Macular Edema and Retinal Hard Exudates in African Americans With Type 1 Diabetes: Title and subTitle BreakThe New Jersey 725 FREE

MACULAR EDEMA secondary to diabetic retinopathy is a well recognized cause of visual loss in persons with diabetes.^{1 - 2} In the population-based Wisconsin Epidemiologic Study of Diabetic Retinopathy(WESDR), the prevalence of macular edema among white patients with type 1 diabetes was 11.1%.³ In that study, macular edema was significantly associated with longer duration of diabetes, male sex, higher glycosylated hemoglobin, presence of proteinuria, and diuretic use.³ In diabetic patients, macular edema is often associated with retinal hard exudates.⁴ Severity of hard exudates has been shown to be significantly associated with elevated serum total cholesterol and low-density lipoprotein cholesterol (LDL-C) levels.^{5 - 6} The above studies have been done in predominantly white patient populations. There are no published data of frequency of, or risk factors for, macular edema and retinal hard exudates among African Americans with type 1 diabetes.

The New Jersey 725 is a hospital-based study of a large cohort of African Americans with type 1 diabetes.⁷ Patients have been identified from lists of hospital admissions in a geographically well-defined are a surrounding the medical school in Newark, NJ, where the study is conducted. The frequency of diabetic retinopathy (63.9%) and associated visual impairment(3.1%) in these patients is high.⁷ We herein report on the frequency of macular edema and retinal hard exudates and associated risk factors in this patient population.

Since 1982, all patients discharged from the 116 New Jersey hospitals with a diagnosis of diabetes mellitus are reported to the New Jersey Department of Health (New Jersey Hospital Discharge Data). From these discharges, 68 455 names of African Americans, with either a primary or secondary diagnosis of diabetes mellitus (International Classification of Diseases, Ninth Revision codes 250.0-250.9, 648.0, and 648.8),⁸ were identified for the years 1982 through 1996. First, a listing of all patients by last hospital admission was generated. This yielded a listing of 34 941 patients for the 31 hospitals located in the 7 counties within a 20-mile (32-km) radius of Newark and that had 100 or more admissions of African Americans with diabetes. A medical chart review was done to identify eligible patients. African Americans with type 1 diabetes mellitus, diagnosed and treated with insulin before 30 years of age, and currently taking insulin, were considered eligible. Ethnicity was determined from the hospital record and later confirmed by self-identification.

Of the 34 941 patients, medical charts of 13 615 were randomly selected and manually reviewed. Of those, 12 489 patients were excluded because they had either type 2 diabetes or insulin-requiring diabetes diagnosed after 30 years of age, or type 1½ diabetes and were not currently taking insulin therapy.^{9 - 10} Also excluded were patients with type 1 diabetes who were deceased (n = 160), resided outside New Jersey (n = 72), had sickle cell disease (n = 4), or were currently incarcerated(n = 15). Of the remaining 875 patients eligible for the study, 725 (82.9%) were enrolled, 38 (4.3%) could not be traced, and 112 (12.8%) declined to participate.

Participants were examined in the eye clinic at University Hospital, Newark. All patients signed an informed consent before being examined. Examination consisted of a detailed structured interview that included patients' sociodemographic factors, medical and ophthalmologic history, family history, diet, current medications, and lifestyle variables (ie, self-report measures of cigarette smoking and alcohol consumption). Weight and height were recorded. Blood pressure was measured twice using a random zero sphygmomanometer following the Hypertension Detection and Follow-up Program protocol.¹¹

A detailed eye examination was performed, including subjective refraction and best-corrected visual acuity using the Early Treatment of Diabetic Retinopathy Study (ETDRS) protocol.^{12 - 13} Intraocular pressure was measured by applanation tonometry and slitlamp examination of the anterior segment checked for corneal abnormalities, depth of the anterior chamber, and presence or absence of iris neovascularization. Pupils were then dilated and lens changes graded at the slitlamp using the Third Lens Opacification Classification System (LOCS III).¹⁴ Fundus examination was performed using direct and indirect ophthalmoscopy, as well as 90-diopter slitlamp examination. Lesions of diabetic retinopathy, including presence of clinically significant macular edema and hard exudates, were noted. Seven standard stereoscopic fundus photographs and red reflex (lens) of each eye were obtained using a 30° Zeiss fundus camera (Carl Zeiss, Oberkochen, Germany).¹⁵

