FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2010053880> ?p ?o ?g. }

• W2010053880 endingPage "721" @default.
• W2010053880 startingPage "716" @default.
• W2010053880 abstract "Purpose: To determine the prevalence of visual impairment, retinopathy and macular oedema, and assess risk factors for retinopathy in persons with diabetes. Methods: The present study included 514 participants with diabetes aged 46–87 years from the Tromsø Eye Study, a sub-study of the population-based Tromsø Study in Norway. Visual acuity was measured using an auto-refractor. Retinal images from both eyes were graded for retinopathy and macular oedema. We collected data on risk factor exposure from self-report questionnaires, clinical examinations, laboratory measurements and case note reviews. Regression models assessed the cross-sectional relationship between potential risk factors and diabetic retinopathy. Results: The prevalence of visual impairment (corrected Snellen visual acuity <20/60 in the better-seeing eye) was 1.6%. The prevalence of diabetic retinopathy was 26.8% and macular oedema 3.9%. In a multivariable logistic regression model, retinopathy was associated with longer diabetes duration (odds ratio, OR 1.07, 95% CI 1.03–1.11), insulin use (OR 2.14, 95% CI 1.19–3.85), nonfasting glucose (OR 1.07, 95% CI 1.00–1.15) and microalbuminuria (OR 1.89, 95% CI 1.28–2.81). Sub-group analyses showed association between retinopathy and even low levels of microalbuminuria (1.16 mg/mmol). Conclusion: The findings suggest that low levels of microalbuminuria may be a useful risk predictor for identifying individuals with diabetes at high risk of retinopathy. The study confirms previous findings that insulin use, longer diabetes duration and higher levels of blood glucose are associated with retinopathy in persons with diabetes. The prevalence of diabetic retinopathy was similar as reported in other studies. Diabetic retinopathy is a major long-term complication of diabetes and a major cause of visual impairment and blindness (Kahn et al. 1977; Buch et al. 2001, 2004). The prevalence of diabetes and diabetic retinopathy increases worldwide owing to changes in lifestyle and an ageing population, and prevalence estimates for the next decades predict an even more dramatic increase (Wild et al. 2004; Saaddine et al. 2008; Danaei et al. 2011). The future public health burden of diabetic retinopathy will therefore be substantial. Randomized trials have found intensive diabetes care to reduce the risk of microvascular complications from diabetes (DCCT Research Group 1995; UKPDS 1998; DCCT/EDIC Research Group 2002). Epidemiologic studies have shown that this has resulted in a decreased prevalence of retinopathy in type 1 diabetes. However, it is less certain that a similar decline in the prevalence of retinopathy among persons with type 2 diabetes has occurred at the population level (Klein & Klein 2010). Previous population-based studies have reported wide range of prevalence rates of diabetic retinopathy in various countries and years (Williams et al. 2004). Recent population-based data on the prevalence of diabetes retinopathy, especially in type 2 diabetes, are therefore warranted. Identification of risk factors for diabetic retinopathy is imperative to implement preventive measures. Further, the presence of retinopathy is associated with cardiovascular disease, all-cause mortality (Kramer et al. 2011), stroke (Cheung et al. 2007) and other systemic vascular complications (Cheung & Wong 2008). Better understanding of risk factors for diabetic retinopathy may therefore provide important insight into the systemic complications of diabetes. Studies on risk factors for diabetic retinopathy have reported conflicting findings, including the role of serum lipids and microalbuminuria (Rema et al. 2006; Wong et al. 2008; Gunnlaugsdottir et al. 2011; Xu et al. 2011). Additional studies are required to examine risk factors for diabetic retinopathy. We aimed to determine the prevalence of visual impairment, retinopathy and macular oedema and assess risk factors for retinopathy in participants with diabetes from the Tromsø Eye Study, a sub-study of the population-based Tromsø Study in Norway. The Tromsø Study and the Tromsø Eye Study design and methodology have been described in detail in previous publications (Bertelsen et al. accepted) (Jacobsen et al. 2011). In short, the Tromsø Eye Study is a sub-study of the Tromsø Study, which is a longitudinal, population-based multipurpose study with repeated cross-sectional surveys from the municipality of Tromsø, Norway. Tromsø Eye Study was conducted from October 2007 through December 2008. The study followed the tenets of the Declaration of Helsinki for research involving humans