June 2011
Volume 52, Issue 7
Free
Clinical and Epidemiologic Research  |   June 2011
Are Obesity and Anthropometry Risk Factors for Diabetic Retinopathy?: The Diabetes Management Project
Author Affiliations & Notes
  • Mohamed Dirani
    From the Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, Australia; and
  • Jing Xie
    From the Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, Australia; and
  • Eva Fenwick
    From the Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, Australia; and
  • Rehab Benarous
    From the Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, Australia; and
  • Gwyneth Rees
    From the Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, Australia; and
  • Tien Yin Wong
    From the Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, Australia; and
    the Singapore Eye Research Institute, National University of Singapore, Singapore.
  • Ecosse L. Lamoureux
    From the Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, Melbourne, Australia; and
    the Singapore Eye Research Institute, National University of Singapore, Singapore.
  • Corresponding author: Ecosse L. Lamoureux, Centre for Eye Research Australia, 32 Gisborne Street, East Melbourne 3002, Victoria, Australia; ecosse@unimelb.edu.au
Investigative Ophthalmology & Visual Science June 2011, Vol.52, 4416-4421. doi:10.1167/iovs.11-7208
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Mohamed Dirani, Jing Xie, Eva Fenwick, Rehab Benarous, Gwyneth Rees, Tien Yin Wong, Ecosse L. Lamoureux; Are Obesity and Anthropometry Risk Factors for Diabetic Retinopathy?: The Diabetes Management Project. Invest. Ophthalmol. Vis. Sci. 2011;52(7):4416-4421. doi: 10.1167/iovs.11-7208.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose.: To investigate the relationship between anthropometric parameters and diabetic retinopathy (DR) in adults with diabetes.

Methods.: Five hundred participants with diabetes were recruited prospectively from ophthalmology clinics in Melbourne, Australia. Each underwent an eye examination, anthropometric measurements, and standardized interview-administered questionnaires, and fasting blood glucose and serum lipids were analyzed. Two-field fundus photographs were taken and graded for DR. Height; weight; body mass index (BMI); waist, hip, neck, and head circumferences; and skinfold measurements were recorded.

Results.: A total of 492 patients (325 men, 66.1%) aged between 26 and 90 years (median, 65) were included in the analysis: 171 (34.8%), 187 (38.0%), and 134 (27.2%) with no DR, nonproliferative DR (NPDR), and proliferative DR (PDR), respectively. After multiple adjustments, higher BMI (odds ratio [OR], 1.06; 95% confidence interval [CI],1.01–1.11; P = 0.02) was significantly associated with any DR. Obese people were 6.5 times more likely to have PDR than were those with normal weight (OR, 6.52; 95% CI, 1.49–28.6; P = 0.013). Neck circumference (OR, 1.05; 95% CI, 1.00–1.10; P = 0.03) and waist circumference (OR, 1.12; 95% CI, 1.03–1.22; P = 0.01) were significantly associated with any DR. BMI (OR, 1.04; 95% CI, 1.00–1.08; P = 0.04) and neck circumference (OR, 1.04 95% CI, 1.01–1.08; P = 0.04) were also positively associated with increasing severity levels of DR.

Conclusions.: Persons with diabetes with higher BMI and larger neck circumference are more likely to have DR and more severe stages of DR. These data suggest that obesity is an independent risk factor for DR.