A nonfasting venous blood sample was obtained for measurement of blood glucose using the hexokinase method, total glycosylated hemoglobin (hemoglobin A₁ and A₂) using high-pressure liquid chromatography(Bio-Rad, Labcorp Laboratory, Hercules, Calif), and plasma C-peptide using a radioimmunoassay. Plasma high-density lipoprotein cholesterol (HDL-C) and total cholesterol levels were measured using an enzymatic assay, and LDL-C levels using separation spectrophotometry (Genzyme Diagnostics, Cambridge, Mass).

A morning first-voided urine specimen collected at home by the patient and a 4-hour timed urine specimen collected in the clinic were assayed for creatinine level using the alkaline picrate method and for albumin using a radioimmunoassay (SmithKline Beecham Clinical Laboratory, Philadelphia, Pa). Creatinine level in the 4-hour timed urine collection was used to determine adequacy of collection. Leukocyte-nitrite dipstick (Boehringer Mannheim Corporation, Indianapolis, Ind) was used to exclude patients with urinary tract infection.

All fundus photographs were graded by the Fundus Photograph Reading Center in Madison, Wis. First, a detailed field-by-field, lesion-by-lesion, grading of each photographic set for each eye was performed using the ETDRS adaptation of the modified Airlie House classification of diabetic retinopathy.^{15 - 16} Subsequently, a computer program was used to analyze the detailed gradings and to derive a general retinopathy level for each eye. In addition to the retinopathy severity level for each eye (according to the ETDRS final scale), the severity of diabetic retinopathy for the patient was determined from the grading of the worse eye.

Macular edema was considered present if any are a of the retina within 1 disc diameter from the center of the macula was thickened with or without loss of transparency. Gradings generated for macular edema, macular thickening size, and center macular thickening were as follows: presence or absence of macular edema was graded as 0, if none; 10, if questionably present; 20, if 1 disc are a (DA) or more within 1 disc diameter from the center of the foveal avascular zone; 30, if 1 DA or more or hard exudates with retinal thickening less than or equal to 500 µ from the center of the foveal avascular zone; and 80, if it could not be graded. Macular thickening size was graded as 0, if none; 10, if questionable; 20, if less than ½ DA; 30, if ½ DA or more but less than 1 DA; 40, if at least 1 DA but less than 2 DAs; 50, if at least 2 DAs; and 80, if could not be graded. Center macular thickening was graded as 0, if absent; 10, if questionable; 20, if less than 1 time the reference thickness (twice the diameter of a major retinal vein at the disc margin); 30, if less than 2 times the reference; 40, less than ½ DA; 50, if at least 2 DAs; 70, if cannot grade because of poor stereophotographic quality; 80, if it could not be graded. In patients who had either questionable or no macular edema at the time of examination, review of all previous ophthalmologic records was done to document any history of laser photocoagulation for macular edema due to diabetes.

Presence or absence of macular edema in each eye was determined from the gradings of either macular edema and/or macular thickening size and/or center macular thickening—absent, if 0 or 10 and no history of laser photocoagulation for macular edema, and present, if either greater than 10 but less than 70 or history of photocoagulation for macular edema. Macular edema grading in each eye was used when examining the relationship between macular edema and ocular characteristics.

Macular edema for each patient was defined as absent, if in both eyes grading of either macular edema, and/or macular thickening size, and/or center thickening size was 0 or 10 and there was no history of laser photocoagulation for macular edema, and as present, if in either eye grading of either macular edema, and/or macular thickening size, and/or center thickening size was either greater than 10 but less than 70 or history of photocoagulation for macular edema. Eyes and patients with gradings of 70 and 80 were excluded from all analyses.

Severity of retinopathy in the worse eye was classified as follows: level 10, none; level 15, questionable retinopathy; levels 20-53, nonproliferative retinopathy of increasing severity; and levels 61-85, proliferative retinopathy of increasing severity.¹⁶ For patients who could not be graded because of media opacities or phthisis but who had a history of panretinal photocoagulation for proliferative diabetic retinopathy or pars plana vitrectomy for diabetic tractional retinal detachment or vitreous hemorrhage secondary to proliferative diabetic retinopathy, the retinopathy level was 85. Proliferative diabetic retinopathy was defined as ETDRS levels equal to or greater than 61 or less than 61 with a history of panretinal photocoagulation.

