- Email: [email protected]
Retinal Nerve Fiber Layer Thickness in a Multiethnic Normal Asian Population
Retinal Nerve Fiber Layer Thickness in a Multiethnic Normal Asian Population The Singapore Epidemiology of Eye Diseases Study Henrietta Ho, FRCOphth,1,2 Yih-Chung Tham, PhD,1 Miao Li Chee, BSc,1 Yuan Shi, PhD,1 Nicholas Y.Q. Tan, MA, BM BCh,1 Kah-Hie Wong, MBBS,1,3 Shivani Majithia, OD,1,4 Carol Y. Cheung, PhD,1,5,6 Tin Aung, FRCOphth, PhD,1,4,6 Tien Yin Wong, FRCS(Ed), PhD,1,4,6 Ching-Yu Cheng, MD, PhD1,4,6 Purpose: To describe variations in retinal nerve fiber layer (RNFL) thickness based on spectral-domain (SD) OCT in a multiethnic Asian population. Design: Population-based, cross-sectional study. Participants: Ethnic Chinese, Malay, and Indian adults older than 48 years without glaucoma who were recruited from the Singapore Epidemiology of Eye Diseases Study. Methods: All participants underwent standardized systemic and ocular examinations. Retinal nerve fiber layer thickness was measured using SD OCT. Participants with poor-quality scans were excluded. Linear regression models were used to investigate the associations of ocular and systemic factors with average RNFL thickness. Generalized estimating equation models were used to account for correlation between both eyes. Main Outcome Measure: Average RNFL thickness. Results: Four thousand four hundred seventy-five participants (8178 eyes) consisting of 1371 Chinese, 1303 Malay, and 1801 Indian adults contributed to this analysis. Average RNFL thickness measured was 95.79.6 mm in Chinese participants, 94.910.6 mm in Malay participants, and 87.310.6 mm in Indian participants (P < 0.001). Multivariate analysis adjusted for age, gender, and ethnicity revealed a reduction in RNFL thickness with increased intraocular pressure and axial length (P < 0.001 for both), as well as a diagnosis of diabetes (P ¼ 0.04); greater RNFL thickness was associated with increased disc area (P < 0.001), signal strength (P < 0.001), and lowdensity lipoprotein cholesterol (P ¼ 0.02). When these significant variables were taken into account, the average RNFL thickness of Indian participants was significantly thinner compared with Chinese participants (7.45 mm thinner on average [95% confidence interval, 6.75e8.15 mm; P < 0.001]), whereas there was no significant difference in average RNFL thickness between Malay and Chinese participants (P ¼ 0.15). Conclusions: Average and regional RNFL thicknesses were significantly thinner in Indian eyes compared with Chinese and Malay eyes. Results of the study highlight the need to acquire more refined normative data for the comparison of individual patients with others of similar ethnic background while accounting for ocular factors that could influence RNFL thickness. This in turn may improve the sensitivity and specificity of glaucoma detection. Ophthalmology 2019;-:1e10 ª 2018 by the American Academy of Ophthalmology Supplemental material available at www.aaojournal.org.
Glaucoma is the leading cause of irreversible vision loss worldwide.1 Because glaucoma may remain asymptomatic until an advanced stage, diagnosis and treatment frequently are delayed. Visual field deficits are detected often only by standard automated perimetry after significant loss of retinal ganglion cells.2 Although structural evaluation of optic nerve head (ONH) features is imperative in the early detection and monitoring of glaucoma, the clinical evaluation and interpretation of ONH features using funduscopy or photographs are known to vary even among glaucoma experts because of interindividual variations of ONH anatomic features.3 ª 2018 by the American Academy of Ophthalmology Published by Elsevier Inc.
In contemporary clinical practice, spectral-domain (SD) OCT is used commonly to evaluate thinning of the peripapillary retinal nerve fiber layer (RNFL) objectively and quantitatively, which is a hallmark for diagnosis of glaucoma.4 New machines and sophisticated algorithms dissect data to allow a more streamlined interpretation. However, the accuracy of what qualifies as statistically significant RNFL thinning depends on the generalizability of normative databases. In this regard, it is problematic that proprietary reference databases of healthy eyes are largely reliant on data from predominantly white populations. Although they account for 60% of glaucoma cases globally,1 Asians are https://doi.org/10.1016/j.ophtha.2018.11.031 ISSN 0161-6420/19
1
Ophthalmology Volume -, Number -, Month 2019 poorly represented in normative OCT databases. For example, in the Cirrus SD OCT RNFL normative database, only 24% of eyes were from Asians. To complicate matters further, previous studies have shown significant differences in RNFL thickness across different ethnic populations.5,6 Therefore, even the consideration of “Asian” as a single group may be insufficient because this may mask ethnic variations (e.g., between Chinese and Indians, the 2 largest Asian ethnic groups) within the heterogeneous Asian population. To date, a comprehensive interethnic evaluation among Asians has not been reported. In multiethnic countries such as Singapore, establishing an ethnic-specific normative database of RNFL thickness is especially important for the diagnosis and follow-up of glaucoma. To address this gap, we examined and described the ethnic variations in peripapillary RNFL thickness using SD OCT (Cirrus HD-OCT; Carl Zeiss Meditec, Inc., Dublin, CA) within a nonglaucomatous multiethnic Asian population. A secondary aim was to identify the systemic (e.g., diabetes, hypertension) or ocular (e.g., axial length) associations with RNFL thickness and to determine if interethnic variations in these systemic parameters may be responsible for the possible interethnic differences in RNFL thickness. For example, Indian adults have a higher prevalence of diabetes, but a lower prevalence of myopia, compared with Chinese adults.7,8 These findings may be useful in understanding ethnic differences in glaucoma prevalence. By promoting the establishment of a more refined, Asian-specific normative database, the detection of glaucoma in Asians may improve. In addition, the results of our study may be able to justify the need to acquire more refined normative data for the comparison of individual patients with others of similar racial background while accounting for ocular factors that could influence RNFL thickness.
participants from the other 2 ethnic groups. To ensure that participants from each ethnic cohort straddled similar age brackets, we removed 391 Chinese participants who were younger than 48 years and 86 Malay and 72 Indian participants who were older than 82 years. After this exclusion, OCT data of ONH and RNFL measures were available for 4879 participants (9166 eyes). Because participants may have had only unilateral OCT data available, we further excluded these participants. In addition, 181 eyes with OCT segmentation errors and poor scan quality with signal strength less than six, 285 eyes with glaucoma and other neurodegenerative diseases and 522 eyes with missing variables were excluded. This left a total of 8178 eyes (4475 participants), including 2316 Chinese eyes (1371 participants), 2445 Malay eyes (1303 participants), and 3417 Indian eyes (1801 participants) for analysis (Supplemental Material S1, available at www.aaojournal.org).
Ophthalmic Assessment Subjects underwent a comprehensive ocular examination at the Singapore Eye Research Institute. Visual acuity and subjective refraction were measured by research optometrists. Slit-lamp biomicroscopy was performed by study ophthalmologists. Intraocular pressure (IOP) measurements were obtained using a Goldmann applanation tonometer. Funduscopy was performed after pupil dilation with tropicamide 1% and phenylephrine hydrochloride 2.5%, and the optic disc was evaluated using þ78-diopter lenses at 16 magnification. Ocular biometry included axial length (AL) measurements using noncontact partial coherence interferometry. Five AL measurements were obtained and the mean was used for analysis. High myopia was defined by a spherical equivalent of less than e6.0 diopters and long AL was defined by AL of more than 26 mm. The diagnosis of glaucoma was made by a glaucoma fellowship-trained ophthalmologist according to the International Society for Geographical and Epidemiological Ophthalmology criteria based on optic disc appearance and static automated perimetry (Swedish interactive threshold algorithm standard 24-2, Humphrey Field Analyzer II; Humphrey Instruments, San Leandro, CA).13
OCT Imaging
Methods Written informed consent was obtained from each participant, and the studies adhered to the tenets of the Declaration of Helsinki. Ethical approval was obtained from the SingHealth Centralized Institutional Review Board.
Study Population We conducted a population-based study using data from the Singapore Epidemiology of Eye Diseases Study of adults 40 to 80 years of age from 3 major Asian ethnic groups: Chinese persons (the Singapore Chinese Eye Study), Indian persons (the Singapore Indian Eye Study), and Malay persons (the Singapore Malay Eye Study). The methodologies of these cohort studies have been reported in detail elsewhere.9e11 Data for the current study were derived as follows. Spectraldomain OCT was incorporated into the examination protocol of the Chinese cohort at baseline (2009e2011; n ¼ 3353; response rate, 72.8%), the Malay cohort at the 6-year follow-up (2011e2014; n ¼ 3280; response rate, 78.7%), and the Indian cohort at the 6-year follow-up (2013e2015; n ¼ 2200; response rate, 75.2%).9,12 Thus, data for the present study were derived from these 3 studies. However, the inclusion of Chinese participants at baseline, compared with Malay and Indian participants at follow-up visits, contributed to a difference among the age distributions of different ethnic groups, with Chinese participants tending to be 4 to 6 years younger than the
2
OCT imaging was performed using a commercially available SD OCT instrument (Cirrus HD-OCT; Carl Zeiss Meditec, Dublin, CA) after pupil dilation, and an optic disc scan was acquired using the optic disc cube 200200 scan protocol, covering a measurement area of 66 mm. The ONH and RNFL algorithms native to Cirrus HD-OCT were used to measure a series of ONH and RNFL parameters (disc area, neuroretinal rim area, cup volume, average and vertical cup-to-disc ratios, and average and per-quadrant peripapillary RNFL thickness) automatically by identifying Bruch’s membrane opening as the disc area, and the reference plane was determined at 200 mm above the level of Bruch’s membrane plane. All images were reviewed using the Cirrus HD-OCT review software, with RNFL and optic disc segmentation checked at this time. Detailed descriptions of the measurement algorithms have been described previously.14 Study eyes with OCT scans showing RNFL algorithm segmentation failure, signal strength less than 6, or artifacts resulting from eye movements or blinking were excluded further from the analysis.
