[Skip to Content]
[Skip to Content Landing]
Table 1.  
Multivariable Association Between Exposure Variables and Presence of Diabetic Retinopathy Stratified by Sexa
Multivariable Association Between Exposure Variables and Presence of Diabetic Retinopathy Stratified by Sexa
Table 2.  
Clinical and Demographic Characteristics of 420 Participants Stratified by the Presence of DR
Clinical and Demographic Characteristics of 420 Participants Stratified by the Presence of DR
Table 3.  
Associations Between BMI and Waist to Hip Ratio With the Presence of DR
Associations Between BMI and Waist to Hip Ratio With the Presence of DR
Table 4.  
Association Between BMI and Severity of DR
Association Between BMI and Severity of DR
Table 5.  
Multivariable Association Between Waist to Hip Ratio and Severity of DR Stratified by Sexa
Multivariable Association Between Waist to Hip Ratio and Severity of DR Stratified by Sexa
1.
Saaddine  JB, Honeycutt  AA, Narayan  KM, Zhang  X, Klein  R, Boyle  JP.  Projection of diabetic retinopathy and other major eye diseases among people with diabetes mellitus: United States, 2005-2050.  Arch Ophthalmol. 2008;126(12):1740-1747.PubMedGoogle ScholarCrossref
2.
Yau  JW, Rogers  SL, Kawasaki  R,  et al; Meta-Analysis for Eye Disease (META-EYE) Study Group.  Global prevalence and major risk factors of diabetic retinopathy.  Diabetes Care. 2012;35(3):556-564.PubMedGoogle ScholarCrossref
3.
Younis  N, Broadbent  DM, Vora  JP, Harding  SP; Liverpool Diabetic Eye Study.  Incidence of sight-threatening retinopathy in patients with type 2 diabetes in the Liverpool Diabetic Eye Study: a cohort study.  Lancet. 2003;361(9353):195-200.PubMedGoogle ScholarCrossref
4.
Ministry of Health.  National Health Survey 2010.http://www.moh.gov.sg/content/moh_web/home/Publications/Reports/2011/national_health_survey2010.html. January 16, 2014.
5.
Vazquez  G, Duval  S, Jacobs  DR  Jr, Silventoinen  K.  Comparison of body mass index, waist circumference, and waist/hip ratio in predicting incident diabetes: a meta-analysis.  Epidemiol Rev. 2007;29:115-128.PubMedGoogle ScholarCrossref
6.
Nguyen  NT, Nguyen  XM, Lane  J, Wang  P.  Relationship between obesity and diabetes in a US adult population: findings from the National Health and Nutrition Examination Survey, 1999-2006.  Obes Surg. 2011;21(3):351-355.PubMedGoogle ScholarCrossref
7.
Cheung  N, Mitchell  P, Wong  TY.  Diabetic retinopathy.  Lancet. 2010;376(9735):124-136.PubMedGoogle ScholarCrossref
8.
World Health Organization.  Obesity and overweight. http://www.who.int/mediacentre/factsheets/fs311/en/. Accessed February 20, 2015.
9.
van Leiden  HA, Dekker  JM, Moll  AC,  et al.  Blood pressure, lipids, and obesity are associated with retinopathy: the Hoorn Study.  Diabetes Care. 2002;25(8):1320-1325.PubMedGoogle ScholarCrossref
10.
Rooney  D, Lye  WK, Tan  G,  et al.  Body mass index and retinopathy in Asian populations with diabetes mellitus.  Acta Diabetol. 2015;52(1):73-80.PubMedGoogle ScholarCrossref
11.
Wong  TY, Klein  R, Islam  FM,  et al.  Diabetic retinopathy in a multi-ethnic cohort in the United States.  Am J Ophthalmol. 2006;141(3):446-455.PubMedGoogle ScholarCrossref
12.
Dowse  GK, Humphrey  AR, Collins  VR,  et al.  Prevalence and risk factors for diabetic retinopathy in the multiethnic population of Mauritius.  Am J Epidemiol. 1998;147(5):448-457.PubMedGoogle ScholarCrossref
13.
Jabbar  A, Irfanullah  A, Akhter  J, Mirza  YK.  Dyslipidemia and its relation with body mass index versus waist hip ratio.  J Pak Med Assoc. 1997;47(12):308-310.PubMedGoogle Scholar
14.
Brown  CD, Higgins  M, Donato  KA,  et al.  Body mass index and the prevalence of hypertension and dyslipidemia.  Obes Res. 2000;8(9):605-619.PubMedGoogle ScholarCrossref
15.
Egbe  EO, Asuquo  OA, Ekwere  EO, Olufemi  F, Ohwovoriole  AE.  Assessment of anthropometric indices among residents of Calabar, South-East Nigeria.  Indian J Endocrinol Metab. 2014;18(3):386-393.PubMedGoogle ScholarCrossref
16.
Brooke  P, Bullock  R.  Validation of a 6 item cognitive impairment test with a view to primary care usage.  Int J Geriatr Psychiatry. 1999;14(11):936-940.PubMedGoogle ScholarCrossref
17.