Presence of retinal hard exudates in the macula in each eye was based on the modified Airlie House classification scheme.¹⁶ Hard exudates were considered absent if they were either absent or questionable in both eyes, and as present, if they were levels 2, 3, 4, or 5 in either eye.¹⁶ Severity of hard exudates was classified as absent if hard exudates were absent, questionable in both eyes, or level 2 in either eye, and classified as present, if they were levels 3, 4, or 5 in either eye.

Visual impairment was classified as absent if the best-corrected distance visual acuity in the better eye was better than 20/40, and as present if less than or equal to 20/40.⁸

The patients' current age was defined as the age at examination or the age at onset of diabetes when the diagnosis was first recorded by a physician in the patient's chart, and the duration of diabetes was defined as the time between the two. Socioeconomic factors included patients' level of education(for those aged ≥25 years), marital and employment status, personal income(for those aged ≥18 years), and family income. To classify patients' socioeconomic status, we used Goldthorpe and Hope Classification of Occupations, which divides subjects into middle (levels 1 through 22) and lower (levels 23 through 36) class based on the occupation of the head of the household, and the Green index.^{17 - 19} Green index scores were divided into tertiles (low, <53; middle, 53-61.1; and high, >61.1). Current and past history of diuretic use was noted. Smoking history was defined as either past or current smoking and number of pack-years.

Mean systolic and diastolic blood pressures were the average of 2 systolic and 2 diastolic blood pressure measurements. Ocular perfusion pressure (PP) in right and left eyes was calculated using mean diastolic (D) and systolic(S) sitting blood pressures and intraocular pressure (IOP) using the following formula: PP = ⅔[D + (S − D) /3] − IOP. Patients were classified as having atherosclerotic vascular disease if they had a history of myocardial infarction, angina pectoris, leg or toe amputation, or stroke, as confirmed by review of the medical records.

Renal disease was considered to be present if (1) the albumin-creatinine ratio in either morning or 4-hour timed urine specimens was greater than 0.03,(2) the albumin excretion rate in the 4-hour timed urine specimen was equal to or greater than 20 µg/min, and/or (3) the patient was receiving dialysis or had a renal transplantation for diabetic renal disease. Microalbuminuria was classified as present if the albumin excretion rate was 20 through 200 µg/min, and macroproteinuria was classified as present if (1) the albumin excretion rate was greater than 200 µg/min and/or (2) the patient was receiving dialysis or had a kidney transplantation.

Patients were classified as currently receiving short- or long-acting insulin or a combination. Information regarding glucose control during the past year was recorded as follows: missing an insulin injection was defined as "often" if at least once a week and as "none" if less than that; frequency of home glucose urine testing was defined as "never" or "rarely" if performed a few times a year, as "sometimes" if done once a month or some months per year, and as "often" if done at least a few times a month per year; frequency of home blood glucose testing was defined as "sometimes" if done not more than once a week, and as "often" if performed at least a few times a week.

Data management and statistical analyses were performed using the Windows-based SPSS statistical software (SPSS Inc, Chicago, Ill). Estimates of the proportion of patients (number with macular edema or hard exudates/number patients) are reported as a function of age at examination and duration of diabetes.

Association between the presence of macular edema (or hard exudates) and hypothesized risk factors (including age, sex, duration of diabetes, age at diagnosis of diabetes, systemic hypertension, glycemic control, presence of proteinuria, socioeconomic status, and behavioral factors) was estimated and tested for significance using logistic regression analysis. For dichotomous variables, odds ratios (ORs) and corresponding 95% confidence intervals (CIs) were used to quantify the association between presence or absence of macular edema (or severity of hard exudates) and each risk factor. The statistical significance of the associations are based on Wald tests. For categorical variables with more than 2 categories, the ORs are presented for each level of the risk factor compared with a reference level thought to represent the lowest risk. Overall tests of association for these categorical variables are based on Wald tests. All tests are 2-sided and use a .05 significance level.

