Associations Between Visual Functions and Severity Gradings, Corneal Scatter, or Higher-Order Aberrations in Fuchs Endothelial Corneal Dystrophy

Abstract

Purpose: The purpose of this study was to investigate the associations between visual function and severity grading, corneal scatter, or higher-order aberrations (HOAs) in patients with Fuchs endothelial corneal dystrophy (FECD).

Methods: This observational case series study included 49 eyes of 27 patients with FECD and 10 eyes of 10 healthy individuals. We evaluated corrected distance visual acuity (CDVA) using Landolt-C and Early Treatment Diabetic Retinopathy Study charts and contrast sensitivity using the CSV-1000E chart and CSV-1000RN letter chart. We analyzed the associations between visual function and explanatory variables, including age, modified Krachmer grade, central corneal thickness (CCT), anterior segment optical coherence tomography (AS-OCT)-based grade, HOAs, intraocular straylight, and corneal densitometry. We additionally conducted receiver operating characteristic (ROC) analysis to identify the corneal densitometry thresholds for decreased visual function.

Results: There were significant associations between visual function and the modified Krachmer grade, CCT, AS-OCT-based grade, HOAs, intraocular straylight, and corneal densitometry. A modified Krachmer grade ≥ 3 was identified as a threshold for decreased visual function. Multivariate analysis showed that corneal densitometry was significantly associated with all visual function parameters, and HOAs were significantly associated with CDVA but not with contrast sensitivity. ROC analysis revealed that corneal densitometry of the posterior layer at 0 to 2 mm ≥ 10 grayscale units (GSU), was identified as a threshold for decreased visual function.

Conclusions: HOAs, forward and backward light scatter affected visual function, with backward light scatter being the most influential. In patients with FECD, modified Krachmer grade ≥ 3 and corneal densitometry ≥ 10 GSU were thresholds for visual disturbance.

Fuchs endothelial corneal dystrophy (FECD) is a bilateral disorder characterized by abnormal deposition of the extracellular matrix, thickening of the Descemet's membrane, and the progressive loss of corneal endothelial cells, which leads to corneal edema and vision impairment.^1–3 Abnormal endothelial changes occur initially in the central cornea and spread peripherally.⁴ 

Patients with FECD have been successfully treated with Descemet membrane endothelial keratoplasty (DMEK) or descemetorhexis without endothelial keratoplasty (DWEK).^5,6 Furthermore, early intervention will be promoted by recent novel therapies, such as Rho-associated protein kinase inhibitor eye drops, cultivated endothelial cell injection, and gene therapy.^7–9 In our daily clinical practice, we often examine patients with FECD with decreased vision at the age of 60 to 70 years who are frequently complicated by both cataracts and FECD. Hence, it is difficult to judge which cataract or corneal lesion affects the visual function of the patient, and it is often challenging to determine the appropriate treatment, such as cataract surgery only, DMEK only, DWEK only, or DMEK triple. Furthermore, there is limited scientific evidence regarding the treatment indications for these therapeutic modalities. 

Some patients with FECD report visual disturbances, despite the absence of corneal edema. The mechanisms underlying visual disturbances in FECD without edema include higher-order aberrations (HOAs)¹⁰ and intraocular forward light scatter, also referred to as straylight.^11,12 Corneal structural changes that increase straylight also induce backward light scatter. Wacker et al. reported that anterior and posterior corneal HOAs and backward light scatter were higher even in the early stages of FECD.¹⁰ It was also assumed that the corneal guttae increase straylight, which causes visual disturbance.¹² Thus, all HOAs and forward and backward light scatter can affect visual function in patients with FECD. 

In the current study, we investigated the associations between visual function and several severity grades, HOAs, and scatter in patients with FECD. 

Methods

Subjects with FECD were recruited at the outpatient clinic of the Department of Ophthalmology at Osaka University Hospital. Cornea specialists diagnosed all cases of FECD based on the presence of longstanding bilateral corneal guttae or a beaten metal appearance. Cases with other causative abnormalities, such as surgery or inflammation, were excluded. Patients with other ocular diseases, previous intraocular surgery, or cataract of grade 3 or more (assessed according to the Lens Opacities Classification System III [LOCS III]) were excluded.^13,14 Healthy individuals without any ocular disorders were also recruited as controls. The Institutional Review Board of the Osaka University Hospital approved the protocol of this observational study (registration number: 14124). Written informed consent was obtained from all study participants after providing them with a thorough explanation of the study design and its risks and benefits. This research adhered to the tenets of the Declaration of Helsinki. 