Consultation  WHOE; WHO Expert Consultation.  Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies.  Lancet. 2004;363(9403):157-163.PubMedGoogle ScholarCrossref
18.
Yeo  CP, Tan  CH, Jacob  E.  Haemoglobin A1c: evaluation of a new HbA1c point-of-care analyser Bio-Rad in2it in comparison with the DCA 2000 and central laboratory analysers.  Ann Clin Biochem. 2009;46(pt 5):373-376.PubMedGoogle ScholarCrossref
19.
Millán  J, Pintó  X, Muñoz  A,  et al.  Lipoprotein ratios: Physiological significance and clinical usefulness in cardiovascular prevention.  Vasc Health Risk Manag. 2009;5:757-765.PubMedGoogle Scholar
20.
Pang  C, Jia  L, Jiang  S,  et al.  Determination of diabetic retinopathy prevalence and associated risk factors in Chinese diabetic and pre-diabetic subjects: Shanghai Diabetic Complications Study.  Diabetes Metab Res Rev. 2012;28(3):276-283.PubMedGoogle ScholarCrossref
21.
Lim  LS, Tai  ES, Mitchell  P,  et al.  C-reactive protein, body mass index, and diabetic retinopathy.  Invest Ophthalmol Vis Sci. 2010;51(9):4458-4463.PubMedGoogle ScholarCrossref
22.
Fujimoto  WY, Abbate  SL, Kahn  SE, Hokanson  JE, Brunzell  JD.  The visceral adiposity syndrome in Japanese-American men.  Obes Res. 1994;2(4):364-371.PubMedGoogle ScholarCrossref
23.
Panagiotakos  DB, Pitsavos  C, Yannakoulia  M, Chrysohoou  C, Stefanadis  C.  The implication of obesity and central fat on markers of chronic inflammation: the ATTICA Study.  Atherosclerosis. 2005;183(2):308-315.PubMedGoogle ScholarCrossref
24.
Anan  F, Takayuki  M, Takahashi  N,  et al.  Diabetic retinopathy is associated with insulin resistance and cardiovascular autonomic dysfunction in type 2 diabetic patients.  Hypertens Res. 2009;32(4):299-305.PubMedGoogle ScholarCrossref
25.
El-Asrar  AM.  Role of inflammation in the pathogenesis of diabetic retinopathy.  Middle East Afr J Ophthalmol. 2012;19(1):70-74.PubMedGoogle ScholarCrossref
26.
Chaturvedi  N, Sjoelie  AK, Porta  M,  et al; EURODIAB Prospective Complications Study.  Markers of insulin resistance are strong risk factors for retinopathy incidence in type 1 diabetes.  Diabetes Care. 2001;24(2):284-289.PubMedGoogle ScholarCrossref
27.
Maneschi  F, Mashiter  K, Kohner  EM.  Insulin resistance and insulin deficiency in diabetic retinopathy of non-insulin-dependent diabetes.  Diabetes. 1983;32(1):82-87.PubMedGoogle ScholarCrossref
28.
Ruderman  N, Chisholm  D, Pi-Sunyer  X, Schneider  S.  The metabolically obese, normal-weight individual revisited.  Diabetes. 1998;47(5):699-713.PubMedGoogle ScholarCrossref
29.
Deepa  M, Farooq  S, Deepa  R, Manjula  D, Mohan  V.  Prevalence and significance of generalized and central body obesity in an urban Asian Indian population in Chennai, India (CURES: 47).  Eur J Clin Nutr. 2009;63(2):259-267.PubMedGoogle ScholarCrossref
30.
Pan  WH, Cedres  LB, Liu  K,  et al.  Relationship of clinical diabetes and asymptomatic hyperglycemia to risk of coronary heart disease mortality in men and women.  Am J Epidemiol. 1986;123(3):504-516.PubMedGoogle Scholar
31.
Lundberg  V, Stegmayr  B, Asplund  K, Eliasson  M, Huhtasaari  F.  Diabetes as a risk factor for myocardial infarction: population and gender perspectives.  J Intern Med. 1997;241(6):485-492.PubMedGoogle ScholarCrossref
32.
Després  JP, Lemieux  I.  Abdominal obesity and metabolic syndrome.  Nature. 2006;444(7121):881-887.PubMedGoogle ScholarCrossref
33.
Benarous  R, Sasongko  MB, Qureshi  S,  et al.  Differential association of serum lipids with diabetic retinopathy and diabetic macular edema.  Invest Ophthalmol Vis Sci. 2011;52(10):7464-7469.PubMedGoogle ScholarCrossref
34.
Man  RE, Sasongko  MB, Wang  JJ,  et al.  The association of estimated glomerular filtration rate with diabetic retinopathy and macular edema.  Invest Ophthalmol Vis Sci. 2015;56(8):4810-4816.PubMedGoogle ScholarCrossref
35.
DeFronzo  RA, Bonadonna  RC, Ferrannini  E.  Pathogenesis of NIDDM: a balanced overview.  Diabetes Care. 1992;15(3):318-368.PubMedGoogle ScholarCrossref
36.
Atkinson  MA, Eisenbarth  GS.  Type 1 diabetes: new perspectives on disease pathogenesis and treatment.  Lancet. 2001;358(9277):221-229.PubMedGoogle ScholarCrossref
Original Investigation
March 2016

