Abstract
Background. Asymmetric dimethylarginine (ADMA) is associated with macrovascular disease and possibly with microangiopathy in type 2 diabetes (T2DM). We tested the hypothesis that ADMA is related to diabetic retinopathy (DR) independently of macrovascular disease. Methods. This cross-sectional study included 127 T2DM patients selected to achieve equal distributions of patients with and without macrovascular disease in the groups with and without DR. Results. Patients with DR had increased ADMA, longer diabetes duration, and reduced glomerular filtration rate (GFR). ADMA correlated with GFR (ρ = -0.35; P < .001), diabetes duration (ρ = 0.19; P = .048), and age (ρ = 0.19; P = .033). Logistic regression analysis revealed an association of ADMA with DR. After adjustment for macrovascular disease, this association remained significant (OR 1.48; 95% CI: 1.02–2.15; P = .039). Inclusion of GFR and T2DM duration into the model abolished this significant relationship. GFR remained the only independent predictor for DR. A 10 mL/min/1.73 m2 GFR decrease was associated with DR in a multivariate model (OR 1.30; 95% CI: 1.08–1.56; P = .006). Conclusions. These findings indicate an association between ADMA and DR in T2DM independent of macrovascular disease. This relationship is modified by GFR, the only parameter significantly related to DR in multivariate analysis.
1. Introduction
Hyperglycemia is closely linked to the risk of microvascular complications in diabetes [1]. Patients suffering from type 2 diabetes (T2DM) and diabetic retinopathy are at increased risk for developing cardiovascular disease [1–3]. Asymmetric dimethylarginine (ADMA), an endogenous competitive nitric oxide synthase inhibitor, is an emerging risk marker for future cardiovascular events in different patient groups including patients with T2DM [4]. In addition to its relation to macrovascular complications, ADMA has also been linked to the development of diabetic microvascular complications. ADMA is elevated in patients with type 1 and type 2 diabetes with diabetic nephropathy and predicts renal impairment in these patients [5–7]. Malecki et al. found increased ADMA concentrations in subjects with T2DM and diabetic retinopathy [8]. However, it remained unclear if macrovascular disease which was associated with significantly elevated circulating ADMA in previous reports influenced this relationship. The current study was planned to test the hypothesis that ADMA is associated with diabetic retinopathy independently of macrovascular disease and other possible confounders.
2. Materials and Methods
This study was conducted in accordance with the guidelines of the Declaration of Helsinki and the institutional ethics committee. It was designed as a cross-sectional study.
One hundred twenty seven patients with T2DM who referred to a specialty outpatient diabetic clinic between September 2003 and March 2004 were included. All patients were suffering from T2DM according to the American Diabetes Association criteria and were on stable antidiabetic treatment for at least 6 months. All patients were treated according to the Guidelines of the American Diabetes Association. At the time of inclusion, demographic and anthropometric data, clinical characteristics, and current medication of all patients were recorded with special attention to present cardiovascular risk factors and co-morbidities. Patients were selected with regard to prevalent retinopathy and macrovascular disease to reach a proportion of around 50% patients with retinopathy and 50% patients with macrovascular disease equally distributed between the retinopathy and the no retinopathy group.
Demographic data and medical history including age, sex, smoking habits, hyperlipidemia, arterial hypertension, coronary artery disease, history of myocardial infarction, and stroke were assessed in all patients. The presence of macrovascular disease was confirmed by hospital records. The study included 60 patients with and 67 patients without diabetic retinopathy. Macrovascular disease was present in 38 patients with and in 30 patients without diabetic retinopathy. Body mass index (BMI) was calculated as weight divided by squared height. Blood pressure was determined in all subjects at rest. Arterial hypertension was diagnosed if resting blood pressure values were >140/90 mmHg or patients were taking antihypertensive drugs. All patients were staged for the presence of diabetic retinopathy according to the Early Treatment of Diabetic Retinopathy Study (ETDRS) classification [9–11]. Patients who had retinal abnormalities other than diabetic retinopathy or who underwent panretinal laser treatment, vitrectomy, or anti-VEGF therapy in the past 12 months were excluded. Before retinal photography, the patient's pupils were dilated with tropicamide 1%; this procedure was repeated if the pupils did not reach at least 5 mm in diameter. Color retinal slides with a suitable 40° retinal camera were taken of two fields in both eyes. For the central field, the center of the optic disc was positioned at the nasal end of the horizontal meridian of the field of view; for the nasal field, the optic disc was positioned 1 disc diameter from the temporal edge of the field on the horizontal meridian. Two pseudostereoscopic images with a horizontal shift of 3° to 5° were then taken of each field in both eyes, resulting in eight images of each patient. The overlap of the fields recorded a retinal view of approximately 75° horizontally by 40° vertically; therefore, clinically significant lesions of diabetic retinopathy were easily detectable. Slides were separately screened by two experienced graders in the same center, according to a validated protocol [12].
