Article Text

Download PDFPDF

Relationship of retinal vascular calibre and diabetic retinopathy in Chinese patients with type 2 diabetes mellitus: the Desheng Diabetic Eye Study
  1. Xiufen Yang1,2,
  2. Yu Deng2,
  3. Hong Gu2,
  4. Xuetao Ren2,3,
  5. Apiradee Lim4,
  6. Torkel Snellingen3,
  7. Xipu Liu3,
  8. Ningli Wang2,
  9. Jeong Won Pak5,
  10. Ningpu Liu2,
  11. Ronald P Danis5
    1. 1Department of Ophthalmology, The Friendship Hospital, Capital Medical University, Beijing, China
    2. 2Beijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical University, Beijing Ophthalmology and Visual Sciences Key Laboratory, Beijing, China
    3. 3Sekwa Research Institute, Beijing, China
    4. 4Department of Mathematics and Computer Science, Faculty of Science and Technology, Prince of Songkla University, Muang Pattani, Thailand
    5. 5Department of Ophthalmology and Visual Sciences, Fundus Photograph Reading Center, University of Wisconsin, Madison, Wisconsin, USA
    1. Correspondence to Dr Ronald P Danis, Department of Ophthalmology and Visual Sciences, Fundus Photograph Reading Center, University of Wisconsin-Madison, 2870 University Avenue, Madison, WI 53705 USA; rpdanis{at}


    Aims To describe the relationship of retinal arteriolar and venular calibre with diabetic retinopathy (DR) and related risk factors, including glucose levels and other biomarkers in a Chinese population with type 2 diabetes mellitus (T2DM).

    Methods A cross-sectional study. Patients with T2DM were recruited from a local community in urban Beijing. Seven fields 30° colour fundus photographs were taken and examined for the presence and severity of DR using a standardised grading system. Retinal vascular calibres were measured and expressed as average central retinal arteriolar and venular equivalent using a computer-based program.

    Results A total of 1340 patients with T2DM were included for analysis. Of these, 472 (35.22%) had DR. Wider retinal venular calibre, but not arteriolar calibre, was associated with increasing glucose and glycosylated haemoglobin A1c levels (p<0.006) and dyslipidaemia (p for trend <0.05). After adjusting for possible covariates, the higher quartile of retinal venular calibre was associated with higher prevalence of any DR (OR 2, 95% CI 1.36 to 2.95). Venular calibre increased from 224.33 μm in those without retinopathy to 231.21 μm in those with mild, 241.01 μm in those with moderate and 235.65 μm in those with severe retinopathy (p for trend <0.001). Arteriolar calibre was not associated with DR.

    Conclusions In the current study, wider venular calibre, but not arteriolar calibre, was shown to be associated with development and increased severity of DR independently from other risk factors in a Chinese diabetic population.

    • Retina
    • Imaging
    • Epidemiology
    • Diagnostic tests/Investigation
    • Physiology

    Statistics from

    Though diabetic retinopathy (DR) is a leading cause of preventable vision loss among working-age people in many developed countries, the exact pathophysiological mechanisms of DR remain unclear and current methods to predict patients with diabetes at risk of DR remain inadequate. Additional predictors of retinopathy need to be identified in screening and monitoring for this complication.

    Changes in retinal vascular calibre are known to be one of the clinically and pathologically noted retinal changes associated with systemic disease and DR.1 Computer-based retinal image analysis has allowed documentation of subtle retinal calibre changes objectively and reliably.2 Alterations in retinal vascular calibre may predict a range of systemic diseases, such as cardiovascular diseases,3 stroke,4 diabetes mellitus and its complications.5 A number of cross-sectional6–12 and prospective studies13–18 have shown associations between variations of retinal vascular calibre and increased risk of development and progression of DR and raised the possibility of retinal vessel assessment as a novel risk marker in predicting the risk of DR. Larger retinal venular diameter, for example, has been found to predict the development,6–9 ,12 and progression of DR in most studies, but not in others.15 ,16 Widened retinal arteriolar calibre has been associated with the development of retinopathy,8 and has been found to be associated with the incidence14–16 and progression13 of DR. Conversely, a narrowed arteriolar calibre was also reported to be related to DR severity level in persons with type 1 diabetes mellitus (T1DM).6 Inconsistencies among studies may be related to methodological and statistical differences between reports, but the possibility of unaccounted confounding ocular and systemic factors, and population differences remains. A few studies have reported retinal vascular dimensions in Asian populations,11 ,12 but there are no data available regarding the association between retinal vascular calibre and DR from Chinese patients with T2DM.