Diabetic retinopathy (DR) is a major microvascular complication of diabetes and a leading cause of vision loss and blindness worldwide. The public health implications associated with DR are expected to magnify, with estimates showing that half of those with diabetes are still undiagnosed. 1 Studies show that the prevalence of DR in individuals with long-standing diabetes (20 years or greater) exceeds 50%. 2  
In the past three decades, population-based studies and clinical trials have demonstrated that longer duration of diabetes and poor glycemic and blood pressure (BP) control are the key risk factors for the development and progression of DR. However, evidence from new trials such as the Action in Diabetes and Vascular Disease (ADVANCE) 3 and the Action to Control Cardiovascular Risk in Diabetes (ACCORD-Eye) 4 has shown a limit to the risk reduction for DR that can be achieved with better glucose and BP management alone, respectively. For this reason, it is important to better understand the role of other modifiable risk factors. 
The evidence supporting a relationship between high body mass index (BMI) and increased risk of DR is inconclusive. 5 9 For instance, the Wisconsin Epidemiologic Study of DR (WESDR) reported a nonsignificant association between obesity and progression and severity of DR. 7 In contrast, the Singapore Malay Eye Study (SiMES) found that those with a high BMI (≥25 kg/m2) were significantly less likely to have DR or clinically significant macular edema (CSME). 10 The lack of consensus may be explained by methodological differences, differences in study participants, lack of comprehensive anthropometric measurements, and inadequate clinical sample size. Moreover, it is known that the anthropometric measurements of BMI, waist–hip ratio (WHR), and waist circumference (WC) do not correlate entirely and are indices of different aspects of obesity. For instance, adults with low BMI may have markedly high anthropometric parameters (e.g., WC), sometimes referred to as metabolically obese normal-weight adults. 11 This phenomenon may also in part explain the variation in the correlation between BMI and DR. For these reasons, it is essential that we consider factors other than BMI when investigating the role of obesity in the development and severity of DR. Finally, to our knowledge no studies have investigated other markers of obesity, such as skinfold thickness, for the presence and severity of DR. 
In this study, we assessed the relationship between obesity (including BMI), height, weight, waist and hip circumference, and skin fold thickness and the presence of DR and PDR in adults with type I or type II diabetes. 
Methods
Study Population
The Diabetes Management Project (DMP) is a clinical study of adults with diabetes residing in Victoria, Australia. English-speaking adults aged 18 years or older with type I or type II diabetes, free of significant hearing and cognitive impairment, and living independently in the community were invited. The participants were recruited from the Royal Victorian Eye and Ear Hospital (RVEEH). 
All study procedures adhered to the tenets of the Declaration of Helsinki, and all privacy requirements were met. Ethical approval for the DMP was provided by the RVEEH Human Research and Ethics Committee (08/815H). Each individual signed a consent form that outlined the purposes and methodology of the DMP, including possible adverse outcomes, confidentiality policies, and data storage procedures. 
DMP Testing Protocol
All examinations were conducted at the Centre for Eye Research Australia (CERA), Melbourne, located at the RVEEH, with a mean test duration of 3 hours per individual. 
Anthropometric Measurements
All individuals had height and weight measured with a wall-mounted, adjustable measuring scale and a calibrated digital scientific weight scale (PRC; Oregon Scientific, Tualatin, OR), respectively. Individuals were instructed to remove any footwear and heavy clothing before testing. BMI was calculated as weight (in kilograms) divided by the height in meters squared (kg/m2). Obesity was defined as BMI ≥ 30 kg/m2. Waist and hip circumferences (in centimeters) were measured with a 151-cm medical tape measure (Birch, Melbourne, VIC, Australia) and waist–hip ratio (WHR) values were then computed. The average of three skin-fold measurements (in millimeters) was taken at the right triceps with a Harpenden skinfold caliper (Baty International, Burgess Hill, UK). 
Blood Pressure Measurements
Each individual underwent a BP assessment with an automated BP machine (model 5200-103Z; Welch Allyn, Rydalmere, NSW, Australia). The average of two separate measurements was recorded for both systolic (SBP) and diastolic (DBP) BP. In cases in which there was a difference greater than 10 mm Hg (millimeters of mercury) for SBP and 5 mm Hg for DBP, a third measurement was taken. The average of the two closest BP measurements was then used for the analysis. Hypertension was defined as a SBP > 140 mm Hg or DBP > 90 mm Hg. 
Fundus Photography
Two-field, 45°, digital, nonstereo color fundus photographs of both eyes were taken from each individual with a nonmydriatic retinal camera (CR6-45NM; Canon Inc, Tokyo, Japan), with all images being stored electronically (Digital Healthcare software; Cambridge, UK). Diabetic retinopathy (DR) grading adhered to the modified two-standard-field color fundus photography procedure (Retinal Vascular Imaging Centre Grading Protocol 01; Assessment of Diabetic Retinopathy); based on the ETDRS and MESA-EYE Digital Grading Protocols. DR was classified into two groups, (1) nonproliferative DR (NPDR) and (2) proliferative DR (PDR), according to the definitions derived from Wilkinson et al. 12 NPDR was further grouped into three progressive stages—mild, moderate and severe—with each stage having defined retinal pathologic signs. Dilation was achieved with a single drop of tropicamide 1% (MINIMS; Chauvin Pharmaceuticals, Ltd, Kingston-on-Thames, UK) in each eye approximately 25 minutes before fundus photography. 
Questionnaires
Data associated with lifestyle and psychosocial risk factors were collected with validated questionnaires. We report on data collected only from the general questionnaire on medical and ocular history, diabetes status and duration, demographics, and medication use. 
Blood Collection
Fasting blood (≥8 hours) was obtained to assess HbA1c levels, fasting glucose and lipids (total cholesterol [C], triglyceride, LDL-C, and HDL-C). All biochemical parameters were analyzed at Melbourne Pathology. Fasting plasma glucose and serum lipids were assessed (Hitachi Modular P analyzer; Roche Diagnostics, Mannheim, Germany). Non-HDL-C was defined by subtracting the HDL-C from the total-C. 
Statistical Analyses
Descriptive statistics were computed for all variables. Normality of the variables was examined using boxplots, Kolmogorov-Smirnov, and Shapiro-Wilks tests. The linearity of nominal and ordinal data were assessed using χ2-based measures. BMI was defined into four categories—(1) normal (18.5–24.9 kg/m2); (2) overweight (25–29.9 kg/m2); and (3) obese (>30 kg/m2)—and was also analyzed as a continuous variable. 
The relationship between DR and body stature measurements was examined using a multivariable logistic regression model adjusting for potential confounders. For each body stature measurement (BMI, WHR, WC, head circumference, neck circumference, and skinfold thickness), we constructed two models: Model 1 adjusted for age, sex, income, education, smoking, and general health status. Model 2 adjusted for all variables in Model 1 plus duration of diabetes, SBP, insulin use, cholesterol, HDL-cholesterol, fasting glucose level and HbA1c. All pertinent variables were examined for correlations and multicollinearity, using the Pearson product moment correlation. Because multicollinearity was present among the examined predictor variables, factor analysis regression (FAR) was used to estimate the regression coefficients of the factor-predicted score. We also examined the association between body stature measurements and severity of DR (no DR, NPDR, and PDR) using an ordinal regression model (ORM). The proportional odds assumption of ordinal regression was checked by an approximate-likelihood ratio test. A two-tailed P < 0.05 was considered statistically significant (all analyses with Stata Corp, College Station, TX). 
Results
A total of 492 patients (325 men, 66.1%) aged between 26 and 90 years (median, 65) were included in the analysis: 171 (34.8%), 187 (38.0%), and 134 (27.2%) patients had no DR, nonproliferative DR (NPDR), and proliferative DR (PDR), respectively. Three hundred twenty-one of the 492 (65.2%) had any DR. Missing data (fundus images) were encountered in 8 cases (1.6%). Of the total number of participants (n = 492), self-reported diabetes consisted of 62 (12.6%) type I diabetes, 391 (79.5%) type II diabetes, and 39 (7.9%) unsure of diabetes type. Compared with those without DR, those with DR were significantly younger and had longer mean duration of diabetes (P < 0.01, Table 1). 
Table 1.
 
