December 2020
Volume 61, Issue 14
Open Access
Retina  |   December 2020
Relationship Between Retinal Capillary Nonperfusion Area and Renal Function in Patients With Type 2 Diabetes
Author Affiliations & Notes
  • Ji Won Min
    Department of Nephrology, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea
  • Hyung Duk Kim
    Department of Nephrology, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea
  • Sang Yoon Park
    Department of Biomedical Engineering, Seoul National University, Seoul, Korea
  • Jun Hyuk Lee
    Department of Ophthalmology, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea
  • Jae Hyun Park
    Department of Ophthalmology, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea
  • Anna Lee
    Department of Ophthalmology, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea
  • Ho Ra
    Department of Ophthalmology, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea
  • Jiwon Baek
    Department of Ophthalmology, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea
  • Correspondence: Jiwon Baek, Department of Ophthalmology and Visual Science, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea, #327 Sosa-ro, Wonmi-gu, Bucheon, Gyeonggi-do 14647, Republic of Korea; md.jiwon@gmail.com
  • Footnotes
     JWM and HDK contributed equally to the work presented here and should therefore be regarded as equivalent authors.
Investigative Ophthalmology & Visual Science December 2020, Vol.61, 14. doi:https://doi.org/10.1167/iovs.61.14.14
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Ji Won Min, Hyung Duk Kim, Sang Yoon Park, Jun Hyuk Lee, Jae Hyun Park, Anna Lee, Ho Ra, Jiwon Baek; Relationship Between Retinal Capillary Nonperfusion Area and Renal Function in Patients With Type 2 Diabetes. Invest. Ophthalmol. Vis. Sci. 2020;61(14):14. doi: https://doi.org/10.1167/iovs.61.14.14.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: We sought to assess the relationship between retinal nonperfusion area (NPA) on ultra-widefield fluorescein angiography (UWFA) and renal function in type 2 diabetes mellitus (DM) patients with diabetic retinopathy (DR) and nephropathy.

Methods: UWFA was performed in 248 eyes (124 patients) with DR, comprising 94 eyes from patients with chronic kidney disease (CKD) caused by diabetes and 154 eyes without CKD (non-CKD). Serum creatinine level (Cr), estimated glomerular filtration rate (eGFR), urine albumin/creatinine ratio (UACR), and urine protein/creatinine ratio (UPCR) were collected. On UWFA, retinal NPA was measured in an automated manner. The correlation between NPA and renal function was analyzed.

Results: The mean NPA value of the total eye was 33.11 ± 45.77-disc diameter (DA) in non-CKD and 100.57 ± 69.52 in CKD (P < 0.001). NPA of posterior pole was 1.21 ± 3.28 DA in non-CKD and 7.99 ± 6.75 in CKD group (P < 0.001). The NPA values of both the total eye and posterior pole were significantly correlated with Cr (r = 0.585 and 0.483), eGFR (r = −0.572 and −0.524), UACR (r = 0.541 and 0.482), and UPCR (r = 0.509 and 0.529, respectively) (all P ≤ 0.001). Linear modeling encompassing all clinical factors and relative clinical factors suggested eGFR as the most important predictor for NPAs of the total eye and posterior pole.

Conclusions: Larger retinal NPA on UWFA is associated with worse renal function in DM patients. Renal function can be used to predict retinal NPA in type 2 DM patients with nephropathy and DR.

