Cohort Profile: The Singapore Epidemiology of Eye Diseases study (SEED)

Abstract

Globally, blindness and visual impairment (VI) are estimated to affect 36 million and 400 million people, respectively.1 Asia is solely responsible for 57.5% of global blindness and 63.4% of global VI.2 Additionally, Asians are disproportionally affected by major age-related eye diseases, such as 60% of glaucoma3 and 35% of age-related macular degeneration (AMD) cases globally.4 As Asia possesses one of the world’s fastest ageing populations, we expect to see an increase in VI and age-related eye diseases in the coming decades. Notably, vision loss is ranked as one of the top three impairments for years lived with disability (YLD),5 thereby greatly impacting an individual’s quality of life. Consequently, age-related eye diseases and VI represent a great burden globally and especially for Asians. Despite increasing concern of a growing ageing Asian population, there is no multi-ethnic longitudinal study where direct inter-ethnic comparisons can be made on the incidence and progression of major age-related eye diseases. Although several Asian studies have reported on prevalence and risk factors associated with major age-related eye diseases, there are limited studies that longitudinally evaluate incidence, progression and long-term risk factors for these diseases.6–13 Additionally, it is known that risk factors for many diseases vary by ethnicity.14–17 However, comparing studies by country or ethnicity provides limited information due to differences in study design, methodology, disease definitions and socio-economic status between countries. Hence, accurate reporting on ethnic differences in eye diseases among Asians remains an important knowledge gap. The Singapore Epidemiology of Eye Diseases (SEED) cohort is a multi-ethnic longitudinal population-based study that evaluates the incidence, prevalence, risk factors and novel biomarkers of age-related eye diseases for Singaporean adults of Malay, Indian and Chinese descent. Singapore provides a unique opportunity where multiple ethnicities can be studied in a relatively uniform environment with similar exposures across the three ethnicities. Additionally, because each ethnicity was evaluated using a standardized examination protocol, SEED allows for a more direct evaluation and comparison of ethnic differences in the incidence and progression rates of major age-related eye diseases. Furthermore, SEED helps in the formulation of normative databases for these three ethnicities as well as guiding public health policy decisions and strategies for screening, diagnosis and early detection of age-related eye diseases not only for Singapore, but for other Asian countries as well. The SEED cohort is based on an age-stratified random sampling strategy of individuals between the ages of 40 and 80+ years from 15 residential districts in the South-Western part of Singapore (Supplementary Figure S1, available as Supplementary data at IJE online). This area was chosen based on the Singapore Population Census 2000, where the residents in this area were a fair representation of the Singapore population in terms of age, housing and socio-economic status. The Ministry of Home Affairs provided a list of 16 069 ethnic Malays, 12 000 ethnic Indians and 12 000 ethnic Chinese. From this list, we derived a final sampling frame of 5600 Malay, 6350 Indian and 6350 Chinese Singaporean residents. A total of 10 033 Singaporean adults participated in the baseline SEED examination. The characteristics of participants who attended SEED-1 can be found in Table 1. The mean age of all baseline participants was 58.9 years, included 50.7% female participants and had almost equal distributions of participants between all three ethnic groups (3280 Malays, 3400 Indians and 3353 Chinese). At baseline, 6393 participants had hypertension (63.9%), 2964 participants had diabetes (29.6%) and the overall spherical equivalent was −0.32 dioptres. A total of 6762 participants (response rate: 78.8%) returned for the 6-year follow-up examination (SEED-2) comprising 1901 Malays, 2200 Indians and 2661 Chinese (Table 2). The overall mean age for SEED-2 was 63.5 years and the study included 52% female participants (Table 2). Additionally, 4698 participants had hypertension (69.8%), 2091 had diabetes (30.9%) and the overall spherical equivalent was −0.21 dioptres. Other participant characteristics for SEED-2 can be found in Table 2. Characteristics of participants who attended SEED-1; data presented are mean (standard deviation) or frequency (%), where appropriate LDL, low-density lipoprotein; SES, socio-economic status. P-value denotes ANOVA comparisons across the three ethnic groups. Education level: O-level, Singapore Cambridge General Certificate of Education, ordinary level; A-level, Singapore Cambridge General