Descriptive analyses, including cross-tabulations, were performed to identify potential confounders for the relationship between macular edema(or severe hard exudates) and the various factors. Multiple logistic regression was used to isolate the impact of specific risk factors by controlling for the effect of potential confounders. The dependent variable in this regression was the presence or absence of macular edema (or severe hard exudates). Models including and excluding proteinuria were used. The generalized estimating equation approach to logistic regression was also used in the multivariate analysis to evaluate ocular variables for both eyes in relation to macular edema and severe hard exudates.

Of the 725 African American patients, macular edema was present in at least 1 eye in 89 (12.3%), questionable in 15 (2.1%), and could not be graded in 12 (1.7%). Fundus photographs were not available in 1 patient who had a documented history of focal laser photocoagulation for macular edema. Table 1 shows the frequency of macular edema in right and left eyes of the 725 patients. At the time of examination, 57 patients had unilateral and 16 had bilateral macular edema. There were 47 right and 42 left eyes with clinically significant macular edema; of those, 25 right (53.2%) and 30 left (71.4%) eyes had foveal involvement, and 41 right(87.2%) and 37 left (88.1%) eyes had no evidence of focal laser photocoagulation.

Compared with patients without macular edema, those with macular edema were significantly older (mean ± SD, 35.3 ± 9.3 years vs 27.3 ± 10.8 years; t₇₁₂ = 6.6, P<.001) and had longer duration of diabetes (17.0 ± 7.4 years vs 10.5 ± 9.2 years; t₇₁₂= 6.4, P<.001). Frequency of macular edema increased from 2.5% in patients aged 15 to 19 years to 33.3% in those aged 45 years or older (Figure 1). The frequency of macular edema also increased with duration of diabetes from 1.0% in those with less than 5 years of diabetes to 10.3% in patients with 10 years of diabetes and to 25.4% in those with 20 years or more (Figure 2).

After adjusting for duration of diabetes, men were on average 1.75 times more likely to have macular edema than women (OR, 1.75; 95% CI, 1.10-2.79)(Table 2). Patients with body mass index (calculated as weight in kilograms divided by the square of height in meters) of 25 or higher were on average twice as likely to have macular edema than those with a body mass index less than 25 (OR, 1.96; 95% CI, 1.21-3.18). Patients who missed their insulin injection at least once a week were also 1.8 times more likely to have macular edema than those who never missed their injections (OR, 1.81; 95% CI, 1.13-2.90).

Patients with hypertension were on average 3.6 times more likely to have macular edema than those who did not have hypertension (OR, 3.59; 95% CI, 2.13-6.05), and those in the highest quartile of either systolic or diastolic blood pressure were 3.5 times more likely to have macular edema than those in the lowest quartile (OR, 3.52; 95% CI, 1.77-6.97, and OR, 3.49; 95% CI, 1.90-6.42, respectively) (Table 2). Patients with microproteinuria were on average 4.6 times more likely to have macular edema than those without proteinuria and those with macroproteinuria were 11 times more likely (OR, 4.61; 95% CI, 2.07-10.24, and OR, 11.0; 95% CI, 5.27-22.97, respectively) (Table 2).Also, patients with evidence of peripheral vascular disease were on average twice as likely to have macular edema than those without peripheral vascular disease (OR, 2.39; 95% CI, 1.45-3.93).

There was no association between the presence of macular edema and age at onset of diabetes, glycosylated hemoglobin, fasting blood glucose, and blood C-peptide levels, frequency of a diabetic diet, number of insulin injections per day, insulin type, daily insulin dose, frequency of blood or urine glucose testing, diuretic use, or number of hospital admission or emergency department visits for high blood glucose levels (data not shown). There was also no association between presence of macular edema and education, personal or family income, socioeconomic status, smoking, or alcohol consumption after adjusting for duration of diabetes (data not shown).

Eyes with macular edema had significantly more visual impairment (χ²₂ = 6.30, P<.04 and χ²₂ = 5.95, P<.001, for right and left eyes, respectively), and proliferative diabetic retinopathy (χ²₁ = 77.3, P<.001).