Diabetes mellitus (DM) is one of the most important and common chronic diseases worldwide and is expected to increase in prevalence alongside continued population growth, aging, and escalating rates of obesity.1,2 Dealing with complications of DM is important because the morbidity and mortality rates and health care costs for diabetic patients are critical socioeconomic issues that must be brought under control.1 Diabetic retinopathy (DR) and diabetic nephropathy are among the most significant complications of DM. 
Both DR and diabetic nephropathy are typical microvascular complications of DM. DR is the most common cause of visual disability in people of working age.3 Meanwhile, diabetic nephropathy, the primary cause of chronic kidney disease (CKD), accounts for 40% of all new cases of end-stage renal disease recorded annually.4 
The association between DR and diabetic nephropathy has been investigated previously under the hypothesis that they share a common pathophysiology because both result from diabetic microangiopathy. The prevalence of DR was increased in DM patients with CKD.5,6 The severity of DR has been linked to a high urine albumin to creatinine ratio (UACR), lower estimated glomerular filtration rate (eGFR), and deterioration of renal function.710 Other studies have suggested that CKD and DR share common risk factors such as smoking, poor glycemic control, systolic hypertension, and dyslipidemia.6,8,11,12 In a recent meta-analysis, DR was proved to be helpful in diagnosing nephropathy in patients with type 2 DM.13 
Most of the previous studies used ophthalmoscopic or conventional fundus photographic findings to define DR, whereas angiographic findings can provide more valuable information on angiopathy in DR eyes. Capillary nonperfusion (NP) is an important sign of microvascular injury seen during fundus angiography of DR eyes. It represents the extent of ischemia, thus signifying impairment of the capillary perfusion status of retinal vessels. The NP area (NPA) is directly correlated with DR severity and neovascularization (NV), which is a pathognomonic finding of proliferative DR (PDR).14,15 Classically, NPA is easily detected by fluorescein angiography. The recent introduction of ultra-widefield imaging has enabled more detailed analyses of NPA by covering most of the retinal area.15,16 
Microvascular pathology also plays an important role in development of renal insufficiency in diabetic nephropathy.17 Considering the proven relationship between DR and nephropathy, microangiopathy associated with DM in the kidney, hence, renal function and proteinuria, can be correlated with microangiopathy in the retina, presented as retinal capillary NPA, proving that this can have clinical advantages. In the viewpoint of nephrologists, maintaining good vision is important for the patients’ quality of life, especially those undergoing dialysis. Quantified correlation between renal function and angiopathy in DR may assist proper referral to DR care. For ophthalmologists, the parameters of renal function can give clues to estimate vascular changes in DM patients who are not suitable for fluorescein angiography because of decreased renal function. 
Because most of the NPAs are located in the peripheral area, NPA measurement ultra-widefield fluorescein angiography (UWFA) can represent global vascular changes in DR eyes.14,18 To the best of our knowledge, there have been no previous studies on the association of DR and nephropathy using quantitative angiographic parameters or ultra-wide field imaging. In this study, we sought to assess the relationship between retinal NPA on UWFA and renal function and the performance of renal function in the prediction of NPA in patients with type 2 DM and DR. 
Methods and Materials
Study Subjects
This was a retrospective cross-sectional study performed by retrospective chart review at Bucheon St. Mary's Hospital, The Catholic University of Korea, Gyeonggi-do, Korea. The institutional review board and ethics committee of the Bucheon St. Mary's Hospital approved this study (HC20RISI0070), which adhered to the tenets of the Declaration of Helsinki. The need for informed consent was waived due to the retrospective nature of the study. 
The study group consisted of all consecutive type 2 DM patients who visited the outpatient departments of ophthalmology and endocrinology and/or nephrology and underwent UWFA, between August 2016 and December 2019. Because of the invasive nature of UWFA imaging, patients with advanced stages of DR (i.e., severe nonproliferative DR [NPDR] or PDR) were included. Patients with decreased renal function were consulted to the internist for medical care before and after UWFA. Type 2 DM was diagnosed if the patient exhibited a fasting plasma glucose level of ≥126 mg/dL or a two-hour post-glucose level of ≥200 mg/dL after a 75-g oral glucose tolerance test.1 Severity of DR was evaluated in accordance with the Early Treatment Diabetic Retinopathy Study (ETDRS) standard grading protocols.19 PDR was manifested by posterior segment NV, which was defined as presence of NV at the disc or elsewhere, with visible angiographic evidence.19 Diagnoses of diabetic nephropathy and CKD were made by a nephrologist according to the National Kidney Foundation's Kidney Disease Outcomes Quality Initiative classification.20 Subjects with eGFR less than 60 ml/min/m2 for at least three months were categorized into the CKD group. A simplified scheme for this study is presented in Figure 1
Figure 1.
 
Simplified flow chart of the study. Consecutive type 2 DM patients who visited the outpatient departments of both ophthalmology and endocrinology and/or nephrology were identified. Of those, severe non-proliferative DR and proliferative DR patients that had ultra-widefield fluorescein (UWFA) angiography were selected. On UWFA, nonperfusion area (NPA) measurement was performed. The NPA was compared between patients with and without chronic kidney disease and correlated with parameters of renal function. Results from correlation was applied to prediction of NPA using renal function.
Figure 1.
 
Simplified flow chart of the study. Consecutive type 2 DM patients who visited the outpatient departments of both ophthalmology and endocrinology and/or nephrology were identified. Of those, severe non-proliferative DR and proliferative DR patients that had ultra-widefield fluorescein (UWFA) angiography were selected. On UWFA, nonperfusion area (NPA) measurement was performed. The NPA was compared between patients with and without chronic kidney disease and correlated with parameters of renal function. Results from correlation was applied to prediction of NPA using renal function.
Exclusion criteria were low-quality images (e.g., significant vitreous hemorrhage, cataract, or other media opacity); any prior treatment for DR including anti-vascular endothelial growth factor therapy, intraocular or periocular steroids, laser photocoagulation, and vitrectomy; presence of any significant retinal pathology other than PDR; renal replacement therapy including dialysis and renal transplant; and CKD caused by an etiology other than DM such as other glomerular disease, vascular disease, tubulointerstitial disease, or cystic disease. 
Measurement of Retinal Capillary NPA on UWFA
Mydriatic UWFA (Optos California P200DTx icg; Optos, Dunfermline, UK) was performed using a standard image view of 200° in a single capture. Early- to mid-phase angiograms (between 40 seconds and one minute after the dye injection) with minimal eyelid or eyelash artifacts, were chosen for analysis of NPA. An NPA was defined as any area of the retina that lacked fluorescein-filled capillaries.21 Because of variation in the overall size, shape, location, and intensity of the lesions, measurement of NPA can be a difficult job. However, NPA is characterized by change in texture and intensity compared to adjacent retinal tissue, and automated measurement is available using these properties. In this study, we used an automated method for the measurement of NPA, using MATLAB R2020a (The MathWorks, Inc., Natick, MA, USA) as described previously.14 The method used four different properties of the images: local brightness was used to compare intensity of a region or an object to the intensity of residual background, and contextual information of the region of interest such as variance, uniformity, and entropy were used. Briefly, preprocessed images using homomorphic filtering were automatically segmented, and NPAs were selected using a threshold of properties including mean image intensity, uniformity, and entropy. The NPA of the total gradable retinal area was calculated following the final modification (total NPA). Separately, the NPA of the posterior pole (circular area of three-disc diameters centered at the fovea) was also measured (pp NPA) (Fig. 2). Measured NPAs were divided by disc area of each eye and presented as disc area (DA). The measurement of NPA was performed individually by two trained ophthalmologists (J. B. and A. L.), and the average of their two measurements in each case was used for analysis. 
Figure 2.
 
