Factors Associated With Visual Acuity Decline in Glaucoma Patients With Loss of Ganglion Cell Complex Thickness

Abstract

Purpose: To investigate the association between ocular/systemic factors and visual acuity decline in glaucoma patients with loss of ganglion cell complex thickness (GCCT).

Methods: In 515 eyes of 515 patients with open-angle glaucoma (mean age, 62.6 ± 12.8 years; mean deviation, −10.95 ± 9.07 dB), we used swept-source optical coherence tomography to measure macular GCCT in sectors classified as corresponding to circumpapillary retinal nerve fiber layer clock-hour sectors from 7 o'clock (inferotemporal) to 11 o'clock (superotemporal). We calculated Spearman's rank correlation coefficient between each sector and best-corrected visual acuity (BCVA), determined cutoff values for BCVA decline (<20/25), and used multivariable linear regression models to determine the correlation between BCVA and biological antioxidant potential (BAP), corneal hysteresis (CH), and temporal-tissue optic nerve head blood flow (represented by temporal mean blur rate, or MBR-T).

Results: Macular GCCT corresponding to the 9 o'clock sector had the highest correlation with BCVA (R_s = −0.454; P < 0.001) and a cutoff of 76.17 µm (area under the receiver operating characteristic curve = 0.891; P < 0.001). Subjects below this cutoff (N = 173) showed significant correlations between BCVA and age, BAP, CH, and MBR-T (β = 0.192, P = 0.033; β = −0.186, P = 0.028; β = −0.217, P = 0.011; and β = −0.222, P = 0.010, respectively).

Conclusions: Multiple factors are involved in BCVA decline in patients with glaucoma with decreased macular GCCT. This suggests that evaluating BCVA may require assessing multiple factors.

Translational Relevance: Multiple factors contribute to BCVA decline.

Introduction

Glaucoma is the world's second most common cause of blindness, and many patients suffer from irreversible visual dysfunction.^1,2 The pathology is due to retinal ganglion cell (RGC) death caused by damage to the lamina cribrosa. In the early to moderate stages of glaucoma, these changes can be detected as localized defects in the retinal nerve fiber layer (RNFL) and subsequent thinning of the RNFL and ganglion cell complex (GCC).³ Generally, visual acuity dysfunction, which directly affects quality of life (QOL), may not occur until the late stages of glaucoma; however, it is possible for visual acuity decline to occur depending on the location of RNFL defects. Also, visual field defects occur more often near the fixation point in patients with normal-tension glaucoma (NTG).^4,5 This is an especially important point to consider in populations with a high incidence of NTG, such as those in Asia.^6,7 At any stage of glaucoma, it is important to monitor central visual function to maintain QOL. 

The advent of optical coherence tomography (OCT) enabled reproducible and objective measurement of the thickness of the RNFL and the GCC, both of which become thinner as glaucoma progresses. The latest OCT technologies—such as swept-source OCT (SS-OCT)—have faster scanning speeds and provide images with good accuracy and reproducibility.⁸ Therefore, OCT is a useful tool for evaluating the visual acuity of glaucoma patients objectively. 

We have previously reported a method for efficiently evaluating visual acuity based on GCC thickness in the papillomacular bundle, as measured with SS-OCT.^9,10 However, the relationship between visual acuity and GCC thickness is not linear, and there is a “floor” past which further thinning does not result in additional loss of visual acuity. In addition, glaucoma is a multifactorial disease, and various risk factors, including systemic factors such as blood pressure¹¹ and oxidative stress,⁶ have been reported to contribute. Previously, we reported that aging,¹² high pulse rate,¹² high blood pressure,¹³ and low biological antioxidant potential (BAP)^14,15 are systemic risk factors for glaucoma. It is likely that these risk factors are also involved in visual acuity decline in patients with glaucoma. Therefore, we believe that it is necessary to consider a variety of factors when evaluating the visual acuity of patients with loss of GCC thickness. 

