●

aa●

●

cpT

zpM

spm●

T.oD1(abfv92aoflea●

C

A

RSRSOo

DA

1

Intravitreal Bevacizumab for Diabetic Macular EdemaAssociated With Severe Capillary Loss: One-Year

Results of a Pilot Study

MARCO BONINI-FILHO, ROGÉRIO A. COSTA, DANIELA CALUCCI, RODRIGO JORGE, LUIZ A. S. MELO, JR,

AND INGRID U. SCOTT

vfc©

Dspwafl“ttribimo2tcoofmomrftm

rprrmb

PURPOSE: To evaluate the effects of intravitreal bev-cizumab in patients with diabetic macular edema (DME)ssociated with severe capillary loss.DESIGN: Multicenter, open-label, nonrandomized study.METHODS: SETTING: Two tertiary ophthalmic referral

enters in Brazil. STUDY POPULATION: Ten consecutiveatients with DME and “severe” capillary loss. OBSERVA-ION PROCEDURES: Intravitreal injection(s) of bevaci-umab (1.5 mg). Standardized ophthalmic evaluation waserformed at baseline and at weeks 8, 16, 24, and 54.AIN OUTCOME MEASURES: Changes in best-corrected vi-ual acuity (BCVA) and in optical coherence tomogra-hy variables (central macular thickness [CMT] and totalacular volume [TMV]).RESULTS: Significant changes in BCVA and in CMT/MV were noted throughout the study (P < .001, P �

009, and P < .001, respectively). The mean logarithmf the minimal angle of resolution Early Treatmentiabetic Retinopathy Study BCVA was 0.786 (�20/25�1) at baseline, 0.646 (�20/80�2) at week 8, 0.58020/80�1) at week 16, 0.574 (�20/80�1) at week 24,nd 0.558 (�20/80�2) at week 54. Compared withaseline, a significant change in BCVA was noted at allollow-up visits (P < .008). The mean CMT/TMValues were, respectively, 472.6/10.9 at baseline, 371.4/.9 at week 8, 359.5/9.8 at week 16, 323.9/9.4 at week4, and 274.6/8.7 at week 54. Compared with baseline,significant change in both CMT and TMV was noted

nly at 24 and 54 weeks (P < .007). At 54 weeks,uorescein angiography demonstrated no change in thextent of macular capillary loss and reduced dye leakages compared with baseline in all patients.CONCLUSIONS: Favorable changes in BCVA and inMT/TMV observed throughout 1 year suggest that intra-

Supplemental Material available at AJO.com.ccepted for publication Jan 12, 2009.From the MacIma–Macular Imaging & Treatment Division (M.B.-F.,

.A.C., D.C., L.A.S.M.), Hospital de Olhos de Araraquara, Araraquara,ão Paulo, Brazil; Retina Section, Department of Ophthalmology (M.B.-F.,.A.C., R.J.), Faculdade de Medicina de Ribeirão Preto, Universidade deão Paulo, Ribeirão Preto, São Paulo, Brazil; and the Departments ofphthalmology and Public Health Sciences (I.U.S.), Penn State College

f Medicine, Hershey, Pennsylvania.Inquiries to Rogério A. Costa, MacIma–Macular Imaging & Treatment

nivision, Hospital de Olhos de Araraquara, Rua Padre Duarte 989 apto 172,raraquara SP 14801-310, Brazil; e-mail: roger.retina@globo.com

© 2009 BY ELSEVIER INC. A022

itreal bevacizumab may be a viable alternative treatmentor the management of patients with DME and severeapillary loss. (Am J Ophthalmol 2009;147:1022–1030.

2009 by Elsevier Inc. All rights reserved.)

IABETIC MACULAR EDEMA (DME) IS THE MOST

frequent cause of visual acuity (VA) loss inpatients with diabetic retinopathy (DR).1 The

afety and efficacy of macular (focal and/or grid) laserhotocoagulation for the treatment of DME have beenell evaluated by the Early Treatment Diabetic Retinop-thy Study (ETDRS) Group.2–4 Based on observationsrom several ETDRS reports, macular (focal and/or grid)aser photocoagulation was recommended for eyes withclinically significant” macular edema, particularly whenhe center of the macula is involved or imminentlyhreatened.2–4 Although eyes with DME across a broadange of baseline edema severity and VA levels werencluded in these studies, results from ETDRS reports muste interpreted cautiously because few eyes that werencluded in these analyses had severe angiographic abnor-alities other than fluorescein leakage.5 More specifically,

ut of the 1,243 eyes with DME included, only 2.25% (n �8; assigned to immediate treatment, n � 13, and assignedo deferral of photocoagulation, n � 15) of eyes werelassified as having “severe” capillary loss within 1500 �mf the center of the macula (Figure 1).5,6 Therefore, pastbservations that laser photocoagulation may be less help-ul in the treatment of DME in eyes with extensiveacular capillary loss remain valid.7 Moreover, evaluation

f the relationship between laser photocoagulation treat-ent effect and fluorescein angiographic (FA) and other

etinal characteristics at baseline performed by the ETDRSor eyes with DME have indeed demonstrated a trendoward decreasing treatment effect in eyes with even aild degree of capillary loss.5

