Six-Year Progression of Retinopathy and Associated Risk Factors in African American Patients With Type 1 Diabetes Mellitus: The New Jersey 725

Objective  To report the 6-year progression of diabetic retinopathy (DR) and associated risk factors among African American patients with type 1 (insulin-dependent) diabetes mellitus.

Methods  Participants from the New Jersey 725 included 483 African American patients with type 1 diabetes who underwent reexamination as part of a 6-year follow-up. Evaluations included a structured clinical interview, ocular examination, 7 stereoscopic fundus photographs, and blood pressure measurements. Severity of DR was determined via masked grading of fundus photographs. Biological evaluation included blood and urine assays.

Results  During the 6-year period, 56.1% of patients at risk showed progression of DR; 15.0% showed progression to proliferative DR; and 15.9% developed macular edema. A baseline high glycosylated hemoglobin level and systemic hypertension were significant risk factors for progression of DR, progression to proliferative DR, and incidence of macular edema. Progression to proliferative DR was significantly associated with baseline older age, renal disease, and severity of DR. The incidence of macular edema was significantly associated with baseline older age, low socioeconomic status, severity of DR, and total serum cholesterol level.

Conclusions  Six-year progression of DR is high in African American patients with type 1 diabetes. Improving glycemic and blood pressure control may reduce the ocular morbidity of diabetes in African Americans.

Diabetic retinopathy (DR) is the leading cause of new cases of legal blindness in Americans aged 20 to 64 years, despite effective treatments to prevent visual loss or to restore useful vision.^1-6 Although the incidence of and risk factors for progression of DR have been reported for white patients with type 1 diabetes mellitus, to our knowledge, few data are available in African American patients with type 1 diabetes.^7-20 We previously reported risk factors for DR in the New Jersey 725 cohort of African Americans diagnosed as having type 1 diabetes before 30 years of age.^21,22 We also reported that a high proportion of these African Americans had poor glycemic control and uncontrolled hypertension, which are the 2 main risk factors for progression of DR in white populations with type 1 diabetes.^{6,10,14,17,21} Our large New Jersey 725 cohort provides an opportunity to examine the progression of DR in African American patients with type 1 diabetes. Therefore, we performed a 6-year follow-up study and examined the association of baseline risk factors with progression of DR.

The original cohort consisted of 725 African Americans with type 1 diabetes mellitus who participated in the New Jersey 725 study from August 25, 1993, through January 6, 1998.²¹ Patients diagnosed as having diabetes, who were treated with insulin before 30 years of age, and who were currently receiving insulin were identified from a random review of 13 615 medical records.²¹ We excluded patients with type 2 diabetes, those diagnosed after age 30 years, and patients with maturity-onset diabetes of youth.^23,24 Of the 875 eligible patients, 725 participated in the baseline examination.

Of the original cohort of 725 patients, 508 participated in the 6-year follow-up examination. At follow-up, 25 (4.9%) of the 508 participants were no longer receiving insulin and had not received a pancreas transplant. Compared with the mean ± SD values in those receiving insulin, patients not receiving insulin had at baseline a shorter duration of diabetes (7.0 ± 6.5 vs 10.4 ± 8.6 years; P<.02), lower glycosylated hemoglobin levels (10.9% ± 4.7% vs 13.5% ± 4.3%; P<.01), and higher C-peptide levels (2.91 ± 1.69 vs 1.06 ± 1.16 ng/mL; P<.001), and more of them had a body mass index (calculated as weight in kilograms divided by the square of height in meters) of 25 or higher (88.0% vs 57.1%; P<.002). There were no significant differences at baseline between patients receiving and those not receiving insulin for sex, age, DR severity, visual impairment, systemic hypertension, macroangiopathy, and renal disease. Because these 25 patients may not be truly insulin dependent, they have been excluded, leaving 483 (95.1%) of the 508 patients available for analysis.

Of the 725 patients, 44 (6.1%) could not be located, 34 (4.7%) refused examination, and 139 (19.2%) had died during the 6-year interval. Compared with the 483 participants, patients who refused to participate or could not be located were more likely at baseline to be working (35.2% vs 47.4%; P = .04) and had been smoking for a shorter time (mean ± SD, 1.8 ± 4.4 vs 3.5 ± 8.3 pack- years; P=.02) (Table 1). Compared with participants, deceased patients were more likely to be older at baseline and at diagnosis of diabetes, to have smoked more, to be unemployed, and to have a longer duration of diabetes, high blood pressure, a lower family income, a lower socioeconomic status, less education, severe DR, macroangiopathy, proteinuria, and a lower body mass index (Table 1). The mean ± SD duration of follow-up was 6.1 ± 0.5 years (median follow-up, 5.96 years).