    The present study therefore aimed to describe the relationship between vessel calibre and DR in a large cohort of Chinese patients with T2DM.

    Materials and methods

    Study population

    Patients with T2DM, as identified from an age-related eye disease screening programme in the Desheng community of urban Beijing, were recruited between November 2009 and April 2012 by using posters, pamphlets and phone calls. Diabetes was defined as either a self-reported history of physician diagnosed T2DM being treated with insulin, oral hypoglycaemic agents, or diet only, or by a fasting plasma glucose (FPG) concentration of 7.0 mmol/L (126 mg/dL) or more in at least two previous examinations or a random plasma glucose concentration of ≥11.1 mmol/L (200 mg/dL). The duration of diabetes was defined as the interval between the first diagnosis and the time of enrolment into the present study. Patients with severe media opacity or with shallow anterior chamber/angle-closure glaucoma precluding mydriasis were excluded. The study protocol was approved by the Ethics Committee of the Beijing Tongren Hospital and adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants before their enrolment.

    Participants underwent a standardised evaluation consisting of a questionnaire, ocular and anthropometric examinations and laboratory investigation. Basic demographic and lifestyle information (age, sex, income, educational level, smoking and alcohol intake) and medical history (such as medication, the use of insulin and history of systemic diseases) were elicited from the interview. The level of education was recorded as high school completed or not and family income was self-reported. Persons currently smoking more than one cigarette/cigar/pipe a day for at least 1 year were classified as current smokers. Persons who never smoked were classified as never smokers. Persons who had stopped smoking before the examination but with a prior smoking history were classified as ex-smokers.

    Anthropometric parameters include body weight and height, waist and hip circumference and three measurements, 5 min apart, of systolic blood pressure (SBP) and diastolic blood pressure (DBP) in a resting status. Hypertension was defined as mean SBP of ≥140 mm Hg, DBP of ≥90 mm Hg or taking antihypertensive medications. Height and weight were measured with subjects in light clothing and not wearing shoes by a trained observer. Body mass index (BMI, kg/m2) was calculated as the ratio between weight and the square of height of the participant. Waist-to-hip ratio was calculated as waist circumference divided by hip circumference. A comprehensive ophthalmological examination included visual acuity with habitual refraction and with pinhole, slit-lamp biomicroscopy and fundus examination. Seven fields 30° colour fundus photographs with stereoscopic images of optic disc and macula were taken through dilated pupils for all patients using a digital fundus camera (Zeiss Visucam Pro, Oberkochen, Germany). Overnight fasting blood samples were collected for measurement of FPG, glycosylated haemoglobin A1c (HbA1c), creatinine, uric acid, C-reactive protein (CRP) and lipid profile (levels of total cholesterol, triglycerides, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol). A first-void, midstream morning spot urine sample was collected, and albuminuria was measured by immunonephelometry (Roche/Cobas C501 Analyzer, Ibaraki, Japan). High albuminuria was defined as ≥20 mg/L.

