Purpose
To report the prevalence and risk factors for retinopathy in African Americans with impaired fasting glucose (IFG) and type 2 diabetes in the Jackson Heart Study and to determine if P-selectin plasma levels are independently associated with retinopathy in this population.
Design
Prospective, cross-sectional observational study.
Methods
setting : Community-based epidemiologic study. study population : Total of 629 patients with type 2 diabetes and 266 participants with impaired fasting glucose. observation procedures : Bilateral, 7-field fundus photographs were scored by masked readers for diabetic retinopathy (DR) level. Covariate data including P-selectin plasma levels and genotypes were collected in a standardized fashion. main outcome measures : Association between risk factors, including P-selectin plasma levels and genotypes, and retinopathy.
Results
The prevalences of any retinopathy among participants with IFG and type 2 diabetes were 9.4% and 32.4%, respectively. Among those with type 2 diabetes, in multivariate models adjusted for age, sex, and other traditional risk factors, higher P-selectin levels were associated with any DR (odds ratio = 1.11, 95% confidence interval = 1.02–1.21, P = .02) and proliferative DR (odds ratio = 1.23, 95% confidence interval = 1.03–1.46, P = .02). To further investigate the relationship between P-selectin and DR, we examined the association between P-selectin genotype and DR. Minor allele homozygotes for the variant rs6128 were less likely to develop DR ( P after Bonferroni correction = 0.03).
Conclusions
Both serologic and genetic data show an association between P-selectin and DR in the Jackson Heart Study. If confirmed in other studies, this association may provide insight into the pathogenesis of retinopathy.
Diabetes is the leading cause of blindness among working-age adults in the United States. There is evidence that diabetic retinopathy (DR) is more prevalent in African Americans than in non-Hispanic whites. Epidemiologic and clinical studies have provided information regarding DR in African Americans with type 2 diabetes. Some of these studies were performed over 20 years ago when diabetes treatment options were limited and patients had poorer glycemic control. Many of these studies only used 1 or 2 photographic fields to ascertain DR. Limited-field photography can lead to inaccurate DR grading as compared with dilated 7-field fundus photography. Nonmydriatic limited-field photographs are also more likely to be ungradable. Studies have shown that subjects with ungradable photographs are more likely to have characteristics consistent with increased retinopathy risk and be African Americans. Therefore, retinopathy may be under-ascertained in African Americans, particularly when a limited number of fields are photographed without pharmacologic pupil dilation.
Longer diabetes duration, hyperglycemia, and hypertension are consistent risk factors for DR, and there are other putative risk factors for DR. Hyperlipidemia and obesity impact DR in some, but not all, studies. Increased urinary albumin has been associated with retinopathy in some populations. C-reactive protein has not been a biomarker for DR in most studies, but a recent prospective investigation found an association with macular edema. P-selectin and E-selectin are molecules involved in leukocyte recruitment and rolling and platelet adhesion. A genetic association between variants in the P-selectin gene, SELP , and DR has been reported in white subjects.
Impaired fasting glucose is defined as a fasting plasma glucose between 100 and 125 mg/dL. Retinopathy develops in 7%–10% of impaired fasting glucose patients, and blood pressure and body mass index have been identified as risk factors for retinopathy in Europeans with impaired fasting glucose. There is limited information regarding retinopathy prevalence and risk factors in African Americans with impaired fasting glucose.
The purpose of our study was to estimate the prevalence of and identify risk factors for retinopathy ascertained with 7-field, dilated fundus photography in African Americans with impaired fasting glucose and type 2 diabetes from the Jackson Heart Study. In particular, we were interested in whether there was an association between P-selectin plasma levels and genotype and retinopathy.
Methods
All aspects of this cross-sectional observational study were prospectively approved by the Institutional Review Boards at the University of Mississippi Medical Center and Massachusetts Eye and Ear Infirmary. All participants provided written informed consent. The study was compliant with the regulations of the Health Insurance Portability and Accountability Act.
The Jackson Heart Study is a community-based observational study of cardiovascular disease among African Americans living in Mississippi. Standardized phenotyping protocols have measured physical characteristics at the 2004 baseline examination and subsequent examinations (Examination 2 in 2005-08 and Examination 3 in 2009-12). We established an ancillary retinopathy study that ran concurrently with Examination 3. Jackson Heart Study participants with type 2 diabetes or impaired fasting glucose were invited to participate in the ancillary study. Diabetes was defined as (a) taking antidiabetic medication or (b) meeting the hemoglobin A1c or fasting plasma glucose criteria for diabetes diagnosis. Type 2 diabetes was defined as age at diabetes diagnosis ≥30 years. Impaired fasting glucose was defined as specified above.
