Hyperglycemia, Blood Pressure, and the 9-Year Incidence of Diabetic Retinopathy

Transcription

1 Hyperglycemia, Blood Pressure, and the 9-Year Incidence of Diabetic Retinopathy The Barbados Eye Studies M. Cristina Leske, MD, MPH, 1 Suh-Yuh Wu, MA, 1 Anselm Hennis, MRCP (UK), PhD, 2,4 Leslie Hyman, PhD, 1 Barbara Nemesure, PhD, 1 Ling Yang, MS, 1 Andrew P. Schachat, MD, 3 for the Barbados Eye Study Group* Objectives: To evaluate factors related to the incidence of diabetic retinopathy (DR) in a population of African descent, after 9 years of follow-up. Design: Population-based cohort study; 81% participation after 9 years. Participants: Three hundred twenty-four participants of the Barbados Eye Studies, with diabetes mellitus (DM) at baseline and at risk for developing DR during follow-up. Methods: Diabetes-related changes were assessed by masked gradings of baseline and follow-up photographs using a standardized system. The 9-year cumulative incidence of DR was based on participants with DM and free of retinopathy at baseline; incidence rates were estimated by the product-limit approach. Cox regression models for discrete-time data were used to evaluate risk factors associated with the 9-year incidence of DR. Results: Multivariate analyses revealed that older age at DM onset decreased the 9-year risk of DR development; for each 10 years of older age at onset, the risk of DR decreased by 30% (risk ratio [RR], 0.7; 95% confidence interval [CI], ). The risk of DR doubled among persons with DM duration between 5 and 9 years (RR, 2.1; 95% CI, ) versus those with shorter durations; it also doubled in those treated with oral medications or insulin at baseline versus those treated with diet only. Antihypertensive treatment halved the risk of DR versus no treatment (RR, 0.5; 95% CI, ) and high systolic or diastolic blood pressure (BP) increased risk. Thus, DR risk increased by 30% for every 10 mmhg of higher systolic BP at baseline (RR, 1.3; 95% CI, ) or of BP increase from baseline to the 4-year follow-up (RR, 1.3; 95% CI, ). Diabetic retinopathy risk similarly increased with each 1% of higher glycosylated hemoglobin level at baseline (RR, 1.3; 95% CI, ). Conclusions: The long-term follow-up of persons with DM in this population of African origin, where disease prevalence is high, identified important potentially modifiable risk factors for DR. Findings suggest that efforts to achieve optimal glycemic and BP control may reduce the vision-threatening complications of DM. Ophthalmology 2005;112: by the American Academy of Ophthalmology. Diabetic retinopathy (DR) is an ocular complication of diabetes mellitus (DM) that remains a leading cause of vision loss. Although results have varied, observational Originally received: August 10, Accepted: November 24, Manuscript no Department of Preventive Medicine, Stony Brook University, Stony Brook, New York. 2 Ministry of Health, Bridgetown, Barbados. 3 Wilmer Eye Institute, The Johns Hopkins University School of Medicine, Baltimore, Maryland. 4 Chronic Disease Research Center, School of Clinical Medicine and Research, University of the West Indies, Bridgetown, Barbados. Supported by the National Eye Institute, Bethesda, Maryland (grant nos.: EY07625, EY07617). Correspondence and reprint requests to M. Cristina Leske, MD, MPH, Department of Preventive Medicine, Stony Brook University, HSC L3 086, Stony Brook, NY cleske@notes.cc.sunysb.edu. *See Appendix. studies have identified several risk factors for DR, including hyperglycemia and high blood pressure (BP). Intervention trials have demonstrated the effectiveness of photocoagulation treatment, 1,2 and glycemic control has been shown to reduce the risk of DR. 3,4 Similar risk reductions from BP control, however, have not been reported consistently. 5 7 Because findings on the risk of DR were based mainly on European-derived populations, their general applicability in other settings is largely unknown. The identification of risk factors for DR is of particular relevance for populations of African origin, who have substantially higher prevalences of DM and hypertension. Among the African-descent participants of the Barbados Eye Studies, approximately 18% had a history of DM, 8 and 55% had hypertension. 9 Furthermore, 30% of persons with DM, mainly type 2, had DR at baseline, 10 a prevalence that increased to 40% at 4 years and remained at similar levels at 9 years, documenting the public health impact of the disease in the population. We have previously reported 2005 by the American Academy of Ophthalmology ISSN /05/$ see front matter Published by Elsevier Inc. doi: /j.ophtha

