Patients with chronic kidney disease (CKD) experience other diseases such as cardiovascular disease (CVD) and retinopathy. The purpose of this study was to assess whether retinopathy predicts future CVD events in a subgroup of the participants of the Chronic Renal Insufficiency Cohort (CRIC) study. In this ancillary investigation, 2,605 participants of the CRIC study were invited to participate, and nonmydriatic fundus photographs were obtained in 1,936 subjects. Using standard protocols, presence and severity of retinopathy (diabetic, hypertensive, or other) and vessel diameter caliber were assessed at a central photograph reading center by trained graders masked to study participant’s information. Patients with a self-reported history of cardiovascular disease were excluded. Incident CVD events were adjudicated using medical records. Kidney function measurements, traditional and nontraditional risk factors, for CVD were obtained. Presence and severity of retinopathy were associated with increased risk of development of any CVD in this population of CKD patients, and these associations persisted after adjustment for traditional risk factors for CVD. We also found a direct relation between increased venular diameter and risk of development of CVD; however, the relation was not statistically significant after adjustment for traditional risk factors. In conclusion, the presence of retinopathy was associated with future CVD events, suggesting that retinovascular pathology may be indicative of macrovascular disease even after adjustment for renal dysfunction and traditional CVD risk factors. Assessment of retinal morphology may be valuable in assessing risk of CVD in patients with CKD, both clinically and in research settings.
Patients with chronic kidney disease (CKD) experience high rates of comorbid illnesses including cardiovascular disease (CVD) and retinopathy. We have previously reported the results of the Retinopathy in Chronic Renal Insufficiency Cohort (RCRIC) study on the cross-sectional association between retinopathy and cardiovascular disease (CVD) in a group of patients with chronic kidney disease (CKD) enrolled in the Chronic Renal Insufficiency Cohort (CRIC) study. In this cohort, the prevalence of CVD was higher in patients with retinopathy and in patients with retinal venular dilation, and this association remained significant after controlling for traditional risk factors for CVD. This present report evaluates the association between retinopathy and venular dilatation and the risk of subsequent CVD events.
Methods
The CRIC study design and methods have been reported previously. Participants for the RCRIC study, an ancillary investigation of the CRIC study, were recruited at 6 of the 7 CRIC clinical centers. All 2,605 participants of the CRIC study from these 6 sites were invited to participate in the RCRIC study. From June 2006 to May 2008, 1936 participants enrolled in the RCRIC study, and eye fundus photography were obtained at median time of 2 years after the CRIC baseline. The protocol was approved by an institutional review board at each of the participating institutions, and all participants provided written informed consent.
Trained nonophthalmic personnel obtained all photographs. To induce a physiologic, nonpharmacologic dilatation of the pupils, participants were seated in a darkened room for 5 minutes. Forty-five degree digital, color fundus photographs were obtained using a Canon CR-DGI, Non-Mydriatic Retinal Camera (Canon Inc, Tokyo, Japan). Two photographs, one centered on the macula and the other on the optic disc, were obtained from each eye. A participant was included in the analysis if either the disc or macula photographs of 1 eye were of sufficient quality for classification by fundus reading center staff.
Trained graders and a retinal specialist, without knowledge of the participant’s clinical and demographic information, assessed all digital fundus photographs at the RCRIC Fundus Photograph Reading Center at the University of Pennsylvania. Fundus pathology including presence and severity of retinopathy of any cause (diabetic, hypertensive, or other) and measurement of the diameter of the major retinal vessels were assessed. Because the readers were unaware of the diabetic or hypertensive status of the participants, retinopathy was evaluated without assumption of cause.
The Atherosclerosis Risk in Communities (ARIC) fundus photographic and the Early Treatment of Diabetic Retinopathy (ETDRS) grading protocols were used to grade retinopathy due to diabetes mellitus, systemic hypertension, and other conditions. The Multi-Ethnic Study of Atherosclerosis protocol was used for the evaluation of macular edema. These grading protocols have been previously validated in diabetic and nondiabetic populations. A single masked reader, using standard protocols with standardized photographic field definitions, evaluated digital fundus photographs displayed on color-calibrated monitors.
