Macular Ganglion Cell–Inner Plexiform Layer and Retinal Nerve Fiber Layer Thickness in Eyes With Primary Open-Angle Glaucoma Compared With Healthy Saudi Eyes

A Cross-Sectional Study

Abstract

Purpose 

We compared the thickness of the ganglion cell–inner plexiform layer at the macula (mGCIPL) and the thickness of the retinal nerve fiber layer (RNFL) in different regions of the retina in eyes with primary open-angle glaucoma (POAG group) and normal eyes (control group).

Design 

This was a cross-sectional study performed in 2014.

Methods 

Spectral domain optical coherence tomography (SD-OCT) was used to measure mGCIPL and RNFL thickness. Age-adjusted means and standard deviation were calculated. Age, sex, refractive status, corneal thickness, and stage/severity of glaucoma (defined by vertical cup-to-disc ratio and visual field changes) were associated to outcomes. Statistical significance was indicated by P < 0.05.

Results 

There were 50 eyes in the POAG group and 52 eyes in the control group. The difference in age between patients in both groups was statistically significant (P < 0.001). The age-adjusted measurements were thinner for POAG with a mean difference (DF) of 11.1 μm for mGCIPL and 8 μm for mRNFL. The mGCIPL to mRFNL ratio was 2.1 in the POAG group and 1.9 in the control group (degrees of freedom = 0.2, P = 0.001). The mGCIPL and RNFL thickness decreased as the severity of glaucoma increased. The mGCIPL to mRNFL ratio was a predictor of the severity of field defects in POAG (AUROC = 0.66, P = 0.0002). Age and myopia were confounders to the association of OCT findings and the visual field changes in POAG (P = 0.07).

Conclusions 

There was generalized thinning of retinal layers in eyes of nondiabetic Arab patients with POAG. OCT parameters can be important for detecting and monitoring glaucoma cases.

Primary open-angle glaucoma (POAG) is one of the priorities of Vision 2020—the global initiative for the prevention of avoidable blindness.¹ Early detection and timely management of POAG are accepted strategies to combat this condition.² A greater knowledge of the pathogenesis will enable caregivers to halt, or perhaps reverse, the damage that leads to visual disabilities in glaucoma. With the advent of newer technologies such as optical coherence tomography (OCT), one can review anatomical changes taking place in different layers of the retina.³ Thinning of the retinal nerve fiber layer (RNFL) at the optic disc has been associated with POAG.⁴ The ganglion cell–inner plexiform layer at the macula (mGCIPL) is also affected in POAG, causing structural and functional changes.⁵ Scientists still debate which occurs first—the thinning of the macular RNFL (mRNFL) or the changes in the mGCIPL.^6,7

Fourier domain, time domain, and spectral domain OCT are used as diagnostic tools for a number of eye diseases.^4,8,9 Optical coherence tomography is mainly used to estimate the thickness of different layers of the retina at the macula and peripapillary region.^10,11 The thicknesses of the RNFL and GCIPL vary with age and race.^12,13 Therefore, the findings of a study area may not be similar to other studies. There is a relative paucity of peer-reviewed literature on structural changes in retinal layers in the Arab population, both in normal eyes and in eyes with POAG. The prevalence of glaucoma in the Arab population is as high as 4.75%.¹⁴ More than half of moderate and severe visual impairment in populations older than age 50 in the Middle East including Arab countries is due to uncorrected refractive error.¹⁵ The life expectancy of the Middle Eastern population is also increasing.¹⁶ Thus, factors such as age, race, refractive error, and glaucoma that can influence retinal layer thickness exist in higher proportions among Arabs and are therefore worth studying.

We conducted a study to evaluate the thickness of different retinal layers in the macula and the RNFL in different quadrants in Arab patients with POAG and compared these measurements with a group of patients free of systemic and ocular pathology. We also aim to provide reference data for the RNFL thickness of healthy eyes and glaucomatous eyes in the Arab population.

MATERIALS AND METHODS

The research and ethical committee of the institution approved this study. All cases with POAG (POAG group) that had undergone treatment for at least 1 year were included in this study. Primary open-angle glaucoma was defined as eyes having optic nerve head changes and field of vision changes suggestive of glaucoma, with an open anterior chamber angle and with or without a rise in intraocular pressure (IOP).¹⁷ Informed consent was obtained from all patients who participated in this study. Patient anonymity was maintained at all times. Patients were excluded if they had diabetes, retinal vascular diseases, or did not consent to participate in the study. The POAG group was compared with a similar number of eyes that did not have glaucoma or any ocular comorbidities but were being evaluated for refractive error correction (control group).

