Article Text

Download PDFPDF

Racial differences and determinants of macular thickness profiles in multiethnic Asian population: the Singapore Epidemiology of Eye Diseases Study
  1. Kah Hie Wong1,2,
  2. Yih-Chung Tham1,
  3. Duc Quang Nguyen1,
  4. Wei Dai1,
  5. Nicholas Y Q Tan1,
  6. Shivani Mathijia1,
  7. Kumari Neelam1,
  8. Carol Yim-lui Cheung1,3,
  9. Charumathi Sabanayagam1,4,
  10. Leopold Schmetterer1,5,6,7,
  11. Tien Yin Wong1,2,4,
  12. Ching-Yu Cheng1,2,4
  1. 1 Singapore National Eye Centre, Singapore Eye Research Institute, Singapore, Singapore
  2. 2 Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore and National University Health System, Singapore, Singapore
  3. 3 Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China
  4. 4 Duke-NUS Medical School, Singapore, Singapore
  5. 5 Lee Kong Chian School of Medicine, Nanyang Technological University, Singapore, Singapore
  6. 6 Department of Clinical Pharmacology, Medical University of Vienna, Vienna, Austria
  7. 7 Center for Biomedical Engineering and Physics, Medical University of Vienna, Vienna, Austria
  1. Correspondence to Professor Ching-Yu Cheng, Singapore Eye Research Institute, Singapore National Eye Centre, Singapore 168751, Singapore; chingyu.cheng{at}


Aim To evaluate racial differences, and ocular and systemic determinants of macular thickness (MT), measured by spectral-domain optical coherence tomography (SD-OCT) in a normal multiethnic Asian population.

Method MT was measured from a 6×6 mm2 central macular area using the Cirrus high-definition OCT (HD-OCT) (Carl Zeiss Meditec, Dublin, CA). The associations between ocular and systemic factors with MT were evaluated using linear regression analyses with generalised estimating equation models to account for intereye correlation.

Results 7447 healthy eyes (2577 Chinese, 2072 Malays and 2798 Indians) of 4510 subjects were included. Multivariable analysis showed that older age (per decade, β=−4.39), female gender (β=−5.74), diabetes (β=−1.10), chronic kidney disease (CKD) (β=−3.21), longer axial length (per mm, β=−2.34), flatter corneal curvature (per mm, β=−1.79) and presence of cataract (β=−0.94) were associated with thinner overall average MT (OMT) (all p≤0.026); higher total cholesterol (β=0.44; p=0.010) was associated with thicker OMT. All these factors were also associated with thinner central subfield MT (CSMT) (all p≤0.001), except for cataract, total cholesterol and CKD. Meanwhile, longer axial length (β=2.51; p<0.001) was associated with thicker CSMT. OMT (mean±SD) was thickest in Chinese (279.9±12.5 µm), followed by Malays (276.5±13.7 µm) and Indians (272.4±13.1 µm), with p≤0.003 for all interethnic comparisons. Similar trend was observed for CSMT.

Conclusion There are interethnic differences in MT profile among Asians, particularly between Chinese and Indians. Ocular and systemic factors affect MT measurements as well. This Asian-specific information may be incorporated into existing clinical interpretation of macular OCT scans to aid in improving the diagnostic and monitoring accuracy of macular diseases among Asians.

  • retina
  • macula
  • imaging

Statistics from


Optical coherence tomography (OCT) has become one of the key imaging modalities used in the clinical ophthalmology. Macular thickness (MT) profiles are a common tool for the diagnosis and monitoring of macular diseases, such as age-related macular degeneration and diabetic macular oedema. Therefore, it is crucial to have a comprehensive understanding on the MT profiles within a normal population. However, the current built-in software of Cirrus macular OCTs is based on an age-adjusted normative database that consisted of 282 subjects (aged 19–84), which were mainly Caucasians and may not well represent Asian populations.1 Furthermore, in current OCT review software, Asian normative database is presented and evaluated as a whole without further taking into account the interethnic difference among Asians.1 This approach is flawed considering the heterogeneity among Asians, and may potentially result in inaccurate clinical assessment when detecting or monitoring macular diseases. Nevertheless, heterogeneity of MT among Asians is currently not well understood. In addition, apart from age, the existing normative database did not take into account any other factors such as gender, which had shown consistent association with MT in previously reported population-based studies.2–5

To date, population data on OCT-measured MT profiles are limited. In addition, there are very few studies which investigated the factors associated with MT.2–4 More importantly, no previous studies investigated the racial differences of MT profiles among Asians. In the previous UK Biobank Study,4 despite its large sample size (n=32 062), the sample size of Asians was relatively small (n=898) as compared with the White subjects, thus did not allow for comprehensive interethnic evaluation among Asians. Taken together, there is currently no clear knowledge on the racial difference of MT among Asians. This information is important to support the role of OCT as the diagnostic and monitoring tool of MT in Asian populations which comprise multiple ethnicities.