Patients were examined in the Institute of Ophthalmology and Visual Science, University of Medicine and Dentistry, in Newark, NJ. On arrival, informed written consent was obtained. Patients underwent a complete eye examination including best corrected visual acuity using the Early Treatment of Diabetic Retinopathy Study (ETDRS) protocol, measurement of intraocular pressure by means of applanation, examination of the iris for neovascularization, dilated retinal examination, and 7 standard stereoscopic Diabetic Retinopathy Study retinal photographs and red reflex lens photographs of each eye using a 30° fundus camera (Carl Zeiss, Thornwood, NY).^25,26 Also obtained were height and weight. Blood pressure was measured twice in the sitting and standing positions using a random-zero sphygmomanometer according to the Hypertension Detection and Follow-up Program protocol.²⁷ The averages of the 2 measurements in each position were used. A structured clinical interview included detailed medical and ophthalmologic histories, sociodemographic factors, and lifestyle variables (ie, self-reported measures of cigarette smoking, alcohol consumption, and illicit drug abuse).

Venous blood was drawn for measurement of total glycosylated hemoglobin level using high-pressure liquid chromatography (LabCorp, Bio-Rad Laboratory, Hercules, Calif) and for measurement of high- and low-density lipoprotein and total cholesterol levels using an enzymatic assay and separation spectrophotometry (Genzyme Diagnostics, Cambridge, Mass). The reference range for total glycosylated hemoglobin level is 4.2% to 7.0% and the intra-assay coefficient of variation is 0.38% to 1.47%. A 4-hour timed urine collection was obtained for measurement of the albumin excretion rate and creatinuria using spectrophotometry (SmithKline Beecham Clinical Laboratory, Philadelphia, Pa). The institutional review board of the University of Medicine and Dentistry, New Jersey Medical School, Newark, approved the study.

Color fundus photographs obtained at baseline and follow-up were graded for DR severity in a masked fashion by the Wisconsin Fundus Photograph Reading Center in Madison. The modified ETDRS Airlie House classification of DR was used^28,29: level 10 indicates no DR; levels 20 to 53, nonproliferative DR of increasing severity; and levels 61 to 85, proliferative DR (PDR) of increasing severity. For each eye, the maximum grade in any of the 7 standard photographic fields was used to define the retinopathy level according to the ETDRS severity scale.²⁹ Subsequently, the retinopathy level for a participant was determined using the severity levels in the right and left eyes, giving greater weight to the worse eye and using a 13-step scale similar to that previously described by Klein et al¹⁰ (10/10, 20/<20, 20/20, 35/<35, 35/35, 43/<43, 43/43, 47/<47, 47/47, 53/<53, 53/53, ≥61/<61, and ≥61/≥61), with all levels of PDR grouped as a single level. If the retinopathy severity could not be graded in one eye, the subject was considered to have a score equivalent to that in the gradable eye.

Macular edema (ME) was considered present if there was thickening of the retina with or without partial loss of retinal transparency within 1 disc diameter from the center of the macula and/or if there were focal laser photocoagulation scars in the macular area and a documented history of ME. Clinically significant macular edema (CSME) was defined as any of the following: thickening of the retina at or within 500 μm of the center of the macula; hard exudates at or within 500 μm of the center of the macula associated with thickening of the adjacent retina; an area of retinal thickening 1 disc diameter or larger, any part of which was within 1 disc diameter of the center of the macula; or focal laser photocoagulation scars of the macular area with a documented history of CSME.³⁰ If CSME or ME could not be graded in one eye, the participant was assigned the score of the other eye.

Eyes that could not be graded—because of opacities of the media, phthisis, or enucleation—were initially classified as “cannot grade.” For such persons, a review of all previous medical records was performed by one of us (M.S.R.). When a history of panretinal photocoagulation for PDR or pars plana vitrectomy for complications of PDR was documented by medical chart review, then the DR level was scored as 85. Persons who had an ETDRS grading less than 61 at the time of examination and had previously received laser photocoagulation for PDR (as documented by medical chart review) received a grade of 61.