Participants' Characteristics, Comparing Those with and Those without DR
Table 1.
 
Participants' Characteristics, Comparing Those with and Those without DR
Without DR (n = 171) With DR (n = 321) P
n % n %
Sex, male 96 56.1 229 71.4 0.001
Income
    <$30,000 111 72.1 198 66.7 0.241
    ≥$30,000 43 27.9 99 33.3
Education
    Primary school or less 24 14.4 43 13.7 0.466
    Secondary school 90 53.9 186 59.4
    14 years or above 53 31.7 84 26.8
Current/past smoker 89 53.0 170 53.8 0.255
Diabetes type*
    I 19 11.8 43 14.7 0.386
    II 142 88.2 249 85.3
Insulin use 38 22.2 172 53.6 <0.001
Number of comorbidities†
    0 17 9.9 53 16.5 0.219
    1 48 28.1 87 27.1
    2 44 25.7 82 25.6
    ≥3 62 36.3 99 30.8
Median or Mean IQR or SD Median or Mean IQR or SD P
Age, y‡ 68.0 17.0 63.0 15.0 <0.001
Current/past smoker, pack/year‡ 25.5 27.5 15.0 25.4 0.003
Systolic blood pressure, mm Hg 138.6 18.6 141.1 18.9 0.175
Diastolic blood pressure, mm Hg 90.8 26.8 93.7 32.7 0.336
Duration of diabetes, y‡ 8.0 9.0 18.0 12.0 <0.001
Weight, kg‡ 80.9 22.2 86.2 27.8 0.044
Height, m‡ 1.66 0.10 1.67 0.10 0.214
Body mass index, kg/m2 30.5 6.6 30.9 6.0 0.216
Waist circumference, cm 104.7 17.5 107.7 15.3 0.051
Hip circumference, cm 106.1 20.8 108.1 18.9 0.287
Waist to hip ratio, cm 0.96 0.09 0.98 0.08 0.127
Head circumference, cm 56.0 3.6 56.5 3.5 0.002
Neck circumference, cm 39.3 6.9 41.4 6.6 0.001
Skin fold, mm‡ 35.5 15.4 33.4 15.4 0.127
Fasting plasma glucose, mg/dL‡ 7.0 2.4 8.3 4.2 <0.001
Haemoglobin Alc, %‡ 6.9 1.3 7.8 2.0 <0.001
Total cholesterol, mg/dL‡ 4.7 1.5 4.4 1.5 0.040
HDL cholesterol, mg/dL‡ 1.4 0.5 1.3 0.5 0.004
Triglycerides, mg/dL‡ 1.5 1.0 1.7 1.3 0.434
LDL‡ 2.3 1.2 2.2 1.2 0.160
In multivariate models, BMI (odds ratio [OR], 1.06; 95% confidence interval [CI], 1.01–1.11; P = 0.021), WC (OR, 1.59; 95% CI, 1.13–2.26; P = 0.008), and neck circumference (OR, 1.05; 95% CI, 1.01–1.10; P = 0.008) were significantly associated with having any DR (model 1, Table 2). After further adjustment for biochemical parameters (model 2), the obese participants were more than three times as likely to have any DR (OR, 3.12; 95% CI, 1.20–8.16; P = 0.02) and 6.5 times as likely to have PDR (OR, 6.52; 95% CI, 1.49–28.6; P = 0.013) compared with those with a normal BMI (reference). Greater WC (OR, 1.09; 95% CI, 1.01–1.21; P = 0.047) and neck circumference (OR, 1.04; 95% CI, 1.00–1.09; P = 0.045) were also still significantly associated with any DR in model 2. 
Table 2.
 
