Prevalence and Risk Factors of Retinopathy in an Asian Population Without Diabetes: Title and subTitle BreakThe Singapore Malay Eye StudyRetinopathy in Asian Population Without Diabetes FREE

Early retinopathy signs, such as isolated microaneurysms, retinal hemorrhages, hard exudates, and cotton wool spots, have often been observed in people without clinically diagnosed diabetes.^{1 - 18} Previous studies in which signs of retinopathy were detected by retinal photography were conducted mostly in white populations; a few studies have been conducted in Hispanic Americans. These have reported a prevalence of retinopathy ranging from 4.8% to 15.5% in people without diabetes.^{3 - 6 ,10 ,14 - 18} To date, however, there are few studies on the frequency of signs of retinopathy in Asian populations without diabetes, with only 1 report from a Japanese population.¹²

Increasingly, signs of retinopathy are recognized as being markers of systemic vascular disease. Studies show that the presence of retinopathy, even in persons without diabetes, is associated with an increased risk of mortality from coronary heart disease,¹⁹ including stroke,²⁰ ischemic heart disease,²¹ and congestive heart failure.²² In addition, signs of retinopathy may precede the clinical diagnosis of diabetes, especially in high-risk populations (eg, persons with a family history of diabetes).² More recently, studies indicate that signs of retinopathy in individuals without diabetes may pose a challenge to the current diagnostic criteria for diabetes.²³ Despite the increasing interest and potential importance, our current knowledge regarding the epidemiology and risk factors for retinopathy in people without diabetes remains limited. Studies have described associations of various cardiovascular risk factors (eg, hypertension and hyperlipidemia) with retinopathy, but the findings have not been consistent across different populations, and to date few have been conducted in Asian populations.^{3 - 6 ,18 ,24} In this study, we examined the prevalence and risk factors of signs of retinopathy in a population-based sample of Malay adults without diabetes.

The Singapore Malay Eye Study is a population-based, cross-sectional study of 3280 Malay adults aged 40 to 80 years randomly selected from 15 districts across southwestern Singapore from 2004 to 2006. Details of the Singapore Malay Eye Study design, sampling plan, and methodology have been described elsewhere.^{25 - 29} In brief, subjects were selected using age-stratified (by 10-year age group) random sampling from a computer-generated list provided by the Singapore Ministry of Home Affairs. Of 4168 eligible individuals, 3280 participated in the study, giving an overall response rate of 78.7%.

For this article, we excluded individuals with diabetes (n = 757), those without retinal photographs or with photographs of insufficient quality for retinopathy grading (n = 19), and those with more advanced retinopathy that was likely a result of diabetes status misclassification or other retinal vascular disorders (eg, retinal vein occlusion) (n = 4), leaving 2500 participants (76.2%) for analysis. Details of participant characteristics are summarized in Table 1.

The tenets of the Declaration of Helsinki were followed, with ethics approval obtained from the Singapore Eye Research Institute Review Board. Written consent was obtained from all study participants.

Retinal photography was performed using a standardized protocol.^{25 ,28 ,30} Following pupil dilation, 2 retinal photographs centered at the optic disc and macula were taken from both eyes of each participant, using a 45° 6.3-megapixel digital nonmydriatic camera with a 10D single-lens reflex back. These photographs were sent to the University of Sydney and graded for signs of retinopathy, according to the Blue Mountains Eye Study protocol.^{31 - 32} For each eye, retinopathy severity was graded using the Early Treatment Diabetic Retinopathy Study adaptation of the modified Airlie House Classification System.^{30 ,33} We defined retinopathy on the basis of the Early Treatment Diabetic Retinopathy Study severity level in the worse eye^{28 ,34} : minimal nonproliferative retinopathy (level 15-20), mild nonproliferative retinopathy (level 35), moderate nonproliferative retinopathy (level 43-47), severe nonproliferative retinopathy (level 53), and proliferative retinopathy (level >60). Macular edema was defined separately as present or absent. Vision-threatening retinopathy was defined as the presence of severe nonproliferative diabetic retinopathy, proliferative retinopathy, or clinically significant macular edema.