All the above coupled with evidence suggesting a possibleole of vascular endothelial growth factor (VEGF) in theathogenesis of blood-retinal barrier breakdown in diabetes inodents and in humans,8–17 boosted by initial safety dataegarding the use of intravitreal anti-VEGF agents in hu-ans,18–21 prompted us to investigate the possible effects of

evacizumab,22 which is a full-length humanized monoclonal

eutralizing antibody against VEGF designed for intravenous

LL RIGHTS RESERVED. 0002-9394/09/$36.00doi:10.1016/j.ajo.2009.01.009

acD

A

pzl

●

PpdJ

ft1avwc�bbvpwcrpo9�c

ebmeapd(SU(imtreeipmwafc“awg(f

ero

F(c(s(aflswcfiw5cs1atmfiE

V

dministration and approved for the treatment of metastaticolorectal cancer,23,24 in the management of patients withME associated with severe capillary loss.

METHODS

PROSPECTIVE, NONRANDOMIZED, OPEN-LABEL TRIAL WAS

erformed to investigate the effects of intravitreal bevaci-umab in patients with DME associated with severe capil-ary loss.

PATIENT ELIGIBILITY AND BASELINE EVALUATION:

atients were recruited at two participating centers (Hos-ital de Olhos de Araraquara and Faculdade de Medicinae Ribeirão Preto, Universidade de São Paulo) from

IGURE 1. Early Treatment Diabetic Retinopathy StudyETDRS) standard photograph 2, used for definition of “severe”apillary loss. According to the ETDRS fluorescein angiogramsFA) grading system, capillary loss may be identified by analy-es of angiographic characteristics within 1 disc diameterconsidered to be 1500 �m) of the center of the macula, takent 1 frame per second from 13 to 28 seconds after beginning theuorescein injection by nonsimultaneous stereoscopic rapideries photographs of standard FA field 2F; the timing sequenceas intended to capture the maximum filling of the perifoveal

apillary net. This region was divided into 5 sectors (“sub-elds”), and severity scales for capillary loss characteristicsere based on the maximum severity grade found in any of thesubfields. “Severe” or grade 4 capillary loss is determined if

apillary loss equals or exceeds that observed in specifiedubfields of ETDRS standard photograph 2, which represents0% to 25% of the expected capillaries in the central subfieldnd about 40% capillary loss of the total area involving inneremporal subfield. Reprinted with permission from Early Treat-ent of Diabetic Retinopathy Study Research Group. Classi-cation of diabetic retinopathy from fluorescein angiograms:TDRS Report No. 11. Ophthalmology 1991;98:807–822.6

anuary 2006 through May 2006. All patients who met the p

BEVACIZUMAB FOR DME AOL. 147, NO. 6

ollowing inclusion criteria were offered study participa-ion: 1) DME associated with “severe” capillary loss within500 �m of the center of the macula (as assessed by FAsccording to ETDRS criteria [Figure 1])6; 2) best-correctedisual acuity (BCVA) (Snellen equivalent) of 20/40 ororse; and 3) central macular thickness (CMT) on opticaloherence tomography (OCT) evaluation greater than 250m. Exclusion criteria included: 1) glycosylated hemoglo-in (HbA1c) level above 10%; 2) history of thromboem-olic event (including myocardial infarction or cerebralascular accident); 3) uncontrolled systemic arterial hy-ertension; 4) bleeding disorders or peptic ulcer diseaseith bleeding; 5) known coagulation abnormalities orurrent use of anticoagulative medication other than aspi-in; 6) surgery (including phacoemulsification) within therior 6 months or planned within the next 28 days; 7) historyf glaucoma or ocular hypertension; 8) history of vitrectomy;) history of previous laser photocoagulation within �3000m of the center of the macula; and 10) panretinal photo-oagulation (PRP) within the prior 6 months.

At the screening visit, a comprehensive ophthalmicvaluation was performed and included a medical history,lood pressure (BP) measurement, BCVA testing (usingodified ETDRS charts 1, 2, and R), applanation tonom-

try, slit-lamp examination, dilated fundus biomicroscopy,nd ophthalmoscopy, as well as stereoscopic color fundushotography and FA (using a high-resolution digital fun-us camera system). Third-generation OCT evaluationStratus tomographer, model 3000; Carl Zeiss Ophthalmicystems Inc, Humphrey Division, Dublin, California,SA) was also performed and consisted of 6 high-density

