Fixed-Interval Versus OCT-Guided Variable Dosing of Intravitreal Bevacizumab in the Management of Neovascular Age-Related Macular Degeneration: A 12-Month Randomized Prospective Study

vfia

wtBwtz(obrrw

Fixed-Interval Versus OCT-Guided Variable Dosing ofIntravitreal Bevacizumab in the Management ofNeovascular Age-Related Macular Degeneration:

A 12-Month Randomized Prospective Study

GEORGES M. EL-MOLLAYESS, ZIYAD MAHFOUD, ALEXANDRE R. SCHAKAL, HAYTHAM I. SALTI,

tbwm

efiA

● PURPOSE: To compare the efficacy of as-needed orariable dosing of intravitreal bevacizumab to continuousxed-interval dosing in the management of neovascularge-related macular degeneration (AMD).

● DESIGN: Prospective, open-label, randomized clinicalstudy.● METHODS: One hundred twenty eyes of 120 patients

ith treatment-naïve subfoveal neovascular AMD par-icipated in this study at the American University ofeirut and Hotel Dieu de France Retina Clinics. Eyesere randomized (1:1) to fixed-interval dosing (every 4

o 6 weeks) or variable dosing with intravitreal bevaci-umab (1.25 mg/0.05 mL). Best-corrected visual acuityBCVA) and central retinal thickness (CRT) usingptical coherence tomography (OCT) were measured ataseline and at each follow-up visit. Presence or recur-ence of fluid on OCT was the main indicator foretreatment in variable dosing. Main outcome measureas improvement in BCVA and CRT at 12 months.

● RESULTS: Compared to baseline, variable dosing had amean improvement in BCVA of 11.0 letters after 12months vs 9.2 letters for fixed-interval dosing (P � .81).Similarly, CRT decreased after 12 months by 80.7 �mfor variable dosing vs 100.5 �m for fixed-interval dosing(P � .37). The average number of injections over 12months was higher for fixed-interval dosing than variabledosing (9.5 vs 3.8 injections, P < .001).● CONCLUSIONS: Fixed-interval and variable dosing reg-imens of intravitreal bevacizumab improved visual acuityand anatomic outcomes after 12 months in eyes withneovascular AMD. However, variable dosing had a re-duced treatment burden. Larger trials are needed toconfirm these results. (Am J Ophthalmol 2012;153:481–489. © 2012 by Elsevier Inc. All rights reserved.)

Accepted for publication Aug 9, 2011.From the Department of Ophthalmology (G.M.E., H.I.S., D.J., Z.F.B.),

American University of Beirut, Beirut, Lebanon; Department of PublicHealth (Z.M.), Weill Cornell Medical College, Doha, Qatar; andDepartment of Ophthalmology (A.R.S.), Hotel Dieu de France (St.Joseph University), Beirut, Lebanon.

Inquiries to Ziad Bashshur, American University of Beirut – Medical

© 2012 BY ELSEVIER INC. A0002-9394/$36.00doi:10.1016/j.ajo.2011.08.018

T HERAPY AGAINST VASCULAR ENDOTHELIAL

growth factor (VEGF) has revolutionized the treat-ment of neovascular age-related macular degeneration

(AMD). In 2006, the US Food and Drug Administrationapproved ranibizumab, an Fab antibody fragment against allisoforms of VEGF-A, for the treatment of neovascular AMD.In the MARINA and ANCHOR trials, monthly intravit-real ranibizumab was able to stabilize visual acuity in over90% of cases and improve visual acuity in about 30% to40% of cases.1,2 The PrONTO study showed it is possibleo stabilize and improve visual acuity with variable dosingased mostly on optical coherence tomography (OCT)hile reducing the total number of injections over 12onths.3

Bevacizumab, a humanized full-length antibody againstall isoforms of VEGF-A, has been used off-label for themanagement of neovascular AMD since 2005.4 Studiesshowed that intravitreal bevacizumab was safe,5 and cur-rently different dosing regimens are used for treatment ofneovascular AMD. In the recently published CATT trial,the analysis between as-needed bevacizumab and monthlybevacizumab or as-needed bevacizumab and monthlyranibizumab was inconclusive.6 This study addresses thefficacy of as-needed bevacizumab treatment compared toxed-interval treatment in the management of neovascularMD.