Logistic Regression Model of the Association between Body Stature Measurements and Presence of DR
Table 2.
 
Logistic Regression Model of the Association between Body Stature Measurements and Presence of DR
Model 1: Adjusted OR (n = 416) Model 2: Adjusted OR (n = 365) Model 2 vs. Model 1* χ2, P
OR (95% CI) P OR (95% CI) P
BMI, kg/m2 1.06 (1.01–1.11) 0.021 1.05 (1.00–1.12) 0.061 LR χ2 7 = 100.14, P < 0.001
    Normal weight, 18.5–24.9 kg/m2 1.0 1.0 LR χ2 7 = 99.30.14, P < 0.001
    Overweight, 25–29.9 kg/m2 2.15 (1.02–4.51) 0.044 2.26 (0.84–6.10) 0.105
    Obesity, >30 kg/m2 2.98 (1.46–6.06) 0.003 3.12 (1.20–8.16) 0.020
Waist circumference, cm (per 5 cm change) 1.12 (1.03–1.22) 0.008 1.09 (1.01–1.21) 0.047 LR χ2 7 = 98.02.14, P < 0.001
Neck circumference, cm 1.05 (1.01–1.10) 0.008 1.04 (1.00–1.09) 0.045 LR χ2 7 = 97.57, P < 0.001
In our ordinal regression model (Table 3), DR categories consisted of no DR, NPDR, and PDR. Using this model, we found that for every 1-kg/m2 increase in BMI (OR, 1.04; 95% CI, 1.00–1.08; P = 0.06), there was a 5-cm increase in WC (OR, 1.07; 95% CI, 1.00–1.15; P = 0.064), but these did not reach statistical significance (model 2). However, for every 1-cm increase in neck circumference (OR, 1.05; 95% CI, 1.00–1.09; P = 0.045), there was a 5% increased risk of severe DR (model 2). 
Table 3.
 
ORM of the Association between Body Stature Measurements and Severity of DR
Table 3.
 
ORM of the Association between Body Stature Measurements and Severity of DR
Model 1: Adjusted OR (n = 416) Model 2: Adjusted OR (n = 365) Model 2 vs. Model 1* χ2, P
OR (95% CI) P OR (95% CI) P
BMI, kg/m2 1.04 (1.01–1.08) 0.037 1.04 (1.00–1.08) 0.068 LR χ2 7 = 94.91 P < 0.001
Waist circumference, cm, (per 5-cm increase) 1.09 (1.02–1.17) 0.015 1.07 (1.00–1.15) 0.064 LR χ2 7 = 92.81.14, P < 0.001
Neck circumference, cm 1.04 (1.01–1.08) 0.038 1.05 (1.00–1.09) 0.045 LR χ2 7 = 89.60, P < 0.001
Using the same statistical models, we stratified the analysis by sex and found that neck circumference, waist circumference, and BMI were significantly associated with DR only in the women (neck circumference: OR, 1.11; 95% CI, 1.01–1.20; P = 0.030; waist circumference: OR, 1.24; 95% CI, 1.07–1.46; P = 0.005; BMI: OR, 1.08; 95% CI, 1.00–1.16; P = 0.029) (model 1; Table 4). In the women, BMI and WC were significantly associated with having DR in each model, with ORs of 1.16 (95% CI, 1.02–1.30; P = 0.019) and 1.28 (95% CI, 1.03–1.59; P = 0.023), respectively, for model 2. 
Table 4.
 
Logistic Regression Model of the Association between Body Stature Measurements and DR by Sex
Table 4.
 