2.1. Laboratory Investigations
Venous blood was drawn after an overnight fast for the determination of ADMA, SDMA, L-arginine, and routine parameters. Blood levels of creatinine, HbA1c, total cholesterol, HDL-cholesterol, LDL-cholesterol, triglycerides, and high-sensitivity C-reactive protein (CRP) were measured by standard laboratory methods in a certified laboratory. Glomerular filtration rate was calculated using the Modification of Diet in Renal Disease Study Group formula [13]. For measurement of ADMA, SDMA, and L-arginine, plasma was subjected to cation exchange solid-phase extraction and analyzed by high-performance liquid chromatography [14, 15]. The coefficients of variation for inter- and intra-assay variations tested with a pooled plasma sample were <3% for all analytes. The detection limit for dimethylarginines was 0.04 μmol/L.
2.2. Statistics
Continuous data are presented as medians (interquartile range). Categorical data are given as counts (%). Mann-Whitney U test was applied for univariate comparison of continuous data and Spearman rank correlation for assessment of associations between continuous variables. Skewed variables were log-transformed for regression analyses. Univariate and multivariate logistic regression analysis including diabetic retinopathy as dependent variable and log-transformed ADMA as well as possible confounders as independent variables was used. To assess the effect of ADMA on the prevalence of diabetic retinopathy, odds ratios (OR) and 95% confidence intervals (95% CI) for a 1 standard deviation (SD) increment in log-transformed ADMA were calculated. Variables associated with ADMA and diabetic retinopathy including the presence of macrovascular disease were treated as possible confounders in the regression analysis. Age and sex were excluded from the logistic regression analysis that included GFR, because these variables are already used in the MDRD formula. A two-sided P < .05 was considered as statistically significant. Calculations were performed with SPSS for Windows (Version 16.0; SPSS, Chicago, Ill).
3. Results
The characteristics of the study population are presented in Table 1. Importantly, the proportion of patients with and without macrovascular disease was comparable in patients with and without diabetic retinopathy. Patients with diabetic retinopathy had significantly higher levels of ADMA and SDMA compared to those without. L-arginine was comparable between the groups. In addition, these patients had longer diabetes duration, lower GFR, and higher HDL-cholesterol. They were less likely to be smokers and tended to be older than patients without retinopathy (Table 1).
Table 1.
Characteristics, L-Arginine, and asymmetric and symmetric dimethylarginine according to the prevalence of diabetic retinopathy.
| No retinopathy | Retinopathy | P value | |
|---|---|---|---|
| N | 60 | 67 | |
| Female | 19 (31.7%) | 30 (44.8%) | .130 |
| Proliferative retinopathy | 32 (47.8%) | ||
| Macrovascular Disease | 30 (50.0%) | 38 (56.7%) | .449 |
| Smokers | 25 (42.4%) | 10 (19.2%) | .009 |
| Age (years) | 62 (54–69) | 64 (56–73) | .051 |
| Diabetes duration (years) | 9 (4–16) | 14 (10–17) | .027 |
| HbA1c (%) | 8.0 (7.0–9.0) | 7.4 (7.0–8.1) | .127 |
| Body mass index (kg/m2) | 29.2 (26.4–32.3) | 29.7 (25.8–34.2) | .529 |
| Systolic blood pressure (mmHg) | 141 (134–157) | 144 (129–163) | .891 |
| Diastolic blood pressure (mmHg) | 83 (77–90) | 80 (73–91) | .164 |
| Total cholesterol (mg/dL) | 188 (157–213) | 193 (164–218) | .686 |
| LDL cholesterol (mg/dL) | 94 (74–129) | 90 (69–116) | .300 |
| HDL cholesterol (mg/dL) | 44 (36–50) | 50 (43–56) | .020 |
| Triglycerides (mg/dL) | 175 (115–265) | 194 (147–284) | .276 |
| Glomerular filtration rate (mL/min/1.73 m2) | 79 (59–100) | 59 (43–78) | <.001 |
| L-Arginine (μmol/L) | 74 (62–94) | 82 (63–119) | .359 |
| Asymmetric dimethylarginine (μmol/L) | 0.55 (0.48–0.67) | 0.62 (0.53–0.96) | .011 |
| Symmetric dimethylarginine (μmol/L) | 0.49 (0.43–0.63) | 0.61 (0.48–0.92) | .005 |
Categorical variables are presented as counts (%) and continuous variables are presented as medians (interquartile range); the χ2-test and Mann-Whitney U test have been applied for comparisons between the groups.