The 6-year incidence of any DR was calculated from all patients who at baseline had no DR (severity level, 10/10). Patients who developed any DR are those of this group with progression to a severity level of 20/<20 or higher at follow-up. Progression to PDR was calculated for all patients who had no DR or nonproliferative DR at baseline (levels 10/10 through 53/53). Patients with progression to PDR were those of this group who had progression at follow-up to PDR (level ≥61/<61 or worse). Progression to PDR with high-risk characteristics (HRCs) was calculated for all patients who had no DR, nonproliferative DR, or PDR without HRCs at baseline (levels 10/10 through 61/61). Patients with progression to PDR with HRCs were those of this group with progression at follow-up to PDR and HRCs (level ≥71/<71 or worse). Progression of DR was calculated for all patients who had no DR or nonproliferative DR at baseline (levels 10/10 through 53/53). Patients who showed progression of DR were those of this group with progression of at least 2 steps at the follow-up examination (from 10/10 to 20/20 or higher, 20/<20 to 35/<35 or higher, etc).¹⁰ Improvement of DR was calculated for patients with a baseline DR severity level of 20/20 or higher who had not received laser photocoagulation. Patients who improved were those of this group in whom DR severity decreased by at least 2 steps at follow-up (from 20/20 to 10/10, etc). When examining 2-step progression or improvement, patients with PDR at baseline (level ≥61/<61 or higher) were excluded. Progression to vision-threatening DR was calculated in patients who had no DR or nonproliferative DR and no CSME at baseline (levels 10/10 through 53/53). Patients with progression to vision-threatening DR were those of this group with progression to PDR (level ≥61/<61 or worse) or CSME at follow-up. The 6-year incidence of ME (or CSME) was calculated for all patients who had no ME (or CSME) or had not received focal laser photocoagulation for ME (or CSME) in either eye at baseline. Patients who developed ME (or CSME) were those of this group who developed ME (or CSME) in at least 1 eye at follow-up or who had received focal laser photocoagulation for ME in either eye since the baseline examination.

Patient age was defined as the age at the time of the baseline examination. Age at diagnosis of diabetes was defined as the age at which the diagnosis of diabetes was first recorded by a physician in the patient's hospital record. Duration of diabetes was the time from the age at diagnosis to the age at baseline. Systemic hypertension was defined as present if at baseline the systolic blood pressure was 140 mm Hg or higher, the diastolic blood pressure was 90 mm Hg or higher, or the patient was taking antihypertensive medication. We defined microproteinuria as being present if the baseline albumin excretion rate was 20 to 200 μg/min, and overt proteinuria as being present if the baseline albumin excretion rate was more than 200 μg/min. Macroangiopathy was considered present if at baseline (1) the patient reported having undergone foot or leg amputation for a circulatory problem (excluding amputation secondary to an infection) or having had a myocardial infarction or a stroke and (2) this diagnosis was confirmed using standardized criteria by review of the medical records of all previous hospital admissions.

Socioeconomic factors recorded included the patient's level of education (for those >25 years old), personal income (for those >18 years old), and family income. Socioeconomic status was classified according to the method of Goldthorpe and Hope³¹ for the social grading of occupations. Smoking was defined as pack-years smoked by dividing the number of cigarettes smoked per day by 20, multiplied by the number of years smoked until the baseline examination.

Statistical analyses were performed using SAS statistical software, version 8.01 (SAS Institute Inc, Cary, NC). We calculated incidence rates with 95% confidence intervals (CIs) for any DR, progression of DR, progression to PDR with and without HRCs, the incidence of ME and CSME, and the incidence of improvement or no change separately for men and women.

The relationships between progression of DR and age, duration of diabetes, or DR severity level at baseline were examined using the Cochran-Armitage trend test. We calculated P values of the Cochran-Mantel-Haenszel statistics for the association of progression of DR with age (after controlling for duration of diabetes at baseline) or with duration of diabetes (after controlling for age at baseline).

We applied univariate logistic regression modeling to selected variables to estimate odds ratio (ORs) and 95% CIs to predict the following 3 end points: progression of DR, progression to PDR, and incidence of ME. Factors with P<.05 from the univariate analyses were included in the multiple logistic regression models to examine independent associations between risk factors and progression of DR using forward and backward selection.