METHODS

THE STUDY WAS CONDUCTED AT THE MEDICAL CENTERS

of the American University of Beirut and Hotel Dieu deFrance. Patients with treatment-naïve neovascular AMDwere enrolled if they met the following criteria: 1) age 50years or older; 2) subfoveal choroidal neovascularization(CNV) attributable to AMD diagnosed by fluoresceinangiography (FA); 3) presence of subretinal fluid, cysticmaculopathy, or central retinal thickness �250 �m onOCT; 4) best-corrected vision, using ETDRS charts, be-tween 20/40 and 20/400 (Snellen equivalent); 5) CNV lessthan 5400 �m in greatest linear dimension; and 6) abilityto understand and sign a consent form. Patients were

LL RIGHTS RESERVED. 481

btFoi

tVTtiU

dM(rawT(lasasSifla

hemorrhage; 2) media opacity that would prevent good-quality retinal imaging; 3) history of uveitis, vitrectomy,diabetic retinopathy, or other condition that may affectvision; and 4) thromboembolic event less than 6 monthsprior to enrollment. All CNV lesion types were includedexcept for retinal angiomatous proliferation (RAP) andpolypoidal choroidal vasculopathy (PCV) since they mayrespond differently to treatment. Eligible eyes were ran-domized into either the continuous fixed-interval dosinggroup or variable dosing group in a 1:1 ratio using arandomization program (GraphPad StatMate, version1.01i; GraphPad Software Inc, San Diego, California,USA). If both eyes of the same patient were eligible, thenthe eye with the worse visual acuity was enrolled. To beable to detect a difference of at least 5 letters in meanvisual acuity using the independent t test with 80% powerand an alpha level of 5%, assuming a standard deviation of10 letters, 60 eyes were needed in each group.

● BASELINE EVALUATION: Baseline evaluation includedest-corrected visual acuity (BCVA), slit-lamp examina-ion of the anterior segment, dilated fundus examination,A, and OCT. Eyes with clinical, fluorescein angiographic,r OCT characteristics suggestive of RAP or PCV had

FIGURE 1. Chart showing the treatment algorithm for theinitial phase of treatment of eyes with neovascular age-relatedmacular degeneration using a variable dosing regimen.

ndocyanine green angiography to more accurately ascer-

AMERICAN JOURNAL OF482

ain these 2 lesion types and exclude them from the study.isual acuity testing was done according to the Earlyreatment Diabetic Retinopathy Study (ETDRS) refrac-

ion protocol using ETDRS charts at 4 meters with anlluminated cabinet (Precision Vision, La Salle, Illinois,SA).One-millimeter central retinal thickness (CRT) was

etermined by spectral-domain OCT (Cirrus; Carl Zeisseditec, Dublin, California, USA) using a macular cube

512 � 128 lines). The internal limiting membrane andetinal pigment epithelium (RPE) lines were identifiedutomatically by the algorithms of the OCT machine,hich then calculated the thickness between these layers.he boundaries were confirmed by 1 of 3 investigators

Z.F.B., A.R.S., H.I.S.) and corrected using the “editayers” function if the algorithms incorrectly traced thectual boundaries. If the patient had difficulty fixating, thecan pattern in the fundus viewport was manually centeredt the fovea. Only good-quality examinations with a signaltrength of 6/10 or better were retained. The “Repeatet-up” option was used to acquire subsequent OCT. Fluid

n the macula was defined as intraretinal cysts or subretinaluid. A fluid-free macula was considered if there wasbsence of intraretinal or subretinal fluid on OCT.

● INITIAL AND FOLLOW-UP TREATMENT: Patients ran-domized to fixed-interval dosing and variable dosing re-

FIGURE 2. Mean gain in best-corrected visual acuity (BCVA)over 12 months of eyes treated with fixed-interval dosing (60eyes) vs OCT-guided variable dosing (60 eyes) of intravitrealbevacizumab for subfoveal choroidal neovascularization in age-related macular degeneration. The table below the graph showsthe absolute values of BCVA (in letters) for every month.

ceived the first and subsequent intravitreal bevacizumab

OPHTHALMOLOGY MARCH 2012

f

et

bfi

(Avastin; Genentech, South San Francisco, California,USA) injections based on a standard protocol.7,8 At each

FIGURE 3. Scatterplot of baseline visual acuity vs visual acuity aevacizumab for subfoveal choroidal neovascularization attributabxed dosing regimen or (Bottom) as-needed variable dosing. All p

ollow-up, BCVA, slit-lamp examination, dilated fundus i

FIXED-DOSING VS VARIABLE-DOSIVOL. 153, NO. 3

xamination, and OCT were performed. The response toreatment on OCT was described as: 1) complete response

onths (expressed in letters) for 60 eyes that received intravitrealage-related macular degeneration using either (Top) a monthlyabove the line imply improved final visual acuity.

t 12 mle tooints

f there was total absence of fluid in the macula; 2) partial

NG BEVACIZUMAB FOR AMD 483

1d

ddtnv1OcrflersO

response if there was a decrease of at least 50% in CRTwith persistent macular fluid; 3) minimal response if therewas a decrease of less than 50% in CRT with persistentmacular fluid; and 4) no response if there was no change orincrease in macular fluid. FA was repeated at the discretionof the treating physician.

