Contents lists available at ScienceDirect

Can J Diabetes 39 (2015) 73e77

Canadian Journal of Diabetesjournal homepage:

www.canadianjournalofdiabetes.com

Original Research

Impact of Insulin Treatment in Diabetic Macular Edema Therapyin Type 2 Diabetes

Simone Matsuda MD a,b, Tiffany Tam MDa, Rishi P. Singh MDa, Peter K. Kaiser MD a, Daniel Petkovsek a,Maria Teresa Zanella MDb, Justis P. Ehlers MD a,*

aOphthalmic Imaging Center, Cole Eye Institute, Cleveland Clinic Foundation, Cleveland, Ohio, United States of AmericabDepartment of Medicine, Division of Endocrinology, Universidade Federal de São Paulo, Sao Paulo, SP, Brazil

a r t i c l e i n f o

Article history:Received 2 April 2014Received in revised form1 June 2014Accepted 27 June 2014

Keywords:afliberceptAvastin�bevacizumabdiabetic macular edemadiabetic retinopathyglycated hemoglobin (A1C)insulin therapymetabolic parametersranibizumabVEGF-inhibitors

Mots clés :afliberceptAvastinMC

bévacizumabœdème maculaire diabétiquehémoglobine glyquée (A1c)inhibiteurs du FCEVinsulinothérapieparamètres métaboliquesranibizumabrétinopathie diabétique

* Address for correspondence: Justis P. Ehlers, MD,Institute, Cleveland Clinic, Cleveland, Ohio 44195, Un

E-mail address: ehlersj@ccf.org

1499-2671/$ e see front matter � 2015 Canadian Diahttp://dx.doi.org/10.1016/j.jcjd.2014.06.005

a b s t r a c t

Objective: To evaluate the impact of insulin therapy on the outcomes of diabetic macular edema (DME)treatment with vascular endothelial growth factor (VEGF) inhibitors in people with type 2 diabetes.Methods: A retrospective consecutive case series of 95 patients with type 2 diabetes and DME who weretreated with anti-VEGF therapy. We examined 2 cohorts: patients taking only oral antidiabetic agentsand patients on insulin therapy. The main outcome measures were change in visual acuity and change incentral subfield macular thickness measured by spectral-domain optical coherence tomography. Theadditional variables analyzed included glycated hemoglobin (A1C), creatinine, blood pressure and bodymass index and their correlations with clinical findings.Results: Both groups had a statistically significant improvement in visual acuity (oral antidiabetic agentsgroup: 20/61 to 20/49, p¼0.003; insulin therapy group: 20/76 to 20/56, p¼0.005). There was no differ-ence between groups at initial or 12-month examination (p¼0.239 and p¼0.489, respectively). From ananatomic standpoint, central subfield macular thickness also improved significantly in both groups: from454.7 mm to 354.9 mm (p<0.001) in the oral antidiabetic agents group and from 471.5 mm to 368.4 mm(p<0.001) in the insulin therapy group. Again, there was no significant difference between groups atinitial or 12-month follow-up examinations (p¼0.586 and p¼0.591, respectively). Mean A1C levelsremained relatively stable during the follow up in both groups.Conclusion: Anti-VEGF therapy is a useful treatment for DME. This study suggests that chronic insulintherapy, compared with oral antidiabetic agents, does not modify the anatomic or functional effective-ness of DME treatment.

