Advertisement

Diabetologia

, Volume 55, Issue 3, pp 671–680 | Cite as

Retinopathy in old persons with and without diabetes mellitus: the Age, Gene/Environment Susceptibility—Reykjavik Study (AGES-R)

  • E. Gunnlaugsdottir
  • S. Halldorsdottir
  • R. Klein
  • G. Eiriksdottir
  • B. E. Klein
  • R. Benediktsson
  • T. B. Harris
  • L. J. Launer
  • T. Aspelund
  • V. Gudnason
  • M. F. Cotch
  • F. Jonasson
Article

Abstract

Aims/hypothesis

We aimed to describe the prevalence of retinopathy in an aged cohort of Icelanders with and without diabetes mellitus.

Methods

The study population consisted of 4,994 persons aged ≥67 years, who participated in the Age, Gene/Environment Susceptibility—Reykjavik Study (AGES-R). Type 2 diabetes mellitus was defined as HbA1c ≥6.5% (>48 mmol/mol). Retinopathy was assessed by grading fundus photographs using the modified Airlie House adaptation of the Early Treatment Diabetic Retinopathy Study protocol. Associations between retinopathy and risk factors were estimated using odds ratios obtained from multivariate analyses.

Results

The overall prevalence of retinopathy in AGES-R was 12.4%. Diabetes mellitus was present in 516 persons (10.3%), for 512 of whom gradable fundus photos were available, including 138 persons (27.0%, 95% CI 23.2, 31.0) with any retinopathy. Five persons (1.0%, 95% CI 0.3, 2.3) had proliferative retinopathy. Clinically significant macular oedema was present in five persons (1.0%, 95% CI 0.3, 2.3). Independent risk factors for retinopathy in diabetic patients in a multivariate model included HbA1c, insulin use and use of oral hypoglycaemic agents, the last two being indicators of longer disease duration. In 4478 participants without diabetes mellitus, gradable fundus photos were available for 4,453 participants, with retinopathy present in 476 (10.7%, 95% CI 9.8, 11.6) and clinically significant macular oedema in three persons. Independent risk factors included increasing age and microalbuminuria.

Conclusions/interpretation

Over three-quarters (78%) of retinopathy cases were found in persons without diabetes and a strong association between microalbuminuria and non-diabetic retinopathy was found. These results may have implications for patient management of the aged.

Keywords

Diabetes mellitus Microalbuminuria Non-diabetic Old age Population sample Random Retinopathy 

Abbreviations

AGES-R

Age Gene/Environment Susceptibility—Reykjavik Study

CSMO

Clinically significant macular oedema

DR

Diabetic retinopathy

IHA

Icelandic Heart Association

MESA

Multi-Ethnic Study of Atherosclerosis

NHANES

National Health and Nutrition Examination Survey

NPDR

Non proliferative diabetic retinopathy

PDR

Proliferative diabetic retinopathy

Introduction

The global burden of diabetes is increasing and trends in fasting plasma glucose have been on the rise since 1980 [1]. The number of people with diabetes in the world is expected to double between 2000 and 2030 [2]. According to the WHO, an estimated 5% of global blindness is due to diabetic retinopathy (DR) [3]. Retinopathy in persons without diabetes, sometimes mistakenly referred to as DR, is estimated to occur in up to 15% of the general population, this prevalence being attributed to older age and systemic hypertension [4, 5, 6].

The experience in Iceland parallels that observed globally. The prevalence of type 2 diabetes mellitus in middle-aged men in Iceland doubled from 1967 to 2002, coinciding with the increasing BMI in the population [7]. In a national diabetic eye screening programme, the prevalence of DR among Icelandic type 2 diabetic patients was reported in 1994 to be 41% [8]. Of those in this screening programme and without retinopathy at baseline, only 1% of patients with type 2 diabetes developed proliferative retinopathy over a 10 year observation period [9]. In a recent study, we reported that the overall prevalence of visual impairment in Icelanders aged 50 years and older was 1% according to WHO criteria (visual acuity <6/18 in the better eye) [10]. No participant had bilateral visual acuity less than 6/12 due to diabetic eye disease, and only one became visually impaired due to DR over a 5 year period [11].

Here, data from the Age, Gene/Environment Susceptibility—Reykjavik Study (AGES-R) [12] are used to provide population-based estimates of the prevalence and severity of retinopathy and associated factors among elderly Icelanders with and without diabetes mellitus.

Methods

Study details

The AGES-R study is an epidemiological, population-based study aimed at investigating genetic and environmental factors contributing to disease as well as to healthy ageing [12]. The study originates from the Reykjavik study, which was conducted by the Icelandic Heart Association (IHA) between 1967 and 1997 [12, 13, 14]. Participants of that study were men and women who were born between 1907 and 1935, and were randomly sampled and living in the Reykjavik area in 1967. At the time of recruitment into AGES-R, 11,549 persons participating in the original Reykjavik Study (38%) were still alive, and a random sample of 5,764 survivors was enrolled. The AGES-R examinations took place from February 2002 to February 2006 [12, 13, 14]. The comprehensive AGES-R examination protocol required each participant to complete three visits to the IHA Research Center over a period of 3 to 6 months. During the assessment at the IHA Research Center (see below), participants completed a detailed interview including questions about diabetes, eye health and cardiovascular history. Blood specimens were drawn and an extensive medical examination, including an eye examination, was performed as described below. Transport to the IHA Clinic was provided for those who asked for it. Participants were asked to bring all medications that they had used in the previous 2 weeks. The AGES-R study methods, examination protocol and characteristics of the cohort have been described in detail elsewhere [12, 13, 14]. Informed consent was obtained from all participants. The study was approved by the National Bioethics Committee in Iceland, which acts as the Institutional Review Board for the IHA (VSN-00-063), and by the National Institute on Aging Intramural Institutional Review Board.