Patients enrolled in the fixed-interval dosing groupreceived bevacizumab injections every 6 weeks unless theeye was not showing a complete response after the 2 initialconsecutive injections. If there was evidence of fluid onOCT, 6 weeks after the second injection, then the thirdinjection was administered and treatment interval wasreduced to every 4 weeks. Once a fluid-free macula wasachieved, the treatment interval could be extended backto 6 weeks at the discretion of the treating physician.

Patients in the variable dosing group were seen 6 weeksafter the initial treatment. If there was a complete re-sponse, the patient was followed every 6 weeks. If there waspersistent fluid in the macula on OCT, a second bevaci-zumab injection was given and follow-up made 6 weekslater. If there was evidence of any fluid in the macula onOCT 6 weeks after the second injection, a third injectionwas administered and further follow-up and treatment wereconducted every 4 weeks until the macula was fluid-free.However, if there was persistent fluid on OCT after 3additional consecutive injections, the treating physicianhad the option to stop injections and observe. In this case

FIGURE 4. Mean change in central retinal thickness (CRT)over 12 months of eyes treated with fixed-interval dosing (60eyes) vs OCT-guided variable dosing (60 eyes) of intravitrealbevacizumab for subfoveal choroidal neovascularization in age-related macular degeneration. The table below the graph showsthe absolute values of CRT (�m) for every month.

the patient would have received a total of 6 bevacizumab

AMERICAN JOURNAL OF484

injections over a period of 24 weeks. Based on personalobservation and other studies, eyes with some persistentfluid in the macula can show stability in BCVA.7,8 Figure

summarizes the initial treatment protocol for the variableosing group.Follow-up was carried out every 6 weeks in the variable

osing group once the macula was dry on OCT or it wasecided to observe persistent fluid in the macula. However,his interval could be reduced to every 4 weeks if deemedecessary by the treating physician. Retreatment in theariable dosing group was based on the following criteria:) recurrence or presence of any fluid in the macula onCT in a previously dry macula; 2) an increase in OCT

entral retinal thickness of at least 50 �m from the lowestecorded value, especially for eyes with persistent macularuid; 3) visual acuity loss of at least 5 letters with OCTvidence of fluid in the macula; 4) new macular hemor-hage; or 5) new area of classic CNV. Persistent and stableerous retinal pigment epithelial detachment (PED) onCT was not a retreatment criterion.Once the 2-year results of the PrONTO trial9 were

available in July 2009, retreatment criteria for variabledosing was revised to include, in addition to those previ-ously mentioned, any OCT changes suggestive of an activeCNV. This included the appearance or increase in size ofa retinal cyst or subretinal fluid. Although it was not ourintention to achieve total resolution of a PED alreadypresent at baseline, the appearance of a new PED orincrease in size of a previously stable PED became aretreatment criterion even in the absence of retinal fluid.These amendments to the retreatment criteria occurredless than 2 months after patient recruitment commenced.High-definition 5-line raster scan centered on the foveawas also obtained for further qualitative assessment of thecentral macular area. In addition, the treating physicianhad the option to do high-definition scans to any suspi-cious areas of the macula that were not clear on the cubescan. In December 2010, OCT software was upgraded toversion 4.5, which allowed “macular change analysis.” Thismade possible side-by-side and line-to-line comparisonbetween different OCT studies, allowing more precisequalitative analysis.

● TREATING PHYSICIANS AND MASKING: The 3 treat-ing physicians (Z.F.B., H.I.S., A.R.S.) are retina specialistsexperienced in the management of neovascular AMD withanti-VEGF therapy using the same treatment protocol asthis study. Treating physicians were not masked to thetreatment regimen of patients under their care and nosham injections were employed. However, visual acuityexaminers were masked to treatment regimen and patientswere instructed not to share this information with theexaminer. Each patient was assigned a unique identifica-tion number on all study documents and imaging modali-ties. The physician reviewing OCT images or other

OPHTHALMOLOGY MARCH 2012

demflmeme

sfari

lp

SU

Mptvppwmlms

itm6Waw

particular patient’s identity and treatment regimen and inno way could be involved in the treatment of that patient.

● ADVERSE EVENTS: Severe ocular adverse events wereefined as sudden loss of 30 letters or more in BCVA,ndophthalmitis, vitreous hemorrhage, retinal detach-ent, cataract formation or progression, intraocular in-

ammation, and intraocular pressure persistently above 21m Hg. Severe systemic adverse events were defined as

levation of arterial blood by more than 10 mm Hg, stroke,yocardial ischemic event, and peripheral thrombotic

vents.

● STATISTICAL ANALYSIS: Sample characteristics wereummarized for the overall sample and for each group usingrequency distribution for sex and lesion type, while meannd standard deviation were used for age. To ensure thatandomization yielded comparable groups, these character-stics were compared between the 2 study arms using �2

test for sex and lesion type and independent t test for age.The main outcome measures, improvement in BCVA

and CRT, were analyzed using linear regression where thedependent variable was BCVA at 12 months (or CRTvalue at 12 months) and the independent variables werebaseline BCVA (or CRT) and dosing regimen. We alsoperformed the independent t test and the Wilcoxon ranksum test to compare relative change in BCVA and CRTbetween the 2 groups. BCVA and CRT were comparedbetween baseline and 12 months within each group usingthe paired t test and Wilcoxon signed rank test.