We assumed that in 90% of eyes with POAG, the macular thickness as measured with OCT would demonstrate changes. In a cross-sectional study with a population of 500 glaucoma cases, to achieve a 10% margin of error and a study design effect of 1.5, at least 50 eyes with POAG should be assessed. A comparable group of at least 50 eyes of persons without glaucoma were also studied to determine the OCT findings of the Saudi population without glaucoma.

One glaucoma specialist, 1 ophthalmologist, and 1 ophthalmic technician were the field investigators. Demographic information included age, sex, duration of glaucoma, eye involved, and the type of glaucoma medication used. Distance visual acuity of each eye was tested using an Early Treatment Diabetic Retinopathy Study chart placed at a 3-m distance. The field of vision was tested using an automated visual field analyzer, by a 76-point, central 30-degree suprathreshold examination with the central reference levels set at 2 or 4 dB lower than the estimated normal median central reference level adjusted for age (Humphrey). Refraction was measured using an autorefractor (Reichert Technologies, Germany) and direct retinoscopy (Heine, Germany) after achieving adequate cycloplegia with instillation of 0.5% tropicamide drops. Myopia was graded as mild [<−2 diopters (D) spherical equivalent], moderate (−2 to −6 D), or high (>−6 D). Hyperopia in adults was defined as those having a spherical equivalent of more than +0.5 D. The eyes without refractive error or up to +0.5 D were termed as emmetropia. Intraocular pressure was measured using an applanation tonometer attached to a slit lamp biomicroscope (Topcon, Japan). A gonio lens (Volk) was used to evaluate the angle of the anterior chamber. Grade 1 was narrow, grade 2 was closed but open with pressure from the gonioscope, grade 3 had an open angle with visible scleral spur, and in grade 4, there was a wide-open angle of the anterior chamber with all structures visible through the gonioscope. Corneal thickness was measured with a pachymeter (Reichert iPac Pachymeter, New York, NY).

The retinal thickness measurements were performed with the Cirrus HD-OCT 5000 (Carl Zeiss Meditec AG, Jena, Germany). We noted the thickness of the mGCIPL and the mGCIPL plus the RNFL (mGCIPL++) superior to the macula. The ratio of mGCIPL thickness to total thickness (mGCIPL++) was calculated. These measurements were performed out to 1 disc diameter inferiorly and superiorly to the fovea. The thickness of the RNFL was automatically measured at various points at 3 separate diameters away from the optic disc, and the average was displayed for superior, inferior, nasal, and temporal quadrants. The unit of measurement for retinal thickness was micrometers.

All glaucoma cases were grouped in consolidated stage (CS) 1 to CS3 based on the classification of the severity of visual field defects.¹⁸ We also grouped all eyes based on vertical cup-to-disc (C/D) ratio.¹⁹

The data were collected using a Microsoft Excel (Microsoft Corp, Redmond, WA) spreadsheet. Statistical Package for the Social Sciences (SPSS 22; IBM Corp, New York, NY) was used to perform univariate and multivariate analysis. All continuous variables, such as thickness of the retinal layer, were reviewed for a normal distribution. If the distribution was skewed, log values were calculated. If the distribution was normal, the mean and standard deviation (SD) were calculated. The OCT findings of eyes with POAG were compared with those of normal eyes by calculating age-adjusted means and SD. We used the “weight cases” function of SPSS to adjust the OCT parameters by age. Then, we estimated the differences in mean and the 95% confidence interval (CI). A similar exercise was undertaken to calculate and compare the thickness of the GCIPL in the superior and inferior macula. The influence of variables such as unadjusted age, sex, refractive status (myopia vs non myopia), eye affected, type of glaucoma medication, extent, and duration of glaucoma on differential thickness of the RNFL and GCIPL was studied using regression analysis and by plotting the area under the receiver operating characteristics curve (AUROC). In the regression model, we inserted the actual age and not the adjusted one.

RESULTS

Our study comprised 50 eyes with glaucoma (POAG group) and 52 eyes without glaucoma (control group). Age and refraction were statistically significantly different between the groups (Table 1). Patients in the POAG group were significantly older than patients in the control group (P < 0.001). Emmetropia was significantly greater in eyes without glaucoma, whereas high myopia was more common in eyes with glaucoma (P < 0.001).