Differential Association of Generalized and Abdominal Obesity With Diabetic Retinopathy in Asian Patients With Type 2 Diabetes

Author Affiliations
  • 1Singapore Eye Research Institute, Singapore National Eye Centre, Singapore
  • 2Duke–National University of Singapore Graduate Medical School, Singapore
  • 3Centre for Health Equity, Melbourne School of Population and Global Health, University of Melbourne, Melbourne, Victoria, Australia
  • 4Centre for Eye Research Australia, University of Melbourne, Melbourne, Victoria, Australia
  • 5Centre for Vision Research, Westmead Millennium Institute for Medical Research, University of Sydney, Sydney, New South Wales, Australia
  • 6Singapore National Eye Centre, Singapore
JAMA Ophthalmol. 2016;134(3):251-257. doi:10.1001/jamaophthalmol.2015.5103
Abstract

Importance  The association between obesity and diabetic retinopathy (DR) is equivocal, possibly owing to the strong interrelation between generalized and abdominal obesity leading to a mutually confounding effect. To our knowledge, no study in Asia has investigated the independent associations of these 2 parameters with DR to date.

Objective  To investigate the associations of generalized (defined by body mass index [BMI], calculated as weight in kilograms divided by height in meters squared) and abdominal obesity (assessed by waist to hip ratio [WHR]) with DR in a clinical sample of Asian patients with type 2 diabetes mellitus.

Design, Setting, and Participants  This cross-sectional clinic-based study was conducted at the Singapore National Eye Centre, a tertiary eye care institution in Singapore, from December 2010 to September 2013. We recruited 498 patients with diabetes. After exclusion of participants with ungradable retinal images and type 1 diabetes, 420 patients (mean [SD] age, 57.8 [7.5] years; 32.1% women) were included in the analyses.

Exposures  Body mass index and WHR as waist/hip circumference (in centimeters).

Main Outcomes and Measures  The presence and severity of DR were graded from retinal images using the modified Airlie House Classification into none (n = 189), mild-moderate (Early Treatment Diabetic Retinopathy Study scale score, 20-41; n = 125), and severe DR (Early Treatment Diabetic Retinopathy Study scale score ≥53; n = 118). The associations of BMI and WHR with DR were assessed using multinomial logistic regression models adjusting for age, sex, traditional risk factors, and mutually for BMI and WHR.

Results  Among the total of 420 patients, the median (interquartile range) for BMI and WHR were 25.7 (5.7) and 0.94 (0.08), respectively. In multivariable models, BMI was inversely associated with mild-moderate and severe DR (odds ratio [OR], 0.90 [95% CI, 0.84-0.97] and OR, 0.92 [95% CI, 0.85-0.99] per 1-unit increase, respectively), while WHR was positively associated with mild-moderate and severe DR (OR, 3.49 [95% CI, 1.50-8.10] and OR, 2.68 [95% CI, 1.28-5.62] per 0.1-unit increase, respectively) in women (P for interaction = .006). No sex-specific associations were found between BMI and DR (P for interaction >.10).

Conclusions and Relevance  In Asian patients with type 2 diabetes, a higher BMI appeared to confer a protective effect on DR, while higher WHR was associated with the presence and severity of DR in women. Our results may inform future clinical trials to determine whether WHR is a more clinically relevant risk marker than BMI for individuals with type 2 diabetes.

Introduction

Diabetic retinopathy (DR) is the most common visual complication of diabetes mellitus1 and is a leading cause of vision loss and blindness worldwide.2,3 Owing to the expected increase in DR prevalence in the coming years,4 implementation of strategies to reduce the burden of DR is imperative and demands an accurate understanding of the factors associated with DR, which remain inadequately understood despite extensive research.

Obesity is an established risk factor for type 2 diabetes5,6 and has also been posited to be a possible risk factor for the pathogenesis of DR.7According to the World Health Organization (WHO),8 2 separate classifications of obesity exist: generalized obesity, defined as body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) of 30 or greater, and central/abdominal obesity, assessed using waist circumference and/or waist to hip ratio (WHR). However, evidence supporting the association between generalized and abdominal obesity with DR has proven inconclusive. For instance, a higher BMI was separately reported to be associated with,9 or protective for,10 the presence of any DR. Similarly, equivocal results exist for the WHR-DR association.11,12

The lack of consensus in the association between these obesity-related parameters with DR may in part be due to the interrelation between the 2 factors.1315 However, to our knowledge, no study has accounted for this mutually confounding effect in its analyses. In this study, we assessed the associations between generalized obesity (defined by BMI) and abdominal (assessed using WHR) obesity with the presence and severity of DR in a well-characterized sample of Asian individuals with type 2 diabetes.

Box Section Ref ID

At a Glance

  • To investigate the associations of generalized (defined by body mass index) and abdominal (assessed as waist to hip ratio) obesity with diabetic retinopathy in a clinical sample of Asian patients with diabetes.

  • The study sample consisted of 420 participants with type 2 diabetes from the Singapore Diabetes Management Project, a clinic-based cross-sectional study.

  • While a higher body mass index was inversely associated with diabetic retinopathy in the overall population, a higher waist to hip ratio was associated with the presence and severity of diabetic retinopathy in women only.