Logistic Regression Model of the Association between Body Stature Measurements and DR by Sex
Model 1: Adjusted OR Model 2: Adjusted OR Model 2 vs. Model 1 χ2, P
OR (95% CI) P OR (95% CI) P
Male
    BMI, kg/m2 1.02 (0.96–1.09) 0.464 1.02 (0.94–1.10) 0.665 LR χ2 7 = 54.86, P < 0.001
    Waist circumference, cm (per 5-cm increase) 1.03 (0.92–1.16) 0.576 1.00 (0.88–1.14) 0.994 LR χ2 7 = 55.42, P < 0.001
    Neck circumference (cm) 1.04 (1.00–1.08) 0.076 1.04 (0.98–1.10) 0.192 LR χ2 7 = 52.72, P < 0.001
Female
    BMI, kg/m2 1.08 (1.00–1.16) 0.029 1.16 (1.02–1.30) 0.019 LR χ2 7 = 56.66, P < 0.001
    Waist circumference, cm (per 5-cm increase) 1.24 (1.07–1.46) 0.005 1.28 (1.03–1.59) 0.023 LR χ2 7 = 51.73, P < 0.001
    Neck circumference, cm 1.11 (1.01–1.20) 0.030 1.09 (0.99–1.19) 0.097 LR χ2 7 = 51.99, P < 0.001
Considering the multicollinearity between variables, we used factor analysis to generate a predicted score that incorporated the information from BMI, WC, and neck circumference. Factor 1 (Eigenvalue >1) explained 74% of the total variance in BMI, WC, and neck circumference. The predicted score of factor 1 was significantly associated with DR for both logistic regression models. In model 2, we showed that the predicted score was significantly associated with DR (OR, 1.49; 95% CI, 1.01–2.21; P = 0.04), after adjustment for all demographic, medical, and biochemical parameters. 
Discussion
In this study, we investigated the relationship between anthropometric parameters and the presence and severity of DR in a clinical sample of adults with diabetes. Our main study findings show that BMI and neck circumference are independently associated with the presence and severity of DR. Obesity was associated with a threefold increased risk of having DR in our diabetic sample population. Although WC was also associated with the presence of DR, the association was nonsignificant for the severity of DR after additional adjustments for biochemical estimates. 
Several epidemiologic studies worldwide have explored the relationship of BMI and anthropometric parameters with DR but, to date, the evidence has been equivocal. Some studies have reported a significant association between low BMI and DR, 7 9,13 16 suggesting a protective role for higher BMI in the development of DR. For example, in an earlier publication by the WESDR group 8 with a sample of 1370 patients aged 30 or older, no significant overall association between higher BMI and DR was reported; instead, small body mass was associated with an increased risk of severity of DR. Almost two decades later, the same group reported a positive association between higher BMI and the progression and severity of DR in those with late-onset, insulin-dependent diabetes. 7  
Our study findings concur with those in several studies that have reported an increased risk of the presence of DR in those with high BMI. 5,17 20 For instance, the Diabetes Control and Complications Trial (DCCT) reported that risk factors such as high BMI were significantly associated with the development of DR in type I diabetics with good metabolic control (HbA1c ≤ 6.87%). 19 Similarly, the Hoorn 20 study examined 2484 adults aged 50 to 74 years and reported that higher prevalence of DR was significantly associated with higher BMI. Again, however, that study did not include other obesity indices such as WHR and WC. Other studies also support our findings of a significant association between BMI and severity of DR, 21 although, because these studies were population based and only included a single measure of obesity, the robustness of their results is restricted. 
In contrast to our study findings, two recent studies 9,10 have reported negative correlations between BMI and DR. In the Singapore Study (SiMES), multivariate analysis showed that those with higher BMI were less likely to have DR or vision-threatening DR. 10 Although the Singapore study undertook multiple adjustments in their multivariate analyses, they were limited in the number of cases of vision-threatening DR (n = 80) due to their population-based study design. Moreover, only random blood glucose was measured in those in whom diabetes had not been diagnosed, which may have led to misclassification. In the SN-DREAMS-I study, combined obesity (BMI ≥ 23 kg/m2 and WC ≥ 90 cm in males and ≥ 80 cm in females) had a protective role in the development of any DR; however, a significant association between abdominal obesity and higher WHR and DR was shown in the females. This population-based study was undertaken in urban India, where only 30% of India's population reside, and the association between anthropometric measurements and DR severity was not assessed. Despite these methodological limitations, the SN-DREAMS-I study clearly demonstrated the importance of accounting for several obesity indices, rather than relying on a single obesity measure, such as BMI. 
Although the pathophysiological mechanisms underpinning the association between higher BMI and DR are yet to be defined, several biological theories have been presented. 22,23 These include the potential involvement of platelet function, blood viscosity, aldose reductase activity, and vasoproliferative parameters, such as vascular endothelial growth factor (VEGF). 23 Apart from these parameters, lifestyle factors, such as physical activity and weight loss, provide some evidence to support the relationship between high BMI and DR. In brief, weight loss has been suggested to delay the onset of diabetic complications, including DR. However, there has also been evidence to show that weight loss in individuals with anorexia nervosa has significantly increased the risk of developing early DR, although this finding is restricted to type I diabetes. 24 Overall, the underlying mechanisms of the association between obesity and DR remain largely unknown and should be further investigated. 
Unlike previous studies, we included other anthropometric measurements, including neck and head circumference. Neck circumference was independently associated with the presence and severity of DR. It may be argued that neck circumference merely reflects BMI or WC; however, correlations between neck circumference and BMI and WC were only 0.40 and 0.60, respectively. Although the underlying mechanisms of the association between neck circumference and DR, other than its relationship with BMI, are unknown, it should be included in future research assessing anthropometric parameters with the development of DR. 
