Prevalence and Associations of Retinal Emboli With Ethnicity, Stroke, and Renal Disease in a Multiethnic Asian Population: The Singapore Epidemiology of Eye Disease Study

Key Points

Question  What are the prevalence and systemic associations of retinal emboli in Asian individuals?

Findings  In this population-based study of 9978 individuals in Singapore with gradable retinal photographs, age-standardized prevalence of retinal emboli was 0.75%, and the prevalence rate was highest in Indian persons, followed by Chinese and Malay persons. Besides its associations with traditional cardiovascular risk factors and stroke, prevalent retinal emboli were also independently associated with chronic kidney disease and its measures.

Meaning  Although based on 88 individuals among 9978 participants, these findings suggest the prevalence of persons with retinal emboli is generally low in Asian populations, but such persons may require both general cardiovascular as well as renal assessment.

Importance  To our knowledge, population-based data on retinal emboli are limited in Asia. Besides its associations with traditional cardiovascular risk factors and stroke, associations between retinal emboli and renal disease and function remain unclear.

Objective  To examine the prevalence of and risk factors for retinal emboli in a large, contemporary, multiethnic Asian population.

Design, Setting, and Participants  This population-based cross-sectional study was conducted from 2004 to 2011 and included a total of 10 033 Chinese, Malay, and Indian persons aged 40 to 80 years residing in the general communities of Singapore. Analyses were performed from November 2016 to February 2017.

Interventions or Exposures  Retinal emboli were ascertained from retinal photographs obtained from both eyes of all participants according to a standardized protocol. Age-standardized prevalence of retinal emboli was calculated using the 2010 Singapore adult population. Risk factors were assessed from comprehensive systemic and ophthalmic examinations, interviews, and laboratory investigations.

Main Outcomes and Measures  Retinal emboli.

Results  Of the 10 033 participants, 9978 (99.5%) had gradable retinal photographs. Of these, 5057 (50.7%) were female, and 3375 (33.8%) were Indian. We identified 88 individuals (0.9%) with retinal emboli; the overall person-specific, age-standardized prevalence of retinal emboli was 0.75% (95% CI, 0.60-0.95), with the highest prevalence seen in the Indian cohort (0.98%), followed by the Chinese (0.73%) and Malay (0.44%) cohorts (P = .03). In multivariable-adjusted analysis, factors associated with prevalent retinal emboli included older age (per 5-year increase; odds ratio [OR], 1.22; 95% CI, 1.05-1.41), Indian ethnicity (compared with Malay ethnicity; OR, 3.58; 95% CI, 1.95-6.60), hypertension (OR, 1.95; 95% CI, 1.03-3.70), chronic kidney disease (OR, 2.05; 95% CI, 1.15-3.64), creatinine level (per SD increase; OR, 1.13; 95% CI, 1.05-1.21), glomerular filtration rate (per SD increase; OR, 0.67; 95% CI, 0.51-0.86), and history of stroke (OR, 3.45; 95% CI, 1.70-6.99).

Conclusions and Relevance  Based on 88 individuals among 9978 participants of 3 major Asian ethnic populations, retinal emboli were most commonly seen in Indian persons and associated with conventional cardiovascular risk factors, stroke, and chronic kidney disease. Therefore, its presence may signal vascular embolic event and damage not only in the brain but also in the kidneys. If these data are confirmed in longitudinal studies, they would suggest that persons with retinal emboli may require both general cardiovascular and renal assessment.

Retinal emboli are discrete plaquelike opacities observed in the lumen of retinal arterioles. While it can be an incidental finding in otherwise asymptomatic persons, the presence of retinal emboli is associated with a higher risk of stroke and stroke mortality, independent of traditional cerebrovascular risk factors, such as smoking, hypertension, diabetes, and dyslipidemia.^1,2

Our current understanding of the epidemiology of retinal emboli remains limited, with only a few studies that have provided population-based data.^3-6 These studies reported prevalence rates of 1.3% to 1.4% in older white populations.^5-7 Notably lower prevalence estimates (0.4%-0.6%) were reported in nonwhite populations.^3,4 Direct comparison of these findings is difficult owing to various methodological differences between these studies.