    DR grading

    One trained ophthalmologist (XY) graded all the images at the University of Wisconsin Fundus Photographic Reading Center, according to the Early Treatment Diabetic Retinopathy Study (ETDRS) standard classification. Retinopathy was considered present if any characteristic lesions as defined by the ETDRS severity grading scale were present, including microaneurysms, haemorrhages, cotton wool spots, intraretinal microvascular abnormalities, hard exudates, venous beading and new vessels.19 A retinopathy severity score was assigned for each eye according to the ETDRS Diabetes Retinopathy Severity Scale and the score of the worse eye was used for analysis. Eyes were graded according to the following criteria: no DR (NDR, level 10) or any DR (levels 14 and above). DR was further divided into mild (levels 14–35), moderate (levels 43 through 47) and severe DR (level 53 and above). Grading reproducibility was assessed by regrading 5% of the eyes by a senior grader at the University of Wisconsin Fundus Photograph Reading Center. Exact agreement on retinopathy severity level was tabulated and weighted κ was calculated and compared with historical reproducibility from reading centre grading.

    Assessment of retinal vessel diameters

    Retinal vascular calibre was measured with a computer-based program (IVAN, developed by Nicola Ferrier, University of Wisconsin, Madison, USA) following a standardised protocol described in detail elsewhere.20 A single grader masked to participant characteristics performed all vessel measurements at the University of Wisconsin Fundus Photograph Reading Center. Photographs from ETDRS field 1 in the right eye (optic nerve head) were selected for measurement. If measurements could not be performed in the right eye, the left eye was chosen and the best quality image of the stereoscopic pair was used. All arterioles and venules coursing through an area 0.5–1 disc diameter from the optic disc margin were measured and summarised into average central retinal arteriolar equivalent and central retinal venular equivalent (CRAE and CRVE). The arteriole-to-venule ratio (AVR) for each subject was also calculated. Vessel diameters are presented in micrometers (µm). Reproducibility of the vessel diameter measurements was evaluated in this study; the intraclass correlation coefficient was 0.90 (95% CI −10.9 to 6.8) for CRAE and 0.95 (95% CI −9.2 to 6.7) for CRVE.

    Statistical analysis

    Statistical analysis was performed using the R statistical analysis package ( Categorical data are presented as percentages and continuous data are presented as mean and SD. Differences in clinical characteristics of participants with or without DR were assessed using the two-sample t test for continuous variables and χ2 test for categorical variables. Retinal vascular calibres (CRAE and CRVE) were analysed as continuous variables. Analysis of covariance (ANCOVA) and linear regression models were used to determine the association of various risk factors with retinal vascular calibre. The ANCOVA was initially used to estimate mean retinal vascular calibre in association with the presence versus absence of categorical variables (eg, high albuminuria) or quartiles of continuous variables (eg, glucose levels) adjusted for age and sex. Tests for trend were determined by treating categorical risk factors (eg, quartiles of glucose) as continuous ordinal variables and the χ2 test statistic for the parameter estimate was computed.

    For the outcome variable of DR, the following multivariable models were constructed. Model 1: adjustment for age and gender. Model 2: adjustment for variables in model 1 and variables with p value ≤0.2 in the univariate analysis shown in tables 1 and 2, including duration of diabetes, smoking status, SBP, DBP, HbA1c, creatinine, total cholesterol, HDL, CRP, use of insulin and albuminuria. Model 3: adjustment for variables in model 2 plus BMI. Results were expressed as p value, OR and 95% CI. Statistical significance was set at p<0.05. Given that laser treatment may affect the vessel calibre, the group of eyes with severe DR and laser treatment were either included or excluded in separate multivariate linear analyses.

    Table 1

    Clinical characteristics of participants with type 2 diabetes

    Table 2

    Relationship of significant characteristics with retinal arteriolar and venular calibre


    This study included 1433 subjects with T2DM, of which, retinal photographs from 93 participants could not be graded due to poor photographic quality and were excluded from the analysis. Consequently, a total of 1340 individuals with a mean age of 64.8±8.23 years were available for the current analysis, including 533 (39.78%) men and 807 (60.22%) women. Most of the participants (92.39%) were self-identified as Han Chinese (the remaining including persons of Manchu, Hui, Mongolian and Miao descent). Any DR was diagnosed in 35.22% (n=472) subjects, of which 383 (28.58%) with mild, 36 (2.69%) with moderate and 53 (3.96%) with severe retinopathy, as assessed by fundus photography. Thirty-four eyes of those with severe retinopathy had been treated with panretinal photocoagulation.