The strengths of our study design include a large clinical sample of diabetics with differing levels of DR, the use of dilated fundus photography, and the inclusion of other anthropometric measurements. With the extensive testing protocol of the DMP, we were able to account for a wide range of covariates in multivariate logistic models. Potential limitations include the higher proportion of males than females in our sample and potential selection biases stemming from our focused recruitment from specialized retinal clinics and the small number of patients with type I diabetes. However, the clinical nature of our study, the extensive testing protocol, and the logistic and financial constraints did not allow us to adequately resolve these issues. 
In conclusion, DR is a complex disease with several known and proposed risk factors. We have shown that BMI and other important anthropometric measurements, such as waist and neck circumferences, are independently associated with the presence and severity of DR in a clinical sample of diabetic patients. Our findings illustrate the importance of not restricting the assessment of obesity to only one measure, such as BMI, particularly with the discrepancies in study findings among different sample populations. Our findings also have potential clinical implications in the management of diabetes and DR, particularly with our novel findings showing a positive association between neck circumference and DR, and our high factor score incorporating neck and waist circumference and BMI. 
Footnotes
 Supported by an Australian Research Council Linkage grant. CERA receives Operational Infrastructure Support from the Victorian Government.
Footnotes
 Disclosure: M. Dirani, None; J. Xie, None; E. Fenwick, None; R. Benarous, None; G. Rees, None; T.Y. Wong, None; E.L. Lamoureux, None
The authors thank their research collaborators, the Royal Victorian Eye and Ear Hospital (RVEEH) and their industry partner, Diabetes Australia, Victoria; co-investigators Jill Keeffe, Jie Jin Wang, Ralph Audehm, and Gabriella Tikellis for significant contributions to the design and implementation of the DMP; and all who agreed to participate in the DMP. 
References
Dunstan DW Zimmet PZ Welborn TA . The rising prevalence of diabetes and impaired glucose tolerance: the Australian Diabetes, Obesity Lifestyle Study Diabetes Care. 2002;25(5):829–834.
Tapp RJ Shaw JE Harper CA . The prevalence of and factors associated with diabetic retinopathy in the Australian population. Diabetes Care. 2003;26(6):1731–1737. [CrossRef] [PubMed]
Beulens JW Patel A Vingerling JR . Effects of blood pressure lowering and intensive glucose control on the incidence and progression of retinopathy in patients with type 2 diabetes mellitus: a randomised controlled trial. Diabetologia. 2009;52:2027–2036. [CrossRef] [PubMed]
Ismail-Beigi F Craven T Banerji MA . Effect of intensive treatment of hyperglycaemia on microvascular outcomes in type 2 diabetes: an analysis of the ACCORD randomised trial. Lancet. 2010;376:419–430. [CrossRef] [PubMed]
Henricsson M Nystrom L Blohme G . The incidence of retinopathy 10 years after diagnosis in young adult people with diabetes: results from the nationwide population-based Diabetes Incidence Study in Sweden (DISS). Diabetes Care. 2003;26(2):349–354. [CrossRef] [PubMed]
The UK Prospective Diabetes Study (UKPDS) Group. Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet 1998;352(9131):854–865. [CrossRef] [PubMed]
Klein R Klein BE Moss SE . Is obesity related to microvascular and macrovascular complications in diabetes? The Wisconsin Epidemiologic Study of Diabetic Retinopathy. Arch Intern Med. 1997;157:650–656. [CrossRef] [PubMed]
Klein R Klein BE Moss SE . 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. 1984;102:527–532. [CrossRef] [PubMed]
Raman R Rani PK Gnanamoorthy P . Association of obesity with diabetic retinopathy: Sankara Nethralaya Diabetic Retinopathy Epidemiology and Molecular Genetics Study (SN-DREAMS Report no. 8). Acta Diabetol. 2010;47:209–215. [CrossRef] [PubMed]
Lim LS Tai ES Mitchell P . C-reactive protein, body mass index, and diabetic retinopathy. Invest Ophthalmol Vis Sci. 2010;51:4458–4463. [CrossRef] [PubMed]
Ruderman N Chisholm D Pi-Sunyer X Schneider S . The metabolically obese, normal-weight individual revisited. Diabetes. 1998;47:699–713. [CrossRef] [PubMed]
Wilkinson CP Ferris FL3rd Klein RE . Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology. 2003;110(9):1677–1682. [CrossRef] [PubMed]
Dowse GK Humphrey AR Collins VR . Prevalence and risk factors for diabetic retinopathy in the multiethnic population of Mauritius. Am J Epidemiol. 1998;47:448–457. [CrossRef]
Rema M Premkumar S Anitha B . Prevalence of diabetic retinopathy in urban India: the Chennai Urban Rural Epidemiology Study (CURES) eye study, Invest Ophthalmol Vis Sci. 2005;46:2328–2333. [CrossRef] [PubMed]
West KM Erdreich LJ Stober JA . A detailed study of risk factors for retinopathy and nephropathy in diabetes. 1980;29:501–508.
Nilsson SV Nilsson JE Frostberg N Emilsson T . The Kristianstad survey. II. Studies in a representative adult diabetic population with special reference to comparison with an adequate control group. Acta Med Scand Suppl. 1967;469:1–42. [PubMed]
Zhang L Krzentowski G Albert A Lefebvre PJ . Risk of developing retinopathy in Diabetes Control and Complications Trial type 1 diabetic patients with good or poor metabolic control. Diabetes Care. 2001;24(7):1275–1279. [CrossRef] [PubMed]
Chaturvedi N Sjoelie AK Porta M . Markers of insulin resistance are strong risk factors for retinopathy incidence in type 1 diabetes. Diabetes Care. 2001;24(2):284–289. [CrossRef] [PubMed]
The UK Prospective Diabetes Study Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998;352(9131):837–853. [CrossRef] [PubMed]
van Leiden HA Dekker JM Moll AC . Blood pressure, lipids, and obesity are associated with retinopathy: the Hoorn study. Diabetes Care. 2002;25:1320–1325. [CrossRef] [PubMed]
Ozmen B Boyvada S . The relationship between self-monitoring of blood glucose control and glycosylated haemoglobin in patients with type 2 diabetes with and without diabetic retinopathy. J Diabetes Complications. 2003;17:128–134. [CrossRef] [PubMed]
Dorchy H Claes C Verougstraete C . Risk factors of developing proliferative retinopathy in type 1 diabetic patients: role of BMI. Diabetes Care. 2002;25:798–799. [CrossRef] [PubMed]
Silha JV Krsek M Sucharda P Murphy LJ . Angiogenic factors are elevated in overweight and obese individuals. Int J Obes (Lond). 2005;29:1308–1314. [CrossRef] [PubMed]
Tamburrino MB McGinnis RA . Anorexia nervosa: a review. Panminerva Med. 2002;44:301–311. [PubMed]
Table 1.
 
