Introduction

Diabetic retinopathy and nephropathy are leading causes of blindness and end-stage renal disease in the working-age population of the western world [1]. Retinopathy is almost universal, and nephropathy is found in approximately 40% of all type 1 diabetic patients with a duration of diabetes of more than 25 years [2–4]. Several studies have investigated the relationship between diabetic retinopathy and nephropathy [5, 6], and similarities in the pathogenesis have been described [7].

Nephropathy without retinopathy is not very common as opposed to retinopathy without nephropathy [4, 5, 8]. For instance, in a cross-sectional study of 2,378 type 1 diabetic patients from the EURODIAB IDDM Complications Study it was found that among patients with macroalbuminuria, only 11.5% had no retinopathy [8]. Similar numbers were 60.9 and 44.9% for patients with normo- and microalbuminuria, respectively. Looking at things the other way around, the prevalence of macroalbuminuria was 1.6, 9.0 and 34.0% for patients with no retinopathy, non-proliferative retinopathy (NPDR) and proliferative retinopathy (PDR), respectively.

Even though many cross-sectional studies [4, 5, 8, 9] have examined the association of retinopathy and nephropathy, only a few studies are prospective [10, 11]. Furthermore, long-term studies on the temporal association between retinopathy and nephropathy are lacking. Thus, it has not been described whether proliferative retinopathy is a marker of nephropathy. This is important in order to clarify the pathogenic similarities between these microvascular complications.

Consequently, the aim of this study was to expand on the knowledge of the association between retinopathy and nephropathy in type 1 diabetes. Particularly, we aimed to investigate the influence of proliferative retinopathy on the long-term incidence of nephropathy in a population-based cohort of Danish type 1 diabetic patients.

Subjects, materials and methods

Participants

As previously described [12, 13], insulin prescriptions were used to identify all type 1 diabetic patients living in Fyn County, Denmark, as of 1 July 1973 with an onset before the age of 30 (n = 727). At that time Fyn County approximately had 450,000 inhabitants and was considered a demographically representative 9% sample of Denmark [13]. One hundred and ninety nine patients participated in clinical examinations in both 1981–1982 (baseline) and 2007–2008 (follow-up). Of these, 15 patients already had proteinuria at baseline and were excluded from further analyses. All patients examined at the follow-up gave a written informed consent.

Baseline examination

At baseline, patients underwent a structured interview, and an ophthalmological as well as a clinical examination was performed [12]. Blood pressure was measured by an Erkameter Sphygmomanometer (Morton Medical Ltd, London, UK) on one arm with the patient in sitting position after 10 min of rest. Measurements of the blood included HbA1c made as total Hb-A1 with resin 70 (Bio-Rad, Hercules, CA, USA) at 20°C and pH = 6.70. Urine protein was considered present if greater than or equal to 0.5 g/l protein was found in a spot urine sample. The smoking habits of the patients were noted. Current and ex-smokers were considered to be smokers for the upcoming models. Using tropicamide 1% both pupils were dilated and a slit lamp examination performed (Haag-Streit, Wedel, Germany). Ophthalmoscopy was performed and retinopathy was described and classified by a single trained retinal specialist (AK Sjølie). The patient’s level of retinopathy was determined by the worse eye and classified as no retinopathy, NPDR or PDR. PDR was defined as newly formed vessels or the presence of photocoagulation scars after panretinal photocoagulation.

Registry based data

In order to evaluate renal failure between baseline and the follow-up examination, data were obtained from The Danish National Patient Registry. Renal failure was defined as hospital contacts due to dialysis or renal transplantations. Data on mortality were provided by The Danish Civil Registration System.

Follow-up examination

At follow-up, nephropathy was evaluated in a single spot urine sample. Definitions were the following: normoalbuminuria 0–19 mg/l, microalbuminuria 20–200 mg/l and macroalbuminuria 201 mg/l and above. Pupils were dilated and digital fundus photographs were taken in both eyes. Nine 45° colour fields were captured with Topcon TRC-NW6S (Topcon, Tokyo, Japan) and auto-mosaicked with IMAGEnet (Topcon, Tokyo, Japan). Photographs were graded with a grading protocol [2, 14, 15] according to the Early Treatment Diabetic Retinopathy Study (ETDRS) adaptation of the modified Airlie House classification of diabetic retinopathy [16, 17]. For the upcoming analyses, ETDRS-levels 10–31 were classified as no or mild NPDR, 37–53 as moderate NPDR and 60–85 as PDR. The final category of the patient was determined by the worst eye. All images were graded at the Ocular Epidemiology Research Services, Madison, Wisconsin. Interobserver and intraobserver variations and the validity of the systems have been presented elsewhere [14–16, 18].