512 A scans/image) linear 6.00-mm scans oriented atntervals of 30 degrees and centered on the foveola. Toinimize bias generated by OCT data, automatic delinea-

ion of the inner and outer boundaries of the neurosensoryetina generated by OCT built-in software was verified forach of the 6 scans using the “retinal thickness (singleye)” analysis protocol and new acquisitions were repeatedf necessary.25,26 Additionally, all OCT evaluations wereerformed in the afternoon (between 13:00 and 17:00) toinimize temporal variations.27,28 For this study, CMTas defined as the average thickness of a central macularrea 1000 �m in diameter centered on the patient’soveola. Total macular volume (TMV) values automati-ally generated by built-in OCT3 software using theretinal thickness (both eyes)” analysis protocol were alsonalyzed. If both eyes were eligible for treatment, the eyeith worse BCVA was included and macular (focal and/orrid) laser photocoagulation performed in the fellow eyeno intravitreal anti-VEGF agents were allowed in theellow eye during the study period).

Of 12 eligible patients evaluated, 2 declined because of thestimated endophthalmitis risk of approximately 1%. Theemaining 10 patients were informed verbally and in writingf the potential risks and benefits of the treatment, and all

atients signed a written informed consent form. For these

ND CAPILLARY LOSS 1023

pi

●

c

(oFt

FsS

1

atients, the screening visit was used as the baseline exam-nation.

TREATMENT ASSIGNMENT: The initial treatment pro-

IGURE 2. FAs at baseline (Top left, Patient 1; Top right, Patievere capillary loss in 4 different patients with diabetic mupplemental Figure 1 available at AJO.com).

TABLE 1. Baseline Characteristics of PatiWith Severe Capillary Loss Tre

Patient No.,

Gender, Age (Y)

Duration of

Diabetes (Y)

BP (mm Hg)

Systolic Diastolic

1, F, 60 13 118 74

2, F, 60 13 115 71

3, M, 59 11 135 82

4, F, 64 10 125 81

5, M, 50 5 132 84

6, F, 63 12 117 80

7, F, 64 17 110 66

8, M, 61 15 132 87

9, F, 59 12 116 72

10, F, 58 12 130 83

BP � blood pressure; DR � diabetic retinopa

M � months; PDR � proliferative diabetic retinop

severe nonproliferative diabetic retinopathy; Y �aBy the time PRP was initially performed.

edure was performed within 1 week of the screening e

AMERICAN JOURNAL OF024

baseline) visit. Intravitreal injection of 1.5 mg (0.06 ml)f bevacizumab (Avastin; Genentech Inc, South Sanrancisco, California, USA) was administered followinghe instillation of topical anesthetic (tetracaine/phenyl-

; Bottom left, Patient 5; Bottom right, Patient 2) exemplifyingr edema (DME) treated with intravitreal bevacizumab (see

With Diabetic Macular Edema AssociatedWith Intravitreal Bevacizumab

c Use of

Insulin

Previous

PRP

Last PRP

Episode (M)

Initial Status of

DRa

Yes Yes 12–18 PDR

Yes Yes 12–18 S-NPDR

Yes Yes 6–12 PDR

No Yes 12–18 PDR

No Yes 18–24 PDR

Yes Yes 12–18 S-NPDR

No Yes 12–18 S-NPDR

No Yes 12–18 S-NPDR

Yes Yes 6–12 S-NPDR

No Yes 12–18 S-NPDR

� female; HbA1c � glycosylated hemoglobin;

PRP � panretinal photocoagulation; S-NPDR �

rs.

ent 9acula

entsated

HbA1

(%)

7.9

8.9

6.2

8

8.1

7.95

8

8.9

9.7

9.2

thy; F

athy;

yea

phrine) drops under sterile conditions (eyelid speculum

OPHTHALMOLOGY JUNE 2009

avtdPflt

●

R

f2zbeifpcT

tzBTfiFBpsrvlw

T

resol

V

nd povidone-iodine). Bevacizumab was injected into theitreous cavity using a 29-gauge half-inch needle insertedhrough the inferotemporal pars plana 3.0 mm (pseu-ophakic) or 3.5 mm (phakic) posterior to the limbus.atients were instructed to instill 1 drop of 0.3% cipro-oxacin into the injected eye 4 times daily for 1 week afterhe procedure.

FOLLOW-UP EXAMINATIONS, RETREATMENT CRITE-

IA, AND OUTCOME MEASURES: Patients were scheduledor follow-up examinations at weeks 8 (� 1), 16 (� 2), and4 (� 2) after the first intravitreal injection of bevaci-umab. The same examination procedures performed ataseline were performed at each study visit. At the end ofach follow-up examination, an additional intravitrealnjection of 1.5 mg (0.06 ml) of bevacizumab was per-ormed if intraretinal and/or subretinal fluid (whetherersistent or recurrent) were disclosed by OCT within aircular area of 1-mm diameter centered on the foveola.

