Diabetic retinopathy (DR) is a typical eye complication caused by diabetes mellitus (DM) and a leading cause of irreversible blindness in adults in developed countries [1]. In 2020, 1.4% of moderate or greater vision loss and 2.5% of blindness among adults aged 50 years and older worldwide were due to DR [2]. Vision loss in DR is related to diabetic macular edema affecting central vision, retinal detachment through the formation of new blood vessels and fibrous tissue around the retina, and vitreous hemorrhage [3]. In a systematic review of 59 studies from 27 countries, the global prevalence of DR was predicted to increase from 103 million in 2020 to 129.8 million and 160.5 million in 2030 and 2045, respectively [4]. In addition, the number of vision-threatening DR cases was predicted to increase continuously from 28.5 million in 2020 to 36.05 million in 2030 and 44.82 million in 2045 [4]. In a systematic review of the prevalence of DR among Asians, most studies were conducted in China, Singapore, and India, and studies conducted in Korea were limited [5]. In a recent study analyzing data from the Korean National Health Insurance Service-National Sample Cohort (KNHIS-NSC), the prevalence of DR in 2015 was 15.9% and that of proliferative DR (PDR) was 1.16% [6], which is relatively low compared to other Asian countries. However, since this study analyzed health insurance data, only those diagnosed with DM who visited medical institutions and underwent retinal examinations were targeted. Accordingly, participants who were not diagnosed with DM or did not undergo retinal examination were excluded. According to the 2020 Diabetes Fact Sheet, 35% of patients with diabetes aged >30 years are unaware of their disease [7]. Considering that the retinal examination rate of Korean patients with diabetes is only 26% [6] to 30% [8], the actual prevalence of DR is estimated to be higher than that reported. Thus, a community-based study of the prevalence of DR, including patients with DM diagnosed by a physician as well as patients with hyperglycemia, is needed.

Risk factors of DR have been reported through several studies [9, 10, 11, 12, 13, 14, 15, 16, 17], but there are some limitations. For example, age has been related to DR in many studies; however, other characteristics have reported unrelated or inconsistent results. Although household income is a major social determinant of health [18] and can affect access to medical care, studies on the relationship between income and DR have been very limited. In addition, limited studies have been conducted to analyze the relationship between health behavior such as drinking or smoking and DR [10, 13, 14, 17]. Obesity has shown inconsistent results depending on the index such as waist circumference, waist-hip ratio, or body mass index (BMI) [16]. Therefore, the purpose of this study was to identify the DR prevalence, and to determine risk factors for DR in two groups of diabetes people using the Korean National Health and Nutrition Examination Survey VII (KNHANES VII), a community-based cohort established in Korea.

This study was conducted to identify the prevalence, to compare the characteristics of people with diabetes with and without DR, and to identify risk factors of DR in people with diabetes aged >40 years using recent published nationwide data in Korea. We analyzed the DM prevalence by dividing it into one group diagnosed by a doctor and the other group that included not only those diagnosed with diabetes, but also treated for diabetes or had high blood sugar. The prevalence of DR was 25.87% in Group 1 and 20.14% in Group 2, and insulin therapy, DM duration, and SBP were found to be factors related to DR. In this study, DR was found in about 26 participants based on 100 diagnosed diabetes people and found in about 20 participants based on 100 diabetes people including potential diabetes people in Korea. Based on previous studies, the global DR prevalence rates were 22.3% [4] and 34.6% [15], and the prevalence of DR in people with diabetes was 28.2% [14] in Singapore, and 40.0% [17] and 35.5% [13] in China. Regional or racial differences are the main factors contributing to the difference in the prevalence of DR [4], and the prevalence of DR is high in people living in North America and the Caribbean, the Middle East, and North Africa. The prevalence of DR is higher in Hispanics and Middle Easterners than in Asians [4]. Among Asian countries, the prevalence of DR observed in this study was relatively lower than in other Asian countries [13, 14, 17]. These differences between studies may be due to the sample size, type of study population, age, DM duration, and various retinal examination methods. In a previous retrospective study conducted in Korea using the KNHIS-NSC database, the prevalence of DR in 2015 was 15.9% [6], and the prevalence of DR in 2011-2012 was 20% [27] in a study using the 5th KNHANES data. This difference seems to be related to differences in the characteristics of the participants, such as the timing of the survey and the source of the data.

Early diagnosis can detect early signs of DR, which are usually asymptomatic [3], and can reduce severe vision loss by more than 90% [28]. For this, it is necessary to establish a screening referral system that can readily undergo retinal examinations at regional healthcare institutions and be evaluated by an ophthalmologist [3]. Various portable fundus cameras have been developed and are reported to have excellent sensitivity and specificity; therefore, they can be used for DR screening and early diagnosis of DR in the local community [29]. Therefore, it is possible to decrease the risk of vision loss by detecting DR at an earlier stage by establishing a referral system as part of the diabetes patient management program in primary health care institutions, including public health centers and public health units. The referral system involves photographing the retina with a portable non-mydriatic fundus camera and referring it to a local ophthalmologist for evaluation.