Retinal abnormalities were graded by comparing participant images to standard photographs. An ETDRS severity score was assigned for each eye. The ETDRS severity score is on an ordinal scale instead of a continuous scale. Scores were classified as normal (<14), very mild nonproliferative retinopathy (NPR 14 to 20), nonproliferative retinopathy (35 to 53), and proliferative retinopathy (PR >60). The score of the eye with more advanced retinopathy or the score from a single eye, if only 1 eye was available, was used as the score of the participant. A total of 116 participants (6%) had photographs in which neither eye could be assessed. Among them, 38 participants had photographs in which retinopathy features could not be assessed because of poor image quality. The remaining 78 participants had photographs that were blurry or dark, and although some mild retinopathy features could be assessed, an accurate grading could not be assigned because more advanced and subtle retinopathy features were not discernible.
As previously reported, grade–regrade reliability was assessed in 200 RCRIC participants. Weighted Kappa for participant’s ETDRS score was 0.77 (95% confidence interval [CI] 0.67 to 0.88), a value consistent with the reproducibility assessment reported by the ETDRS study.
Image measurements of vascular arteriolar and venular calibers were performed according to the ARIC protocol, using interactive vessel analysis software (IVAN) developed at the University of Wisconsin. Vessels were measured within an annulus spanning 0.5 to 1 disc diameter from the edge of the disc. Graders identified major arterioles and venules and chose segments most suitable for measurement. The diameter of up to 6 arterioles and 6 venules were averaged.
The participants of CRIC study were queried about any cardiovascular events at the time of enrollment to the CRIC study. This information was not ascertained by investigating patient’s medical records. Subsequently, during semiannual CRIC study visits, participants were queried again about possible cardiovascular events, onset of end-stage renal disease (ESRD), all hospitalizations, and a selected set of diagnostic tests or procedures. International Classification of Diseases, Ninth Revision, discharge codes were recorded for all hospitalizations. When codes included any from a preselected list suggesting a cardiovascular event (congestive heart failure [CHF], myocardial infraction [MI], stroke, atrial fibrillation [AF], and peripheral arterial disease [PAD]) or participants died during a hospitalization, medical records were retrieved for detailed review using event-specific criteria. Two physicians performed these reviews and classified each hospitalization with respect to the probability of the events of CHF, MI, stroke, and AF. Trained study staff reviewed medical records classified with International Classification of Diseases, Ninth Revision, codes that suggested PAD and abstracted data onto a study form.
The criteria used for adjudication of CHF were based on clinical symptoms, radiographic evidence of pulmonary congestion, physical examination of the heart and lungs, central venous hemodynamic monitoring data, and echocardiographic imaging. The criteria for MI were based on symptoms of angina, cardiac biomarkers, and electrocardiographic data. The criteria for AF were based on electrocardiographic data, rhythm strips, and selected medical record notes. Two neurologists reviewed all hospitalizations suggestive of stroke. The criteria for stroke were based on neurologic symptoms, tests/procedures including magnetic resonance imaging of the brain, computerized axial tomography of the brain, transthoracic echocardiographic and transesophageal echocardiographic imaging, and any history of previous stroke. The guidelines for MI were based on symptoms of angina, cardiac biomarkers, and electrocardiographic data. The guidelines for AF were based on electrocardiographic data, rhythm strips, and selected medical record notes. Hypertension was defined as either systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or use of antihypertensive medications. Diabetes was defined as either fasting glucose ≥126 mg/dl, random glucose ≥200 mg/dl, or use of insulin or antidiabetic medication. ESRD was defined as the initiation of long-term dialysis or kidney transplantation, and ascertainment of ESRD was supplemented by linkage to the US Renal Data System. Estimated glomerular filtration rate (eGFR) was calculated using the CRIC eGFR equation.
Only the 1,245 patients who did not report a history of CVD at baseline were considered at risk of developing a CVD event. Participant’s characteristics at the time of fundus photography were analyzed for all participants and stratified by diabetes status. Data values from the annual visit closest in time to fundus photography were used as baseline characteristics for the RCRIC study. The associations between retinopathy and incidence of each cardiovascular condition (AF, CHF, stroke, MI, and PAD), as well as a composite of all conditions with the exception of AF, termed “any CVD,” were assessed using Kaplan–Meier curves and Cox proportional hazards models to estimate hazard ratios (HR) and their 95% CIs. AF was excluded from the composite of any CVD because it is most probably caused by a different pathophysiological pathway. p Values were calculated for any difference between retinopathy severity categories and also for linear trends across categories among subjects with gradable photographs. Multivariate models were adjusted by traditional risk factors for CVD including age, gender, low-density lipoprotein, high-density lipoprotein, previous CVD history, systolic blood pressure, smoking status (never/former/current), diabetes, hypertension, hemoglobin A1C, eGFR, and log-transformed 24-hour urine protein. Similar analyses were performed for the association between any retinopathy or baseline retinal vascular calibers (categorized into quartiles) and any CVD (the composite cardiovascular outcome), for all participants together, and then, after a stratification by diabetes status. All data analyses were performed in SAS version 9.4 (SAS Institute Inc., Cary, North Carolina), and 2-sided p value <0.05 was considered to be statistically significant.