Definition of diabetes mellitus and laboratory tests

The criteria used for type 2 diabetes diagnosis were self-reported diabetes in the questionnaire, and/or use of diabetes medication or HbA1c ≥6.5% (≥48 mmol/mol). The latter is a recently recommended approach and cut point for diagnosis of diabetes, and is intended to better avoid the problem of day-to-day variability of glucose values [15]. To be able to compare our results with results from earlier publications, we also, in a separate analysis, used fasting serum glucose of ≥7 mmol/l for undiagnosed diabetes mellitus, based on earlier WHO recommendations [16, 17]. These last results are briefly mentioned in the text. The criteria for type 1 diabetes included age ≤25 years at diagnosis and insulin dependence.

Blood samples were drawn after overnight fasting. Total cholesterol, HDL-cholesterol, triacylglycerol, high-sensitivity C-reactive protein, serum glucose and HbA1c were analysed on an Hitachi 912 analyser (Mito, Japan) using reagents from Roche Diagnostics (Mannheim, Germany) and following the manufacturer’s instructions. LDL-cholesterol was calculated using the Friedewald equation [18, 19]. The coefficient of variation (%) was: for serum glucose 1.8, for total cholesterol 1.4, for HDL 2.3 and for urinary albumin 4.8.

Eye examination

The visual acuity of each eye was measured using an auto-refractokeratometer (ARK-760A; Nidek, Tokyo, Japan) with a built-in Snellen chart. Pupils were then dilated with tropicamide eye drops 10 mg/ml and photographs of the retina taken. For each eye, two 45-degree digital retinal images, centred on the optic disc (field 1) and the macula (field 2), respectively, were taken using a CR6 non-mydriatic camera (Canon, Tokyo, Japan) with a Canon D60 camera back. Images were evaluated at the Ocular Epidemiology Reading Center, University of Wisconsin using the modified Airlie House Classification system [20]. EyeQ Lite image processing software was used [21]. All graders were masked to the health status of participants.

Definition of retinopathy and macular oedema

If the grader was at least 90% certain that a retinopathy lesion was present, the lesion was marked as definite. The grader then assigned a retinopathy level according to the Modified Airlie House adaptation from the Early Treatment Diabetic Retinopathy Study protocol [20]. The following grading criteria were used: no DR (levels 10–13), mild non-proliferative DR (NPDR) (levels 14–20), moderate to severe NPDR (levels 41–51) or proliferative DR (PDR) (levels 60–80). Macular oedema was defined by hard exudates in the presence of microaneurysm and blot haemorrhage within 1 disc diameter from the foveal centre, or if focal or macular grid photocoagulation scars were present in the macula. Clinically significant macular oedema (CSMO) was considered present when oedema involved the fovea or was within 500 μm of the fovea, or if photocoagulation scars were present in the macula.

Assessment of factors associated with retinopathy

The duration of diabetes was calculated as the difference between reported age of diabetes diagnosis from the interview and the year of the AGES-R examination. Use of medications was noted from medications brought to the clinic and from self-report questionnaire. Blood pressure was assigned using the mean value of two blood pressure measurements with a large-cuff mercury sphygmomanometer. Hypertension was defined by self-reported doctor’s diagnosis of hypertension, use of hypertensive medication, or measured systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg. BMI was calculated from measured height and weight (kg/m2). Microalbuminuria was noted as present if the albumin/creatinine ratio in a random urine sample was >300 mg/g. GFR (ml min−1 1.73 m−2) was estimated using the Modification of Diet in Renal Disease (MDRD) formula [22].

HbA1c was imputed for 7% of the participants with missing data by applying the expectation-maximisation (EM) algorithm in the statistical program PROC MI (SAS/STAT Institute, Cary, NC, USA). The imputation was performed separately for participants without and those with diabetes, depending on medication use, and using age, fasting serum glucose and duration of diabetes as the predictors for HbA1c. Neither diabetes status nor retinopathy status was associated with missing HbA1c. HbA1c was imputed for five persons with undiagnosed diabetes mellitus. However, their fasting serum glucose was 8.9 mmol/l or higher.

Data handling and statistical analysis

For statistical analysis we used SAS Enterprise Guide software, version 9.1 (SAS Institute, Cary, NC, USA). Analysis of eyes was based on the participant’s eye with the most severe retinopathy lesion (worse eye). If images from an eye were ungradable or if fundus photographs were unavailable, scores from the contralateral gradable eye were used. The overall prevalence of retinopathy was calculated using the entire sample with information on retinopathy in at least one eye.

Differences between groups were tested using χ 2 test for categorical variables and t test for continuous variables. Confidence intervals for ratios were calculated using binominal distribution. Logistic regression models adjusting for age and sex were run for each risk factor. In the multivariate models, we included age, sex, systolic blood pressure, hypertension as a categorical variable, HbA1c and microalbuminuria; and, for participants with diabetes, duration of diabetes and use of glucose-lowering medication.

Results

Study cohort

Of the 5,764 AGES-R participants, 4,994 completed an eye examination. Of these, 4,965 had gradable photographs for at least one eye, as well as clinical examination findings, along with blood and urine test results. A comparison of participants and non-participants is reported elsewhere [12]. Of the 4,994 participants, 516 (10.3%) fulfilled the criteria for diagnosis of diabetes mellitus. Two persons (both women) were younger than 25 years at diabetes diagnosis and insulin-dependent. These two were therefore classified as having type 1 diabetes mellitus, but were included in the analysis. If we had classified by onset of diabetes mellitus <35 years and insulin dependence, three more persons would also have been classified as having type 1 diabetes mellitus.