The number of eyes avoiding loss of 15 letters and thosethat gained at least 15 letters was compared between the 2

TABLE. Comparison of Eyes With NeovasShowed a Partial vs Complete R

Pa

Sex, n (%)

Male

Female

Lesion type, n (%)

Occult

Minimally classic

Predominantly classic

Age, years

Baseline BCVA (letters)

Baseline CRT (�m)

Change in BCVA at 12 months (letters)

Change in CRT at 12 months (�m)

BCVA � best-corrected visual acuity; CRT �aP value, significant at 5% level.b�2 test.cWilcoxon rank sum test.

groups using the �2 test or Fisher exact test for cell counts a

FIXED-DOSING VS VARIABLE-DOSIVOL. 153, NO. 3

ess than 5. The average number of injections was com-ared between the 2 groups using Wilcoxon rank sum test.All analyses were done using Statistical Package for Social

ciences (SPSS, version 17; SPSS Inc, Chicago, Ilinois,SA). A P value of .05 or less was considered statistically

significant.

RESULTS

● DEMOGRAPHICS: Recruitment for the study started inay 7, 2009 and was completed by October 12, 2009, with 60

atients randomized to each group. All patients completedhe 12 months of the study and were able to make scheduledisits with no greater than a 7-day delay. The majority of thearticipants of both groups combined were female (78/120atients, 65.0%) and had an average age of 76.8 years. Thereere no significant differences in sex distribution (P � .25) orean age (P � .60) between the 2 groups. The most frequent

esion type was pure occult (50/120 eyes, 41.7%), followed byinimally classic (46/120 eyes, 38.3%). Lesion type did not

ignificantly differ between the 2 groups (P � .17).

● VISUAL ACUITY AND CENTRAL RETINAL THICK-

NESS: At baseline, mean BCVA was 56.6 and 53.3 lettersn the fixed-interval and variable dosing groups, respec-ively (t test, P � .21; Wilcoxon, P � .14). At 12 months,ean BCVA increased by 11.0 letters from baseline, to

4.3 letters, in the variable dosing group (t test andilcoxon, P � .001). The fixed-interval dosing group

chieved a mean BCVA of 65.8 letters at 12 months,hich was a 9.2-letter increase compared to baseline (t test

Age-Related Macular Degeneration Thatnse to Intravitreal Bevacizumab

esponders

19

Complete Responders

n � 101 P Valuea

2.6%) 32 (31.7%) .08b

7.4%) 69 (68.3%)

6.3%) 45 (44.6%) .25b

2.1%) 38 (37.6%)

1.6%) 18 (17.8%)

76.3 76.9 .81c

51.8 55.5 .31c

14.3 337.7 �.001c

8.6 10.4 .58c

02.0 �88.4 .50c

ral retinal thickness.

cularespo

rtial R

n �

10 (5

9 (4

5 (2

8 (4

6 (3

4

�1

cent

nd Wilcoxon, P � .001). Increase in mean BCVA by 12

NG BEVACIZUMAB FOR AMD 485

Pt

olel(

dW1fibt1(P1

.fiad6id

(ag.fiird

gbbsCpfB2i

wsafiemca

Ptmm

months did not differ significantly between the 2 groups(regression analysis, P � .81; t test, P � .37; Wilcoxon,

� .50). Figures 2 and 3 illustrate the gain in BCVA overhe course of the study.

Sixty of 60 eyes (100%) in fixed-interval dosing and 59f 60 eyes (98.3%) in variable dosing avoided losing ateast 15 letters of BVCA (P � .99). Twenty-four of the 60yes (40%) in the variable dosing group gained at least 15etters vs 21 of 60 eyes (35%) for fixed-interval treatmentP � .57).

Baseline mean CRT was 350.0 �m for fixed-intervalosing and 341.9 �m for variable dosing (t test, P � .31;

ilcoxon, P � .25). Compared to baseline, mean CRT at2 months decreased by 80.7 �m to 261.2 �m for thexed-interval group (t test and Wilcoxon, P � .001) andy 100.5 �m to 268.9 �m for the variable dosing group (test and Wilcoxon, P � .001). Decrease in mean CRT at2 months did not differ significantly between the 2 groupsregression analysis, P � .37; t test, P � .13; Wilcoxon,� .26). Figure 4 shows the mean decrease in CRT over

2 months.