and was approved by the Regional Committee for Medical Research Ethics. All participants gave an informed written consent. The recruitment strategies and sampling for the different visits are described in detail in previous publications (Bertelsen et al. accepted) (Jacobsen et al. 2011). The study sample consists of a combination of whole birth cohorts supplemented with random sample from the Tromsø population. A total of 7307 subjects participated in the Tromsø Eye Study (TES) (attendance rate 64% of eligible population). Participants were interviewed about eye diseases, and corrected visual acuity was measured by a Nidek AR 660A auto-refractor (Nidek Co., Ltd., Gamagori, Japan) using the built-in Snellen charts ranging from 20/200 to 20/20. Retinal photography was performed in both eyes after mydriasis with a Visucam PRONM (Carl Zeiss Meditec, Jena, Germany) digital retinal camera. Five field’s 45 degree colour retinal photographs were taken using the camera preset internal fixation and one 30 degree was taken centred on the macula. All images were graded for retinopathy according to ‘The International Clinical Diabetic Retinopathy and Diabetic Macular Edema Disease Severity Scales’ (Wilkinson et al. 2003), with minor modifications owing to the lack of stereo photographs. Diabetic macular oedema was therefore defined by the presence or absence of hard exudates and the distance from the centre of the fovea or the presence of grid laser photocoagulation burns in the macular area. In addition, micro aneurisms, haemorrhages, hard exudates, soft exudates, new vessels on the disc (NVD) and new vessels elsewhere (NVE) were quantified. Photographs with presence of laser photocoagulation burns indicating pan retinal laser photocoagulation were classified as proliferative retinopathy. Other eye diseases known to cause retinal haemorrhages or other findings similar to diabetic retinopathy were also classified and both eyes on each participant were graded consecutively. In case of retinal vessel occlusions, the grading of the fellow eye was used in the analyses. The grading was performed by one single grader (GB) and masked for other variables except for visual acuity, self-reported cataract, glaucoma and age-related macular degeneration (AMD). For intragrader assessment, a random sample of 200 participants (400 eyes) was generated with all stages of pathology represented mixed with 42.5% normal photographs. Exact agreement was 89% and kappa 0.81 when compared with the final grading. The Tromsø Study holds a diabetic registry where possible cases of diabetes are validated by reviewing medical records at the only local hospital in the region. In the present study, diabetes was defined as self-reported diabetes, nonfasting blood glucose ≥11.1 mmol/l, Hb1Ac >6.5% or a diabetes diagnosis in the diabetes registry. Information on diabetes type was collected from the diabetes registry. Information on diabetes treatment and diabetes duration was collected from both the diabetes registry and questionnaires. In case of discrepancy, the most extensive treatment and longest diabetes duration were used. Anthropometric measurements and blood pressure were obtained by clinical examination. Blood pressure (BP) was measured by trained technicians, using an automatic device (Dinamap Vital Signs Monitor, Tampa, FL, USA). Three consecutive blood pressure measurements were carried out at 1-min intervals and the mean of the two last measurements was used in the analyses. Body mass index (BMI) was calculated by dividing body weight (kg) with the square of height (m). Smoking habits, diet, supplements and medical history were obtained by questionnaires. Blood and urine samples were collected at the clinical examination. All laboratory measurements were performed at the University Hospital of North Norway. Microalbuminuria: Renal albumin excretion was assessed as albumin-creatinine ratio (ACR). Three separate urine samples of morning spot urine from three consecutive days were collected and analysed within 20 hr. Urine creatinine was measured using colometric methods (Jaffes reaction) with an autoanalyzer (ABX PENTRA; Horiba ABX, Montpellier, France). Urine albumin concentration was measured with an immunoturbidimetric method, on an ABX PENTRA autoanalyzer (Horiba ABX). ACR was calculated by dividing urine albumin concentration (mg/l) by creatinine (mmol/l). Mean of the three ACRs from three different days was defined as ACR. Microalbuminuria was defined as ACR >3.4 mg/mmol. Owing to the extremely skewed distribution of ACR with mainly low values and some very high, we also analysed log-transformed ACR. HbA1c: Nonfasting HbA1c was measured in EDTA whole blood by high performance liquid chromatography (HPLC) using an automated