All participants underwent a standardized comprehensive interview, physical examination, and laboratory investigation.^{25 - 28 ,35 - 38} Socioeconomic measures (eg, income and education), lifestyle risk factors (eg, current smoking), medication use, and self-reported history of systemic diseases (eg, angina, heart attack, and stroke) were elicited from the interview.²⁸

Body mass index (BMI) was calculated as body weight in kilograms divided by body height in meters squared. Underweight was defined as a BMI of less than 18.5; normal weight as a BMI of 18.5 to less than 25; overweight as a BMI of 25 to less than 30; and obese as a BMI of 30 or higher.³⁹ Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured using a digital automatic BP monitor 2 times 5 minutes apart, and the mean of the 2 measurements was used for analyses. Hypertension was defined as an SBP of 140 mm Hg or greater, DBP of 90 mm Hg or greater, or use of antihypertensive medication.³⁰

Nonfasting blood samples were drawn and sent for analysis of serum lipids (total cholesterol, high-density lipoprotein cholesterol, and low-density lipoprotein cholesterol), hemoglobin A_1c, creatinine, and glucose at the National University Hospital Reference Laboratory. Urine samples were collected to determine urinary creatinine levels at the Alexandra Hospital Laboratory.

Diabetes mellitus was defined as either a self-reported history of diabetes, current use of diabetic medications, or a nonfasting glucose of 200 mg/dL or higher (to convert to millimoles per liter, multiply by 0.0555) at examination.^{28 ,40} A nonfasting glucose level of less than 200 mg/dL defined normal fasting glucose. Chronic kidney disease was defined as an estimated glomerular filtration rate of less than 60 mL/min/1.73 m². Hyperlipidemia was defined as either total cholesterol of 240 mg/dL or higher (to convert to millimoles per liter, multiply by 0.0259) or the use of lipid-lowering drugs.

Statistical analyses were carried out using SPSS, version 15 (SPSS Inc, Chicago, Illinois). Any retinopathy was analyzed as a binary outcome variable. All potential risk factors were analyzed either as binary (eg, hypertension) or linear (eg, BMI) traits. Body mass index was also analyzed as a categorical variable (normal, underweight, overweight, or obese). Logistic regression models were used to estimate the odds ratio (OR) (and 95% confidence interval [CI]) for any retinopathy associated with various risk factors. Statistically significant risk factors (P < .05) in the age- and sex-adjusted model were selected for inclusion in the multivariate models. Only factors that contributed significantly to the model were included in the final parsimonious models. Goodness of fit of the final logistic regression models was assessed by the Nagelkerke R², which is a pseudo R² measure (range, 0-1), approximating the R² measure used in linear regression models.

Table 1 shows the characteristics of the participants without diabetes by sex. Compared with women, men without diabetes were more likely to be older and much more likely to smoke, to have higher income and education, higher serum levels of hemoglobin A_1c and triglycerides, and higher SBP, and lower total cholesterol, lower HDL cholesterol, and lower BMI.

Figure 1 shows the prevalence of retinopathy in persons without diabetes by age and sex. Among persons without diabetes, there were more men with retinopathy than women across all age groups. The prevalence of retinopathy peaked between the ages of 50 and 69 years and then declined slightly in both men and women. Table 2 presents the prevalence and severity of retinopathy in all subjects without diabetes and in men and women separately. Retinopathy was found in 6.0% of people (n = 149), with all cases being minimal (generally isolated microaneurysms or hemorrhages) to mild (relatively few coexistent microaneurysms and hemorrhages in most cases). There were no cases of vision-threatening retinopathy, macular edema, or clinically significant macular edema. Of participants with neither diabetes nor hypertension (n = 921), 42 (4.7%) had hemorrhages or microaneurysms.

Of the 149 with retinopathy, 142 had evidence of retinal hemorrhages and/or microaneurysms (5.7%), and only 0.3% had other retinopathy signs. One hundred thirty-nine (93.3%) participants had at most 1 retinopathy lesion (hemorrhage/microaneurysm) and 3 others had 2 retinopathy lesions (hemorrhage/microaneurysm and cotton wool spots). Cotton wool spots were relatively uncommon (0.2% in men and 0% in women). Retinopathy was more frequent in men than in women (6.7% vs 4.8%; P = .03).