In this regard, Singapore Epidemiology of Eye Diseases (SEED) Study is a cross-sectional population-based study that included a large sample size of multiethnic adult Asian populations, consisting of Chinese, Malays and Indians,6–8 providing a unique opportunity to perform comprehensive interethnic evaluation in a relatively standardised study setting. Hence, the aim of this study was to evaluate the racial difference and determinants of MT measured by spectral-domain OCT (SD-OCT). Findings from this study may potentially benefit clinical interpretation of OCT scans, which may in turn help to improve detection and monitoring of macular diseases.


Study population

Study subjects were enrolled from the SEED Study, comprising three major ethnic groups in Singapore, including Chinese, Malay and Indians. Methodology and details of the SEED Study were reported previously.6–8 The study was conducted in Singapore using a standardised study protocol across the three ethnic groups of subjects. The data for this study were derived from 3353 Chinese participants (year 2009–2011, response rate 72.8%),5 1901 Malays (year 2010–2014, response rate 72.1%)9 and 2200 Indians (year 2013–2015, response rate 75.5%).8

Ocular examinations

Every participant underwent standardised ocular examinations. Subjective refraction was performed. Intraocular pressure (IOP) was measured with Goldmann applanation tonometer (Haag-Streit, Bern, Switzerland) before pupil dilation. Lens opacities were graded using Lens Opacity Classification System (LOCS III) during slit-lamp examinations.10 The presence of cataract was defined as nuclear opalescence ≥ grade 4, nuclear colour ≥ grade 4, cortical cataract ≥ grade 2 or posterior subcapsular cataract ≥ grade 2, based on LOCS III grading.10Central corneal thickness (CCT) was measured using an ultrasound pachymeter (Advent; Mentor O & O, Norwell, MA).6–8 Axial length was measured with non-contact partial coherence laser interferometer (IOL Master V.3.01; Carl Zeiss Meditec, Jena, Germany).

OCT imaging

The subjects underwent Cirrus high-definition OCT (HD-OCT) (software V.6.0; Carl Zeiss Meditec) imaging after pupil dilation. A macular cube scan of an area of 6×6 mm2 (online supplementary figure 1A) was acquired in each study eye using the macular cube scan 512×128 protocol. MT maps were generated by the built-in algorithm with nine macular subfields presented in a 6 mm diameter circle that was centred at the fovea, as defined by the Early Treatment Diabetic Retinopathy Study.11

MT was measured from the internal limiting membrane to retinal pigment epithelium (online supplementary figure 1B). The standard nine macular subfields included (1) central subfield, (2) inner macular subfields divided into four quadrants (inner superior, nasal, inferior and temporal), and (3) outer macular subfields segregated into four quadrants (outer superior, nasal, inferior and temporal) (online supplementary figure 1C). Central subfield, inner macular subfield and outer macular subfield were defined as the area within circles of the radii 0.5 mm, 0.5–1.5 mm and 1.5–3 mm, respectively. Average inner and outer MT were defined as the average of MT in the four quadrants of inner and outer macula, respectively. The built-in computational software was used to calculate overall average MT (OMT) which averaged MT measurements across the entire measurement area (6×6 mm2).5

OCT image evaluation

Quality control was performed to isolate OCT scans with imaging errors or artefacts that might affect MT measurements, for example, poor signal strength (<6), scan misalignment, motion artefact and segmentation error. These details had been reported elsewhere previously.5

Systemic examinations

The body mass index (BMI) was calculated as body weight (in kilograms) divided by body height (in metres) squared. Smoking status was assessed through self-reported questionnaires and defined as current and non-current smokers. Diabetes mellitus (DM) was defined by random glucose ≥11.1, glycosylated haemoglobin (HbA1C) ≥6.5%, use of diabetic medication(s) or self-reported history. Hypertension was defined by systolic blood pressure (BP) ≥140 mm Hg, diastolic BP ≥90 mm Hg, physician’s diagnosis or self-reported history of hypertension. Kidney function was assessed using estimated glomerular filtration rate (GFR) from serum creatinine using the Chronic Kidney Disease-Epidemiology Collaboration equation.8 Subjects were defined to have chronic kidney disease (CKD) if GFR was <60 mL/min/1.73 m2.