Characteristics of participants by diabetes status are presented in Table 1. Men were significantly more likely than women to have diabetes mellitus. Persons with diabetes were more likely than those without to have hypertension, larger BMI, higher fasting serum glucose, lower total cholesterol, GFR <60 ml min–1 1.73 m–2, a positive history of cardiovascular disease, higher levels of HbA1c and microalbuminuria. The 83 participants with undiagnosed diabetes mellitus had a mean age of 76.3 years.
Table 1

Comparison of characteristics of participants in AGES-R

Characteristic

Total group (n = 4,994)

Diabetes mellitus by diagnosis time (n = 516)

Diabetes mellitus by treatment (n = 516)

No diabetes

Diabetes

p value

Undiagnosed

Previously known

p value

Insulin ± orala

Only orala

Diet or no treatment

p valueb

n (%)

4,478 (89.7)

516 (10.3)

83 (16.1)

433 (83.9)

37 (6.2)

273 (52.9)

206 (40.0)

Male sex, n (%)

1,880 (42.0)

272 (52.7)

<0.001

36 (43.4)

236 (54.5)

0.06

21 (56.8)

164 (60.1)

87 (42.2)

0.61

Age (years)

76.4 ± 5.5

76.4 ± 5.3

0.97

76.6 ± 5.3

76.5 ± 5.4

0.89

75.7 ± 4.6

76.3 ± 5.2

76.9 ± 5.6

0.35

Age at T2DM diagnosis (years)

64.9 ± 12.6

76.3 ± 5.4

63.7 ± 12.6

<0.001

53.1 ± 13.5

64.7 ± 10.2

67.7 ± 13.7

<0.001

DM duration (years)

10.3 (11.2)

0

12.4 (11.2)

<0.001

24.9 (13.7)

11.1 (9.7)

6.8 (10.4)

<0.001

Hypertension, n (%)

3,579 (79.9)

466 (90.3)

<0.001

72 (86.8)

394 (91.0)

0.23

33 (89.2)

250 (91.6)

184 (89.3)

0.81

Systolic BP (mmHg)

141.9 ± 20.2

146.0 ± 20.4

<0.001

141.6 ± 21.9

146.0 ± 20.4

0.10

145.6 ± 19.8

145.2 ± 20.6

145.4 ± 21.0

0.93

BMI (kg/m²)

26.8 ± 4.3

28.9 ± 4.7

<0.001

27.8 ± 4.4

28.6 ± 4.6

0.14

28.5 ± 5.8

28.6 ± 4.4

28.2 ± 4.5

0.94

Measured HbA1c (%)

5.6 ± 0.3

6.5 ± 0.9

<0.001

7.0 ± 0.9

6.5 ± 0.9

<0.001

7.4 ± 0.9

6.6 ± 0.9

6.5 ± 1.0

<0.001

Imputed HbA1c (%)

5.9 ± 0.3

6.5 ± 0.9

<0.001

7.0 ± 0.8

6.5 ± 0.9

0.007

7.3 ± 0.9

6.6 ± 0.9

6.5 ± 1.0

<0.001

Measured HbA1c (mmol/mol)

37.6 ± 3.5

48.8 ± 10.4

<0.001

53.1 ± 9.5

48.0 ± 10.3

<0.001

56.9 ± 9.5

48.6 ± 9.7

47.6 ± 10.8

<0.001

Imputed HbA1c (mmol/mol)

37.6 ± 3.4

48.7 ± 10.1

<0.001

53.0 ± 9.3

47.8 ± 10.1

0.007

56.5 ± 9.4

48.6 ± 9.4

47.3 ± 10.6

<0.001

Fasting glucose (mmol/l)

5.5 ± 0.6

7.8 ± 2.1

<0.001

7.6 ± 2.6

7.8 ± 2.4

0.50

9.2 ± 3.8

8.1 ± 2.1

7.1 ± 2.3

<0.001

Total cholesterol (mmol/l)

5.7 ± 1.1

5.2 ± 1.1

<0.001

5.4 ± 1.2

5.1 ± 1.1

0.007

4.6 ± 1.1

5.0 ± 1.1

5.4 ± 1.1

0.008

GFR <60 ml min−1 1.73−2, n (%)

1,336 (29.8)

206 (39.9)

<0.001

34 (41.0)

172 (39.7)

0.83

17 (45.9)

105 (38.5)

84 (40.8)

0.33

Current smoker, n (%)

546 (12.2)

60 (11.6)

0.71

10 (12.0)

50 (11.6)

0.90

5 (13.5)

34 (12.5)

21 (10.2)

0.71

Ever smoker, n (%)

2,550 (56.9)

305 (59.1)

0.34

47 (56.6)

258 (59.7)

0.25

25 (67.6)

161 (59.0)

119 (57.8)

0.28

Previous CVD, n (%)

921 (20.6)

160 (31.0)

<0.001

21 (25.3)

139 (22.3)

0.20

14 (37.8)

95 (34.8)

51 (24.8)

0.30

Microalbuminuria, n (%)

434 (7.4)

101 (19.6)

<0.001

10 (12.0)

91 (21.3)

0.06

11 (29.7)

66 (24.2)

24 (11.7)

0.12

Values are mean±SD unless otherwise indicated

CVD, cardiovascular disease according to hospital records; DM, diabetes mellitus; T2DM, type 2 diabetes mellitus

aOral hypoglycaemic agents

bFor insulin users compared with the two other treatment groups combined

Prevalence and factors associated with retinopathy in participants with diabetes mellitus

Gradable fundus photographs from at least one eye were available from 512 of the 516 persons with diabetes mellitus. Any DR was present in one or both eyes in 138 persons (27.0%). Table 2 shows the prevalence of DR and macular oedema by age, sex, categories of diabetes duration and diabetes treatment. Among those with diabetes mellitus, there were no statistically significant age- or sex-specific differences in the prevalence of any DR measure. Of the 83 persons with undiagnosed type 2 diabetes mellitus, one did not have readable fundus photographs and nine (11.0%) had any retinopathy (eight mild, one moderate NPDR). The prevalence of any DR and CSMO was higher in those who were being treated with insulin than in persons receiving other forms of treatment. Their mean duration of diabetes was 24.9 (SD 13.7) years, compared with 11.1 (SD 9.7) years for participants receiving oral treatment only and 6.8 (SD 10.4) years for participants without medical treatment.
Table 2

