Diabetes mellitus (DM) is a major health problem worldwide, and is one of the most common and rapidly increasing chronic diseases. The prevalence of diabetes for all age groups worldwide was estimated to be 2.8% in 2000 and is predicted to reach 4.4% in 2030, meaning that the total number of people with diabetes is projected to rise from 171 million in 2000 to 366 million in 2030 (1). This increasing prevalence is due to an aging population, changes in lifestyle with economic development, and increasing rates of obesity associated with decreased physical activity (2). The prevalence of diabetes has also increased rapidly in Korea. Estimated to be 7.6% in 2001, the prevalence of diabetes rose to 9.1% of the adult population according to an analysis of the 2005 Korea National Health and Nutrition Examination Survey (KNHANES), with a prevalence of impaired fasting glucose of 17.4% (3). The economic burden continues to rise and, according to the KNHANES, the aggregate annual direct cost of diabetes in Korea was 260 million USD in 2009 (4, 5).

Diabetes is always associated with long-term complications. Diabetic retinopathy (DR) is the leading cause of blindness among working-aged adults worldwide. The prevalence of any DR and vision-threatening DR were variably reported as 15.8%-46.9% and 4.6%-10%, respectively, showing that considerable numbers of patients are at risk of DR (6). Although early retinopathy does not have a significant effect on vision, it can progress to a more advanced stage, termed proliferative DR (PDR), putting patients at a substantially increased risk of serious complications that can result in permanent vision loss. Therefore, prevention, early detection, and effective management of DR and the identification of associated risk factors are very important for diabetes care to improve public health and ease the financial burden (5).

Several large population-based studies, including the Diabetes Control and Complications Trial (DCCT), the UK Prospective Diabetes Study (UKPDS), the Wisconsin Epidemiologic Study of Diabetic Retinopathy (WESDR), and the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study, have identified other risk factors associated with the development or progression of DR (7, 8, 9, 10, 11). From the results of these studies, age, sex, socioeconomic status, and comorbid systemic arterial hypertension are considered important determinants of retinopathy risk. On the other hand, one study showed that metabolic control as measured by HbA1c and disease duration accounts for only 11% of the risk of DR, leaving 89% of the risk of the development of DR to other factors (12). As these variables are time and ethnicity dependent (13), an analysis based on recent data in Koreans is required. Although there have been several cross-sectional studies regarding the prevalence and associated risk factors in Korea (6, 14, 15), cohort studies on the development and progression of DR in recent decades are not available, to our knowledge. The purpose of this study was to explore the incidence of and risk factors for the development of DR and progression to PDR in Korean patients with type 2 DM since 2000.

Patients who were diagnosed with type 2 DM and followed for more than 5 yr annually or more often at a hospital-based diabetic clinic (Asan Medical Center, Seoul, Korea) were selected by medical record review in a consecutive manner. In this retrospective cohort study, patients with PDR at the initial examination, with concomitant ocular disease other than DR, or history of ocular trauma or intraocular surgery were excluded. If the patient became pregnant or underwent intravitreal anti-vascular endothelial growth factor (VEGF) or steroid injections, only data obtained during the period before the event were analyzed. At baseline, a detailed medical history, including previous treatment for diabetes, hypertension, smoking habits, and alcoholic intake, was taken. At baseline and every visit, arterial blood pressure (BP), body weight, and height were measured, and blood tests, including plasma glucose, HbA1c, and lipid profiles, were performed. Urine tests were also performed and urine albumin-to-creatinine ratios (UACR) were used in the analysis. Body weight and height were used to calculate the body mass index (BMI), which was used for analysis; diabetes medication information was described. Patients were followed every 3 or 6 months depending on their glucose control.

At their initial visit and at each visit to a retina clinic, all patients underwent a comprehensive ophthalmologic examination that included a review of their ophthalmologic history, measurement of visual acuity and intraocular pressure (IOP), slit lamp biomicroscopy, and funduscopic examinations through dilated pupils by retinal specialists. DR was evaluated on the basis of fundus photography taken on every patient at every visit. When necessary, fluorescein angiography was performed to assess vascular abnormalities and the extent of the capillary nonperfusion area. The frequency of the eye examination was at least once a year or more, determined by the ophthalmologist depending on the degree of DR. The degree of DR was classified into 5 stages by the following criteria of the Early Treatment Diabetic Retinopathy Study (ETDRS): 1) no diabetic retinopathy-" no DR"; 2) mild nonproliferative diabetic retinopathy-" mild NPDR"; 3) moderate nonproliferative diabetic retinopathy-" moderate NPDR"; 4) severe nonproliferative diabetic retinopathy-"severe NPDR"; and 5) proliferative diabetic retinopathy-" PDR". If both eyes were rated at different stages, the grade of the worse eye was used.

Based on the DR classification at the initial and final visits, the development of DR and progression to PDR were defined as follows: 1) the development of any type of DR in the patients who were free of DR at baseline-"development of DR"; and 2) the progression to PDR in patients who had NPDR at baseline-"progression to PDR". The incidences of the development of DR and progression to PDR were defined as the number of patients with those events divided by the total sum of person-years of the follow-up period.