To assess visual function, we evaluated corrected distance visual acuity (CDVA) and contrast sensitivity. CDVA was determined using two charts: Landolt-C decimal visual acuity chart and CSV-1000 Early Treatment Diabetic Retinopathy Study (ETDRS) letter chart. The decimal visual acuity with the Landolt-C chart was converted into the logarithm of the minimum angle of resolution (logMAR) for further evaluation. Contrast sensitivity was measured using 2 charts: the CSV-1000E chart and the CSV-1000RN letter chart (VectorVision, Houston, TX, USA) for letter contrast sensitivity (LCS). The area under the log contrast sensitivity function (AULCSF) was calculated using the contrast sensitivity measured with the CSV-1000E chart. 

As explanatory parameters for visual function, we analyzed age, modified Krachmer grade,¹⁵ central corneal thickness (CCT), anterior segment optical coherence tomography (AS-OCT)-based grade,¹⁶ HOAs, intraocular straylight, and corneal densitometry. Intraocular straylight and corneal densitometry parameters represent forward and backward light scatter, respectively.¹⁷ Corneal specialists determined the modified Krachmer grade for each patient using slit-lamp microscopy: grade 0 = no guttae; grade 1 = 1 to 12 central or paracentral non-confluent corneal guttae; grade 2 = more than 12 central/paracentral non-confluent corneal guttae; grade 3 = 1 to 2 mm of confluent central/paracentral corneal guttae at the widest diameter of the confluence after rotating the slit beam and measuring the diameter by narrowing the length of the beam and recording the length in mm; grade 4 = greater than 2 mm and up to 5 mm; grade 5 = greater than 5 mm of confluent central/paracentral guttae; and grade 6 = over 5 mm of confluent central/paracentral guttae with clinically apparent stromal and/or epithelial edema.¹⁸ We evaluated CCT, elevation map of the posterior corneal surface, and pachymetry map using AS-OCT, SS-1000 (TOMEY, Nagoya, Japan).¹⁹ This swept-source AS-OCT can acquire 10 µm axial and 30 µm lateral high-resolution images and scan at high speeds of up to 30,000 A-scans per second. We classified AS-OCT-based grades into 1 to 3 as follows: grade 1 (guttae only), a positive value in the central cornea, detected by the posterior elevation map; grade 2 (subclinical stromal edema), a negative value in the central cornea, detected by the posterior elevation map; and grade 3 (epithelial and stromal edema), a corneal thickness of > 700 µm within 3 mm of the central cornea.¹⁶ We examined HOAs using a wavefront analyzer, KR-1W (Topcon, Tokyo, Japan).²⁰ Corneal and ocular HOAs were obtained for the central cornea of 4-mm and 6-mm diameters, and the root mean square values of the HOAs were calculated by Zernike terms up to the sixth order. Intraocular straylight was evaluated using the C-Quant Straylight Meter (Oculus, Wetzlar, Germany), which quantifies the amount of intraocular forward light scatter using the compensation method.^12,21,22 The corneal densitometry, as a parameter of backward light scatter, was quantified using the Pentacam HR (Oculus).²³ Corneal densitometry was measured in four annular zones centered on the apex of the cornea (0–2, 2–6, 6–10, and 10–12 mm in diameter) and in three layers: anterior (anterior 120 µm), center (from the anterior 120 µm to the posterior 60 µm), and posterior (the posterior 60 µm) layers. Densitometry was expressed in grayscale units (GSU) ranging from 0 to 100. 

The associations between the visual functions and modified Krachmer grade or AS-OCT-based grade were analyzed using the Kruskal–Wallis test and Spearman's correlation coefficients. Steel–Dwass analysis was used to compare the corneal parameters for each pair of grades. We also performed univariate analysis to assess the relationships between the visual functions and explanatory factors using Spearman's rank correlation coefficient, except for the modified Krachmer and AS-OCT-based grades. We calculated the point-biserial correlation coefficients for age and lens status. Further, multivariate analysis was performed using a multiple linear regression model, and the adjusted coefficients of determination (adjusted R²) were performed. Based on the results of the univariate analysis, we included age, CCT, 4-mm ocular HOAs, intraocular straylight, and corneal densitometry of the posterior layer in 0 to 2 mm as explanatory variables in the multivariate analysis. 