ADMA did not differ significantly between women and men (0.61 (0.52–0.93) μmol/L versus 0.59 (0.50–0.68) μmol/L; P = .207) as well as smokers and nonsmokers (0.58 (IQR: 0.50–0.76) μmol/L versus 0.58 (0.49–0.69) μmol/L; P = .568). ADMA was significantly related to SDMA, L-arginine, age, diabetes duration, and glomerular filtration rate in the whole study cohort. No association between ADMA and other clinical parameters could be observed (Table 2).
Table 2.
Spearman correlations of ADMA with other variables.
| ρ | P value | |
|---|---|---|
| Age | 0.189 | .033 |
| Diabetes duration | 0.190 | .048 |
| HbA1c | −0.079 | .412 |
| Body mass index | 0.016 | .870 |
| Systolic blood pressure | −0.133 | .172 |
| Diastolic blood pressure | −0.183 | .060 |
| Total cholesterol | −0.068 | .539 |
| LDL cholesterol | −0.040 | .727 |
| HDL cholesterol | 0.024 | .830 |
| Triglycerides | −0.141 | .203 |
| Glomerular filtration rate | −0.347 | <.001 |
| L-Arginine | 0.548 | <.001 |
| Symmetric dimethylarginine | 0.633 | <.001 |
The univariate logistic regression analysis showed that ADMA was significantly associated with the prevalence of diabetic retinopathy before and after adjusting for macrovascular disease. However, including age and sex into the multivariate model raised the P-level of this association above the level of significance. Multivariate adjustment for prevalence of macrovascular disease, diabetes duration, and GFR completely abolished the association between ADMA and diabetic retinopathy (Table 3).
Table 3.
Univariate and multivariate logistic regression analysis including diabetic retinopathy as dependent variable and asymmetric dimethylarginine (ADMA) and possible confounders as independent variables.
| Odds ratio (95% CI) | P value | ||
|---|---|---|---|
| Univariate model | ADMA (per 1 SD log increase) | 1.50 (1.03–2.17) | .034 |
| Multivariate model 1 | ADMA (per 1 SD log increase) | 1.48 (1.02–2.15) | .039 |
| MVD versus no MVD | 1.22 (0.60–2.50) | .581 | |
| Multivariate model 2 | ADMA (per 1 SD log increase) | 1.36 (0.93–2.00) | .113 |
| Age | 1.04 (1.00–1.07) | .078 | |
| Sex (male versus female) | 0.70 (0.33–1.52) | .370 | |
| MVD versus no MVD | 1.27 (0.61–2.64) | .522 | |
| Multivariate model 3 | ADMA (per 1 SD log increase) | 1.17 (0.73–1.90) | .518 |
| DM2 duration (per 1 SD log increase) | 1.27 (0.80–2.01) | .306 | |
| GFR (per 10 mL/min/1.73 m2 decrease) | 1.30 (1.08–1.56) | .006 | |
| MVD versus no MVD | 0.88 (0.37–2.10) | .774 | |
GFR: glomerular filtration rate calculated with the MDRD formula; 95% CI: 95% confidence interval; DM2: diabetes mellitus type 2; MVD: macrovascular disease.
Because of the significant independent association between GFR and diabetic retinopathy, a secondary analysis including GFR as dependent variable was performed. GFR was significantly related to SDMA, L-arginine, diabetes duration, and HbA1c (Table 4). GFR was comparable between women and men (60 (52–86) mL/min/1.73 m2 versus 72 (52–93) mL/min/1.73 m2; P = .255) as well as smokers and nonsmokers (72 (43–72) mL/min/1.73 m2 versus 67 (53–90) mL/min/1.73 m2; P = .697). Both the univariate and multivariate logistic regression models showed that GFR was the only parameter independently related to the prevalence of diabetic retinopathy (Table 5).