In this study, we focused on a localized macular area that is closely associated with visual acuity by classifying sectors corresponding to clock-hour circumpapillary retinal nerve fiber layer (cpRNFL) sectors (from 7 o'clock to 11 o'clock). Then we investigated various ocular/systemic factors that may contribute to the decline in visual acuity observed in this area. We identified factors that are associated with visual acuity disturbance other than GCC thickness. 

Methods

Subjects

This study included 515 eyes of 515 patients with primary open-angle glaucoma (POAG) or NTG, including preperimetric glaucoma (PPG). The initial diagnosis was made by a glaucoma specialist (T.N.) between February 2019 and September 2021 at Tohoku University Hospital. Glaucoma was defined in this study as the presence of an abnormal glaucomatous optic disc (with diffuse or focal thinning of the neuroretinal rim) and corresponding glaucomatous visual field defects, defined by the Anderson–Patella criteria¹⁶ as the presence of one or more of the following: (1) a cluster of three points with reduced sensitivity at a probability of <5% on the pattern deviation map in at least one hemifield (including ≥1 point at a probability of <1% or a cluster of two points at a probability of <1%), (2) glaucomatous hemifield test results outside the normal limits, and (3) a pattern standard deviation beyond 95% of normal limits, as confirmed in at least two reliable examinations. NTG was diagnosed if peak IOP was 21 mmHg or less without any glaucoma medication, and POAG was diagnosed if peak IOP was more than 21 mmHg. Eyes with cataract (i.e., a lens nucleus with a grade of 3 or more in the Emery–Little classification, posterior subcapsular cataracts, or cortical cataracts), any ocular surface disease, any corneal disease, any vitreoretinal or optic nerve diseases other than glaucoma, or a history of intraocular surgery other than cataract surgery were excluded from the current study. When both eyes of a patient were eligible, the eye with worse visual acuity was included. Comorbidities were self-reported for this study. 

This study adhered to the tenets of the Declaration of Helsinki. The protocols were approved by the Clinical Research Ethics Committee of the Tohoku University Graduate School of Medicine (Study 2021-1-430). 

Measurement of Ocular Variables

Best-corrected visual acuity (BCVA) was measured with a standard Japanese decimal visual acuity chart and converted to the logarithm of the minimum angle of resolution (logMAR). Intraocular pressure (IOP) was measured with Goldmann applanation tonometry. Axial length was measured with ocular biometry (OA-2000; Tomey Corp., Nagoya, Japan). Central corneal thickness was measured with anterior-segment OCT (Casia2; Tomey Corp.). Optic nerve head blood flow was assessed with laser speckle flowgraphy (LSFG-NAVI; Softcare Co., Ltd., Fukutsu, Japan), which measures mean blur rate (MBR) in arbitrary units (AUs). In the current study, we used the temporal tissue-area MBR (MBR-T) to assess capillary blood flow. Corneal hysteresis (CH) was measured with a specialized tonometer (Ocular Response Analyzer; Reichert Technologies, Depew, NY). Mean deviation (MD) was measured with the Humphrey Field Analyzer (HFA; Carl Zeiss Meditec, Dublin, CA) with the Swedish Interactive Threshold Algorithm (SITA)-standard strategy of the 24-2 program. Only reliably measured MD values were used (<20% fixation errors, <15% false-positive results, and <33% false-negative results). All of the examination data were obtained within a 2-month period, without interrupting the use of medication for glaucoma. 

Measurement of Clinical Variables

Clinical parameters for each patient were recorded, including age, gender, and body mass index (BMI). BMI was calculated with the following formula: weight (kg)/height² (m²). Blood pressure and pulse rate were conventionally measured in the brachial artery at the height of the heart (HBP-1300; Omron Colin Co., Ltd., Tokyo, Japan). Diacron reactive oxygen metabolites (dROMs) and BAPs were measured with a Free Carpe Diem (Wismerll Co., Ltd., Tokyo, Japan) after at least 3 hours of fasting. 