We additionally conducted receiver operating characteristic (ROC) analysis to identify the threshold of corneal densitometry of the posterior layer in 0 to 2 mm for the diagnosis of visual dysfunction in patients with FECD. Specifically, we defined visual dysfunction as a value inferior to the mean by more than two standard deviations for the control eyes. The threshold was determined based on the Youden index, which optimizes sensitivity and specificity as high as possible. 

We used a Venn diagram to analyze the overlap of eyes with a modified Krachmer grade of ≥ 3 and those with a densitometry of the posterior layer in the 0 to 2 mm range of ≥ 10. We compared the visual functions between densitometry of ≥ 10 GSU (high densitometry group) and < 10 GSU (low densitometry group) in eyes with modified Krachmer grade ≥ 3 using Wilcoxon rank sum test. 

All statistical analyses were conducted under the guidance of a biostatistician (author R.K.) using JMP Pro Software (SAS Inc., Cary, NC, USA). The P values less than 0.05 were considered statistically significant. 

Results

The Institutional Review Board of the Osaka University Hospital approved the protocol of this observational study on January 15, 2019. This study was conducted between January 15, 2019, and October 2, 2020, the date when we ended the data collection. We enrolled 49 eyes of 27 patients with FECD (FECD group) and 10 eyes of 10 healthy individuals (control group). The characteristics of the participants are presented in Table 1 and Supplementary Table S1. There were significant differences in the rates of female eyes and phakic eyes between the control and FECD groups (Fisher's exact test, P = 0.047 and P = 0.01, respectively). 

Table 1.

 

Patient Characteristics

 

The associations between the modified Krachmer grade and visual function are shown in Figure 1. There were significant differences in all types of visual function among the grades: CDVA (Landolt-C), P = 0.0002; CDVA (ETDRS), P = 0.001; AULCSF, P < 0.0001; and LCS, P = 0.001. In particular, there were significant differences between grades 0 and 3 in CDVA (Landolt-C; average = −0.17 for grade 0 (95% confidence interval [CI] = −0.19 to −0.15), average = −0.014 for grade 3 (95% CI = −0.085 to 0.057), AULCSF, average = 1.41 for grade 0 (95% CI = 1.35 to 1.46), average = 0.91 for grade 3 (95% CI = 0.74 to 1.07), and LCS average = 19.4 for grade 0 (95% CI = 17.4 to 21.4), average = 13.6 for grade 3 (95% CI = 11.8 to 15.4). 

Figure 1.

 

Association between the modified Krachmer grade and visual functions. Visual functions deteriorated in eyes with modified Krachmer grade 3 or worse. The numbers of eyes were 10 in grade 0, 4 in grade 1, 6 in grade 2, 10 in grade 3, 21 in grade 4, 5 in grade 5, and 3 in grade 6. CDVA = corrected distance visual acuity; ETDRS = Early Treatment Diabetic Retinopathy Study; AULCSF = area under the log-contrast sensitivity function; LCS = letter contrast sensitivity. * P < 0.05, Steel–Dwass analysis. ρ = Spearman's correlation coefficient.

The associations between AS-OCT-based grades and visual functions are shown in Figure 2. There were significant differences among the grades in all visual functions (CDVA [Landolt-C], P < 0.0001; CDVA [ETDRS], P < 0.0001; AULCSF, P < 0.0001; and LCS, P = 0.0002). Visual function deteriorated even in patients with grade 1 (guttae only). 

Figure 2.

 

Association between the anterior segment optical coherence tomography (AS-OCT)-based grade and visual functions. Visual functions were affected even in eyes with grade 1. The numbers of eyes were 10 in grade 0, 33 in grade 1, and 16 in grade 2. CDVA = corrected distance visual acuity; ETDRS = Early Treatment Diabetic Retinopathy Study; AULCSF = area under the log-contrast sensitivity function; LCS = letter contrast sensitivity. *P < 0.05, Steel–Dwass analysis. ρ = Spearman's correlation coefficient.