After adjusting for duration of diabetes, patients with higher ocular perfusion pressure were on average twice as likely to have macular edema compared with those with lower perfusion pressure (OR, 2.11; 95% CI, 1.15-3.86, and OR, 2.16; 95% CI, 1.20-3.88, for right and left eyes, respectively). Patients with a myopic refractive error of at least −2 diopter spherical equivalent were at a higher risk for macular edema compared with those with a lower refractive error (OR, 1.19; 95% CI, 0.62-2.28, and OR, 2.07; 95% CI, 1.15-3.71), for right and left eyes, respectively). However, the association was only significant for left eyes. There was no association between presence of macular edema and intraocular pressure (data not shown).

Among the 660 patients without either dialysis or renal transplantation, the mean total cholesterol was 5.38 ± 1.46 mmol/L (208.0 ± 56.5 mg/dL), mean HDL-C was 1.43 ± 0.47 mmol/L (55.2 ± 18.1 mg/dL), and mean LDL-C was 2.84 ± 1.09 mmol/L (109.7 ± 42.0 mg/dL). Of the 604 patients in whom both total and HDL-C values were available, 31.3% had a total cholesterol–HDL-C ratio equal to or greater than 4.5.

After adjusting for duration of diabetes, patients in the highest quartile of either total or LDL-C were on average 2 to 3 times more likely to have macular edema (OR, 2.90; 95% CI, 1.36-6.17, and OR, 2.38; 95% CI, 1.12-5.05, respectively) than patients in the lowest quartile (Table 3). Patients with a total cholesterol–HDL-C ratio of 4.5 or more were also at higher risk of having macular edema than those with a ratio less than 4.5 (OR, 1.70; 95% CI, 1.01-2.88). There was no significant association between HDL-C level and macular edema.

In the multivariate analyses, presence of microproteinuria or macroproteinuria, missing insulin injections at least once a week, and longer duration of diabetes were significantly and independently associated with the presence of macular edema (Table 4). When proteinuria was excluded from the model, missing insulin injections at least once a week and longer duration of diabetes were the only 2 variables significantly and independently associated with macular edema. Male sex and LDL-C levels were of borderline significance. The effect of high ocular perfusion pressure (>46 mm Hg) using the generalized estimating equation approach did not significantly enhance prediction of macular edema.

Among the 725 patients, 149 (20.6%; 116 right and 105 left eyes) had retinal hard exudates. Hard exudates could not be graded in 18 patients). Frequency of hard exudates increased significantly with age from 18.9% in those aged 20 to 29 years to 42.4 % in patients 45 years or older (χ²₅ = 70.4, P<.001) (Figure 3). Patients with hard exudates had a significantly longer duration of diabetes than patients without hard exudates(17.7 ± 8.5 years vs 9.5 ± 8.6 years; t₆₉₃ = 10.4, P<.001). Frequency of hard exudates increased from 1% in patients with less than 5 years of diabetes to 22.9% in patients with 10 years of diabetes, and 43.2% in those with 30 years or more (χ²₂ = 47.3, P<.001)(Figure 4).

After adjusting for duration of diabetes, men were on average twice as likely to have severe retinal hard exudates than women (OR, 2.04; 95% CI, 1.27-3.28) (Table 5). Patients with hypertension were on average twice as likely to have severe hard exudates than those who did not have hypertension (OR, 2.24; 95% CI, 1.38-3.65; P <.001), and those in the highest quartile of either systolic or diastolic blood pressure were 2 to 3 times more likely to have severe hard exudates than those in the lowest quartile (OR, 2.31; 95% CI, 1.18-4.51, and OR, 3.78; 95% CI, 1.98-7.23, respectively). Patients with macroproteinuria were 5 times more likely to have severe hard exudates than those without proteinuria(OR, 5.04; 95% CI, 2.69-9.44). Patients in the upper 2 quartiles of glycosylated hemoglobin values were 2 to 3 times more likely to have severe hard exudates than those in the lowest quartile (OR, 2.73; 95% CI, 1.33-5.60, and OR, 2.92; 95% CI, 1.40-6.09, respectively). Also, patients with evidence of peripheral vascular disease were on average 2½ times more likely to have severe hard exudates than those without peripheral vascular disease (OR, 2.40; 95% CI, 1.45-4.0).