TABLE 2. Mean Best-Corrected Visual Acuitvalues), Central Macular Thickness, and To

Compared With Baseline of Patients With DCapillary Loss Treated W

LogMAR Best-C

Week n Mean SD

Baseline 10 0.786 0.383

8 10 0.646 0.436

16 10 0.580 0.396

24 10 0.574 0.378

54 10 0.558 0.380

Friedman Test, P � .001

Central Ma

Week n Mean SD

Baseline 10 472.6 205.2

8 10 371.4 148.2

16 10 359.5 140.3

24 10 323.9 115.7

54 10 274.6 55.3

Friedman Test, P � .009

Total Ma

Week N Mean SD

Baseline 10 10.883 3.297

8 10 9.901 2.404

16 10 9.787 2.109

24 10 9.385 2.043

54 10 8.751 1.618

Friedman Test, P � .001

LogMAR � logarithm of the minimal angle of

he final study visit was performed at week 54 (� 2). a

BEVACIZUMAB FOR DME AOL. 147, NO. 6

Local and systemic adverse events were monitoredhroughout the study. The effects of intravitreal bevaci-umab treatment were evaluated by changes observed inCVA and in OCT variables (CMT, measured in �m; andMV, measured in mm3). Qualitative changes in FAndings from baseline were also evaluated at week 54.riedman test was performed to analyze the changes inCVA and OCT variables throughout the 54-week studyeriod. In the situations where the Friedman test result wastatistically significant (P � .01), the Wilcoxon signed-ank test was then used to compare the BCVA and OCTariables at each follow-up visit with those at baseline. Inight of the multiple comparisons, the significance levelas set at .01 rather than .05.

RESULTS

HIS STUDY INCLUDED 7 WOMEN (70%) AND 3 MEN (30%),

logarithm of the minimal angle of resolutionacular Volume Throughout the Study andic Macular Edema Associated With Severetravitreal Bevacizumab

d Visual Acuity

Minimum Maximum

Wilcoxon

Signed-Rank Test

0.26 1.16 —

0.02 1.14 0.008

0.00 1.06 0.005

0.02 1.06 0.007

0.02 0.94 0.005

hickness

Minimum Maximum

Wilcoxon

Signed-Rank Test

257 762 —

188 559 0.07

215 603 0.03

210 584 0.005

211 410 0.007

Volume

Minimum Maximum

Wilcoxon

Signed-Rank Test

7.27 15.92 —

6.87 14.06 0.11

7.16 13.63 0.17

7.14 12.84 0.007

6.65 11.58 0.005

ution; SD � standard deviation.

y (intal Miabetith In

orrecte

cular T

cular

nd all patients completed all follow-up visits. The mean

ND CAPILLARY LOSS 1025

(((m7lapretfAbs(

t

aww2wamE

atPr�2(c

F(S

FDA

1

standard deviation [SD]) age of the patients was 59.84.0) years. Duration of diabetes prior to injection was 12.03.2) years. Five patients (50%) were using insulin. Theean (SD) systolic and diastolic BP was 123.0 (8.8) and

8.0 (6.8) mm Hg, respectively. The mean (SD) HbA1cevel was 8.28% (0.97%) at baseline, and 8.13% (1.47%)t last follow-up visit. All patients had been treatedreviously with mild or full scatter PRP. Based on aeview of patients’ medical records, DR was nonprolif-rative (severe) in 6 eyes (60%) and early proliferative inhe remaining 4 eyes (40%) at the time PRP was per-ormed. Baseline characteristics are summarized in Table 1.ll eyes had DME with center involvement as confirmed

y a CMT greater than 250 �m on OCT, and associatedevere capillary loss according to ETDRS FA criteriaFigures 1 and 2).6

A significant change in BCVA was noted throughout

IGURE 3. Color fundus photography and optical coherence toweek 52) of 1 patient with DME associated with severe capupplemental Figure 2 available at AJO.com).

IGURE 4. Red-free fundus photography and OCT evaluationME associated with severe capillary loss treated with intravitrJO.com).

he study (P � .001). The mean logarithm of the minimal 2

AMERICAN JOURNAL OF026

ngle of resolution ETDRS (Snellen equivalent) BCVAas 0.786 (�20/125�1) at baseline, 0.646 (�20/80�2) at 8eeks, 0.580 (20/80�1) at 16 weeks, 0.574 (�20/80�1) at4 weeks, and 0.558 (�20/80�2) at 54 weeks. Comparedith baseline, a significant change in BCVA was noted atll follow-up visits (P � .008) (Table 2). At week 54, theean improvement in BCVA from baseline was 2.28TDRS lines (range, �0.2 line to �4.0 lines).The mean CMT and TMV were, respectively, 472.6 �m

nd 10.88 mm3 at baseline. Significant changes were notedhroughout the study in CMT and TMV (P � .009, and� .001, respectively). The mean CMT and TMV were,

espectively, 371.4 �m and 9.90 mm3 at 8 weeks, 359.5m and 9.79 mm3 at 16 weeks, 323.9 �m and 9.39 mm3 at4 weeks, and 274.6 �m and 8.75 mm3 at 54 weeksFigures 3 and 4). Compared with baseline, a significanthange in both CMT and TMV was noted only at weeks

raphy (OCT) evaluation at baseline and at final follow-up visity loss treated with intravitreal bevacizumab (Patient 6) (see

aseline and at final follow-up visit (week 54) of 1 patient withevacizumab (Patient 3) (see Supplemental Figure 3 available at

mogillar

at beal b

4 and 54 (P � .007) (Table 2).