Results
Characteristics of the participants recruited have been described before. All 1,245 participants who had baseline retinal photographs and did not self-report any CVD at baseline were included in the study. Of these participants, 1,081 had retinal vascular caliber measurements. Median age at the time of photography was 60 years (range 22 to 77 years), 53.1% were white, 40.4% were black, 52.3% were men, and 39.8% had diabetes. At baseline, 896 participants (72.0%) had no retinopathy, 77 (6.2%) had mild nonproliferative retinopathy, 135 (10.8%) had nonproliferative retinopathy, 87 (7.0%) had proliferative retinopathy, and the remaining 50 (4.0%) were not assessed because of poor image quality. Table 1 provides participant’s characteristics overall and by presence or absence of diabetes mellitus.
| Characteristic | Non-diabetics (n=750) | Diabetics (n=495) | Total (n=1245) | |
|---|---|---|---|---|
| Age (years) | Mean (SD) | 56.5 (11.8) | 59.5 (10.4) | 57.7 (11.4) |
| Sex | Male | 378 (50.4%) | 273 (55.2%) | 651 (52.3%) |
| Female | 372 (49.6%) | 222 (44.8%) | 594 (47.7%) | |
| Smoking status | Non-smoker | 410 (54.7%) | 230 (46.5%) | 640 (51.4%) |
| Former smoker | 271 (36.1%) | 209 (42.2%) | 480 (38.6%) | |
| Current smoker | 69 (9.2%) | 56 (11.3%) | 125 (10.0%) | |
| Hypertension | Yes | 589 (78.5%) | 464 (93.7%) | 1053 (84.6%) |
| Systolic Blood Pressure (mmHg) | Mean (SD) | 121.8 (19.0) | 130.0 (21.4) | 125.0 (20.4) |
| High-density Lipoprotein (mg/dL) | Mean (SD) | 52.2 (17.5) | 46.8 (14.1) | 50.1 (16.4) |
| Low-density Lipoprotein (mg/dL) | Mean (SD) | 110.3 (31.3) | 93.1 (31.7) | 103.5 (32.5) |
| Triglycerides (mg/dL) | Mean (SD) | 143.2 (107.2) | 151.9 (105.3) | 146.7 (106.5) |
| Hemoglobin A1C (%) ∗ | Mean (SD) | 5.6 (0.49) | 7.3 (1.50) | 6.3 (1.33) |
| eGFR (ml/min per 1.73 m 2 ) | Mean (SD) | 50.4 (19.7) | 43.1 (17.9) | 47.6 (19.3) |
| 24-h Urine Protein (g/24H) | Mean (SD) | 0.55 (1.25) | 0.93 (1.71) | 0.70 (1.46) |
| Retinal Level | No NPR | 658 (87.7%) | 238 (48.1%) | 896 (72.0%) |
| Mild NPR | 49 (6.5%) | 28 (5.7%) | 77 (6.2%) | |
| NPR | 20 (2.7%) | 115 (23.2%) | 135 (10.8%) | |
| PR | 7 (0.9%) | 80 (16.2%) | 87 (7.0%) | |
| Ungradable | 16 (2.1%) | 34 (6.9%) | 50 (4.0%) | |
| Arteriole diameter (μm) | Mean (SD) | 148.6 (14.0) | 151.0 (14.0) | 149.4 (14.1) |
| Venule diameter (μm) | Mean (SD) | 218.0 (21.8) | 221.6 (23.9) | 219.3 (22.6) |
| Arteriole/Venule ratio | Mean (SD) | 0.68 (0.063) | 0.69 (0.073) | 0.69 (0.067) |
∗ Hemoglobin A1C was not available at baseline fundus photography, and therefore, the CRIC baseline was used.
In participants with gradable fundus photographs, retinopathy was present in 223 of 461 participants (48%) with diabetes compared to 76 of 734 participants (10%) without diabetes (p <0.001), and 283 of 1,008 participants (28%) with hypertension compared to 16 of 187 (10%) without hypertension (p <0.001). In 161 participants with neither diabetes mellitus nor hypertension, 3 (2%) had mild retinopathy, and another 4 had ungradable photographs.