Assessment of nonperfusion area (NPA) on ultra-widefield fluorescein angiography (UWSA). (Left) Two regions of interest were selected (total retina and posterior pole) on UWFA and saved as Bitmap files, which were preprocessed with a homomorphic filter. (Middle) The images were then segmented automatically. (Right) NPAs were chosen using a threshold of area properties, and the sum of the selected areas was calculated.
Figure 2.
 
Assessment of nonperfusion area (NPA) on ultra-widefield fluorescein angiography (UWSA). (Left) Two regions of interest were selected (total retina and posterior pole) on UWFA and saved as Bitmap files, which were preprocessed with a homomorphic filter. (Middle) The images were then segmented automatically. (Right) NPAs were chosen using a threshold of area properties, and the sum of the selected areas was calculated.
Clinical Information and Laboratory Analysis
Clinical information of the patient was collected and comprised age, sex, DM duration, hemoglobin A1c, comorbid hypertension, hypertension duration, systolic blood pressure, diastolic blood pressure, heart rate, body mass index, and comorbid cardio- or cerebrovascular disease (CVD). Serum creatinine level (Cr), eGFR, urine albumin–creatinine ratio (UACR), and urine protein–creatinine ratio (UPCR) were chosen as parameters for renal function. Additionally, serum levels of hemoglobin, albumin, total cholesterol, triglycerides, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, calcium, and phosphate were recorded. All data were obtained within one month from the date of UWFA examination. 
Statistical Analysis
Statistical analyses were performed with the Statistical Package for the Social Sciences version 22.0.1 for Windows (IBM Corp., Armonk, NY, USA). Continuous variables are presented as mean ± standard deviation. Independent t-test was used to compare continuous variables between groups. The Mann-Whitney U test was used when normal distribution could not be confirmed. The χ2 test was used for comparison of categorical variables. Intraclass correlation coefficient (ICC) between two graders for NPA measurements were calculated. Analysis of variance (ANOVA) was performed considering NPAs of the eGFR categories, and the Bonferroni test was applied as the post-hoc test. Pearson's correlation analysis was used to determine the correlations between NPA and clinical parameters. Multiple linear regression analysis was performed for covariate adjustment. A P value <0.05 was considered statistically significant. 
Results
Baseline Demographic and Clinical Features
In total, 248 eyes of 124 type 2 DM patients with treatment-naïve DR were enrolled. The mean age was 57.90 ± 11.21 years, and 54% of the study population was male. One-hundred sixty-six eyes were classified as non-CKD, and 82 eyes were classified as CKD. For non-CKD and CKD eyes, the mean Cr, eGFR, UACR, and UPCR values were 0.84 ± 0.27 versus 2.74 ± 1.93 mg/dL, 91.27 ± 26.94 versus 34.44 ± 19.66 mL/min/1.73 m2, 211.23 ± 398.53 versus 2757.01 ± 2861.67 mg/g, and 0.54 ± 1.04 versus 4.07 ± 3.78 mg/g, respectively (all P < 0.001). Furthermore, the non-CKD and CKD groups differed in age, DM duration, DR grade, comorbid hypertension, hypertension duration, and comorbid CVD (all P ≤ 0.041). The demographic and clinical features of the study eyes are summarized and compared between the non-CKD and CKD groups in Table 1
Table 1.
 
Baseline Demographic and Clinical Characteristics of the Enrolled Patients
Table 1.
 
Baseline Demographic and Clinical Characteristics of the Enrolled Patients
Comparison of NPAs and Correlations Between Retinal NPA and Clinical Parameters
The mean total and pp NPA values were significantly larger in the CKD group than in the non-CKD group (33.11 ± 45.77 DA and 100.57 ± 69.52 DA versus 1.21 ± 3.28 DA and 7.99 ± 6.75 DA; both P < 0.001) (Fig. 3A). The ICCs for total and pp NPAs were 0.956 and 0.951, respectively. ANOVA revealed that the NPA values differed according to eGFR category, whereas post-hoc analysis indicated that NPAs of eGFR grades 3, 4, and 5 were significantly larger than those of eGFR grades 1 and 2 (all P ≤ 0.002) (Fig. 3B). 
Figure 3.
 
Comparisons of total retina and posterior pole nonperfusion according to CKD group and eGFR category. (A) Comparison of NPA according to CKD status shows larger NPAs in the CKD group compared to the non-CKD group. (B) Comparison of NPA according to eGFR category shows that the NPAs of eGFR grades 3, 4, and 5 were significantly larger than those of eGFR grades 1 and 2. o indicates outlier and * indicates extreme outlier. a. Independent t-test between non-CKD and CKD group. b. Independent t-test between two eGFR categories. c. One-way analysis of covariance between eGFR categories 3, 4, and 5. d. Independent t-test between eGFR category 1, 2 and eGFR category 3,4,5.
Figure 3.
 