2 Ophthalmology Volume 112, Number 5, May and 9-year incidence estimates of DR in the cohort. 11,12 The present report evaluates the factors related to the 9-year incidence of DR, thus providing new data on characteristics that affect the long-term risk of DR in a population with the same ancestral origin as African Americans. Materials and Methods Background The Barbados Eye Studies ( ), funded by the National Eye Institute, are epidemiologic investigations of the prevalence, incidence, and risk factors for major eye diseases, including DR, in a predominantly African-origin population The baseline prevalence study, the Barbados Eye Study ( ), was based on a simple random sample of Barbadian-born citizens, years old (4709 persons; 84% participation; 93% black by selfreport). 13 The surviving cohort was invited for reexaminations to determine the incidence and progression of these eye diseases after 4 years (85% participation) and 9 years (81% participation). 14,15 Protocols for Diabetic Retinopathy The study protocols to determine the incidence of DR and the definitions used have been reported previously, 11,12 and a summary follows. Data Collection. Standardized protocols at baseline 13 and both follow-up visits 14,15 included an extensive interview to determine risk factors and various ocular, anthropometric, and BP measurements, as well as a comprehensive ophthalmologic examination with dilatation for a systematic 10% sample and participants with positive screening findings (e.g., history of ocular disease or diabetes, intraocular pressure 21 mmhg, best-corrected visual acuity 20/30). Assays of total glycosylated hemoglobin (GHb) using affinity chromatography of venous whole blood 16 by Glyc-Affin GHb kits (Isolab, Akron, OH) were available for 3754 (81%) of the participants (GHb was not measured at the beginning of the study), with good reproducibility shown by duplicate testing of a random sample (n 264; intraclass correlation coefficient, 0.89). No significant differences were found between participants with and without data on GHb with respect to baseline age, gender, BP, and DM duration. The protocol included 30 color stereo fundus photography of the disc and macula (standard fields 1 and 2 of the Diabetic Retinopathy Study 17 ), which was graded at the Fundus Photography Reading Center 18 using an adapted version of the modified Early Treatment Diabetic Retinopathy Study Airlie House classification. 17,19,20 All photographs were classified independently by 2 graders; discrepancies were resolved by consensus or adjudication by the study retinal specialist (AS). Similar to the results obtained at baseline 10 and the 4-year examination, 14 good intergrader agreement for diabetic changes was achieved at 9 years (field 1: ; field 2: ), with no evidence of grading drift over time. Definitions 1. Diabetes mellitus: self-reported history of physician-diagnosed DM and/or GHb levels of 10% ( 2 standard deviations [SDs] above the population mean of persons without a diabetes history [mean, 7%; SD, 1.5%] at baseline). 2. Younger-onset DM: diagnosis at age of 30 years; participants were considered to have type 1 DM if they were also receiving insulin. 3. Older-onset DM: diagnosis at age of 30 years or new diagnosis in the study (no DM history, but GHb 10%). 4. Diabetic retinopathy: diabetic changes in persons with DM. These changes included at least 3 microaneurysms, retinal hemorrhages, hard and soft exudates, intraretinal microvascular abnormalities, new vessels within 1 disc diameter of the disc, and new vessels originating elsewhere, as well as other abnormalities including venous beading, focal narrowing, venous loops, and clinically significant macular edema. 