� 2015 Canadian Diabetes Association

r é s u m é

Objectif : Évaluer l’effet de l’insulinothérapie sur les résultats du traitement de l’œdème maculaire dia-bétique (OMD) par les inhibiteurs du facteur de croissance endothélial vasculaire (FCEV) chez les per-sonnes souffrant du diabète de type 2.Méthodes : Il s’agit d’une série rétrospective de 95 cas consécutifs de patients souffrant du diabète de type2 et d’OMD ayant reçu un traitement anti-FCEV. Nous avons examiné 2 cohortes: les patients prenantseulement des antidiabétiques oraux et les patients recevant une insulinothérapie. Les principaux cri-tères d’évaluation étaient les variations de l’acuité visuelle et de l’épaisseur maculaire centrale mesuréespar tomographie par cohérence optique en domaine spectral. Les variables supplémentaires analyséescomprenaient l’hémoglobine glyquée (A1c), la créatinine, la pression artérielle et l’indice de massecorporelle, ainsi que leurs corrélations avec les résultats cliniques.Résultats : Une amélioration statistiquement significative de l’acuité visuelle (groupe prenant les anti-diabétiques oraux: 20/61 à 20/49, p ¼ 0,003, groupe recevant l’insulinothérapie: 20/76 à 20/56,p ¼ 0,005) a été observée. Aucune différence entre les groupes à l’examen initial ou après 12 mois(p ¼ 0,239 et p ¼ 0,489, respectivement) n’a été notée. D’un point de vue anatomique, l’épaisseurmaculaire centrale a aussi augmenté de manière significative dans les deux groupes: de 454,7 mm à354,9 mm (p < 0,001) dans le groupe prenant les antidiabétiques oraux et de 471,5 mm à 368,4 mm(p < 0,001) dans le groupe recevant l’insulinothérapie. De nouveau, aucune différence significative entre

9500 Euclid Ave/i32, Cole Eyeited States of America.

betes Association

S. Matsuda et al. / Can J Diabetes 39 (2015) 73e7774

les groupes n’a été notée lors des examens de suivi initiaux ou après 12 mois (p ¼ 0,586 et p ¼ 0,591,respectivement). Les concentrations moyennes de l’A1c des deux groupes sont demeurées relativementstables durant le suivi.Conclusion : Le traitement anti-FCEV s’est avéré utile pour traiter l’OMD. Cette étude suggère que l’in-sulinothérapie à long terme comparativement aux antidiabétiques oraux ne modifie pas l’efficacité sur leplan anatomique ou fonctionnel du traitement de l’OMD.

� 2015 Canadian Diabetes Association

Introduction

Diabetes mellitus is a major public health problem affectingapproximately 285 million people worldwide in 2010 (1). Theproblem is only increasing; data from the Framingham Heart Studyalso indicate that the incidence of type 2 diabetes has doubled overthe past 30 years (2). Diabetic retinopathy (DR) and diabeticmacular edema (DME) are among the most significant anddisabling chronic complications of diabetes mellitus. DME is animportant cause of severe vision loss in type 2 diabetes. Hyper-permeability of retinal blood vessels and subsequent formation ofedema and hard exudates are the key clinical features. As estab-lished in several clinical trials (3e5), strict metabolic control stillremains the standard care for prevention of DR, which in manycases is achieved only by intense insulin therapy. Regarding DMEtreatment, antivascular endothelial growth factors (anti-VEGF)intravitreal injections with or without laser photocoagulation hasbecome the gold standard for reducing macular edema andimproving visual acuity (6,7).

Previous studies have demonstrated that insulin treatment ofdiabetes may result in increased retinal vascular permeability andthen induce diabetic retinopathy progression and visual impair-ment (8e11). These possible alterations are similar to the changesseen in diabetic retinopathy with increased VEGF load and mayhave an impact on the responsiveness of an eye to anti-VEGFtherapy. The aim of the present study was, therefore, to assesswhether chronic insulin therapy impacts DME response toanti-VEGF therapy.