Furthermore, given the rarity of the condition, previous studies had limited power to assess some of the risk factors or associations. In particular, although the associations of retinal emboli with conventional cardiovascular risk factors have been described, to our knowledge, the potential link with renal function and disease has not been evaluated. This is relevant in the context of atheroembolic renal disease, which is increasingly recognized as an important clinical syndrome.⁸ Atheroembolic renal disease is a form of renal dysfunction secondary to embolic occlusion of the renal circulation that can occur spontaneously or after intravascular trauma or anticoagulation. Retinal emboli are thought to be one of the key clinical manifestations of atheroembolic renal disease,⁸ but to our knowledge, there are no population-based data on their associations with renal disease or its measures.

Moreover, about 4.5 billion people, more than half of the world population, live in Asia.⁹ There is a lack of epidemiological data on retinal emboli in Asian populations. In this report, we examined the prevalence and associations of retinal emboli in the Singapore Epidemiology Eye Disease (SEED) study. The large combined sample size of the SEED study and the identical methods used in the study cohorts allows for more direct and valid intraethnic data comparison and offers greater power to assess conventional and novel risk factors, including the possible association of retinal emboli with renal disease.

The SEED study is the largest epidemiological eye study in Asia, including 3 major Asian ethnic populations (Chinese, Malay, and Indian individuals) living in Singapore.^10,11 Its baseline examinations were conducted from 2004 through 2011: the Singapore Malay Eye Study (2004-2006), the Singapore Indian Eye Study (2007-2009), and the Singapore Chinese Eye Study (2009-2011). All of these studies were conducted at the Singapore Eye Research Institute with approval from the SingHealth Institutional Review Board and in accordance with the Declaration of Helsinki. All participants provided written informed consent.

Detailed methods of these studies have been described previously and were identical to allow direct comparison between the studies.^10,11 In brief, we selected persons aged 40 to 80 years residing in southwest Singapore during each stipulated study period from a computer-generated list provided by the Singapore Ministry of Home Affairs using an age-stratified (by 10-year age groups) random sampling method. The numbers of eligible persons were 4168 Malay persons, 4497 Indian persons, and 4606 Chinese persons. Of these, 3280 Malay persons (78.7%), 3400 Indian persons (75.6%), and 3353 Chinese persons (72.8%) participated in the study. Participants who had moved from their residential address, had not lived there in the previous 6 months, or were deceased or terminally ill were ineligible. In each ethnic cohort, nonparticipants were older than participants (P < .001), but there was no sex difference. Of the 10 033 participants, 55 were excluded owing to ungradable retinal photographs, leaving 9978 participants (99.5%) for analysis. Baseline characteristics of the SEED study populations have been described previously.^10,12

We used a digital fundus camera (Canon CR-Dgi with 10D single-lens reflex camera digital camera back; Canon) to capture color photographs of Early Treatment Diabetic Retinopathy Study standard field 1 (centered on the optic disc) and Early Treatment Diabetic Retinopathy Study standard field 2 (centered on the fovea) of each eye after pupil dilation. Photographs were sent to the University of Sydney in Sydney, Australia, and were graded for retinal emboli based on the Blue Mountains Eye Study (BMES) protocol.^3,6 The photographs were read by one of 2 graders who were masked to participants’ information to detect the presence of emboli. Each photograph with retinal emboli was classified as cholesterol, fibrin-platelet, or calcific type for the embolus, and 2 retinal specialists (N.C. and K.T.) adjudicated all the cases. Examples of the cases for retinal emboli are illustrated in the Figure (eFigure in the Supplement).