● NUMBER OF INJECTIONS AND QUALITATIVE OPTICAL

COHERENCE TOMOGRAPHY FINDINGS: By 12 months,the average number of injections received per patient was3.8 for variable dosing vs 9.5 for fixed-interval dosing (P �001). From the initiation of treatment, eyes in thexed-dosing group showed a fluid-free macula after anverage of 2.1 injections vs 1.8 injections for variableosing (P � .11). In the fixed-interval dosing group, 49 of0 eyes (81.7%) achieved a fluid-free macula after 3njections or less; 47 of 60 eyes (78.3%) in the variableosing group showed a similar response (P � .82). In the

variable dosing group, the mean treatment-free intervalonce the macula became fluid-free was 4.6 months.

None of the eyes in either group showed minimal or noresponse. However, 9 of 60 eyes (15.0%) in variable dosingand 10 of 60 eyes (16.7%) in fixed-interval dosing showeda partial response (P � .80) during the 12 months of thestudy. Anatomic and functional outcomes of partial re-sponders were similar in both treatment groups. By 12months, partial responders in the fixed-interval dosinggroup improved BCVA by a mean of 10.0 letters whilethose with variable dosing improved by a mean of 7.1letters (P � .68). Similarly, CRT decreased after 12months in the partial responders of the fixed-interval groupby a mean of 91.0 �m vs 114.2 �m for variable dosingP � .84). Over the 12 months, partial responders receivedn average of 10.1 injections in the fixed-interval dosingroup and 6.3 injections in the variable dosing group (P �001). Of the 10 eyes that were partial responders in thexed-interval group, 3 eyes were complete respondersnitially (ie, after the first 3 injections) and became partialesponders later in the course of treatment. This situation

AMERICAN JOURNAL OF486

Partial responders of both treatment regimens wererouped together and compared to complete responders ofoth groups combined. There was no significant differenceetween partial and complete responders with respect toex, age, lesion type, and baseline BCVA. Only baselineRT was significantly greater in partial responders com-ared to the complete responders. This was usually in theorm of large intraretinal cysts. The mean improvement inCVA and CRT at 12 months was similar between theseresponse patterns. The Table summarizes the character-

stics of partial and complete responders.

● ADVERSE EVENTS: No severe ocular adverse eventsere noted in both groups over 12 months. Similarly, no

ystemic adverse events were reported. However, 3 monthsfter the completion of the study, 5 patients in thexed-interval dosing group had major thromboembolicvents. Two patients had a stroke and 3 patients had ayocardial ischemic event that necessitated insertion of a

ardiac stent. All 5 patients had prior history of coronaryrtery disease.

DISCUSSION

THIS IS A PROSPECTIVE, OPEN-LABEL, RANDOMIZED CLINI-

cal study that compares fixed-interval dosing to variabledosing of intravitreal bevacizumab in the management ofneovascular AMD. By 12 months, gain in BCVA anddecrease in CRT were similar in both regimens. Thisimprovement occurred after the first injection, then stabi-lized through the study. Over 95% of eyes in bothtreatments avoided loss of 15 letters in BCVA, andimprovement by at least 15 letters occurred in a similarpercentage of eyes in both groups.

Both dosing regimens required a similar number ofinjections to achieve a fluid-free macula at the initiation oftreatment and the majority did not require more than 3consecutive injections. Arias and associates showed thatan induction phase with bevacizumab gave better visualoutcome over a 6-month period.10 Similarly, Gupta andassociates showed that induction with 3 injections ofranibizumab resulted in more visual gains compared toas-needed treatment without a loading phase.11 We didnot have a fixed induction phase in the variable dosinggroup. We assumed that further injections once the maculawas fluid-free may not be of benefit.

Most trials that investigated dosing regimens were con-ducted with ranibizumab and seemed to favor frequentfixed-interval dosing for greatest visual acuity gains.12 The

IER13 and EXCITE14 studies showed that quarterly fixedreatments may not maintain the gains of the initial 3onthly injections. Even monthly follow-up during theaintenance phase in the PrONTO3 and SUSTAIN15

studies did not preserve or increase the gain of the initial

OPHTHALMOLOGY MARCH 2012

dtammatbi

Cvssfpei

db

erpeigBmtaCbvtwhda

retreatment threshold based on subjective OCT changes,which improved visual acuity gains. Variable dosing in theCATT6 had a low retreatment threshold for any sign ofisease activity with no minimum threshold for retinalhickness. At the end of the first year of the CATT,s-needed ranibizumab treatment was noninferior toonthly ranibizumab treatment. We amended our retreat-ent criteria at the early stages of the study to take

dvantage of this potential benefit. Our low retreatmenthreshold for the OCT-guided variable dosing group maye the reason for similar results to the frequent fixed-nterval dosing group.