analyser (Variant II, Bio-Rad Laboratories Inc., Hercules, CA, USA). Cholesterol and triglycerides: nonfasting serum cholesterol and triglycerides were analysed by standard enzymatic colorimetric assay using an automated clinical chemistry analyser (Modular P, Roche Diagnostics, Mannheim, Germany). Descriptive statistics were calculated for demographic characteristics by diabetes treatment and retinopathy status. Student’s t-test was used for comparisons of means and chi-square test for comparisons of proportions. Multivariable and age- and sex-adjusted logistic regression models were used to calculate odds ratios (OD) and 95% confidence intervals for associations between retinopathy and risk factors. Post hoc analyses found the multivariable model to fit the data well (Hosmer–Lemeshow goodness of fit). StataSE version 12 (Stata Corp LP, College Station, TX, USA) was used for all statistical analysis. Of the 7307 participants in TES, 608 (8.3%) had diabetes. Participants without diabetes were excluded (n = 6699). A total of 81 participants with diabetes did not participate in photographic examination owing to technical and logistic reasons (n = 55) and no consent (n = 26). Additional 13 participants had ungradable images on both eyes, mainly owing to media opacities. The final sample consisted of 514 participants aged 46–87 years, of whom 18 participants had type 1 diabetes. The excluded diabetic participants were older (69.1 versus 66.4 years), had a higher proportion of smokers (24% versus 15%) and longer diabetes duration (7.69 versus 5.77 years) compared to the final sample. Table 1 shows the characteristics of the study population stratified by diabetes treatment. The insulin users were younger and had lower diastolic blood pressure and lower total cholesterol as compared with the two other groups combined. Further, the insulin users had higher frequency of lipid-lowering medication, microalbuminuria, diabetic retinopathy and diabetic macular oedema than the others. The overall prevalence of visual impairment (corrected Snellen visual acuity <20/60) was 1.64% in the better-seeing eye. Corrected Snellen visual acuity <20/40 was 4.11% in the better-seeing eye. We found no legally blind participants. The total prevalence of diabetic retinopathy was 26.8% and macular oedema 3.9% in participants with diabetes. Among participants with type 1 diabetes, 78% had retinopathy and the mean diabetes duration was 25.2 years, as compared to 25% with retinopathy and mean diabetes duration of 5.1 years in type 2 diabetes. The insulin-treated group had more than twice the frequency of retinopathy when compared with the oral-treated group (53.6% versus 19.6%). The overall prevalence of proliferative retinopathy was 1.2%. Table 2 shows the distribution of diabetic retinopathy and macular oedema by gender and age. There was no overall difference in diabetic retinopathy frequency between men and women, but men aged 55–64 had more retinopathy than women. Men had overall more macular oedema (p = 0.015), although numbers were small. There were too few cases of macular oedema to analyse by age groups. Significant differences in microalbuminuria, insulin use and diabetes duration were observed in participants with and without diabetic retinopathy (Table 3). Participants with diabetic retinopathy had lower levels of cholesterol but higher frequency of lipid-lowering medication. In logistic regression models adjusted for sex and age, diabetes duration, insulin use, pulse pressure, microalbuminuria, nonfasting glucose and HbA1c were associated with increased odds for retinopathy (Table 4). In contrast, higher total cholesterol and LDL cholesterol were associated with decreased odds for retinopathy. Statistical significant variables in the age- and sex-adjusted model were included in the multivariable model. LDL cholesterol was not included together with total cholesterol in the multivariable model owing to high correlation (Pearson’s r = 0.95). In the final multivariable model (Table 4), retinopathy was associated with insulin use, diabetes duration, microalbuminuria and nonfasting glucose, while total cholesterol and pulse pressure were no longer associated with retinopathy. When substituting total cholesterol with LDL cholesterol, the main effects did not change. When substituting nonfasting glucose with HbA1c in the multivariable model, HbA1c was no longer associated with retinopathy (p = 0.521). We also analysed associations between retinopathy and self-reported stroke, self-reported coronary heart disease, high-sensitive CRP, omega-3 fatty acid supplements, cod liver oil supplements, waist-to-hip ratio, anthropometric measurements, previous smokers and pack years of smoking, but did not find any significant