Certificate of Education, advanced level. Low SES, primary or lower education, individual monthly income <S$2000 and 1–2 room Housing and Development Board flat. Characteristics of participants who attended SEED-1; data presented are mean (standard deviation) or frequency (%), where appropriate LDL, low-density lipoprotein; SES, socio-economic status. P-value denotes ANOVA comparisons across the three ethnic groups. Education level: O-level, Singapore Cambridge General Certificate of Education, ordinary level; A-level, Singapore Cambridge General Certificate of Education, advanced level. Low SES, primary or lower education, individual monthly income <S$2000 and 1–2 room Housing and Development Board flat. Characteristics of participants who attended SEED-2; data presented are mean (standard deviation) or frequency (%), where appropriate HDL, high-density lipoprotein; LDL, low-density lipoprotein; SES, socio-economic status. P-value denotes ANOVA comparisons across the three ethnic groups. Education level: O-level, Singapore Cambridge General Certificate of Education, ordinary level; A-level, Singapore Cambridge General Certificate of Education, advanced level. Low SES: primary or lower education, individual monthly income <S$2000 and 1–2 room Housing and Development Board flat. Characteristics of participants who attended SEED-2; data presented are mean (standard deviation) or frequency (%), where appropriate HDL, high-density lipoprotein; LDL, low-density lipoprotein; SES, socio-economic status. P-value denotes ANOVA comparisons across the three ethnic groups. Education level: O-level, Singapore Cambridge General Certificate of Education, ordinary level; A-level, Singapore Cambridge General Certificate of Education, advanced level. Low SES: primary or lower education, individual monthly income <S$2000 and 1–2 room Housing and Development Board flat. To date, we have completed baseline (SEED-1) and 6-year follow-up (SEED-2) examinations. The 12-year follow-up (SEED-3) examination is currently ongoing and will be completed by 2023. SEED-1 was conducted from 2004 to 2011 and SEED-2 was conducted from 2011 to 2017. Malays were the first ethnicity to undergo the baseline examination from 2004 to 2006,18 followed by the Indian ethnic group from 2007 to 2009 and the Chinese ethnic group from 2009 to 2011.19 All eligible participants who attended the baseline examination were invited to the 6-year follow-up study. Of the 10 033 participants from SEED-1, 1450 were ineligible to participate in SEED-2 due to death, severe cognitive impairment, severe mobility impairment, migration, imprisonment or psychiatric illness. Of the remaining 8583 eligible participants for the follow-up examination, 6762 participants attended SEED-2 with a response rate of 78.8% (Figure 1). The 6-year follow-up examination was conducted from 2011 to 2013 for Malays,20 2013 to 2015 for Indians21 and 2015 to 2017 for Chinese.22 As of 2019, we had further completed the 12-year follow-up examination for Malays and are currently examining the 12-year follow-up for Indian participants. The 12-year follow-up examination for Chinese participants will commence in 2021 and is expected to be completed by 2023. SEED-2 recruitment flow chart SEED-2 recruitment flow chart As with any longitudinal study there was loss in follow-up data due to death and uncontactable participants. However, when compared with other population-based eye studies,7,12,23 the overall response rate of SEED was higher. Moreover, we analysed the impact of attrition in Table 3. Compared with individuals who did not return for SEED-2, participants who did return for SEED-2 were more likely to be younger, female and have a higher socio-economic status (all P < 0.001) and less likely to be a current smoker, have diabetes, hypertension, history of cardiovascular disease and chronic kidney disease (CKD) (all P < 0.001, Table 3). From SEED-1, 1450 participants were ineligible for SEED-2, thus we compare eligible participants who attended SEED-2 with participants who were eligible but did not return for SEED-2. We found that eligible participants who did not return for the follow-up examination were more likely to have a lower socio-economic status, be a current smoker and have diabetes, hypertension or CKD (all P ≤ 0.001 Supplementary Table S1, available as Supplementary data at IJE online). However, the absolute differences in the characteristics were quite small, which therefore indicates the absence of obvious selection bias. Baseline characteristics of participants who attended SEED-2 compared with those who did not attend; data presented are mean (standard deviation) or frequency (%), where appropriate P-value was based on chi-square or t-test where appropriate and denotes attending participants vs those who did not attend follow-up exam (i.e. either non-eligible