Prevalence and severity of retinopathy and macular oedema in persons with diabetes mellitus

Group

Any retinopathy

Retinopathy

  

Macular oedema

NPDR

 

Any

CSMO

Mild

Moderate

PDR

Persons with DM (n = 512)

n

138

120

13

5

9

5

%

27.0

23.4

2.5

1.0

18

1.0

95% CI

23.2, 31.0

19.8, 27.4

1.4, 4.3

0.3, 2.3

0.8, 3.3

0.3, 2.3

Age group (years)

       

 67–74 (n = 190)

n

52

45

4

3

3

2

%

27.7

23.7

2.1

1.6

1.6

1.1

95% CI

21.2, 34.3

17.8, 30.4

0.6, 5.3

0.3, 4.5

0.3, 4.5

0.1, 3.7

 ≥75 (n = 322)

n

86

75

9

2

6

3

%

26.7

23.3

2.8

0.6

1.9

0.9

95% CI

22.0, 31.9

18.8, 28.3

1.3, 5.2

0.1, 2.2

0.7, 4.0

0.2, 2.7

Sex

       

 Male (n = 272)

n

78

74

8

1

4

1

%

28.7

26.4

2.9

0.4

1.5

0.4

95% CI

23.4, 34.4

20.3, 31.0

1.3, 5.7

0.0, 2.0

0.4, 3.7

0.0, 2.0

 Female (n = 240)

n

60

51

5

4

5

4

%

25.0

21.3

2.1

1.7

2.1

1.7

95% CI

19.8, 31.0

16.3, 27.0

0.7, 4.8

0.5, 4.2

0.7, 4.8

0.5, 4.2

DM

       

 Undiagnosed (n = 82)

n

9

8

1

0

0

0

%

11.0

9.8

1.2

0

0

0

95% CI

5.1, 19.8

4.3, 18.3

0.0, 6.6

0.0, 4.4

0.0, 4.4

0.0, 4.4

 Previously known (n = 430)

n

129

112

12

5

9

5

%

30.0

26.1

2.8

1.2

2.1

1.2

95% CI

25.7, 34.6

22.0, 30.5

1.5, 4.8

0.4, 2.7

1.0, 3.9

0.4, 2.7

Treatment

       

 Insulin ± orala (n = 37)

n

24

14

7

3

4

2

%

64.9

37.8

18.9

8.1

10.8

5.4

95% CI

47.5, 79.8

22.5, 55.2

8.5, 35.2

1.7, 21.9

3.0, 25.4

0.7, 18.2

 Only orala (n = 270)

n

82

76

4

2

4

3

%

30.4

28.1

1.5

0.7

1.5

1.1

95% CI

24.9, 36.2

22.9, 33.9

0.4, 3.8

0.1, 2.7

0.4, 3.7

0.2, 3.2

 Diet or no treatment (n = 205)

n

32

30

2

0

1

0

%

15.6

14.6

1.0

0

0.5

0

95% CI

10.9, 21.3

10.1, 20.2

0.1, 3.5

0.0, 1.8

0.0, 2.7

0.0, 1.8

DM, diabetes mellitus

aOral hypoglycaemic agents

Persons with any DR were younger at the time of diabetes mellitus diagnosis (mean age 61.1 [SD 14.4] vs 66.3 [SD 11.6] years) and had longer duration of diabetes mellitus, i.e. mean 13.7 (SD 12.6) vs mean 9.1 (SD 10.5) years.

In the multivariate analysis of participants with diabetes mellitus (Table 3), we found HbA1c, insulin use and use of oral hypoglycaemic agents to be associated with increased risk of any retinopathy. Systolic blood pressure was associated with increased risk of retinopathy in the multivariate model. However, the categorical variable for hypertension was not associated with increased risk of retinopathy. As shown above, duration of diabetes mellitus was highly confounded with type of treatment; in other words, the use of insulin or oral hypoglycaemic agents was an indicator of longer duration of diabetes mellitus. The duration variable was therefore included in the multivariate model, even though the p value (p = 0.0645) did not reach statistical significance. When the multivariate analysis was limited to persons with diabetes of duration more than 10 years, however, the association with the use of insulin or oral hypoglycaemic agents was of similar magnitude and statistically significant. We found HbA1c to be significantly associated with retinopathy if fasting glucose was left out of the model. Additionally, the fasting glucose level was found to be significantly associated with retinopathy when selected before HbA1c, but after adjustment for diabetes duration and medication use. Both variables could not be retained in the model due to co-linearity; the correlation between fasting glucose and HbA1c in participants with diabetes was 0.73. The final variables in the multivariate model were: age, sex, duration of diabetes mellitus, insulin use, oral hypoglycaemic agent use, systolic blood pressure and HbA1c.
Table 3

Univariate and multivariate logistic regression results on risk factors for any retinopathy in persons with and without diabetes mellitus

Logistic regression

n

With diabetes mellitusa

Without diabetes mellitusb

OR (95% CI)

p value

 

OR (95% CI)

p value

Multivariate

 Age, per 10 years

 

0.86 (0.57, 1.31)

0.4687

 

1.30 (1.09, 1.55)

0.0036

 Sex, male vs female

 

0.96 (0.62, 1.51)

0.871

 

0.97 (0.79, 1.18)

0.7245

 Systolic BP, per 10 mmHg

 

1.16 (1.04, 1.29)

0.0095

 

1.05 (1.00, 1.10)

0.0699

 Duration per 10 years

 

1.21 (0.99, 1.48)

0.0645

   

 Insulin use, yes vs no

 

3.51 (1.59, 7.77)

0.002

   

 Oral hypoglycaemic, yes vs no

 