In this study, insulin-treated diabetes people had a higher risk of DR, which was consistent with findings from previous studies [11, 12]. These results can be explained in the following three ways: first, the rapid decrease in plasma glucose concentration due to insulin therapy lowers intravascular osmotic pressure, leading to water retention in the ocular vessels [30]. Second, high-dose exogenous insulin can exacerbate vascular proliferation and DR by acting on vascular endothelial growth factors generated in the ischemic retina [30]. Third, in addition to the possibility of DR induced by insulin therapy, as insulin therapy is mainly prescribed for patients with hyperglycemia that is not well controlled with oral drugs, it may be related to hyperglycemia rather than the effect of insulin therapy itself [10, 11, 12, 13, 14, 17].

DM duration was a representative risk factor for DR [10, 11, 12, 13, 14, 17], and it was identified as a major risk factor in this study. The longer DM duration, the higher the risk of DR. The prevalence of DR was about twice higher in diabetes people with a duration of 5-9 years and about 2~3 times higher in those with duration of more than 10 years compared to those with duration of less than 5 years. According to a systematic review on the occurrence of DR [31], the annual incidence of DR differs depending on the DM duration. In India [32], the mean DM duration was 5.3 years, and the annual incidence of DR was 2.4% after 4 years of follow-up. In the United States [33], when 47% of participants had a DM duration of more than 10 years, the annual incidence of DR was 10.4% during the 4-year follow-up period. In China [34], the annual incidence of DR was 2.2% when participants with an average DM duration of 5.7 years were followed for 10 years. On the other hand, in another Chinese study [35], the annual incidence of DR was 12.7% when participants with a DM duration of 11 years were followed for 5 years. Overall, it was found that the annual incidence of DR was approximately five times or more different between people with <6 years and >10 years of DM duration. Frequent eye examinations are required for people with diabetes for more than five years, and education on blood sugar management is required.

High blood pressure was identified a major risk factor for DR, which was consistent with the results of previous studies [10, 14, 15]. The results of this study were supported by the results of a large-scale experiment [36] that strictly controlled blood pressure had a significant effect on DR reduction.

This study has several strengths. First, this study used the most recent nationwide community-based data to identify the prevalence and risk factors of DR in diabetes people living in Korea. Therefore, the results of this study can be representative of Korea. Second, the prevalence and risk factors of DR were analyzed by defining the DM group diagnosed by the doctor and the DM group including all potential diabetes people, so it was possible to compare whether the results were overestimated. Third, retinal examination and survey data used as primary data sources in this study were measured or investigated in a standardized manner by trained investigators. Fourth, various factors related to DR can be identified by analyzing socioeconomic variables and health-related factors related to DR based on previous studies.

Despite these strengths, this study has limitations, and caution is required when interpreting the results. First, DR prevalence may be underestimated because participants could not stare at a point on the screen in the camera or had vision loss with visual loss from the examination. Second, since KNHANES VII data did not include the severity of DR, we could not compare the characteristics according to the severity of DR with DR. Therefore, further studies including DR severity are needed. Third, for Group 2, DM was defined based on various criteria to minimize exclusion due to undiagnosed or unknown diabetes. However, subjects with temporarily high blood sugar in a blood glucose test may also be selected as DM, and thus there may be a risk of selection bias. Fourth, DM duration was defined from time of DM diagnosis to time of investigation. For those who diagnosed with DM due to hyperglycemia in Group 2, DM duration could be underestimated because the time of investigation was the time of diagnosis of DM. Fifth, since KNHANES VII data is a cross-sectional study, it is difficult to determine whether the risk factors occurred before the DR. Future longitudinal studies are recommended to identify the causal relationship between risk factors and DR incidence. Finally, there were significant differences in some characteristics including SBP between subjects included and excluded in the final analysis. Therefore, the results of this study are applied to the subjects included in this analysis, and caution is needed in generalization to other subjects.

This study is significant in that it provided evidence that would be helpful in managing risk factors for DR to prevent DR and delay its onset in diabetes people aged 40 years or older. Specifically, to prevent, detect, and treat DR early, it is necessary to promote education programs and regular eye examinations to recognize and manage the high likelihood of DR in people receiving insulin therapy, DM duration of more than 5 years, or high SBP of 140 mmHg or higher. In addition, it is possible to decrease the risk of vision loss by detecting DR at an earlier stage by establishing a referral system as part of the diabetes patient management program in primary health care institutions, including public health centers and public health units. In addition, the results of this study can be used as evidence to establish a referral system as part of a diabetes patient management program in primary medical institutions such as public health centers or units to detect DR early and reduce the risk of vision loss.

The DR prevalence among people with diabetes aged ≥ 40 years in Korea was approximately 20.14~25.87%, and showed higher in people with insulin therapy, DM duration ≥5 years, and high SBP. Therefore, eye examinations education is highly recommended as part of a diabetes management programs in the community. In particular, we propose to shorten the eye examination cycle for early detection of DR for the high-risk group identified in this study and establish a referral system according to the eye examination results that can be linked to treatment.