OCT Examination

Macular GCC thickness (GCCT) and cpRNFL thickness (cpRNFLT) were measured in a 12 mm × 9-mm-wide scan with SS-OCT (DRI OCT Triton; Topcon Corp., Tokyo, Japan). Macular GCCT was measured in a 6-mm × 6-mm area of the wide scan and classified into sectors corresponding to the clock-hour sectors of the cpRNFL, ranging from 7 o'clock (the inferotemporal area) to 11 o'clock (the superotemporal area). The sectors were calculated from 10 × 10 grids laid over the macular scans, as previously reported,¹⁷ with slight modification for SS-OCT (Fig. 1). 

Figure 1.

 

Division of each macular grid point (in a right eye). A 10 × 10 grid overlaid on the macular GCC was classified into sectors corresponding to each clock-hour sector of the cpRNFL from 7 o'clock (the inferotemporal area) to 11 o'clock (the superotemporal area).

Statistical Analysis

The correlation coefficient between sectoral thickness and visual acuity was calculated with Spearman's rank correlation coefficient. Logistic regression analysis was performed to calculate the receiver operating characteristics area under the curve (ROC-AUC) to evaluate the diagnostic ability of specific factors to predict visual acuity decline (<20/25).^18,19 Cutoff values were calculated with the Youden index. AUCs were compared with the Delong test. The correlation coefficients between clinical/ocular parameters and BCVA were calculated with univariable and multivariable linear regression models. The multivariable regression analysis was performed using the significant parameters revealed in the univariable linear regression analysis (P < 0.05) and was further adjusted for the number of medications and the stage of glaucoma (mild, MD > −6 dB; moderate–severe, MD ≤ −6 dB). The analyses used JMP 16.2.0 (SAS Institute Japan, Tokyo, Japan) or R 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria). P < 0.05 was considered statistically significant. 

Results

Tables 1 and 2 show the systemic and ocular characteristic of the subjects (age = 62.6 ± 12.8 years; female/male = 254/261; MD = −10.95 ± 9.07; NTG/POAG/PPG = 357/117/41). 

Table 1.

 

Patient Systemic Characteristics

 

Table 2.

 

Patient Ocular Characteristics

 

Figure 2 shows scatterplots comparing clock-hour cpRNFLT sectors and corresponding macular sectors. All of the sectors had high R² values. (7 o'clock, 0.68; 8 o'clock, 0.66; 9 o'clock, 0.58; 10 o'clock, 0.73; 11 o'clock, 0.56; all P < 0.001). 

Figure 2.

 

Scatterplots showing the relationship between cpRNFLT in clock-hour sectors and macular GCCT in corresponding sectors: (A) 7 o'clock; (B) 8 o'clock; (C) 9 o'clock; (D) 10 o'clock; (E) 11 o'clock. R² is the coefficient of determination.

Table 3 shows Spearman's rank correlation coefficient for BCVA and GCCT in each macular sector. All of the sectors showed a significant correlation to BCVA, especially the 9 o'clock sector, which is located temporally (R_s = −0.454, P < 0.001). As shown in Table 4, we determined cutoff values to predict visual acuity decline (<20/25). Among the macular sectors, 9 o'clock GCCT showed the highest diagnostic ability for visual acuity decline (AUC = 0.891; 95% confidence interval [CI], 0.853–0.929; P < 0.001) with a cutoff value of 76.17 µm. 

Table 3.

 

Spearman's Rank Correlation Coefficient (R_s) Between BCVA and Macular GCCT in Each Sector

 

Table 4.

 

ROC-AUC and Cutoff Values for GCCT to Predict BCVA Decrease (<20/25) in Each Macular Sector

 

Figure 3 shows a scatterplot of the relationship between GCCT in the 9 o'clock sector and BCVA. There were few cases of BCVA decline in patients with GCCT above the cutoff value (76.17 µm). However, in patients with GCCT below the cutoff value, there was a mix of cases with poor and good visual acuity. 

Figure 3.