1.93 (1.21, 3.07)

0.0056

   

 HbA1c per percentage point

 

1.35 (1.08, 1.68)

0.0078

 

0.93 (0.68, 1.27)

0.6431

 Hypertension

 

1.23 (0.53, 2.86)

0.6235

 

1.01 (0.76, 1.33)

0.9665

 Microalbuminuria, yes vs no

 

1.48 (0.88, 2.48)

0.1356

 

1.77 (1.30, 2.40)

0.0003

Univariate, adjusted for age and sex

 Age, per 10 years

512

0.89 (0.61, 1.28)

0.5175

4,453

1.38 (1.16, 1.63)

0.0002

 Sex, male vs female

512

1.21 (0.81, 1.79)

0.3498

4,453

1.01 (0.83, 1.22)

0.9216

 Systolic BP, per 10 mmHg

512

1.18 (1.07, 1.29)

0.001

4,452

1.06 (1.01, 1.11)

0.0128

 Duration per 10 years

498

1.41 (1.20, 1.66)

<0.0001

   

 Insulin use, yes vs no

512

5.79 (2.85, 11.76)

<0.0001

   

 Oral hypoglycaemic, yes vs no

512

2.15 (1.40, 3.29)

0.0004

   

 HbA1c per percentage point

512

1.42 (1.19, 1.68)

0.0005

4,453

0.94 (0.69, 1.27)

0.6699

 Hypertension

512

1.52 (0.74, 3.14)

0.2592

4,453

1.12 (0.87, 1.43)

0.3923

 Microalbuminuria, yes vs no

506

1.97 (1.22, 3.17)

0.0056

4,403

1.82 (1.34, 2.47)

0.0001

a n = 496 for multivariant analysis, n = 512 for univariant analysis

b n = 4,402 for multivariant analysis, n = 4,453 for univariant analysis

Prevalence and factors associated with retinopathy in non-diabetic participants

Of the 4,478 non-diabetic persons in the sample, 4,453 had gradable fundus photographs. Of these, retinopathy was found in 476 persons (10.7%), as shown in Table 4. Macular oedema was present in five persons (0.1%), all of whom had branch retinal vein occlusion. There was no sex-specific difference in the prevalence of retinopathy among non-diabetic persons. Factors associated with any retinopathy in a multivariate model were older age and microalbuminuria (p = 0.0003) (Table 3). Higher systolic blood pressure was significantly associated with retinopathy in the age- and sex-adjusted model, but its significance was only marginal in the multivariate model.
Table 4

Prevalence and severity of retinopathy and macular oedema in persons without diabetes mellitus

Group

Any retinopathy

Retinopathy

  

Macular oedema

NPDR

Any

CSMO

Mild

Moderate

PDR

Persons without DM (n = 4,453)

n

476

472

4

0

5

3

%

10.7

10.6

0.1

0

0.1

0.1

95% CI

9.8, 11.6

9.7, 11.5

0.0, 0.2

0.0, 0.1

0.0, 0.3

0.0, 0.2

Age group (years)

       

 67–74 (n = 1,828)

n

169

167

2

0

3

2

%

9.2

9.1

0.1

0

0.2

0.1

95% CI

8.0, 10.7

7.9, 10.6

0.0, 0.4

0.0, 0.2

0.1, 0.5

0.0, 0.4

 ≥75 (n = 2,625)

n

307

305

2

0

2

1

%

11.7

11.6

0.1

0

0.1

0.0

95% CI

10.5, 13.0

10.4, 12.9

0.0, 0.3

0.0, 0.1

0.0, 0.3

0.0, 0.2

Sex

       

 Male (n = 272)

n

201

199

2

0

2

0

%

10.7

10.6

0.1

0

0.1

0

95% CI

9.4, 12.2

9.3, 12.1

0.0, 0.4

0.0, 0.2

0.0, 0.4

0.0, 0.2

 Female (n = 240)

n

275

273

2

0

3

3

%

10.7

10.6

0.1

0

0.1

0.1

95% CI

9.5, 11.9

9.4, 11.8

0.0, 0.3

0.0, 0.1

0.0, 0.3

0.0, 0.3

DM, diabetes mellitus

Discussion

This large population-based study of old persons reports an overall prevalence of retinopathy of 12.4%. The prevalence of DR was 27.0% and the prevalence of retinopathy in non-diabetic participants was 10.7%. We used the recently recommended and presumably more accurate HbA1c ≥6.5% (≥48 mmol/mol) for diagnosis of diabetes. This classification is expected to be better at capturing chronic hyperglycaemia than the older classification using fasting glucose ≥7 mmol/l [15]. Persons with diabetes mellitus are 2.5 times more likely to have retinopathy than non-diabetic participants of comparable age, but over 75% of retinopathy cases in our study occurred in persons without diabetes. HbA1c, duration of diabetes mellitus, use of insulin and use of oral hypoglycaemic agents were associated with increased risk of retinopathy in persons with diabetes in a multivariate model. Use of insulin and hypoglycaemic agents was also associated with longer duration of diabetes in a multivariate model. In persons without diabetes, increased risk of retinopathy was associated with advancing age and microalbuminuria. Microalbuminuria data show a sort of dose–response, with a prevalence of 7.4% in those with no diabetes mellitus, 12.0% in those undiagnosed and 21.3% in those with previously diagnosed diabetes mellitus and mean disease duration of 12.4 years. These results support the hypothesis that microalbuminuria may be a marker of microvascular dysfunction. Studies of DR have until recently used fasting blood glucose (7 mmol/l) to diagnose diabetes mellitus. To facilitate comparison with previous work, we additionally did uni- and multivariate analysis using this definition, which is based on a one-time event, in contrast to our present classification of type 2 diabetes mellitus (HbA1c), which is expected to capture cases with chronic hyperglycaemia. Using the new classification, the group of participants with undiagnosed type 2 diabetes mellitus included only 50% of the number of persons classified, based on fasting serum glucose. In general, results were similar using the two classifications.