This study appears in the shadow of the much largerATT. In the CATT, the comparison of bevacizumabariable dosing to fixed-dosing bevacizumab was inconclu-ive.6 In our study, variable and fixed dosing achievedimilar gains in BCVA. We were able to achieve aollow-up or retreatment interval of 6 weeks with mostatients while the CATT had follow-up and retreatmentvery 4 weeks. We assumed that a retreatment/follow-upnterval of 6 weeks would be appropriate16 but made

provision to decrease this interval to 4 weeks if there wasinsufficient response after 6 weeks. It seems that PCV andRAP may not have been excluded in the CATT, nor wasthere a limitation on lesion size.17 Furthermore, some eyesid not have subfoveal CNV while eyes with subfoveallood were not excluded.6 These factors may influence

treatment outcome, which is why we excluded these eyesfrom our study. The CATT administered a treatment inthe variable dosing group if there was any fluid on OCT,which included sub-RPE fluid.18 Possibly stable or persis-tent PED were treated in the CATT trial, and this mayincrease the frequency or total number of injections sincePED tend to have a slower response to anti-VEGF ther-apy.7 A stable PED without associated retinal fluid mayhave little effect on visual acuity. Therefore, we onlytreated a newly appearing PED or pre-existing PED thatincreased in size, assuming this represents an active CNV.This may help decrease the total number of injectionswithout compromising visual acuity gains or stability.

In this study, the treating physician and patient werenot masked to the type of treatment regimen, which maybe a cause of bias. The fact that vision examiners and thephysician reviewing OCT or other study data were maskedto treatment regimen may have decreased this bias. Wealso did not have a reading center to confirm that treatingphysicians were adhering to the study protocol, which is asignificant drawback. We used spectral-domain OCT,which provides a superior resolution to time-domain OCTwith a larger sample of the macula and smaller gapsbetween scans. This may have improved our ability todetect very early recurrences and subtle changes that maybe overlooked by time-domain OCT used in the CATT.Unlike the CATT, which did not have a minimumincrease in retinal thickness to warrant retreatment, our

FIXED-DOSING VS VARIABLE-DOSIVOL. 153, NO. 3

if there was an increase in CRT of at least 50 �m, even inthe absence of subjective increase of retinal fluid. This mayact as a fail-safe in case the treating physician missed anysubjective OCT changes. However, the ability to makeside-by-side comparisons of 2 OCT studies helped detectminute OCT changes well before there was an objectiveincrease of 50 �m in CRT. We believe that spectral-domain OCT with the additional use of high-definitionraster scans made adherence to the study protocol easier forthe treating physicians. It will be interesting to see howspectral-domain OCT will impact the second-year resultsof the CATT.6

As in previous studies, we found that some eyes con-tinue to manifest fluid in the macula despite treatmentwith intravitreal bevacizumab.7,8 This occurred equally inither a fixed-interval or variable dosing protocol. Partialesponders of both treatment regimens had similar im-rovement in BCVA and CRT. However, partial respond-rs in variable dosing received fewer bevacizumabnjections. Partial responders of both treatment strategiesrouped together had similar improvement in CRT andCVA compared to complete responders of both treat-ent groups combined. Engelbert and associates suggested

hat type 1 (sub-RPE) CNV may play a protective rolegainst geographic atrophy.19 Therefore, allowing occultNV to persist under a controlled situation may haveenefits against atrophic changes. Based on our results,ariable dosing with close follow-up and low retreatmenthreshold may be effective even for partial respondershile reducing the treatment burden. Similarly, CATTad between 50% and 80% of eyes with fluid on OCT butid not seem to have a significant effect on gain in visualcuity.6 Since CATT considered sub-RPE fluid as “fluid on

OCT,” it is not clear how many eyes had only sub-RPEfluid with no subretinal or intraretinal fluid.

It seems that partial responders can appear at the onsetof treatment, in which case this may be attributable tosome intrinsic resistance of the CNV to anti-VEGF ther-apy.20–23 Partial responders in our study tended to havegreater baseline CRT in the form of large intraretinal cysts,indicating chronic CNV with mature VEGF-independentvessels. This was similarly observed by Byun and associ-ates.24 Partial response can also develop later in the courseof treatment because of a type of tachyphylaxis to anti-VEGF treatment.25 Three eyes in fixed-interval treatmentshowed this phenomenon. Reducing the number of bev-acizumab treatments with variable dosing may delay theonset of late resistance.

We did not observe any ocular or systemic adverseevents over 12 months with either treatment group. Fivepatients in the fixed dosing group had a thromboembolicevent at least 3 months after the end of the study. At thattime, these patients were under a variable dosing regimen.It is difficult to draw a causal relationship between frequentfixed-interval dosing of intravitreal bevacizumab and these

NG BEVACIZUMAB FOR AMD 487

when the patients were not in the more controlledenvironment of the study. It seems reasonable to try andminimize the potential systemic complications of thistreatment by reducing the number of injections, at least inpatients with known risks for thromboembolic disease.

In conclusion, intravitreal bevacizumab for neovascular

Ophthalmol 2008;92(12):1636–1641.