OPHTHALMOLOGY JUNE 2009

z(tptontwlp3n

ww

T

r2accg

Fw

Fw

V

Of a total of 4 possible intravitreal injections of bevaci-umab (ie, at baseline and at weeks 8, 16, and 24), a meanSD) of 2.9 (1.2) treatment procedures were performedhroughout the study (1 patient received 1 treatment, 4atients received 2 treatments, and 5 patients received 4reatments). No serious drug-related adverse events werebserved. The treatment procedure was well tolerated, ando clinical evidence of inflammation, uveitis, endoph-halmitis, or ocular toxicity was observed. Further, thereere no significant changes in BP, intraocular pressure, or

ens status in any of the patients during the study follow-uperiod. At the last follow-up visit, evaluation of FA data byretinal specialists (M.B.F., R.A.C., R.J.) demonstrated no

IGURE 5. FA at baseline and at final follow-up visit (week 52ith intravitreal bevacizumab (same patient illustrated in Figu

IGURE 6. FA at baseline and at final follow-up visit (week 54ith intravitreal bevacizumab (same patient illustrated in Figu

oticeable qualitative change in the status of capillary loss g

BEVACIZUMAB FOR DME AOL. 147, NO. 6

ithin the macula and reduced dye leakage as comparedith baseline in all patients (Figures 5 and 6).

DISCUSSION

HE ETDRS ADDRESSED IN DETAIL SEVERAL MAJOR ISSUES

elated to diagnosis and management of DME.2–6 For over0 years, the ETDRS recommendation that macular (focalnd/or grid) photocoagulation should be considered forlinically significant macular edema, particularly when theenter is involved or imminently threatened, remains theold-standard treatment for DME. In 1995, the ETDRS

patient with DME associated with severe capillary loss treated(see Supplemental Figure 4 available at AJO.com).

patient with DME associated with severe capillary loss treated(see Supplemental Figure 5 available at AJO.com).

) of 1re 3

) of 1re 4)

roup analyzed the possible influence of several baseline FA

ND CAPILLARY LOSS 1027

cseecalmnaeTtwoh

uciPlwptmob(rsandtybiw

foiafaocwBumobvatrim

cemtporNoTtccjrs

TdIdIH

a

1

haracteristics (such as degree of macular capillary closure,everity or source of fluorescein leakage, extent of retinaldema, presence of cystoid changes, and severity of hardxudates) on the efficacy of macular photocoagulation,5 andoncluded that the beneficial treatment effect extendedcross virtually all the aforementioned subgroups. Neverthe-ess, probably related to the fact that only eyes with mild oroderate nonproliferative DR were included, only a smallumber of the eyes included in these analyses had severengiographic abnormalities; for example, only 28 out of 1,243yes (2.25%) presented with severe capillary loss at baseline.5

herefore, the same ETDRS group suggested interpretation ofhis analysis with caution, bearing in mind that their resultsere not necessarily inconsistent with the observations ofthers that macular photocoagulation of DME might be lesselpful in eyes with extensive capillary loss.7

Among the procedures available to date for investigativese in the management of DME associated with severeapillary loss, local angiogenic pharmacomodulation utilizingntravitreal anti-VEGF agents seems a reasonable approach.revious studies have reported that VEGF levels are upregu-

ated in the aqueous humor, vitreous, and retina of patientsith DR, either nonproliferative or proliferative, as well as inatients with DME.8–17 Favorable preliminary results ob-ained with the anti-VEGF agent pegaptanib for the manage-ent of DME further supports the premise of the possible role

f VEGF in the pathogenesis of DME.19 In the current study,ecause of local unavailability of the selective anti-VEGFisoform 165) aptamer pegaptanib and promising initialesults reported with intravitreal bevacizumab at the time thistudy was designed, we decided to study the latter anti-VEGFgent. Bevacizumab is a full-length humanized monoclonaleutralizing antibody that binds to all subtypes of VEGF,22

esigned for intravenous administration and approved for thereatment of metastatic colorectal cancer.23,24 In the past 2ears, the clinical use of and research involving intravitrealevacizumab for ocular diseases has spread worldwide,29–34

ncluding its utilization for several complications associated

ith DR.35–41 o

Report No. 1. Arch Ophthalmol 1985;103:1796–1806.