Retinopathy in patients with diabetes mellitus

The overall prevalence of retinopathy using HbA1c ≥6.5% for diagnosis was 27.0% (95% CI 23.0, 31.0). The prevalence of DR in the present study is lower than the 41% for any retinopathy and 7% for PDR reported in 1994 in a previous study on DR in Iceland [8]. This would also hold true if we used the older classification of type 2 diabetes mellitus, which includes fasting glucose ≥7 mml/l and would give an overall prevalence of 25.3% in the present study, whereby sex distribution in both studies is similar. However, the previous study was based on clinical referrals and is therefore likely to reflect selection bias. We attribute the lower cholesterol to the use of lipid-lowering agents among the group with type 2 diabetes.

The 25.3% overall prevalence of DR in AGES-R using the older diagnostic criteria is similar to recent reports of: (1) 24.8% in the Multi-Ethnic Study of Atherosclerosis (MESA) in a white population with mean age of 64 years [23]; and (2) the age- and sex-adjusted rate of 26.4% in whites aged 40 years and older in the National Health and Nutrition Examination Survey (NHANES) 2005–2008 [24], the latter diagnosed using the HbA1c criteria.

The overall DR prevalence in the Beaver Dam Study in the USA and the Blue Mountains Study in Australia was 36.8% and 32.4%, respectively [25, 26]. These older studies, however, used different definitions of type 2 diabetes mellitus, including higher fasting serum glucose values than our additional analysis, and therefore may have included cases with more advanced disease than in the present study. Differences in photographic techniques may also complicate a direct comparison of the studies because the AGES-R and MESA studies used digital non-stereoscopic 45° photographs of two fields, while the Beaver Dam and the Blue Mountains Eye studies used 30° stereoscopic photographs of up to seven fields and might therefore have detected more peripheral retinopathy lesions.

We confirmed findings from previous studies [23, 24, 25, 26, 27, 28, 29, 30], including a significant correlation between any DR in persons with diabetes mellitus and insulin use, oral hypoglycaemic medications, HbA1c ≥6.5% (≥48 mmol/mol) and longer duration of diabetes mellitus. In agreement with NHANES 2005–2008 [24], we found that higher systolic blood pressure was associated with any DR in a multivariate model. In a multivariate model, we found no correlation between DR and BMI, hypertension or diastolic blood pressure; this is in agreement with results from other studies [23, 24, 26, 28]. Due to the extremely low prevalence of PDR (1.0%; n = 5) and CSMO (1.0%; n = 5), statistical analysis of associated factors was not attempted.

Retinopathy in non-diabetic persons

We found a 10.7% (n = 476) prevalence of retinopathy in non-diabetic persons, which is comparable to the 11.9% prevalence of retinopathy reported in white non-diabetic persons in the MESA Study [5]. In fact, although the prevalence of retinopathy in diabetic patients is higher than in those without diabetes, the actual number of individuals with retinopathy was threefold higher in the non-diabetic group in our sample. Retinopathy lesions among non-diabetic participants have been associated in other studies with older age and hypertension [4, 6]. Non-diabetic participants with signs of retinopathy in our study were not more likely to be hypertensive, nor did they have higher values of fasting serum glucose or HbA1c than those without retinopathy. We did, however, find a strong association between microalbuminuria and non-diabetic retinopathy. While microalbuminuria has been confirmed to be a reliable marker of retinopathy in diabetic patients [31], we are not aware of previous data to confirm this relationship in non-diabetic persons. Several studies [32, 33] have suggested that since microvascular damage manifests outside the eye in other organ systems, retinopathy and nephropathy may share a common pathogenesis and be signs of systemic microvascular dysfunction.

Patients who have been diagnosed with retinopathy in eye clinics or screening programmes may want to know how this affects management of their condition. Should these elderly persons be referred for a diabetic work-up, which commonly happens, or for a cardiovascular work-up, or for both or neither? Interestingly, the Blue Mountains Eye Study [5] describes most retinopathy lesions as being transient, with only 3.5% of persons developing diabetes over 5 years. We will need longitudinal data to further understand the pathophysiology underlying retinopathy in non-diabetic patients and to provide an evidence basis for clinical management.

Study strengths and weaknesses

A strength of the AGES-R study is that the sample was drawn from a large, randomly selected, population-based cohort. We also used a recently recommended diagnosis criterion for diabetes mellitus, namely HbA1c ≥6.5% (≥48 mmol/mol), thus avoiding day-to-day variability of glucose values. HbA1c may be better at predicting the development of retinopathy [34], although this classification also has limitations [35]. This is a multidisciplinary study, providing detailed information on various physiological systems, including metabolic regulation and ophthalmic health. The diabetes cases in our sample included 83 individuals, i.e. 16.1% of the diabetes mellitus sample, who were first diagnosed at the baseline examination. All participants are survivors from the Reykjavik study, which took place in several stages over a 30 year period from 1967 to 1997, with testing for diabetes occurring at each examination, using the same fasting serum glucose cut-off (≥7 mmol/l) as was used in our additional analysis. Another strength of our study includes the use of digital photographic documentation of each fundus through dilated pupils, as well as the masked grading of retinopathy using standardised methods. Limitations of this study include the relatively few cases of diabetes mellitus and DR, resulting in reduced power to detect factors associated with increased risk of retinopathy.

Summary

This Icelandic population-based study reports that more than one in seven elderly persons has retinopathy. Although the prevalence of retinopathy among diabetic patients is 2.5 times higher than among non-diabetic persons, the actual number of individuals with retinopathy is more than threefold higher in the non-diabetic group. Retinopathy occurs in persons with HbA1c <6.5% (≥48 mmol/mol) and also in persons with normal levels of fasting serum glucose, and in these were significantly associated with microalbuminuria in the present study. Further study of the association between retinopathy and microalbuminuria is needed to provide insight into underlying systemic disease processes, which could help in the management of elderly patients to preserve vision and health, and to maintain quality of life in old age.