During a median follow-up of 5.01 years, 103 of the 1,245 participants (8.3%) developed one or more cardiovascular events. They included 67 AF events, 52 CHF events, 22 stroke events, 32 MI events, and 17 PAD. New any CVD events were more common in participants with diabetes; 62 of 495 patients (12.5%) with diabetes and 41 of 750 subjects (5.5%) without diabetes had a new any CVD event.
Retinopathy was associated with increased risk of development of new any CVD event in the univariate analysis (p <0.001; Table 2 , Figure 1 ). Although less strong, this association remained statistically significant after adjustment for other CVD risk factors such as age, LDL, HDL, systolic pressure, urinary protein, eGFR, gender, diabetes, hypertension, smoking (multivariate analysis, p = 0.002; Table 2 ). Presence of any retinopathy (mild NPR, NPR, or PR) was significantly associated with increased risk of any CVD (HR 1.76 [1.03 to 2.99], p = 0.02, Table 2 ).
| Characteristic § | # of events / N (%) | Univariate | Multivariate ¶ | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | p-value | HR (95% CI) | p-value | |||
| Retinal Level | No NPR | 48/896 (5%) | 1 | <0.001 ∗ < 0.001 † | 1 | 0.002 ∗ 0.008 † |
| Mild NPR | 7/77 (9%) | 1.76 (0.80, 3.89) | 1.20 (0.47, 3.08) | |||
| NPR | 16/135 (12%) | 2.40 (1.36, 4.22) | 1.33 (0.64, 2.75) | |||
| PR | 23/87 (26%) | 6.05 (3.68, 9.96) | 3.40 (1.71, 6.78) | |||
| Ungradable | 9/50 (18%) | 4.22 (2.07, 8.61) | 3.14 (1.40, 7.06) | |||
| Any Retinopathy | No Retinopathy | 48/896 (5%) | 1 | <0.001 ∗ < 0.001 ‡ | 1 | 0.02 ∗ 0.04 ‡ |
| Any Retinopathy | 46/299 (15%) | 3.18 (2.12, 4.77) | 1.76 (1.03, 2.99) | |||
| Ungradable | 9/50 (18%) | 4.20 (2.06, 8.57) | 2.92 (1.31, 6.53) | |||
| Arteriole diameterquartile (μm) | 1 (≤140) | 20/272 (7%) | 1 | 0.86 ∗ 0.91 † | 1 | 0.61 ∗ 0.69 † |
| 2 (>140, ≤149) | 15/269 (6%) | 0.76 (0.39, 1.49) | 0.59 (0.28, 1.26) | |||
| 3 (>149, ≤158) | 19/270 (7%) | 0.94 (0.50, 1.76) | 0.78 (0.38, 1.60) | |||
| 4 (>158) | 20/270 (7%) | 0.98 (0.52, 1.81) | 0.78 (0.37, 1.62) | |||
| Venule diameterquartile (μm) | 1 (≤204) | 14/271 (5%) | 1 | 0.04 ∗ 0.009 † | 1 | 0.49 ∗ 0.12 † |
| 2 (<204, ≤218) | 13/270 (5%) | 0.93 (0.44, 1.98) | 1.23 (0.53, 2.83) | |||
| 3 (>218, ≤234) | 19/270 (7%) | 1.36 (0.68, 2.72) | 1.45 (0.64, 3.28) | |||
| 4 (>234) | 28/270 (10%) | 2.08 (1.10, 3.96) | 1.79 (0.83, 3.88) | |||
| Arteriole-Venuleratio quartile | 1 (≤0.64) | 28/271 (10%) | 1 | 0.02 ∗ 0.02 † | 1 | 0.23 ∗ 0.37 † |
| 2 (>0.64, ≤0.68) | 19/270 (7%) | 0.67 (0.37, 1.20) | 0.83 (0.43, 1.62) | |||
| 3 (>0.68, ≤0.73) | 10/270 (4%) | 0.33 (0.16, 0.67) | 0.42 (0.18, 0.97) | |||
| 4 (>0.73) | 17/270 (6%) | 0.58 (0.32, 1.07) | 0.89 (0.45, 1.76) | |||
∗ p value for evaluation of differences between categories.
† p value for evaluation of linear trend across categories, excluding ungradable.
‡ p value for evaluation of a difference between any retinopathy and no retinopathy.
§ Retinal level based on worst eye, retinal vascular calibers based on mean of both eyes.