TABLE 1: Comparison of Demographic and Refractive Data of Eyes With and Without Glaucoma

The IOP was 15 mm Hg or less in 25 eyes. In the rest of the eyes with glaucoma, it ranged from 16 to 22 mm Hg. The vertical C/D ratio was more than 0.7 in 22 eyes, 0.5 to 0.7 in 25 eyes, and less than 0.5 in 5 eyes. Visual field defects were of 1 SD or less in 19 eyes, 1 SD to 2 SDs in 20 eyes, and more than 2 SDs in 11 eyes. Thirty-eight eyes were treated with 1 glaucoma medication, whereas 12 eyes were treated with a combination of glaucoma medications. Gonioscopy revealed that, in glaucomatous eyes, the anterior chamber angle was grade 2, 3, and 4 in 17, 26, and 7 eyes, respectively.

We compared the RNFL in 4 quadrants in both groups (Table 2). The RNFL was statistically significantly thinner in the POAG group compared with the control group. This difference was highest in the inferior quadrant and lowest in the temporal quadrant.

TABLE 2: Retinal Nerve Fiber Layer Thickness in Eyes With and Without Glaucoma

The OCT measurements of the macula are presented in Table 3. Both the mGCIPL and mRNFL were thinner in the POAG group compared with the control group. In addition, the mGCIPL and mRNFL superior to the macula were thicker than the corresponding measurements inferior to the macula in both groups.

TABLE 3: Macular Parameters in Eyes With and Without Glaucoma

The 102 eyes were grouped according to the grade of visual field changes. The measurements of different layers of the macula and in the 4 quadrants of the retina based on visual field defects are presented in Table 4. Different retinal layers both at the macula and in different quadrants of the retina were thinner in eyes with severe visual field defects (CS3 and CS4) compared with early field changes (CS1 and CS2).

TABLE 4: OCT Parameters by Field Defect Severity Classification of Glaucoma

The measurements of different layers of the macula and in the 4 quadrants of the retina of the 102 eyes grouped by vertical C/D ratio are presented in Table 5. In eyes with a higher C/D ratio (>0.7), different retinal layers both at the macula and in different quadrants of the retina were thinner than in eyes with C/D ratios less than 0.7.

TABLE 5: OCT Parameters by C/D Ratio

In the POAG group, we calculated the ratio of mGCIPL thickness to total retinal thickness at the macula (GT ratio). The linear regression equation suggested that the variation in GT ratio among POAG cases could be explained with age using the formula GT = 1.7 (0.008 × age) (P < 0.05). Variations in GT ratio by type of refractive error (P = 0.7) and sex (P = 0.5) were not statistically significant.

The mean deviation in visual field defects in the eyes with POAG could be predicted with the GT ratio noted on OCT, MD = [2.3 + (−2.1 × GT)] (P = 0.0006). However, predicting the vertical C/D ratio in eyes with POAG on the basis of the GT ratio (found on OCT) was not statistically significant, C/D ratio = [0.6+ (0.06 × GT)] (P = 0.3).

By plotting the AUROC, we noted that the ratio of the mGCIPL to the mRNFL was a weak but significant predictor of the severity of field defects in the eyes with POAG (AUROC = 0.66, P = 0.0002).

Refractive error is a known confounder that affects the association of glaucoma with retinal thickness parameters. Therefore, we conducted stratified analyses. In 26 myopic eyes with glaucoma, the RNFL (P = 0.01), mRNFL (P = 0.01), and mGCIPL (P = 0.003) were thinner than in 10 eyes without glaucoma. Among 24 emmetropic and hypermetropic eyes with glaucoma, the RNFL (P < 0.001), mRNFL (P < 0.001), and mGCIPL (P < 0.001) were significantly thinner than in 42 eyes without glaucoma.

We carried out linear regression to study the interaction of factors on retinal layer thickness. The variation of RNFL was associated with cases (P < 0.001), age (P = 0.07), myopia (P = 0.2), and males (P = 0.1). The variation of mRNFL was associated with cases (P < 0.001), age (P = 0.07), myopia (P = 0.2), and males (P = 0.1).

DISCUSSION

In this study, we established the importance of OCT in measuring retinal layer thickness at the macula and in 4 quadrants of eyes with POAG. There was generalized thinning of different layers of the retina in POAG compared with Saudi eyes without glaucoma. The ratio of mGCIPL thickness to mRNFL thickness was a predictor of visual field defects in eyes with POAG. Age and myopia were confounders to the association of retinal layer thickness measured by OCT with the grades of field changes in POAG.