Methods
Study Population

The Singapore Diabetes Management Project is a clinic-based cross-sectional study investigating the clinical, behavioral, and environmental barriers associated with optimal diabetes care in patients with diabetes with and without DR. In brief, we recruited a total of 498 individuals with types 1 and 2 diabetes, aged 21 years and older, from the Singapore National Eye Centre from December 2010 to September 2013. All participants were free from cognitive impairment (assessed using the 6-item Cognitive Impairment Test),16 were of sufficient hearing to be able to conduct normal conversations, and lived independently in the community (ie, not living in assisted care facilities). The presence of diabetes was defined as physician-diagnosed diabetes, with the information retrieved from participants’ case notes. Written informed consent was obtained from all participants and the study was approved by the Singapore Centralized Institutional Review Board (reference 2010/470/A) and adhered to the tenets of the Declaration of Helsinki. For this study, we included participants with type 2 diabetes of Asian ethnicity (Chinese, Malay, and Indian) who had gradable fundus photographs and optical coherence tomographic (OCT) images of 6-signal strength or greater (n = 420).

Assessment of BMI and WHR

For all measurements, participants were required to remove their shoes and heavy objects such as belts, cell phones, keys, and wallets. Height was measured in centimeters using a wall-mounted measuring tape and weight in kilograms using a digital scale. Body mass index, calculated as weight in kilograms divided by height in meters squared, was categorized into underweight (<18.5), normal (18.5-4.9), overweight (25-29.9), and obese (≥30), according to WHO-defined international BMI cut points.8 However, owing to the small sample size of individuals who were underweight (n = 4), we had to combine the underweight and normal weight categories. For supplementary analyses, we also categorized BMI using the Asian cut points for obesity as recommended by the WHO17 (<23 for under/normal weight, 23-27.5 for overweight, and >27.5 for obese).

Both waist and hip circumferences were assessed using a nonstretchable medical tape. Waist circumference (in centimeters) was taken at the smallest horizontal girth between the costal margins and the iliac crests at the end of tidal expiration, while hip measurements (in centimeters) were made at the maximal protuberance of the buttocks. The WHR was calculated by dividing the waist by the hip circumferences.

Assessment of DR and Diabetic Macular Edema

The presence and severity of DR was graded from 2-field fundus photographs (Canon CR6-45NM; Canon Inc) using the modified Airlie House classification system into none (Early Treatment Diabetic Retinopathy Study levels 10-15), mild nonproliferative DR (NPDR) (level 20), moderate NPDR (levels 31-43), severe NPDR (levels 53-60), and proliferative DR (levels 61-80). Diabetic macular edema (DME) was defined by hard exudates in the presence of microaneurysms and blot hemorrhage within 1 disc diameter from the foveal center or the presence of focal photocoagulation scars in the macular area. Clinically significant macular edema (CSME) was considered present when the macular edema was within 500 mm of the foveal center or if focal laser photocoagulation scars were present in the macular area. These were confirmed using central macular thickness measurements using OCT (described here). For analytical purposes, we reclassified the severity of DR as none (levels 10-15), mild-moderate (levels 20-43), and severe DR (≥level 53 and/or presence of CSME). Data from the worst eye were used, and if images from both eyes were ungradable, the participant was excluded from analyses.

Assessment of Macular Thickness

Macular thickness measurements were assessed by spectral-domain OCT (Cirrus Version 3.0; Carl Zeiss Meditec) using the macular thickness cube scan protocol (512 × 128). Only scans with signal strength of 6 or greater were included. Central macular thickness was defined as the central circular zone of 1 mm in the scan.

Assessment of Covariates

Information on participants’ demographic and socioeconomic characteristics (eg, age, sex, income, and education), lifestyle factors (eg, smoking), and medical history (eg, duration of diabetes) was collected using a standardized questionnaire. Trained personnel performed blood pressure measurements using a digital blood pressure machine (Dinamap Pro 100 V2; GE Heathcare). Nonfasting venous blood samples were also collected to assess glycated hemoglobin (HbA1c), serum total cholesterol, high-density lipoprotein cholesterol (HDL), low-density lipoprotein cholesterol, and triglyceride levels. All samples were analyzed at the Singapore General Hospital Hematology Laboratory. The HbA1c level was assessed via immunoassay conducted using the Roche Cobas c501 (Roche Diagnostics).18 Likewise, serum total, HDL, and low-density lipoprotein cholesterol and triglyceride levels were assessed via spectrophotometry conducted using the Beckman Coulter Unicel DxC 800 (Beckman Coulter Inc). Total cholesterol to HDL cholesterol ratio was used as a measure of dyslipidemia in analyses because this measure has been demonstrated to have greater predictive value in terms of physiological and clinical significance than isolated parameters used independently.19

Statistical Analysis

All analyses were done using intercooled Stata version 12.1 for Windows (StataCorp). We included all participants with type 2 diabetes and gradable fundus and OCT images in analyses; 420 (84.3%) satisfied these criteria. Participants’ characteristics with and without DR were compared using the χ2 statistic for proportions and a t test and/or Mann-Whitney U test for means or median as appropriate. Body mass index was analyzed continuously and categorically (using both the international and Asian WHO definitions for generalized obesity as described here), and WHR was assessed continuously and categorically in tertiles. Binomial and multinomial logistic regression models were then used to assess the associations between BMI and WHR with the presence and severity of DR (outcome), respectively, adjusted for potential confounding factors established in previous research, regardless of whether they were significant in univariate analyses. These variables included age, sex, ethnicity, insulin use, lipid-lowering medication use, triglyceride level, HbA1c level, systolic blood pressure, diabetes duration, total to HDL cholesterol ratio, BMI, and WHR. Ordinal logistic regression models were also used to estimate the overall trend of the categorical exposures with the presence and severity of DR. Because a significant interaction between sex and WHR was found (P = .006; see the Results section), we also conducted sex-stratified analyses of the 2 exposures with the presence of DR. Because only WHR displayed sex-specific differences in associations with DR presence, we did sex-stratified analyses for WHR with DR severity (Table 1).