Other Measurements Relevant sociodemographic and medical information was collected through a detailed interviewer-administered questionnaire that was conducted in the patient’s language of choice (English, Chinese, Malay, or Tamil) by bilingual interviewers. Systemic examination was performed and blood samples were collected for biochemistry analysis.9,15 Blood pressure (BP) was measured using a digital
Ho et al
RNFL Thickness among Asian Ethnic Groups
Table 1. Demographic and Systemic Characteristics among Participants from the 3 Ethnic Groups P Value Chinese (n [ 1371 Participants)
Malay (n [ 1303 Participants)
Indian (n [ 1801 Participants)
57.47.0 683 (49.8) 23.63.5 192 (14.0) 726 (53.0) 3.40.9
60.68.6 600 (46.1) 27.05.1 404 (31.0) 907 (69.7) 3.41.1
186 (13.6)
246 (18.9)
Age (yrs) Male gender Body mass index (kg/m2) Diabetes Hypertension Low-density lipoprotein cholesterol (mmol/l) Current smoker
P Value*
Indian vs. Malay
Chinese vs. Malay
Indian vs. Chinese
60.77.8 910 (50.5) 26.54.5 740 (41.1) 1170 (65.0) 3.51.0
<0.001 0.03 <0.001 <0.001 <0.001 <0.001
0.90 0.01 0.02 <0.001 0.01 0.04
<0.001 0.05 <0.001 <0.001 <0.001 0.16
<0.001 0.69 <0.001 <0.001 <0.001 <0.001
237 (13.2)
<0.001
<0.001
<0.001
0.75
Data presented are mean standard deviation for continuous variables and number (%) for categorical variables. *Difference across the 3 groups.
automatic BP monitor, using a standardized protocol.16 Hypertension was defined as systolic BP of 140 mmHg or more or diastolic BP of 90 mmHg or more, by physician diagnosis, or by use of antihypertensive medication. Nonfasting venous blood samples were sent for analysis of serum lipid levels of lowdensity lipoprotein cholesterol (LDL-C), hemoglobin A1c (HbA1c) levels, and random glucose levels. Diabetes mellitus was defined as random glucose levels of 11.1 mmol/1 or more, selfreported physician diagnosis of diabetes, self-reported use of hypoglycemic medication, or an HbA1c value of 6.5% or more.15 Hyperlipidemia was defined as total cholesterol of 6.2 mmol/1 or more or by use of lipid-lowering drugs.15 Each participant’s height was measured in centimeters using a wall-mounted measuring tape, and weight was measured in kilograms using a digital scale. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared.
Statistical Analysis Data analysis was performed using commercial analytic R software version 3.22 (R Development Core Team, Vienna, Austria). For comparison across the 3 ethnic groups, 1-way analysis of variance was used to compare continuous variables, whereas chi-square tests were used to compare categorical variables. Systemic factors, including diabetes, hypertension, LDL cholesterol, and higher BMI, were chosen based on their associations with RNFL thickness, as previously reported in other studies.14,17e19 Ocular factors evaluated in the analysis included IOP, AL, and signal strength. The effect of lens status also was evaluated, because significant change in RNFL thickness after cataract surgery has been reported previously.20 Multivariate linear regression was performed to evaluate the independent effects of age, gender, and ocular and systemic factors on RNFL thickness. The interaction between
Table 2. Comparison of Ocular Characteristics from the 3 Ethnic Groups P Value
Intraocular pressure (mmHg) RNFL thickness (mm) Average Temporal Superior Nasal Inferior ONH parameters Rim area (mm2) Disc area (mm2) Average CDR Vertical CDR Cup volume (mm3) Signal strength Myopia High myopia Axial length (mm) Long axial length (>26 mm) Cataract surgery
Chinese (n [ 2316 Eyes)
Malay (n [ 2445 Eyes)
Indian (n [ 3417 Eyes)
P Value*
Indian vs. Malay
Chinese vs. Malay
Indian vs. Chinese
14.13.0
14.63.0
15.32.8
<0.001
<0.001
0.001
<0.001
95.79.6 71.412.2 119.916.7 68.110.9 123.416.7
94.910.6 67.811.3 118.617.3 71.010.9 122.418.3
87.310.6 59.210.9 108.816.5 69.111.1 112.317.3
<0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
0.03 <0.001 0.02 <0.001 0.06
<0.001 <0.001 <0.001 0.01 <0.001
1.270.23 1.970.38 0.550.15 0.500.14 0.180.15 8.21.1 844 (36.7) 137 (6.0) 24.01.3 171 (7.4) 117 (5.1)
1.340.26 2.050.41 0.550.15 0.510.14 0.190.17 8.11.1 581 (24.6) 58 (2.5) 23.61.0 57 (2.3) 135 (5.5)
1.230.23 1.980.38 0.580.14 0.550.13 0.230.19 7.81.1 741 (24.1) 49 (1.6) 23.41.1 82 (2.4) 457 (13.4)
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.58 0.09 <0.001 0.84 <0.001
<0.001 <0.001 0.39 0.05 0.16 <0.001 <0.001 <0.001 <0.001 <0.001 0.43
<0.001 0.52 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
CDR ¼ cup-to-disc ratio; ONH ¼ optic nerve head; RNFL ¼ retinal nerve fiber thickness. Data are mean standard deviation for continuous variables and number (%) for categorical variables. *Difference across the 3 groups.
3
Ophthalmology Volume -, Number -, Month 2019 age and ethnicity on RNFL thickness was assessed by including ethnicity, age, and the product of the 2 in the regression model. Generalized estimating equations were applied in the linear regression model to account for intereye correlation within individuals. Sensitivity analyses were performed in the following 3 subsets of data: (1) excluding pseudophakic eyes, (2) excluding high myopia or eyes with AL longer than 26 mm, and (3) including only eyes with a signal strength of more than 7. A Gaussian model was assumed for defining the 1st, 5th, and 95th percentile cutoff values for average RNFL thickness across each ethnic group.
Results The mean ages for the respective ethnic groups were 57.47.0 years for Chinese participants, 60.68.6 years for Malay participants, and 60.77.8 years for Indian participants (P < 0.001). There was a significantly greater proportion of male participants in the Indian population compared with the Chinese and Malay populations (P ¼ 0.03). Among the 3 ethnic groups, Malay participants had a higher BMI and a greater proportion of current smokers, whereas the Indian
participants had a higher proportion with hypertension, diabetes, and higher levels of LDL cholesterol (Table 1). Ocular characteristics of the participants are described in Table 2. Axial length measurements were longer in Chinese participants compared with Malay and Indian participants. The IOP measurements were greater in Indian participants compared with Chinese and Malay participants (P < 0.001). A greater proportion of Indian participants (13.4%) underwent cataract surgery compared with Chinese (5.1%) and Malay (5.5%) participants. The Indian participants had the smallest mean rim area measurements and the largest average and vertical cup-todisc ratio and mean cup volume. Signal strength was notably the lowest in Indian participants compared with Chinese and Malay participants. Average RNFL thickness was 95.79.6 mm in Chinese participants, 94.910.6 mm in Malay participants, and 87.310.6 mm in Indian participants (P < 0.001; Table 2). The normative distribution of RNFL thickness for each ethnic group is shown in Figure 1, color-coded according to the color used in Cirrus OCT.21 A significant proportion of Indian eyes with RNFL thickness that falls within the top 95% of the Indian distribution would be placed in the lower 5% of the population when
Figure 1. Normative distribution of retinal nerve fiber layer (RNFL) thickness by ethnic group. White represents the thickest 5% of measurements (95th percentile), yellow represents the thinnest 5% of measurements (5th percentile), and red represents the thinnest 1% of measurements (1st percentile). Red line represents that cutoff value for 5% or less of Chinese RNFL thickness (reference).
4
Ho et al
RNFL Thickness among Asian Ethnic Groups
considering the Chinese distribution for RNFL thickness. The average RNFL thickness in percentiles across each ethnic group based on 10-year age brackets is shown in Figure 2, and the RNFL thickness for the lowest first percentile and fifth percentile by age and ethnic groups are shown in Supplemental Material S2 (available at www.aaojournal.org). Comparing RNFL thickness by quadrant among the 3 ethnic groups, Indian participants showed thinner RNFL in all except for the nasal quadrant when compared with Chinese and Malay participants (Fig 3A). A general trend toward thinner RNFL was noted with older age (Fig 3B) across all ethnic groups (b ¼ e1.26, e1.72, and e1.31 for Chinese, Malay, and Indian participants, respectively; P < 0.001 for all, where b represents the change in RNFL thickness in micrometers with every 5-year increase in age). In Table 3, multivariate linear regression models were used to examine the independent associations between ocular (IOP, AL, previous cataract surgery, disc area, and signal strength) and systemic (BMI, hypertension and diabetes status, and LDL cholesterol) parameters with average RNFL thickness. In model 1, after adjustment for age, gender, and ethnicity, reduction in RNFL thickness was found to be associated with increased IOP (b ¼ e0.20; 95% CI, e0.29 to 0.11), AL (b ¼ e1.89; 95% CI, e2.14 to e1.65), and diabetes (e0.88; 95% CI, e1.54 to e0.22), and greater RNFL thickness was found to be associated with increased disc area (b ¼ 4.10; 95% CI, 3.49e4.72), signal strength (b ¼ 1.31; 95% CI, 1.15e1.48), and LDL cholesterol levels (0.45; 95% CI, 0.15e0.74), with P < 0.001 for all. In model 2, after additional adjustments for variables identified as significant (P < 0.05) in model 1, Indian participants showed significantly thinner average RNFL (7.45 mm thinner on average; 95% CI, 6.75e8.15; P < 0.001) compared with Chinese participants, whereas there was no significant difference in average RNFL thickness between Malay and Chinese participants (P ¼ 0.15). We performed further analysis examining the interaction between age and ethnicity with regard to RNFL thickness and did not find a statistically significant interaction between the two (P ¼ 0.13). We further performed the following sensitivity analyses to evaluate potential factors that could influence the observed difference in RNFL thickness among the 3 ethnic groups (Table 4). First, we examined average RNFL thickness in phakic eyes only and found consistently lower RNFL thickness in Indian compared with Chinese and Malay participants. Second, when eyes with high myopia of (spherical equivalent less than e6.0 diopters and long AL of more than 26.0 mm were excluded), Indian participants still were found to have lower average RNFL thickness compared with Chinese and Malay participants. Finally, we included only eyes with an OCT signal strength of more than 7 in the analysis and consistently observed thinner RNFL thickness in Indians.