    Table 1 shows the characteristics of the study population.

    In this study cohort, mean±SD of CRAE was 151.31±16.1 µm, CRVE was 226.59±24.12 µm and AVR was 0.67±0.07. Table 2 shows adjusted associations. Smaller CRAE and CRVE were related to older age (p for trend ≤0.005). Higher levels of CRP and current/ex smoking were related to both widened CRAE and CRVE (p for trend ≤0.002). Higher levels of FPG and HbA1c and dyslipidaemia (higher triglycerides and lower levels of HDL cholesterol) and presence of high albuminuria were related to larger CRVE, but not to CRAE. For retinal arteriolar calibre, there is a gender-difference with a strong inverse correlation with DBP, SBP and hypertension, but not for venular calibre. Neither CRAE nor CRVE was related to duration of diabetes, BMI, total cholesterol and LDL.

    After adjusting for possible risk factors (table 3), including age, sex, duration of diabetes, HbA1c, blood pressure, cholesterol, triacylglycerol, CRVE was wider in the group with any DR compared with NDR (p<0.001 for trend). The risk of presence of retinopathy was approximately two times greater for the participants with CRVE in the fourth quartile compared with the first quartile (95% CI 1.36 to 2.95). After adjusting for the same covariates, the relationship was not statistically significant for CRAE in the DR group compared with the NDR group.

    Table 3

    Associations of retinal arteriolar and venular calibres with diabetic retinopathy (DR)

    Table 4 shows the relationship of vessels calibre with different severity levels of DR, adjusted for covariates in different models. As laser treatment for proliferative DR may affect the calibre of retinal vessels, analysis was also performed with laser-treated eyes excluded from the severe DR group (data not shown), which did not affect the statistical results.

    Table 4

    Associations of quantitative retinal calibre signs with the severity of DR

    The mean of CRAE and CRVE (figure 1) varied by the severity level of retinopathy. Increasing severity of DR was associated with wider CRVE (p for trend<0.001), but not CRAE (p for trend=0.87) in the univariate analysis. A smaller retinal vascular calibre was observed in those subjects with severe retinopathy, which differed slightly but not significantly between those who did or did not receive laser treatment.

    Figure 1

    Mean CRVE (A) and CRAE (B) by diabetic retinopathy grade. Level 10, no retinopathy; Level 20 microaneurysms only; Level 35 mild NPDR; Level 43 moderate NPDR; Level 47 moderate-severe NPDR; Level 53+ severe NPDR and untreated early PDR; Level 60, PDR treated by panretinal laser photocoagulation. CRAE, central retinal arteriolar equivalent; CRVE, central retinal venular equivalent; DR, diabetic retinopathy; NPDR, nonproliferative DR; PDR, proliferative DR.


    The present study is a detailed analysis of the relationship between quantitatively measured retinal vessel diameter and the classification of DR associated with T2DM in a cohort of predominantly Han Chinese. Wider venular calibre, but not arteriolar calibre, was associated with development of DR and more severe retinopathy, independent of other risk factors. Larger venular calibre was additionally associated with higher glucose levels and dyslipidaemia (higher level of triglycerides and lower level of HDL cholesterol). Retinal arteriolar calibre was strongly related to blood pressure but not DR.