Results

A total of 420 patients with type 2 diabetes were included in this analysis, consisting of 317 Chinese (75.5%), 36 Malay (8.6%), and 67 Indian (16.0%) individuals. The median age was 59 years (interquartile range, 9.9 years) and of these patients, 237 (56.4%) had DR, consisting of 123 (51.9%) with mild-moderate DR and 114 (48.1%) with severe DR. Table 2 summarizes the clinical and demographic characteristics of the study population in those with and without DR. Patients with DR were likely to be younger, taking insulin, have a larger WHR, have higher HbA1c level, and have longer duration of diabetes (all P < .05) compared with patients without DR. We also looked at patients’ characteristics stratified by ethnicity and found that among the 2 exposures, only BMI was significantly different between the 3 different ethnicities (eTable 1 in the Supplement).

Association of BMI With Any DR

In multivariable adjustments, we observed that increasing categories of BMI were associated with lower odds of DR (P for trend = .02; Table 3; model 2). This inverse association persisted when BMI was analyzed as a continuous variable (odds ratio [OR], 0.91; 95% CI, 0.86-0.97 per unit increase; Table 3; model 2). Sex-stratified analyses found similar inverse BMI-DR associations in both sexes, although the association was attenuated in women (Table 1).

In addition, supplementary analyses for the BMI-DR categorical associations using the Asian cut points for obesity revealed that although the association was attenuated (P for trend >.05; eTable 2 in the Supplement), the inverse direction of the association remained unchanged. Ethnicity-stratified analyses of BMI revealed similar inverse BMI-DR associations between the 3 different ethnicities, although the associations were attenuated in Malays and Indians (P > .05 for both; data not shown).

Association of BMI With the Severity of DR

Table 4 shows the associations between BMI and the severity of DR. In multivariable-adjusted models, categorical analysis of BMI failed to reveal any significant trends with increasing categories of generalized obesity (all P > .05), although per unit increase in BMI was still associated with a lower likelihood of having both mild-moderate DR (OR, 0.90; 95% CI, 0.84-0.97 per unit increase), and severe DR (OR, 0.92; 95% CI, 0.85-0.99; model 2). Supplementary analyses using the Asian cut points for obesity showed similar nonsignificant categorical trend results for mild-moderate DR and severe DR (P > .05; eTable 2 in the Supplement).

Association of WHR With Any DR

After categorization of WHR into tertiles, we identified in multivariable analyses that those in the highest tertile of WHR measurements were more likely to have any DR compared with those participants in the lowest WHR tertiles (OR, 1.88; 95% CI, 1.01-3.51) with a significant trend (P for trend = .04; Table 3; model 2). In addition, we demonstrated a linear association of increasing WHR with a greater likelihood of having any DR (OR, 1.47; 95% CI, 1.01-2.13 per 0.1-unit increase) after multivariable adjustments (Table 3; model 2).

Moreover, we further observed a significant interaction between sex and WHR (P = .006). Sex-stratified analyses revealed that in women, higher WHR was associated with the presence of DR (OR, 2.85; 95% CI, 1.48-5.48 per 0.1-unit increase; Table 1). However, the WHR-DR association was attenuated and not significant in men (P > .05; Table 1).

Association of WHR With Severity of DR

In sex-stratified multivariable-adjusted WHR-DR severity associations, a positive association was observed in men for both mild-moderate and severe DR, although these associations were not significant (all P > .05; Table 5). However, higher WHR was significantly associated with greater odds of both mild-moderate (OR, 3.49; 95% CI, 1.50-8.10 per 0.1-unit increase) and severe DR (OR, 2.68; 95% CI, 1.28-5.62 per 0.1-unit increase; Table 5) in women.

Discussion

The current study investigated the association between generalized obesity (in the form of BMI) and abdominal obesity (assessed using WHR) with the presence and severity of DR in a clinical sample of multiethnic Asian adults with type 2 diabetes in Singapore. We identified an inverse association between BMI and the presence of any and mild-moderate, but not severe, DR, while higher WHR was associated with the presence of any, mild-moderate, and severe DR, although the WHR-DR association was attenuated in men. Our results suggest that abdominal, and not generalized, obesity may play an important role in the pathophysiology of DR in patients with type 2 diabetes.