Comparisons of total retina and posterior pole nonperfusion according to CKD group and eGFR category. (A) Comparison of NPA according to CKD status shows larger NPAs in the CKD group compared to the non-CKD group. (B) Comparison of NPA according to eGFR category shows that the NPAs of eGFR grades 3, 4, and 5 were significantly larger than those of eGFR grades 1 and 2. o indicates outlier and * indicates extreme outlier. a. Independent t-test between non-CKD and CKD group. b. Independent t-test between two eGFR categories. c. One-way analysis of covariance between eGFR categories 3, 4, and 5. d. Independent t-test between eGFR category 1, 2 and eGFR category 3,4,5.
Total NPA showed positive correlations with pp NPA, DM duration, DR grade, comorbid hypertension, Cr, UACR, and UPCR (r = 0.737, 0.147, 0.277, 0.585, 0.541, and 0.509; all P ≤ 0.021), but a negative correlation with eGFR (r = −0.572; P < 0.001). Meanwhile, pp NPA showed positive correlations with total NPA, DR grade, comorbid hypertension, Cr, UACR, and UPCR (r = 0.737, 0.243, 0.209, 0.483, 0.482, and 0.529; all P ≤ 0.001) and negative correlations with HbA1c and Cr (r = −0.194 and −0.524; both P ≤ 0.002). Correlation coefficients and P values are summarized in Table 2
Table 2.
 
Correlation Between Retinal Capillary Nonperfusion Area and Clinical Parameters in Total Subjects
Table 2.
 
Correlation Between Retinal Capillary Nonperfusion Area and Clinical Parameters in Total Subjects
Detailed Analysis of the Association Between NPA and Renal Function
The clinical factors that differed between non-CKD and CKD eyes (i.e., DM duration, DR grade, hypertension, and CVD) did not affect the finding that NPAs in the CKD group were significantly larger than those in the non-CKD group (all P ≤ 0.007), except with respect to the impact of CVD on total NPA (P = 0.107) (Fig. 4). Two-way ANOVA also revealed no significant effect of DM duration, DR grade, hypertension, and CVD on NPA differences between non-CKD and CKD eyes (F = 1.349, 0.99, 0.267, and 1.122 and P = 0.247, 0.753, 0.606, and 0.291 for total NPA; F = 0.340, 0.629, 1.060, and 1.720 and P = 0.560, 0.428, 0.304, and 0.191 for pp NPA, respectively). 
Figure 4.
 
Comparisons of NPAs between the CKD and non-CKD groups with regard to factors that differed between the groups. (A) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to disease duration. (B) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to DR grade. (C) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to presence of hypertension. (D) Posterior polar NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to presence of comorbid cardio-/cerebrovascular disease. In the presence of CVD, total NPA was larger in the CKD group, but it failed to reach statistical significance. Blue bar indicates non-CKD and green bar indicates CKD group. P-values are obtained from independent t-test between non-CKD and CKD group.
Figure 4.
 
Comparisons of NPAs between the CKD and non-CKD groups with regard to factors that differed between the groups. (A) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to disease duration. (B) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to DR grade. (C) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to presence of hypertension. (D) Posterior polar NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to presence of comorbid cardio-/cerebrovascular disease. In the presence of CVD, total NPA was larger in the CKD group, but it failed to reach statistical significance. Blue bar indicates non-CKD and green bar indicates CKD group. P-values are obtained from independent t-test between non-CKD and CKD group.
Separately, correlations between NPAs and all parameters concerning renal function (i.e., Cr, eGFR, UACR, and UPCR) remained significant after adjusting for clinical factors that showed correlation with NPA (i.e. DM duration, DR grade, and hypertension) (all P < 0.001). Linear regression analysis including all these factors resulted in good prediction of NPA (R2 = 0.444 for total NPA and R2 = 0.448 for pp NPA; both P < 0.001). Overall, the coefficients were significant for hypertension, Cr, and eGFR (P = 0.009, 0.001, and 0.018 for total NPA, respectively, and P = 0.006, < 0.001, and 0.002 for pp NPA) (Tables 3A3B). 
Table 3A.
 
Correlation Between Nonperfusion Areas and Parameters for Renal Function Adjusted by Retinopathy Grade, Disease Duration, and Comorbid Hypertension: Total NPA
Table 3A.
 
Correlation Between Nonperfusion Areas and Parameters for Renal Function Adjusted by Retinopathy Grade, Disease Duration, and Comorbid Hypertension: Total NPA
Table 3B.
 
Correlation Between Nonperfusion Areas and Parameters for Renal Function Adjusted by Retinopathy Grade, Disease Duration, and Comorbid Hypertension: Posterior Polar NPA
Table 3B.
 
Correlation Between Nonperfusion Areas and Parameters for Renal Function Adjusted by Retinopathy Grade, Disease Duration, and Comorbid Hypertension: Posterior Polar NPA
Automatic linear modeling with all clinical factors revealed accuracy of 51.8% for total NPA and 48.0% for pp NPA. The eGFR category was the most important predictor for both total and pp NPAs (Fig. 5A). Meanwhile, automatic linear modeling including only parameters showing correlation with NPA displayed accuracy of 48.7% for total NPA and 46.7% for pp NPA. Again, the eGFR category was the most important predictor for both total and pp NPAs (Fig. 5B). 
Figure 5.
 