Methods

Study population

After Cleveland Clinic Institutional Review Board approval wasobtained, a retrospective consecutive case series was performed inpatients with type 2 diabetes and DME treated with intravitrealanti-VEGF injections between January 2010 and January 2013 at theCole Eye Institute. The tenets of the Declaration of Helsinki werefollowed. Patients were included in the study if they met thefollowing inclusion criteria: type 2 diabetes, center-involving DMEas defined below, treatment with any intravitreal anti-VEGF ther-apy, baseline visual acuity less than or equal to 20/25, and spectral-domain optical coherence tomography (OCT) evaluation at allfollow-up time points (Cirrus; Carl Zeiss Meditec, Dublin,California). Exclusion criteria included type 1 diabetes (defined asjuvenile-onset diabetes due to loss of insulin production); intra-ocular surgery within 3 months of initiation of anti-VEGF therapy;intravitreal steroid injection within 3 months of initiation of anti-VEGF therapy; presence of significant media opacity that wouldlimit vision recovery (e.g. significant cataract, vitreous hemorrhage,corneal scar); presence of coexisting macular disease (e.g. age-related macular degeneration, vascular occlusive disease); vitre-omacular traction as determined by spectral-domain OCT; macularischemia if noted by the treating physician based on fluoresceinangiography, previous vitreoretinal surgery (e.g. vitrectomy); andless than 1 year of follow up after initial injection. Subjects were

divided in 2 groups: those taking oral antidiabetic agents and thoseon insulin therapy. Patients from both groups were matched atbaseline according to glycated hemoglobin (A1C) levels to avoidinterference by glycemic control. If both eyes had treated DME, 1eye of each patient was randomly selected for study inclusion.

Eye-specific clinical variables were collected and analyzedincluding Snellen visual acuity (VA), intraocular pressure,ophthalmic examination features and postinjection complications.Snellen VA was converted to logMAR notation for statisticalanalysis.

Assessment of metabolic parameters

Metabolic parameters were assessed through review of thecomprehensive medical records, including blood pressure(mm Hg); body mass index (BMI); and levels of A1C, serum glucoseand creatinine. All values aforementioned were assessed both atbaseline and at final follow up. In addition, confounding medicaldiagnoses, medications, gender, age and duration of DME (years)were also recorded. Glomerular filtration rate was calculated basedon appropriate parameters.

Assessment and treatment of DME

Center-involving DME was considered to be present based onthe presence of foveal intraretinal fluid on spectral-domain OCT inassociation with the clinical diagnosis of diabetic retinopathy andconcurrent appropriate fluorescein angiographic findings (e.g.leaking microaneurysms, if performed). At each visit, spectral-domain OCT (Cirrus, Carl Zeiss Meditec) data were collected toassess the central subfield macular thickness (CST).

All patients underwent intravitreal anti-VEGF therapy for DMEwith bevacizumab (1.25 mg). Patients were treated with a pro renata (PRN) protocol, with intravitreal bevacizumab given forcenter-involving DME. Follow-up periods were typically every 4 to8 weeks. However, given the retrospective nature of this study, thespecific treatment regimens were ultimately at the treating physi-cians’ discretion. Panretinal photocoagulation was performed inaccordance with the Early Treatment Diabetic Retinopathy Study(ETDRS) guidelines for high-risk proliferative DR, and focal/gridlaser photocoagulationwas allowed during the treatment period atthe treating physician’s discretion (12,13). The main outcomesassessed were mean changes in VA and CST in order to assess theimpact of insulin therapy on DME treatment.

Statistical analysis

Data processing and analysis were performed using SPSS v 20.0software (SPSS, Chicago, IL, US). Patients were separated into 2cohorts to assess the influence of diabetes therapy. In addition,these groups were also separated into 2 subcohorts according toA1C levels to assess the impact of glycemic control: serum A1Cvalues �7.0% (acceptable control) and serum A1C values >7.0%(poor control). The normal distribution of the variables was verifiedusing the Kolmogorov-Smirnov test. Comparisons between groupsand primary outcome variables were performed using paired t test

Table 2Visual acuity and central subfield macular thickness at the initial and the finalfollow up

Oral antidiabeticagents group (n¼46)

Insulin therapygroup (n¼49)

p

S. Matsuda et al. / Can J Diabetes 39 (2015) 73e77 75

for continuous variables. Categoric variables were expressed inproportions and analyzed by the Pearson c2 test. For all statisticaltests, p<0.05 was considered statistically significant after Bonfer-roni correction to adjust for multiple comparisons. Data wereexpressed as mean � standard deviation.