As noted previously, current research on the association between BMI and DR has been equivocal, which could be owing to their failure to account for the interdependence between generalized and abdominal obesity. Our data demonstrate that BMI is inversely associated with the presence and severity of DR independent of WHR. Interestingly, our results concur with studies conducted in Asian populations, which demonstrate either a nonsignificant or inverse BMI-DR association. For instance, the Shanghai Diabetic Complications Study reported no significant BMI-DR associations,20 while Lim et al21 and Rooney et al10 found evidence of an inverse association of BMI with the presence and severity of DR, using data from the Singapore Epidemiology of Eye Diseases cohort studies. In contrast, Western studies have reported opposing results: the Australian Diabetes Management Project, our sister study conducted in Victoria, Australia, found a detrimental association of higher BMI with DR; similarly, the Hoorn Study conducted in the Netherlands again demonstrated a significant association between higher BMI and odds of having DR.9 While it may be argued that this discrepancy could partially be because both Asian studies used obesity classifications meant for white8 instead of Asian17 individuals, supplementary analyses with the Asian categorizations demonstrated that the BMI-DR association remained largely unchanged.10 However, as these studies did not account for the mutually confounding effect of generalized and abdominal obesity, the robustness of their findings may have been affected. Therefore, more research on the effect of ethnicity on the association between BMI and DR may be warranted.

The exact mechanism underlying the protective BMI-DR association that we observed is still unclear. It is possible that patients with uncontrolled or poorly controlled diabetes and multiple comorbidities may have unintentional weight loss resulting in a lower BMI, as compared with individuals with well-controlled diabetes. However, BMI may not be an accurate measure of generalized obesity because, as its name alludes, it also includes both muscle and bone mass. Therefore, further longitudinal studies using more accurate measures of total body fat (eg, using dual-energy x-ray absorptiometry) are needed to determine the temporal nature of the generalized obesity–DR association.

In contrast to the protective association between BMI and DR, we found that WHR was associated with overall worse DR outcomes (presence and severity), although sex-stratified analyses revealed that these associations were attenuated in men. The pathophysiological mechanisms underpinning the detrimental WHR-DR association are unclear; however, abdominal obesity has been found to contribute to insulin resistance22 and inflammation,23 both of which are believed to contribute to the pathogenesis of DR.2427 To our knowledge, only one other study has reported these sex-specific WHR-DR findings: the SN-DREAMS Study found that women with isolated abdominal obesity (ie, metabolically obese, normal-weight individuals28) were more likely to have DR. The authors posited that in women, the presence of abdominal obesity was a better marker of obesity-related metabolic risk,29 leading to higher odds of DR. This is further supported by research showing that women in general are at a higher risk (than men) for cardiovascular complications resulting from insulin resistance30 and diabetes,31 both of which have been linked to abdominal obesity.32 Therefore, it is conceivable that this sex-specific susceptibility may also be a factor in the detrimental WHR-DR association in women observed in our study. Thus, further longitudinal and experimental research (eg, animal models) is warranted to elucidate the temporal nature of, as well as the mechanisms underlying, this association.

Interestingly, recent research has revealed dissimilar associated factors in the development of DR and DME, possibly implying differing underlying pathogenic mechanisms. For instance, Benarous et al33 found that serum lipids were associated with the presence of CSME, but not DR, while our group demonstrated a significant association between estimated glomerular filtration rate and DR that was not present with DME in a clinical sample of patients with type 2 diabetes in Australia.34 Owing to the small number of patients with CSME (n = 25), we grouped CSME together with severe NPDR-PDR cases for our analyses. However, supplementary analyses of BMI and WHR with CSME and central macular thickness revealed no significant associations (eTable 3 in the Supplement), despite the strong associations of the 2 exposures with the severity spectrum of DR. Hence, our results support the theory of different underlying pathophysiological processes between DR and DME. However, caution must be taken when interpreting our findings because our small sample size of patients with CSME may mean that our results were simply not powered enough to obtain significance. Therefore, further studies to determine the longitudinal impact of BMI and WHR on DME are warranted.

The strengths of this study included the assessment of DR following standardized grading protocols and a comprehensive clinical examination protocol. However, there were some limitations. First, because of the cross-sectional nature of the study, causality and the temporal sequence of these associations cannot be determined. Second, because of the clinical sample, our results may not be generalizable to the general population. However, our results are relevant to the clinic population, which has a high proportion of cases with DR. Third, 2-field fundus photographs were used to grade for the presence of DR, possibly leading to an underestimation of DR presence. Last, owing to the low number of participants with type 1 diabetes in our study (n = 10), we did not assess the association of BMI and WHR with DR in patients with type 1 diabetes. Owing to the differences in pathogenesis between the 2 types of diabetes,35,36 the effect of obesity-related changes may differ as well. Therefore, further studies to confirm the differential association between BMI and WHR with DR in patients with type 1 diabetes are warranted.

Conclusions

This study provides evidence that a higher BMI is inversely associated with the presence and severity of DR in type 2 diabetes, while a greater WHR is associated with increased likelihood and severity of the disease, particularly in Asian women. Longitudinal studies are warranted to confirm the role of abdominal obesity in the pathogenesis of DR and our results may also inform future clinical trials to determine whether WHR is a more clinically relevant risk marker than BMI for individuals with type 2 diabetes.

Back to top
Article Information

Corresponding Author: Charumathi Sabanayagam, MD, PhD, Singapore Eye Research Institute, The Academia, 20 College Rd, Singapore 169856 (charumathi.sabanayagam@seri.com.sg).

Submitted for Publication: September 10, 2015; final revision received October 8, 2015; accepted October 18, 2015.

Published Online: December 17, 2015. doi:10.1001/jamaophthalmol.2015.5103.