Enrolled participants had 1 study visit at the University of Mississippi Medical Center Department of Ophthalmology. A short questionnaire about ocular history and diabetes diagnosis was administered. Bilateral, dilated, 7-standard-field fundus photographs including macular stereoscopic pairs were obtained with a TRC-50DX camera (Topcon, Tokyo, Japan). The photographs were scored contemporaneously by 2 independent, masked ophthalmologist-investigators with the Early Treatment Diabetic Retinopathy Study (ETDRS) scale. Clinically significant macular edema was deemed present if ETDRS criteria were met and/or if focal laser treatment scars were present. Disagreements were arbitrated by a third masked ophthalmologist-investigator and/or by joint review by the ophthalmologist-investigators. Participants were excluded only if they had another ocular disease that precluded photograph grading.
Quality metrics for contemporaneous and temporal reproducibility were assessed. Intergrader agreement was measured in terms of percentage agreement and weighted kappa. The contemporaneous intergrader exact agreement percentages for ETDRS grade and presence of clinically significant macular edema were 96.3% and 96.8%, respectively. The associated weighted kappas were 0.76 and 0.52, respectively. One hundred photographs graded in the first year were randomly selected and regraded at the study’s end with readers masked to original grade assigned. For these regraded photographs, the percentages of exact agreement between the grades assigned in the first year and last year of the study were 95.5% and 94.0%, respectively, for ETDRS grade and presence of clinically significant macular edema. The corresponding weighted kappas were 0.84 and 0.48, respectively. These metrics suggest there was no systematic temporal drift in evaluating ETDRS grade and presence of clinically significant macular edema.
Covariate Data
Covariates were chosen because of association in other populations and data availability from Jackson Heart Study examinations. The risk factors considered were: age, sex, hemoglobin A1c, diabetes duration, body mass index, waist circumference, systolic blood pressure, diastolic blood pressure, total cholesterol, triglycerides, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, C-reactive protein, urine albumin, smoking status, diabetic medication use, E-selectin, and P-selectin.
Covariate data were selected from the Jackson Heart Study examination closest to the retinopathy study visit for which the participant had available data. Diabetes duration was calculated from the onset date to the retinopathy examination date. Seated blood pressure was measured 3 times and the mean was used in the analyses. Smoking status was defined as ever having smoked (current or past smoker) vs never having smoked. Urine albumin was obtained from a spot urine specimen. Diabetic medication use was defined as taking any diabetic medication. Plasma E-selectin and P-selectin were quantified by enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, Minnesota, USA). The interassay coefficients of variation for the E-selectin and P-selectin detection methods were 9.78% and 5.14%, respectively. A subset of patients in the Jackson Heart Study had been genotyped as part of a larger consortium for several vascular disease candidate genes, as previously described. This list of candidate genes included 3 single nucleotide polymorphisms in SELP : rs6128, rs6133, and rs3917779.
Statistical Methods
Retinopathy status was based on the eye with the higher ETDRS level. Absence of retinopathy was defined as ETDRS level <14. We examined 3 different outcomes: any retinopathy (ETDRS level ≥14), proliferative DR (ETDRS level ≥60), and clinically significant macular edema. Clinically significant macular edema was deemed present if at least 1 eye demonstrated clinically significant macular edema and/or focal laser scars. If 1 eye was ungradable, the scores for the other eye were used to define these outcomes.
We estimated the prevalence of any retinopathy in impaired fasting glucose participants and of any retinopathy, proliferative diabetic retinopathy (PDR), and clinically significant macular edema in type 2 diabetes participants. We examined risk factors in participants in our ancillary study compared with those who did not participate. Univariate analyses comparing participants with and without any retinopathy were performed separately for impaired fasting glucose and type 2 diabetes participants. For type 2 diabetes participants, univariate analyses comparing participants (1) with PDR (ETDRS level ≥60) and without PDR (ETDRS level <60) and (2) with and without clinically significant macular edema were also performed. Differences in means and proportions were tested by the t test and χ 2 test, respectively. We used the subset of participants with complete information for the covariate of interest in that particular analysis to maximize generalizability and power.