There was no association between severity of hard exudates and body mass index, diuretic use, frequency of a diabetic diet and of missing insulin injections, daily insulin dose, education, family income, socioeconomic status, smoking, or alcohol consumption (data not shown).

Severity of hard exudates was significantly associated with presence of either proliferative diabetic retinopathy (χ²₁= 39.23, P<.001) or macular edema (χ²₁ = 252.7, P<.001). After adjusting for duration of diabetes, patients with higher ocular perfusion pressure were on average twice as likely to have severe hard exudates compared with those with lower ocular perfusion pressure (OR, 1.70; 95% CI, 0.97-3.0, and OR, 2.08; 95% CI, 1.11-3.88, for right and left eyes, respectively). There was no association between severity of hard exudates and visual impairment, refractive error, or intraocular pressure (data not shown).

After adjusting for duration of diabetes, patients in the highest quartile of either total or LDL-C levels were on average 5 to 6 times more likely to have severe retinal hard exudates than those in the lowest quartile (OR, 6.75; 95% CI, 2.67-17.07, and OR, 5.30; 95% CI, 2.17-12.95, respectively) (Table 3). Patients with a total cholesterol–HDL-C ratio of 4.5 or more were on average twice as likely to have severe hard exudates compared with those with a ratio less than 4.5 (OR, 2.28; 95% CI, 1.34-3.90(Table 3). There was no significant association between severity of hard exudates and HDL-C.

In the multivariate analyses, variables significantly and independently associated with the severity of hard exudates included presence of macroproteinuria, male sex, higher blood LDL-C levels, and longer duration of diabetes (Table 4). When proteinuria was excluded from the model, male sex, higher LDL-C levels, and longer duration of diabetes remained significantly and independently associated with the severity of hard exudates. Including ocular perfusion pressure in both eyes in the model using the generalized estimating equation approach did not alter the results.

The results of this study indicate that frequencies of macular edema(12.3%) and retinal hard exudates (20.6%) are high in hospitalized African Americans with type 1 diabetes. Presence of macular edema is significantly and independently associated with presence of proteinuria, missing insulin injections at least once a week, and longer duration of diabetes. Severity of hard exudates is significantly and independently associated with presence of macroproteinuria, male sex, higher LDL-C levels, and longer duration of diabetes.

This is the first study of the frequency of macular edema and retinal hard exudates and associated risk factors among a large (N = 725) geographically well-defined cohort of African Americans with type 1 diabetes. In this patient population, frequency of macular edema (12.3%) is similar to the 11.1% reported for whites with type 1 diabetes in the WESDR.³ In the Barbados Eye Study, 41.7% of the 12 patients with younger-onset diabetes had clinically significant macular edema.²⁰ Higher frequencies of macular edema have also been reported in clinic-based studies of mostly white patients with diabetes, but these may not be representative of the diabetic population.^{4 ,21 - 23} African American patients with macular edema are at a higher risk of having other retinal lesions, including proliferative diabetic retinopathy, as previously reported for whites with type 1 diabetes.³ They also are at a higher risk of being visually impaired, consistent with the fact that a large proportion of eyes (53.2% of right and 71.4% of left eyes) with clinically significant macular edema at the time of examination have foveal involvement, and 87% to 88% of them have not received focal laser photocoagulation. In the WESDR in 1980-1982, 54.5% of patients with macular edema present in at least 1 eye at the baseline examination and at the follow-up visit had not received laser photocoagulation.²⁴

In our patient population, macular edema is not seen in those younger than 15 years, but its frequency in patients 45 years or older (33.3%) is higher (but not statistically so) than the 26.1% reported among whites with type 1 diabetes of the same age.³ Macular edema is encountered after a relatively short (4 years) duration of diabetes, peaks to 25.7% for those with 22 to 24 years of diabetes, and then levels off. This again differs from white patients with type 1 diabetes in whom macular edema is rare in the first 10 years of diabetes, but then increases steadily to reach 32% in those with at least 30 years of diabetes.³ In our patients, severity of retinal hard exudates is strongly associated with macular edema and severity of diabetic retinopathy as previously noted.^{4 ,24 - 25} This is expected since both lesions of diabetic retinopathy are considered to be the result of a breakdown of the blood-retinal barrier.²⁶ Like macular edema, retinal hard exudates are encountered in African American patients with a short (5 years) duration of diabetes.

