[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 34.204.0.181. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Table 1.  
Baseline Characteristics of Patients Classified by Diabetic Retinopathy
Baseline Characteristics of Patients Classified by Diabetic Retinopathy
Table 2.  
Baseline Characteristics of Patients Classified by Diabetic Macular Edema
Baseline Characteristics of Patients Classified by Diabetic Macular Edema
Table 3.  
Cox Regression Analysis for the New Onset of Any Diabetic Retinopathy
Cox Regression Analysis for the New Onset of Any Diabetic Retinopathy
Table 4.  
Cox Regression Analysis for the New Onset of Proliferative Diabetic Retinopathy
Cox Regression Analysis for the New Onset of Proliferative Diabetic Retinopathy
Table 5.  
Cox Regression Analysis for the New Onset of Diabetic Macular Edema
Cox Regression Analysis for the New Onset of Diabetic Macular Edema
1.
Kollias  AN, Ulbig  MW.  Diabetic retinopathy: early diagnosis and effective treatment.  Dtsch Arztebl Int. 2010;107(5):75-83.PubMedGoogle Scholar
2.
Bhagat  N, Grigorian  RA, Tutela  A, Zarbin  MA.  Diabetic macular edema: pathogenesis and treatment.  Surv Ophthalmol. 2009;54(1):1-32.PubMedGoogle ScholarCrossref
3.
Chen  H, Zheng  Z, Huang  Y,  et al.  A microalbuminuria threshold to predict the risk for the development of diabetic retinopathy in type 2 diabetes mellitus patients.  PLoS One. 2012;7(5):e36718.PubMedGoogle ScholarCrossref
4.
He  B-B, Wei  L, Gu  Y-J,  et al.  Factors associated with diabetic retinopathy in chinese patients with type 2 diabetes mellitus.  Int J Endocrinol. 2012;2012:157940.PubMedGoogle ScholarCrossref
5.
Stratton  IM, Kohner  EM, Aldington  SJ,  et al.  UKPDS 50: risk factors for incidence and progression of retinopathy in type II diabetes over 6 years from diagnosis.  Diabetologia. 2001;44(2):156-163.PubMedGoogle ScholarCrossref
6.
Wu  J, Geng  J, Liu  L, Teng  W, Liu  L, Chen  L. The relationship between estimated glomerular filtration rate and diabetic retinopathy.  J Ophthalmol. 2015;2015:326209. doi:10.1155/2015/326209Google Scholar
7.
Rodríguez-Poncelas  A, Mundet-Tudurí  X, Miravet-Jiménez  S,  et al.  Chronic kidney disease and diabetic retinopathy in patients with type 2 diabetes.  PLoS One. 2016;11(2):e0149448.PubMedGoogle ScholarCrossref
8.
Hammes  H-P, Welp  R, Kempe  H-P, Wagner  C, Siegel  E, Holl  RW; DPV Initiative—German BMBF Competence Network Diabetes Mellitus.  Risk factors for retinopathy and DME in type 2 diabetes—results from the German/Austrian DPV database.  PLoS One. 2015;10(7):e0132492.PubMedGoogle ScholarCrossref
9.
Man  RE, Sasongko  MB, Wang  JJ,  et al.  The association of estimated glomerular filtration rate with diabetic retinopathy and macular edema.  Invest Ophthalmol Vis Sci. 2015;56(8):4810-4816.PubMedGoogle ScholarCrossref
10.
Wilkinson  CP, Ferris  FL  III, Klein  RE,  et al; Global Diabetic Retinopathy Project Group.  Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales.  Ophthalmology. 2003;110(9):1677-1682.PubMedGoogle ScholarCrossref
11.
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(8):e0161897.PubMedGoogle ScholarCrossref
12.
Stevens  PE, Levin  A; Kidney Disease: Improving Global Outcomes Chronic Kidney Disease Guideline Development Work Group Members.  Evaluation and management of chronic kidney disease: synopsis of the kidney disease: improving global outcomes 2012 clinical practice guideline.  Ann Intern Med. 2013;158(11):825-830.PubMedGoogle ScholarCrossref
13.
Klein  R, Knudtson  MD, Lee  KE, Gangnon  R, Klein  BE.  The Wisconsin epidemiologic study of diabetic retinopathy: XXII the twenty-five-year progression of retinopathy in persons with type 1 diabetes.  Ophthalmology. 2008;115(11):1859-1868.PubMedGoogle ScholarCrossref
14.
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(9):1119-1125.PubMedGoogle ScholarCrossref
15.
Maharaj  AS, Saint-Geniez  M, Maldonado  AE, D’Amore  PA.  Vascular endothelial growth factor localization in the adult.  Am J Pathol. 2006;168(2):639-648.PubMedGoogle ScholarCrossref
16.
Schrijvers  BF, Flyvbjerg  A, De Vriese  AS.  The role of vascular endothelial growth factor (VEGF) in renal pathophysiology.  Kidney Int. 2004;65(6):2003-2017.PubMedGoogle ScholarCrossref
17.
Cha  DR, Kang  YS, Han  SY,  et al.  Vascular endothelial growth factor is increased during early stage of diabetic nephropathy in type II diabetic rats.  J Endocrinol. 2004;183(1):183-194.PubMedGoogle ScholarCrossref
18.
Pawlak  K, Mysliwiec  M, Pawlak  D.  Oxidative stress, phosphate and creatinine levels are independently associated with vascular endothelial growth factor levels in patients with chronic renal failure.  Cytokine. 2008;43(1):98-101.PubMedGoogle ScholarCrossref
19.
Ziyadeh  FN, Wolf  G.  Pathogenesis of the podocytopathy and proteinuria in diabetic glomerulopathy.  Curr Diabetes Rev. 2008;4(1):39-45.PubMedGoogle ScholarCrossref
20.
Sung  SH, Ziyadeh  FN, Wang  A, Pyagay  PE, Kanwar  YS, Chen  S.  Blockade of vascular endothelial growth factor signaling ameliorates diabetic albuminuria in mice.  J Am Soc Nephrol. 2006;17(11):3093-3104.PubMedGoogle ScholarCrossref
21.
Satchell  SC, Tooke  JE.  What is the mechanism of microalbuminuria in diabetes: a role for the glomerular endothelium?  Diabetologia. 2008;51(5):714-725.PubMedGoogle ScholarCrossref
22.
Knudsen  ST, Bek  T, Poulsen  PL, Hove  MN, Rehling  M, Mogensen  CE.  Macular edema reflects generalized vascular hyperpermeability in type 2 diabetic patients with retinopathy.  Diabetes Care. 2002;25(12):2328-2334.PubMedGoogle ScholarCrossref
23.
Klaassen  I, Van Noorden  CJ, Schlingemann  RO.  Molecular basis of the inner blood-retinal barrier and its breakdown in diabetic macular edema and other pathological conditions.  Prog Retin Eye Res. 2013;34:19-48.PubMedGoogle ScholarCrossref
24.
Levitt  DG, Levitt  MD.  Human serum albumin homeostasis: a new look at the roles of synthesis, catabolism, renal and gastrointestinal excretion, and the clinical value of serum albumin measurements.  Int J Gen Med. 2016;9:229-255.PubMedGoogle ScholarCrossref
25.
Wang  YT, Tadarati  M, Wolfson  Y, Bressler  SB, Bressler  NM.  Comparison of prevalence of diabetic macular edema based on monocular fundus photography vs optical coherence tomography.  JAMA Ophthalmol. 2016;134(2):222-228.PubMedGoogle ScholarCrossref
26.
Thomas  RL, Dunstan  FD, Luzio  SD,  et al.  Prevalence of diabetic retinopathy within a national diabetic retinopathy screening service.  Br J Ophthalmol. 2015;99(1):64-68.PubMedGoogle ScholarCrossref
Original Investigation
January 2018