To examine the association between risk factors and retinopathy, we constructed multivariate models using backward stepwise logistic regression to determine the odds ratios and 95% confidence intervals. All models were adjusted for age and sex. A P value <.05 was considered statistically significant. For the genetic analyses of the association between single nucleotide polymorphisms in SELP and retinopathy, the χ 2 test was used to compare the frequency of minor allele homozygotes between cases and controls. The Bonferroni method was used to correct for multiple hypothesis testing. All analyses were performed using Stata/IC version 12.1(Stata, College Station, Texas, USA).
Results
From Jackson Heart Study examinations 1 and 2, we identified 1303 type 2 diabetes participants and 689 impaired fasting glucose participants. A total of 629 type 2 diabetes and 266 impaired fasting glucose participants enrolled in the retinopathy study. Table 1 compares known DR risk factors between enrolled and nonenrolled participants. Enrolled impaired fasting glucose participants had a higher mean age and lower mean diastolic blood pressure. On average, enrolled type 2 diabetes participants had shorter diabetes duration, lower hemoglobin A1c, and lower systolic blood pressure.
| Variable | Enrolled in Retinopathy Study | Not Enrolled in Retinopathy Study | P Value | ||
|---|---|---|---|---|---|
| N | % or Mean (SD) | N | % or Mean (SD) | ||
| Impaired fasting glucose participants | |||||
| Age, y | 266 | 65.7 (9.6) | 423 | 63.0 (12.4) | .003 |
| Percent female | 266 | 59.8% | 423 | 52.7% | .07 |
| Hemoglobin A 1C , % | 266 | 5.9 (0.38) | 422 | 5.9 (0.58) | .34 |
| Body mass index, kg/m 2 | 266 | 31.3 (6.0) | 422 | 32.3 (7.0) | .06 |
| Waist circumference, cm | 266 | 102.6 (13.6) | 423 | 104.1 (16.0) | .20 |
| Systolic BP, mm Hg | 266 | 128.5 (16.6) | 423 | 130.8 (18.9) | .11 |
| Diastolic BP, mm Hg | 266 | 74.6 (9.9) | 423 | 76.8 (11.1) | .01 |
| Total cholesterol, mg/dL | 266 | 199.3 (37.3) | 423 | 202.1 (39.8) | .36 |
| Triglycerides, mg/dL | 266 | 103.4 (65.7) | 423 | 102.5 (58.3) | .86 |
| LDL cholesterol, mg/dL | 266 | 121.1 (34.4) | 422 | 125.9 (37.3) | .09 |
| HDL cholesterol, mg/dL | 266 | 57.6 (17.5) | 423 | 55.8 (16.5) | .16 |
| C-reactive protein, mg/L | 266 | 4.4 (5.3) | 423 | 5.5 (11.0) | .12 |
| Urine albumin, mg/L | 259 | 33.7 (119.3) | 391 | 39.0 (152.5) | .64 |
| Ever smoked, % | 266 | 26.6% | 423 | 29.3% | .35 |
| Type 2 diabetes participants | |||||
| Age, y | 629 | 66.5 (10.1) | 674 | 66.4 (11.0) | .95 |
| Percent female | 629 | 66.1% | 674 | 66.3% | .94 |
| Diabetes duration, y | 629 | 9.6 (9.0) | 655 | 12.0 (9.8) | 7.4 × 10 −6 |
| Hemoglobin A 1C , % | 623 | 7.2 (1.6) | 668 | 7.5 (1.8) | .002 |
| Body mass index, kg/m 2 | 627 | 34.1 (6.9) | 674 | 34.0 (7.5) | .89 |
| Waist circumference, cm | 627 | 109.0 (15.2) | 673 | 109.1 (16.2) | .94 |
| Systolic BP, mm Hg | 627 | 130.0 (20.2) | 674 | 132.9 (21.1) | .01 |
| Diastolic BP, mm Hg | 627 | 73.6 (11.1) | 674 | 74.1 (11.7) | .44 |
| Total cholesterol, mg/dL | 612 | 188.9 (44.6) | 642 | 190.3 (46.2) | .58 |
| Triglycerides, mg/dL | 612 | 112.2 (65.4) | 642 | 123.5 (145.2) | .07 |
| LDL cholesterol, mg/dL | 612 | 111.7 (39.6) | 632 | 113.6 (40.7) | .42 |
| HDL cholesterol, mg/dL | 612 | 54.7 (14.8) | 642 | 53.4 (14.9) | .12 |
| C-reactive protein, mg/L | 626 | 5.5 (7.9) | 672 | 5.4 (9.1) | .81 |
| Urine albumin, mg/L | 610 | 150.2 (621.5) | 583 | 146.8 (526.5) | .92 |
| Ever smoked, % | 625 | 33.0% | 673 | 33.1% | .95 |
| Diabetic medication use | 624 | 78.2% | 667 | 80.7% | .28 |
The distribution of retinopathy grades is shown in Table 2 . The prevalences of any retinopathy in impaired fasting glucose and type 2 diabetes participants were 9.4% and 32.4%, respectively. No impaired fasting glucose participants had clinically significant macular edema. Among type 2 diabetes participants, 48 (7.8%) had clinically significant macular edema and 28 (4.5%) had PDR. In impaired fasting glucose participants, none of the examined covariates were associated with retinopathy in univariate or multivariate analyses.