In our current study, the development of DR and progression to PDR and associated risk factors in Koreans were examined based on the experiences at our tertiary institution. In patients receiving regular care and management of diabetes, the rate of development to DR from no DR was 32.1/1,000 person-years, and the rate of progression from NPDR to PDR was 26.2/1,000 person-years. These incidence rates were slightly lower than the 44.4/1,000 person-years rate of development of DR and the 37.5/1,000 person-years rate of progression to PDR reported by Kim et al. (16), who studied the rates of progression of DR in Koreans with type 2 diabetes in a similar manner in our center from 1990 to 1996. Baseline HbA1C level was 11.2-11.5%, which was also higher than our current study. It is not accurate, of course, to simply compare the results of different studies based on participants with unequal baseline DRs and socioeconomic backgrounds with different follow-up frequencies. However, the recent decreasing trend in diabetic micro-vasculopathy can be seen in other previous reports (17). In a previously published meta-analysis that reviewed the rates of progression of DR, rates of progression to PDR and severe visual loss were lower among participants in 1986-2008 than in 1975-1985, showing that diabetic patients have lower rates of progression since 1985 (18). By generating a more precise estimate of the prevalence of DR and its relationship with major modifiable risk factors based on many previous large-scale studies, public health education and awareness have improved and the optimal clinical prevention and management of DR could be possible, ultimately lowering the incidence rate of DR and related complications (8, 10, 11).

It is well known that inadequate glycemic control highly correlates with both the development and progression of DR in previous studies (7, 8, 9, 10, 11, 16). Our current findings confirmed these observations and showed that a higher HbA1c is independently associated with an increased risk of DR, i.e. for every 1% increase in HbA1c, the risk of DR development increased 1.16-fold (16%) and the risk of progression to PDR increased 1.19-fold (19%). Whereas the fasting blood sugar and HbA1c values at baseline did not differ in terms of the development of DR or progression to PDR, the mean HbA1c concentration was higher in patients who developed DR and progressed to PDR. This result is in line with many earlier reports that demonstrated that a long-lasting high blood glucose level could aggravate DR and, meanwhile, that strict glucose control could slow the rates of DR development and progression in patients with previously poor glycemic control (16, 19). Of note, the SD of HbA1c, which is considered to show the long-term fluctuation in glycemic control, was significantly higher in the progressor group. Although some previous studies demonstrated that variability in the HbA1c was a risk factor for diabetes-related microvascular diseases, there are still inconsistent results for its correlation with DR in type 2 DM (20). The duration of diabetes, which is related to the duration of high glucose levels, is an important risk factor for DR (21). Rema et al. (22) reported that for every 5-yr increase in the duration of diabetes, the risk of DR increased 1.89-fold. In our present study, for every 1 yr increase, the risk of DR increased by 1.13-fold (13%). Therefore, emphasizing continuous follow-up for DR assessment is important for diabetic patients.

DR and diabetic nephropathy are the 2 major diabetic microvascular complications and the most important causes of blindness and end-stage renal disease in Korea (3). Our findings suggest that patients with diabetic nephropathy show a more rapid progression to PDR, which is in line with the results of most previous studies (23). Albuminuria is one of the indicators of generalized microvascular disease in diabetic patients and is therefore a risk factor for combination DR (24). Nonetheless, inconsistent data also have been reported. In patients with type 2 diabetes, the association between nephropathy and retinopathy is less predictable than that of type 1 diabetes (14). In our present study, after adjustment for other risk factors, patients with albuminuria showed an increased risk of progression to PDR. Thus, more careful ophthalmologic follow-up is recommended when albuminuria develops in patients with DR.

Other significant risk factors for the development of DR and progression to PDR in previous studies were a younger age at DM diagnosis (21), male sex (25), hypertension (19, 23), high systolic BP (26), low diastolic BP (21), hypercholesterolemia (27), anemia (27), elevated systemic neutrophil count (28), use of insulin (29), high waist-to-hip ratio (21), smoking status, and diabetic foot disease (30). In our current study of Korean patients, age, sex, BP, hyperlipidemia, and smoking status did not correlate with DR progression (P>0.05); caloric intake, blood cell count, and diabetic foot disease were not included in our analyses. Some of these differences might be explained that most of our study cohort was in good control of BP and hyperlipidemia. And some might be due to unique characteristics of Korean patients in that a significant portion of Korean diabetic patients are nonobese with type 2 DM, and their body weights often decrease during the disease course (6).

This study has several limitations in terms of generalizing the natural course of DR because it was a tertiary clinic-based study. It is quite possible that only patients with poorly controlled diabetes or a more severe form of diabetes with complications attend our hospital-based tertiary medical center. The mean baseline HbA1c level of study participants was relatively high, indicating this cohort consisted of poorly controlled diabetic patients. In addition, patients with severe forms of diseases generally tend to follow for a longer period. Taken together, the inclusion of patients with poor glycemic control may have overestimated the incidence of DR and PDR. On the other hand, this setting makes it possible to maintain a greater portion of patients who are more concerned about their health and adopt a more active attitude to the control of chronic diseases, such as diabetes. Byun et al. (5) recently reported that socioeconomic factors, including occupation, family income, and education, were associated with the frequency of screening for diabetic complications. In other words, patients with less education or low family income underwent less frequent screening for retinopathy or nephropathy. Thus, the participants of this study who had regular check-ups in the tertiary clinic are considered to have better compliance for screening and management of DR and show lower rates of diabetic complications than the general population. Moreover, patients with high-risk severe NPDR undergo panretinal photocoagulation treatment in most cases, affecting the natural history of DR progression. Therefore, when interpreting the results of our current study, these possible sources of bias should be considered. Large-scale population-based longitudinal studies are needed to examine the natural history of the development of DR and progression to PDR in Korean diabetic patients.

In conclusion, the rates of the development of DR and progression to PDR in Koreans with type 2 diabetes are lower than those reported over the last decade. And this study shows traditional, known risk-factors are still the most influential in the development and progression of DR in cohort of Korean patients with type 2 DM. Poor glycemic control is the common risk factor for the development of DR and progression to PDR. In addition, a longer duration of diabetes and a combination with diabetic nephropathy are significant risk factors for the development of DR and progression to PDR, respectively. Our findings may be used as a reference when planning screening tests and the subsequent follow-up of DR in Koreans with type 2 DM.