Discussion In this multiethnic Asian population representing the 3 largest ethnic groups in Asia (Chinese, Malays, and Indians), we demonstrated significant interethnic variations in RNFL thickness among participants without glaucoma, with Indian participants having significantly thinner average and regional RNFL thickness compared with Chinese and Malay participants. Although higher IOP, AL, and diabetes were associated with thinner average RNFL, and greater systemic LDL cholesterol levels, disc area, and signal strength were associated with thicker RNFL, these parameters did not account significantly for the ethnic differences in average
Figure 2. Graph showing the average retinal nerve fiber layer (RNFL) thickness by ethnicity based on 10-year age brackets.
and regional RNFL thickness in our study population. Although we found overlaps between Chinese and Malay RNFL data, we showed that applying the criterion of 95% or more of RNFL thickness (Fig 1, green area) based on data from Chinese eyes would mislabel a significant proportion of the Indian population (Fig 1). Our findings underscore the need for further investigations of racial differences in RNFL thickness, as well as for the development of more refined normative databases worldwide. Despite smaller patient numbers in other studies, racial differences in RNFL thickness have been demonstrated previously. In a multiethnic population with a European majority, significantly thinner RNFL has been reported among European eyes compared with that of Chinese, African, and Hispanic eyes.22 Although another study revealed no significant difference between white and black persons,23 comparatively thicker RNFL was found in Asian and Hispanic eyes.5,23 In our population, Chinese and Malay participants demonstrated similar average RNFL thickness, whereas Indian participants demonstrated significantly thinner RNFL. Compared with Chinese eyes, Indian eyes were approximately 8.18 mm thinner in average RNFL thickness, with the greatest difference in thickness of approximately 12.0 mm being measured in the inferior quadrant (Table 3A). In Malay and Indian participants, RNFL in the inferior quadrant was the thickest, followed by the superior, nasal, and temporal quadrants, which is compatible with descriptions of RNFL configurations reported previously.14 Although Chinese eyes also showed a double-hump configuration of greater inferior and superior quadrants, the nasal quadrant was found to be the thinnest among the 3 ethnic groups. The analysis of RNFL thickness with age suggested that the rate of RNFL thinning per 5-year increase in age ranged between 1.3 and 1.7 mm. The decline in RNFL thickness for every 10-year increase in age has been reported to range from 1.5 to 2.5 mm.24e26 Studies have shown that a maximum rate of reduction in RNFL thickness becomes apparent at a certain age, and this has been shown to differ
5
Ophthalmology Volume -, Number -, Month 2019
Figure 3. Graphs comparing, by ethic group, (A) regional retinal nerve fiber layer (RNFL) thickness and (B) association of average RNFL thickness with age. Distribution of RNFL thickness is measured in 4 peripapillary sectors by spectral-domain OCT. Graphs show mean and 95% confidence intervals.
by ethnicity. In the Taiwanese Chinese population, significant age-related RNFL thinning was found at 41 years of age,27 whereas the rate of thinning was faster after 50 years of age in Indian eyes.26 Comparison of age-related RNFL
6
decline in our population with that in others is difficult because of a more limited age range included in our study, especially after excluding participants to match better between ethnic groups.
Ho et al
RNFL Thickness among Asian Ethnic Groups
Table 3. Association between Ocular and Systemic Factors with Average Peripapillary Retinal Nerve Fiber Layer Thickness Model 1 Risk Factors
b*
Age (per 5 yrs) Female gender Ethnicity Malay (Chinese as reference) Indian (Chinese as reference) Body mass index (kg/m2) Hypertension (yes) Diabetes (yes) Low-density lipoprotein cholesterol (mmol/l) Current smoker Intraocular pressure (mmHg) Axial length (mm) Disc area (mm2) Signal strength History of cataract surgery
Model 2
95% Confidence Interval
P Value
b*
e1.43 0.52
e1.62 to e1.25 e0.05 to 1.09
<0.001 0.07
0.05 e7.45 e0.01 e0.36 e0.88 0.45 0.55 e0.20 e1.89 4.1 1.31 0.40
e0.69 e8.14 e0.08 e0.99 e1.54
to 0.78 to e6.77 to 0.06 to 0.27 to e0.22 0.15e0.74 e0.33 to 1.43 e0.29 to e0.11 e2.14 to e1.65 3.49e4.72 1.15e1.48 e0.51 to 1.30
0.90 <0.001 0.81 0.26 0.01 0.003 0.22 <0.001 <0.001 <0.001 <0.001 0.39
95% Confidence Interval
P Value
e1.23 e0.02
e1.41 to e1.05 e0.58 to 0.55
<0.001 0.96
e0.52 e7.45 d d e0.67 0.35 d e0.18 e1.43 3.46 1.20 d
e1.22 to 0.19 e8.15 to e6.75 d d e1.32 to e0.02 0.06e0.65 d e0.26 to e0.09 e1.69 to e1.18 2.85e4.08 1.04e1.37 d
0.15 <0.001 d d 0.04 0.02 d <0.001 <0.001 <0.001 <0.001 d
e ¼ analysis not done. *b, regression coefficient, represents the change in retinal nerve fiber layer (RNFL) thickness (in micrometers) per unit change in continuous variables, or the change in RNFL thickness (in micrometers) compared with the reference group for categorical variables. In model 1, all bs presented were adjusted for age, gender, and ethnicity, whereas in model 2, b was adjusted for age, gender, ethnicity, and variables found to be significant (P < 0.05) in model 1.
Although ocular and systemic parameters differed among our participants, these parameters did not impact significantly the observed ethnic differences in average and regional RNFL thickness. Axial length was found to be a significant determinant of average RNFL thickness in our population. The correlation between thinner RNFL and AL has been studied extensively. Differences in measured RNFL thickness in eyes with greater AL have been speculated to be caused by mechanical elongation and stretching of the sclera,28 although other reports support an ocular magnification effect resulting in factitious RNFL thinning.28e30 When a
correction factor is applied to the latter, one study showed a reversal of association to reveal thicker RNFL in participants with longer AL.31 In our study participants, Chinese eyes demonstrated longer AL and thicker RNFL compared with Indian eyes. However, the observed ethnic difference remained significant when AL was taken into account in our multivariate model (Table 3). Similarly, although disc area was reported previously to be the most important determinant of ONH and RNFL thickness measurements in Chinese persons,14 the ethnic difference in RNFL remains significant in our multivariate regression models even after
Table 4. Comparison of Average Retinal Nerve Fiber Layer Thickness by Ethnic Group All Included Eyes No. of Eyes All included eyes Chinese 2316 Malay 2445 Indian 3417 Pseudophakic eyes removed Chinese 2199 Malay 2308 Indian 2959 High myopia and long AL removed Chinese 2069 Malay 2271 Indian 2990 Including only eyes with a signal strength >7 Chinese 1750 Malay 1735 Indian 1994
Mean (Standard Deviation)
Difference* (95% Confidence Interval)
P Value
95.7 (9.6) 94.9 (10.6) 87.3 (10.6)
d e0.5 (e1.2 to 0.2) e7.5 (e8.2 to 6.8)
d 0.15 <0.001
95.8 (9.6) 95.1 (10.5) 87.8 (10.5)
d e0.4 (e1.2 to 0.3) e7.3 (e8.1 to e6.6)
d 0.60 <0.001
96.2 (9.9) 95.2 (10.7) 87.5 (10.6)
d e0.5 (e1.2 to 0.3) e7.5 (e8.3 to e6.8)
d 0.20 <0.001
96.6 (9.5) 96.5 (10.4) 88.5 (10.4)
d e0.3 (e1.1 to 0.5) e7.8 (e8.6 to e7.0)
d 0.436 <0.001
e ¼ analysis not done. *Difference estimated by linear regression models from model 2 of Table 3; Chinese is the reference ethnic group, there is no difference for comparison.
7
Ophthalmology Volume -, Number -, Month 2019 adjusting for disc area. Likewise, IOP did not affect ethnic differences significantly in average and regional RNFL thickness, as previously reported in other studies of participants with and without glaucoma.14,32 We found a small difference in OCT signal strength in the Indian participants (approximately 0.3e0.4 lower) compared with the other 2 ethnic groups. The reason is not known exactly, but a possible explanation for lower signal strength in our Indian participants could be a difference in pigmentation within the RPE of our study population, because melanin within the RPE strongly scatters light, which in turn attenuates the downstream signal that returns to the OCT detector.33 Previous studies have shown that a greater signal strength is associated with thicker RNFL measurements on OCT,34,35 which is consistent with what we found in our study (Table 3, showing that 1 unit higher in the signal strength was associated with 1.31-mm thicker RNFL on average). Nevertheless, our observed interethnic RNFL findings remained robust and similar even after excluding scans with poor signal strength (less than 7; Table 4) and adjusting for signal strength (Table 3) in the multivariate regression analysis, indicating that the difference in image quality is not the underlying explanation for thinner mean RNFL in our Indian participants. Although microvascular ischemia in vascular diseases has been proposed as a possible cause of RNFL thinning, the results of various population-based studies have yielded conflicting results regarding the association between systemic vascular disease and RNFL thickness.14,18,36e38 In our study participants, the presence of diabetes was found to be associated with thinner RNFL (P ¼ 0.04), whereas hypertension was not. In contrast, the European Eye Epidemiology consortium recently reported that hypertension (P ¼ 0.03), and not diabetes, was associated significantly with reduced RNFL in European eyes.18 Ethnic differences could account for the inconsistent findings. However, the association was significant only marginally in both cases. Association between RNFL thickness and LDL cholesterol was not assessed in the European Eye Epidemiology consortium. To our knowledge, there are no prior reports from other populations showing a significant relationship between LDL cholesterol and RNFL thickness. Thus, studies are warranted to examine these associations further. The strengths of our study include the use of 3 ethnically distinct population-based cohorts with identical study protocols. This allowed for comparisons and pooled analyses across the 3 major Asian ethnic groups. Furthermore, extensive standardized systemic and ocular examinations and investigations have allowed us to control for multiple potential confounders in our analyses. However, our study had several limitations. First, the inclusion of Chinese participants at baseline and Malay and Indian participants at the follow-up studies led to a difference in age distributions across the 3 ethnic groups. This was reduced first by removing Chinese participants younger than 48 years and Malay and Indian participants older than 82 years. Despite adjustment of age in our models, a residual significant difference in age distribution remained, which potentially may impact the comparison of RNFL profiles despite adjustment of age in our models. A large number of participants without
8
OCT data were excluded from the study (n ¼ 2026). When we compared the demographic details of participants with and without OCT data (Supplemental Material S3, available at www.aaojournal.org), we found significant differences in age, BMI, hypertension, LDL cholesterol, and current smoking status (P < 0.001 for all), which could have influenced our results. Because some eyes have undergone cataract surgery, some high myopia cases might have been missed. We chose AL over spherical equivalence in our analysis because this is not influenced by cataract surgery. Finally, ocular magnification was not corrected for by OCT, which could influence the accuracy of RNFL measurements. In summary, we demonstrated significantly thinner average and regional RNFL thicknesses in Indian participants compared with Chinese and Malay participants without glaucoma. These differences were not related to systemic or ocular parameters. Our results make a strong global argument for the need to acquire more refined normative data for comparing individual patients with others of similar racial background beyond Asia.