 

Scatterplot showing the relationship between BCVA and GCCT in the GCCT sector that corresponded to 9 o'clock cpRNFLT. The cutoff value to predict BCVA decline was 76.17 µm (logistic regression analysis).

Table 5 shows the correlation between BCVA and clinical/ocular parameters in 173 patients (NTG/POAG = 125/48) with GCCT in the 9 o'clock sector below the cutoff value (i.e., ≤76.17 µm). Overall, the multivariable regression analysis showed that age, BAP, CH, and MBR-T were significantly correlated to BCVA (β = 0.192, P = 0.032; β = −0.186, P = 0.028; β = −0.217, P = 0.011; and β = −0.222, P = 0.010, respectively). In the NTG subjects, the multivariable regression analysis showed that BCVA was significantly correlated to age, CH, and MBR-T (β = 0.258, P = 0.016; β = −0.217, P = 0.042; and β = −0.260, P = 0.019, respectively). In the POAG subjects, BMI and MBR-T were significantly correlated to BCVA in the univariable analysis (β = −0.293, P = 0.046 and β = −0.341, P = 0.019, respectively), but not the multivariable analysis (β = −0.104, P = 0.412 and β = −0.190, P = 0.167, respectively). BAP and CH were significantly correlated to BCVA in the multivariable regression analysis (β = −0.389, P = 0.022 and β = −0.311, P = 0.019, respectively). In addition, we performed a severity-stratified analysis (Supplementary Table S1). In the early stage (MD > −6 dB), age was significantly correlated with BCVA (β = 0.440, P = 0.037). On the other hand, in the moderate–severe stage (MD ≤ −6 dB), CH and MBR-T were significantly correlated to BCVA in the multivariable regression analysis (β = −0.228, P = 0.009 and β = −0.226, P = 0.013, respectively). 

Table 5.

 

Factors Associated With Visual Acuity Below the Cutoff Value for Visual Acuity Decline in 9 O'Clock GCCT

 

Figure 4 shows a comparison of ROC-AUCs calculated with a crude model and an adjusted model to predict visual acuity decline (<20/25) in the macular GCCT sector corresponding to the 9 o'clock sector of the cpRNFL. The adjusted model was adjusted for age, BAP, CH, and MBR-T. The AUC for the adjusted model (0.908; 95% CI, 0.864–0.952) was significantly higher than for the crude model (0.891; 95% CI, 0.853–0.929; P = 0.034). 

Figure 4.

 

ROC curves to predict visual acuity decline based on GCCT in the macular sector corresponding to the 9 o'clock sector of the cpRNFL. The red line shows results from a crude model (AUC = 0.891; 95% CI, 0.853–0.929), and the blue line shows results from a model adjusted for age, BAP, CH, and optic nerve head blood flow in the temporal optic nerve head tissue (AUC = 0.908; 95% CI, 0.864–0.952). The AUC of the adjusted model was significantly higher than that for the crude model (P = 0.034).

Discussion

In the current study, we classified the macula into sectors corresponding to the clock-hour sectors of the cpRNFL from 7 o'clock (the inferotemporal area) to 11 o'clock (the superotemporal area) and found that macular GCCT in the 9 o'clock sector showed the highest correlation coefficient to BCVA (R_s = −0.454; P < 0.001). The scatterplot in Figure 3 shows that there were many cases with good visual acuity in the reduced GCCT group. To further investigate this finding, we decided to focus exclusively on cases with GCCT loss in the 9 o'clock sector and to explore various clinical and ocular parameters that are associated with visual acuity decline. Interestingly, the results showed that age, BAP, CH, and MBR-T were significantly associated with BCVA, and the ROC-AUC significantly improved (from 0.819 to 0.908) after adjusting for these risk factors. Thus, our results suggest that various parameters other than GCCT might be involved in visual acuity disturbance. 