Table 3 shows the associations of various risk factors with retinopathy in subjects without diabetes. In model 2, independent risk factors for any retinopathy were higher serum levels of glucose within the nondiagnostic range for diabetes (ie, random serum glucose <200 mg/dL; OR, 1.13 per mmol/L increase; 95% CI, 1.00-1.28), higher SBP (OR, 1.15 per 10–mm Hg increase; 95% CI, 1.06-1.25), higher BMI (OR, 1.04 per point increase; 95% CI, 1.00-1.07), obesity (OR, 1.95; 95% CI, 1.22-3.11), and history of heart attack (OR, 2.68; 95% CI, 1.48-4.83).

Table 4 shows the risk factors for retinopathy by sex. Elevated SBP (OR, 1.17 per 10–mm Hg increase; 95% CI, 1.06-1.31) and previous heart attack (OR, 2.45; 95% CI, 1.23-4.88) were factors associated with retinopathy in men, while elevated SBP (OR, 1.12 per 10–mm Hg increase; 95% CI, 1.00-1.26) was associated with retinopathy in women.

Figure 2 shows the relationship between the level of hemoglobin A_1c and prevalence of retinopathy by tertiles of SBP. Among participants without diabetes who had the highest tertile of SBP, an increase in hemoglobin A_1c from the lowest to the highest tertile was associated with a 2-fold increase in retinopathy prevalence from 4.5% to 9.5%.

Retinopathy lesions, as detected from fundus photographs, were common in this Asian Malay population without diabetes. All cases were early lesions, being classified as minimal (5.8%) or mild (0.2%) retinopathy, with nearly all (5.7%) composed of isolated retinal hemorrhages and/or microaneurysms. In persons without diabetes, higher levels of serum glucose, blood pressure, and BMI and a history of heart attack were independent factors associated with retinopathy. Men had a higher prevalence of retinopathy, but there was no apparent difference in the risk factors of retinopathy between men and women.

Figure 3 compares the prevalence of retinopathy in persons without diabetes reported in different adult population-based studies.^{3 - 8 ,10 ,14 - 17} The 6.0% overall prevalence of retinopathy among participants without diabetes in our Asian cohort was comparable with those reported in the Los Angeles Latino Eye Study (6.6%)¹⁰ and the Australian Diabetes Obesity and Lifestyle Study (6.7%),¹⁴ though it was lower than prevalences in previous population-based studies, such as the Hoorn Study (7.3% 9-year incidence),⁶ Funagata Study (7.7%),¹² Beaver Dam Eye Study (7.8%),³ Diabetes Prevention Program (7.9%),¹⁶ Cardiovascular Health Study (8.3%),⁸ Blue Mountains Eye Study [BMES] (9.8%),⁴ and Proyecto VER (15.5%).¹⁵ This disparity of results likely stems, at least in part, from different population characteristics (eg, age and ethnicity), the prevalence of other risk factors (eg, BP levels), the number of retinal photographs taken, and the grading methodology. Regardless, these data suggest that retinopathy signs are frequent in the general adult population.

Further comparison of our results with those from BMES is useful. Retinopathy was graded by the same staff using the same grading protocol as in BMES,⁴ but there was a lower prevalence of retinopathy in this study. This is likely related to differences in methodology (the use of only 2 nonstereoscopic digital images per eye in our study vs 6- to 7-field stereoscopic film–based photographs in BMES). Our findings also differed from BMES in that the retinopathy prevalence peaked in individuals aged from 50 to 69 years, in contrast with the age-related increase found in the BMES population.⁴ Retinopathy signs, however, were consistently more prevalent among men than women in both our study and BMES. We investigated if there were different patterns of association of retinopathy signs with risk factors (Table 4). After sex stratification, the pattern of associations was largely similar, though only elevated SBP and history of heart attack were significantly associated with retinopathy in men, and elevated SBP was significantly associated with retinopathy in women. Further research is needed to understand possible sex differences in the development of early signs of retinopathy.