| Worst Eye ETDRS Grade | Impaired Fasting Glucose Participants (Mean Age = 65.7 Years) | Type 2 Diabetes Participants (Mean Age = 66.5 Years) |
|---|---|---|
| Unable to grade | 1 | 6 |
| ≤14 | 240 | 421 |
| 15 | 12 | 31 |
| 20 | 9 | 50 |
| 35 | 4 | 76 |
| 43 | 0 | 11 |
| 47 | 0 | 4 |
| 53 | 0 | 2 |
| 71 | 0 | 27 |
| 81 | 0 | 1 |
| Total | 266 | 629 |
In type 2 diabetes participants, longer diabetes duration, higher hemoglobin A1c, higher systolic blood pressure, greater waist circumference, higher urine albumin, diabetic medication use, and higher plasma P-selectin levels were associated with retinopathy in univariate analyses ( Table 3 ). In the multivariate model ( Table 4 ), P-selectin levels remained significantly associated with presence of any retinopathy when controlling for other risk factors, including C-reactive protein, another inflammatory biomarker. The odds of having retinopathy were 71% higher for every additional 5 years of diabetes ( P = 5.7 × 10 −13 ), 25% higher for each percentage point increase in hemoglobin A1c ( P = .002), 13% higher for every 10 mm Hg increase in systolic blood pressure ( P = .03), 2-fold higher for those using a diabetic medication ( P = .04), and 11% higher for each 5 ng/mL increase in plasma P-selectin ( P = .02). Because higher P-selectin levels have been associated with cardiovascular events in other populations, we examined whether inclusion of history of a cardiovascular event (defined as coronary heart disease, stroke, or heart failure) in the model would alter the association of P-selectin levels with presence of retinopathy, but this did not alter the results, with higher P-selectin levels remaining significantly associated with retinopathy despite adjusting for cardiovascular events ( P = .01).
| Variable | Any Retinopathy (ETDRS Grade ≥14) | Proliferative Diabetic Retinopathy (ETDRS Grade ≥60) | Clinically Significant Macular Edema | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No Retinopathy | Retinopathy | P Value a | No PDR | PDR | P Value a | No CSME | CSME | P Value a | |||||||
| N | % or Mean (SD) | N | % or Mean (SD) | N | % or Mean (SD) | N | % or Mean (SD) | N | % or Mean (SD) | N | % or Mean (SD) | ||||
| Age, y | 421 | 66.6 (9.9) | 202 | 65.8 (10.4) | .31 | 595 | 66.2 (10.0) | 28 | 70.2 (9.9) | .04 | 571 | 66.2 (9.9) | 48 | 67.5 (11.0) | .38 |
| Percent female | 421 | 64.6% | 202 | 69.8% | .20 | 595 | 65.9% | 28 | 75.0% | .32 | 571 | 66.5% | 48 | 64.6% | .78 |