Table 4.
Spearman correlations of GFR with other variables.
| ρ | P value | |
|---|---|---|
| Diabetes duration | −0.266 | .006 |
| HbA1c | 0.307 | .001 |
| Body mass index | 0.085 | .402 |
| Systolic blood pressure | −0.017 | .862 |
| Diastolic blood pressure | 0.247 | .011 |
| Total cholesterol | 0.077 | .492 |
| LDL cholesterol | 0.128 | .268 |
| HDL cholesterol | −0.136 | .233 |
| Triglycerides | 0.060 | .590 |
| L-Arginine | −0.252 | .005 |
| Symmetric dimethylarginine | −0.684 | <.001 |
Table 5.
Univariate and multivariate logistic regression analysis including diabetic retinopathy as dependent variable and glomerular filtration rate (GFR) and possible confounders as independent variables.
| Odds ratio (95% CI) | P value | ||
|---|---|---|---|
| Univariate model | GFR (10 mL/min/1.73 m2 decrease) | 1.32 (1.13–1.54) | .001 |
| Multivariate model | GFR (10 mL/min/1.73 m2 decrease) | 1.30 (1.08–1.56) | .006 |
| ADMA (per 1 SD log increase) | 1.17 (0.73–1.90) | .518 | |
| DM2 duration (per 1 SD log increase) | 1.27 (0.80–2.01) | .306 | |
| MVD versus no MVD | 0.88 (0.37–2.10) | .881 | |
GFR: glomerular filtration rate calculated with the MDRD formula; 95% CI: 95% confidence interval; DM2: diabetes mellitus type 2.
4. Discussion
The results of the present study show that circulating ADMA is related to diabetic retinopathy independently of macrovascular disease. These findings do not confirm the hypothesis that this association is independent of other possible confounders. Interestingly, GFR was shown to be the only independent predictor for the prevalence of diabetic retinopathy in this cross-sectional analysis.
The finding that ADMA is not independently related to diabetic retinopathy in patients with T2DM was surprising regarding the recently published data by Malecki et al. [8]. Nevertheless this study missed to evaluate the prevalence of macrovascular disease which would be important. This shortfall could introduce bias, as ADMA is known to be related to macrovascular disease [4]. Macrovascular disease itself has been shown to be related to diabetic retinopathy as shown by Reaven et al. [16]. For this reason, the present investigation included patients with and without retinopathy who were carefully selected with regard to the prevalence of macrovascular disease. Nearly equal proportions of patients had macrovascular disease in the groups with and without retinopathy. Macrovascular disease was not a major confounder and did not abolish the significant association between ADMA and retinopathy in the multivariate regression analysis. These results differ from those reported by Malecki et al. due to the effects of sex, age, and GFR that eliminated the significant relationship between retinopathy and ADMA. GFR was the best independent predictor for the prevalence of retinopathy in our study. Interestingly, calculated creatinine clearance did not differ between patients with and without retinopathy in the study by Malecki et al. which included diabetic subjects with better renal function [8]. Only 22 patients had GFR > 90 mL/min/1.73 m2 in the present study. It is possible that ADMA is independently related to retinopathy only in patients with normal renal function. The notion that moderate renal failure leads to a significant ADMA increase [17] could at least partly explain the discrepancies between the reported findings. In addition, our study included 32 patients with T2DM and proliferative retinopathy, which reflects more prevalence of serious microvascular disease compared with the study population investigated by Malecki et al., which included only 3 patients with proliferative retinopathy [8]. Proliferative diabetic retinopathy and impaired renal function are often interrelated in patients with T2DM [18, 19]. These findings and the fact that both study cohorts differed substantially regarding the included number of patients with proliferative retinopathy provide another potential explanation for the discrepant results between the studies. In a large number of patients with type 1 diabetes, Tarnow et al. [5] did not observe a relationship between ADMA and simplex or proliferative diabetic retinopathy, findings which are in accordance with our results obtained in patients with T2DM.
Nevertheless, the divergent findings on circulating ADMA and diabetic retinopathy do not exclude a pathophysiologic role of ADMA for the development of diabetic retinopathy. As published by Sugai et al., evidence exists that ADMA concentrations in aqueous humor are associated with severe diabetic retinopathy. Nevertheless, it was also mentioned in this report that ADMA concentrations found in aqueous humor are not related to circulating ADMA [20]. This could be another explanation why circulating ADMA was not associated with diabetic retinopathy. Intraocular but not circulating ADMA might be of pathophysiologic relevance for diabetic retinopathy.