Association of Abnormal Renal Profiles and Proliferative Diabetic Retinopathy and Diabetic Macular Edema in an Asian Population With Type 2 Diabetes

Author Affiliations
  • 1Department of Ophthalmology, National Taiwan University Hospital, Taipei, Taiwan
  • 2Department of Ophthalmology, Taipei Tzu Chi Hospital, Buddhist Tzu Chi Medical Foundation, New Taipei, Taiwan
  • 3Division of Endocrinology and Metabolism, Department of Internal Medicine, Changhua Christian Hospital, Changhua, Taiwan
  • 4Division of Endocrinology and Metabolism, Department of Internal Medicine, Yuanlin Christian Hospital, Changhua, Taiwan
  • 5Department of Molecular Biotechnology, Da-Yeh University, Changhua, Taiwan
JAMA Ophthalmol. 2018;136(1):68-74. doi:10.1001/jamaophthalmol.2017.5202
Key Points

Question  What is the association of chronic kidney disease with the development of diabetic retinopathy in patients with type 2 diabetes?

Findings  This 8-year cohort study of a Chinese population showed that an estimated glomerular filtration rate of less than 60 mL/min/1.73m2 and a urinary albumin to creatinine ratio of more than 30 mg/g at baseline were both associated with the development of proliferative diabetic retinopathy, while a baseline urinary albumin to creatinine ratio of more than 30 mg/g was associated with the development of diabetic macular edema. After adjusting the baseline values, abnormal mean follow-up renal profiles were still correlated with new-onset proliferative diabetic retinopathy.

Meaning  These data suggest that an abnormal renal profile is associated with proliferative diabetic retinopathy and diabetic macular edema in patients from this population with type 2 diabetes.