5. Nine-year cumulative incidence of DR: development of retinopathy in at least one eye by the 9-year follow-up among persons with DM and free of retinopathy at baseline. The product-limit approach 21 was used to estimate the incidence rates, thus allowing the use of data from persons with 4-year follow-up only. Risk Factors Persons with preexisting DR were not included in the risk factor evaluations. These analyses were based on participants of African descent only, given the small size of the other groups. Possible risk factors included age, gender, education, and lifetime occupation; body mass index (weight in kilograms/height in meters, squared) and waist/hip ratio; systolic BP (SBP) and diastolic BP (DBP) (measured twice with the Hawksley random zero sphygmomanometer, and the average used in the analysis); hypertension (SBP 140 mmhg, DBP 90 mmhg, and/or antihypertensive treatment); smoking and alcohol use; age of onset of DM, as well as its duration and treatment; self-reported family history of diabetes; and GHb levels. Changes in BP and GHb after 4 years were also investigated as risk factors for 9-year incidence, being defined as the respective BP or GHb value at the 4-year follow-up examination minus the value at baseline. Potential risk factors were first evaluated individually in a Cox regression model for discrete-time data. 22 Factors with a P value of 0.1 from the univariate investigation were included simultaneously in a multivariate model. Results were presented as risk ratios (RRs) and corresponding 95% confidence intervals (CIs) based on the Cox regression model. Results This report is based on 324 cohort participants with DM at baseline, who were at risk for developing DR during the 9 years of follow-up. 12 Their median age was 58 years, and 64% were female. Table 1 presents the 9-year incidence of DR in this group by various baseline features. Incidence rates did not vary by age, gender, and other characteristics presented in the table, except for age of DM onset. Based on the median age of onset, persons diagnosed with DM before 53 years of age were 1.5 times more likely to develop DR than those diagnosed at older ages. Hypertension was frequent in the study cohort, with a prevalence of 55% overall among participants of African origin. 9 Mean ( SD) and median SBPs were 137 ( 23) mmhg and 134 mmhg, and the respective values for DBP were 81 ( 12) mmhg and 80 mmhg; 30% of participants reported antihypertensive treatment, which mainly consisted of diuretics and adrenergic inhibitors. Table 2 describes the relationships between DR and BP, stratifying by the presence/absence of antihypertensive medication usage. A comparison of the first 2 columns of Table 2 indicates that incidence of DR among hypertensives seemed to vary with antihypertensive treatment, with a trend toward decreased risks with treatment. In particular, treated hypertensives with SBP between 140 and 160 mmhg, or DBP higher than 80 mmhg (the largest sample sizes), had markedly lower incidences than the respective untreated groups. 800

3 Leske et al Hyperglycemia, Blood Pressure, and Incident Diabetic Retinopathy Table 1. 9-Year Incidence of Diabetic Retinopathy by Various Baseline Characteristics Status of Diabetes at Baseline No. at Risk Incidence (%) Crude RR (95% Confidence Interval)* Age (yrs) ( ) Gender Male Female ( ) Age of diabetes onset (yrs) ( ) Body mass index (kg/m 2 ) ( ) Education (yrs) ( ) Cigarette smoking Yes ( ) No Alcohol use Yes ( ) No RR risk ratio. *Based on Cox regression model, adjusting for antihypertensive treatment. P As evident from the third column of Table 2, the overall incidence of DR tended to increase as SBP and DBP increased. After adjusting for antihypertensive treatment, univariate evaluations shown in the fourth column of Table 2 yielded either statistically significant or marginally significant (P 0.1) results for the associations of DR with SBP or DBP. As compared to the lowest SBP or DBP categories, these analyses suggested that the risk of DR doubled when SBP was 160 mmhg (RR, 2.0; P 0.10) or DBP was 70 mmhg (RR, 2.1; P 0.5; RR, 2.0; P 0.10). Similarly, analyses that considered BP as a continuous variable suggested that the risk of DR increased by 10% with every 10 mmhg of higher SBP or DBP at baseline (RR, 1.1 per 10 mmhg of higher baseline pressure; P 0.10). Additional univariate analyses found that persons with increases in SBP from baseline to the 4-year follow-up examination had significantly increased incidence of DR (P 0.05, data not shown). Univariate analyses also revealed a positive association between DR risk and the duration of DM. As compared to a DM duration of 5 years, the crude RRs were 2.8 (95% CI, ) for duration 5 9 years, 2.6 (95% CI, ) for duration years, and 1.3 (95% CI, ) for duration 15 years (data not shown). In addition, the use of any insulin or oral DM treatment was significantly (P 0.05) associated with 9-year incidence. No other factors investigated reached P 0.1 in the univariate analyses. Table 3 presents the results of multivariate analyses that include all the factors listed. The top panel of the table shows the main model, which includes SBP-related variables; the lower panel presents a model with DBP-related variables. The main model results indicate that an older onset of DM decreased the risk of DR development. In fact, for each 10 years of older age at DM onset, the risk of DR decreased by 30% (RR, 0.7). The incidence of DR was also related to the duration of DM, as was found in the univariate analyses. Persons with DM duration between 5 and 9 years had a 2-fold increase in the risk of DR (RR, 2.1) as compared with persons with a shorter DM duration, an increase that was not sustained for longer durations, although the sample size with such durations was small. Few persons at risk in this cohort used insulin at baseline. This small group of insulin users had a 6.6-fold increased risk of DR when compared with those without treatment or treated with diet only; those treated with oral medications had only a 2.2-fold increase in risk. Results of the main model also demonstrated that antihypertensive treatment halved the risk of DR (RR, 0.5) in this cohort of persons with DM, substantiating the results of Table 2. The risk of DR also increased as SBP increased. Relative risks rose from 1.8 for SBP levels between 120 and 140 mmhg, to 3.3 and 4.6 for higher SBP levels, both statistically significant at P Using SBP as a continuous rather than categorical variable, the risk of Table 2. 9-Year Incidence of Diabetic Retinopathy and Blood Pressure at Baseline Antihypertensive Medications Incidence (%) Yes (n*) No (n*) Overall (n*) RR (95% CI) SBP at baseline (mmhg) (14) 34.3 (56) 35.0 (70) SBP (40) 41.1 (80) 39.2 (120) 1.4 ( ) 140 SBP (44) 62.2 (37) 42.1 (81) 1.7 ( ) (38) 42.9 (14) 41.7 (52) 2.0 ( ) Per 10-mmHg-higher SBP at baseline 1.1 ( ) DBP at baseline (mmhg) (12) 32.8 (43) 28.8 (55) DBP (29) 41.2 (67) 44.1 (96) 2.1 ( ) 80 DBP (46) 49.7 (54) 41.6 (100) 2.0 ( ) (19) 45.5 (24) 38.9 (73) 2.0 ( ) Per 10-mmHg-higher DBP at baseline 1.1 ( ) CI confidence interval; DBP diastolic blood pressure; RR risk ratio; SBP systolic blood pressure. *At risk. Based on discrete hazard model, adjusting for antihypertensive treatment P 0.1. P

4 Ophthalmology Volume 112, Number 5, May 2005 Table 3. Associations with 9-Year Incidence of Diabetic Retinopathy Factors No. at Risk* Adjusted RR 95% CI P Value Main model with SBP-related variables Age of onset of diabetes per 10-yr-higher age Duration of diabetes (yrs) Diabetes treatment None/diet only Oral medication Insulin Antihypertensive medication SBP (mmhg) at baseline SBP SBP Change in SBP, per 10-mmHg increase Separate model with DBP-related variables Antihypertensive medication DBP (mmhg) at baseline DBP DBP Change in DBP, per 10-mmHg increase CI confidence interval; DBP diastolic blood pressure; RR risk ratio; SBP systolic blood pressure. *Eleven of the overall 324 persons at risk had missing values for variables included in the multivariate regression models. Based on multivariate discrete hazard regression model with covariates as specified in this table. DR increased by 30% for every 10 mmhg of higher SBP at baseline (RR, 1.3). In addition, an increase in SBP from baseline to the 4-year follow-up also was associated independently with the development of DR 9 years later (RR, 1.3 for each 10- mmhg increase in SBP from baseline). In a separate model substituting DBP for SBP, which is