From the present data, it could be assumed that renal dysfunction, which is often found in patients with T2DM, is an important factor associated with diabetic retinopathy. This is in line with findings from previous studies, which showed a significant and independent association between retinopathy and declining renal function in T2DM. Moreover, it was shown that diabetic retinopathy can predict the deterioration of renal function [21, 22]. This suggests that reduced GFR rather appears after than before the development of retinopathy. As GFR significantly influences circulating ADMA, it is likely that elevated ADMA in patients with diabetic retinopathy appeared after the emergence of diabetic retinopathy in our study cohort.
The present investigation is limited by the cross-sectional study design which does not allow statements on causality. The strength of the present study design is the high prevalence of macrovascular disease in patients with and without retinopathy. This allowed the evaluation of the effect of macrovascular disease on the relationship between ADMA and diabetic retinopathy. Although no independent relationship between ADMA and the prevalence of diabetic retinopathy could be detected in this study, a pathophysiological link between ADMA, impaired renal function, and diabetic retinopathy cannot be excluded. Further studies are needed to evaluate the influence of ADMA on microvascular complication.
5. Conclusion
In conclusion, this study demonstrated that ADMA is associated with retinopathy in patients with T2DM independent of macrovascular disease. This effect is substantially modified by adjusting for GFR which remained the single independent parameter related to diabetic retinopathy in a logistic regression analysis. Our findings cannot support the hypothesis that circulating ADMA is independently associated with diabetic retinopathy but do not exclude that ADMA lies in the causative path between renal dysfunction and diabetic retinal disease.
Conflict of Interests
The authors declared that there is no conflict of interests.
References
- 1.Van Hecke MV, Dekker JM, Stehouwer CDA, et al. Diabetic retinopathy is associated with mortality and cardiovascular disease incidence: the EURODIAB prospective complications study. Diabetes Care. 2005;28(6):1383–1389. doi: 10.2337/diacare.28.6.1383. [DOI] [PubMed] [Google Scholar]
- 2.Cheung N, Wang JJ, Klein R, Couper DJ, Sharrett AR, Wong TY. Diabetic retinopathy and the risk of coronary heart disease: the Atherosclerosis Risk in Communities Study. Diabetes Care. 2007;30(7):1742–1746. doi: 10.2337/dc07-0264. [DOI] [PubMed] [Google Scholar]
- 3.Hiller R, Sperduto RD, Podgor MJ, Ferris FL, III, Wilson PWF. Diabetic retinopathy and cardiovascular disease in type II diabetics. The framingham heart study and the framingham eye study. American Journal of Epidemiology. 1988;128(2):402–409. doi: 10.1093/oxfordjournals.aje.a114980. [DOI] [PubMed] [Google Scholar]
- 4.Krzyzanowska K, Mittermayer F, Wolzt M, Schernthaner G. Asymmetric dimethylarginine predicts cardiovascular events in patients with type 2 diabetes. Diabetes Care. 2007;30(7):1834–1839. doi: 10.2337/dc07-0019. [DOI] [PubMed] [Google Scholar]
- 5.Tarnow L, Hovind P, Teerlink T, Stehouwer CDA, Parving HH. Elevated pasma asymmetric dimethylarginine as a marker of cardiovascular morbidity in early diabetic nephropathy in type 1 diabetes. Diabetes Care. 2004;27(3):765–769. doi: 10.2337/diacare.27.3.765. [DOI] [PubMed] [Google Scholar]
- 6.Hanai K, Babazono T, Nyumura I, et al. Asymmetric dimethylarginine is closely associated with the development and progression of nephropathy in patients with type 2 diabetes. Nephrology Dialysis Transplantation. 2009;24(6):1884–1888. doi: 10.1093/ndt/gfn716. [DOI] [PubMed] [Google Scholar]