Automatic linear modeling using clinical factors and renal function parameters for total and posterior polar nonperfusion areas. (A) Linear modeling with all clinical factors revealed an accuracy of 51.8% for total NPA and 48.0% for pp NPA. eGFR category was the most important predictor for both total and pp NPAs. (B) Linear modeling including only parameters showing correlation with NPA revealed accuracy of 48.7% for total NPA and 46.7% for pp NPA. The eGFR category was the most important predictor for both total and pp NPA.
Figure 5.
 
Automatic linear modeling using clinical factors and renal function parameters for total and posterior polar nonperfusion areas. (A) Linear modeling with all clinical factors revealed an accuracy of 51.8% for total NPA and 48.0% for pp NPA. eGFR category was the most important predictor for both total and pp NPAs. (B) Linear modeling including only parameters showing correlation with NPA revealed accuracy of 48.7% for total NPA and 46.7% for pp NPA. The eGFR category was the most important predictor for both total and pp NPA.
Discussion
The association between DR and diabetic nephropathy, two important microvascular complications of DM, has been actively investigated, and the results have implied an intimate relationship between the manifestations and severity of DR and renal function and its progression in patients with DM.5,6,8,9,2226 Retinal capillary nonperfusion is a major sign of microvascular insufficiency of DR, reflecting the extent of ischemia and is directly involved in development of PDR, the end stage of DR.14,16,18,27 Ultra-widefield imaging has enabled more thorough assessments of retinal vascularity including the far peripheral retina, therefore enabling a more precise assessment of the NPA in DR eyes.1416,18,28 In this study, we measured retinal NPAs quantitatively on UWFA in an objective manner and analyzed their correlation with renal function in patients with type 2 DM with DR. The results demonstrated a significant association between retinal NPA and renal function and indicated that eGFR grade was the most potent predictor of retinal NPA. 
The mean NPA values were significantly larger in the CKD group than in the non-CKD group and, when stratified according to eGFR category, the NPA values of eGFR grade 3 or more were significantly larger than those of eGFR grades 1 and 2. Retinal capillary nonperfusion is an irreversible microvascular change in DR.29 Retinal hypoxia is a powerful stimulator for capillary abnormalities and formation of NV, while nonperfusion is the leading cause of hypoxia in DR eyes.14,30 Therefore, assessment of nonperfusion can highlight more severe stages of microangiopathy in DM patients and this can reflect the status of renal function in these patients. Although most previous studies have revealed an association between DR and diabetic nephropathy on the basis of color fundus photography, the current study adopted angiographic imaging, which enabled assessment of NPA. In addition, in contrast to conventional color fundus photography, which can only show the posterior pole of the retina, the UWFA imaging used in this study can reveal more than 80% of the total retinal area. This is a major evolution relative to the outcomes of recent reports on values of peripheral lesions in DR.28,31,32 
DM duration, DR grade, and comorbid hypertension and CVD are well-known risk factors for CKD in DM.5,8,33 In this study, these characteristics were also revealed as baseline clinical factors that significantly varied between non-CKD and CKD eyes. The difference in NPA values between the non-CKD and CKD conditions was significant after adjusting for DM duration, DR grade, and comorbid hypertension, suggesting an independent association between NPA and CKD regardless of other relevant factors. This is in accordance with previous research contending an association between presence of DR and renal function deterioration independent of associated DM or hypertension.7 When divided by comorbid CVD, NPA values between the non-CKD and CKD groups did not differ significantly. A strong association between DR and CVD was previously suggested by Grunwald et al.,33 even following adjustments for renal dysfunction and traditional CVD risk factors. The ischemic status of the posterior pole can be affected not only by renal function, but also by other factors associated with CVD, such as atherosclerosis. Nonetheless, total NPA, which represents the global ischemic status of DR, was still correlated with presence of CKD. 
Both total and pp NPA showed positive correlations with renal function. Other involved clinical factors were DR grade, DM duration, and hypertension for total NPA and DR grade, hypertension, and HbA1c for pp NPA, respectively. The correlations between NPA and Cr, eGFR, UACR, and UPCR, respectively, remained significant after adjusting these factors, suggesting an independent correlation between NPA and renal function. 