References 1. Tham Y-C, Li X, Wong TY, et al. Global prevalence of glaucoma and projections of glaucoma burden through 2040. Ophthalmology. 2014;121(11):2081e2090. 2. Kuang TM, Zhang C, Zangwill LM, et al. Estimating lead time gained by optical coherence tomography in detecting glaucoma before development of visual field defects. Ophthalmology. 2015;122(10):2002e2009. 3. Varma R, Steinmann WC, Scott IU. Expert agreement in evaluating the optic disc for glaucoma. Ophthalmology. 1992;99(2):215e221. 4. Leung CK-S, Yu M, Weinreb RN, et al. Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography. Ophthalmology. 2012;119(9):1858e1866. 5. Girkin CA, McGwin G, Sinai MJ, et al. Variation in optic nerve and macular structure with age and race with spectraldomain optical coherence tomography. Ophthalmology. 2011;118(12):2403e2408. 6. Zangwill LM, Weinreb RN, Berry CC, et al. Racial differences in optic disc topography: baseline results from the confocal scanning laser ophthalmoscopy ancillary study to the ocular hypertension treatment study. Arch Ophthalmol. 2004;122(1): 22e28. 7. He L, Tuomilehto J, Qiao Q, et al. Impact of classical risk factors of type 2 diabetes among Asian Indian, Chinese and Japanese populations. Diabetes Metab. 2015;41(5):401e409. 8. Pan CW, Zheng YF, Anuar AR, et al. Prevalence of refractive errors in a multiethnic Asian population: the Singapore Epidemiology of Eye Disease Study. Invest Ophthalmol Vis Sci. 2013;54(4):2590e2598. 9. Foong AWP, Saw S-M, Loo J-L, et al. Rationale and methodology for a population-based study of eye diseases in Malay people: the Singapore Malay Eye Study (SiMES). Ophthalmic Epidemiol. 2007;14(1):25e35. 10. Sabanayagam C, Yip W, Gupta P, et al. Singapore Indian Eye Study-2: methodology and impact of migration on systemic and eye outcomes. Clin Exp Ophthalmol. 2017;45(8):779e789. 11. Rosman M, Zheng Y, Wong W, et al. Singapore Malay Eye Study: rationale and methodology of 6-year follow-up study (SiMES-2). Clin Exp Ophthalmol. 2012;40(6):557e568.
Ho et al
RNFL Thickness among Asian Ethnic Groups
12. Lavanya R, Jeganathan VSE, Zheng Y, et al. Methodology of the Singapore Indian Chinese Cohort (SICC) eye study: quantifying ethnic variations in the epidemiology of eye diseases in Asians. Ophthalmic Epidemiol. 2009;16(6):325e336. 13. Foster PJ, Buhrmann R, Quigley HA, Johnson GJ. The definition and classification of glaucoma in prevalence surveys. Br J Ophthalmol. 2002;86(2):238e242. 14. Cheung CY, Chen D, Wong TY, et al. Determinants of quantitative optic nerve measurements using spectral domain optical coherence tomography in a population-based sample of non-glaucomatous subjects. Invest Ophthalmol Vis Sci. 2011;52(13):9629e9635. 15. Su DHW, Wong TY, Wong W-L, et al. Diabetes, hyperglycemia, and central corneal thickness: the Singapore Malay Eye Study. Ophthalmology. 2008;115(6):964e968.e1. 16. Kramer H, Han C, Post W, et al. Racial/ethnic differences in hypertension and hypertension treatment and control in the Multi-Ethnic Study of Atherosclerosis (MESA). Am J Hypertens. 2004;17(10):963e970. 17. Khawaja AP, Chan MPY, Garway-Heath DF, et al. Associations with retinal nerve fiber layer measures in the EPIC-Norfolk eye study. Invest Ophthalmol Vis Sci. 2013;54(7):5028e5034. 18. Mauschitz MM, Bonnemaijer PWM, Diers K, et al. Systemic and ocular determinants of peripapillary retinal nerve fiber layer thickness measurements in the European Eye Epidemiology (E3) population. Ophthalmology. 2018;125(10):1526e1536. 19. Embong Zunaina YN-S, Zunaina E, Shatriah I, et al. Correlation of retinal nerve fibre layer thickness with HbA1c and oxidised LDL in non-proliferative diabetic retinopathy. J Diabetes Metab. 2013;4(8):1e6. 20. Mwanza JC, Bhorade AM, Sekhon N, et al. Effect of cataract and its removal on signal strength and peripapillary retinal nerve fiber layer optical coherence tomography measurements. J Glaucoma. 2011;20(1):37e43. 21. Carl Zeiss Meditec, Inc. Cirrus HD-OCT 5.0 User Manual. Dublin, CA: Carl Zeiss Meditec Inc.; 2010. 22. Knight OJ, Girkin CA, Budenz DL, et al. Effect of race, age, and axial length on optic nerve head parameters and retinal nerve fiber layer thickness measured by Cirrus HD-OCT. Arch Ophthalmol. 2012;130(3):312e318. 23. Alasil T, Wang K, Keane PA, et al. Analysis of normal retinal nerve fiber layer thickness by age, sex, and race using spectral domain optical coherence tomography. J Glaucoma. 2013;22(7):532e541. 24. Budenz DL, Anderson DR, Varma R, et al. Determinants of normal retinal nerve fiber layer thickness measured by Stratus OCT. Ophthalmology. 2007;114(6):1046e1052. 25. Bendschneider D, Tornow RP, Horn FK, et al. Retinal nerve fiber layer thickness in normals measured by spectral domain OCT. J Glaucoma. 2010;19(7):475e482.
26. Parikh RS, Parikh SR, Sekhar GC, et al. Normal age-related decay of retinal nerve fiber layer thickness. Ophthalmology. 2007;114(5):921e926. 27. Peng PH, Hsu SY, Wang WS, Ko ML. Age and axial length on peripapillary retinal nerve fiber layer thickness measured by optical coherence tomography in nonglaucomatous Taiwanese participants. PLoS One. 2017;12(6). 28. Lam DSC, Leung KS, Mohamed S, et al. Regional variations in the relationship between macular thickness measurements and myopia. Invest Ophthalmol Vis Sci. 2007;48(1):376e382. 29. Nowroozizadeh S, Cirineo N, Amini N, et al. Influence of correction of ocular magnification on spectral-domain OCT retinal nerve fiber layer measurement variability and performance. Invest Ophthalmol Vis Sci. 2014;55(6):3439e3446. 30. Huang D, Chopra V, Lu AT-H, et al. Does optic nerve head size variation affect circumpapillary retinal nerve fiber layer thickness measurement by optical coherence tomography? Invest Ophthalmol Vis Sci. 2012;53(8):4990e4997. 31. Bae SH, Kang SH, Feng CS, et al. Influence of myopia on size of optic nerve head and retinal nerve fiber layer thickness measured by spectral domain optical coherence tomography. Korean J Ophthalmol. 2016;30(5):335. 32. Chang PT, Sekhon N, Budenz DL, et al. Effect of lowering intraocular pressure on optical coherence tomography measurement of peripapillary retinal nerve fiber layer thickness. Ophthalmology. 2007;114(12):2252e2258. 33. Wolbarsht M, Walsh A, George G. Melanin, a unique biological absorber. Appl Opt. 1981;20(13):2184e2186. 34. Cheung CYL, Leung CKS, Lin D, et al. Relationship between retinal nerve fiber layer measurement and signal strength in optical coherence tomography. Ophthalmology. 2008;115(8): 1347e1351.e2. 35. Wu Z, Vazeen M, Varma R, et al. Factors associated with variability in retinal nerve fiber layer thickness measurements obtained by optical coherence tomography. Ophthalmology. 2007;114(8):1505e1512. 36. Chhablani J, Sharma A, Goud A, et al. Neurodegeneration in type 2 diabetes: evidence from spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci. 2015;56(11):6333e6338. 37. Lamparter J, Schmidtmann I, Schuster AK, et al. Association of ocular, cardiovascular, morphometric and lifestyle parameters with retinal nerve fibre layer thickness. PLoS One. 2018;13(5):e0197682. 38. Schuster AK-G, Fischer JE, Vossmerbaeumer C, Vossmerbaeumer U. Determinants of peripapillary retinal nerve fiber layer thickness regarding ocular and systemic parametersdthe MIPH Eye & Health Study. Graefes Arch Clin Exp Ophthalmol. 2016;254(10):2011e2016.
Footnotes and Financial Disclosures Originally received: June 25, 2018. Final revision: November 26, 2018. Accepted: November 27, 2018. Available online: ---.
4
Ophthalmology and Visual Science Academic Clinical Program, DukeNUS Medical School, Singapore, Republic of Singapore.
5
Manuscript no. 2018-1453.
1
Singapore Eye Research Institute, Singapore National Eye Center, Singapore, Republic of Singapore.
2
Department of Ophthalmology, St. Thomas’ Hospital, London, United Kingdom. 3
Department of Ophthalmology, National University Hospital, Singapore, Republic of Singapore.
Department of Ophthalmology and Visual Science, The Chinese University of Hong Kong, Hong Kong, China.
6
Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Republic of Singapore. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.