Participants' Characteristics, Comparing Those with and Those without DR
Table 1.
 
Participants' Characteristics, Comparing Those with and Those without DR
Without DR (n = 171) With DR (n = 321) P
n % n %
Sex, male 96 56.1 229 71.4 0.001
Income
    <$30,000 111 72.1 198 66.7 0.241
    ≥$30,000 43 27.9 99 33.3
Education
    Primary school or less 24 14.4 43 13.7 0.466
    Secondary school 90 53.9 186 59.4
    14 years or above 53 31.7 84 26.8
Current/past smoker 89 53.0 170 53.8 0.255
Diabetes type*
    I 19 11.8 43 14.7 0.386
    II 142 88.2 249 85.3
Insulin use 38 22.2 172 53.6 <0.001
Number of comorbidities†
    0 17 9.9 53 16.5 0.219
    1 48 28.1 87 27.1
    2 44 25.7 82 25.6
    ≥3 62 36.3 99 30.8
Median or Mean IQR or SD Median or Mean IQR or SD P
Age, y‡ 68.0 17.0 63.0 15.0 <0.001
Current/past smoker, pack/year‡ 25.5 27.5 15.0 25.4 0.003
Systolic blood pressure, mm Hg 138.6 18.6 141.1 18.9 0.175
Diastolic blood pressure, mm Hg 90.8 26.8 93.7 32.7 0.336
Duration of diabetes, y‡ 8.0 9.0 18.0 12.0 <0.001
Weight, kg‡ 80.9 22.2 86.2 27.8 0.044
Height, m‡ 1.66 0.10 1.67 0.10 0.214
Body mass index, kg/m2 30.5 6.6 30.9 6.0 0.216
Waist circumference, cm 104.7 17.5 107.7 15.3 0.051
Hip circumference, cm 106.1 20.8 108.1 18.9 0.287
Waist to hip ratio, cm 0.96 0.09 0.98 0.08 0.127
Head circumference, cm 56.0 3.6 56.5 3.5 0.002
Neck circumference, cm 39.3 6.9 41.4 6.6 0.001
Skin fold, mm‡ 35.5 15.4 33.4 15.4 0.127
Fasting plasma glucose, mg/dL‡ 7.0 2.4 8.3 4.2 <0.001
Haemoglobin Alc, %‡ 6.9 1.3 7.8 2.0 <0.001
Total cholesterol, mg/dL‡ 4.7 1.5 4.4 1.5 0.040
HDL cholesterol, mg/dL‡ 1.4 0.5 1.3 0.5 0.004
Triglycerides, mg/dL‡ 1.5 1.0 1.7 1.3 0.434
LDL‡ 2.3 1.2 2.2 1.2 0.160
Table 2.
 