At the 6-year examination, 72.3% of the patients at risk for incidence of DR had developed any DR,56.1% showed progression of DR, 15.0% had progression to PDR, 5.1% had progression to PDR with HRCs, 15.9% had developed ME, 13.0% had developed CSME, 21.6% had progression to vision-threatening DR, 6.9% had improved, and 42.0% showed no change in severity of DR (Table 2). There were no significant sex differences for any category of progression of DR (Table 2).

Both progression to PDR and incidence of ME increased significantly with increasing age at baseline: for PDR from 2.6% to 35.3%, and for ME from 7.7% to 32.0% in patients 10 to 14 years of age and those 45 years or older at baseline, respectively (Table 3). No patient younger than 10 years at baseline had progression to PDR or developed ME. After controlling for duration of diabetes at baseline, the incidence of ME remained significantly associated with age at baseline (tests for general association, P<.001) (data not shown). There was no significant association between progression of DR and age at baseline (test for trend, P = .7), although the highest progression rate of DR (61.5%) was observed in patients aged 10 to 14 years at baseline (Table 3).

Progression of DR and progression to PDR were significantly and directly associated with duration of diabetes at baseline (test for trend, P = .02 and P<.001, respectively) (Table 3). The highest rates of progression of DR (68.5%) and progression to PDR (27.8%) were observed for patients with 15 to 19 years of diabetes duration at baseline, and rates declined thereafter.

After controlling for age at baseline, progression of DR, progression to PDR, and incidence of ME were all significantly but not linearly associated with duration of diabetes at baseline (tests for general association, P = .01, P = .004, and P<.001, respectively) (Table 4).

Progression to PDR and incidence of ME were significantly associated with severity of DR at baseline (P<.001 and P = .004, respectively) (Table 5). Among patients with moderate nonproliferative DR in the worse eye at baseline (ETDRS level, 43-53), 32 (54.2%) progressed to PDR, 14 (23.7%) to PDR with HRC, and 18 (22.8%) of those with ETDRS levels of 43 to 61 or higher developed ME during the 6-year follow-up (Table 5).

Baseline Glycosylated Hemoglobin. Patients in the upper 2 quartiles of glycosylated hemoglobin values at baseline had 3 to 8 times the odds for progression of DR, 8 to 20 times the odds for progression to PDR, and 4 to 7 times the odds for developing ME at the 6-year follow-up than patients in the lowest quartile (Table 6). The strongest association was seen in relation to progression to PDR. Patients in the lowest quartile of glycosylated hemoglobin values at baseline had 9 times the odds for improving than patients in the upper 3 quartiles (OR, 9.33; 95% CI, 2.71-32.12).

Baseline Systemic Hypertension and Proteinuria. Patients with baseline systemic hypertension were more likely to have progressed to PDR or to have developed ME at the 6-year follow-up than patients without baseline hypertension (OR, 3.68 [95% CI, 2.11-6.42; P<.001]; and OR, 2.53 [95% CI, 1.50-4.27; P<.001], respectively). Also, patients in the upper quartile of systolic blood pressure and those in the upper 2 quartiles of diastolic blood pressure at baseline were more likely to show progression to PDR or ME than those in the lowest quartile (Table 6). Patients with untreated systemic hypertension at baseline were more likely to show progression than patients without hypertension (OR, 1.77 [95% CI, 1.08-2.89; P<.007]). Patients with microproteinuria or overt proteinuria at baseline had 3 to 5 times the odds for progression to PDR than patients with no baseline proteinuria (Table 6).

Other Baseline Risk Factors. Patients who were 13 years or older at diagnosis of diabetes had twice the odds for progression of DR than patients younger than 13 years at diagnosis (OR, 2.20 [95% CI, 1.12-4.33]). None of the patients with a diagnosis before 13 years of age showed progression to PDR or developed ME, while of those with a diagnosis at 13 years or older, 63 (16.6%) showed progression to PDR and 69 (17.5%) developed ME during the 6-year follow-up (P<.006 and P<.004, respectively).

Patients with low baseline socioeconomic status had 2 times the odds for developing ME than patients with middle to high socioeconomic status (Table 6). Patients in the upper 2 quartiles of total serum cholesterol values at baseline had 3 to 4 times the odds for progression to PDR and 3 to 5 times the odds for developing ME than patients in the lowest quartile. Patients in the upper quartile of low-density lipoprotein cholesterol levels at baseline had 3 times the odds for progression to PDR and 4 times the odds for developing ME at the 6-year follow-up than patients in the lowest quartile (Table 6).