Abstract

Importance  The comorbidity of chronic kidney disease and diabetic retinopathy (DR) is well known. However, to our knowledge, no cohort study has demonstrated the effect of chronic kidney disease on the development or progression of DR.

Objective  To investigate the association of chronic kidney disease with the development of DR and diabetic macular edema (DME) in type 2 diabetes.

Design, Setting, and Participants  This 8-year prospective cohort study that was conducted in 2 medical centers in Taiwan included 2135 patients with type 2 diabetes.

Exposures  The baseline and mean follow-up renal profiles including serum creatinine level, estimated glomerular filtration rate (eGFR), and urinary albumin/creatinine ratio (ACR).

Main Outcomes and Measures  Diabetic retinopathy and DME were detected with nonmydriatic fundus photography. Cox regression analyses was used to evaluate the hazard ratios (HRs) for the renal profiles of new-onset DR, proliferative DR, and DME.

Results  The mean (SD) age of the study participants was 63.4 (11.9) years and 1205 (56%) were women. A higher serum creatinine level (HR of 2.358 for an increase of 1 mg/dL [to convert to micromoles per liter, multiply by 76.25]; 95% CI, 1.901-2.924; P < .001), an estimated glomerular filtration rate of less than 60 mL/min/1.73m2 (40-60: HR, 2.235; 95% CI, 1.351-4.035; P = .002; 30-45: HR, 2.625; 95% CI, 1.436-4.798; P = .002; <30: HR, 5.488; 95% CI, 2.739-10.993; P < .001), and a urinary albumin to creatinine ratio (ACR) of more than 30 mg/g (31-300: HR, 3.202; 95% CI, 2.029-5.053; P < .001; >300: HR, 6.652; 95% CI, 3.922-11.285; P < .001) at baseline were all associated with the development of proliferative DR. A baseline urinary ACR of more than 30 mg/g (31-300: HR, 1.563; 95% CI, 1.078-2.267; P = .02; >300: HR, 2.707; 95% CI, 1.640-4.470; -2.707; P < 0.001) was associated with the development of DME. After adjusting the baseline values, the mean follow-up renal profiles, including a higher serum creatinine level (HR, 2.369 per mg/dL; 95% CI, 1.704-3.293; P < .001), an estimated glomerular filtration rate of less than 30 mL/min/1.73m2 (HR, 4.215; 95% CI, 1.265-14.039; P = .02), and a urinary ACR of more than 30 mg/g (31-300: HR, 2.344; 95% CI, 1.200-4.503; P = .01; >300: HR, 4.193; 95% CI, 1.638-10.735; P = .003) were still correlated with new-onset PDR during the follow-up periods.

Conclusions and Relevance  Abnormal renal profiles at baseline, including a high serum creatinine level, low estimated glomerular filtration rate, and high urinary ACR, were associated with the development of PDR in patients with type 2 diabetes. A high baseline urinary ACR was associated with DME. Abnormal mean follow-up renal profiles were still correlated with new-onset PDR after adjusting for baseline values. Aggressive treatment for chronic kidney disease may have a role in preventing the deterioration of DR.

Introduction

Diabetic retinopathy (DR) has been a leading cause of blindness worldwide.1 Proliferative diabetic retinopathy (PDR), which may result in vitreous hemorrhage, tractional retinal detachment, and neovascular glaucoma, is one main cause of blindness in patients with diabetes. However, diabetic macular edema (DME) may also result in the visual loss because of the blood-retinal barrier breakdown, increased permeability of macular retinal vessels, and exudation of serous fluid and lipids into the macula.2 Diabetic macular edema could happen during any stage of DR, including nonproliferative diabetic retinopathy (NPDR) and PDR.

Chronic kidney disease (CKD) is also a main microvascular complication in patients with diabetes. It is characterized by increased serum creatinine levels, persistent albuminuria, and decreased glomerular filtration rates. Previous studies have shown that albuminuria and a low estimated glomerular filtration rate (eGFR) were both risk factors for DR.3-7 However, most previous studies were cross-sectional, not longitudinal. As for DME, only limited studies have reported its association with renal function. Macroalbuminuria, not eGFR, has been shown to be associated with DME.8,9 In this study, we conducted a prospective cohort in a Chinese population with type 2 diabetes to investigate the role of CKD in the development of DR and DME.

Methods

Patients who received a diagnosis of type 2 diabetes and underwent treatment in the outpatient clinic of the Metabolism Division at Changhua Christian Hospital and Kaohsiung Medical University Hospital between April 2002 and September 2004 were prospectively enrolled as a cohort. All patients obtained regular follow-up for systemic checkups every 2 to 6 months during the follow-up period. Those who were lost to follow-up within 6 months were excluded from the study. A total of 2198 patients were enrolled intially, and 37 patients were excluded because of loss of follow-up within 6 months. This cohort study ended on December 31, 2010. The human research ethics committees at both hospitals approved the protocol for this study (Changhua Christian Hospital institutional review board approval number, P201411-15; Kaohsiung Medical University Hospital institutional review board approval number, KUMH-IRB-95-00-74), and written informed consent was obtained from each patient. This study adhered to the tenets of the Declaration of Helsinki.