The results of the univariate analysis of visual function and explanatory factors are shown in Table 2. There were significant associations between visual functions and HOAs, straylight, or densitometry of the central cornea up to 6 mm. 

Table 2.

 

Results of Univariate Analysis of Visual Functions and Explanatory Factors

 

For the multivariate analysis, the selection of explanatory valuables was conducted as follows. The variables were classified in five categories: age, CCT, HOAs, intraocular straylight, and corneal densitometry. There was only one parameter for age, CCT, and intraocular straylight. For HOAs, ocular (4 mm) had the highest correlation coefficient with LCS. Moreover, ocular HOAs included information regarding the posterior surface of the cornea as well as the anterior surface, whereas corneal HOAs included information only for the anterior surface of the cornea. The corneal changes in patients with FECD start from the posterior surface of the central cornea. Thus, we considered it adequate to choose ocular (4 mm) HOAs as an explanatory variable. For densitometries, the posterior (0–2 mm) had the highest correlation coefficient with CDVA (Landolt-C), AULCSF, and LCS in the univariate analysis. Thus, we selected it as an explanatory variable. Furthermore, it makes clinical sense because the corneal guttae or opacity occurs predominantly in the posterior segment of the central cornea. The results of the multivariate analysis are shown in Table 3. The explanatory variables of HOAs and corneal densitometry were significantly associated with CDVA (Landolt-C). Age, HOAs, and corneal densitometry values were significantly associated with CDVA (ETDRS). Corneal densitometry was significantly associated with AULCSF. CCT and corneal densitometry were both significantly associated with LCS. In summary, corneal densitometry was significantly associated with all parameters of visual function. HOAs were significantly associated with CDVA. 

Table 3.

 

Results of the Multivariate Analysis

 

The results of ROC analysis are shown in Figure 3. The threshold values of corneal densitometry to classify visual dysfunction based on the Youden index were 9.7 GSU for CDVA (Landolt-C; area under the curve [AUC] = 0.82, sensitivity = 0.82, and specificity = 0.80), 11.6 GSU for CDVA (ETDRS; AUC = 0.71, sensitivity = 0.59, and specificity = 0.78), 9.7 for AULCSF (AUC = 0.91, sensitivity = 0.83, and specificity = 1.00), and 11.3 GSU for LCS (AUC = 0.77, sensitivity = 0.72, and specificity = 0.79). Based on the estimated threshold values derived from the four visual outcomes, we adopted 10 GSU as the suggested threshold value for corneal densitometry to identify visual dysfunction. 

Figure 3.

 

ROC analysis for the threshold optimization of corneal densitometry (posterior layer in 0–2 mm) for visual dysfunction. Youden index, optimized sensitivity, and optimized specificity were determined based on ROC analysis. Visual dysfunction thresholds were −0.11 for CDVA (Landolt-C), 0.19 for CDVA (ETDRS), 1.27 for AULCSF, and 14.0 for LCS. ROC = receiver operating characteristic; CDVA = corrected distance visual acuity; ETDRS = Early Treatment Diabetic Retinopathy Study; AULCSF = area under the log-contrast sensitivity function; LCS = letter contrast sensitivity.

Using a Venn diagram (Fig. 4), we illustrated the overlap between eyes with a modified Krachmer grade of 3 or worse and those with densitometry of the posterior layer at 0 to 2 mm of 10 or worse. Most eyes with a modified Krachmer grade ≥ 3 overlapped consistently with those with densitometry ≥ 10 GSU. Among eyes with modified Krachmer grade ≥ 3, a densitometry equal or above 10 GSU was observed in 29 eyes (74.4%), and below 10 GSU in 10 eyes (25.6%). In order to analyze which of the two thresholds of densitometry for 10 GSU or modified Krachmer grade of 3 reflected visual deterioration, we intended to make a comparison to investigate which of them reflected visual deterioration. However, there were only 2 eyes with densitometry ≥ 10 and modified Krachmer grade <3, as shown in Figure 4. Therefore, we decided to compare patients with modified Krachmer grade ≥ 3 who had a densitometry ≥ 10 (N = 29) and those with a densitometry < 10 (N = 10). CDVA (Landolt-C) and AULCSF were significantly worse in the high densitometry group than in the low densitometry group among eyes with a modified Krachmer grade ≥ 3 (Table 4). 

