Vitamin D Deficiency Is Not Associated with Diabetic Retinopathy or Maculopathy

Background. Experimental and clinical studies suggest a possible association between vitamin D deficiency and both diabetic retinopathy and maculopathy. Methods. We have performed a cross-sectional study in adults with types 1 and 2 diabetes mellitus. The relationship between the presence and severity of diabetic retinopathy and maculopathy with serum 25-hydroxyvitamin D concentration was evaluated using logistic regression analyses in the presence of demographic and clinical covariates. Results. 657 adults with diabetes were stratified based on retinopathy grading: No Diabetic Retinopathy (39%), Background Diabetic Retinopathy (37%), Preproliferative Diabetic Retinopathy (21%), and Proliferative Diabetic Retinopathy (3%), respectively. There were no differences in serum 25-hydroxyvitamin D concentrations (25(OH)D) between the groups (15.3 ± 9.0 versus 16.4 ± 10.5 versus 15.9 ± 10.4 versus 15.7 ± 8.5 ng/mL, P = NS). Logistic regression analysis demonstrated no statistically significant relationship between the severity of retinopathy and serum 25(OH)D. Furthermore, there was no difference in serum 25(OH)D between those with (n = 94, 14%) and those without (n = 563, 86%) Diabetic Maculopathy (16.2 ± 10.0 versus 15.8 ± 9.8, P = NS) and no relationship was demonstrated by logistic regression analyses between the two variables. Conclusions. This study has found no association between serum 25(OH)D and the presence and severity of diabetic retinopathy or maculopathy.


Introduction
The prevalence of diabetic retinopathy (DR) approaches 93 million people worldwide [1] and is one of the leading causes of premature visual loss in the UK and worldwide [2]. Indeed, the World Health Organization estimates that whilst diabetic retinopathy accounts for approximately 5% of the global prevalence of blindness, the prevalence rises sharply to 15-17% in developed countries [3]. Several risk factors are implicated in the aetiology of DR with hyperglycemia and hypertension showing the strongest association [4], yet interventions aimed at correcting these risk factors have demonstrated moderate success [5,6]. Therefore, the interactions between neural and retinal vascular dysfunction and the mechanisms resulting in retinal pathology including neovascularisation have been questioned recently [7]. Furthermore, micronutrients including vitamin C, vitamin E, and magnesium have been postulated to play a role in DR [8].
Vitamin D deficiency has been linked to a host of cardiovascular diseases including diabetes and hypertension [9,10]. Vitamin D receptor (VDR) genotypes have been associated with the cumulative prevalence of diabetic retinopathy [11]. In two separate studies of the VDR gene in the French population, FokI and TaqI single nucleotide polymorphisms have been associated with DR [12,13]. In a study of Caucasians with C-peptide-negative type 1 diabetes, there was a novel association between the functional FokI VDR polymorphism and severe DR [12]. VDR dependent calcium binding proteins have been isolated in the human retina, particularly in the photoreceptor layer of the cones [14], and immunostaining in animal models has shown that VDR is expressed in the ganglion cells, the inner and outer plexiform layer, and the photoreceptor layer [15]. In an in vitro study of retinoblastoma tissue expressing VDR, supplementation with vitamin D resulted in a reduction of growth and apoptosis of the retinoblastoma cells [16]. 1,25-Dihydroxyvitamin D 3 (1,25(OH) 2 D 3 ) closely regulates Vascular Endothelial Growth Factor in experimental models [17] and there is an inverse correlation of 25(OH)D with Vascular Endothelial Growth Factor, postulated to be related to tissue hypoxia [18]. In a mouse model of ischaemic retinopathy, 1,25(OH) 2 D 3 was shown to inhibit neovascularisation in retinal tissue [19]. Vitamin D may also have a direct effect on the reninangiotensin-aldosterone-system and the renin-angiotensinaldosterone-system is known to be overexpressed in patients with type 1 diabetes and retinopathy [20] and blockade of this system reduces DR progression [21]. A Vitamin D analogue (paricalcitol) has shown an improvement in microalbuminuria through a mechanism related to inhibition of renin-angiotensin-aldosterone-system [22].
Aksoy et al. demonstrated an inverse correlation in a Turkish cohort between worsening diabetic retinopathy and lower 1,25-dihydroxyvitamin D 3 (active vitamin D) in a population of 66 subjects [23]. Furthermore, severe vitamin D deficiency has been shown to predict not only mortality but the development of nephropathy and retinopathy in type 1 diabetes mellitus [4]. In a recent cross-sectional study of children and adolescents with type 1 diabetes, retinopathy prevalence was higher in children and adolescents with lower levels of vitamin D [24]. Other cross-sectional studies which have assessed vitamin D status in relation to DR in adults either have had small numbers [25] or have been based on retrospective analysis of data collected from the National Health and Nutrition Examination Survey between 1988 and 1994 [26]. However, since then, the targets for glycaemia, blood pressure, and lipids have changed and also this study made no assessment of Diabetic Maculopathy [26]. Therefore, we have undertaken a study to establish the relationship between vitamin D status and the severity of DR and maculopathy in a large adult population with type 1 and type 2 diabetes.