Notes

Acknowledgements

We thank the AGES-R participants for making this study possible. The Age, Gene/Environment Susceptibility-Reykjavik Study is funded by NIH contract N01-AG-12100, the NIH/NIA Intramural Research Program, the National Eye Institute (NEI) of the NIH (ZIAEY000401), Hjartavernd (the Icelandic Heart Association), the Althingi (the Icelandic Parliament) and the University of Iceland Research Fund.

Contribution statement

EG contributed to the analysis, interpretation of results, and manuscript drafting and revision. SH and TA contributed to the interpretation of results, statistical analysis and revision of the draft manuscript. RK contributed to the analysis, interpretation of results and manuscript revision. GE contributed to data collection and management, analysis, and revision of the manuscript. BEK contributed to the analysis, interpretation of results and manuscript revision. TBH and LJL contributed to data acquisition and study design, results interpretation and revision of the draft manuscript. VG contributed to data acquisition and study design, the interpretation of results and manuscript revision. MFC contributed to study design and analysis, data collection and acquisition, the interpretation of results, and manuscript drafting and revision. FJ contributed to study design, analysis, interpretation of results, manuscript drafting and revision of the manuscript. All the authors approved the final version of the manuscript.

Duality of interest

The authors declare that there is no duality of interest associated with this manuscript.