presented in the lower panel of Table 3, higher baseline DBP doubled or tripled the risk of DR. The RRs were 2.4 for DBP between 70 and 80 mmhg and 2.6 for DBP between 80 and 90 mmhg, further increasing to 3.4 for DBP of 90 mmhg. Additional multivariate analyses were conducted among the subgroup of persons with hypertension at baseline. A comparison of treated and untreated hypertensives also revealed significant protective RRs for antihypertensive treatment (RR, 0.47 [95% CI, ] for a model including SBP; RR, 0.51 [95% CI, ] for a model including DBP). Table 4 presents the associations of 9-year DR incidence and GHb among the subgroup of persons with baseline GHb measurements (n 285). The incidence gradually increased from 15.4% in persons with GHb of 8% to 34.0%, 40.5%, and 62.2% in those with higher GHb levels at baseline. Relative to persons with GHb of 8% at baseline, adjusted analyses found that the RR increased progressively from 2.2 in persons with GHb of 8% to 10% to 5.6 and 8.9 in persons with higher values. The increasing trend was further confirmed in a separate model using GHb as a continuous variable. Those results showed that the RR increased by 30% with each 1% of higher baseline GHb level. A borderline significant association was also noted between DR incidence and an increase in GHb from baseline to the 4-year follow-up (RR, 2.4). Results for other covariates in this subgroup remained similar to those found in the main cohort. Discussion This long-term longitudinal study has identified several factors related to the development of DR in a population of African origin. Such information, which was not previously available, is valuable for understanding the reasons why DR develops, as well as planning strategies for prevention and targeting groups at high risk. Table 4. Associations* of Glycosylated Hemoglobin (GHb) and 9-Year Incidence of Diabetic Retinopathy No. at risk Incidence (%) Adjusted RR (95% CI)* P Value GHb level at baseline 8% % 10% ( ) % 11.5% ( ) % ( ) 0.01 RR per 1% increase 1.3 ( ) 0.01 Change in GHb per 1% increase 2.4 ( ) CI confidence interval; RR risk ratio. *Based on multivariate discrete hazard regression model including other covariates: age of onset, duration, and treatment of diabetes; antihypertensive treatment; and baseline and 4-year changes of systolic blood pressure. 802

5 Leske et al Hyperglycemia, Blood Pressure, and Incident Diabetic Retinopathy Diabetes: Age of Onset, Duration, and Treatment Our findings substantiate the adverse impact of early DM onset and severity on the long-term development of DR. In the multivariate analyses, persons with an older age at DM onset had decreased risks of DR development. In this population, in which type 2 diabetes predominates, there was an estimated 30% reduction in risk for every 10 years of older age at onset. The association with age of DM onset has been reported in other populations, 23,24 as well as an association with duration of DM, the latter including a nonlinear relationship. 24,31,32 Data from our previously reported 4-year follow-up 11 and the current longer follow-up also suggest a nonlinear association between DR development and DM duration. In our study, the adjusted risk of developing DR in 9 years was approximately doubled in persons with 5 to 9 years of DM duration, relative to those with shorter duration. An association with DM duration of 10 to 14 years was also found in the univariate analyses but was not confirmed in the multivariate analyses, possibly because of the adjustment for the seemingly correlated variables of age of DM onset and its type of treatment. In the group with the longest DM duration ( 15 years), selective mortality and small sample sizes may have contributed to the non statistically significant results. This long-term follow-up study also confirmed our previous findings 11 that DM treatment, another variable that may reflect disease severity, remained as an independent predictor of DR development in the study. Hyperglycemia A strong relationship between increasing level of glycemia, as measured by GHb, and incidence of DR was found over the 9-year period of the study. Our data further suggested that a 1% increase in GHb from baseline to the 4-year follow-up was associated with a 2-fold risk of DR. Our previous evaluations indicated that the 4-year incidence increased from 10% in persons with GHb of 8% to 42% in those with GHb of 11.5%. 