AMERICAN JOURNAL OF028

In the current study, we prospectively evaluated theeasibility of intravitreal bevacizumab for the managementf DME associated with severe capillary loss. Significantmprovement in BCVA was observed as early as 8 weeksfter initial treatment, and was sustained up to the lastollow-up visit at 1 year. Favorable changes in macularrchitecture (ie, in CMT and TMV values) were alsobserved throughout the study period; however, whenompared with baseline, changes were significant only ateeks 24 and 54. Of note, favorable changes observed inCVA and in OCT variables at week 24 were maintainedp to week 54, even after a period of approximately 6onths (�30 weeks) with no additional re-injections. In

rder to investigate for a possible association of intravitrealevacizumab and induction of retinal ischemia, we re-iewed the FA findings extensively and did not find anyngiographic evidence of such an association, at leasthroughout the 1-year period of this study. Of note, in aecently reported series of 19 patients with DR undergoingntravitreal bevacizumab treatment, short-term improve-ent in peripheral retinal ischemia was demonstrated.42

In conclusion, intravitreal bevacizumab for DME asso-iated with severe capillary loss was associated with ben-ficial effects on vision and favorable remodeling of theacular architecture, and no major adverse event related

o treatment was observed during the 54-week follow-uperiod of this study. No definitive conclusions can be maden the basis of this small, uncontrolled study. Recentesults from the Diabetic Retinopathy Clinical Researchetwork have shown that macular laser photocoagulation

utperformed intravitreal injections of triamcinolone.43

he long-term performance of macular laser photocoagula-ion vs anti-VEGF agents is not known yet, particularlyonsidering the particular subset of patients included in theurrent study. Therefore, our findings should not be used asustification to treat patients in an uncontrolled fashion, butather should be used to establish initial safety data for furthertudies on intravitreal anti-VEGF agents for the management

f DME associated with severe capillary loss.

HE AUTHORS INDICATE NO FINANCIAL SUPPORT OR FINANCIAL CONFLICT OF INTEREST. INVOLVED IN CONCEPTION ANDesign (M.B.-F., R.A.C., I.U.S.); analysis and interpretation of data (M.B.-F., R.A.C., R.J., L.A.S.M., I.U.S.); writing the article (M.B.-F., R.A.C., L.A.S.M.,.U.S.); critical revision of the article (M.B.-F., R.A.C., D.C., R.J., L.A.S.M., I.U.S.); final approval of the article (M.B.-F., R.A.C., D.C., R.J., L.A.S.M., I.U.S.);ata collection (M.B.-F., D.C.); supply of materials, patients, or resources (M.B.-F., R.A.C., R.J.); literature search (M.B.-F., R.A.C., D.C., R.J., L.A.S.M.,.U.S.); and administrative, technical, or logistical support (M.B.-F., R.A.C., D.C., R.J.). This study was approved by the Comitê de Ética em Pesquisa doospital das Clínicas de Ribeirão Preto—Universidade de São Paulo (USP) e da Faculdade de Medicina de Ribeirão Preto—Universidade de São Paulo.The authors thank Luciana Miranda, MacIma-Macular Imaging & Treatment Division, Hospital de Olhos de Araraquara, São Paulo, Brazil, for crucial

dministrative and technical support.

REFERENCES1. Klein R, Klein BE, Moss SE, et al. The Wisconsin Epidemi-

ologic Study of Diabetic Retinopathy, IV: diabetic macularedema. Ophthalmology 1984;91:1464–1474.

2. Early Treatment of Diabetic Retinopathy Study ResearchGroup. Photocoagulation for diabetic macular edema: ETDRS

3. Early Treatment of Diabetic Retinopathy Study ResearchGroup. Photocoagulation for diabetic macular edema: ETDRSReport No. 4. Int Ophthalmol Clin 1987;27:265–272.

4. Early Treatment of Diabetic Retinopathy Study ResearchGroup. Early photocoagulation for diabetic retinopathy:ETDRS Report No. 9. Ophthalmology 1991;98:766–785.

5. Early Treatment of Diabetic Retinopathy Study Research

Group. Focal photocoagulation treatment of diabetic macu-

OPHTHALMOLOGY JUNE 2009

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

V

lar edema: Relationship of treatment effect to fluoresceinangiographic and other retinal characteristics at baseline:ETDRS Report No. 19. Arch Ophthalmol 1995;113:1144 –1155.

6. Early Treatment of Diabetic Retinopathy Study ResearchGroup. Classification of diabetic retinopathy from fluoresceinangiograms: ETDRS Report No. 11. Ophthalmology 1991;98:807–822.

7. Ticho U, Patz A. The role of capillary perfusion in themanagement of diabetic macular edema. Am J Ophthalmol1973;76:880–886.

8. Adamis AP, Miller JW, Bernal MT, et al. Increased vascularendothelial growth factor levels in the vitreous of eyes withproliferative diabetic retinopathy. Am J Ophthalmol 1994;118:445–450.