Method
All patients attending clinics were assessed for the level of 25(OH)D, irrespective of a history suggestive of vitamin D deficiency. Written informed consent and ethical approval were not required as the data were extracted retrospectively and did not extend beyond standard clinical practice. All patient records and information were anonymised and deidentified prior to analysis. 25(OH)D was added as a standard routine test from June 2009 due to the high levels of deficiency noted. This was a retrospective analysis of data which had been collected already in our clinic for clinical rather than research reasons; that is, the patients with diabetes attending clinic underwent assessment of vitamin D as the clinical practice was to assess vitamin D in all patients and subsequently treat those who are deficient. These same patients were also undergoing retinal assessment as part of the annual review under the English retinal screening programme. The data (vitamin D and retinopathy grade) were not collected specifically for this analysis. There were a sample of 657 subjects in this retrospective study and prospective sample size analyses were inappropriate as all data available were assessed.

Subjects
All participants were aged ≥ 18 years attending clinics at the Central Manchester Foundation Trust, Manchester, and the assessment was conducted from August 2009 to May 2011. Those with renal impairment (eGFR <30 mL/min/1.73 m 2 (CKD stages 4 and 5)), granulomatous diseases (tuberculosis, sarcoidosis, etc.), and malabsorption syndromes (coeliac disease, bacterial overgrowth, and concomitant orlistat treatment), pregnant and lactating women, and those currently on vitamin D supplementation were excluded from the analysis. Biases were limited by using an unselected cohort of subjects, not based on symptomatology of vitamin D deficiency.

Blood Pressure and Anthropometric Measurements. Body
Mass Index (BMI) was measured as per the standard equation mass (kg)/(height(m)) 2 . Weight was measured with a digital scale (Seca 701, Seca, Hamburg, Germany) to the nearest 0.1 kg and height was measured to the nearest 0.1 cm. Blood pressure measurements were obtained with the use of an automated device (Dinamap pro 100v2, GE Medical Systems, Freiburg, Germany) with an appropriate cuff size. A minimum of two measurements of systolic and diastolic blood pressures were made five minutes apart with the lowest reading recorded and the mean of the preceding 2-year blood pressure results was used. Metabolic variables were also recorded with a mean of 2-year retrospective readings for glycosylated haemoglobin A1c (HbA1c) and components of the lipid profile (total cholesterol (CHL), high-density lipoprotein cholesterol [HDL-C], and triglycerides). The following measurements were taken as "spot readings" at the same date as baseline 25(OH)D measurements: Body Mass Index (BMI), bone profile markers such as corrected calcium (CCa), alkaline phosphatase (ALP), and estimated glomerular filtration rate (eGFR).

Assessment of Demographics, Cardiovascular Disease, and
Medications. An assessment of patient demographics, previous cardiovascular events, and medications were made through analysis of medical records and an in-hospital medical record database (Diamond database, Hicom, Surrey, UK). Subject demographics extracted were age, sex, ethnicity (Caucasian, South Asian, Far East Asian, and Afro-Caribbean descent), smoking status (never, previous, and current), and type (types 1 and 2 diabetes) and duration of diabetes. Dates of baseline 25(OH)D were used to obtain respective retinopathy screening data. Only retinopathy screening data within 1 year of the baseline 25(OH)D and prior to vitamin D supplementation were included.
Baseline 25(OH)D status and retinopathy data were collected for 657 patients who had attended their retinopathy screening appointments. The retinopathy data were collected according to the grading criteria of the National Screening Committee [27]. Previous studies have shown acceptable level of quality and accuracy of grading compared to expert graders within the English National Screening Committee [28]. The national guidelines do not contain R1.5 or M0.5 grades and are categorised as Preproliferative Diabetic Retinopathy and Diabetic Maculopathy, respectively. These subgradings were used locally in screening centres and have been included. Retinopathy was graded as follows:

25(OH)D Assay.
Serum was separated from whole blood and stored at −20 ∘ C until assay. The assay used was an automated platform assay (ImmunoDiagnostic Systems Ltd., Bolden, Tyne and Wear, UK) and is based on chemiluminescence technology. The assay was performed exactly as per the manufacturer's instructions. Briefly, samples were subjected to a pretreatment step to denature the vitamin D binding protein. The treated samples were then neutralised in assay buffer and a specific anti-25(OH)D antibody labelled with acridinium was added. Following an incubation step, magnetic particles linked to 25(OH)D were added. Following a further incubation step, the magnetic particles were "captured" using a magnet. After a washing step and addition of trigger reagents, the light emitted by the acridinium label was inversely proportional to the concentration of 25(OH)D in the original sample. Concentration of 25(OH)D was calculated automatically using a 4-point logistic curve. The reportable range of the assay was 5-140 ng/mL. Inter-and intra-assay variation of the in-house control were 5.6% and 9.7%, respectively.

Discussion
Vitamin D deficiency has wide ranging implications for insulin resistance, beta cell dysfunction, and hypertension and therefore provides a potential link with diabetic complications. Experimental studies have postulated an important link between vitamin D deficiency and retinopathy [30]  and an increased risk of diabetic retinopathy has been demonstrated in the presence of VDR polymorphisms [12]. However, our study has shown no relationship between the vitamin D status and the severity of diabetic retinopathy or maculopathy in a large cohort of patients with predominantly type 2 diabetes, after correcting for glycaemic control, blood pressure, and lipids. We confirm that the "usual culprits" of longer duration of diabetes, higher HbA1c, and systolic blood pressure are directly related to retinopathy and maculopathy [1], thereby providing confidence in the validity of our data. Furthermore, the metabolic and anthropometric measurements used in the regression analysis were taken over an extended period of time as opposed to "spot" readings taken in other studies [23,25]. A possible explanation for the lack of relationship between vitamin D deficiency and retinopathy could be the striking extent of vitamin D deficiency in this population, although this is consistent with our previous data [29]. Thus, the majority of patients demonstrated deficiency (∼90%) and indeed severe deficiency (∼31%). Therefore, any relationship between retinopathy and adequacy of vitamin D could not be explored adequately. Furthermore, there were only a small number of patients with Diabetic Maculopathy ( = 94, 14%) in this study, which ultimately limits the power of the analysis. Only a limited number of clinical studies have investigated the role of vitamin D deficiency in DR. In one of the earliest studies, Aksoy et al. showed an inverse relationship between 1,25(OH) 2 D 3 and worsening retinopathy, although the short half-life of 1,25(OH) 2 D 3 may limit the interpretation of any such relationship [23]. Another smaller North American study has shown that subjects with DR, in particular PDR, have lower levels of 25(OH)D [25]. Whilst in a recent study the percentage of individuals with vitamin D deficiency increased with the severity of retinopathy, regression analysis did not demonstrate a statistically significant relationship between retinopathy severity and serum 25(OH)D concentration [26]. In a prospective observational follow-up study of a cohort of patients with type 1 diabetes, although severe vitamin D deficiency independently predicted allcause mortality, it was not related to the development of either retinopathy or nephropathy [4]. In a recently published study of 715 patients with type 2 diabetes, serum 25(OH)D levels decreased significantly in relation to the severity of either retinopathy or nephropathy or both [31]. However, in the prospective EURODIAB study conducted in subjects with type 1 diabetes, both higher 25(OH)D 2 and 25(OH)D 3 were associated with a lower prevalence of macroalbuminuria, but not retinopathy and cardiovascular disease [32].
This large cross-sectional study found no association of vitamin D status with diabetic retinopathy or maculopathy. A population with a larger spread of vitamin D levels may provide further insight into a possible association, but this may not be possible due to the high prevalence of vitamin D deficiency. In the long term, randomised controlled trials of adequate vitamin D intervention and diabetic microvascular outcomes are required to truly assess any potential therapeutic benefit.