Outcome Measurements: DR and DME

All participants underwent fundus photography with a 45° nonmydriatic fundus camera (CanonCR-2; Canon) at baseline and every 6 to 12 months during the follow-up period. Two images were obtained for each eye: one that centered on the optic disc, and the other that centered on the fovea. The images were then graded by well-trained graders according to the international clinical diabetic retinopathy and diabetic macular edema disease severity scales as proposed by the Global Diabetic Retinopathy Project Group.10 The severity of DR in this study was categorized as no apparent retinopathy (no DR), NPDR, and PDR. As for the DME outcome in this study, we defined it as hard exudates in the presence of microaneurysms and blot hemorrhages within 1500 μm from the foveal center or the presence of focal photocoagulation scars in the macular area. For patients with gradable image results from both eyes, we recorded the eye with a more severe result; for those who had gradable image results available from only 1 eye, the results from the gradable eye were recorded. Patients with ungradable image results from both eyes at baseline were excluded from this study.

Renal Profile and Other Parameters

An interview and a comprehensive assessment of disease status, complications, and risk factors were performed for each patient on enrollment in this study. All patients had their blood samples obtained at 8 am after an overnight fast and before taking their morning medication; overnight first-void urine samples were also obtained. We measured renal profiles, including serum creatinine levels, eGFR, urinary albumin to creatinine ratios (ACR), and other parameters, including body mass index (calculated as weight in kilograms divided by height in meters squared), systolic blood pressure (SBP), diastolic blood pressure, hemaglobin A1c (HbA1c), fasting glucose level, total cholesterol level, high-density lipoprotein cholesterol level, low-density lipoprotein cholesterol level, and triglyceride level. All serum levels, except HbA1c, were measured using a biochemistry automatic analyzer (Bechman-Coulter Inc); HbA1c was measured in whole blood using ion-exchange high-performance liquid chromatography (BIo-Rad). Urinary ACR was calculated as the ratio of the urine albumin level over the creatinine level, and eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation.

All patients obtained regular follow-up every 2 to 6 months. At each follow-up, patients underwent the same tests as they had at baseline. The mean serum creatinine levels, eGFR, and urinary ACR values during the follow-up periods were computed for each patient.

Statistical Analysis

For patients with NPDR and those with PDR at baseline, the renal profile (serum creatinine, eGFR, and urinary ACR) and other parameters were compared with those without DR at baseline using t tests or χ2 tests. These parameters were also compared between patients with DME and those without DME at baseline. For those without DR/PDR/DME at baseline, Cox regression models were used to calculate the hazard ratios (HRs) for the baseline and mean follow-up renal profiles for any new-onset DR/PDR/DME. The mean follow-up values were calculated as the means of all examination results during the follow-up periods. The eGFR levels were divided into 5 groups for regression analysis: (1), more than 90 mL/min/1.73m2, (2) 61 to 90 mL/min/1.73m2, (3) 46 to 60 mL/min/1.73m2, (4) 30 to 45 mL/min/1.73m2, and (5) less than 30 mL/min/1.73m2. As for ACR, it was divided into 4 groups for regression analysis: (1) less than 10 mg/g, (2) 10 to 30 mg/g, (3) 31 to 300 mg/g, and (4) more than 300 mg/g. While analyzing the baseline renal profile, the following covariates were adjusted in the regression models using forward selection: age, sex, duration of diabetes, baseline body mass index, SBP, fasting glucose levels, HbA1c, total cholesterol levels, high-density lipoprotein cholesterol levels, low-density lipoprotein cholesterol levels, and triglyceride levels. While analyzing the mean follow-up renal profile, their baseline values were also adjusted in the regression models, along with the previously mentioned covariates. Diastolic blood pressure was not adjusted in the models because it was highly correlated with SBP. Because the triglyceride distribution was highly right-skewed, a natural log transformation was taken for regression analysis. Participants who were lost to follow-up due to death or any other reasons were treated as censored data. Log-minus-log plots and residual plots were used to examine the proportional hazards assumption and fitness of the Cox regression models. All statistical operations were performed using SAS, version 9.4 (SAS Institute).

Results

Of the 2161 patients in this cohort, 26 of them had ungradable fundus images. Therefore, a total of 2135 patients were included in this study. The mean (SD) age was 63.4 (11.9) years; 1205 (56%) were female, and 930 (44%) were male. The mean (SD) duration of diabetes at baseline was 15.1 (7.0) years. The mean (SD) follow-up period was 68.7 (23.4) months (26-101 months). At baseline, 751 patients (35.2%) had NPDR, 39 (1.8%) had PDR, and 34 (1.6%) had DME. Other baseline characteristics are shown in the eTable in the Supplement.