or refused, n = 3271). Education level: O-level, Singapore Cambridge General Certificate of Education, ordinary level; A-level, Singapore Cambridge General Certificate of Education, advanced level. Low SES (socio-economic status): primary or lower education, individual monthly income <S$2000 and 1–2 room Housing and Development Board flat. Baseline characteristics of participants who attended SEED-2 compared with those who did not attend; data presented are mean (standard deviation) or frequency (%), where appropriate P-value was based on chi-square or t-test where appropriate and denotes attending participants vs those who did not attend follow-up exam (i.e. either non-eligible or refused, n = 3271). Education level: O-level, Singapore Cambridge General Certificate of Education, ordinary level; A-level, Singapore Cambridge General Certificate of Education, advanced level. Low SES (socio-economic status): primary or lower education, individual monthly income <S$2000 and 1–2 room Housing and Development Board flat. A standardized examination protocol was used for the three ethnic groups during baseline and follow-up visits. Comparisons of examination components and questionnaires between SEED-1 and SEED-2 are listed in Table 4. Additionally, in-depth disease assessment methods and definitions are listed in Supplementary Tables S2 and S3, available as Supplementary data at IJE online. Comparison of examination components between SEED-1 and SEED-2; ‘+’ denotes done, ‘-’ denotes not done OCT-A, OCT angiography; SD-OCT, spectral domain OCT. Subjective refraction performed on all SEED-1 participants and only on SEED-2 participants when presenting visual acuity was worse than 0.30 LogMAR. Comparison of examination components between SEED-1 and SEED-2; ‘+’ denotes done, ‘-’ denotes not done OCT-A, OCT angiography; SD-OCT, spectral domain OCT. Subjective refraction performed on all SEED-1 participants and only on SEED-2 participants when presenting visual acuity was worse than 0.30 LogMAR. Blood pressure (BP) and pulse rate measurements were collected for all participants with an automatic BP monitor (Dinamap model Pro Series DP110X-RW, 100V2; GE Medical Systems Information Technologies Inc., Milwaukee, USA). Participant’s height was measured using a wall-mounted measuring tape and weight was measured using a digital weighing scale (SECA, model 782 2321009; Vogel & Halke, Hamburg, Germany). Body mass index (BMI) for each participant was then calculated. A non-fasting blood sample (37.5 ml) was drawn for biochemistry tests which included glycated haemoglobin (HbA1c), lipid profile, serum creatinine, full blood count and glucose level. DNA and RNA were extracted for genetic testing from either blood or buccal swab samples. A 20 ml urine sample was also collected to test for creatinine level. Objective refraction, keratometry and ocular biometry were measured using an auto-refractor machine (Canon RK-5 Auto Ref-Keratometer, Canon Inc. Ltd, Japan). Current optical correction (if any) was measured by an auto lensometer. Axial length and anterior chamber depth was measured by a non-contact partial coherence laser interferometer (IOLMaster V3.01, Carl Zeiss; Meditec AG Jena, Germany). Presenting distance visual acuity (VA) with participant’s current optical correction (spectacles or contact lenses, if any) was measured using the logarithm of the minimum angle of resolution (LogMAR) ETDRS numerical chart (Lighthouse International, NY, USA), at a distance of 4 m in standardized lighting conditions. Binocular near VA was measured using the logarithmic Near Visual Acuity Chart ‘2000’ at 40 cm. Subjective refraction and best-corrected VA (BCVA) was conducted by trained and certified study optometrist when presenting VA was worse than LogMAR 0.30. The study ophthalmologist examined participants’ anterior segment eye health by slit-lamp examination (Haag-Streit model BQ-900; Haag-Streit, Koeniz, Swtzerland) and measured intraocular pressure (IOP) using the Goldmann Applanation Tonometer (Haag-Streit, Switzerland) method before dilation. All participants were dilated unless they presented with narrow angles, allergy to mydriatic eye drops, IOP > 21 mmHg or any other contraindication. Pachymetry (Advent, Mentor O&O, Norwell, MA, USA) was used to measure central corneal thickness. Additionally, diagnosed glaucoma participants and glaucoma suspects underwent gonioscopy and visual field testing using the Humphrey Visual Analyzer II’s (model 750, Carl Zeiss Meditec, Switzerland) 24–2 Swedish Interactive Threshold Algorithms-Fast (SITA-FAST) test, prior to pupil dilation. Post-dilation examination included fundus examination, optic nerve head assessment (including cup to disc ratio measurement) and cataract grading was based on either the Wisconsin cataract grading system and/or the Lens Opacities Classification