There are a few limitations to this cross-sectional study. The causal association of glaucoma with the variation in macular thickness parameters could not be established. Longitudinal studies are recommended to identify the sequence of damage to the different layers as glaucoma progresses. The subgroup of refractive status was small, and statistical error cannot be ruled out in its association with the observed changes in macular parameters. Further studies with an adequate sample size could establish the role of refractive status on the variability of macular thickness parameters in glaucoma. All cases of POAG had macular changes, and less than 5% of eyes without glaucoma had changes in the macular parameters (perhaps due to refractive error or other comorbidities that could not be determined at the beginning of the study). In such cases, the study sample allows greater validity of the results.

A study in Riyadh, Saudi Arabia, revealed that the vertical C/D ratio was 0.49 in 29 normal eyes and 0.83 in 32 glaucomatous eyes.²⁰ The mean RNFL thickness in all 4 quadrants in the study by Zeried et al²⁰ was 101 versus 63.9 μm. In our study, it was 101 and 80.7 μm. Perhaps differences in the number of cases of advanced glaucoma could be the reason for variation in retinal measurements between studies. In 30 normal Egyptian eyes, the RNFL thickness was 110 μm, and in 150 glaucomatous eyes, it was 80 μm.²¹ Thus, even among Arabs, RNFL thickness varied between the Saudi and Egyptian communities. In normal and glaucomatous eyes of the Arab population, macular thickness parameters measured by OCT are presented perhaps for the first time in our study.

The RNFL is uniformly thinner in glaucomatous eyes compared with nonglaucomatous eyes in some recent studies.^22,23 Sectorial variation in the thinning of the RNFL in glaucomatous eyes compared with normal eyes was noted in our study. It was marked in the inferior quadrant followed by the superior quadrant. Similar variations were also observed by Gyatsho et al.²⁴ The postural effect of high IOP being greater in the lower part of the globe could explain the sectorial variation observed in our study.²⁵

The mGCIPL thickness in glaucomatous eyes was significantly thinner than in normal eyes. Francoz et al²⁶ also noted this difference while comparing such parameters in normal and glaucomatous eyes. Functional changes due to altered pathophysiology at the cellular level could be the key to the thinning of the mGCIPL. Kita et al⁷ used the ratio of the mGCIPL to total retinal thickness and demonstrated their role in improving diagnostic ability. Clinicians dealing with glaucoma patients in the Arab population could use different retinal layer ratios at the macula to predict the presence of the disease and monitor its progression.

Disc changes seem to be structural damage due to glaucoma, whereas visual field changes represent pathophysiological damage due to disease. The association of changes in macular parameters with visual field changes was more accurate and reliable than that with disc changes. Shen et al²² noted a similar linear relationship. Perhaps thinning of the retinal layers at the macula reflects field of vision changes as short-term and disc changes as long-term outcomes.

The normative retinal layer thickness found in our study matched with data from Indian and Chinese populations.^26,27 In the Belgian population, it was 105 μm compared with 102 μm in the Japanese population.²⁸ We noted nearly similar retinal thicknesses using similar equipment. It seems that the variation could be racial or it could be due to differences in the refractive error and age of participants.

Age is a known confounder to the association of retinal thickness parameters with glaucoma.¹³ The RNFL thickness was different in our study population compared with glaucoma patients in different countries.^10,29,30 Even among Arab countries, the parameters varied in Saudi Arabia and Egypt.^20,21 Therefore, clinicians should account for race and age before interpreting glaucoma status based only on retinal layer thickness.

We had differential representation of myopes, hyperopes, and emmetropes in eyes with glaucoma compared with healthy eyes. Although the retinal layers were thin in eyes with myopia in the normal population, they were marginally thinner in eyes with glaucoma. The influence of refractive error (mainly myopia) on the association of glaucoma and retinal layer thickness thus was not conclusive. Further studies with adequate subgroups of different types of refractive error are recommended.

High-quality OCT is a reliable tool for assessing the presence and severity of POAG. Anatomical and functional progression of POAG-related damage can be studied if OCT is provided to caregivers. Further studies are needed to determine the validity of individual parameters, such as mGCIPL and total retinal thickness, and the role of myopia on such association to predict the severity of POAG.

ACKNOWLEDGMENTS

The authors thank the staff of Specialized Medical Center for assisting in assessing and investigating the participants in this study. Dr Randy Craven and Dr Deepak Edward provided suggestions to improve the manuscript. The authors also thank the patients for their cooperation and consent to participate in this research project.