The study was approved by the Regional Committee on Biomedical Research Ethics of Southern Denmark.

Statistical analyses

Continuous data are presented as median, and categorical data are presented as per cent. All statistical differences were tested at the 0.05 level. Confidence intervals that do not cross 1.0 were regarded statistically significant. Chi-square tests were used for categorical data, and for continuous data Mann–Whitney test was used to test for differences between two groups.

A multivariate logistic regression was performed to calculate odds ratio for incident nephropathy (micro- or macroalbuminuria) according to the level of PDR at baseline. All patients without proteinuria at baseline who participated at follow-up were considered at risk. Stata Intercooled 9.2 was used for all analyses.

Results

One hundred and ninety nine patients were examined in 1981–1982 and again in 2007–2008. An additional 265 patients were examined at baseline but died prior to the follow-up examination. In general, non-survivors were older, had a longer duration of diabetes, a higher systolic blood pressure, and were more likely to be smokers and to have PDR (Table 1). On the other hand, there was no difference in gender, glycaemic regulation and diastolic blood pressure between patients who participated at the follow-up or died prior to this. Furthermore, there was no statistically significant difference (p = 0.19) between non-survivors (6.8%, 18 of 265) and follow-up participants (4.0%, 8 of 201) in the risk of developing renal failure between baseline and follow-up.

Table 1 Baseline characteristics for patients who participated at the follow-up examination and patients who died prior to this
Full size table

Fifteen patients already had proteinuria at baseline and were thus excluded. Of the 184 patients included in the study, 111 (60.3%) were men and 73 (39.7%) were women. Median age and duration of diabetes at baseline was 33.0 and 16 years, respectively.

At baseline, PDR was present in 29 patients (15.8%). As compared to patients without PDR at baseline, patients with PDR had a longer duration of diabetes (23 vs. 15 years, p < 0.01) (Table 2). There was no statistically significant difference between the groups according to age, sex, smoking, HbA1, systolic or diastolic blood pressure.

Table 2 Characteristics according to level of proliferative diabetic retinopathy at the baseline examination in 1981–1982
Full size table

At the follow-up examination, the median age and duration of diabetes was 58.8 and 43 years, respectively, and the number of patients with normo-, micro- and macroalbuminuria was 127 (69.0%), 41 (22.3%) and 16 (8.7%), respectively (Table 3). The level of albuminuria was higher for patients with PDR at baseline (p = 0.04). For instance, macroalbuminuria was present in 20.7 and 6.5% of patients with and without PDR at baseline, respectively. In a multivariate analysis adjusted for baseline-levels of age, sex, duration of diabetes, smoking, HbA1, systolic and diastolic blood pressure, the odds ratio of nephropathy (micro- and macroalbuminuria combined) was 2.98 (95% confidence interval 1.18–7.51, p = 0.02) for patients with baseline PDR as compared to those without.

Table 3 Level of nephropathy at follow-up according to presence or absence of proliferative diabetic retinopathy at baseline
Full size table

At the follow-up examination, the level of nephropathy was correlated to the level of retinopathy (p < 0.01) (Table 4). For instance, normoalbuminuria was found in 82.9 and 81.8% of patients with no or mild NPDR, and moderate NPDR, respectively. Patients with PDR, however, were more likely to have albuminuria. Macroalbuminuria was found in 13.0% of patients with PDR, as compared to 4.3 and 4.6% of patients with no or mild NPDR, and moderate NPDR, respectively.

Table 4 Level of nephropathy according to level of retinopathy at follow-up
Full size table

Discussion

Diabetes may lead to various micro- and macrovascular complications [19–22]. A positive association between diabetic retinopathy and nephropathy has been well established in several studies [2, 4–6, 8, 10]. However, this is as far as we know the first long-term prospective study to examine a prospective link between retinopathy and nephropathy in type 1 diabetes. In the present population-based 25-year prospective follow-up study, we found that patients with proliferative retinopathy were three times as likely to develop macroalbuminuria as those without. These findings could not be explained by differences in other possible risk factors for nephropathy and suggest an underlying link between proliferative retinopathy per se and macroalbuminuria in type 1 diabetes.