associations (data not shown). To evaluate microalbuminuria cut-off level for increased retinopathy odds, we substituted log-transformed microalbuminuria with tertiles of ACR as a categorical variable in the final multivariable model. The lowest ACR tertile was used as reference. In this model, the highest ACR tertile was associated with retinopathy (OR third tertile; 2.90, 95% CI 1.62–5.21). When excluding participants with proteinuria (ACR >30 mg/mmol), the results were similar (OR third tertile; 2.80, 95% CI 1.55–5.05). The third ACR tertile ranged from 1.26 to 75.3 mg/mmol when proteinuria was included and 1.16–28.5 mg/mmol when proteinuria was excluded. The present study reports prevalence of visual impairment, retinopathy and macular oedema among the diabetic population in Tromsø. The study confirms results from previous studies on classical risk factors for retinopathy, and in addition found correlation between retinopathy and even very low levels of microalbuminuria. Microalbuminuria was highly significantly associated with retinopathy. Similar findings have been reported in other studies (Gunnlaugsdottir et al. 2011; Xu et al. 2011). When excluding participants with proteinuria (ACR >30 mg/mmol), ACR >1.16 mg/mmol was associated with increased risk of retinopathy. Other studies have reported association between low levels of microalbuminuria and increased risk of vascular disease (Hillege et al. 2002; Arnlov et al. 2005). A previous publication from the Tromsø Study found an increased risk of stroke at ACR concentration >0.43 mg/mmol in a nondiabetes sample (Solbu et al. 2009). When dividing the study population in quartiles or quintiles of ACR, we found an association with retinopathy at levels even lower than 1.16 mg/mmol, but owing to few cases in each group the estimates had large confidence intervals and were not significant for all groups. This suggests that the current cut-off for microalbuminuria (3.4 mg/mmol) may be too high and that even very low levels of microalbuminuria may be useful as a clinical tool for identifying diabetics with increased risk of retinopathy. The prevalence of diabetic retinopathy was 26.8%. Several other population-based studies reported 24.7–28.4% using similar methodology for retinopathy grading (Zhang et al. 2010; Gunnlaugsdottir et al. 2011; Kilstad et al. 2011; Xu et al. 2011), but the study sample was older in the AGES-R study. The Blue Mountains Eye Study found slightly higher prevalence (32.4%) (Mitchell et al. 1998), and The Wisconsin Study Epidemiologic Study of Diabetic Retinopathy (WESDR) found an even higher prevalence, but the participants in WESDR had a longer duration of diabetes (Klein et al. 1984). There was a low prevalence of proliferative diabetic retinopathy in the present study (1.2%), comparable to other population-based studies. In the Blue Mountains Eye Study, 1.6% had proliferative diabetic retinopathy (Mitchell et al. 1998), and the AGES-R study found 1% (Gunnlaugsdottir et al. 2011). The low prevalence of proliferative diabetic retinopathy may partly be explained by the cross-sectional design where participants without previously known diabetes were included, resulting in a high proportion of cases with short diabetes duration and less severe disease. In the age- and sex-adjusted analyses, we found significant associations between retinopathy and well-established risk factors such as insulin use, diabetes duration, HbA1c and nonfasting glucose (Klein et al. 1984; Mitchell et al. 1998; Stratton et al. 2001; Varma et al. 2007). We did not find any association between systolic or diastolic blood pressure and retinopathy. The reason for this may be the cross-sectional study design. In the present study, 56% of the participants used blood pressure treatment compared with 25.7% in nondiabetes participants in the Tromsø Eye Study. This will attenuate any effect of measured blood pressure on retinopathy. In addition, according to national diabetes care guidelines, more intensive blood pressure treatment in the most severe diabetes cases with nephropathy further attenuates the blood pressure effect on retinopathy. The type 1 diabetes participants also contribute to this finding by being younger and having lower diastolic BP (74.6 versus 77.8 mmHg) but five times longer diabetes duration compared with the type 2 diabetes. The literature on the association between serum lipids and retinopathy is more inconsistent. Total cholesterol was associated with increased odds for retinopathy in The Chennai Urban Rural Epidemiology Study (CURES) Eye Study and decreased odds in the Singapore Malay Eye Study (Rema et al. 2006; Wong et al. 2008). Other studies have not found any association between cholesterol and retinopathy, but some on triglycerides (Brown et al. 2003; Emanuele et al. 2005; Cikamatana et al. 2007; Xu et al. 2011). In the age- and sex-adjusted model, we found that increased level of total cholesterol were associated with decreased odds for retinopathy. Similarly, use of lipid-lowering medication was associated with increased odds for retinopathy. These somewhat unexpected effects were lost in the multivariable model. We believe the reason for the apparent protective effect of higher cholesterol in the age- and sex-adjusted analysis may be that lipid-lowering treatment was correlated with the diabetes severity shown in Table 1. In the multivariable model, HbA1c was, unexpectedly, not associated with retinopathy. The reason for this may be that the participants in this study were relatively well treated with anti-glycaemic medication (mean HbA1c 7.16%) and therefore other factors such as diabetes duration become more important as predictors for retinopathy. In addition, the effect of HbA1c on retinopathy is attenuated because we included some screening-detected and therefore untreated diabetes cases with very high HbA1c (up to 15.6%) and presumably short diabetes duration. In contrast, nonfasting glucose was associated with retinopathy in both regression models. This may support that both chronic elevated glucose levels as well as large variations in blood glucose are important factors in the development of retinopathy. Visual impairment was low and this is most likely underestimated owing to technical reasons. Visual acuity was measured with an auto-refractor and participants with visual impairment and blindness more often generated missing values compared with participants with normal visual acuity. On the other hand, we must expect that a subjective refraction might have given better visual acuity in some cases. If all missing values of visual acuity were owing to visual acuity below 20/40, the prevalence of visual acuity <20/40 would be 9.1%. The strengths of our study include the population-based design and comprehensive assessment of risk factors, relatively high attendance rate and high proportion of gradable retinal images. The cross-sectional design may be a limitation because it is not able to determine causality. A general concern in health surveys is the healthy participant effect and this may lead to underestimation of the prevalence of retinopathy, visual impairment and blindness. In conclusion, the present study provides prevalence estimates of retinopathy that are similar to a recent screening study in Norway and other epidemiologic studies in Caucasian and Asian populations. Visual impairment was low. To the best of our knowledge, this is the first study providing comprehensive data on risk factors of retinopathy in a Norwegian population. Our risk factor analyses were consistent with previous epidemiologic studies on retinopathy. An important finding is the strong association between microalbuminuria and retinopathy, and subgroup analyses suggest that even low levels of microalbuminuria are associated with retinopathy. The analysis of microalbuminuria is cheap and safe, and our data suggest that the clinical use may expand to evaluate risk for retinopathy even at very low microalbuminuria levels to identify patients with higher risk for retinopathy. TES has been financially supported by the Norwegian Extra Foundation for Health and Rehabilitation through EXTRA funds, the Research Council of Norway, University of Tromsø and the North Norway Regional Health Authority and Simon Fougner Hartmanns Familiefond. Tunde Peto would like to say thank you to NIHR BMRC in Ophthalmology for funding her participation in this study." @default.
• W2010053880 created "2016-06-24" @default.
• W2010053880 creator A5002705709 @default.
• W2010053880 creator A5018227698 @default.
• W2010053880 creator A5028457187 @default.
• W2010053880 creator A5042266966 @default.
• W2010053880 creator A5048696419 @default.
• W2010053880 creator A5067883594 @default.
• W2010053880 creator A5071483338 @default.
• W2010053880 creator A5079412313 @default.
• W2010053880 date "2012-09-20" @default.
• W2010053880 modified "2023-10-14" @default.
• W2010053880 title "Tromsø eye study: prevalence and risk factors of diabetic retinopathy" @default.
• W2010053880 cites W1910188440 @default.
• W2010053880 cites W1965211223 @default.
• W2010053880 cites W1966206262 @default.
• W2010053880 cites W1984790912 @default.
• W2010053880 cites W1988293663 @default.
• W2010053880 cites W1991096675 @default.
• W2010053880 cites W1993992694 @default.
• W2010053880 cites W2010456451 @default.
• W2010053880 cites W2016052123 @default.
• W2010053880 cites W2020267609 @default.
• W2010053880 cites W2025923126 @default.
• W2010053880 cites W2042363066 @default.