Furthermore, linear modeling with all clinical factors and relative clinical factors revealed eGFR category to be the most important predictor for both total and pp NPA. The eGFR category is influenced by all parameters of retinal function. Thus this result may be interpreted to mean that renal function, among other clinical factors, has the most intimate relationship with NPA. Prediction of NPA using renal function can improve management of DM patients with CKD by assisting proper referral of these patients and enabling prompt treatment for NPAs which deteriorates DR state and causes eventual vision loss. Preservation of vision is especially important in these patients because many CKD patients require either peritoneal dialysis or hemodialysis. For patients on peritoneal dialysis, preservation of vision is mandatory to reduce possibilities of infection that increases mortality of the patient. And patients on hemodialysis also require enough vision for three times per week ambulation to the hospital. In addition, the strong correlation between NPA and renal function may offer a rationale for estimation of renal function using NPA observed via angiography in DM patients. 
There are several limitations to consider in this study. First, treatment regimens for DM and diabetic nephropathy were not controlled because of the retrospective nature of the study. This factor requires further validation. Second, the patients involved in this study had advanced stages of DR, so the results of this investigation may not be applicable to earlier stages of DR. Third, angiographic findings other than nonperfusion, such as extent of microaneurysm, leakage, and neovascularization, should be considered in DR. This also requires further research, despite inherent difficulties in quantification of these parameters. Forth, selection bias can occur by omitting patients with advanced DR and not followed-up for systemic conditions in our institution. Further, NPA at the posterior pole can be more precisely visualized using OCT angiography, so additional validation of our results regarding the posterior pole warrants further assessment by OCT angiography. And finally, the subject of this study was restricted to advanced stages of DR because of the invasive nature of UWFA imaging. Prospective studies involving patients at earlier stages of DR are warranted to encompass this correlation in the general DR population. 
In conclusion, DR eyes of CKD patients showed larger retinal NPA compared with those of non-CKD patients, and larger retinal NPA was associated with worse renal function in DM patients independent of other risk factors. These results imply a strong relationship between retinal NPA and renal function in DM. Also, eGFR category was the most potent predictor of retinal NPA, suggesting that renal function can be used to predict retinal NPA. Further studies relying on other methods for angiography and other angiographic findings are warranted. 
Acknowledgments
Supported by Busan Sungmo Eye Hospital Sodam Scholarship Committee. 
Disclosure: J.W. Min, None; H.D. Kim, None; S.Y. Park, None; J.H. Lee, None; J.H. Park, None; A. Lee, None; H. Ra, None; J. Baek, None 
References
Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: Diagnosis and classification of diabetes mellitus provisional report of a who consultation. Diabet Med. 1998; 15: 539–553. [CrossRef] [PubMed]
Danaei G, Finucane MM, Lu Y, et al. National, regional, and global trends in fasting plasma glucose and diabetes prevalence since 1980: Systematic analysis of health examination surveys and epidemiological studies with 370 country-years and 2·7 million participants. Lancet. 2011; 378: 31–40. [CrossRef] [PubMed]
Aiello LP, Gardner TW, King GL, et al. Diabetic retinopathy. Diabet Care. 1998; 21: 143–156. [CrossRef]
Fong DS, Aiello LP, Ferris FL, 3rd, Klein R. Diabetic retinopathy. Diabet Care. 2004; 27: 2540–2553. [CrossRef]
Wong TY, Coresh J, Klein R, et al. Retinal microvascular abnormalities and renal dysfunction: the atherosclerosis risk in communities study. J Am Soc Nephrol. 2004; 15: 2469–2476. [CrossRef] [PubMed]
Grunwald JE, Alexander J, Maguire M, et al. Prevalence of ocular fundus pathology in patients with chronic kidney disease. Clin J Am Soc Nephrol. 2010; 5: 867–873. [CrossRef] [PubMed]
Edwards MS, Wilson DB, Craven TE, et al. Associations between retinal microvascular abnormalities and declining renal function in the elderly population: the cardiovascular health study. Am J Kidney Dis. 2005; 46: 214–224. [CrossRef] [PubMed]
Grunwald JE, Alexander J, Ying G-S, et al. Retinopathy and chronic kidney disease in the chronic renal insufficiency cohort (cric) study. Arch Ophthalmol. 2012; 130: 1136–1144. [CrossRef] [PubMed]
Lee WJ, Sobrin L, Kang MH, et al. Ischemic diabetic retinopathy as a possible prognostic factor for chronic kidney disease progression. Eye (Lond). 2014; 28: 1119–1125. [CrossRef] [PubMed]
Rodriguez-Poncelas A, Mundet-Tuduri X, Miravet-Jimenez S, et al. Chronic kidney disease and diabetic retinopathy in patients with type 2 diabetes. PLoS One. 2016; 11: e0149448. [CrossRef] [PubMed]