- 7.Krzyzanowska K, Mittermayer F, Shnawa N, et al. Asymmetrical dimethylarginine is related to renal function, chronic inflammation and macroangiopathy in patients with type 2 diabetes and albuminuria. Diabetic Medicine. 2007;24(1):81–86. doi: 10.1111/j.1464-5491.2007.02018.x. [DOI] [PubMed] [Google Scholar]
- 8.Malecki MT, Undas A, Cyganek K, et al. Plasma asymmetric dimethylarginine (ADMA) is associated with retinopathy in type 2 diabetes. Diabetes Care. 2007;30(11):2899–2901. doi: 10.2337/dc07-1138. [DOI] [PubMed] [Google Scholar]
- 9.Early treatment diabeticretinopathy study design and baseline patient characteristics. ETDRS report number 7. Ophthalmology . 1991;98(5):741–756. doi: 10.1016/s0161-6420(13)38009-9. [DOI] [PubMed] [Google Scholar]
- 10.Fundus photographic risk factors for progression of diabetic retinopathy. ETDRS report number 12. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology. 1991;98(5):823–833. [PubMed] [Google Scholar]
- 11.Grading diabetic retinopathy from stereoscopic color fundus photographs—an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology. 1991;98(5):786–806. [PubMed] [Google Scholar]
- 12.Davis MD, Fisher MR, Gangnon RE, et al. Risk factors for high-risk proliferative diabetic retinopathy and severe visual loss: early Treatment Diabetic Retinopathy Study Report #18. Investigative Ophthalmology and Visual Science. 1998;39(2):233–252. [PubMed] [Google Scholar]
- 13.Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Annals of Internal Medicine. 1999;130(6):461–470. doi: 10.7326/0003-4819-130-6-199903160-00002. [DOI] [PubMed] [Google Scholar]
- 14.Teerlink T, Nijveldt RJ, De Jong S, Van Leeuwen PAM. Determination of arginine, asymmetric dimethylarginine, and symmetric dimethylarginine in human plasma and other biological samples by high-performance liquid chromatography. Analytical Biochemistry. 2002;303(2):131–137. doi: 10.1006/abio.2001.5575. [DOI] [PubMed] [Google Scholar]
- 15.Mittermayer F, Namiranian K, Pleiner J, Schaller G, Wolzt M. Acute Escherichia coli endotoxaemia decreases the plasma L-arginine/asymmetrical dimethylarginine ratio in humans. Clinical Science. 2004;106(6):577–581. doi: 10.1042/CS20030363. [DOI] [PubMed] [Google Scholar]
- 16.Reaven PD, Emanuele N, Moritz T, et al. Proliferative diabetic retinopathy in type 2 diabetes is related to coronary artery calcium in the Veterans Affairs Diabetes Trial (VADT) Diabetes Care. 2008;31(5):952–957. doi: 10.2337/dc07-1926. [DOI] [PubMed] [Google Scholar]
- 17.Kielstein JT, Böger RH, Bode-Böger SM, et al. Marked increase of asymmetric dimethylarginine in patients with incipient primary chronic renal disease. Journal of the American Society of Nephrology. 2002;13(1):170–176. doi: 10.1681/ASN.V131170. [DOI] [PubMed] [Google Scholar]
- 18.Parving HH, Gall MA, Nielsen FS. Dyslipidaemia and cardiovascular disease in non-insulin-dependent diabetic patients with and without diabetic nephropathy. Journal of Internal Medicine, Supplement. 1994;236(736):89–94. [PubMed] [Google Scholar]
- 19.Nelson RG, Wolfe JA, Horton MB, Pettitt DJ, Bennett PH, Knowler WC. Proliferative retinopathy in NIDDM. Incidence and risk factors in Pima Indians. Diabetes. 1989;38(4):435–440. doi: 10.2337/diab.38.4.435. [DOI] [PubMed] [Google Scholar]
- 20.Sugai M, Ohta A, Ogata Y, et al. Asymmetric dimethylarginine (ADMA) in the aqueous humor of diabetic patients. Endocrine Journal. 2007;54(2):303–309. doi: 10.1507/endocrj.k06-140. [DOI] [PubMed] [Google Scholar]
- 21.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. American Journal of Kidney Diseases. 2005;46(2):214–224. doi: 10.1053/j.ajkd.2005.05.005. [DOI] [PubMed] [Google Scholar]
- 22.Wong TY, Coresh J, Klein R, et al. Retinal microvascular abnormalities and renal dysfunction: the atherosclerosis risk in communities study. Journal of the American Society of Nephrology. 2004;15:2469–2476. doi: 10.1097/01.ASN.0000136133.28194.E4. [DOI] [PubMed] [Google Scholar]