VA Initial 20/61�20/39 20/76�20/55 0.239Final 20/49�20/36 20/56�20/60 0.489

CST (mm) Initial 454.7�142.8 471.5�145.3 0.586Final 354.9�120.3 368.4�122.9 0.591

CST, central subfield macular thickness; VA, visual acuity.

Results

Demographics and systemic metabolic parameters

A total of 95 eyes of 95 patients with type 2 diabetes wereincluded in the study. Of the patients, 57 (60%) were male, and 38(40%) were female; the mean age was 65.8 years (range, 38 to88 years). Of all patients, 49 (51.7%) were taking insulin therapy (31on insulin therapy only,11 on insulin therapy plusmetformin, and 7on insulin therapy plus sitagliptin/metformin); 46 (48.3%) weretaking oral antidiabetic agents only. In the group taking insulin, thelength of time since initiating the therapy was 6.4�6.7 years (1 to26 years). At baseline examination, the mean A1C of all enrolledpatients was 7.2% (range, 5.2 to 13.5); the mean creatinine was1.4 mg/dL (range. 0.4 to 5.9 mg/dL); the mean BMI was 30.9 kg/m2

(range, 20.8 to 60.8 kg/m2); the mean systolic blood pressure was136.1 mm Hg (range, 106 to 203 mm Hg); the mean diastolic bloodpressurewas 74.5mmHg (range, 50 to 100mmHg). Of the patients,84 (88.4%) were taking antihypertensive medications. The meanduration of diabetes was 14.3 years. The mean number of intra-vitreal injections was 5.9�2.7, and the mean follow up was13.1 months.

At baseline, the mean A1C level was 7.2%�1.62 in the groupreceiving oral antidiabetic agents and 7.3%�0.96 in the groupreceiving insulin therapy (p¼0.774). The mean duration of DM was12.4�8.6 years and 15.9�8.3 years in the oral antidiabetic agentsgroup and the insulin therapy group, respectively. The meancreatinine value was 1.1�0.72 mg/dL in the oral antidiabetic agentsgroup and was 1.63�1.29 mg/dL in the insulin therapy group. Therewere no statistically significant differences between the 2 groups atbaseline. Epidemiologic and baseline clinical findings of patientsdivided in oral antidiabetics medication group and the insulintherapy group are summarized and compared in Table 1.

At final follow-up examination, most metabolic parameterswere stable. The mean A1C was 6.83 (range, 5.5 to 8.4); the meancreatinine was 1.6 mg/dL (range, 0.6 to 11.4 mg/dL); the mean BMIwas 30.1 kg/m2 (range, 19 to 54.8 kg/m2); the mean systolic bloodpressure was 134.7 mm Hg (range, 103 to 182 mm Hg); and themean diastolic blood pressure was 72.0 mm Hg (range, 40 to98 mm Hg). At final follow up, the A1C was 6.7% in the oral

Table 1Epidemiologic and clinical findings at baseline examination

Oral antidiabeticagents group n¼46

Insulin therapygroup n¼49

p

Age (years) 66.4�10.7 65.2�9.3 0.545Sex (m/f) 25/21 32/17Duration of DM (years) 12.4�8.6 15.9�8.3 0.054A1C (%) 7.2�1.62 7.3�0.96 0.774Creatinine (mg/dL) 1.1�0.72 1.63�1.29 0.061BMI (kg/m2) 29.6�5.45 32.3�7.85 0.153CKD EPI (mL/min) 66.3�23.2 45.4�30.4 0.072Hypertension (n) 38 46SBP (mm Hg) 132.1�19.5 140�17.8 0.064DBP (mm Hg) 74.6�11.9 74.3�10.7 0.886