9
Ophthalmology Volume -, Number -, Month 2019 Supported by the National Medical Research Council, Republic of Singapore (grant nos.: NMRC/1249/2010, NMRC/CIRG/1371/2013, NMRC/CIRG/1417/2015, and NMRC/CIRG/1488/2018; and NMRC/CSASI/0012/2017 [C.-Y.C.]); and the Biomedical Research Council, Singapore, Republic of Singapore (grant no.: 08/1/35/19/550). HUMAN SUBJECTS: Human subjects were included in this study. The human ethics committees at SingHealth Centralized Institutional Review Board approved the study. All research adhered to the tenets of the Declaration of Helsinki. All participants provided informed consent. No animal subjects were included in this study. Author Contributions: Conception and design: Ho, Tham, Shi, Cheng Analysis and interpretation: Ho, Tham, Chee, Shi, Tan, K.H.Wong, Majithia, Cheung, Aung, T.Y.Wong, Cheng
10
Data collection: Ho, Tham, Chee, Shi, Cheung, Aung, T.Y.Wong, Cheng Obtained funding: Cheng, study was performed at the Singapore Eye Research Institute. Overall responsibility: Ho, Tham, Chee, Shi, Tan, K.H.Wong, Majithia, Cheung, Aung, T.Y.Wong, Cheng Abbreviations and Acronyms: AL ¼ axial length; BMI ¼ body mass index; BP ¼ blood pressure; CDR ¼ cup-to-disc ratio; HbA1c ¼ hemoglobin A1c; IOP ¼ intraocular pressure; LDL-C ¼ low density lipoprotein cholesterol; ONH ¼ optic nerve head; RNFL ¼ retinal nerve fiber layer; SD ¼ spectral-domain. Correspondence: Ching-Yu Cheng, MD, PhD, Singapore Eye Research Institute, 20 College Road, The Academia, Level 6, Singapore 169856, Republic of Singapore. E-mail: [email protected].
Glaucoma is the leading cause of irreversible vision loss worldwide.1 Because glaucoma may remain asymptomatic until an advanced stage, diagnosis and treatment frequently are delayed. Visual field deficits are detected often only by standard automated perimetry after significant loss of retinal ganglion cells.2 Although structural evaluation of optic nerve head (ONH) features is imperative in the early detection and monitoring of glaucoma, the clinical evaluation and interpretation of ONH features using funduscopy or photographs are known to vary even among glaucoma experts because of interindividual variations of ONH anatomic features.3 ª 2018 by the American Academy of Ophthalmology Published by Elsevier Inc.
In contemporary clinical practice, spectral-domain (SD) OCT is used commonly to evaluate thinning of the peripapillary retinal nerve fiber layer (RNFL) objectively and quantitatively, which is a hallmark for diagnosis of glaucoma.4 New machines and sophisticated algorithms dissect data to allow a more streamlined interpretation. However, the accuracy of what qualifies as statistically significant RNFL thinning depends on the generalizability of normative databases. In this regard, it is problematic that proprietary reference databases of healthy eyes are largely reliant on data from predominantly white populations. Although they account for 60% of glaucoma cases globally,1 Asians are https://doi.org/10.1016/j.ophtha.2018.11.031 ISSN 0161-6420/19
1
Ophthalmology Volume -, Number -, Month 2019 poorly represented in normative OCT databases. For example, in the Cirrus SD OCT RNFL normative database, only 24% of eyes were from Asians. To complicate matters further, previous studies have shown significant differences in RNFL thickness across different ethnic populations.5,6 Therefore, even the consideration of “Asian” as a single group may be insufficient because this may mask ethnic variations (e.g., between Chinese and Indians, the 2 largest Asian ethnic groups) within the heterogeneous Asian population. To date, a comprehensive interethnic evaluation among Asians has not been reported. In multiethnic countries such as Singapore, establishing an ethnic-specific normative database of RNFL thickness is especially important for the diagnosis and follow-up of glaucoma. To address this gap, we examined and described the ethnic variations in peripapillary RNFL thickness using SD OCT (Cirrus HD-OCT; Carl Zeiss Meditec, Inc., Dublin, CA) within a nonglaucomatous multiethnic Asian population. A secondary aim was to identify the systemic (e.g., diabetes, hypertension) or ocular (e.g., axial length) associations with RNFL thickness and to determine if interethnic variations in these systemic parameters may be responsible for the possible interethnic differences in RNFL thickness. For example, Indian adults have a higher prevalence of diabetes, but a lower prevalence of myopia, compared with Chinese adults.7,8 These findings may be useful in understanding ethnic differences in glaucoma prevalence. By promoting the establishment of a more refined, Asian-specific normative database, the detection of glaucoma in Asians may improve. In addition, the results of our study may be able to justify the need to acquire more refined normative data for the comparison of individual patients with others of similar racial background while accounting for ocular factors that could influence RNFL thickness.
participants from the other 2 ethnic groups. To ensure that participants from each ethnic cohort straddled similar age brackets, we removed 391 Chinese participants who were younger than 48 years and 86 Malay and 72 Indian participants who were older than 82 years. After this exclusion, OCT data of ONH and RNFL measures were available for 4879 participants (9166 eyes). Because participants may have had only unilateral OCT data available, we further excluded these participants. In addition, 181 eyes with OCT segmentation errors and poor scan quality with signal strength less than six, 285 eyes with glaucoma and other neurodegenerative diseases and 522 eyes with missing variables were excluded. This left a total of 8178 eyes (4475 participants), including 2316 Chinese eyes (1371 participants), 2445 Malay eyes (1303 participants), and 3417 Indian eyes (1801 participants) for analysis (Supplemental Material S1, available at www.aaojournal.org).
Ophthalmic Assessment Subjects underwent a comprehensive ocular examination at the Singapore Eye Research Institute. Visual acuity and subjective refraction were measured by research optometrists. Slit-lamp biomicroscopy was performed by study ophthalmologists. Intraocular pressure (IOP) measurements were obtained using a Goldmann applanation tonometer. Funduscopy was performed after pupil dilation with tropicamide 1% and phenylephrine hydrochloride 2.5%, and the optic disc was evaluated using þ78-diopter lenses at 16 magnification. Ocular biometry included axial length (AL) measurements using noncontact partial coherence interferometry. Five AL measurements were obtained and the mean was used for analysis. High myopia was defined by a spherical equivalent of less than e6.0 diopters and long AL was defined by AL of more than 26 mm. The diagnosis of glaucoma was made by a glaucoma fellowship-trained ophthalmologist according to the International Society for Geographical and Epidemiological Ophthalmology criteria based on optic disc appearance and static automated perimetry (Swedish interactive threshold algorithm standard 24-2, Humphrey Field Analyzer II; Humphrey Instruments, San Leandro, CA).13
OCT Imaging
Methods Written informed consent was obtained from each participant, and the studies adhered to the tenets of the Declaration of Helsinki. Ethical approval was obtained from the SingHealth Centralized Institutional Review Board.
Study Population We conducted a population-based study using data from the Singapore Epidemiology of Eye Diseases Study of adults 40 to 80 years of age from 3 major Asian ethnic groups: Chinese persons (the Singapore Chinese Eye Study), Indian persons (the Singapore Indian Eye Study), and Malay persons (the Singapore Malay Eye Study). The methodologies of these cohort studies have been reported in detail elsewhere.9e11 Data for the current study were derived as follows. Spectraldomain OCT was incorporated into the examination protocol of the Chinese cohort at baseline (2009e2011; n ¼ 3353; response rate, 72.8%), the Malay cohort at the 6-year follow-up (2011e2014; n ¼ 3280; response rate, 78.7%), and the Indian cohort at the 6-year follow-up (2013e2015; n ¼ 2200; response rate, 75.2%).9,12 Thus, data for the present study were derived from these 3 studies. However, the inclusion of Chinese participants at baseline, compared with Malay and Indian participants at follow-up visits, contributed to a difference among the age distributions of different ethnic groups, with Chinese participants tending to be 4 to 6 years younger than the
2
OCT imaging was performed using a commercially available SD OCT instrument (Cirrus HD-OCT; Carl Zeiss Meditec, Dublin, CA) after pupil dilation, and an optic disc scan was acquired using the optic disc cube 200200 scan protocol, covering a measurement area of 66 mm. The ONH and RNFL algorithms native to Cirrus HD-OCT were used to measure a series of ONH and RNFL parameters (disc area, neuroretinal rim area, cup volume, average and vertical cup-to-disc ratios, and average and per-quadrant peripapillary RNFL thickness) automatically by identifying Bruch’s membrane opening as the disc area, and the reference plane was determined at 200 mm above the level of Bruch’s membrane plane. All images were reviewed using the Cirrus HD-OCT review software, with RNFL and optic disc segmentation checked at this time. Detailed descriptions of the measurement algorithms have been described previously.14 Study eyes with OCT scans showing RNFL algorithm segmentation failure, signal strength less than 6, or artifacts resulting from eye movements or blinking were excluded further from the analysis.
Other Measurements Relevant sociodemographic and medical information was collected through a detailed interviewer-administered questionnaire that was conducted in the patient’s language of choice (English, Chinese, Malay, or Tamil) by bilingual interviewers. Systemic examination was performed and blood samples were collected for biochemistry analysis.9,15 Blood pressure (BP) was measured using a digital
Ho et al
RNFL Thickness among Asian Ethnic Groups
Table 1. Demographic and Systemic Characteristics among Participants from the 3 Ethnic Groups P Value Chinese (n [ 1371 Participants)
Malay (n [ 1303 Participants)
Indian (n [ 1801 Participants)
57.47.0 683 (49.8) 23.63.5 192 (14.0) 726 (53.0) 3.40.9
60.68.6 600 (46.1) 27.05.1 404 (31.0) 907 (69.7) 3.41.1
186 (13.6)
246 (18.9)
Age (yrs) Male gender Body mass index (kg/m2) Diabetes Hypertension Low-density lipoprotein cholesterol (mmol/l) Current smoker
P Value*
Indian vs. Malay
Chinese vs. Malay
Indian vs. Chinese
60.77.8 910 (50.5) 26.54.5 740 (41.1) 1170 (65.0) 3.51.0
<0.001 0.03 <0.001 <0.001 <0.001 <0.001
0.90 0.01 0.02 <0.001 0.01 0.04
<0.001 0.05 <0.001 <0.001 <0.001 0.16
<0.001 0.69 <0.001 <0.001 <0.001 <0.001
237 (13.2)
<0.001
<0.001
<0.001
0.75
Data presented are mean standard deviation for continuous variables and number (%) for categorical variables. *Difference across the 3 groups.
automatic BP monitor, using a standardized protocol.16 Hypertension was defined as systolic BP of 140 mmHg or more or diastolic BP of 90 mmHg or more, by physician diagnosis, or by use of antihypertensive medication. Nonfasting venous blood samples were sent for analysis of serum lipid levels of lowdensity lipoprotein cholesterol (LDL-C), hemoglobin A1c (HbA1c) levels, and random glucose levels. Diabetes mellitus was defined as random glucose levels of 11.1 mmol/1 or more, selfreported physician diagnosis of diabetes, self-reported use of hypoglycemic medication, or an HbA1c value of 6.5% or more.15 Hyperlipidemia was defined as total cholesterol of 6.2 mmol/1 or more or by use of lipid-lowering drugs.15 Each participant’s height was measured in centimeters using a wall-mounted measuring tape, and weight was measured in kilograms using a digital scale. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared.