Visual acuity is primarily determined by the density of RGCs and their ability to detect spatial frequencies. The fovea, where RGCs are most densely packed, is capable of detecting details finer than can be resolved by the lens and cornea. Therefore, some redundancy exists in this part of the visual system, and visual acuity may not decrease significantly if RGC density is only slightly reduced.^20,21 Nevertheless, when density falls below a certain threshold, visual acuity decreases. Although visual acuity decline can be caused by various factors, such as abnormalities in the photoreceptors, cataracts, or corneal disease, glaucomatous visual acuity decline depends specifically on the RGCs in the macula and their axons, which make up the papillomacular bundle. Interestingly, visual acuity may remain good in some cases even when the GCC becomes thinner, whereas in other cases it may deteriorate.²² This nonlinear relationship between structure and function highlights the need to consider factors other than GCCT in assessing glaucomatous damage and visual acuity. 

In this study, we focused on cases in which the loss of GCCT had occurred and investigated the correlation between various ocular and systemic parameters and visual acuity. The parameters were selected based on various factors that we have previously reported that affect visual acuity and that we routinely measure in our outpatient clinic, including aging,¹² corneal thickness,¹³ CH,¹⁵ ocular blood flow,¹² blood pressure,¹³ pulse rate,¹² and oxidative stress, as represented by dROM level and BAP.^14,23 The results showed that age, BAP, CH, and MBR-T were significant parameters even after conducting a multiple linear regression analysis. These factors have been reported as glaucoma risk factors. Aging is considered an important risk factor for glaucoma, and it is believed that one of the mechanisms involves the accumulation of mitochondrial dysfunction, which is thought to be a cause of RGC death.²⁴ BAP is a biomarker of antioxidant capacity, and it has been suggested that lower BAP levels are correlated with worse visual field sensitivity²⁵ and decreased RGC count.¹⁴ CH, which reflects the ability of corneal tissue to absorb and release energy during bidirectional flattening, is related to glaucoma progression,²⁶ progressive RNFL loss,²⁷ and anterior displacement of the LC²⁸ when CH levels are low. Low blood flow has also been reported to be associated with the progression of glaucoma,²⁹ particularly in the temporal sector (i.e., MBR-T), where it has been suggested that reduced blood flow precedes thinning of the RNFL in elderly patients or those with a high pulse rate.¹² In the patients below the cutoff value for GCCT, we observed no significant correlations among these parameters (data unpublished), even though they are all considered relevant to glaucoma progression. This suggests that various mechanisms are involved in the visual acuity disturbance associated with glaucoma, making it difficult to explain the cause of visual acuity decline in a unified manner. 

In the model shown in Table 5, the β values for each parameter are not high, and the P values are not particularly low, so it is unlikely that all events can be explained by these parameters alone. However, by including these factors in a logistic model, we were able to significantly improve the AUC for visual acuity decline (Fig. 4). Importantly, these are factors other than GCC thickness that are related to decreased visual acuity, and they should also be evaluated in glaucoma treatment. Findings in the severity-stratified analysis (Supplementary Table S1) have clinical implications, particularly for moderate to severe glaucoma: The significance of the results on thickness suggests that, as glaucoma worsens, measurements of thickness might be less relevant to visual acuity than systemic factors, thus strengthening the importance of early intervention. Additionally, the significant parameters differed between NTG and POAG patients even when they were examined separately by disease type. In the future, a study should be performed with patient stratification to assess the risk of glaucoma in individual cases, considering both ocular and systemic factors. 

Our results did not show a significant association between blood pressure and the study outcomes. Information on coexisting conditions, including hypertension, relied on patient self-reports, and a considerable number of patients had already undergone therapeutic intervention. As a result, it would have been challenging to construct a model that took into account disease severity and duration. However, MBR-T emerged as a significant factor associated with visual acuity, suggesting its potential utility in evaluating vascular risk factors. 