• W2010053880 cites W2049036672 @default.
• W2010053880 cites W2072690677 @default.
• W2010053880 cites W2081255185 @default.
• W2010053880 cites W2114774559 @default.
• W2010053880 cites W2120462273 @default.
• W2010053880 cites W2121176463 @default.
• W2010053880 cites W2123831918 @default.
• W2010053880 cites W2127508155 @default.
• W2010053880 cites W2130189158 @default.
• W2010053880 cites W2140651657 @default.
• W2010053880 cites W2141864382 @default.
• W2010053880 cites W2155336724 @default.
• W2010053880 cites W2159389777 @default.
• W2010053880 cites W2199339317 @default.
• W2010053880 cites W2337454357 @default.
• W2010053880 cites W2757208104 @default.
• W2010053880 cites W28087213 @default.
• W2010053880 cites W4248904215 @default.
• W2010053880 doi "https://doi.org/10.1111/j.1755-3768.2012.02542.x" @default.
• W2010053880 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/22994366" @default.
• W2010053880 hasPublicationYear "2012" @default.
• W2010053880 type Work @default.
• W2010053880 sameAs 2010053880 @default.
• W2010053880 citedByCount "37" @default.
• W2010053880 countsByYear W20100538802013 @default.
• W2010053880 countsByYear W20100538802014 @default.
• W2010053880 countsByYear W20100538802015 @default.
• W2010053880 countsByYear W20100538802016 @default.
• W2010053880 countsByYear W20100538802017 @default.
• W2010053880 countsByYear W20100538802018 @default.
• W2010053880 countsByYear W20100538802019 @default.
• W2010053880 countsByYear W20100538802020 @default.
• W2010053880 countsByYear W20100538802021 @default.
• W2010053880 countsByYear W20100538802022 @default.
• W2010053880 countsByYear W20100538802023 @default.
• W2010053880 crossrefType "journal-article" @default.
• W2010053880 hasAuthorship W2010053880A5002705709 @default.
• W2010053880 hasAuthorship W2010053880A5018227698 @default.
• W2010053880 hasAuthorship W2010053880A5028457187 @default.
• W2010053880 hasAuthorship W2010053880A5042266966 @default.
• W2010053880 hasAuthorship W2010053880A5048696419 @default.
• W2010053880 hasAuthorship W2010053880A5067883594 @default.
• W2010053880 hasAuthorship W2010053880A5071483338 @default.
• W2010053880 hasAuthorship W2010053880A5079412313 @default.
• W2010053880 hasBestOaLocation W20100538801 @default.
• W2010053880 hasConcept C118487528 @default.
• W2010053880 hasConcept C119767625 @default.
• W2010053880 hasConcept C134018914 @default.
• W2010053880 hasConcept C2779829184 @default.
• W2010053880 hasConcept C555293320 @default.
• W2010053880 hasConcept C71924100 @default.
• W2010053880 hasConceptScore W2010053880C118487528 @default.
• W2010053880 hasConceptScore W2010053880C119767625 @default.
• W2010053880 hasConceptScore W2010053880C134018914 @default.
• W2010053880 hasConceptScore W2010053880C2779829184 @default.
• W2010053880 hasConceptScore W2010053880C555293320 @default.
• W2010053880 hasConceptScore W2010053880C71924100 @default.
• W2010053880 hasIssue "8" @default.
• W2010053880 hasLocation W20100538801 @default.
• W2010053880 hasLocation W20100538802 @default.
• W2010053880 hasOpenAccess W2010053880 @default.
• W2010053880 hasPrimaryLocation W20100538801 @default.
• W2010053880 hasRelatedWork W1563850031 @default.
• W2010053880 hasRelatedWork W2367206690 @default.
• W2010053880 hasRelatedWork W2401570230 @default.
• W2010053880 hasRelatedWork W2415759662 @default.
• W2010053880 hasRelatedWork W2898850272 @default.
• W2010053880 hasRelatedWork W3036934084 @default.
• W2010053880 hasRelatedWork W3139175759 @default.
• W2010053880 hasRelatedWork W3190746054 @default.
• W2010053880 hasRelatedWork W4318422993 @default.
• W2010053880 hasRelatedWork W943310924 @default.
• W2010053880 hasVolume "91" @default.

• next