Statistical Analysis Data analysis was performed using commercial analytic R software version 3.22 (R Development Core Team, Vienna, Austria). For comparison across the 3 ethnic groups, 1-way analysis of variance was used to compare continuous variables, whereas chi-square tests were used to compare categorical variables. Systemic factors, including diabetes, hypertension, LDL cholesterol, and higher BMI, were chosen based on their associations with RNFL thickness, as previously reported in other studies.14,17e19 Ocular factors evaluated in the analysis included IOP, AL, and signal strength. The effect of lens status also was evaluated, because significant change in RNFL thickness after cataract surgery has been reported previously.20 Multivariate linear regression was performed to evaluate the independent effects of age, gender, and ocular and systemic factors on RNFL thickness. The interaction between
Table 2. Comparison of Ocular Characteristics from the 3 Ethnic Groups P Value
Intraocular pressure (mmHg) RNFL thickness (mm) Average Temporal Superior Nasal Inferior ONH parameters Rim area (mm2) Disc area (mm2) Average CDR Vertical CDR Cup volume (mm3) Signal strength Myopia High myopia Axial length (mm) Long axial length (>26 mm) Cataract surgery
Chinese (n [ 2316 Eyes)
Malay (n [ 2445 Eyes)
Indian (n [ 3417 Eyes)
P Value*
Indian vs. Malay
Chinese vs. Malay
Indian vs. Chinese
14.13.0
14.63.0
15.32.8
<0.001
<0.001
0.001
<0.001
95.79.6 71.412.2 119.916.7 68.110.9 123.416.7
94.910.6 67.811.3 118.617.3 71.010.9 122.418.3
87.310.6 59.210.9 108.816.5 69.111.1 112.317.3
<0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
0.03 <0.001 0.02 <0.001 0.06
<0.001 <0.001 <0.001 0.01 <0.001
1.270.23 1.970.38 0.550.15 0.500.14 0.180.15 8.21.1 844 (36.7) 137 (6.0) 24.01.3 171 (7.4) 117 (5.1)
1.340.26 2.050.41 0.550.15 0.510.14 0.190.17 8.11.1 581 (24.6) 58 (2.5) 23.61.0 57 (2.3) 135 (5.5)
1.230.23 1.980.38 0.580.14 0.550.13 0.230.19 7.81.1 741 (24.1) 49 (1.6) 23.41.1 82 (2.4) 457 (13.4)
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.58 0.09 <0.001 0.84 <0.001
<0.001 <0.001 0.39 0.05 0.16 <0.001 <0.001 <0.001 <0.001 <0.001 0.43
<0.001 0.52 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
CDR ¼ cup-to-disc ratio; ONH ¼ optic nerve head; RNFL ¼ retinal nerve fiber thickness. Data are mean standard deviation for continuous variables and number (%) for categorical variables. *Difference across the 3 groups.
3
Ophthalmology Volume -, Number -, Month 2019 age and ethnicity on RNFL thickness was assessed by including ethnicity, age, and the product of the 2 in the regression model. Generalized estimating equations were applied in the linear regression model to account for intereye correlation within individuals. Sensitivity analyses were performed in the following 3 subsets of data: (1) excluding pseudophakic eyes, (2) excluding high myopia or eyes with AL longer than 26 mm, and (3) including only eyes with a signal strength of more than 7. A Gaussian model was assumed for defining the 1st, 5th, and 95th percentile cutoff values for average RNFL thickness across each ethnic group.
Results The mean ages for the respective ethnic groups were 57.47.0 years for Chinese participants, 60.68.6 years for Malay participants, and 60.77.8 years for Indian participants (P < 0.001). There was a significantly greater proportion of male participants in the Indian population compared with the Chinese and Malay populations (P ¼ 0.03). Among the 3 ethnic groups, Malay participants had a higher BMI and a greater proportion of current smokers, whereas the Indian
participants had a higher proportion with hypertension, diabetes, and higher levels of LDL cholesterol (Table 1). Ocular characteristics of the participants are described in Table 2. Axial length measurements were longer in Chinese participants compared with Malay and Indian participants. The IOP measurements were greater in Indian participants compared with Chinese and Malay participants (P < 0.001). A greater proportion of Indian participants (13.4%) underwent cataract surgery compared with Chinese (5.1%) and Malay (5.5%) participants. The Indian participants had the smallest mean rim area measurements and the largest average and vertical cup-todisc ratio and mean cup volume. Signal strength was notably the lowest in Indian participants compared with Chinese and Malay participants. Average RNFL thickness was 95.79.6 mm in Chinese participants, 94.910.6 mm in Malay participants, and 87.310.6 mm in Indian participants (P < 0.001; Table 2). The normative distribution of RNFL thickness for each ethnic group is shown in Figure 1, color-coded according to the color used in Cirrus OCT.21 A significant proportion of Indian eyes with RNFL thickness that falls within the top 95% of the Indian distribution would be placed in the lower 5% of the population when
Figure 1. Normative distribution of retinal nerve fiber layer (RNFL) thickness by ethnic group. White represents the thickest 5% of measurements (95th percentile), yellow represents the thinnest 5% of measurements (5th percentile), and red represents the thinnest 1% of measurements (1st percentile). Red line represents that cutoff value for 5% or less of Chinese RNFL thickness (reference).
4
Ho et al
RNFL Thickness among Asian Ethnic Groups
considering the Chinese distribution for RNFL thickness. The average RNFL thickness in percentiles across each ethnic group based on 10-year age brackets is shown in Figure 2, and the RNFL thickness for the lowest first percentile and fifth percentile by age and ethnic groups are shown in Supplemental Material S2 (available at www.aaojournal.org). Comparing RNFL thickness by quadrant among the 3 ethnic groups, Indian participants showed thinner RNFL in all except for the nasal quadrant when compared with Chinese and Malay participants (Fig 3A). A general trend toward thinner RNFL was noted with older age (Fig 3B) across all ethnic groups (b ¼ e1.26, e1.72, and e1.31 for Chinese, Malay, and Indian participants, respectively; P < 0.001 for all, where b represents the change in RNFL thickness in micrometers with every 5-year increase in age). In Table 3, multivariate linear regression models were used to examine the independent associations between ocular (IOP, AL, previous cataract surgery, disc area, and signal strength) and systemic (BMI, hypertension and diabetes status, and LDL cholesterol) parameters with average RNFL thickness. In model 1, after adjustment for age, gender, and ethnicity, reduction in RNFL thickness was found to be associated with increased IOP (b ¼ e0.20; 95% CI, e0.29 to 0.11), AL (b ¼ e1.89; 95% CI, e2.14 to e1.65), and diabetes (e0.88; 95% CI, e1.54 to e0.22), and greater RNFL thickness was found to be associated with increased disc area (b ¼ 4.10; 95% CI, 3.49e4.72), signal strength (b ¼ 1.31; 95% CI, 1.15e1.48), and LDL cholesterol levels (0.45; 95% CI, 0.15e0.74), with P < 0.001 for all. In model 2, after additional adjustments for variables identified as significant (P < 0.05) in model 1, Indian participants showed significantly thinner average RNFL (7.45 mm thinner on average; 95% CI, 6.75e8.15; P < 0.001) compared with Chinese participants, whereas there was no significant difference in average RNFL thickness between Malay and Chinese participants (P ¼ 0.15). We performed further analysis examining the interaction between age and ethnicity with regard to RNFL thickness and did not find a statistically significant interaction between the two (P ¼ 0.13). We further performed the following sensitivity analyses to evaluate potential factors that could influence the observed difference in RNFL thickness among the 3 ethnic groups (Table 4). First, we examined average RNFL thickness in phakic eyes only and found consistently lower RNFL thickness in Indian compared with Chinese and Malay participants. Second, when eyes with high myopia of (spherical equivalent less than e6.0 diopters and long AL of more than 26.0 mm were excluded), Indian participants still were found to have lower average RNFL thickness compared with Chinese and Malay participants. Finally, we included only eyes with an OCT signal strength of more than 7 in the analysis and consistently observed thinner RNFL thickness in Indians.
Discussion In this multiethnic Asian population representing the 3 largest ethnic groups in Asia (Chinese, Malays, and Indians), we demonstrated significant interethnic variations in RNFL thickness among participants without glaucoma, with Indian participants having significantly thinner average and regional RNFL thickness compared with Chinese and Malay participants. Although higher IOP, AL, and diabetes were associated with thinner average RNFL, and greater systemic LDL cholesterol levels, disc area, and signal strength were associated with thicker RNFL, these parameters did not account significantly for the ethnic differences in average
Figure 2. Graph showing the average retinal nerve fiber layer (RNFL) thickness by ethnicity based on 10-year age brackets.
and regional RNFL thickness in our study population. Although we found overlaps between Chinese and Malay RNFL data, we showed that applying the criterion of 95% or more of RNFL thickness (Fig 1, green area) based on data from Chinese eyes would mislabel a significant proportion of the Indian population (Fig 1). Our findings underscore the need for further investigations of racial differences in RNFL thickness, as well as for the development of more refined normative databases worldwide. Despite smaller patient numbers in other studies, racial differences in RNFL thickness have been demonstrated previously. In a multiethnic population with a European majority, significantly thinner RNFL has been reported among European eyes compared with that of Chinese, African, and Hispanic eyes.22 Although another study revealed no significant difference between white and black persons,23 comparatively thicker RNFL was found in Asian and Hispanic eyes.5,23 In our population, Chinese and Malay participants demonstrated similar average RNFL thickness, whereas Indian participants demonstrated significantly thinner RNFL. Compared with Chinese eyes, Indian eyes were approximately 8.18 mm thinner in average RNFL thickness, with the greatest difference in thickness of approximately 12.0 mm being measured in the inferior quadrant (Table 3A). In Malay and Indian participants, RNFL in the inferior quadrant was the thickest, followed by the superior, nasal, and temporal quadrants, which is compatible with descriptions of RNFL configurations reported previously.14 Although Chinese eyes also showed a double-hump configuration of greater inferior and superior quadrants, the nasal quadrant was found to be the thinnest among the 3 ethnic groups. The analysis of RNFL thickness with age suggested that the rate of RNFL thinning per 5-year increase in age ranged between 1.3 and 1.7 mm. The decline in RNFL thickness for every 10-year increase in age has been reported to range from 1.5 to 2.5 mm.24e26 Studies have shown that a maximum rate of reduction in RNFL thickness becomes apparent at a certain age, and this has been shown to differ
5
Ophthalmology Volume -, Number -, Month 2019
Figure 3. Graphs comparing, by ethic group, (A) regional retinal nerve fiber layer (RNFL) thickness and (B) association of average RNFL thickness with age. Distribution of RNFL thickness is measured in 4 peripapillary sectors by spectral-domain OCT. Graphs show mean and 95% confidence intervals.
by ethnicity. In the Taiwanese Chinese population, significant age-related RNFL thinning was found at 41 years of age,27 whereas the rate of thinning was faster after 50 years of age in Indian eyes.26 Comparison of age-related RNFL
6
decline in our population with that in others is difficult because of a more limited age range included in our study, especially after excluding participants to match better between ethnic groups.