PDR (n) 14 22NPDR (n) 32 27

BMI, body mass index; CKD-EPI, chronic kidney disease epidemiology collaborationequation; DBP, diastolic blood pressure; DM, diabetes mellitus; NPDR, non-proliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy; SBP,systolic blood pressure.

antidiabetic agents group and 6.9% in the insulin therapy group(p¼0.328). All systemic parameters did not reach a statisticallysignificant difference between groups either at the baseline or atfinal examination. The number of anti-VEGF intravitreal injectionswas the same in both groups (5.9�3.4 in the insulin therapy groupand 5.9�2.6 in the oral antidiabetic agents group).

Anatomic and functional response to anti-VEGF therapy

Regarding functional response, the mean initial best-correctedvisual acuity (BCVA) was 20/61, and it improved to 20/49(p¼0.003) at the 12-month follow up in the oral antidiabetic agentsgroup. In the insulin therapy group, the initial BCVA also improvedfrom 20/76 to 20/56 (p¼0.005); there was no significant differencebetween the groups at initial or 12-month examination (p¼0.239and p¼0.489, respectively) (Table 2). Similar to the functionalanalysis, both groups had significant anatomic improvement in CSTreduction following anti-VEGF therapy (454.7 mm to 354.9 mm,p<0.001, in the oral antidiabetic agents group; and 471.5 mm to368.4 mm, p<0.001, in the insulin therapy group). There was nodifference between the 2 groups at baseline or at the 12-monthfollow up in CST (Table 2).

When glycemic control was considered, patients with A1Clevels �7% showed a statistically significant improvement in CSTin both insulin therapy and oral antidiabetic agents groups(Table 3). On the other hand, in patients with poor glycemiccontrol (A1C levels >7%), CST improved significantly only in theinsulin therapy group (Table 3). Considering visual acuity, therewas a trend toward improvement in the well-controlled insulintherapy group and a statistically significant improvement in thepoorly controlled insulin therapy group and the well-controlledoral antidiabetic agents group. Similar to the CST findings, therewas no improvement in VA in the poorly controlled oral antidia-betic agents group (Table 3).

Discussion

DME affects approximately 7% of all people with diabetes (1,14).Although the disruption of the blood-retinal barrier (BRB), whichisolates the retina from the bloodstream, is an important feature ofDR, the basic physiologic defect that causes retinal vascular leakageis still unknown (8). The most well-known risk factor for DME ischronic hyperglycemia. It is believed that sustained hyperglycemiadamages the BRB directly, thereby leading to increased fluidextravasation and DME (15). Specifically, hyperglycemia leads to anaccumulation of free radicals and advanced glycemic end-products.Additionally, the upregulation of VEGF, prostaglandins and othercytokines alter the structure and function of the BRB (16). Othersystem-associated risk factors include hypertension, nephropathy,anemia, sleep apnea and pregnancy (15).

Many patients with type 2 diabetes require insulin therapybecause of the progressive loss of pancreatic beta cell function.However, insulin use may also increase the risk for DR and DME incertain patients (17). Poulaki et al (10) reported that acute,intensive insulin treatment in rats with diabetes causes BRB

Table 3Mean and standard deviation of visual acuity and central subfield macular thickness according to A1C levels

Initial CST (mm) Final CST (mm) p Initial VA Final VA p

Insulin therapy group A1C �7% 522.7 393.5 0.004* 20/82 20/65 0.151A1C >7% 444.7 355.8 0.003* 20/70 20/50 0.054

Oral antidiabetic agents group A1C �7% 452.5 310.2 0.000* 20/58 20/43 0.010*

A1C >7% 462.8 441.5 0.519 20/69 20/60 0.409

A1C, glycated hemoglobin; CST, central subfield macular thickness; SD, standard deviation; VA, visual acuity.* p<0.05