¶ Adjusted by age, gender, low-density lipoprotein, high-density lipoprotein, systolic blood pressure, smoking status, diabetes, hypertension, triglycerides, hemoglobin A1C, eGFR, and log of 24-hour urine protein.
We detected no significant association between retinopathy and risk of development of AF ( Table 3 ). For CHF and PAD, although strong associations were detected in the univariate analysis (p <0.001, Table 3 ), these associations were not statistically significant after adjustment for other CVD risk factors (p = 0.34 for CHF and p = 0.07 for PAD). Among the outcomes that we studied, the risk of stroke showed the strongest relation with retinopathy. Participants with proliferative retinopathy had a 9.09 (95% CI 2.18 to 37.8) times greater risk of future stroke than participants without retinopathy ( Table 3 ). For MI and PAD, participants with proliferative retinopathy had a 5.43 (95% CI 1.37 to 21.5) and 5.73 (95% CI 1.14 to 28.9) times greater risk than participants without retinopathy.
| Condition | Retinal Level ‡ | # of events / N (%) | Univariate | Multivariate § | ||
|---|---|---|---|---|---|---|
| HR (95% CI) | p-value | HR (95% CI) | p-value | |||
| Atrial fibrillation (n=67) | No NPR | 52/896 (6%) | 1 | 0.76 ∗ 0.42 † | 1 | 0.49 ∗ 0.26 † |
| Mild NPR | 5/77 (6%) | 1.14 (0.46, 2.86) | 1.26 (0.49, 3.22) | |||
| NPR | 6/135 (4%) | 0.81 (0.35, 1.88) | 0.59 (0.22, 1.59) | |||
| PR | 3/87 (3%) | 0.63 (0.20, 2.03) | 0.51 (0.14, 1.87) | |||
| Ungradable | 1/50 (2%) | 0.37 (0.05, 2.68) | 0.29 (0.04, 2.13) | |||
| Congestive heart failure (n=52) | No NPR | 26/896 (3%) | 1 | 0.001 ∗ < 0.001 † | 1 | 0.34 ∗ 0.63 † |
| Mild NPR | 4/77 (5%) | 1.85 (0.65, 5.31) | 1.71 (0.56, 5.26) | |||
| NPR | 8/135 (6%) | 2.18 (0.99, 4.82) | 0.48 (0.15, 1.49) | |||
| PR | 9/87 (10%) | 4.02 (1.88, 8.59) | 1.11 (0.39, 3.18) | |||
| Ungradable | 5/50 (10%) | 3.96 (1.52, 10.3) | 1.68 (0.53, 5.26) | |||
| Stroke (n=22) | No NPR | 7/896 (1%) | 1 | <0.001 ∗ < 0.001 † | 1 | 0.04 ∗ 0.004 † |
| Mild NPR | 2/77 (3%) | 3.39 (0.70, 16.3) | n/a ¶ | |||
| NPR | 4/135 (3%) | 4.04 (1.18, 13.8) | 3.17 (0.70, 14.4) | |||
| PR | 7/87 (8%) | 11.7 (4.10, 33.4) | 9.09 (2.18, 37.8) | |||
| Ungradable | 2/50 (4%) | 5.77 (1.20, 27.9) | 6.49 (1.14, 37.0) | |||
| Myocardial infarction (n=32) | No NPR | 12/896 (1%) | 1 | <0.001 ∗ < 0.001 † | 1 | 0.06 ∗ 0.01 † |
| Mild NPR | 2/77 (3%) | 2.03 (0.45, 9.05) | 2.26 (0.48, 10.6) | |||
| NPR | 8/135 (6%) | 4.89 (2.00, 12.0) | 5.49 (1.54, 19.5) | |||
| PR | 7/87 (8%) | 6.66 (2.62, 16.9) | 5.43 (1.37, 21.5) | |||
| Ungradable | 3/50 (6%) | 5.23 (1.47, 18.6) | 5.36 (1.05, 27.4) | |||
| Peripheral arterial disease (n=17) | No NPR | 5/896 (1%) | 1 | <0.001 ∗ < 0.001 † | 1 | 0.07 ∗ 0.06 † |
| Mild NPR | 0/77 (0%) | n/a ¶ | n/a ¶ | |||
| NPR | 2/135 (1%) | 2.83 (0.55, 14.6) | 1.37 (0.20, 9.51) | |||
| PR | 7/87 (8%) | 16.5 (5.23, 52.2) | 5.73 (1.14, 28.9) | |||
| Ungradable | 3/50 (6%) | 12.9 (3.06, 54.2) | 7.30 (1.47, 36.3) | |||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