The diabetes control and complications trial (dcct). Design and methodologic considerations for the feasibility phase. The dcct research group. Diabetes. 1986; 35: 530–545. [CrossRef] [PubMed]
Reichard P, Nilsson BY, Rosenqvist U. The effect of long-term intensified insulin treatment on the development of microvascular complications of diabetes mellitus. N Engl J Med. 1993; 329: 304–309. [CrossRef] [PubMed]
Jiang S, Yu T, Zhang Z, et al. Diagnostic performance of retinopathy in the detection of diabetic nephropathy in type 2 diabetes: A systematic review and meta-analysis of 45 studies. Ophthalmic Res. 2019; 62: 68–79. [CrossRef] [PubMed]
Ra H, Park JH, Baek JU, Baek J. Relationships among retinal nonperfusion, neovascularization, and vascular endothelial growth factor levels in quiescent proliferative diabetic retinopathy. J Clin Med. 2020; 9: 1462. [CrossRef]
Nicholson L, Ramu J, Chan EW, et al. Retinal nonperfusion characteristics on ultra-widefield angiography in eyes with severe nonproliferative diabetic retinopathy and proliferative diabetic retinopathy. JAMA Ophthalmol. 2019; 137: 626–631. [CrossRef] [PubMed]
Baxter SL, Ashir A, Nguyen BJ, Nudleman E. Quantification of retinal nonperfusion associated with posterior segment neovascularization in diabetic retinopathy using ultra-widefield fluorescein angiography. Ophthalmic Surg Lasers Imaging Retina. 2019; 50: 86–92. [CrossRef] [PubMed]
Barrett EJ, Liu Z, Khamaisi M, et al. Diabetic microvascular disease: An endocrine society scientific statement. J Clin Endocrinol Metab. 2017; 102: 4343–4410. [CrossRef] [PubMed]
Fan W, Nittala MG, Velaga SB, et al. Distribution of non-perfusion and neovascularization on ultra-wide field fluorescein angiography in proliferative diabetic retinopathy (recovery study): Report 1. Am J Ophthalmol. 2019; 206: 154–160. [CrossRef] [PubMed]
Fundus photographic risk factors for progression of diabetic retinopathy. Etdrs report number 12. Early treatment diabetic retinopathy study research group. Ophthalmology. 1991; 98: 823–833. [CrossRef] [PubMed]
National Kidney F. K/doqi clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002; 39: S1–S266. [PubMed]
Cardillo Piccolino F, Zingirian M, Mosci C. Classification of proliferative diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol. 1987; 225: 245–250. [CrossRef] [PubMed]
He F, Xia X, Wu XF, Yu XQ, Huang FX. Diabetic retinopathy in predicting diabetic nephropathy in patients with type 2 diabetes and renal disease: a meta-analysis. Diabetologia. 2013; 56: 457–466. [CrossRef] [PubMed]
Hung CC, Lin HY, Hwang DY, et al. Diabetic retinopathy and clinical parameters favoring the presence of diabetic nephropathy could predict renal outcome in patients with diabetic kidney disease. Sci Rep. 2017; 7: 1236. [CrossRef] [PubMed]
Jeng CJ, Hsieh YT, Yang CM, Yang CH, Lin CL, Wang IJ. Diabetic retinopathy in patients with diabetic nephropathy: development and progression. PLoS One. 2016; 11: e0161897. [CrossRef] [PubMed]
Kotlarsky P, Bolotin A, Dorfman K, Knyazer B, Lifshitz T, Levy J. Link between retinopathy and nephropathy caused by complications of diabetes mellitus type 2. Int Ophthalmol. 2015; 35: 59–66. [CrossRef] [PubMed]
Park HC, Lee Y-K, Cho A, et al. Diabetic retinopathy is a prognostic factor for progression of chronic kidney disease in the patients with type 2 diabetes mellitus. PloS one. 2019; 14: e0220506–e0220506. [CrossRef] [PubMed]
Niki T, Muraoka K, Shimizu K. Distribution of capillary nonperfusion in early-stage diabetic retinopathy. Ophthalmology. 1984; 91: 1431–1439. [CrossRef] [PubMed]
Silva PS, Dela Cruz AJ, Ledesma MG, et al. Diabetic retinopathy severity and peripheral lesions are associated with nonperfusion on ultrawide field angiography. Ophthalmology. 2015; 122: 2465–2472. [CrossRef] [PubMed]
Kohner EM. Dynamic changes in the microcirculation of diabetics as related to diabetic microangiopathy. Acta Med Scand Suppl. 1975; 578: 41–47. [PubMed]
Merin S, Ber I, Ivry M. Retinal ischemia (capillary nonperfusion) and retinal neovascularization in patients with diabetic retinopathy. Ophthalmologica. 1978; 177: 140–145. [CrossRef] [PubMed]
Brown DM, Ou WC, Wong TP, Kim RY, Croft DE, Wykoff CC. Targeted retinal photocoagulation for diabetic macular edema with peripheral retinal nonperfusion: three-year randomized dave trial. Ophthalmology. 2018; 125: 683–690. [CrossRef] [PubMed]
Torp TL, Kawasaki R, Wong TY, Peto T, Grauslund J. Peripheral capillary non-perfusion in treatment-naive proliferative diabetic retinopathy associates with postoperative disease activity 6 months after panretinal photocoagulation. Br J Ophthalmol. 2018; 103: 816–820. [CrossRef] [PubMed]
Grunwald JE, Ying GS, Maguire M, et al. Association between retinopathy and cardiovascular disease in patients with chronic kidney disease (from the chronic renal insufficiency cohort [cric] study). Am J Cardiol. 2012; 110: 246–253. [CrossRef] [PubMed]
Figure 1.
 