Non-fasting venous blood samples were collected for biochemistry tests including plasma cholesterol (total cholesterol, low-density lipoprotein and high-density lipoprotein (HDL)), serum triglyceride, HbA1c, creatinine and random glucose.

Inclusion and exclusion criteria

Only subjects who underwent OCT imaging were initially included. We further excluded subjects with neurodegenerative diseases, followed by eyes with low vision (best corrected visual acuity worse than 20/40); eyes with macular or vitreoretinal diseases, including epiretinal membrane, diabetic retinopathy, macular oedema, age-related macular degeneration, glaucoma and other retinopathy that might affect the retinal thickness; eyes with previous retinal surgery or laser treatment; and eyes with suboptimal OCT scan quality (as described above).

Statistical analysis

For individual-level analysis, one-way analysis of variance (ANOVA) was performed to compare continuous variables among ethnic groups, and Χ2 tests were used for categorical variables. For eye-level analysis, repeated measures ANOVA was performed to compare continuous variables across the three ethnic groups, and logistic regression model with estimating equation was used to compare categorical variables.

MT measurements from the central subfield, inner, outer, and the overall 6×6 mm2 macular areas (online supplementary figure 1) were evaluated as the outcome variables separately. Associations between ocular and systematic factors with MT were investigated using univariate and multivariable linear regression analyses, with generalised estimating equations to account for correlation between eyes. Ocular and systemic parameters with p<0.10 found in univariate linear regression analysis were subsequently included in the multivariable regression model.

To compare MT profiles across the three ethnicities, the comparisons were adjusted for relevant ocular and systematic factors found to be significantly associated with MT parameters. All statistical analyses were performed using Stata V.14 (StataCorp, College Station, TX).


Of the original 10 049 eyes of 5333 subjects who underwent OCT imaging, we excluded eyes with low vision (n=592), presence of any retinopathy (n=1632), glaucoma (n=164), poor OCT image quality (n=212) and eyes from subjects with neurodegenerative diseases (n=2). This left 7447 eyes from 4510 subjects (1661 Chinese, 1218 Malays and 1631 Indians) included for the final analysis.

Online supplementary table 1 summarises the demographic and systemic characteristics of included subjects. The mean age±SD of Chinese, Malay and Indian subjects were 55.3±7.6, 60.5±8.9 and 60.4±7.8 years, respectively. The gender distribution was similar across all ethnicities. There was higher proportion of hypertension and DM in Indians, and higher proportion of CKD and current smokers in Malays. Malays have the highest BMI level whereas Chinese have the highest total cholesterol level (all p<0.001).

Online supplementary table 2 shows the ocular characteristics of the included eyes across all ethnicities. Chinese were more myopic, had longer axial length, thicker CCT and lower IOP compared with Malays and Indians (all p<0.001). More Indians had cataracts and had undergone cataract operations.

Online supplementary table 3 shows the univariate analyses of the ocular and systemic factors with OMT and central subfield MT (CSMT). Online supplementary table 4 showed the univariate analyses between average inner and average outer MT with ocular and systemic factors. Ocular and systemic factors significantly associated with MT parameters (p<0.10) in univariate analyses were then included in the subsequent multivariable analysis. Only axial length, but not spherical equivalent, was included in the multivariable model due to collinearity between them.

Multivariable analyses in table 1 show that older age (per decade, β=−4.39; p<0.001), female gender (β=−5.74; p<0.001), Malays (compared with Chinese, β=−1.55; p=0.004), Indians (compared with Chinese, β=−5.67; p<0.001), diabetes (β=−1.10; p=0.026), CKD (β=−3.21; p<0.001), longer axial length (per mm, β=−2.34; p<0.001), flatter corneal curvature (per mm, β=−1.79; p=0.018), presence of cataract (β=−0.94; p=0.005) and higher signal strength (β=−0.94, p<0.001) were associated with thinner OMT. Higher total cholesterol (per mmol/L, β=0.44; p=0.010) was associated with thicker OMT. On the other hand, older age, female gender, Indians, diabetes, flatter corneal curvature and higher signal strength were significantly associated with thinner CSMT (all p≤0.001), whereas higher HDL (p=0.045) and longer axial length (p<0.001) were significantly associated with thicker CSMT. Similarly, the above-mentioned determinants were also significantly associated with average inner and outer MT (online supplementary table 5). Figures 1 and 2 further illustrate the associations between age and axial length with OMT and CSMT, respectively, across the three ethnic groups.