In African Americans, the presence of proteinuria is strongly associated with macular edema, an association we also reported in the same group of patients in relation to proliferative diabetic retinopathy.²⁷ This association is also found in cross-sectional, but not incidence, data from the WESDR.^{3 ,24 - 25} It may reflect coexisting advanced microangiopathy or represent fluid overload, since renal transplantation and dialysis have been shown to reverse macular edema in diabetic patients with proteinuria.^{28 - 29} In our patients, proteinuria is also strongly associated with severe retinal hard exudates. African American patients with either macular edema or severe hard exudates should be referred for evaluation of kidney function early on so that they may benefit from therapies directed at improving renal function.³⁰

In the present study, male sex is a strong risk factor for severe retinal hard exudates. Men who have proteinuria are also at a higher risk for macular edema. This association with male sex is consistent with our previously reported data indicating that, among African Americans with type 1 diabetes, men are a higher risk for both hypertension and renal disease than women, as well as with data from other studies in white patients with diabetes.^{3 ,27 ,31}

In our patients, missing insulin injections at least once a week is the only variable indicating that poor glycemic control is associated with macular edema. Specifically, there is no "threshold" glycosylated hemoglobin value for either macular edema or severe hard exudates. This differs from other studies in which higher glycosylated hemoglobin values are found in patients with macular edema and may be due to the fact that 90% of our patients have poor glycemic control or that factors other than poor glycemic control, such as longer duration of diabetes or presence of renal disease, are more important determinants of macular edema in our cohort.^{3 ,24 - 25 ,31}

Among our patients, those with high LDL-C levels are at a high risk of having macular edema and severe retinal hard exudates, as found in white patients with type 1 diabetes.^{5 - 6 ,32 - 33} While these diabetic lesions result from leakage of plasma and lipids from retinal capillaries, lipid abnormalities in diabetic patients are also thought to contribute to vascular endothelial damage.³⁴ The relationship between LDL-C values and both macular edema and hard exudates is of practical importance since lowering lipids either through diet or pharmacological agents has a beneficial effect not only in preventing development of hard exudates but also in reducing hard exudates already present, although this is not always associated with improvement in visual acuity.^{35 - 38} In our patient population, total and HDL-C levels are high compared with previous reports in white diabetic patients, and HDL-C levels are also higher than in African Americans without diabetes.^{39 - 40} However, neither total nor HDL-C levels are significantly associated with either macular edema or severe hard exudates on multiple regression analysis, as previously found in whites with type 1 diabetes in relation to hard exudates.⁵

In our patients, frequency of macular edema and hard exudates may have been underestimated because these changes were graded from stereoscopic fundus photographs, which have been shown to be less sensitive than slitlamp biomicroscopy in evaluating macular edema.⁴¹ They may also have been overestimated since patients were recruited from among hospital admissions and may thus have more severe diabetes than the diabetic population at large. Finally, results of this study may have been affected by selective mortality of patients with renal disease, leading us not to find previously reported associations.⁴²

In summary, our data indicate that in African Americans with type 1 diabetes macular edema and retinal hard exudates occur early on in the disease and are strongly associated with the presence of renal disease and longer duration of diabetes and to a lesser degree with male sex and higher LDL-C values. Longitudinal studies, however, are needed to further clarify the relationship between these lesions of diabetic retinopathy and presence of renal disease in this patient population.

Accepted for publication June 16, 2000.

This study was supported by grant RO1 EY09860 from the National Eye Institute, Bethesda, Md.

We thank Michael Parides, PhD, Division of Biostatistics, Joseph L. Mailman, School of Public Health, Columbia University, New York, NY, for statistical help; Lisa Schoenherr, study coordinator, for technical assistance; Robert Nelson and Jim Besra, MPH, research assistants; Clara Baker for secretarial help; the New Jersey Department of Health; and the 31 New Jersey hospitals that participated in the study.

Corresponding author and reprints: Monique S. Roy, MD, University of Medicine and Dentistry, New Jersey Medical School, Department of Ophthalmology, 90 Bergen St, Room 6164, Newark, NJ 07103.