| Diabetes duration, y | 421 | 6.8 (6.9) | 202 | 15.3 (10.1) | 1.5 × 10 −31 | 595 | 8.8 (8.1) | 28 | 26.7 (9.2) | 4.7 × 10 −27 | 571 | 8.6 (8.2) | 48 | 20.5 (10.1) | 6.0 × 10 −20 |
| Hemoglobin A 1C , % | 415 | 6.9 (1.2) | 202 | 7.9 (1.9) | 3.5 × 10 −14 | 589 | 7.2 (1.6) | 28 | 7.8 (1.7) | .03 | 565 | 7.1 (1.5) | 48 | 8.2 (2.0) | 2.4 × 10 −6 |
| Body mass index, kg/m 2 | 419 | 33.8 (7.0) | 202 | 34.6 (6.4) | .17 | 593 | 34.0 (6.8) | 28 | 36.7 (7.0) | .04 | 569 | 34.0 (6.8) | 48 | 34.6 (6.9) | .56 |
| Waist circumference, cm | 419 | 108.2 (15.4) | 202 | 111.0 (14.7) | .03 | 593 | 108.7 (14.9) | 28 | 117.4 (19.0) | .003 | 569 | 108.8 (15.0) | 48 | 112.4 (16.6) | .11 |
| Systolic blood pressure, mm Hg | 419 | 128.7 (18.4) | 202 | 132.4 (23.4) | .03 | 593 | 129.7 (19.8) | 28 | 134.9 (28.3) | .18 | 569 | 129.8 (20.1) | 48 | 132.3 (22.2) | .42 |
| Diastolic blood pressure, mm Hg | 419 | 74.0 (10.3) | 202 | 72.9 (12.5) | .23 | 593 | 73.9 (10.9) | 28 | 69.2 (13.1) | .03 | 569 | 74.0 (10.9) | 48 | 70.3 (12.4) | .03 |
| Total cholesterol, mg/dL | 412 | 190.0 (42.9) | 194 | 186.5 (48.2) | .37 | 579 | 189.6 (44.6) | 27 | 171.8 (41.6) | .04 | 556 | 189.0 (44.0) | 46 | 191.2 (51.6) | .74 |
| Triglycerides, mg/dL | 412 | 110.2 (62.4) | 194 | 116.6 (71.6) | .27 | 579 | 112.0 (65.9) | 27 | 117.7 (56.4) | .66 | 556 | 112.6 (66.8) | 46 | 106.8 (48.9) | .57 |
| LDL cholesterol, mg/dL | 412 | 112.9 (39.3) | 194 | 109.2 (40.3) | .29 | 579 | 112.5 (39.6) | 27 | 95.3 (35.5) | .03 | 556 | 111.9 (39.0) | 46 | 113.1 (45.9) | .85 |
| HDL cholesterol, mg/dL | 412 | 55.0 (15.2) | 194 | 54.1 (13.9) | .46 | 579 | 54.8 (14.9) | 27 | 53.0 (12.3) | .54 | 556 | 54.6 (14.9) | 46 | 56.8 (12.5) | .33 |
| C-reactive protein, mg/L | 418 | 5.9 (8.7) | 202 | 4.9 (5.9) | .13 | 592 | 5.5 (8.0) | 28 | 6.4 (5.0) | .55 | 568 | 5.6 (8.2) | 48 | 4.8 (4.5) | .47 |
| Urine albumin, mg/L | 412 | 88.0 (424.6) | 193 | 283.3 (901.3) | .0003 | 581 | 135.8 (595.8) | 24 | 502.1 (1064.8) | .005 | 555 | 131.0 (579.3) | 46 | 393.0 (1013) | .006 |
| Ever smoked | 418 | 33.2% | 201 | 31.0% | 0.59 | 592 | 32.3% | 27 | 37.0% | .60 | 568 | 32.0% | 47 | 38.3% | .38 |
| Diabetic medication use | 416 | 70.2% | 202 | 94.6% | 6.2 ×10 −12 | 590 | 77.1% | 28 | 100% | .004 | 566 | 76.3% | 48 | 97.9% | .001 |
| E-selectin, ng/mL | 393 | 47.7 (23.8) | 189 | 48.7 (24.5) | 0.16 | 556 | 46.8 (24.2) | 26 | 43.4 (19.9) | .47 | 531 | 46.5 (24.5) | 47 | 46.9 (18.5) | .91 |
| P-selectin, ng/mL | 393 | 32.8 (12.2) | 189 | 36.6 (13.2) | 0.0007 | 556 | 33.8 (12.5) | 26 | 38.9 (15.7) | .04 | 531 | 33.7 (12.6) | 47 | 37.3 (12.8) | .06 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