Brownlee has presented a possible unifying pathogenic mechanism in which the endothelium in retinas capillaries and the mesangial cells in the kidneys glomeruli are both affected [7]. The hyperglycaemia-induced process that leads to tissue damage is caused by an overproduction of superoxide. This leads to oxidative stress and an increased production of reactive oxygen species that causes the damage to the affected cells.

Several studies have found a striking difference in prevalence of retinopathy and nephropathy in type 1 diabetes. Nephropathy rarely occurs without the retinopathy, but retinopathy is often seen in patients without nephropathy [4, 5, 8]. This could be explained by a higher vulnerability of the retina as compared to the kidneys. It is likely that the retina has a greater ability to produce these harmful metabolites as compared to the kidneys, or a reduced ability to repair tissue damage. Another explanation could be that retinal lesions (i.e. microaneurysms) are easier to detect, given that microvasculature is only available for in vivo inspection in the eye. Microalbuminuria reflects a structural damage of the kidneys as compared to retinal microaneurysms which are to some extent reversible. Even though it has been demonstrated that microaneurysms are positively related to progression of retinopathy [23], it has also been found that only 58% of all microaneurysms are still present after 1 year [24]. This reversibility might explain, at least partly, the earlier detection of retinopathy than nephropathy.

Despite the differences in the prevalence rates, several studies have demonstrated a close connection between proteinuria and severe retinopathy [2, 5, 10]. In a 25-year follow-up of the Wisconsin Epidemiological Study of Diabetic Retinopathy, Klein et al. found that the risk of incident PDR was twice as high for patients with microalbuminuria at baseline as in those who had normoalbuminuria [2]. Similar findings were demonstrated in a Spanish follow-up study where the presence of microalbuminuria was a close marker for severe forms of retinopathy (odds ratio 6.31) [10]. Both studies analysed the effect of microalbuminuria as a predictor of proliferative retinopathy, but studies of the reverse relationship are few.

A cross-sectional study from Saudi Arabia found similar results as ours [25]. Thus, the estimated relative risk of nephropathy was 45.5 in patients with PDR compared with patients without retinopathy, suggesting that proliferative retinopathy is a risk indicator for nephropathy. A small Swedish retrospective follow-up study demonstrates results to the contrary. None of the patients with PDR at the baseline developed persistent microalbuminuria within the observation period of 10 years [11]. An interesting aspect of this Swedish study was that all of the patients with PDR at baseline had a diastolic blood pressure under 85 mmHg. Several studies have shown that hypertension is involved in the association between retinopathy and nephropathy [4, 5, 8], which might explain the lack of relationship between retinopathy and nephropathy in the Swedish study.

The major strengths of the current study are the population-based design and the long follow-up. It has been stated in several studies that macroalbuminuria often does not appear during the first 15 years of diabetes [1, 4]. Therefore, a long follow-up and a long duration of diabetes are important to be able accurately to evaluate the long-term effect of retinopathy on nephropathy. The present study also has some limitations. First, survival bias must be considered. All patients at follow-up had survived more than 30 years of diabetes. Since proliferative retinopathy has been proposed as an independent marker of mortality rate among type 1 diabetic patients [26, 27], selective mortality is a concern in the present study. Second, the methods for evaluating retinopathy and proteinuria may lead to inaccuracy. At baseline, retinopathy was evaluated from ophthalmoscopy, which has a lower sensitivity as compared to fundus photographs that are considered the gold standard of today [28]. At both examinations, proteinuria was measured from a single spot urine. The gold standard for evaluating proteinuria is either to use a 24 h collection of urine or to use multiple collections due to the substantial intra-individual and day variability in urinary albumin excretion rate [29]. This was, unfortunately, not possible in the present study.

In conclusion, a close association was found between proliferative diabetic retinopathy and subsequent development of diabetic nephropathy in type 1 diabetic patients. This study suggests that proliferative retinopathy in type 1 diabetic patients may be an independent marker for the long-term incidence of nephropathy. Further studies are needed to examine the applications of this finding.