1

1

1

1

1

1

1

1

1

2

2

AMERICAN JOURNAL OF488

improvement. This was equally possible with frequentfixed-interval dosing or OCT-guided variable dosing. Themain advantage of variable dosing is reduced treatmentburden with a potentially higher safety profile. Variabledosing requires frequent follow-up with meticulous subjec-tive assessment of the OCT and a low retreatment

PUBLICATION OF THIS ARTICLE WAS SUPPORTED BY THE DEPARTMENT OF OPHTHALMOLOGY AND UNIVERSITY RESEARCHBoard of American University of Beirut Medical Center, Beirut, Lebanon. The authors indicate no financial interest in any product discussed in this study.Involved in design of study (Z.F.B., Z.M.); conduct of study (Z.F.B., G.M.E., A.R.S., H.I.S.); data collection (G.M.E., D.J.); management, analysis, andinterpretation of the data (Z.F.B., Z.M.); and preparation (Z.F.B., Z.M., G.M.E.), review (Z.F.B., Z.M., G.M.E., A.R.S., H.I.S., D.J.), and approval of themanuscript (Z.F.B., Z.M., G.M.E., A.R.S., H.I.S., D.J.). This study was approved prospectively by the institutional review board at the American Universityof Beirut Medical Center and was in adherence to the tenets of the Declaration of Helsinki. All patients signed a study consent form in which the procedureand treatment options were explained thoroughly. Patients were also reminded of the off-label use of intravitreal bevacizumab.

REFERENCES

1. Rosenfeld PJ, Brown DM, Heier JS, et al. Ranibizumab forneovascular age-related macular degeneration. N Engl J Med2006;355(14):1419–1431.

2. Brown DM, Kaiser PK, Michels M, et al. Ranibizumab versusverteporfin for neovascular age-related macular degenera-tion. N Engl J Med 2006;355(14):1432–1444.

3. Fung AE, Lalwani GA, Rosenfeld PJ, et al. An opticalcoherence tomography-guided, variable dosing regimen withintravitreal ranibizumab (Lucentis) for neovascular age-re-lated macular degeneration. Am J Ophthalmol 2007;143(4):566–583.

4. Rosenfeld PJ, Moshfeghi AA, Puliafito CA. Optical coher-ence tomography findings after an intravitreal injection ofbevacizumab (Avastin) for neovascularization age-relatedmacular degeneration. Ophthalmic Surg Lasers Imaging2005;36(4):331–335.

5. Wu L, Martínez-Castellanos MA, Quiroz-Mercado H, et al.Twelve-month safety of intravitreal injections of bevaci-zumab (Avastin®): results of the Pan-American Collabora-tive Retina Study Group (PACORES). Graefes Arch ClinExp Ophthalmol 2008;246(1):81–87.

6. The CATT Research Group. Ranibizumab and bevacizumabfor neovascular age-related macular degeneration. N EnglJ Med 2011;364:1897–1908.

7. Bashshur ZF, Haddad ZA, Schakal AR, Jaafar RF, Saad A,Noureddin BN. Intravitreal bevacizumab for treatment ofneovascular age related macular degeneration: the secondyear of a prospective study. Am J Ophthalmol 2009;148(1):59–65.

8. Bashshur ZF, Haddad ZA, Schakal A, Jaafar RF, Saab M,Noureddin BN. Intravitreal bevacizumab for treatment ofneovascular age related macular degeneration: a one yearprospective study. Am J Ophthalmol 2008;145(2):249–256.

9. Lalwani GA, Rosenfeld PJ, Fung AE, et al. A variable-dosingregimen with intravitreal ranibizumab for neovascular age-related macular degeneration: year 2 of the PrONTO Study.Am J Ophthalmol 2009;148(1):43–58.

10. Arias L, Caminal JM, Casas L, et al. A study comparing twoprotocols of treatment with intravitreal bevacizumab (Avas-tin) for neovascular age-related macular degeneration. Br J

1. Gupta B, Adewoyin T, Patel SK, Sivaprasad S. Comparisonof two intravitrealranibizumab treatment schedules for neo-vascular age-related macular degeneration. Br J Ophthalmol2011;95(3):386–390.

2. Mitchel P, Korobelnik JF, Lanzetta P, et al. Ranibizumab(Lucentis) in neovascular age-related macular degeneration:evidence from clinical trials. Br J Ophthalmol 2010;94(1):2–13.

3. Regillo CD, Brown DM, Abraham P, et al. Randomized,double-masked, sham-controlled trial of ranibizumab forneovascular age-related macular degeneration: PIER Studyyear 1. Am J Ophthalmol 2008;145(2):239–248.

4. Schmidt-Erfurth U, Eldem B, Guymer R, et al; EXCITEStudy Group. Efficacy and safety of monthly versus quarterlyranibizumab treatment in neovascular age-related maculardegeneration: The EXCITE Study. Ophthalmology 2011;118(5):831–839.