Ho et al
RNFL Thickness among Asian Ethnic Groups
Table 3. Association between Ocular and Systemic Factors with Average Peripapillary Retinal Nerve Fiber Layer Thickness Model 1 Risk Factors
b*
Age (per 5 yrs) Female gender Ethnicity Malay (Chinese as reference) Indian (Chinese as reference) Body mass index (kg/m2) Hypertension (yes) Diabetes (yes) Low-density lipoprotein cholesterol (mmol/l) Current smoker Intraocular pressure (mmHg) Axial length (mm) Disc area (mm2) Signal strength History of cataract surgery
Model 2
95% Confidence Interval
P Value
b*
e1.43 0.52
e1.62 to e1.25 e0.05 to 1.09
<0.001 0.07
0.05 e7.45 e0.01 e0.36 e0.88 0.45 0.55 e0.20 e1.89 4.1 1.31 0.40
e0.69 e8.14 e0.08 e0.99 e1.54
to 0.78 to e6.77 to 0.06 to 0.27 to e0.22 0.15e0.74 e0.33 to 1.43 e0.29 to e0.11 e2.14 to e1.65 3.49e4.72 1.15e1.48 e0.51 to 1.30
0.90 <0.001 0.81 0.26 0.01 0.003 0.22 <0.001 <0.001 <0.001 <0.001 0.39
95% Confidence Interval
P Value
e1.23 e0.02
e1.41 to e1.05 e0.58 to 0.55
<0.001 0.96
e0.52 e7.45 d d e0.67 0.35 d e0.18 e1.43 3.46 1.20 d
e1.22 to 0.19 e8.15 to e6.75 d d e1.32 to e0.02 0.06e0.65 d e0.26 to e0.09 e1.69 to e1.18 2.85e4.08 1.04e1.37 d
0.15 <0.001 d d 0.04 0.02 d <0.001 <0.001 <0.001 <0.001 d
e ¼ analysis not done. *b, regression coefficient, represents the change in retinal nerve fiber layer (RNFL) thickness (in micrometers) per unit change in continuous variables, or the change in RNFL thickness (in micrometers) compared with the reference group for categorical variables. In model 1, all bs presented were adjusted for age, gender, and ethnicity, whereas in model 2, b was adjusted for age, gender, ethnicity, and variables found to be significant (P < 0.05) in model 1.
Although ocular and systemic parameters differed among our participants, these parameters did not impact significantly the observed ethnic differences in average and regional RNFL thickness. Axial length was found to be a significant determinant of average RNFL thickness in our population. The correlation between thinner RNFL and AL has been studied extensively. Differences in measured RNFL thickness in eyes with greater AL have been speculated to be caused by mechanical elongation and stretching of the sclera,28 although other reports support an ocular magnification effect resulting in factitious RNFL thinning.28e30 When a
correction factor is applied to the latter, one study showed a reversal of association to reveal thicker RNFL in participants with longer AL.31 In our study participants, Chinese eyes demonstrated longer AL and thicker RNFL compared with Indian eyes. However, the observed ethnic difference remained significant when AL was taken into account in our multivariate model (Table 3). Similarly, although disc area was reported previously to be the most important determinant of ONH and RNFL thickness measurements in Chinese persons,14 the ethnic difference in RNFL remains significant in our multivariate regression models even after
Table 4. Comparison of Average Retinal Nerve Fiber Layer Thickness by Ethnic Group All Included Eyes No. of Eyes All included eyes Chinese 2316 Malay 2445 Indian 3417 Pseudophakic eyes removed Chinese 2199 Malay 2308 Indian 2959 High myopia and long AL removed Chinese 2069 Malay 2271 Indian 2990 Including only eyes with a signal strength >7 Chinese 1750 Malay 1735 Indian 1994
Mean (Standard Deviation)
Difference* (95% Confidence Interval)
P Value
95.7 (9.6) 94.9 (10.6) 87.3 (10.6)
d e0.5 (e1.2 to 0.2) e7.5 (e8.2 to 6.8)
d 0.15 <0.001
95.8 (9.6) 95.1 (10.5) 87.8 (10.5)
d e0.4 (e1.2 to 0.3) e7.3 (e8.1 to e6.6)
d 0.60 <0.001
96.2 (9.9) 95.2 (10.7) 87.5 (10.6)
d e0.5 (e1.2 to 0.3) e7.5 (e8.3 to e6.8)
d 0.20 <0.001
96.6 (9.5) 96.5 (10.4) 88.5 (10.4)
d e0.3 (e1.1 to 0.5) e7.8 (e8.6 to e7.0)
d 0.436 <0.001
e ¼ analysis not done. *Difference estimated by linear regression models from model 2 of Table 3; Chinese is the reference ethnic group, there is no difference for comparison.
7
Ophthalmology Volume -, Number -, Month 2019 adjusting for disc area. Likewise, IOP did not affect ethnic differences significantly in average and regional RNFL thickness, as previously reported in other studies of participants with and without glaucoma.14,32 We found a small difference in OCT signal strength in the Indian participants (approximately 0.3e0.4 lower) compared with the other 2 ethnic groups. The reason is not known exactly, but a possible explanation for lower signal strength in our Indian participants could be a difference in pigmentation within the RPE of our study population, because melanin within the RPE strongly scatters light, which in turn attenuates the downstream signal that returns to the OCT detector.33 Previous studies have shown that a greater signal strength is associated with thicker RNFL measurements on OCT,34,35 which is consistent with what we found in our study (Table 3, showing that 1 unit higher in the signal strength was associated with 1.31-mm thicker RNFL on average). Nevertheless, our observed interethnic RNFL findings remained robust and similar even after excluding scans with poor signal strength (less than 7; Table 4) and adjusting for signal strength (Table 3) in the multivariate regression analysis, indicating that the difference in image quality is not the underlying explanation for thinner mean RNFL in our Indian participants. Although microvascular ischemia in vascular diseases has been proposed as a possible cause of RNFL thinning, the results of various population-based studies have yielded conflicting results regarding the association between systemic vascular disease and RNFL thickness.14,18,36e38 In our study participants, the presence of diabetes was found to be associated with thinner RNFL (P ¼ 0.04), whereas hypertension was not. In contrast, the European Eye Epidemiology consortium recently reported that hypertension (P ¼ 0.03), and not diabetes, was associated significantly with reduced RNFL in European eyes.18 Ethnic differences could account for the inconsistent findings. However, the association was significant only marginally in both cases. Association between RNFL thickness and LDL cholesterol was not assessed in the European Eye Epidemiology consortium. To our knowledge, there are no prior reports from other populations showing a significant relationship between LDL cholesterol and RNFL thickness. Thus, studies are warranted to examine these associations further. The strengths of our study include the use of 3 ethnically distinct population-based cohorts with identical study protocols. This allowed for comparisons and pooled analyses across the 3 major Asian ethnic groups. Furthermore, extensive standardized systemic and ocular examinations and investigations have allowed us to control for multiple potential confounders in our analyses. However, our study had several limitations. First, the inclusion of Chinese participants at baseline and Malay and Indian participants at the follow-up studies led to a difference in age distributions across the 3 ethnic groups. This was reduced first by removing Chinese participants younger than 48 years and Malay and Indian participants older than 82 years. Despite adjustment of age in our models, a residual significant difference in age distribution remained, which potentially may impact the comparison of RNFL profiles despite adjustment of age in our models. A large number of participants without
8
OCT data were excluded from the study (n ¼ 2026). When we compared the demographic details of participants with and without OCT data (Supplemental Material S3, available at www.aaojournal.org), we found significant differences in age, BMI, hypertension, LDL cholesterol, and current smoking status (P < 0.001 for all), which could have influenced our results. Because some eyes have undergone cataract surgery, some high myopia cases might have been missed. We chose AL over spherical equivalence in our analysis because this is not influenced by cataract surgery. Finally, ocular magnification was not corrected for by OCT, which could influence the accuracy of RNFL measurements. In summary, we demonstrated significantly thinner average and regional RNFL thicknesses in Indian participants compared with Chinese and Malay participants without glaucoma. These differences were not related to systemic or ocular parameters. Our results make a strong global argument for the need to acquire more refined normative data for comparing individual patients with others of similar racial background beyond Asia.