Figure 4.

 

Venn diagram that depicts the relationship between eyes with modified Krachmer grade of ≥ 3 and those with densitometry ≥ 10 grayscale units (GSU). Most eyes with modified Krachmer grade ≥ 3 overlapped with those with densitometry ≥ 10 GSU.

Table 4.

 

Comparison of Visual Function Between the Low and High Densitometry Groups in Eyes With Modified Krachmer Grade 3 or Worse

 

Discussion

In the current study, we investigated the association between visual function and severity grading, corneal scatter, or HOAs in patients with FECD. We showed that the threshold for decreased visual function was confluent corneal guttae equivalent to the modified Krachmer grading scale ≥ 3 and the central 0 to 2 mm corneal densitometry of the posterior layer ≥ 10 GSU. We also demonstrated that HOAs and forward and backward light scatter affected visual functions, with backward light scatter being the most influential factor. 

Modified Krachmer grade 3 is the grade where corneal guttae become confluent, and we previously reported that corneal guttae without edema can affect the quality of vision (QOV) in patients with FECD.¹² Moreover, we demonstrated in the current study that corneal guttae affect visual function once they become confluent. We found that the visual function of patients with AS-OCT-based grade 1 was significantly impaired. Patients with AS-OCT-based grade 1 included those with modified Krachmer grades 2, 3, and 4 (see Table 1). Therefore, AS-OCT-based grading appears not to be particularly sensitive for the evaluation of visual dysfunction. 

Using univariate analysis, we demonstrated that all HOAs and forward and backward light scatter significantly correlated with visual function. Additionally, corneal densitometry was significantly associated with all parameters of visual function in the multivariate analysis, whereas HOAs were only associated with CDVA. Moreover, in the group of patients with modified Krachmer grade ≥ 3, those with the 0 to 2 mm central corneal densitometry values of the posterior layer ≥ 10 GSU had significantly decreased CDVA (Landolt-C) and AULCSF (see Table 4). This might suggest that corneal densitometry would be more valuable than the modified Krachmer grade to determine patients with significantly decreased visual function. Therefore, we believe that corneal backward light scatter is the most influential factor affecting visual function. Kobashi et al. evaluated the factors affecting CDVA in patients with FECD. They suggested that both forward light scatter and anterior backward light scatter, especially the former, play a more important role in visual performance than corneal HOAs in FECD.²⁴ The results regarding forward light scatter were somehow different from our results, possibly due to the difference in the measurement method (subjective measurement in our study, and objective measurement in their study). 

To date, there have been other reports regarding the association between severity and visual function in patients with FECD. Shah et al. reported an association of decreased CDVA in FECD with both central corneal densitometry values at 0 to 2 mm and severity of guttae.²⁵ Sun et al. and Patel et al. also reported that the presence of these tomographic features of edema in pseudophakic FECD eyes, such as loss of parallel isopachs, displacement of the thinnest point of the cornea, and focal posterior corneal surface depression, was associated with worse visual acuity, disability glare, and patient-reported visual disability.^26,27 

This study had some limitations. The sample size was relatively small. A larger sample size is necessary to determine the most relevant explanatory variable affecting visual functions. In Japan, a nationwide registry for patients with FECD is being established currently to investigate the detailed clinical and genetic characteristics as well as the etiology in government-funded research groups to establish standardized diagnostic criteria and clinical practice guidelines for intractable corneal diseases. It is anticipated that these parameters can be analyzed in a larger number of patients. Nevertheless, the findings of this study are based on a robust analysis of the available sample size. 

In conclusion, our investigation into the associations between several severity grades of FECD and visual function revealed a significant impairment in visual function with a modified Krachmer grade of ≥ 3 and corneal densitometry of ≥ 10 GSU. 

Acknowledgments

Supported by Japan Agency for Medical Research and Development (AMED; Grant Number JP22ek0109590), and the MHLW Research on Rare and Intractable Diseases Program (Grant Number 20FC1032). 

Disclosure: C. Kai, None; Y. Oie, None; N. Nishida, None; S. Doi, None; C. Fujimoto, None; S. Asonuma, None; S. Maeno, None; T. Soma, None; S. Koh, None; V. Jhanji, None; R. Kawasaki, None; K. Nishida, None