S. Matsuda et al. / Can J Diabetes 39 (2015) 73e7776

breakdown via VEGF expression. Henricsson et al (17) described a100% increased risk for DME with insulin treatment compared tooral medication in more than 300 insulin-treated patients. Inaddition, Zapata et al (11) found that people with type 2 diabetesreceiving insulin therapy had greater macular thickness than didthe control group. Possible mechanisms of action include upre-gulation of VEFG expression and the vasoactive effects of insulinitself as well as sudden improvement in glycemic control thatfurther compromises an already damaged BRB (10,11,15). More-over, a recent study has shown that insulin can disrupt tight celljunction in the retinal pigment epithelium cells that regulates theouter BRB (8).

To our knowledge, no previous study has examined the role ofchronic insulin therapy on the responsiveness of DME treatment byanti-VEGF intravitreal injections. In our study, patients taking oralantidiabetic medications experienced similar improvements inboth VA and anatomic parameters compared to patients on chronicinsulin therapy. Indeed, the satisfactory outcomes in both groupsare in agreement with previous clinical trials that have demon-strated a rapid improvement in both VA and CST in patients treatedwith anti-VEGF therapy for DME (6,7). Both groups of patients hadsimilar levels of A1C at the initial and final follow-up examinations,removing glycemic control as a variable. Moreover, regardless ofthe modality of DM treatment (oral antidiabetic agents or insulintherapy), patients with better glycemic control tended to achievemore satisfactory outcomes after anti-VEGF therapy at 1-yearfollow-up examinations.

The present study reveals adequate effectiveness of anti-VEGFtherapy in DME treatment, regardless of insulin use. A previousstudy by our group demonstrated that patients with initial lowerA1C levels and better glycemic control during the DME treatmentachieved better final outcomes as measured by both VA and CST(18). It is interesting that in this study, when glycemic control wasexamined, there appeared to be a particularly significant impact inthe oral antidiabetic agent group on the anatomic and functionalresponses. It is important to recognize that none of our patientswere being acutely started on insulin. Therefore, our patients didnot have an abrupt decrease of the glycemic level, considered asignificant factor for the worsening of DR and DME (9).

It is also significant to note that this research has a number oflimitations, some inherent in the retrospective nature of the study.Although DME therapy utilizes an OCT-based treatment regimenwith follow up every 4 to 8 weeks, this was not mandated, andvariability may exist, such as the possibility of under-treatment insome patients. Patients were treated with a PRN treatment proto-col; however, given the retrospective nature of the study, variabilitymay exist from physician to physician in terms of decisions to treat.Nonetheless, the number of treatments in the 2 groups was virtu-ally identical. Despite the lack of a strict standard treatment pro-tocol, our study involved 2 groups of individuals with equivalentepidemiologic and metabolic parameters who were treated with asimilar number of anti-VEGF injections. In addition, an ideal situ-ation would include only treatment-naive DME patients. Anotherlimitation consists of limited information concerning DME durationand VA assessment without ETDRS standardization. However, webelieve the study brings some value information to the pursuit of

understanding the factors involved in the responsiveness of DMEtreatment to anti-VEGF therapy.

Based on our findings, anti-VEGF therapy for DME is effectiveand provides similar results in patients taking oral antidiabeticagents or chronic insulin therapy. There is a particularly robustresponse to anti-VEGF therapy in both groups when glycemiccontrol is optimized, highlighting the critical importance ofcommunication between the physicians managing the DME andthose managing systemic diabetes control. Improving our under-standing of the factors that contribute to anti-VEGF response forDME therapy may help to enhance DME treatment outcomes,particularly through the concerted coordinated efforts of thetreating retina specialist and the endocrinologist or primary carephysician.

Author Disclosures

This work was supported in part by Coordenação de Aperfei-çoamento de Pessoal de Nível Superior (CAPES #1541-12-9) (SM);NIH/NEI #K23-EY022947-01A1 (JPE); and Research to PreventBlindness (PKK). Dr. Ehlers has served as a speaker for Regeneronand has received research grants from Genentech. Dr. Kaiser is aconsultant for Genentech, Bayer, Alcon, Novartis, Regeneron andKanghong Biotech.