We previously described a method of dividing the macular map into a grid pattern and dividing the grid into sectors corresponding to the 7 o'clock to 11 o'clock positions of the cpRNFL. As expected, this study showed that the area corresponding to the 9 o'clock sector of the cpRNFL had the highest correlation with visual acuity. Therefore, when evaluating visual acuity based on a macular map, it may be better to focus on localized changes rather than the entire averaged area. In addition, we also previously reported a method for efficiently evaluating visual acuity based on examining GCCT in a section of the vertical line passing through the disc/fovea midpoint.⁹ Using this method, we identified a narrow band in the RNFL that was closely associated with good visual acuity.¹⁰ This RNFL band was located slightly above the disc/fovea line, which is also consistent with our current findings, as the macular 9 o'clock sector is located slightly above the horizontal line of the fovea. 

The present study found that patients with GCCT loss sometimes had poor visual acuity and sometimes had good visual acuity, which is consistent with previous studies. The structure–function relationship in perimetry is widely accepted to have a “floor effect”: As glaucoma progresses, the RNFL becomes thin, but the progression of this thinning (i.e., RNFL reduction) stops at a certain point. ³⁰ Similarly, a nonlinear relationship has been reported between macular GCCT and visual function, suggesting the presence of a floor effect similar to that in the RNFL.^31,32 Findings from basic research can explain this phenomenon. Wang et al.³³ studied the retinas of organ donors who had glaucoma and found that they had significantly more glial cells in the nerve fiber layer than healthy controls. Animal models of high intraocular pressure also show glial proliferation in the optic nerve head.³⁴ These findings indicate that, as the RNFL becomes thinner, glial cells proliferate, and the apparent thickness of the RNFL reaches a floor and does not decrease. It is important to note that these reports focused on the optic nerve head or the cpRNFL, whereas the current study targeted the macular GCC. Thus, further investigation is needed to clarify whether glial proliferation also occurs in the macular GCC. 

We acknowledge that there were several limitations in this study. It had a small number of patients, all of whom were of a single ethnicity, and it was single center, cross-sectional, and retrospective. Furthermore, OCT measurements of RNFLT and GCCT can be affected by OCT image quality. This can create a bias toward thickness values that are lower than the real values. To prevent this issue, we excluded images with image quality (IQ) of less than 30.³ The macular sectors may not account for anatomical differences, such as disc tilt or rotation, that may lead to individual variation. However, macular sector differences due to disc tilt have been reported to have no substantial effect on the reproducibility of the measurements.³⁵ Additionally, the potential impact of cataracts on visual acuity decline cannot be completely dismissed. However, only a few cases of mild cataracts resulting in a decline in visual acuity have been reported.³⁶ In our study, we defined visual acuity decline as 20/25, which is considered a relatively mild decline, and the AUC for visual acuity decline was high. This indicates that the effect of mild cataracts was not substantial. 

In conclusion, this study used classification of the macula into sectors that corresponded with clock-hour sectors of the cpRNFL from 7 o'clock to 11 o'clock to show that macular 9 o'clock GCCT had the highest correlation with BCVA. Moreover, we found that multiple factors were associated with visual acuity decline in glaucoma patients with GCCT loss. Thus, our results suggest that research on the structure–function relationship in the retina requires not only measurement of retinal thickness but also evaluation of various factors, such as aging, ocular blood flow, oxidative stress, and CH. 

Acknowledgments

The authors thank Tim Hilts for editing this manuscript and Takehiro Miya of the Tohoku University Graduate School of Medicine for technical support. 

Supported in part by a Japan Science and Technology grant from the Japan Society for the Promotion of Science Kakenhi Grants-in-Aid for Scientific Research (20H03838 to TN), by the Japan Science and Technology Center for Revitalization Promotion and Kakenhi Grants-in-Aid for Scientific Research (21K09690 to KO), by COI-NEXT (JPMJPF2201), and by a Kitazawa Yoshiaki Glaucoma Research Award. 

Disclosure: N. Takahashi, None; K. Omodaka, None; T. Kikawa, Topcon (E); T. Ninomiya, None; N. Kiyota, None; S. Tsuda, None; N. Himori, None; T. Nakazawa, None