All participants underwent a standardized questionnaire and examination, including collection of blood samples. Demographic data, socioeconomic measures, and self-reported history of systemic diseases, including stroke, acute myocardial infarction, and angina, were elicited from the interview. Surgery for heart attack was defined as self-reported history of having had coronary artery bypass surgery, coronary angioplasty, or cardiac valve replacement surgery. Overweight was defined as a body mass index (calculated as weight in kilograms divided by height in meters squared) of 25 or greater. Hypertension was defined as a systolic blood pressure of 140 mm Hg or greater, a diastolic blood pressure of 90 mm Hg or greater, or use of antihypertensive medications according to the protocol used in the Multi-Ethnic Study of Atherosclerosis.¹³ Diabetes was self-reported or defined as a random glucose level of 200 mg/dL (to convert to millimoles per liter, multiply by 0.0555) or greater or use of oral hypoglycemic medication or insulin. Hypercholesterolemia was defined as a total cholesterol level of 239.38 mg/dL (to convert to millimoles per liter, multiply by 0.0259) or greater, and abnormally high low-density lipoprotein cholesterol was defined as 158.3 mg/dL or greater. An estimated glomerular filtration rate (GFR) was used to assess renal function. Chronic kidney disease was defined as an estimated GFR less than 60 mL/min/1.73 m² from serum creatinine.¹⁴

Retinal embolus was analyzed as a binary outcome variable. All potential risk factors were analyzed either as binary traits or as per unit or SD change for continuous variables. Prevalence estimates and 95% CIs of retinal emboli were calculated and standardized to the population distribution from the 2000 Singapore Census.¹⁵ The distribution of each predictor variable was compared between individuals with and without retinal emboli using χ² and t tests for categorical and continuous variables, respectively. Logistic regression models were used to estimate the odds ratios and 95% CIs for retinal emboli for each risk factor, initially adjusting for age and sex (model 1) and additionally adjusted for ethnicity, smoking status, systolic blood pressure, body mass index, total cholesterol level, glycated hemoglobin A_1c, and GFR in multivariate analysis (model 2). All P values were 2-sided, and statistical significance was set at P < .05. All analyses were performed using Stata version 13 (StataCorp).

Of the 9978 participants, 5066 (50.8%) were female. With 88 individuals with retinal emboli identified, the overall person-specific and age-standardized prevalence of retinal emboli was 0.75% (95% CI, 0.60-0.95), with prevalence rates of 0.73% (95% CI, 0.47-1.11) among Chinese persons, 0.44% (95% CI, 0.24-0.76) among Malay persons, and 0.98% (95% CI, 0.68-1.39) among Indian persons (P = .02). Of the 88 participants with retinal emboli, 2 participants (2%) had emboli found in both eyes. Three participants (3%) had multiple emboli in 1 eye. One participant (1%) with retinal emboli was also found to have retinal artery occlusion. Of the participants with retinal emboli, 52 (59%) had the cholesterol type, 17 (19%) had the platelet-fibrin type, and 8 (9%) had the calcific type. Eleven participants (13%) had retinal emboli ungradable for a definitive type. There was no ethnic difference in the type of retinal emboli in our study (P = .29).

Table 1 summarizes the characteristics of the participants with and without retinal emboli. Compared with participants without retinal emboli, those with retinal emboli were more likely to be older, of Indian ethnicity, and have hypertension, higher systolic blood pressure, diabetes, chronic kidney disease, cardiovascular disease, angina, or stroke.

Table 2 shows the systemic and ocular associations of retinal emboli. In multivariable-adjusted analysis, factors associated with retinal emboli were older age, Indian ethnicity, current smoking, hypertension, chronic kidney disease and its related measures (lower GFR and higher creatinine levels), and history of previous stroke. Further analysis shows the associations of these factors with retinal emboli in each of the ethnic cohorts. The pattern of the observed associations was similar across the ethnic groups for some but not all the factors (eTable in the Supplement). Of note, the association of retinal emboli with chronic kidney disease was positive among the Chinese and Indian cohorts only, while smoking was most strongly associated with retinal emboli in the Malay cohort.

In an additional supplementary multivariable-adjusted analysis (data not shown), we controlled for hypertension status rather than systolic blood pressure, which could be affected by use of antihypertensive medications. The associations of retinal emboli with the renal outcomes remained similar. In analysis stratified by hypertension or diabetes status, the associations of retinal emboli with renal function and chronic kidney disease were also similar. In particular, the association between retinal emboli and lower GFR remained in those with or without hypertension or diabetes.