There was no association between progression of DR and smoking or use of diuretics, angiotensin-converting enzyme inhibitors, or statin medications (data not shown).

To evaluate the relative role of the various risk factors in relation to progression of DR, we used multiple logistic regression models. Baseline glycosylated hemoglobin values were significantly and independently associated with progression of DR, progression to PDR, and incidence of ME (all, P<.001) (Table 7). Baseline systemic hypertension was significantly and independently associated with progression to PDR and incidence of ME (P = .001 and P = .03, respectively). When systolic and diastolic blood pressure were included in the model instead of systemic hypertension, patients in the upper 2 quartiles of diastolic blood pressure at baseline were more likely to have progression to PDR than those in the lowest quartile, independently of any other factor (OR, 2.50 [95% CI, 1.04-6.00] and OR, 3.00 [95% CI, 1.30-7.00], respectively) (data not shown). Patients with baseline proteinuria had 2 to 4 times the odds for progression to PDR than patients with no proteinuria at baseline (P = .009) (Table 7).

Older age at baseline was significantly and independently associated with progression to PDR and incidence of ME. Patients 13 years or older at diagnosis of diabetes had twice the odds for progression of DR than patients younger than 13 years at diagnosis (P=.03). Patients in the upper quartile of total serum cholesterol level and those with low socioeconomic status at baseline were more likely to develop ME at the 6-year follow-up than patients in the lowest quartile of total cholesterol level or with middle to high socioeconomic status. When baseline severity of DR was included in the model instead of age, patients with baseline moderate nonproliferative DR (ETDRS levels 43-53) had on average 71 times the odds for progression to PDR and twice the odds for developing ME at the 6-year follow-up than patients with baseline ETDRS levels of 10 to 15 (OR, 70.85 [95% CI, 17.66-284.22; P<.001] and OR, 2.42 [95% CI, 1.04-5.65; P<.001], respectively).

This is a large data set about the natural history of DR in African American patients with type 1 diabetes mellitus. The results show that during the 6-year follow-up, 72.3% of the patients at risk for incidence of DR developed any DR, 56.1% showed progression of DR, 15.0% showed progression to PDR, and 15.9% developed ME.

In the Wisconsin Epidemiologic Study of Diabetic Retinopathy, the 4-year progression of DR among white patients with type 1 diabetes was high, at 41.2%, 10.5%, and 8.2% for progression of DR, progression to PDR, and incidence of ME, respectively.^10,12 Our slightly longer 6-year data show that progression of DR in African American patients with type 1 diabetes is also very high, at 56.1%, 15.0%, and 15.9%, respectively. The 6-year incidence of ME was 27.2% for patients with 5 to 10 years of diabetes duration and 8.4% for those with less than 5 years of diabetes at baseline compared with 7.8% and 2.5%, respectively, in the Wisconsin Epidemiologic Study of Diabetic Retinopathy 4-year follow-up.¹²

Poor glycemic control is a well-documented risk factor for progression of DR in persons with type 1 diabetes.^{8,9,11-20,32,33} In our African American patients, it is strongly associated with progression of DR. Patients in the upper quartile of glycosylated hemoglobin values at baseline had 20.12 (95% CI, 4.67-86.66) times the odds of progression to PDR at the 6-year follow-up than patients in the lowest quartile. This is particularly noteworthy because our patients had poor glycemic control at baseline.²² Furthermore, although patients' glycosylated hemoglobin values at the second visit were significantly lower than at the first visit (12.2% ± 3.3% vs 13.5% ± 4.3% [t₄₇₇ = 6.83; P<.001]), these values were still high compared with reference levels (4.2%-7.0% in our laboratory). Like previous reports, there was no evidence that a particular level of glycosylated hemoglobin at baseline led to a significant increase in the progression or significant improvement of DR at follow-up.^17,18,33 However, it is noteworthy that 80% of study patients with baseline glycosylated hemoglobin levels of 16.2% or higher showed progression of DR during the 6-year follow-up.