References 1. Tham Y-C, Li X, Wong TY, et al. Global prevalence of glaucoma and projections of glaucoma burden through 2040. Ophthalmology. 2014;121(11):2081e2090. 2. Kuang TM, Zhang C, Zangwill LM, et al. Estimating lead time gained by optical coherence tomography in detecting glaucoma before development of visual field defects. Ophthalmology. 2015;122(10):2002e2009. 3. Varma R, Steinmann WC, Scott IU. Expert agreement in evaluating the optic disc for glaucoma. Ophthalmology. 1992;99(2):215e221. 4. Leung CK-S, Yu M, Weinreb RN, et al. Retinal nerve fiber layer imaging with spectral-domain optical coherence tomography. Ophthalmology. 2012;119(9):1858e1866. 5. Girkin CA, McGwin G, Sinai MJ, et al. Variation in optic nerve and macular structure with age and race with spectraldomain optical coherence tomography. Ophthalmology. 2011;118(12):2403e2408. 6. Zangwill LM, Weinreb RN, Berry CC, et al. Racial differences in optic disc topography: baseline results from the confocal scanning laser ophthalmoscopy ancillary study to the ocular hypertension treatment study. Arch Ophthalmol. 2004;122(1): 22e28. 7. He L, Tuomilehto J, Qiao Q, et al. Impact of classical risk factors of type 2 diabetes among Asian Indian, Chinese and Japanese populations. Diabetes Metab. 2015;41(5):401e409. 8. Pan CW, Zheng YF, Anuar AR, et al. Prevalence of refractive errors in a multiethnic Asian population: the Singapore Epidemiology of Eye Disease Study. Invest Ophthalmol Vis Sci. 2013;54(4):2590e2598. 9. Foong AWP, Saw S-M, Loo J-L, et al. Rationale and methodology for a population-based study of eye diseases in Malay people: the Singapore Malay Eye Study (SiMES). Ophthalmic Epidemiol. 2007;14(1):25e35. 10. Sabanayagam C, Yip W, Gupta P, et al. Singapore Indian Eye Study-2: methodology and impact of migration on systemic and eye outcomes. Clin Exp Ophthalmol. 2017;45(8):779e789. 11. Rosman M, Zheng Y, Wong W, et al. Singapore Malay Eye Study: rationale and methodology of 6-year follow-up study (SiMES-2). Clin Exp Ophthalmol. 2012;40(6):557e568.
Ho et al
RNFL Thickness among Asian Ethnic Groups
12. Lavanya R, Jeganathan VSE, Zheng Y, et al. Methodology of the Singapore Indian Chinese Cohort (SICC) eye study: quantifying ethnic variations in the epidemiology of eye diseases in Asians. Ophthalmic Epidemiol. 2009;16(6):325e336. 13. Foster PJ, Buhrmann R, Quigley HA, Johnson GJ. The definition and classification of glaucoma in prevalence surveys. Br J Ophthalmol. 2002;86(2):238e242. 14. Cheung CY, Chen D, Wong TY, et al. Determinants of quantitative optic nerve measurements using spectral domain optical coherence tomography in a population-based sample of non-glaucomatous subjects. Invest Ophthalmol Vis Sci. 2011;52(13):9629e9635. 15. Su DHW, Wong TY, Wong W-L, et al. Diabetes, hyperglycemia, and central corneal thickness: the Singapore Malay Eye Study. Ophthalmology. 2008;115(6):964e968.e1. 16. Kramer H, Han C, Post W, et al. Racial/ethnic differences in hypertension and hypertension treatment and control in the Multi-Ethnic Study of Atherosclerosis (MESA). Am J Hypertens. 2004;17(10):963e970. 17. Khawaja AP, Chan MPY, Garway-Heath DF, et al. Associations with retinal nerve fiber layer measures in the EPIC-Norfolk eye study. Invest Ophthalmol Vis Sci. 2013;54(7):5028e5034. 18. Mauschitz MM, Bonnemaijer PWM, Diers K, et al. Systemic and ocular determinants of peripapillary retinal nerve fiber layer thickness measurements in the European Eye Epidemiology (E3) population. Ophthalmology. 2018;125(10):1526e1536. 19. Embong Zunaina YN-S, Zunaina E, Shatriah I, et al. Correlation of retinal nerve fibre layer thickness with HbA1c and oxidised LDL in non-proliferative diabetic retinopathy. J Diabetes Metab. 2013;4(8):1e6. 20. Mwanza JC, Bhorade AM, Sekhon N, et al. Effect of cataract and its removal on signal strength and peripapillary retinal nerve fiber layer optical coherence tomography measurements. J Glaucoma. 2011;20(1):37e43. 21. Carl Zeiss Meditec, Inc. Cirrus HD-OCT 5.0 User Manual. Dublin, CA: Carl Zeiss Meditec Inc.; 2010. 22. Knight OJ, Girkin CA, Budenz DL, et al. Effect of race, age, and axial length on optic nerve head parameters and retinal nerve fiber layer thickness measured by Cirrus HD-OCT. Arch Ophthalmol. 2012;130(3):312e318. 23. Alasil T, Wang K, Keane PA, et al. Analysis of normal retinal nerve fiber layer thickness by age, sex, and race using spectral domain optical coherence tomography. J Glaucoma. 2013;22(7):532e541. 24. Budenz DL, Anderson DR, Varma R, et al. Determinants of normal retinal nerve fiber layer thickness measured by Stratus OCT. Ophthalmology. 2007;114(6):1046e1052. 25. Bendschneider D, Tornow RP, Horn FK, et al. Retinal nerve fiber layer thickness in normals measured by spectral domain OCT. J Glaucoma. 2010;19(7):475e482.
26. Parikh RS, Parikh SR, Sekhar GC, et al. Normal age-related decay of retinal nerve fiber layer thickness. Ophthalmology. 2007;114(5):921e926. 27. Peng PH, Hsu SY, Wang WS, Ko ML. Age and axial length on peripapillary retinal nerve fiber layer thickness measured by optical coherence tomography in nonglaucomatous Taiwanese participants. PLoS One. 2017;12(6). 28. Lam DSC, Leung KS, Mohamed S, et al. Regional variations in the relationship between macular thickness measurements and myopia. Invest Ophthalmol Vis Sci. 2007;48(1):376e382. 29. Nowroozizadeh S, Cirineo N, Amini N, et al. Influence of correction of ocular magnification on spectral-domain OCT retinal nerve fiber layer measurement variability and performance. Invest Ophthalmol Vis Sci. 2014;55(6):3439e3446. 30. Huang D, Chopra V, Lu AT-H, et al. Does optic nerve head size variation affect circumpapillary retinal nerve fiber layer thickness measurement by optical coherence tomography? Invest Ophthalmol Vis Sci. 2012;53(8):4990e4997. 31. Bae SH, Kang SH, Feng CS, et al. Influence of myopia on size of optic nerve head and retinal nerve fiber layer thickness measured by spectral domain optical coherence tomography. Korean J Ophthalmol. 2016;30(5):335. 32. Chang PT, Sekhon N, Budenz DL, et al. Effect of lowering intraocular pressure on optical coherence tomography measurement of peripapillary retinal nerve fiber layer thickness. Ophthalmology. 2007;114(12):2252e2258. 33. Wolbarsht M, Walsh A, George G. Melanin, a unique biological absorber. Appl Opt. 1981;20(13):2184e2186. 34. Cheung CYL, Leung CKS, Lin D, et al. Relationship between retinal nerve fiber layer measurement and signal strength in optical coherence tomography. Ophthalmology. 2008;115(8): 1347e1351.e2. 35. Wu Z, Vazeen M, Varma R, et al. Factors associated with variability in retinal nerve fiber layer thickness measurements obtained by optical coherence tomography. Ophthalmology. 2007;114(8):1505e1512. 36. Chhablani J, Sharma A, Goud A, et al. Neurodegeneration in type 2 diabetes: evidence from spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci. 2015;56(11):6333e6338. 37. Lamparter J, Schmidtmann I, Schuster AK, et al. Association of ocular, cardiovascular, morphometric and lifestyle parameters with retinal nerve fibre layer thickness. PLoS One. 2018;13(5):e0197682. 38. Schuster AK-G, Fischer JE, Vossmerbaeumer C, Vossmerbaeumer U. Determinants of peripapillary retinal nerve fiber layer thickness regarding ocular and systemic parametersdthe MIPH Eye & Health Study. Graefes Arch Clin Exp Ophthalmol. 2016;254(10):2011e2016.
Footnotes and Financial Disclosures Originally received: June 25, 2018. Final revision: November 26, 2018. Accepted: November 27, 2018. Available online: ---.
4
Ophthalmology and Visual Science Academic Clinical Program, DukeNUS Medical School, Singapore, Republic of Singapore.
5
Manuscript no. 2018-1453.
1
Singapore Eye Research Institute, Singapore National Eye Center, Singapore, Republic of Singapore.
2
Department of Ophthalmology, St. Thomas’ Hospital, London, United Kingdom. 3
Department of Ophthalmology, National University Hospital, Singapore, Republic of Singapore.
Department of Ophthalmology and Visual Science, The Chinese University of Hong Kong, Hong Kong, China.
6
Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Republic of Singapore. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.
9
Ophthalmology Volume -, Number -, Month 2019 Supported by the National Medical Research Council, Republic of Singapore (grant nos.: NMRC/1249/2010, NMRC/CIRG/1371/2013, NMRC/CIRG/1417/2015, and NMRC/CIRG/1488/2018; and NMRC/CSASI/0012/2017 [C.-Y.C.]); and the Biomedical Research Council, Singapore, Republic of Singapore (grant no.: 08/1/35/19/550). HUMAN SUBJECTS: Human subjects were included in this study. The human ethics committees at SingHealth Centralized Institutional Review Board approved the study. All research adhered to the tenets of the Declaration of Helsinki. All participants provided informed consent. No animal subjects were included in this study. Author Contributions: Conception and design: Ho, Tham, Shi, Cheng Analysis and interpretation: Ho, Tham, Chee, Shi, Tan, K.H.Wong, Majithia, Cheung, Aung, T.Y.Wong, Cheng
10
Data collection: Ho, Tham, Chee, Shi, Cheung, Aung, T.Y.Wong, Cheng Obtained funding: Cheng, study was performed at the Singapore Eye Research Institute. Overall responsibility: Ho, Tham, Chee, Shi, Tan, K.H.Wong, Majithia, Cheung, Aung, T.Y.Wong, Cheng Abbreviations and Acronyms: AL ¼ axial length; BMI ¼ body mass index; BP ¼ blood pressure; CDR ¼ cup-to-disc ratio; HbA1c ¼ hemoglobin A1c; IOP ¼ intraocular pressure; LDL-C ¼ low density lipoprotein cholesterol; ONH ¼ optic nerve head; RNFL ¼ retinal nerve fiber layer; SD ¼ spectral-domain. Correspondence: Ching-Yu Cheng, MD, PhD, Singapore Eye Research Institute, 20 College Road, The Academia, Level 6, Singapore 169856, Republic of Singapore. E-mail: [email protected].