Our study demonstrated associations between retinopathy and several cardiovascular risk factors (eg, hyperglycemia, hypertension, obesity, and history of previous heart attack) in persons without diabetes. As in BMES, our study confirmed a strong and consistent association between elevated blood pressure and retinopathy among participants without diabetes.⁴ These observations are also consistent with the Cardiovascular Health Study, which demonstrated that retinopathy in persons without diabetes was associated with hypertension and coronary heart disease.¹⁸ Furthermore, BMES found that retinopathy in persons without diabetes signaled an increased risk of cardiovascular mortality that was equivalent to the increased risk associated with only the presence of diabetes (without diabetic retinopathy).¹⁹ Thus, our findings and others suggest, given the documented associations of retinopathy signs with cardiovascular disease, that isolated retinopathy signs may represent subclinical manifestation of early microvascular damage from glucose and blood pressure dysregulation.^{3 - 5 ,18 ,41 - 42} These signs could indicate intermediate pathology in the pathway to cardiovascular disease and may signal a more adverse cardiovascular profile in people without diabetes. Retinopathy signs are therefore likely markers of generalized microcirculatory dysfunction, not only in the eye but also in vital organs elsewhere.^{19 - 22}

While our study and others identified major vascular risk factors as being related to signs of retinopathy in people without diabetes, the specific underlying pathogenesis remains unclear. In persons without diabetes who also had normal BP levels, retinopathy prevalence was still around 5%, suggesting that BP (at least as measured at 1 time) only explains a small proportion of cases with retinopathy. We did not capture the effects of long-standing hypertension or elevated BP. Markers of chronic low-grade inflammation and endothelial dysfunction, including C-reactive protein, have been associated with macrovascular disease in persons without diabetes.^{43 - 45} The Hoorn Study showed that C-reactive protein was associated with retinopathy in individuals without diabetes.⁴⁶ However, in our study, we did not find an association of serum C-reactive protein with retinopathy signs (data not shown). Thus, the pathophysiology of these early retinopathy lesions deserve further investigation.⁴⁷

Strengths of the present study include its large population-based sample, high response rate, use of standardized photographic grading protocols to define retinopathy lesions in a masked fashion, a high proportion of gradable photographs, and adjustment for potential confounders. Limitations should also be noted. First, potential misclassification of diabetes status cannot be completely excluded, as its definition was based on nonfasting glucose levels rather than on fasting glucose (eg, as in BMES).⁴ Second, we used the Airlie House Classification System to define retinopathy severity. It remains unclear whether this classification system, originally designed for grading of diabetic retinopathy,³³ is suitable to define retinopathy in people without diabetes. Third, because the associations observed in our study are cross-sectional, they need replication by longitudinal data. Finally, the possibility that some detected retinopathy cases were due to other rare causes (eg, past exposure to radiation, blood dyscrasias, or systemic infection) cannot be excluded.

In conclusion, as in white populations, retinopathy signs are frequent in Asian Malays without clinical diabetes and are associated with several modifiable cardiovascular and metabolic risk factors.^{4 - 7} While it is now well established that signs of retinopathy in people without diabetes represent an adverse vascular risk profile and are markers of a range of major cardiovascular disorders, including stroke and cardiovascular mortality, further research is needed to clarify their pathophysiology, classification, and natural evolution.

Correspondence: Tien Yin Wong, MBBS, PhD, Singapore Eye Research Institute, Singapore National Eye Centre, 11 Third Hospital Ave, Singapore 168751 (ophwty@nus.edu.sg).

Submitted for Publication: December 23, 2008; final revision received March 31, 2009; accepted April 4, 2009.

Financial Disclosure: None reported.

Funding/Support: This study was supported by grants 0796/2003, 0863/2004, and CSI/0002/2005 from the National Medical Research Council and 501/1/25-5 from the Biomedical Research Council. Additional support was provided by the Singapore Tissue Network and the Ministry of Health, Singapore.

Additional Contributions: We thank the staff and participants in the Singapore Malay Eye Study for their important contributions.