9. Dvorak HF, Brown LF, Detmar M, Dvorak AM. Vascularpermeability factor/vascular endothelial growth factor, mi-crovascular hyperpermeability, and angiogenesis. Am JPathol 1995;146:1029–1039.

0. Murata T, Nakagawa K, Khali A, et al. The relation betweenexpression of vascular endothelial growth factor and break-down of the blood retinal barrier in diabetic rat retina. LabInvest 1996;74:819–825.

1. Amin RH, Frank RN, Kennedy A, et al. Vascular endothe-lium growth factor is present in glial cels of the retina and theoptic nerve of human subjects with non-proliferative diabeticretinopathy. Invest Ophthalmol Vis Sci 1997;38:36–47.

2. Funatsu H, Yamashita H, Noma H, et al. Increased levels ofvascular endothelial growth factor and interleukin-6 in theaqueous humor of diabetics with macular edema. Am JOphthalmol 2002;133:70–77.

3. Funatsu H, Yamashita H, Ikeda T, et al. Angiotensin II andvascular endothelial growth factor in the vitreous fluid ofpatients with diabetic macular edema and other retinaldisorders. Am J Ophthalmol 2002;133:537–543.

4. Funatsu H, Yamashita H, Ikeda T, et al. Vitreous levels ofinterleukin-6 and vascular endothelial growth factor arerelated to diabetic macular edema. Ophthalmology 2003;110:1690–1696.

5. Funatsu H, Yamashita H, Sakata K, et al. Vitreous levels ofvascular endothelial growth factor and intercellular adhesionmolecule 1 are related to diabetic macular edema. Ophthal-mology 2005;112:806–816.

6. Funatsu H, Yamashita H, Nakamura S, et al. Vitreous levelsof pigment epithelium-derived factor and vascular endothe-lial growth factor are related to diabetic macular edema.Ophthalmology 2006;113:294–301.

7. Nguyen QD, Tatlipinar S, Shah SM, et al. Vascular endo-thelial growth factor is a critical stimulus for diabetic macularedema. Am J Ophthalmol 2006;142:961–969.

8. Gragoudas ES, Adamis AP, Cunningham ET Jr, et al; VEGFInhibition Study in Ocular Neovascularization Clinical TrialGroup. Pegaptanib for neovascular age-related macular de-generation. N Engl J Med 2004;351:2805–2816.

9. Cunningham ET Jr, Adamis AP, Altaweel M, et al. A phaseII randomized double-masked trial of pegaptanib, an anti-vascular endothelial growth factor aptamer, for diabeticmacular edema. Ophthalmology 2005;112:1747–1757.

0. Rosenfeld PJ, Moshfeghi AA, Puliafito CA. Optical coher-

ence tomography findings after an intravitreal injection of

BEVACIZUMAB FOR DME AOL. 147, NO. 6

bevacizumab (Avastin) for neovascular age-related maculardegeneration. Ophthalmic Surg Lasers Imaging 2005;36:331–335.

1. Rosenfeld PJ, Fung AE, Puliafito CA. Optical coherencetomography findings after an intravitreal injection of bevaci-zumab (Avastin) for macular edema from central retinal veinocclusion. Ophthalmic Surg Lasers Imaging 2005;36:336–339.

2. Ferrara N, Hillan KJ, Gerber HP, Novotny W. Discovery anddevelopment of bevacizumab, an anti-VEGF antibody fortreating cancer. Nat Rev Drug Discov 2004;3:391–400.

3. Yang JC, Haworth L, Sherry RM, et al. A randomized trial ofbevacizumab, an anti-vascular endothelial growth factorantibody, for metastatic renal cancer. N Engl J Med 2003;349:427–434.

4. Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumabplus irinotecan, fluorouracil, and leucovorin for metastaticcolorectal cancer. N Engl J Med 2004;350:2335–2342.

5. Costa RA, Calucci D, Skaf M, et al. Optical coherencetomography 3: Automatic delineation of the outer neuralretinal boundary and its influence on retinal thickness measure-ments. Invest Ophthalmol Vis Sci 2004;45:2399–2406.

6. Costa RA, Skaf M, Melo LA Jr, et al. Retinal assessmentusing optical coherence tomography. Prog Retin Eye Res2006;25:325–353.

7. Sternberg P Jr, Fitzke F, Finkelstein D. Cyclic macularedema. Am J Ophthalmol 1982;94:664–669.

8. Frank RN, Schulz L, Abe K, Iezzi R. Temporal variation indiabetic macular edema measured by optical coherencetomography. Ophthalmology 2004;111:211–217.

9. Avery RL, Pieramici DJ, Rabena MD, et al. Intravitrealbevacizumab (Avastin) for neovascular age-related maculardegeneration. Ophthalmology 2006;113:363–372.

0. Iturralde D, Spaide RF, Meyerle CB, et al. Intravitrealbevacizumab (Avastin) treatment of macular edema in cen-tral retinal vein occlusion: a short-term study. Retina 2006;26:279–284.