Author Contributions: Drs Man and Sabanayagam had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Sabanayagam, Chiang, Wong, Lamoureux.

Acquisition, analysis, or interpretation of data: Man, Chiang, Li, Noonan, Wang, Cheung, Tan, Lamoureux.

Drafting of the manuscript: Man, Wong, Lamoureux.

Critical revision of the manuscript for important intellectual content: Sabanayagam, Chiang, Li, Noonan, Wang, Cheung, Tan, Lamoureux.

Statistical analysis: Man, Sabanayagam, Li, Noonan, Tan, Lamoureux.

Obtained funding: Lamoureux.

Administrative, technical, or material support: Chiang, Wong, Cheung, Tan.

Study supervision: Chiang, Wang, Tan, Lamoureux.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. The Singapore National Eye Centre has received consulting fees, travel expenses, fees related to review activities, and payment for lectures from Abbott, Allergan, Novartis, and Bayer (Dr Wong). No other disclosures were reported.

Funding/Support: The Singapore Diabetes Management project was funded by the National Medical Research Council (Singapore) as part of its Centre/Programmatic Project grant.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Saaddine  JB, Honeycutt  AA, Narayan  KM, Zhang  X, Klein  R, Boyle  JP.  Projection of diabetic retinopathy and other major eye diseases among people with diabetes mellitus: United States, 2005-2050.  Arch Ophthalmol. 2008;126(12):1740-1747.PubMedGoogle ScholarCrossref
2.
Yau  JW, Rogers  SL, Kawasaki  R,  et al; Meta-Analysis for Eye Disease (META-EYE) Study Group.  Global prevalence and major risk factors of diabetic retinopathy.  Diabetes Care. 2012;35(3):556-564.PubMedGoogle ScholarCrossref
3.
Younis  N, Broadbent  DM, Vora  JP, Harding  SP; Liverpool Diabetic Eye Study.  Incidence of sight-threatening retinopathy in patients with type 2 diabetes in the Liverpool Diabetic Eye Study: a cohort study.  Lancet. 2003;361(9353):195-200.PubMedGoogle ScholarCrossref
4.
Ministry of Health.  National Health Survey 2010.http://www.moh.gov.sg/content/moh_web/home/Publications/Reports/2011/national_health_survey2010.html. January 16, 2014.
5.
Vazquez  G, Duval  S, Jacobs  DR  Jr, Silventoinen  K.  Comparison of body mass index, waist circumference, and waist/hip ratio in predicting incident diabetes: a meta-analysis.  Epidemiol Rev. 2007;29:115-128.PubMedGoogle ScholarCrossref
6.
Nguyen  NT, Nguyen  XM, Lane  J, Wang  P.  Relationship between obesity and diabetes in a US adult population: findings from the National Health and Nutrition Examination Survey, 1999-2006.  Obes Surg. 2011;21(3):351-355.PubMedGoogle ScholarCrossref
7.
Cheung  N, Mitchell  P, Wong  TY.  Diabetic retinopathy.  Lancet. 2010;376(9735):124-136.PubMedGoogle ScholarCrossref
8.
World Health Organization.  Obesity and overweight. http://www.who.int/mediacentre/factsheets/fs311/en/. Accessed February 20, 2015.
9.
van Leiden  HA, Dekker  JM, Moll  AC,  et al.  Blood pressure, lipids, and obesity are associated with retinopathy: the Hoorn Study.  Diabetes Care. 2002;25(8):1320-1325.PubMedGoogle ScholarCrossref
10.
Rooney  D, Lye  WK, Tan  G,  et al.  Body mass index and retinopathy in Asian populations with diabetes mellitus.  Acta Diabetol. 2015;52(1):73-80.PubMedGoogle ScholarCrossref
11.
Wong  TY, Klein  R, Islam  FM,  et al.  Diabetic retinopathy in a multi-ethnic cohort in the United States.  Am J Ophthalmol. 2006;141(3):446-455.PubMedGoogle ScholarCrossref
12.
Dowse  GK, Humphrey  AR, Collins  VR,  et al.  Prevalence and risk factors for diabetic retinopathy in the multiethnic population of Mauritius.  Am J Epidemiol. 1998;147(5):448-457.PubMedGoogle ScholarCrossref
13.
Jabbar  A, Irfanullah  A, Akhter  J, Mirza  YK.  Dyslipidemia and its relation with body mass index versus waist hip ratio.  J Pak Med Assoc. 1997;47(12):308-310.PubMedGoogle Scholar
14.
Brown  CD, Higgins  M, Donato  KA,  et al.  Body mass index and the prevalence of hypertension and dyslipidemia.  Obes Res. 2000;8(9):605-619.PubMedGoogle ScholarCrossref
15.
Egbe  EO, Asuquo  OA, Ekwere  EO, Olufemi  F, Ohwovoriole  AE.  Assessment of anthropometric indices among residents of Calabar, South-East Nigeria.  Indian J Endocrinol Metab. 2014;18(3):386-393.PubMedGoogle ScholarCrossref
16.