DR vs No DR at Baseline

Table 1 shows our comparisons regarding renal profiles and other baseline characteristics between patients with NPDR/PDR and those without DR. Both patients with NPDR and those with PDR had higher serum creatinine levels, lower eGFR, higher urinary ACR, and higher SBP than those without DR. Furthermore, patients with NPDR were older, had diabetes for a longer time, and had higher HbA1c levels than those without DR. Patients with PDR had lower high-density lipoprotein cholesterol levels than those without DR.

DME vs No DME at Baseline

Table 2 shows our comparisons regarding renal profiles and other baseline characteristics between patients with and without DME. Patients with DME had higher serum creatinine levels and lower eGFRs than those without DME.

New Onset of Any DR

Of the 1345 patients without DR at baseline, 685 patients (50.9%) developed DR during the follow-up period. No baseline serum creatinine level or baseline ACR level correlated with new-onset DR, but those with a baseline eGFR between 30 and 45 mL/min/1.73m2 had a higher risk of developing any DR, as compared with those with an eGFR of more than 90 mL/min/1.73m2 (HR, 1.465; 95% CI, 1.122-1.914; P = .01). For the mean follow-up renal profiles, serum creatinine was correlated with new-onset DR with an HR of 1.239 per mg/dL (95% CI, 1.012-2.477; P = .04). Compared with patients who had a mean follow-up eGFR of more than 90 mL/min/1.73m2, patients in the following eGFR groups were at higher risk of developing new-onset DR: those with mean follow-up eGFRs between 61 and 90 mL/min/1.73m2 (HR, 1.517; 95% CI, 1.118-2.059; P = .01), those with eGFRs between 46 and 60 mL/min/1.73m2 (HR, 1.782; 95% CI, 1.152-2.756; P = .01), those with eGFRs between 30 and 45 mL/min/1.73m2 (HR, 2.149; 95% CI, 1.250-3.697; P = .01), and those with an eGFRof less than 30 mL/min/1.73m2 (HR, 2.213; 95% CI, 1.036-4.726; P = .04). As for the mean follow-up ACR, it was not correlated with new-onset DR (Table 3).

New-Onset PDR

Of the 2096 patients without DR or with only NPDR at baseline, 179 patients (8.5%) developed PDR during the follow-up period. The Kaplan-Meier plots are shown in the eFigure in the Supplement. For the baseline renal profiles, serum creatinine levels (HR, 2.358 per mg/dL; 95% CI, 1.901-2.924; P < .001), eGFR, and urinary ACR were all assoicated with new-onset PDR. Compared with patients who had an eGFR of more than 90 mL/min/1.73m2, patients in the following eGFR groups were at higher risk of new-onset PDR: those with eGFRs between 46 and 60 mL/min/1.73m2 (HR, 2.335; 95% CI, 1.351-4.035; P = .002), those with eGFRs between 30 and 45 mL/min/1.73m2 (HR, 2.625; 95% CI, 1.436-4.798; P = .002), and those with eGFR of less than 30 mL/min/1.73m2 (HR, 5.488; 95% CI, 2.739-10.993; P < .001). Compared with patients who had an ACR of less than 10 mg/g, those with an ACR between 31 to 300 mg/g (HR, 3.202; 95% CI, 2.029-5.053; P < .001) and those with an ACR of more than 300 mg/g (HR, 6.652; 95% CI, 3.922-11.285; P < .001) both had a higher risk of developing PDR. As for the mean follow-up renal profiles, serum creatinine (HR, 2.369 per mg/dL [to convert to micromoles per liter, multiply by 76.25]; 95% CI, 1.704-3.293; P < .001), eGFR, and urinary ACR were still correlated new-onset PDR after adjusting the baseline values. Compared with patients who had an eGFR of more than 90 mL/min/1.73m2, those with an eGFR of less than 30 mL/min/1.73m2 were at a higher risk of developing new-onset PDR (HR, 4.215; 95% CI, 1.265-14.039; P = .02). Compared with patients who had an ACR of less than 10 mg/g, patients with an ACR between 31 and 300 mg/g (HR, 2.344; 95% CI, 1.200-4.503; P = .01) and those with an ACR of more than 300 mg/g (HR, 4.193; 95% CI, 1.638-10.735; P = .003) both had a higher risk of developing PDR (Table 4).