To our knowledge, the SEED study represents the largest population-based study on the prevalence and risk factors of retinal emboli in a multiethnic population in Asia. Our data suggest that the person-specific and age-standardized prevalence of retinal emboli was 0.75% among middle-aged to older persons in Singapore. The prevalence varied by ethnicity and was higher in our Indian cohort (0.98%) compared with our Chinese (0.6%) or Malay (0.43%) cohorts. After adjusting for potential confounders, factors found to be associated with retinal emboli were older age, Indian ethnicity, smoking, hypertension, chronic kidney disease and its related measures (lower GFR and higher creatinine levels), and history of stroke.

As previously reported from our Malay cohort,³ our Chinese cohort similarly had a lower prevalence of retinal emboli (0.6%) than that of the previous studies in white populations. The Beaver Dam Eye Study reported a prevalence of 1.3% for retinal emboli in the United States, while the BMES reported a prevalence of 1.4% in Australia.^5,6 Retinal emboli in our study were ascertained from 2 digital images per eye, a simpler method than that used in the Beaver Dam Eye Study and the BMES. However, our prevalence rate in Chinese persons was similar to that of the Los Angeles Latino Eye Study (0.4%),⁴ which assessed retinal emboli from stereoscopic photographs of more retinal photographic fields. This may suggest that the lower prevalence seen in our Chinese cohort may not be entirely because of methodological differences in retinal photography.

Our data indicate that the prevalence of retinal emboli was highest among Indian participants. This was not likely owing to differences in the distribution of risk factors across the ethnic populations in our study. Age, hypertension, and cigarette smoking are the most consistent risk factors previously reported for retinal emboli. The prevalence of retinal emboli was consistently highest among the Indian cohort across all age groups (data not shown). Second, the proportion of older participants (70 years or older) was lowest in the Indian cohort.¹⁰ Third, rather than the Indian cohort, the Malay cohort had the highest prevalence of hypertension, chronic kidney disease, current smokers, and higher creatinine level despite having the lowest prevalence of retinal emboli.¹⁰ Furthermore, compared with Malay ethnicity, Indian ethnicity was associated with more than 3-fold increased odds of retinal emboli, even after adjusting for age, smoking, and levels of blood pressure, cholesterol, and renal function. All these collectively suggest that the potential confounding effects of these risk factors may not account for the association of retinal emboli with Indian ethnicity among Asian individuals. Additional studies are needed to determine the reason for this finding.

Several cardiovascular risk factors have been associated with retinal emboli in previous population-based studies.^4,5,7 Consistent with these studies, we found associations of retinal emboli with older age, smoking, and hypertension. Although the cholesterol type of retinal emboli was the most prevalent type found in our study, our data showed no independent association between retinal emboli and hypercholesterolemia or levels of individual cholesterol subtypes. Similarly, the BMES did not find any association between prevalent retinal emboli and levels of total cholesterol, triglycerides, and high-density lipoprotein cholesterol.⁷

Previous studies have shown that people with retinal emboli were more likely to have carotid artery disease and had higher risk of stroke.^{1,2,5,7,16,17} Our data further reinforce the association between retinal emboli and stroke. By virtue of its anatomic orientation, the central retinal artery represents an end artery of the cerebral circulation. Emboli seen in the retina may therefore reflect thromboembolic events that occurred in the brain or elsewhere in the body (eg, carotid artery). About 10% of the patients with retinal emboli may have significant ipsilateral carotid artery stenosis, and of these, about half of them required carotid endarterectomy.¹⁸ Therefore, although the temporality of the relationship could not be determined in our cross-sectional study, it is generally recommended that people with asymptomatic retinal emboli should have a thorough cardiovascular workup.¹