Logistic Regression Model of the Association between Body Stature Measurements and Presence of DR
Table 2.
 
Logistic Regression Model of the Association between Body Stature Measurements and Presence of DR
Model 1: Adjusted OR (n = 416) Model 2: Adjusted OR (n = 365) Model 2 vs. Model 1* χ2, P
OR (95% CI) P OR (95% CI) P
BMI, kg/m2 1.06 (1.01–1.11) 0.021 1.05 (1.00–1.12) 0.061 LR χ2 7 = 100.14, P < 0.001
    Normal weight, 18.5–24.9 kg/m2 1.0 1.0 LR χ2 7 = 99.30.14, P < 0.001
    Overweight, 25–29.9 kg/m2 2.15 (1.02–4.51) 0.044 2.26 (0.84–6.10) 0.105
    Obesity, >30 kg/m2 2.98 (1.46–6.06) 0.003 3.12 (1.20–8.16) 0.020
Waist circumference, cm (per 5 cm change) 1.12 (1.03–1.22) 0.008 1.09 (1.01–1.21) 0.047 LR χ2 7 = 98.02.14, P < 0.001
Neck circumference, cm 1.05 (1.01–1.10) 0.008 1.04 (1.00–1.09) 0.045 LR χ2 7 = 97.57, P < 0.001
Table 3.
 
ORM of the Association between Body Stature Measurements and Severity of DR
Table 3.
 
ORM of the Association between Body Stature Measurements and Severity of DR
Model 1: Adjusted OR (n = 416) Model 2: Adjusted OR (n = 365) Model 2 vs. Model 1* χ2, P
OR (95% CI) P OR (95% CI) P
BMI, kg/m2 1.04 (1.01–1.08) 0.037 1.04 (1.00–1.08) 0.068 LR χ2 7 = 94.91 P < 0.001
Waist circumference, cm, (per 5-cm increase) 1.09 (1.02–1.17) 0.015 1.07 (1.00–1.15) 0.064 LR χ2 7 = 92.81.14, P < 0.001
Neck circumference, cm 1.04 (1.01–1.08) 0.038 1.05 (1.00–1.09) 0.045 LR χ2 7 = 89.60, P < 0.001
Table 4.
 
Logistic Regression Model of the Association between Body Stature Measurements and DR by Sex
Table 4.
 
Logistic Regression Model of the Association between Body Stature Measurements and DR by Sex
Model 1: Adjusted OR Model 2: Adjusted OR Model 2 vs. Model 1 χ2, P
OR (95% CI) P OR (95% CI) P
Male
    BMI, kg/m2 1.02 (0.96–1.09) 0.464 1.02 (0.94–1.10) 0.665 LR χ2 7 = 54.86, P < 0.001
    Waist circumference, cm (per 5-cm increase) 1.03 (0.92–1.16) 0.576 1.00 (0.88–1.14) 0.994 LR χ2 7 = 55.42, P < 0.001
    Neck circumference (cm) 1.04 (1.00–1.08) 0.076 1.04 (0.98–1.10) 0.192 LR χ2 7 = 52.72, P < 0.001
Female
    BMI, kg/m2 1.08 (1.00–1.16) 0.029 1.16 (1.02–1.30) 0.019 LR χ2 7 = 56.66, P < 0.001
    Waist circumference, cm (per 5-cm increase) 1.24 (1.07–1.46) 0.005 1.28 (1.03–1.59) 0.023 LR χ2 7 = 51.73, P < 0.001
    Neck circumference, cm 1.11 (1.01–1.20) 0.030 1.09 (0.99–1.19) 0.097 LR χ2 7 = 51.99, P < 0.001
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×