1. Rich RM, Rosenfeld PJ, Puliafito CA, et al. Short-term safetyand efficacy of intravitreal bevacizumab (Avastin) for neo-vascular age-related macular degeneration. Retina 2006;26:495–511.

2. Bashshur ZF, Bazarbachi A, Schakal A, et al. Intravitrealbevacizumab for the management of choroidal neovascular-ization in age-related macular degeneration. Am J Ophthal-mol 2006;142:1–9.

3. Costa RA, Jorge R, Calucci D, et al. Intravitreal bevacizumabfor choroidal neovascularization caused by AMD (IBeNAStudy): results of a phase 1 dose-escalation study. InvestOphthalmol Vis Sci 2006;47:4569–4578.

4. Costa RA, Jorge R, Calucci D, et al. Intravitreal bevacizumab(avastin) for central and hemicentral retinal vein occlusions:IBeVO study. Retina 2007;27:141–149.

5. Spaide RF, Fisher YL. Intravitreal bevacizumab (Avastin)treatment of proliferative diabetic retinopathy complicatedby vitreous hemorrhage. Retina 2006;26:275–278.

6. Bakri SJ, Donaldson MJ, Link TP. Rapid regression of discneovascularization in a patient with proliferative diabeticretinopathy following adjunctive intravitreal bevacizumab.Eye 2006;20:1474–1475.

7. Chen E, Park CH. Use of intravitreal bevacizumab as apreoperative adjunct for tractional retinal detachment repairin severe proliferative diabetic retinopathy. Retina 2006;26:

699–700.

ND CAPILLARY LOSS 1029

3

3

4

4

4

4

1

8. Mason JO III, Nixon PA, White MF. Intravitreal injection ofbevacizumab (Avastin) as adjunctive treatment of proliferativediabetic retinopathy. Am J Ophthalmol 2006;142:685–688.

9. Avery RL, Pearlman J, Pieramici DJ, et al. Intravitrealbevacizumab (Avastin) in the treatment of proliferativediabetic retinopathy. Ophthalmology 2006;113:1695.e1–15.

0. Haritoglou C, Kook D, Neubauer A, et al. Intravitrealbevacizumab (Avastin) therapy for persistent diffuse diabetic

macular edema. Retina 2006;26:999–1005.

AMERICAN JOURNAL OF030

1. Jorge R, Costa RA, Calucci D, et al. Intravitreal bevacizumab(Avastin) for persistent new vessels in diabetic retinopathy(IBePe study). Retina 2006;26:1006–1013.

2. Neubauer AS, Kook D, Haritoglou C, et al. Bevacizumab andretinal ischemia. Ophthalmology 2007;114:2096.

3. Diabetic Retinopathy Clinical Research Network. A ran-domized trial comparing intravitreal triamcinolone acetonideand focal/grid photocoagulation for diabetic macular edema.

Ophthalmology 2008;115:1447–1449; 1449.e1–10.

OPHTHALMOLOGY JUNE 2009

S9e

V

UPPLEMENTAL FIGURE 1. Fluorescein angiograms (FA) (; Bottom left, Patient 5; Bottom right, Patient 2) exemplifyindema (DME) treated with intravitreal bevacizumab (same pat

negative images) at baseline (Top left, Patient 1; Top right, Patientg severe capillary loss in 4 different patients with diabetic macularients illustrated in Figure 2 in the text).

BEVACIZUMAB FOR DME AND CAPILLARY LOSSOL. 147, NO. 6 1030.e1

Sw

1

UPPLEMENTAL FIGURE 2. Color fundus photography andeeks 8, 16, 24, and 52 of 1 patient with DME associated with

optical coherence tomography (OCT) evaluation at baseline and atsevere capillary loss treated with intravitreal bevacizumab (Patient 6).

AMERICAN JOURNAL OF OPHTHALMOLOGY030.e2 JUNE 2009

S1

V

UPPLEMENTAL FIGURE 3. Red-free fundus photography and OCT evaluation at baseline and at weeks 8, 16, 24, and 54 ofpatient with DME associated with severe capillary loss treated with intravitreal bevacizumab (Patient 3).

BEVACIZUMAB FOR DME AND CAPILLARY LOSSOL. 147, NO. 6 1030.e3

Sc

1

UPPLEMENTAL FIGURE 4. FA at baseline and at weeks 8, 16, 24, and 52 of 1 patient with DME associated with severeapillary loss treated with intravitreal bevacizumab (same patient illustrated in Figure 2).

AMERICAN JOURNAL OF OPHTHALMOLOGY030.e4 JUNE 2009

Sc

V

UPPLEMENTAL FIGURE 5. FA at baseline and at weeks 8, 16, 24, and 54 of 1 patient with DME associated with severeapillary loss treated with intravitreal bevacizumab (same patient illustrated in Figure 3).

BEVACIZUMAB FOR DME AND CAPILLARY LOSSOL. 147, NO. 6 1030.e5