Another important new finding, to our knowledge not examined in previous population-based studies, is the association of retinal emboli with renal function and chronic kidney disease. Our results show that higher creatinine levels and lower GFR, 2 indicators of renal dysfunction, were both consistently associated with retinal emboli, independent of age, smoking, blood pressure (or hypertension status), and other risk factors. In stratified analysis, the association between retinal emboli and lower GFR (ie, renal function) remained similar, even in participants without hypertension or diabetes. In addition, participants with retinal emboli had double the odds of having chronic kidney disease compared with those without retinal emboli. None of the 3 previous population-based studies of retinal emboli examined these relationships.^4,5,7 Our findings suggest that retinal emboli may represent a sign of vascular embolic event and damage, not only in the brain (stroke) but also in the kidneys. In support of this theory, retinal emboli have been observed in 6% to 25% of patients with atheroembolic renal disease, a clinical syndrome gaining increasing recognition for its importance.^8,19 It is often difficult to diagnose unless kidney biopsy is performed, and it is precipitated by acute or subacute renal failure and is associated with poor prognosis.^8,20 In view of our findings, patients with asymptomatic retinal emboli may benefit from not only general cardiovascular workup but also assessment of renal function to detect this potentially fatal complication.²¹ Future longitudinal studies are needed to determine whether retinal emboli could predict risk of renal disease and its prognosis.

Strengths of our study include its large multiethnic population-based sample, masked grading of retinal photographs according to a standardized protocol, low number of ungradable photographs, and standardized and comprehensive assessment of risk factors. Limitations should also be noted. First, the cross-sectional design of our study prevented inferring causality. Second, any prevalence estimate of retinal emboli, in our study or others, could be an underestimate owing to the transient nature of emboli. Third, retinal emboli in our study were ascertained from 2 digital retinal images per eye, a simpler method than that used in previous population-based studies. This could be another source of underestimation, but it should not affect the intraethnic comparison of prevalence among the different Asian populations examined in our study. Fourth, given the large number of risk factors examined, the possibility of chance findings should be discussed. In particular, the results of the supplementary analysis stratified by ethnic groups should be interpreted with caution (eTable in the Supplement). It is well known that chance findings are more likely in subgroup analysis. While false-positive findings owing to chance are possible, the chance of false-negative results may also increase owing to smaller sample size in each group. Of note, the odds ratio for retinal emboli in association with chronic kidney disease was positive among Chinese and Indian individuals but not among Malay individuals. Furthermore, the odds ratio for retinal emboli in association with current smoking was positive among Malay and Indian persons but not among Chinese persons. The possibility of false-negative findings for these inconsistencies cannot be excluded. Lastly, our grading method was based on embolus appearance and did not account for the anatomical location of the embolus. However, there is no strong evidence that anatomical location of the embolus along the retinal vasculature is a reliable determinant for the type or source of embolus.

In summary, our study represents the largest cross-sectional study of retinal emboli in 3 major contemporary Asian populations, to our knowledge. Based on 88 individuals among 9978 participants, we found a relatively low age-standardized prevalence of asymptomatic retinal emboli in Malay and Chinese individuals but a higher prevalence in Indian individuals. We confirmed the associations of retinal emboli with some of the common cardiovascular risk factors, including older age, smoking, and hypertension. Besides stroke, we also found associations of retinal emboli with chronic kidney disease and its measures. These findings, if confirmed in longitudinal studies, suggest that patients with retinal emboli may benefit from general cardiovascular workup with a focus on cerebrovascular risk factors as well as renal assessment for atheroembolic renal disease.

Corresponding Author: Ning Cheung, MD, Singapore Eye Research Institute, Singapore National Eye Centre, 11 Third Hospital Ave, Singapore 168751 (cheung.ning@snec.com.sg).

Accepted for Publication: July 4, 2017.

Published Online: August 24, 2017. doi:10.1001/jamaophthalmol.2017.2972

Author Contributions: Drs Cheung and Wong had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Cheung, Teo, Cheng, Wong.

Acquisition, analysis, or interpretation of data: Teo, Zhao, Wang, Neelam, Tan, Mitchell.

Drafting of the manuscript: Teo, Tan.

Critical revision of the manuscript for important intellectual content: Cheung, Zhao, Wang, Neelam, Mitchell, Cheng, Wong.

Statistical analysis: Teo, Zhao, Wong.

Obtained funding: Mitchell.

Administrative, technical, or material support: Tan, Mitchell, Cheng.

Supervision: Wang, Wong.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Funding/Support: This study was supported by grants 0796/2003, 0863/2004, and CSI/0002/2005 from the National Medical Research Council and grant 501/1/25-5 from the Biomedical Research Council.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.