Figure 1

(A) Overall average macular thickness and (B) central macular thickness by age (years) and ethnic groups. Data presented are in mean±SE.

Figure 2

(A) Overall average macular thickness and (B) central macular thickness by axial length and ethnic groups. Data presented are in mean±SE.

Table 1

Multivariable analyses of associations between systemic and ocular factors with overall average and central subfield macular thickness parameters

Table 2 presents the distribution of all MT parameters in Chinese, Malays and Indians. We observed that Chinese generally have thicker MT profiles compared with Malays and Indians, whereas Malays have thicker macula than Indians. After further adjusting for age, gender, ethnicity, DM, total cholesterol, CKD, axial length, corneal curvature, presence of cataract and signal strength, these interethnic differences were still significant (all p≤0.042). As there were more Chinese in the age group below 50 years, and more Indians and Malays in the age group above 80 years, there might be residual confounding effect by age which may impact the comparison of MT profiles between ethnic groups. To minimise this potential residual confounding effect, we further performed a sensitivity analysis limited to subjects aged 50–80 years but still observed significant differences across ethnic groups (all p≤0.030), except for CSMT comparison between Malays and Indians. Similar trends of ethnic differences were also observed for the quadrants of inner and outer MT (online supplementary table 6).

Table 2

Distribution of macular thickness profiles in overall sample and respective ethnic groups

Table 3 shows the summary of determinant findings of previously reported population-based studies.2–4 In comparison with the previously reported studies, our study revealed consistent associations between MT with age, female gender, myopia, axial length and DM. On top of that, we have discovered novel associations between MT with ethnic difference, corneal curvature, total cholesterol and CKD.

Table 3

Summary of determinants of macular thickness parameters in SEED and previous population-based studies


In this large multiethnic Asian population, we found that Chinese and Malays have thicker macular profiles compared with Indians. Older age, female gender, diabetes, CKD, longer axial length, flatter corneal curvature, cataracts and higher signal strength were significantly associated with thinner overall macula, whereas higher total cholesterol was significantly associated with thicker overall macula. This is the first population-based study that comprehensively assessed MT profiles in Asians, providing new and valuable information on interethnic differences among Asians. Our comprehensive findings will be useful in establishing a more refined, Asian-specific normative database in MT measured from SD-OCT, which may in turn help to improve the diagnosis and monitoring of macular diseases in Asians.

Our study found ethnic differences of MT between Singaporean Chinese, Malays and Indians. Previous studies have reported on racial differences, but either had a small sample size,12 13 or a sample with an unbalanced proportion of different ethnic groups. These limitations may have introduced biases.4 On the contrary, our study’s large sample size and comparable numbers of each ethnic group have overcome the previous shortcomings and successfully proven the ethnic differences of MT in Asians. In particular, the differences in MT between Indians and Chinese were found to be statistically significant in overall, central subfield, average inner and average outer macula, and potentially clinically relevant due to its high standardised β values. Our findings may contribute to the normative database of MT of each ethnicity and introduce the concept of ethnic differences while interpreting macular OCTs. Across all three ethnicities, nasal quadrants of central subfield, inner macula and outer macula are the thickest. This is consistent in previous studies2–5 and is compatible with the known anatomic relation of papillomacular bundle of retinal nerve fibre layer (RNFL) and the converging nerve fibre from the rest of the retina to optic nerve.

Ageing has been found to be strongly and consistently related to thinner overall, central subfield, inner and outer macula.2–5 This may be the normal end of physiological spectrum, whereas those who reached the pathological spectrum via the overactivation of complement pathway would result in geographic atrophy.14 The understanding of effect of ageing on MT may contribute to the reduction of rate of macular thinning process and the progression to geographic atrophy. This, however, requires longitudinal studies to shed more light.