    The association of retinal venule widening with the presence of DR in our study confirms other cross-sectional results.6–12 The lack of association between CRAE and DR also has been reported in previous studies,6–8 In contrast to CRVE, CRAE has been inconsistently reported to be associated with DR in the previous studies. In accordance with what has been shown in other studies, a strong and consistent association between retinal arteriolar narrowing and hypertension has been found in the present study.21 ,22

    Our data also showed that wider venular calibre, but not arteriolar calibre, was related to glucose levels and dyslipidaemia (higher triglycerides and lower levels of HDL cholesterol), consistent with observations from the Multi-Ethnic Study of Atherosclerosis study23 and the Singapore Malay Eye Study.24 ,25 Nguyen has suggested that the retinal arteriolar and venular calibre changes may reflect different pathophysiological processes.26

    The biological mechanisms that underlie the association of wider retinal venular calibre with hyperglycaemia and DR are unclear. Retinal venular widening may indicate increased retinal blood flow associated with hyperglycaemia and retinal hypoxia and resultant lactate accumulation or autoregulatory endothelial dysfunction.27 ,28 Second, it may reflect systemic inflammatory processes that are implicated in the pathogenesis of impaired glucose metabolism and DR.29–31

    Clinical experience, in conjunction with epidemiological studies,7 ,17 provides robust and consistent evidence that narrower retinal vessels are correlated with laser treatment. A trend was noted for generalised narrowing of retinal arterioles and venules in eyes that had been treated with laser (figure 1), as was reported by Wisconsin Epidemiologic Study of Diabetic Retinopathy (WESDR).7 In our study, there was also a slightly decreased width of CRVE and CRAE in participants with proliferative diabetic retinopathy (PDR). The decreased width was slightly more prominent in persons with laser treatment. This phenomenon may reflect a reduction in blood flow secondary to improved oxygenation of the inner retina after effective laser treatment, which is known to cause vasoconstriction of retinal blood vessels due to retinal blood flow autoregulation. The cause of decreased vascular width in eyes with PDR but no laser treatment compared with severe nonproliferative diabetic retinopathy (NPDR) is not known, but we speculate that retinas with severe ischaemia and consequent PDR may have slightly smaller vessels due to decreased overall perfusion and volumetric blood flow.

    The relative strengths of our study are that the participants with T2DM were recruited from a relatively homogeneous population living in one stable community, validated methods were applied in the measurement of retinal vascular calibre, standardised protocols to document potential confounding variables and the use of standardised methods to photograph and grade retinopathy. Limitations include that, being a cross-sectional survey, the findings only support an association between vessel calibre and DR but do not prove a causal relationship. The uniformity of the study population may limit the generalisability of the results to all populations. The study participants were not randomly selected and selection bias may be introduced.

    In summary, findings from this study confirm the association of venular dilatation and risk of DR presence and severity in a cohort of Han Chinese patients with T2DM in adjusted analyses. No association of DR with retinal arteriolar calibre was observed. This information further supports the concept that measurement of retinal vessel calibre, particularly venular calibre, has potentially significant clinical value.


    The authors thank James L Reimers, Sapna Ganqaputra, Ashwini Narkar and Jill C Kubiak, staff at the Fundus Photograph Reading Center, Department of Ophthalmology and Visual Sciences, University of Wisconsin, for their important contributions on the diabetic retinopathy grading and vessel measurement training and quality control.



    • Collaborators James L Reimers Sapna Gangaputra Ashwini Narkar Jill C Kubiak.

    • Contributors All the authors are justifiably credited with authorship, according to the authorship criteria. In detail, XY: data acquisition, analysis and interpretation of data, drafting of manuscript, final approval given. TS, XL, NL, NW, RPD: conception, design, data acquisition, analysis and interpretation of the data, critical revision of manuscript, final approval given. YD, HG, JWP: data acquisition, analysis and interpretation of data, critical revision of manuscript, final approval given.

    • Funding Supported by the Beijing Natural Science Foundation Grant 7131007.

    • Competing interests None declared.

    • Ethics approval Capital Medical University.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement Data from this project will be shared upon request to NL.

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Linked Articles

    • At a glance
      Keith Barton James Chodosh Jost Jonas