New-Onset DME

Of the 2101 patients without DME at baseline, 193 (9.2%) developed DME during the follow-up period. For the baseline renal profile, only urinary ACR was correlated with new-onset DME. Compared with patients with an ACR of less than 10 mg/g, patients with an ACR between 31 and 300 mg/g (HR, 1.563; 95% CI, 1.078-2.267; P = .02) and those with an ACR of more than 300 mg/g (HR, 2.707; 95% CI, 1.640-4.470; P < .001) both had a higher risk of developing DME. Neither baseline serum creatinine levels nor eGFR was correlated with new-onset DME. As for the mean follow-up renal profile, serum creatinine levels were correlated with new-onset DME with an HR of 1.402 per mg/dL (95% CI, 1.087-1.808; P = .01). For those with a mean follow-up eGFR between 30 and 45 mL/min/1.73m2, the HR for developing DME was 2.106, as compared with those with an eGFR of more than 90 mL/min/1.73m2 (95% CI, 1.268-7.609; P = .01). Mean follow-up ACRs were no longer correlated with new-onset DME after adjusting for the baseline ACR (Table 5).

Discussion

The association between CKD and DR has been extensively investigated in previous studies.3,7,8,11 The vascular endothelial dysfunction of small vessels secondary to prolonged hyperglycemia contributed to both the development and progression of DR and CKD. To our knowledge, our study was the first large-scale prospective cohort study that provided a more in-depth analysis of both eGFR and albuminuria, which are the 2 most important markers for assessing renal function, as suggested by the Kidney Disease: Improving Global Outcomes guideline.12 Similar to previous cross-sectional studies,7,9 we found that those with NPDR and PDR had a higher baseline serum creatinine level, higher baseline ACR, and a lower baseline eGFR level as compared with those without any type of DR. We also found that patients with lower eGFRs at baseline and during follow-up tended to have a progression of DR. Jeng et al11 also concluded that CKD was an independent risk factor for the development and progression of DR in a nationwide data-based cohort study. In addition to a lower eGFR, we also found that an ACR of more than 30 mg/g at baseline and during follow-up was correlated with new-onset PDR, but not any DR, suggesting that more severe proteinuria be associated with advanced DR. This agreed with previous results, indicating that albuminuria was strongly associated with the presence of PDR.13 On the other hand, one study demonstrated that a more extensive nonperfusion area in the retina increased the risk of progression of CKD over a 2-year period.14 These studies strengthened the link between diabetic retinopathy and nephropathy, suggesting that either one may progress along with the other.

Chronic kidney disease has been implicated in the production of vascular endothelial growth factors (VEGF) that are normally detected in podocytes in glomeruli and tubular epithelial cells.15,16 A marked expression of VEGF secondary to glomerular injury was observed in experimental diabetic rats with early nephropathy.17 Local production of VEGF may be further released into the systemic circulation. Elevated serum VEGF levels were also found in patients with CKD, and it had an inverse relationship with eGFR levels.18 This could be the possible mechanism explaining why CKD with a lower eGFR was correlated with a new onset of PDR. Moreover, after the baseline eGFR value was adjusted, patients with higher mean follow-up eGFRs still had a higher risk for developing now-onset DR or PDR. This implicates that treatment for CKD may have a role in preventing the progression of DR.

Vascular endothelial growth factors were also a possible mediator in the occurrence of albuminuria, causing increased vascular permeability and protein filtration.19,20 Vascular endothelial growth factors and other inflammatory cytokines damage glomerular endothelial cells and disturb podocyte-endothelial cell communication, leading to microalbuminuria.21 While microalbuminuria was an early manifestation of nephropathy, the presence of macroalbuminuria may indicate progressive CKD. The gross destruction of glycocalyx, which acted as a glomerular filtration barrier, may have caused additional podocyte loss and mesangial expansion that eventually led to irreversible glomerulosclerosis and overt proteinuria.21 Therefore, a high VEGF level may result in both overt albuminuria and PDR, and this support our findings that both high initial and follow-up ACR values were associated with new PDR.

While both eGFR and the presence of albuminuria indicated the severity of CKD and were associated with DR, they may not have the same role in explaining the progression of DME occurrence, as few patients with DR developed DME over the course of follow-up in this study, accounting for 9.2% of patients without DME at baseline in our study. In this study, patients with DME at baseline also tended to have a low eGFR at baseline. On the other hand, patients with baseline ACRs of more than 30 mg/g were more likely to develop DME over the follow-up period, in accordance with a previous study that demonstrated that both microalbuminuriaand macroalbuminuria were risk factors for DME, with macroalbuminuria carrying a stronger risk.8 The mechanisms for DME may provide some insights on what we observed. First, the breakdown of the blood-retinal barrier caused by VEGF and other cytokines resulted in capillary leakage and fluid accumulation in the retinal parenchyma. Diabetic macular edema reflected a process of a more generalized vascular hyperpermeability and leakage.22 We inferred that a lower eGFR and DME may both result from an elevated VEGF and therefore are correlated with each other. Second, advanced macroalbuminuria with marked protein loss may result in lower oncotic pressure that drives fluid leakage into the extravascular space according to Starling’s23 rule that states that fluid movement is governed by changes in pressure gradients. While serum albumin deficits can be compensated by an increased production of albumin molecules in the liver and may not lead to hypoalbuminemia, sustained and massive protein loss may impair this homeostasis in the long run.24 According to this deduction, patients with macroalbuminuria may be prone to more fluid leakage from the damaged retinal vessels, which results in DME eventually.