System III (LOCS III). Multiple anterior segment imaging tests were included in SEED. Before pupil dilation, iris photographs were taken on all follow-up participants using a slit-lamp camera (Topcon DC-3 camera, Tokyo, Japan) to document iris surface features such as Fuch’s crypts, contraction furrows and iris colour. A slit-lamp camera (Neitz CT-S, Neitz Instruments Co., Ltd, Tokyo, Japan) was also used to photograph the lens through a dilated pupil for cataract assessment for all baseline participants and during the follow-up examination for the Malay ethnicity. Lastly, anterior segment optical images using Visante optical coherence tomography (OCT; Carl Zeiss Meditec, Jena, Germany) were taken to assess anterior chamber angle depth for almost all participants. Upon dilation, posterior segment imaging included fundus photography (Canon CR-1 Mark-II Nonmydriatic Digital Retinal Camera) on all baseline and follow-up participants. Macula and optic nerve head OCT scans using Cirrus HD-OCT (Carl Zeiss Meditec, Germany) were introduced for the Chinese participant group during baseline examination and were subsequently continued for all follow-up participants. Additionally, the acquisition of macula and optic nerve head OCT angiography scans (RTVue XR 100 Avanti®, Optovue® Inc., Fremont, CA, USA) only commenced during the follow-up examination for the Chinese group. Lastly, Spectralis OCT (Heidelberg Engineering, Germany) was introduced during the follow-up examination for the three ethnicities. During each study visit participants completed an interviewer-administered standardized questionnaire. Participants’ information such as: demographic data, socio-economic status data, lifestyle-related questionnaire, medical history, current medication use, falls history, cognitive function/mental assessment test (for individuals aged ≥60 years), ocular disease history and family medical history. Additionally, the vision function questionnaire (VF-14), patient health questionnaire (PHQ-8 or 9), impact of vision impairment profile (IVI), the modified life space questionnaire, and participant’s knowledge, awareness and attitude towards existing eye care services were also collected. SEED has played a key role in addressing gaps in the epidemiology of eye diseases in Asia by providing novel findings on age-related eye diseases for the three major ethnic groups Indians and in Additionally, SEED has (Supplementary Figure available as Supplementary data at IJE online). The top from SEED are listed in Supplementary Table available as Supplementary data at IJE online. The prevalence of VI and major age-related eye diseases from baseline SEED can be found in Table and Supplementary Table available as Supplementary data at IJE the factors associated with these eye conditions. The overall prevalence of blindness < based on the and VI < to based on the was and Malays the prevalence of VI compared with Indians and Chinese Indians had the prevalence of blindness among the three ethnicities and the prevalence of blindness compared with Chinese prevalence of visual impairment and major age-related eye diseases from baseline SEED to the Singapore Population on the best-corrected visual acuity of the on the visual impairment was as VA to was as VA visual impairment participants with visual impairment in one eye and visual in the other on the visual impairment was as VA to was as VA visual impairment participants with visual impairment in one eye and visual in the other of any existing cataract or any history of cataract in either eye Wisconsin prevalence of visual impairment and major age-related eye diseases from baseline SEED to the Singapore Population on the best-corrected visual acuity of the on the visual impairment was as VA to was as VA visual impairment participants with visual impairment in one eye and visual in the other on the visual impairment was as VA to was as VA visual impairment participants with visual impairment in one eye and visual in the other of any existing cataract or any history of cataract in either eye Wisconsin The overall prevalence of macular degeneration was with Chinese the prevalence of and Indians the prevalence among the three ethnicities (Table The age-related eye disease with the prevalence was cataract Table and it was also the of best-corrected VI and for of cases of VI and of cases of blindness (Supplementary Table available as Supplementary data at IJE is the of was the of blindness (Supplementary Table available as Supplementary data at IJE online). age and female were risk factors across the three ethnicities for (Supplementary Table available as Supplementary data at IJE online). the other risk factors associated with best-corrected VI or blindness across the three ethnicities were age, cognitive impairment and no (Supplementary Table available as Supplementary data at IJE