5. Holz FG, Amoaku W, Donate J, et al; SUSTAIN Study Group.Safety and efficacy of a flexible dosing regimen of ranibizumabin neovascular age-related macular degeneration: The SUS-TAIN Study. Ophthalmology 2011;118(4):663–671.

6. Shah AR, Del Priore LV. Duration of action of intravitrealranibizumab and bevacizumab in exudative AMD eyes basedon macular volume measurements. Br J Ophthalmol 2009;93(8):1027–1032.

7. Patient Selection. In: Comparison of Age-related Macular Degen-eration Treatments Trials (CATT). CATT: Lucentis-Avastintrial: manual of procedures. May 2010: 3.1-3.4. Available at: http://www.med.upenn.edu/cpob/studies/documents/CATTManualofProceduresMay2010_000.pdf. Accessed: June 27, 2011.

8. Study Drug and Treatment. In: Comparison of Age-related Mac-ular Degeneration Treatments Trials (CATT). CATT: Lucentis-Avastin trial: manual of procedures. May 2010: 4.1-4.10. Availableat: http://www.med.upenn.edu/cpob/studies/documents/CATTManualofProceduresMay2010_000.pdf. Accessed: June 27, 2011.

9. Engelbert M, Zweifel SA, Freund KB. Long-term follow-upfor type 1 (subretinal pigment epithelium) neovasculariza-tion using a modified “treat and extend” dosing regimen ofintravitreal antivascular endothelial growth factor therapy.Retina 2010;30(9):1368–1375.

0. Ton NC, Jayson GC. Resistance to anti-VEGF agents. CurrPharm Des 2004;10(1):51–64.

1. Ferrara N, Kerbel RS. Angiogenesis as a therapeutic target.

OPHTHALMOLOGY MARCH 2012

2

2

22. Abdollahi A, Klatky L, Huber PE. Endostatin: the logic ofantiangiogenic therapy. Drug Resist Updat 2005;8(1-2):59 –74.

23. Van Kempen LC, Leenders WP. Tumors can adapt toanti-angiogenic therapy depending on the stromal context:lessons from endothelial cell biology. Eur J Cell Biol 2006;

FIXED-DOSING VS VARIABLE-DOSIVOL. 153, NO. 3

4. Byun YJ, Lee SJ, Koh HJ. Predictors of response afterintravitreal bevacizumab injection for neovascular age-re-lated macular degeneration. Jpn J Ophthalmol 2010;54(6):571–577.

5. Forooghian F, Cukras C, Meyerle CB, Chew EY, Wong WT.Tachyphylaxis after intravitreal bevacizumab for exudative age-

85(2):61–68. related macular degeneration. Retina 2009;29(6):723–731.

REPORTING VISUAL ACUITIES

The AJO encourages authors to report the visual acuity in the manuscript using the same nomenclature that was used ingathering the data provided they were recorded in one of the methods listed here. This table of equivalent visual acuitiesis provided to the readers as an aid to interpret visual acuity findings in familiar units.

Table of Equivalent Visual Acuity Measurements

Snellen Visual Acuities

Decimal Fraction LogMAR4 Meters 6 Meters 20 Feet

4/40 6/60 20/200 0.10 �1.0

4/32 6/48 20/160 0.125 �0.9

4/25 6/38 20/125 0.16 �0.8

4/20 6/30 20/100 0.20 �0.7

4/16 6/24 20/80 0.25 �0.6

4/12.6 6/20 20/63 0.32 �0.5

4/10 6/15 20/50 0.40 �0.4

4/8 6/12 20/40 0.50 �0.3

4/6.3 6/10 20/32 0.63 �0.2

4/5 6/7.5 20/25 0.80 �0.1

4/4 6/6 20/20 1.00 0.0

4/3.2 6/5 20/16 1.25 �0.1

4/2.5 6/3.75 20/12.5 1.60 �0.2

4/2 6/3 20/10 2.00 �0.3

From Ferris FL III, Kassoff A, Bresnick GH, Bailey I. New visual acuity charts for clinical research. Am J Ophthalmol 1982;94:91–96.

NG BEVACIZUMAB FOR AMD 489

Biosketch

Georges M El-Mollayess, MD, is currently a medical retina fellow in the Department of Ophthalmology at the AmericanUniversity of Beirut in Beirut, Lebanon. He is interested in research regarding the use of anti-vascular endothelial growthfactor molecules for the treatment of neovascular age-related macular degeneration and has been involved in studies whichwere published on this subject. Dr. El-Mollayes is looking forward to pursuing a surgical vitreoretinal fellowship in the nearfuture followed by a career in academic medicine and research.

FIXED-DOSING VS VARIABLE-DOSING BEVACIZUMAB FOR AMDVOL. 153, NO. 3 489.e1