Strengths and Limitations

The strength of this study is that it was an 8-year prospective cohort study. We analyzed both eGFR and albuminuria for CKD not only at baseline but also during the follow-up period as dynamic markers for retinal disease progression. We found that eGFR and albuminuria, while probably intercorrelated, may not contribute equally to the progression of DR and DME. To ou knowledge, this distinction was not specifically addressed in previous studies. This study also had some limitations. The diagnosis of DME was based on fundus photography only, which is less sensitive than optical coherence tomography.25 However, most previous studies also adopted the same diagnostic tools. As for the diagnosis of DR, the 2-field nonmydriatic fundus photography that was used in this study has been widely used in previous studies and proven to be valid.9,26 Furthermore, this is an observational study. Although we tried to adjust for possible confounding factors in multiple regression models, there were still residual confounding factors that could not be controlled. This study could reveal associations but not causations. Further randomized clinical trials are necessary for investigating the causal relationships between CKD and DR. Finally, all the participants in this study were of the Chinese population in Taiwan. We cannot be sure if such results can be generalized to other populations.

Conclusions

Abnormal renal profiles, including a high serum creatinine level, low eGFR, and high urinary ACR, both at baseline and during the follow-up period, were associated with a new onset of PDR. A high ACR at baseline was associated with the development of DME. This connection implies that aggressive treatment for CKD may have a role in preventing the deterioration of DR.

Back to top
Article Information

Accepted for Publication: October 5, 2017.

Corresponding Author: Ming-Chia Hsieh, Division of Endocrinology and Metabolism, Department of Internal Medicine, Changhua Christian Hospital, Changhua, Taiwan, 135 Nanhsiao Street. Changhua City, Taiwan 50006 (mingchia570531@gmail.com).

Correction: This article was corrected online December 20, 2018, for data errors in the Abstract and Results section.

Published Online: November 22, 2017. doi:10.1001/jamaophthalmol.2017.5202

Author Contributions: Dr Hsieh had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Y. Hsieh, Tu, M. Hsieh.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Y. Hsieh, Tsai.

Critical revision of the manuscript for important intellectual content: Tu, M. Hsieh.

Statistical analysis: Y. Hsieh.

Administrative, technical, or material support: Y. Hsieh, M. Hsieh.

Supervision: Tu, M. Hsieh.