Author Contributions

All authors have made substantial contributions to the infor-mation and materials in this article.

References

1. Yau JW, Rogers SL, Kawasaki R, et al. Global prevalence and major risk factors ofdiabetic retinopathy. Diabetes Care 2012;35:556e64.

2. Fox CS, Pencina MJ, Meigs JB, et al. Trends in the incidence of type 2 diabetesmellitus from the 1970s to the 1990s: The Framingham Heart Study. Circula-tion 2006;113:2914e8.

3. Implications of the Diabetes Control and Complications Trial. American Dia-betes Association. Diabetes 1993;42:1555e8.

4. Nilsson PM. ACCORD and risk-factor control in type 2 diabetes. N Engl J Med2010;362:1628e30.

5. Nathan DM. Some answers, more controversy, from UKPDS. United KingdomProspective Diabetes Study. Lancet 1998;352:832e3.

6. Mitchell P, Bandello F, Schmidt-Erfurth U, et al. The RESTORE study: Ranibi-zumab monotherapy or combined with laser versus laser monotherapy fordiabetic macular edema. Ophthalmology 2011;118:615e25.

7. Nguyen QD, Brown DM, Marcus DM, et al. Ranibizumab for diabetic macularedema: results from 2 phase III randomized trials: RISE and RIDE. Ophthal-mology 2012;119:789e801.

8. Sugimoto M, Cutler A, Shen B, et al. Inhibition of EGF signaling protects thediabetic retina from insulin-induced vascular leakage. Am J Pathol 2013;183:987e95.

9. Henricsson M, Berntorp K, Fernlund P, Sundkvist G. Progression of retinopathyin insulin-treated type 2 diabetic patients. Diabetes Care 2002;25:381e5.

10. Poulaki V, Qin W, Joussen AM, et al. Acute intensive insulin therapy exacer-bates diabetic blood-retinal barrier breakdown via hypoxia-inducible factor-1alpha and VEGF. J Clin Invest 2002;109:805e15.

11. Zapata MA, Badal J, Fonollosa A, et al. Insulin resistance and diabetic macularoedema in type 2 diabetes mellitus. Br J Ophthalmol 2010;94:1230e2.

12. Do DV, Schmidt-Erfurth U, Gonzalez VH, et al. The DA VINCI Study: phase 2primary results of VEGF trap-eye in patients with diabetic macular edema.Ophthalmology 2011;118:1819e26.

S. Matsuda et al. / Can J Diabetes 39 (2015) 73e77 77

13. Rajendram R, Fraser-Bell S, Kaines A, et al. A 2-year prospective randomizedcontrolled trial of intravitreal Bevacizumab or Laser Therapy (BOLT) in themanagement of diabetic macular edema: 24-month data: Report 3. ArchOphthalmol 2012;130:972e9.

14. Ding J, Wong TY. Current epidemiology of diabetic retinopathy and diabeticmacular edema. Curr Diab Rep 2012;12:346e54.

15. Diep TM, Tsui I. Risk factors associated with diabetic macular edema. DiabetesRes Clin Pract 2013;100:298e305.

16. Romero-Aroca P. Managing diabetic macular edema: The leading cause ofdiabetes blindness. World J Diabetes 2011;2:98e104.

17. Henricsson M, Nilsson A, Janzon L, Groop L. The effect of glycaemic control andthe introduction of insulin therapy on retinopathy in non-insulin-dependentdiabetes mellitus. Diabet Med 1997;14:123e31.

18. Matsuda S, Tam T, Singh RP, et al. The impact of metabolic parameters onclinical response to VEGF inhibitors for diabetic macular edema. J DiabetesComplications 2014;28:166e70.