online). Additionally, ratio = P = was only associated with best-corrected VI or blindness among Malays, CKD = P = was only associated with best-corrected VI or blindness among (Supplementary Table available as Supplementary data at IJE online). The overall prevalence of any was with Indians the prevalence followed by Chinese and Malays Additionally, Indians had the prevalence of and among the three ethnicities (Table Furthermore, the three ethnicities had similar associated risk factors for any which included of diabetes, higher higher lower BP and of (Supplementary Table available as Supplementary data at IJE online). The overall prevalence of was with Chinese the prevalence of followed by Indians and Malays (Table Chinese also the prevalence of early and Malays had the prevalence of early Notably, Indians and Malays had the prevalence of of (Table the three age years and were associated with early (Supplementary Table available as Supplementary data at IJE online). The overall prevalence of glaucoma was with Malays the prevalence of followed by Chinese and Indians (Table Additionally, higher IOP was associated with among Indians = P = and Chinese = P < (Supplementary Table available as Supplementary data at IJE online). SEED findings have also to other such as public machine and the data collected have used by and such as the Ministry of Singapore, the the Asian Eye Epidemiology and the of SEED data have also used in global prevalence and disease burden for glaucoma3 and SEED has in providing key data for the of the and has for in SEED has also made important the between VI and as well as cognitive We found that participants with VI had a higher risk of participants ≥60 years, those with VI were more likely to have cognitive Furthermore, SEED data have also used to a for detection of among normative OCT were also from SEED Lastly, SEED genetic data have to novel genetic in age-related eye and the of the world’s Asian genetic The SEED study is the first longitudinal multi-ethnic population-based eye study in We have collected ocular data, images of the and optic We have also collected a data of diseases, genetic and quality of life and were for ocular and across the three ethnic groups Indian and Chinese). The in data provides an opportunity and to compare data across the three ethnicities with similar and socio-economic population-based sample provides to evaluate the impact of and socio-economic risk factors associated with major age-related eye diseases. Additionally, data from Singapore can be on Asian as Singaporean Chinese Singaporean Malays Malays and and Singaporean Indians Furthermore, the SEED examination protocol was the Eye and the Eye which are for examination and grading A of study be study Malay, Chinese and Indians from an and findings not be of these ethnicities in Additionally, we individual monthly income the for baseline, and 12-year follow-up for the of direct comparison with the baseline However, we that income are likely to the years and this a if in income status is evaluated as of Furthermore, the and 12-year incidence rates for major age-related eye diseases be by early detection during baseline or 6-year follow-up examinations. participants with ocular hypertension in the baseline visit have which affect the incidence of glaucoma at the follow-up Lastly, because acquisition of OCT data only commenced at the 6-year follow-up there was no comparison data from the baseline examination. However, OCT data collected from the ongoing 12-year follow-up visit can be used to compare with the 6-year follow-up Although the study is not available in the public any are and on a can be to the The Singapore Epidemiology of Eye Diseases (SEED) cohort is a multi-ethnic Indians and longitudinal population-based study comprising 000 participants aged years at baseline and included 6762 participants at the 6-year follow-up (response rate: 78.8%) examination. The 12-year follow-up examination is currently ongoing and is expected to be completed by 2023. The SEED study the prevalence, incidence, progression and risk factors of major age-related eye diseases and visual impairment (VI) among Asians. from SEED have to important ethnic differences among Asians. Malays had the prevalence of VI and blindness when compared with Indians and Chinese. Additionally, Indians had the prevalence of from SEED have provided novel to genetic in age-related eye diseases and have used as a to a system in To date, the SEED study has for with SEED and a list of can be found at SEED was conducted in with the of and was by the Board was from all participants. Supplementary data are available at IJE online. This was by the Medical and the for and The the ocular epidemiology group and data of the Singapore Epidemiology for and and the study. and analysed and the and the All and the final

Topics & Concepts