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

References
1.
Kollias  AN, Ulbig  MW.  Diabetic retinopathy: early diagnosis and effective treatment.  Dtsch Arztebl Int. 2010;107(5):75-83.PubMedGoogle Scholar
2.
Bhagat  N, Grigorian  RA, Tutela  A, Zarbin  MA.  Diabetic macular edema: pathogenesis and treatment.  Surv Ophthalmol. 2009;54(1):1-32.PubMedGoogle ScholarCrossref
3.
Chen  H, Zheng  Z, Huang  Y,  et al.  A microalbuminuria threshold to predict the risk for the development of diabetic retinopathy in type 2 diabetes mellitus patients.  PLoS One. 2012;7(5):e36718.PubMedGoogle ScholarCrossref
4.
He  B-B, Wei  L, Gu  Y-J,  et al.  Factors associated with diabetic retinopathy in chinese patients with type 2 diabetes mellitus.  Int J Endocrinol. 2012;2012:157940.PubMedGoogle ScholarCrossref
5.
Stratton  IM, Kohner  EM, Aldington  SJ,  et al.  UKPDS 50: risk factors for incidence and progression of retinopathy in type II diabetes over 6 years from diagnosis.  Diabetologia. 2001;44(2):156-163.PubMedGoogle ScholarCrossref
6.
Wu  J, Geng  J, Liu  L, Teng  W, Liu  L, Chen  L. The relationship between estimated glomerular filtration rate and diabetic retinopathy.  J Ophthalmol. 2015;2015:326209. doi:10.1155/2015/326209Google Scholar
7.
Rodríguez-Poncelas  A, Mundet-Tudurí  X, Miravet-Jiménez  S,  et al.  Chronic kidney disease and diabetic retinopathy in patients with type 2 diabetes.  PLoS One. 2016;11(2):e0149448.PubMedGoogle ScholarCrossref
8.
Hammes  H-P, Welp  R, Kempe  H-P, Wagner  C, Siegel  E, Holl  RW; DPV Initiative—German BMBF Competence Network Diabetes Mellitus.  Risk factors for retinopathy and DME in type 2 diabetes—results from the German/Austrian DPV database.  PLoS One. 2015;10(7):e0132492.PubMedGoogle ScholarCrossref
9.
Man  RE, Sasongko  MB, Wang  JJ,  et al.  The association of estimated glomerular filtration rate with diabetic retinopathy and macular edema.  Invest Ophthalmol Vis Sci. 2015;56(8):4810-4816.PubMedGoogle ScholarCrossref
10.
Wilkinson  CP, Ferris  FL  III, Klein  RE,  et al; Global Diabetic Retinopathy Project Group.  Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales.  Ophthalmology. 2003;110(9):1677-1682.PubMedGoogle ScholarCrossref
11.
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(8):e0161897.PubMedGoogle ScholarCrossref
12.
Stevens  PE, Levin  A; Kidney Disease: Improving Global Outcomes Chronic Kidney Disease Guideline Development Work Group Members.  Evaluation and management of chronic kidney disease: synopsis of the kidney disease: improving global outcomes 2012 clinical practice guideline.  Ann Intern Med. 2013;158(11):825-830.PubMedGoogle ScholarCrossref
13.
Klein  R, Knudtson  MD, Lee  KE, Gangnon  R, Klein  BE.  The Wisconsin epidemiologic study of diabetic retinopathy: XXII the twenty-five-year progression of retinopathy in persons with type 1 diabetes.  Ophthalmology. 2008;115(11):1859-1868.PubMedGoogle ScholarCrossref
14.
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(9):1119-1125.PubMedGoogle ScholarCrossref
15.
Maharaj  AS, Saint-Geniez  M, Maldonado  AE, D’Amore  PA.  Vascular endothelial growth factor localization in the adult.  Am J Pathol. 2006;168(2):639-648.PubMedGoogle ScholarCrossref
16.
Schrijvers  BF, Flyvbjerg  A, De Vriese  AS.  The role of vascular endothelial growth factor (VEGF) in renal pathophysiology.  Kidney Int. 2004;65(6):2003-2017.PubMedGoogle ScholarCrossref
17.
Cha  DR, Kang  YS, Han  SY,  et al.  Vascular endothelial growth factor is increased during early stage of diabetic nephropathy in type II diabetic rats.  J Endocrinol. 2004;183(1):183-194.PubMedGoogle ScholarCrossref
18.
Pawlak  K, Mysliwiec  M, Pawlak  D.  Oxidative stress, phosphate and creatinine levels are independently associated with vascular endothelial growth factor levels in patients with chronic renal failure.  Cytokine. 2008;43(1):98-101.PubMedGoogle ScholarCrossref
19.
Ziyadeh  FN, Wolf  G.  Pathogenesis of the podocytopathy and proteinuria in diabetic glomerulopathy.  Curr Diabetes Rev. 2008;4(1):39-45.PubMedGoogle ScholarCrossref
20.
Sung  SH, Ziyadeh  FN, Wang  A, Pyagay  PE, Kanwar  YS, Chen  S.  Blockade of vascular endothelial growth factor signaling ameliorates diabetic albuminuria in mice.  J Am Soc Nephrol. 2006;17(11):3093-3104.PubMedGoogle ScholarCrossref
21.
Satchell  SC, Tooke  JE.  What is the mechanism of microalbuminuria in diabetes: a role for the glomerular endothelium?  Diabetologia. 2008;51(5):714-725.PubMedGoogle ScholarCrossref
22.
Knudsen  ST, Bek  T, Poulsen  PL, Hove  MN, Rehling  M, Mogensen  CE.  Macular edema reflects generalized vascular hyperpermeability in type 2 diabetic patients with retinopathy.  Diabetes Care. 2002;25(12):2328-2334.PubMedGoogle ScholarCrossref
23.
Klaassen  I, Van Noorden  CJ, Schlingemann  RO.  Molecular basis of the inner blood-retinal barrier and its breakdown in diabetic macular edema and other pathological conditions.  Prog Retin Eye Res. 2013;34:19-48.PubMedGoogle ScholarCrossref
24.
Levitt  DG, Levitt  MD.  Human serum albumin homeostasis: a new look at the roles of synthesis, catabolism, renal and gastrointestinal excretion, and the clinical value of serum albumin measurements.  Int J Gen Med. 2016;9:229-255.PubMedGoogle ScholarCrossref
25.
Wang  YT, Tadarati  M, Wolfson  Y, Bressler  SB, Bressler  NM.  Comparison of prevalence of diabetic macular edema based on monocular fundus photography vs optical coherence tomography.  JAMA Ophthalmol. 2016;134(2):222-228.PubMedGoogle ScholarCrossref
26.
Thomas  RL, Dunstan  FD, Luzio  SD,  et al.  Prevalence of diabetic retinopathy within a national diabetic retinopathy screening service.  Br J Ophthalmol. 2015;99(1):64-68.PubMedGoogle ScholarCrossref
×