References

  1. 1.
    Danaei G, Finucane MM, Lu Y et al (2011) National, regional, and global trends in fasting plasma glucose and diabetes prevalence since 1980: systematic analysis of health examination surveys and epidemiological studies with 370 country-years and 2.7 million participants. Lancet 378:31–40PubMedCrossRefGoogle Scholar
  2. 2.
    Wild S, Roglic G, Green A, Sicree R, King H (2004) Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 27:1047–1053PubMedCrossRefGoogle Scholar
  3. 3.
    World Health Organization (2006) Prevention of blindness from diabetes mellitus. Report of a WHO consultation. WHO, Geneva. http://whqlibdoc.who.int/publications/2006/924154712x_eng.pdf. Accessed 13 May 2011
  4. 4.
    Klein R, Klein BE, Moss SE, Wong TY (2006) The relationship of retinopathy in persons without diabetes to the 15-year incidence of diabetes and hypertension: Beaver Dam Eye Study. Trans Am Ophthalmol Soc 104:98–107PubMedGoogle Scholar
  5. 5.
    Ojaimi E, Nguyen TT, Klein R et al (2011) Retinopathy signs in people without diabetes: the multi-ethnic study of atherosclerosis. Ophthalmology 118:656–662PubMedCrossRefGoogle Scholar
  6. 6.
    Cugati S, Cikamatana L, Wang JJ, Kifley A, Liew G, Mitchell P (2006) Five-year incidence and progression of vascular retinopathy in persons without diabetes: the Blue Mountains Eye Study. Eye 20:1239–1245PubMedCrossRefGoogle Scholar
  7. 7.
    Thorsson B, Aspelund T, Harris TB, Launer LJ, Gudnason V (2009) Trends in body weight and diabetes in forty years in Iceland. Laeknabladid 95:259–266PubMedGoogle Scholar
  8. 8.
    Kristinsson JK, Stefansson E, Jonasson F, Gislason I, Bjornsson S (1994) Screening for eye disease in type 2 diabetes mellitus. Acta Ophthalmol (Copenh) 72:341–346CrossRefGoogle Scholar
  9. 9.
    Olafsdottir E, Stefansson E (2007) Biennial eye screening in patients with diabetes without retinopathy: 10-year experience. Br J Ophthalmol 91:1599–1601PubMedCrossRefGoogle Scholar
  10. 10.
    Gunnlaugsdottir E, Arnarsson A, Jonasson F (2008) Prevalence and causes of visual impairment and blindness in Icelanders aged 50 years and older: the Reykjavik Eye Study. Acta Ophthalmol 86:778–785PubMedCrossRefGoogle Scholar
  11. 11.
    Gunnlaugsdottir E, Arnarsson A, Jonasson F (2010) Five-year incidence of visual impairment and blindness in older Icelanders: the Reykjavik Eye Study. Acta Ophthalmol 88:358–366PubMedGoogle Scholar
  12. 12.
    Harris TB, Launer LJ, Eiriksdottir G et al (2007) Age, gene/environment susceptibility—Reykjavik Study: multidisciplinary applied phenomics. Am J Epidemiol 165:1076–1087PubMedCrossRefGoogle Scholar
  13. 13.
    Jonasson F, Arnarsson A, Eiriksdottir G et al (2011) Prevalence of age-related macular degeneration in old persons: age, gene/environment susceptibility—Reykjavik Study. Ophthalmology 118:825–830PubMedCrossRefGoogle Scholar
  14. 14.
    Qiu C, Cotch MF, Sigurdsson S et al (2009) Microvascular lesions in the brain and retina: the age, gene/environment susceptibility—Reykjavik Study. Ann Neurol 65:569–576PubMedCrossRefGoogle Scholar
  15. 15.
    World Health Organization (2011) Use of glycated haemoglobin (HbA1c) in the diagnosis of diabetes mellitus. Abbreviated report of a WHO consultation. WHO/NMH/CPM/11.1 2011. WHO, GenevaGoogle Scholar
  16. 16.
    World Health Organization (1999) Expert Committee Definition dacodmaic. Report of a WHO consultation, part 1: diagnosis and classification of diabetes mellitus. WHO, GenevaGoogle Scholar
  17. 17.
    American Diabetes Association (2010) Diagnosis and classification of diabetes mellitus. Diabetes Care 33(Suppl 1S):62–69CrossRefGoogle Scholar
  18. 18.
    Warnick GR, Knopp RH, Fitzpatrick V, Branson L (1990) Estimating low-density lipoprotein cholesterol by the Friedewald equation is adequate for classifying patients on the basis of nationally recommended cutpoints. Clin Chem 36:15–19PubMedGoogle Scholar
  19. 19.
    Olafsdottir E, Aspelund T, Sigurdsson G et al (2011) Effects of statin medication on mortality risk associated with type 2 diabetes in older persons: the population-based AGES—Reykjavik Study. BMJ Open 1:e000132PubMedCrossRefGoogle Scholar
  20. 20.
    Early Treatment Diabetic Retinopathy Study Research Group (1991) Grading diabetic retinopathy from stereoscopic color fundus photographs—an extension of the modified Airlie House classification. ETDRS report number 10. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology 98:786–806Google Scholar
  21. 21.
    Klein R, Davis MD, Magli YL, Segal P, Klein BE, Hubbard L (1991) The Wisconsin age-related maculopathy grading system. Ophthalmology 98:1128–1134PubMedGoogle Scholar
  22. 22.
    Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D (1999) A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 130:461–470PubMedGoogle Scholar
  23. 23.
    Wong TY, Klein R, Islam FM et al (2006) Diabetic retinopathy in a multi-ethnic cohort in the United States. Am J Ophthalmol 141:446–455PubMedCrossRefGoogle Scholar
  24. 24.
    Zhang X, Saaddine JB, Chou CF et al (2010) Prevalence of diabetic retinopathy in the United States, 2005–2008. JAMA 304:649–656PubMedCrossRefGoogle Scholar
  25. 25.
    Klein R, Klein BE, Moss SE, Linton KL (1992) The Beaver Dam Eye Study. Retinopathy in adults with newly discovered and previously diagnosed diabetes mellitus. Ophthalmology 99:58–62PubMedGoogle Scholar
  26. 26.
    Mitchell P, Smith W, Wang JJ, Attebo K (1998) Prevalence of diabetic retinopathy in an older community. The Blue Mountains Eye Study. Ophthalmology 105:406–411PubMedCrossRefGoogle Scholar
  27. 27.
    Klein R, Klein BE, Moss SE, Davis MD, DeMets DL (1984) The Wisconsin Epidemiologic Study of Diabetic Retinopathy III. Prevalence and risk of diabetic retinopathy when age at diagnosis is 30 or more years. Arch Ophthalmol 102:527–532PubMedCrossRefGoogle Scholar
  28. 28.
    Varma R, Macias GL, Torres M, Klein R, Pena FY, Azen SP (2007) Biologic risk factors associated with diabetic retinopathy: the Los Angeles Latino Eye Study. Ophthalmology 114:1332–1340PubMedCrossRefGoogle Scholar
  29. 29.
    Klein R, Klein BE, Moss SE, Cruickshanks KJ (1994) Relationship of hyperglycemia to the long-term incidence and progression of diabetic retinopathy. Arch Intern Med 154:2169–2178PubMedCrossRefGoogle Scholar
  30. 30.
    Danielsen R, Helgason T, Jonasson F (1983) Prognostic factors and retinopathy in type 1 diabetics in Iceland. Acta Med Scand 213:323–326PubMedCrossRefGoogle Scholar
  31. 31.
    Wong TY, Coresh J, Klein R et al (2004) Retinal microvascular abnormalities and renal dysfunction: the Atherosclerosis Risk in Communities Study. J Am Soc Nephrol 15:2469–2476PubMedCrossRefGoogle Scholar
  32. 32.
    Qiu C, Cotch MF, Sigurdsson S et al (2008) Retinal and cerebral microvascular signs and diabetes: the Age, Gene/Environment Susceptibility—Reykjavik Study. Diabetes 57:1645–1650PubMedCrossRefGoogle Scholar
  33. 33.
    Cheung N, Wong TY (2008) Diabetic retinopathy and systemic vascular complications. Prog Retin Eye Res 27:161–176PubMedCrossRefGoogle Scholar
  34. 34.
    Tapp RJ, Tikellis G, Wong TY et al (2008) Longitudinal association of glucose metabolism with retinopathy: results from the Australian Diabetes Obesity and Lifestyle Study. Diabetes Care 31:1349–1354PubMedCrossRefGoogle Scholar
  35. 35.
    Gallagher EJ, LeRoith D, Bloomgarden ZT (2009) Review of hemoglobin A(1c) in the management of diabetes. J Diabetes 1:9–17PubMedCrossRefGoogle Scholar

Copyright information

© Springer-Verlag 2011

Authors and Affiliations

  • E. Gunnlaugsdottir
    • 1
    • 2
  • S. Halldorsdottir
    • 3
  • R. Klein
    • 4
  • G. Eiriksdottir
    • 3
  • B. E. Klein
    • 4
  • R. Benediktsson
    • 2
    • 5
  • T. B. Harris
    • 6
  • L. J. Launer
    • 6
  • T. Aspelund
    • 2
    • 3
  • V. Gudnason
    • 2
    • 3
  • M. F. Cotch
    • 7
  • F. Jonasson
    • 1
    • 2
  1. 1.University Eye DepartmentLandspitalinnReykjavikIceland
  2. 2.Faculty of MedicineUniversity of IcelandReykjavikIceland
  3. 3.Icelandic Heart AssociationKopavogurIceland
  4. 4.Department of Ophthalmology and Visual SciencesUniversity of Wisconsin School of Medicine and Public HealthMadisonUSA
  5. 5.Department of Endocrinology and MetabolismLandspitali University HospitalReykjavikIceland
  6. 6.Laboratory of Epidemiology, Demography and Biometry, Intramural Research ProgramNational Institute of AgeingBethesdaUSA
  7. 7.Division of Epidemiology and Clinical ApplicationsBethesdaUSA

Personalised recommendations