattern was centered on the BMO with 24 radially equidistant B-scans. Each B-scan delineated the rim area using a circular sector of 7.5 degrees; a total of 48 rim measurements were obtained. The BMO and inner limiting membrane (ILM) were automatically segmented and confirmed by an experienced ophthalmologist (K.I.N.) to rule out segmentation errors. BMO-MRW was defined as the minimal distance from the BMO point to the ILM within each B-scan image. BMO-MRA was calculated as the total sum of the areas of the 48 trapeziums in the minimum surface area between the BMO and ILM, at an angle θ above the BMO plane. The height of the trapezium was set to equal the rim width (W) at this angle. The base of the trapezium was measured by the circumference length of the BMO in each sector, that is, 2πr/48, where r is defined as the length from the BMO centroid to each BMO point. The top of the trapezium was measured using the following formula: 2π/48*(r −W*cos(θ)). The area of each trapezium was calculated using the following formula 7 : Area = (2πr/48+2π/48*(r-W* cos(θ)))/2×W.
Clinical characteristics of normal and glaucomatous eyes were compared using the Student t test for continuous variables and the χ 2 test for categorical variables. Scatter plots of VFI, RNFLT, and BMO-MRA were used to determine the relationship between the parameters. A series of broken-stick regression models were used to fit these scatter plots and estimate the tipping point. The tipping point was determined using Davies' test 13 and the above and below slopes were reported with their 95% confidence intervals (CIs).
To represent the relationship between VFI, BMO-MRA, and RNFLT, the scale of VFI and BMO-MRA was adjusted based on the subject groups between 1 and 2 SD and −1 to −2 SD from the mean of RNFLT. Then, polynomial regression analysis (third-order) of VFI and BMO-MRA with RNFLT was performed to fit the scatter plots demonstrated on the same graph.
All statistical analyses were performed using R language software (http://www.R-project.org) and the segmented R library. Statistical significance was set at P < 0.05.

RESULTS
A total of 121 eyes (48 normal and 73 glaucomaaffected) were included in the study. The demographic and clinical characteristics of the study participants are summarized in Table 1.
The VF parameters (mean deviation, pattern SD, and VFI) for glaucoma patients were significantly worse than those of normal subjects (P < 0.001). The RNFLT was significantly thinner in the glaucoma group than the healthy group (72.16 ± 15.16 vs. 101.44 ± 9.60, P < 0.001). The BMO-MRA also showed a significant difference  Figures 1 and 2 show scatter plots of VFI and BMO-MRA with RNFLT and broken-stick models fitted to the data. The tipping points for VFI and BMO-MRA were 88.62 μm (95% CI: 79.59-97.65; P = 0.001) and 60.00 μm (95% CI: 48.28-71.72; P = 0.220), respectively, and these points were statistically determined using Davies' test (Tables 2, 3). Using the mean RNFLT values for the healthy subjects in our study, the percentages of RNFLT loss required to reach the tipping points were also calculated for VFI and BMO-MRA.
The tipping points for VFI and BMO-MRA occurred after a 12.63% and 40.85% loss from the mean normative RNFLT, respectively.
The slopes for VFI and BMO-MRA with the corresponding RNFLT above and below the tipping points and the differences between the slopes are summarized in Tables 2 and 3 (Fig. 2). Figure 3 shows scatter plots of VFI with BMO-MRA and broken-stick models fitted to the data. The tipping points for VFI were 1.00 mm 2 (95% CI: 0.92-1.07; P < 0.001) ( Table 4). Using the mean BMO-MRA values for the healthy subjects in our study, the percentages of BMO-MRA loss required to reach the tipping point were also calculated for VFI. The tipping points for VFI occurred after a 20.95% loss from the mean normative BMO-MRA.
The slope for VFI with the corresponding BMO-MRA above and below the tipping points and the differences between the slopes are summarized in Table 4. Above the tipping point, VFI showed a plateau regardless of the decrease in the BMO-MRA (slope = 8.72; 95% CI: −9.56 to 27.01; P = 0.346). In contrast, below the tipping point, the VFI was significantly related to the BMO-MRA (slope = 113.22; 95% CI: 90.65-135.79; P < 0.001) (Fig. 3).