Simplified flow chart of the study. Consecutive type 2 DM patients who visited the outpatient departments of both ophthalmology and endocrinology and/or nephrology were identified. Of those, severe non-proliferative DR and proliferative DR patients that had ultra-widefield fluorescein (UWFA) angiography were selected. On UWFA, nonperfusion area (NPA) measurement was performed. The NPA was compared between patients with and without chronic kidney disease and correlated with parameters of renal function. Results from correlation was applied to prediction of NPA using renal function.
Figure 1.
 
Simplified flow chart of the study. Consecutive type 2 DM patients who visited the outpatient departments of both ophthalmology and endocrinology and/or nephrology were identified. Of those, severe non-proliferative DR and proliferative DR patients that had ultra-widefield fluorescein (UWFA) angiography were selected. On UWFA, nonperfusion area (NPA) measurement was performed. The NPA was compared between patients with and without chronic kidney disease and correlated with parameters of renal function. Results from correlation was applied to prediction of NPA using renal function.
Figure 2.
 
Assessment of nonperfusion area (NPA) on ultra-widefield fluorescein angiography (UWSA). (Left) Two regions of interest were selected (total retina and posterior pole) on UWFA and saved as Bitmap files, which were preprocessed with a homomorphic filter. (Middle) The images were then segmented automatically. (Right) NPAs were chosen using a threshold of area properties, and the sum of the selected areas was calculated.
Figure 2.
 
Assessment of nonperfusion area (NPA) on ultra-widefield fluorescein angiography (UWSA). (Left) Two regions of interest were selected (total retina and posterior pole) on UWFA and saved as Bitmap files, which were preprocessed with a homomorphic filter. (Middle) The images were then segmented automatically. (Right) NPAs were chosen using a threshold of area properties, and the sum of the selected areas was calculated.
Figure 3.
 
Comparisons of total retina and posterior pole nonperfusion according to CKD group and eGFR category. (A) Comparison of NPA according to CKD status shows larger NPAs in the CKD group compared to the non-CKD group. (B) Comparison of NPA according to eGFR category shows that the NPAs of eGFR grades 3, 4, and 5 were significantly larger than those of eGFR grades 1 and 2. o indicates outlier and * indicates extreme outlier. a. Independent t-test between non-CKD and CKD group. b. Independent t-test between two eGFR categories. c. One-way analysis of covariance between eGFR categories 3, 4, and 5. d. Independent t-test between eGFR category 1, 2 and eGFR category 3,4,5.
Figure 3.
 
Comparisons of total retina and posterior pole nonperfusion according to CKD group and eGFR category. (A) Comparison of NPA according to CKD status shows larger NPAs in the CKD group compared to the non-CKD group. (B) Comparison of NPA according to eGFR category shows that the NPAs of eGFR grades 3, 4, and 5 were significantly larger than those of eGFR grades 1 and 2. o indicates outlier and * indicates extreme outlier. a. Independent t-test between non-CKD and CKD group. b. Independent t-test between two eGFR categories. c. One-way analysis of covariance between eGFR categories 3, 4, and 5. d. Independent t-test between eGFR category 1, 2 and eGFR category 3,4,5.
Figure 4.
 
Comparisons of NPAs between the CKD and non-CKD groups with regard to factors that differed between the groups. (A) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to disease duration. (B) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to DR grade. (C) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to presence of hypertension. (D) Posterior polar NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to presence of comorbid cardio-/cerebrovascular disease. In the presence of CVD, total NPA was larger in the CKD group, but it failed to reach statistical significance. Blue bar indicates non-CKD and green bar indicates CKD group. P-values are obtained from independent t-test between non-CKD and CKD group.
Figure 4.
 
Comparisons of NPAs between the CKD and non-CKD groups with regard to factors that differed between the groups. (A) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to disease duration. (B) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to DR grade. (C) NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to presence of hypertension. (D) Posterior polar NPAs were larger in the CKD group than in the non-CKD group when the subjects were divided according to presence of comorbid cardio-/cerebrovascular disease. In the presence of CVD, total NPA was larger in the CKD group, but it failed to reach statistical significance. Blue bar indicates non-CKD and green bar indicates CKD group. P-values are obtained from independent t-test between non-CKD and CKD group.
Figure 5.
 
Automatic linear modeling using clinical factors and renal function parameters for total and posterior polar nonperfusion areas. (A) Linear modeling with all clinical factors revealed an accuracy of 51.8% for total NPA and 48.0% for pp NPA. eGFR category was the most important predictor for both total and pp NPAs. (B) Linear modeling including only parameters showing correlation with NPA revealed accuracy of 48.7% for total NPA and 46.7% for pp NPA. The eGFR category was the most important predictor for both total and pp NPA.
Figure 5.
 
Automatic linear modeling using clinical factors and renal function parameters for total and posterior polar nonperfusion areas. (A) Linear modeling with all clinical factors revealed an accuracy of 51.8% for total NPA and 48.0% for pp NPA. eGFR category was the most important predictor for both total and pp NPAs. (B) Linear modeling including only parameters showing correlation with NPA revealed accuracy of 48.7% for total NPA and 46.7% for pp NPA. The eGFR category was the most important predictor for both total and pp NPA.
Table 1.
 
Baseline Demographic and Clinical Characteristics of the Enrolled Patients
Table 1.
 
Baseline Demographic and Clinical Characteristics of the Enrolled Patients
Table 2.
 
Correlation Between Retinal Capillary Nonperfusion Area and Clinical Parameters in Total Subjects
Table 2.
 
Correlation Between Retinal Capillary Nonperfusion Area and Clinical Parameters in Total Subjects
Table 3A.
 
Correlation Between Nonperfusion Areas and Parameters for Renal Function Adjusted by Retinopathy Grade, Disease Duration, and Comorbid Hypertension: Total NPA
Table 3A.
 
Correlation Between Nonperfusion Areas and Parameters for Renal Function Adjusted by Retinopathy Grade, Disease Duration, and Comorbid Hypertension: Total NPA
Table 3B.
 
Correlation Between Nonperfusion Areas and Parameters for Renal Function Adjusted by Retinopathy Grade, Disease Duration, and Comorbid Hypertension: Posterior Polar NPA
Table 3B.
 
Correlation Between Nonperfusion Areas and Parameters for Renal Function Adjusted by Retinopathy Grade, Disease Duration, and Comorbid Hypertension: Posterior Polar NPA
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×