Another risk factor for progression to PDR or incidence of ME in white patients with type 1 diabetes is systolic or diastolic hypertension.^{11-13,15,17,18,20} In this study, high diastolic blood pressure was significantly associated only with progression to PDR. At follow-up, the prevalence of hypertension among our African American patients was high (44.4%). Furthermore, 25.5% of the 212 patients with hypertension were not receiving antihypertensive medications, whereas in 49.4% of those receiving antihypertensive medications, blood pressure was not controlled.

Progression of DR and progression to PDR have been previously shown to increase in a nonlinear fashion with duration of diabetes at baseline.^{9,10,14,17,20} Similarly, in our African American patients, progression of DR (68.5%) and progression to PDR (27.8%) were highest for those with a relatively short duration (15-19 years) of diabetes at baseline. The decline thereafter may have been due to selective mortality of patients with severe diabetic complications, as we have previously found in this African American cohort.³⁴

Studies have also shown that progression of DR increases with severity of DR at baseline.^{10,11,14,17,35} For example, in the ETDRS, 21.0% of mostly white patients with type 1 diabetes and moderate nonproliferative DR (ETDRS level 43) at the initial evaluation showed progression to PDR in 5 years compared with 11.0% of patients with mild nonproliferative DR (ETDRS level 35).²⁹ Also, in the Wisconsin Epidemiologic Study of Diabetic Retinopathy, Klein et al¹⁷ reported a 14-year 57.4% progression to PDR in patients with a similar DR severity in the worse eye compared with 31.7% of patients with an ETDRS level of 21. We found that 46.5% of our African American patients with a similar ETDRS level of 43 showed progression to PDR during 6 years of follow-up. This suggests the need for careful follow-up retinal examinations in such patients to detect DR requiring treatment.

In the Wisconsin Epidemiologic Study of Diabetic Retinopathy 4-year follow-up, progression to PDR increased significantly with age at baseline, reaching 15.8% in patients aged 14 to 30 years, and leveled off thereafter.¹⁰ This finding differs from that in our African American patients, in whom progression to PDR increased steadily with age at baseline to 35.3% in those older than 45 years. However, as in the study by Klein et al,¹⁰ progression of DR was the highest among our young patients aged 10 to 19 years at baseline, and no African American patient younger than 10 years at baseline developed PDR or ME.^10,14,17 Also in agreement with some but not all previous studies, our African American patients who were diagnosed as having diabetes at 13 years or older were more likely to show DR progression than those who received a diagnosis before that age.^18,36,37

Although proteinuria at baseline was common among the 725 patients (49.8%) and associated with the presence of ME, there was no association between baseline proteinuria and the development of ME among patients participating in the 6-year follow-up, as similarly reported in white patients with type 1 diabetes.^12,38,39 It is possible that ME may have resolved as patients with renal disease received diuretics or underwent dialysis.⁴⁰ Another possible explanation is selective mortality of patients with diabetic kidney disease, because we have previously shown that baseline proteinuria is a strong predictor of death.³⁴ Of the 139 study patients who died before the follow-up reexamination, at baseline 89 (64.0%) already had overt proteinuria and 28 (20.1%) had microproteinuria.

Low socioeconomic status has been linked to increased morbidity and mortality in a number of diseases, including diabetes, although its association with progression of DR has so far not been clearly established.^41-43 In our African American patients, low socioeconomic status was significantly associated with incidence of ME. However, education, income, medical or eye care, and health insurance at baseline were not significantly different between patients with and without ME at follow-up. The association between baseline high serum cholesterol levels and incidence of ME in our patients lends support to the possible role of lipid abnormalities in vascular endothelial damage in diabetes.⁴⁴

In summary, the results show that progression of DR in African American patients with type 1 diabetes is high. Poor glycemic control and systemic hypertension are 2 modifiable risk factors significantly associated with progression of DR. Because glycemic and blood pressure control in this population are poor, measures to improve medical care and ensure regular dilated eye examination to detect vision-threatening DR may reduce morbidity from the disease.

Correspondence: Monique S. Roy, MD, Department of Ophthalmology, University of Medicine and Dentistry, New Jersey Medical School, 90 Bergen St, Room 6164, Newark, NJ 07101-1709 (Roymo@umdnj.edu).

Submitted for Publication: September 23, 2005; final revision received April 6, 2006; accepted April 23, 2006.

Financial Disclosure: None reported.

Funding/Support: This research was supported by grant RO1 EY 09860 from the National Eye Institute and a Lew Wasserman Merit Award from Research to Prevent Blindness, Inc.