We represent the relationship between VFI, BMO-MRA, and RNFLT according to the extent of glaucomatous damage. Polynomial regression analysis (third-order) of VFI and BMO-MRA with RNFLT were conducted separately (adjusted R 2 = 0.653 and 0.705, respectively) (Fig. 4). It can be seen from the graph that as the severity of glaucoma increased, BMO-MRA first decreased more rapidly and then decreased in the order of RNFLT and VFI.

DISCUSSION
Glaucoma is the second leading cause of blindness worldwide. 14 Efforts have been made to understand the relationship between structural changes and functional visual loss in order to preserve vision in patients with glaucoma. It has been reported that considerable retinal ganglion cell loss or RNFL defects obtained using SD-OCT were observed before VF loss was detectable and was associated with locations of functional impairment. [15][16][17][18][19][20] In previous studies, the broken-stick model was used to determine the threshold values of RNFLT and BMO-MRW according to VF defects. [21][22][23][24][25] Wollstein et al 25 determined the RNFLT at which VF loss becomes detectable, and 17.3% of RNFLT loss appears to be necessary for the detection of functional loss. Similarly, Alasil et al 21 found that a 8.4% loss in global RNFLT was necessary for detecting VF damage.
Recently, BMO-MRW was introduced as a new structural parameter that represents the minimum distance from the BMO to the ILM and offered a more accurate evaluation of the neuroretinal rim and better diagnostic capability than traditional disc parameters. [26][27][28][29] The structure-function relationship between BMO-MRW and functional defects was explored instead of using traditional RNFLT. Park et al 24 reported that VF defects occurred after a 25.9% loss in global BMO-MRW, and Li et al 23 determined the tipping points of global BMO-MRW loss as 14.9% in OAG, 9.3% in NTG, and 17.2% in POAG groups, respectively.
Moreover, Gardiner and colleagues introduced the 2-dimensional parameter BMO-MRA, which has been shown to provide benefits compared with 1-dimensional parameters analyzing the neuroretinal rim tissue by SD-OCT. 7,[9][10][11] In this study, we used BMO-MRA instead of BMO-MRW as the ONH parameter.
We explored the relationship between ONH, RNFL, and VF using the broken-stick model to determine the tipping point of global RNFLT according to BMO-MRA and VFI, and the tipping point of BMO-MRA according to VFI. WE found that 12.42% of RNFLT loss was needed to detect functional VF defects, which was in similar ranges to the RNFLT tipping points reported in previous studies. 21,25 The BMO-MRA decreased rapidly as glaucoma severity increased, but there was no significant decrease after 40.85% loss of global RNFLT. The tipping point of BMO-MRA according to VFI was 20.95%, which is higher than that of RNFLT, suggesting that BMO-MRA shows a relatively large change in the early stage of glaucoma.
BMO-MRA decreases faster than RNFLT and reaches a floor earlier can be understood by considering the characteristic differences between the 2 parameters. Patel et al 30 increased the IOP of 6 monkeys from 10 to 60 mm Hg in 10 mm Hg steps every 10 minutes and observed changes in the ONH and RNFLT with SD-OCT. The BMO-MRW decreased with an increase in IOP, but RNFLT did not change significantly. In addition, when the IOP was lowered to 10 mm Hg, BMO-MRW was thinner than that at baseline. Sharma et al 31 reported that BMO-MRW significantly decreased when IOP   32 At the beginning of the glaucomatous damage to the eye, it can be assumed that the ONH change occurs first because of neural rim compression, although the RNFL has not changed yet.
To compare the changes in BMO-MRA, RNFLT, and VFI together, we had to express the three parameters in one graph; therefore, we adjusted the scales of BMO-MRA and VFI based on RNFLT. The graph was then drawn using polynomial regression analysis. When comparing the damage of BMO-MRA, RNFLT, and VFI according to glaucoma progression in early glaucoma, BMO-MRA shows a curve that decreases more rapidly than RNFLT. These findings suggest that BMO-MRA may be a more sensitive indicator of early glaucomatous damage than RNFLT.
Meanwhile, VFI decreased more slowly than BMO-MRA and RNFLT in early glaucoma. This reconfirms the commonly known statement that structural damage precedes functional damage in glaucoma. In contrast, in advanced glaucoma, the decrease in VFI was more rapid than that of RNFLT, and BMO-MRA showed the least change.
A limitation of this study is that it was a cross-sectional study. Better research methods to analyze how each glaucoma parameter changes with glaucoma progression would include long term in glaucoma patients. However, it was