Previous reports have consistently shown that women consistently have thinner overall, central subfield, inner and outer macula than men in Chinese and Caucasians.2–5 In animal model, females were found to have thinner retina than males, and this was proposed to be due to females’ higher ratio of parvocellular retina ganglion cells to magnocellular retina ganglion cells, which are thicker and larger.15 In our study, female gender has the highest standardised β in the multivariable analysis for overall, central subfield, inner and outer macula. This shows that gender difference plays a significant role in the variations of MT and the incorporation of gender factor in normative database may be improving the diagnostic and monitoring performance of OCT.

Diabetics without diabetic retinopathy (DR) have thinner overall, central subfield and inner macula. Previous studies have reported thinner macular ganglion cell-inner plexiform layers,16 17 inner plexiform layer,18 RNFL17 and inner retinal layer18 in eyes of diabetics without DR, as compared with healthy controls. This suggests that neuroretinal degeneration may occur early in diabetics and may precede the microvascular damages.

High total cholesterol was associated with thicker overall, inner and outer macula. In animal models fed with cholesterol-enriched diet, lipid was found to accumulate in the retina.19 This possibly explains our observed finding in this study.

Eyes of CKD subjects were found to have thinner macula in our study. Previous study similarly reported that subjects with end-stage renal failure had thinner macula and RNFL than healthy controls, with ischaemia cited as a possible reason.20 Nevertheless, the exact mechanism of macular thinning in patients with CKD remains largely unclear. Further evaluation in this aspect is still required.

Myopia has thicker central subfield macula, but thinner overall, inner and outer macula, similar to previous studies.3 4 21 This may be explained by the thinning of RNFL, which is higher in myopia than emmetropia and increasingly thinner with severity of myopia22; photoreceptor cell degeneration via apoptosis in pathologic myopia23; and the mechanical forces. Increasing axial length in patients with myopia results in mechanical stretching of sclera and hence retina thinning.21 On the other hand, the elevation of foveola and fovea may be explained by the stretching and flattening tendency of the internal limiting membrane and the centripetal force of the posterior vitreous.21

Flatter corneal curvature was associated with thinner overall, central subfield and inner MT. This may be due to ocular magnification effect.24 In addition, the presence of cataract was associated with thinner overall, central subfield, inner and outer macula. Similar finding was also reported in Beaver Dam Eye Study, although not statistically significant.3

The strengths of our study are our large Asian population-based sample size of study eyes without retinal diseases with the composition of three main Asian ethnicities. In addition, our study adopted standardised assessments of a comprehensive range of ocular, systemic and biochemical factors, thus allowing relatively direct and objective comparisons across the three ethnic groups. The age group of our study sample involved the age range of which common macular diseases are most prevalent, hence increasing the generalisability of our findings to the normative database for comparison with diseased macula of each ethnicity. Our study has some limitations. Due to the cross-sectional nature of our study, the causal associations between the determined factors and MT cannot be ascertained. Further longitudinal evaluation is required to elucidate this aspect. In addition, there might be residual confounding of the systemic factors that were not yet discovered and thus not taken into account in our analysis.


In this multiethnic Asian population, Chinese have the thickest macula, and Indians the thinnest. Older age, female gender, diabetes, CKD, longer axial length, flatter corneal curvature and cataract were significantly associated with thinner OMT, whereas higher total cholesterol was associated with thicker OMT. Collectively, this information may be taken into account in establishing a more refined, Asian-specific normative database for MT measurements from SD-OCT, which may in turn improve the diagnostic and monitoring performances of OCT for macular diseases among Asians.



  • KHW and Y-CT are joint first authors.

  • Contributors Conception and design: KHW, YCT, TYW, CYC. Data collection: KHW, YCT, WD, NYQT, SM, KN, CYLC, CS, CYC. Analysis and interpretation: KHW, YCT, DQN, CS, LS, TYW, CYC. Drafting of manuscript: KHW, YCT, CYC. Final revision of manuscript: KHW, YCT, DQN, WD, NYQT, SM, KN, CYLC, CS, LS, TYW, CYC.

  • Funding The study is funded by National Medical Research Council (grants 0796/2003, IRG07nov013, IRG09nov014, STaR/0003/2008; CG/SERI/2010) and Biomedical Research Council (grants 08/1/35/19/550, 09/1/35/19/616), Singapore. The sponsor or funding organization had no role in the design or conduct of this research. C-YC is supported by National Medical Research Council (NMRC/CSA/033/2012).

  • Competing interests None declared.

  • Patient consent Not required.

  • Ethics approval SingHealth Centralised Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost B Jonas