11 After 9 years of follow-up, the corresponding rates increased to 15% and 62%, respectively, with the adjusted RR being close to 9 for the highest GHb category. Over three fifths of persons with GHb of 11.5% developed DR in 9 years. Many epidemiological studies have confirmed that hyperglycemia, assessed by plasma glucose 23 25,27,28,33,34 or GHb, 29,30,34 39 is an important risk factor for DR. In the Wisconsin Epidemiologic Study of Diabetic Retinopathy, the GHb level at baseline was a significant predictor of incidence of any retinopathy over the 10-year period. 38 Our findings on GHb are to some extent consistent with the Wisconsin Epidemiologic Study of Diabetic Retinopathy evaluations after 14 years of follow-up, in which a higher GHb at baseline and a GHb increase between the baseline and 4-year follow-up visits were associated with an increased risk of proliferative retinopathy or incidence of macular edema in type 1 DM patients. The benefit of intensive blood glucose control was reported by the United Kingdom Prospective Diabetes Study, which found a 25% risk reduction in microvascular end points over 10 years among patients with type 2 diabetes. 4 The results from our study further support the importance of glycemic control in the early course of DM, particularly in populations with a high prevalence of the disease. Blood Pressure The relationship between BP and risk for DR has been controversial. Elevated SBP 27,28,33,35,36,39 or DBP 25 was associated with the development of DR in some studies, whereas no significant relation was found among olderonset diabetic patients in the Wisconsin Epidemiologic Study of Diabetic Retinopathy 40 and other studies. 23,34,41,42 In the Barbados Eye Studies, we found that higher SBP and DBP at baseline were associated with the development of DR over a 9-year period, after controlling for antihypertensive treatment and other confounding variables. In addition, an increase in SBP from baseline to 4 years later also increased the 9-year incidence of DR. The associations remained after adjusting for GHb in the subgroup of persons with GHb data (data not shown). These results suggest the potential value and beneficial implications of BP control in this African-origin population, where 50% have hypertension. 9 These results are further supported by our finding of a lower DR incidence in persons receiving treatment to reduce BP, as well as a halving of the risk among hypertensives who were treated versus those who were not. This result is of interest, as the effectiveness of antihypertensives to reduce microvascular complications in persons with type 2 DM has not been shown consistently in clinical trials. The United Kingdom Prospective Diabetes Study showed that after a median follow-up of 8.4 years, tight BP control resulted in 37% reduction in microvascular end points, which was predominantly related to a reduced risk of retinal photocoagulation. 5 In contrast, the Appropriate Blood Pressure Control in Diabetes Trial found no difference in DR progression over 5 years when comparing hypertensive patients with intensive (DBP goal of 75 mmhg) and moderate (DBP goal of mmhg) treatment. 6 The increased risk of DR seemed to start at levels lower than those traditionally considered as high BP, a result consistent with an earlier finding of a 6-year follow-up study in Pima Indians. 27 Another report from the Appropriate Blood Pressure Control in Diabetes Trial demonstrated that intensive BP control in normotensive type 2 diabetic patients decreased the progression of DR. 7 From this overview, the efficacy of antihypertensive treatment to prevent DR in patients with DM remains unclear. Additionally, the extent of the reduction needed to achieve beneficial effects has not been determined. Our results suggesting lower incidence in treated hypertensives and an increased risk of DR at higher SBP and DBP levels are consistent with the beneficial effects of BP reduction on DR development. In conclusion, DM and its complications are especially serious health problems in persons of African descent. Our study has determined that indicators of disease severity, such as younger age at DM onset, duration, and treatment, are related to the long-term risk of DR in this population, along with poor glycemic control. Results also support the role of high BP in DR development, as well as a decreased risk with hypertension treatment. These results further reinforce the importance of glycemic and hypertension control in populations of African origin. 803