Brooke  P, Bullock  R.  Validation of a 6 item cognitive impairment test with a view to primary care usage.  Int J Geriatr Psychiatry. 1999;14(11):936-940.PubMedGoogle ScholarCrossref
17.
Consultation  WHOE; WHO Expert Consultation.  Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies.  Lancet. 2004;363(9403):157-163.PubMedGoogle ScholarCrossref
18.
Yeo  CP, Tan  CH, Jacob  E.  Haemoglobin A1c: evaluation of a new HbA1c point-of-care analyser Bio-Rad in2it in comparison with the DCA 2000 and central laboratory analysers.  Ann Clin Biochem. 2009;46(pt 5):373-376.PubMedGoogle ScholarCrossref
19.
Millán  J, Pintó  X, Muñoz  A,  et al.  Lipoprotein ratios: Physiological significance and clinical usefulness in cardiovascular prevention.  Vasc Health Risk Manag. 2009;5:757-765.PubMedGoogle Scholar
20.
Pang  C, Jia  L, Jiang  S,  et al.  Determination of diabetic retinopathy prevalence and associated risk factors in Chinese diabetic and pre-diabetic subjects: Shanghai Diabetic Complications Study.  Diabetes Metab Res Rev. 2012;28(3):276-283.PubMedGoogle ScholarCrossref
21.
Lim  LS, Tai  ES, Mitchell  P,  et al.  C-reactive protein, body mass index, and diabetic retinopathy.  Invest Ophthalmol Vis Sci. 2010;51(9):4458-4463.PubMedGoogle ScholarCrossref
22.
Fujimoto  WY, Abbate  SL, Kahn  SE, Hokanson  JE, Brunzell  JD.  The visceral adiposity syndrome in Japanese-American men.  Obes Res. 1994;2(4):364-371.PubMedGoogle ScholarCrossref
23.
Panagiotakos  DB, Pitsavos  C, Yannakoulia  M, Chrysohoou  C, Stefanadis  C.  The implication of obesity and central fat on markers of chronic inflammation: the ATTICA Study.  Atherosclerosis. 2005;183(2):308-315.PubMedGoogle ScholarCrossref
24.
Anan  F, Takayuki  M, Takahashi  N,  et al.  Diabetic retinopathy is associated with insulin resistance and cardiovascular autonomic dysfunction in type 2 diabetic patients.  Hypertens Res. 2009;32(4):299-305.PubMedGoogle ScholarCrossref
25.
El-Asrar  AM.  Role of inflammation in the pathogenesis of diabetic retinopathy.  Middle East Afr J Ophthalmol. 2012;19(1):70-74.PubMedGoogle ScholarCrossref
26.
Chaturvedi  N, Sjoelie  AK, Porta  M,  et al; EURODIAB Prospective Complications Study.  Markers of insulin resistance are strong risk factors for retinopathy incidence in type 1 diabetes.  Diabetes Care. 2001;24(2):284-289.PubMedGoogle ScholarCrossref
27.
Maneschi  F, Mashiter  K, Kohner  EM.  Insulin resistance and insulin deficiency in diabetic retinopathy of non-insulin-dependent diabetes.  Diabetes. 1983;32(1):82-87.PubMedGoogle ScholarCrossref
28.
Ruderman  N, Chisholm  D, Pi-Sunyer  X, Schneider  S.  The metabolically obese, normal-weight individual revisited.  Diabetes. 1998;47(5):699-713.PubMedGoogle ScholarCrossref
29.
Deepa  M, Farooq  S, Deepa  R, Manjula  D, Mohan  V.  Prevalence and significance of generalized and central body obesity in an urban Asian Indian population in Chennai, India (CURES: 47).  Eur J Clin Nutr. 2009;63(2):259-267.PubMedGoogle ScholarCrossref
30.
Pan  WH, Cedres  LB, Liu  K,  et al.  Relationship of clinical diabetes and asymptomatic hyperglycemia to risk of coronary heart disease mortality in men and women.  Am J Epidemiol. 1986;123(3):504-516.PubMedGoogle Scholar
31.
Lundberg  V, Stegmayr  B, Asplund  K, Eliasson  M, Huhtasaari  F.  Diabetes as a risk factor for myocardial infarction: population and gender perspectives.  J Intern Med. 1997;241(6):485-492.PubMedGoogle ScholarCrossref
32.
Després  JP, Lemieux  I.  Abdominal obesity and metabolic syndrome.  Nature. 2006;444(7121):881-887.PubMedGoogle ScholarCrossref
33.
Benarous  R, Sasongko  MB, Qureshi  S,  et al.  Differential association of serum lipids with diabetic retinopathy and diabetic macular edema.  Invest Ophthalmol Vis Sci. 2011;52(10):7464-7469.PubMedGoogle ScholarCrossref
34.
Man  RE, Sasongko  MB, Wang  JJ,  et al.  The association of estimated glomerular filtration rate with diabetic retinopathy and macular edema.  Invest Ophthalmol Vis Sci. 2015;56(8):4810-4816.PubMedGoogle ScholarCrossref
35.
DeFronzo  RA, Bonadonna  RC, Ferrannini  E.  Pathogenesis of NIDDM: a balanced overview.  Diabetes Care. 1992;15(3):318-368.PubMedGoogle ScholarCrossref
36.
Atkinson  MA, Eisenbarth  GS.  Type 1 diabetes: new perspectives on disease pathogenesis and treatment.  Lancet. 2001;358(9277):221-229.PubMedGoogle ScholarCrossref
×