6 Ophthalmology Volume 112, Number 5, May 2005 References 1. Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema: Early Treatment Diabetic Retinopathy Study report number 1. Arch Ophthalmol 1985;103: Diabetic Retinopathy Study Research Group. Indications for photocoagulation treatment of diabetic retinopathy: Diabetic Retinopathy Study Report no. 14. Int Ophthalmol Clin 1987; 27: Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulindependent diabetes mellitus. N Engl J Med 1993;329: UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998;352: UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes UKPDS 38. BMJ 1998;317: Estacio RO, Jeffers BW, Gifford N, Schrier RW. Effect of blood pressure control on diabetic microvascular complications in patients with hypertension and type 2 diabetes. Diabetes Care 2000;23(suppl):B Schrier RW, Estacio RO, Esler A, Mehler P. Effects of aggressive blood pressure control in normotensive type 2 diabetic patients on albuminuria, retinopathy and strokes. Kidney Int 2002:61; Hennis A, Wu SY, Nemesure B, et al, Barbados Eye Studies Group. Diabetes in a Caribbean population: epidemiological profile and implications. Int J Epidemiol 2002;31: Hennis A, Wu SY, Nemesure B, Leske MC, Barbados Eye Studies Group. Hypertension prevalence, control and survivorship in an Afro-Caribbean population. J Hypertens 2002; 20: Leske MC, Wu SY, Hyman L, et al. Diabetic retinopathy in a black population:the Barbados Eye Study. Ophthalmology 1999;106: Leske MC, Wu SY, Hennis A, et al, Barbados Eye Studies Group. Incidence of diabetic retinopathy in the Barbados Eye Studies. Ophthalmology 2003;110: Leske MC, Wu SY, Hennis A, et al. 9-year incidence of diabetic retinopathy in the Barbados Eye Studies. Arch Ophthalmol. In press. 13. Leske MC, Connell AM, Schachat A, Hyman L. The Barbados Eye Study: prevalence of open angle glaucoma. Arch Ophthalmol 1994;112: Leske MC, Connell AM, Wu SY, et al, Barbados Eye Studies Group. Incidence of open-angle glaucoma: the Barbados Eye Studies. Arch Ophthalmol 2001;119: Leske MC, Wu SY, Nemesure B, et al, Barbados Eye Studies Group. Nine-year incidence of lens opacities in the Barbados Eye Studies. Ophthalmology 2004;111: Willey DG, Rosenthal MA, Caldwell S. Glycosylated haemoglobin and plasma glycoprotein assays by affinity chromatography. Diabetologia 1984;27: Diabetic Retinopathy Study Group. Diabetic Retinopathy Study. Report 7: a modification of the Airlie House classification of diabetic retinopathy. Invest Ophthalmol Vis Sci 1981;21: Schachat AP, Hyman L, Leske MC, et al, Barbados (West Indies) Eye Study Group. Comparison of diabetic retinopathy detection by clinical examinations and photograph gradings. Arch Ophthalmol 1993;111: Klein BE, Davis MD, Segal P, et al. Diabetic retinopathy: assessment of severity and progression. Ophthalmology 1984; 91: National Eye Institute. Early Treatment Diabetic Retinopathy Study Research Group. Manual of Operations. Classification of Diabetic Retinopathy. Detailed Grading of Stereo Color Fundus Photographs. Springfield, VA: National Technical Information Service; NTIS Publication PB Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53: Cox DR. Regression models and life-tables. J R Stat Soc Ser B 1972;34: Sasaki A, Horiuchi N, Hasewgawa, Uehara M. Development of diabetic retinopathy and its associated risk factors in type 2 diabetic patients in Osaka district, Japan: a long-term prospective study. Diabetes Res Clin Pract 1990;10: Ballard DJ, Melton LJ III, Dwyer MS, et al. Risk factors for diabetic retinopathy: a population-based study in Rochester, Minnesota. Diabetes Care 1986;9: Agardh E, Agardh CD, Koul S, Torffvit O. A four-year follow-up study on the incidence of diabetic retinopathy in older onset diabetes mellitus. Diabetic Med 1994;11: Dorf A, Ballintine EJ, Bennett PH, Miller M. Retinopathy in Pima Indians: relationships to glucose level, duration of diabetes, age at diagnosis of diabetes, and age at examination in a population with a high prevalence of diabetes mellitus. Diabetes 1976;25: Knowler WC, Bennett PH, Ballintine EJ. Increased incidence of retinopathy in diabetics with elevated blood pressure: a six-year follow-up study in Pima Indians. N Engl J Med 1980;302: Dowse GK, Humphrey AR, Collins VR, et al. Prevalence and risk factors for diabetic retinopathy in the multiethnic population of Mauritius. Am J Epidemiol 1998;147: McKay R, McCarty CA, Taylor HR. Diabetic retinopathy in Victoria, Australia: the Visual Impairment Project. Br J Ophthalmol 2000;84: West SK, Klein R, Rodriguez J, et al. Diabetes and diabetic retinopathy in a Mexican-American population: Proyecto VER. Diabetes Care 2001;24: Klein R, Klein BE, Moss SE, et al. The Wisconsin Epidemiologic Study of Diabetic Retinopathy. X. Four-year incidence and progression of diabetic retinopathy when age at diagnosis is 30 years or more. Arch Ophthalmol 1989;107: Mitchell P. Development and progression of diabetic eye disease in Newcastle ( ): rates and risk factors. Aust N Z J Ophthalmol 1985;13: Teuscher A, Schnell H, Wilson PWF. Incidence of diabetic retinopathy and relationship to baseline plasma glucose and blood pressure. Diabetes Care 1988;11: Janghorbani M, Amini M, Ghanbari H, Safaiee H. Incidence of and risk factors for diabetic retinopathy in Isfahan, Iran. Ophthalmic Epidemiol 2003;10: Janghorbani M, Jones RB, Murray KJ, Allison SP. Incidence of and risk factors for diabetic retinopathy in diabetic clinic attenders. Ophthalmic Epidemiol 2001;8: Tudor SM, Hamman RE, Baron A, et al. Incidence and progression of diabetic retinopathy in Hispanics and non-hispanic whites with type 2 diabetes. Diabetes Care 1998;21: Klein R, Klein BE, Moss SE, et al. Glycosylated hemoglobin predicts the incidence and progression of diabetic retinopathy. JAMA 1988;260: Klein R, Klein BE, Moss SE, Cruickshanks KJ. Relationship of hyperglycemia to the long-term incidence and pro- 804

7 Leske et al Hyperglycemia, Blood Pressure, and Incident Diabetic Retinopathy gression of diabetic retinopathy. Arch Intern Med 1994; 154: West KM, Ahuja MM, Bennett PH, et al. Interrelationships of microangiopathy, plasma glucose and other risk factors in 3583 diabetic patients: a multinational study. Diabetologia 1982;22: Klein R, Klein BEK, Moss SE, et al. Is blood pressure a predictor of the incidence or progression of diabetic retinopathy? Arch Intern Med 1989;149: Yanko L, Goldbourt U, Michaelson IC, et al. Prevalence and 15-year incidence of retinopathy and associated characteristics in middle-aged and elderly diabetic men. Br J Ophthalmol 1983;67: Mitchell P, Smith W, Wang JJ, Attebo K. Prevalence of diabetic retinopathy in an older community. The Blue Mountains Eye Study. Ophthalmology 1998;105: Appendix: The Barbados Eye Studies Group Principal Investigator M. Cristina Leske, MD, MPH. Coordinating Center Stony Brook University, Stony Brook, New York: M. C. Leske, MD, MPH, Barbara Nemesure, PhD, Suh-Yuh Wu, MA, Leslie Hyman, PhD, Xiaowei Li, PhD, Lixin Jiang, MS, Ling Yang, MS, Kasthuri Sarma, Karen Kelleher, Melinda Santoro. Data Collection Center Ministry of Health, Bridgetown, Barbados: Anthea M. S. Connell, FRCS, FRCOphth (deceased), Anselm Hennis, MRCP(UK), PhD, Ann Bannister, MB,BS, DO, Muthu A. Thangaraj, MB,BS, DO, Coreen Barrow, Patricia Basdeo, Kim Bayley, Anthanette Holder. Fundus Photography Reading Center The Johns Hopkins University, Baltimore, Maryland: Andrew P. Schachat, MD, Judith A. Alexander, Cheryl J. Hiner, Noreen B. Javornik, MS, Deborah A. Phillips, Reva W. Strozykowski, Terry W. George. Advisory Committee Trevor Hassell, FRCP, FACC (Department of Cardiology), Henry Fraser, PhD, GCM (Chronic Diseases Research Centre), Clive Gibbons, FRCS(Ed), FRCP (Department of Ophthalmology), School of Clinical Medicine and Research, University of the West Indies, and Queen Elizabeth Hospital, Barbados. 805