Seven-Year Outcomes in Ranibizumab-Treated

Patients in ANCHOR, MARINA, and

HORIZON
A Multicenter Cohort Study (SEVEN-UP)
Soraya Rofagha, MD, MPH,1 Robert B. Bhisitkul, MD, PhD,1 David S. Boyer, MD,2 SriniVas R. Sadda, MD,3
Kang Zhang, MD, PhD,4 for the SEVEN-UP Study Group*

Purpose: To assess long-term outcomes 7 to 8 years after initiation of intensive ranibizumab therapy in
exudative age-related macular degeneration (AMD) patients.
Design: Multicenter, noninterventional cohort study.
Participants: Sixty-five AMD patients originally treated with ranibizumab in the phase 3 Anti-VEGF Antibody
for the Treatment of Predominantly Classic Choroidal Neovascularization in AMD (ANCHOR) trial, Minimally
Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular AMD (MARINA) trial,
and Open-Label Extension Trial of Ranibizumab for Choroidal Neovascularization Secondary to Age-Related
Macular Degeneration (HORIZON).
Methods: Fourteen clinical trial sites recruited their original subjects for a return evaluation. Individual subject
comparisons were obtained from the ANCHOR, MARINA, and HORIZON databases.
Main Outcome Measures: The primary end point was percentage with best-corrected visual acuity (BCVA)
of 20/70 or better; secondary outcomes included mean change in letter score compared with previous time points
and anatomic results on fluorescein angiography, spectral-domain ocular coherence tomography (OCT), and
fundus autofluorescence (FAF).
Results: At a mean of 7.3 years (range, 6.3e8.5 years) after entry into ANCHOR or MARINA, 37% of study
eyes met the primary end point of 20/70 or better BCVA, with 23% achieving a BCVA of 20/40 or better. Thirty-
seven percent of study eyes had BCVA of 20/200 or worse. Forty-three percent of study eyes had a stable or
improved letter score (
0-letter gain) compared with ANCHOR or MARINA baseline measurements, whereas
34% declined by 15 letters or more, with overall a mean decline of 8.6 letters (P<0.005). Since exit from the
HORIZON study, study eyes had received a mean of 6.8 anti-vascular endothelial growth factor (VEGF) injections
during the mean 3.4-year interval; a subgroup of patients who received 11 or more anti-VEGF injections had
a significantly better mean gain in letter score since HORIZON exit (P<0.05). Active exudative disease was
detected by spectral-domain OCT in 68% of study eyes, and 46% were receiving ongoing ocular anti-VEGF
treatments. Macular atrophy was detected by FAF in 98% of eyes, with a mean area of 9.4 mm2; the area of
atrophy correlated significantly with poor visual outcome (P<0.0001).
Conclusions: Approximately 7 years after ranibizumab therapy in the ANCHOR or MARINA trials, one third of
patients demonstrated good visual outcomes, whereas another third had poor outcomes. Compared with
baseline, almost half of eyes were stable, whereas one third declined by 15 letters or more. Even at this late stage
in the therapeutic course, exudative AMD patients remain at risk for substantial visual decline.
Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references.
Ophthalmology 2013;120:2292-2299 ª 2013 by the American Academy of Ophthalmology.

*Group members listed online in Appendix 1 (available at http://aaojournal.org).

Exudative age-related macular degeneration (AMD) is
a chronic, progressive disease.1e4 Ocular anti-vascular
endothelial growth factor (VEGF) therapy, the current
standard for exudative AMD, inhibits the end results of the
diseasedchoroidal neovascularization (CNV) and vascular
leakagedbut does not seem to act on the causative mech-
anisms underlying VEGF overexpression and the onset and
perpetuation of active disease. It is unclear whether ocular
anti-VEGF therapy leads to a cure and how many patients
require long-term treatment for recurrent exudation.
Ocular anti-VEGF therapy has been in wide clinical use
for many years. Ranibizumab (Lucentis; Genentech, Inc.,
South San Francisco, CA) was approved in the United
States in June 2006. Even earlier than that, bevacizumab

2292  2013 by the American Academy of Ophthalmology ISSN 0161-6420/13/$ - see front matter
Published by Elsevier Inc. http://dx.doi.org/10.1016/j.ophtha.2013.03.046
 Rofagha et al 
7-Year Outcomes in ANCHOR, MARINA, and HORIZON
(Avastin; Genentech, Inc.) was described as an off-label
ocular therapy5 (Rosenfeld PJ. Avastin for AMD. Paper
presented at: AAO Annual Meeting Retina Subspecialty
Day, October 14, 2005; Chicago, IL). At this point, many
AMD patients are at 7 years or more in their treatment
course with anti-VEGF agents, yet scant evidence is avail-
able on long-term outcomes for treatment of this chronic
disease. The pivotal phase 3 trials of monthly ranibizumab
for exudative AMD, the Anti-VEGF Antibody for the
Treatment of Predominantly Classic Choroidal Neo-
vascularization in AMD (ANCHOR) trial and the Mini-
mally Classic/Occult Trial of the Anti-VEGF Antibody
Ranibizumab in the Treatment of Neovascular AMD
(MARINA) trial,6e8 had primary end points at 1 year with
study termination at 2 years. Similarly, for other major
randomized controlled trials of ocular anti-VEGF therapy
for exudative AMD, the results are limited to 1 to 2
years.9e12 The exception is the phase 3b Open-Label
Extension Trial of Ranibizumab for Choroidal Neo-
vascularization Secondary to Age-Related Macular Degen-
eration (HORIZON)13 that assessed ranibizumab treatment
for 2 additional years among patients who previously had
completed 2 years within the ANCHOR, MARINA, or
RhuFab V2 Ocular Treatment Combining the Use of
Visudyne to Evaluate Safety (FOCUS) trials.
The participants in the original ANCHOR and MARINA
trials are among the earliest AMD patients to receive rani-
bizumab therapy. Subjects enrolled in the MARINA study
from March 2003 through December 2003; those in the
ANCHOR study enrolled from June 2003 through
September 2004. Upon completion of the ANCHOR or
MARINA trials, subjects were eligible to enter the
HORIZON extension study.13 A unique AMD patient
cohort therefore is available from these study subjects,
with the longest ranibizumab treatment period within
a clinical trial protocol (4 years) and the longest duration
of follow-up available at this time (7e8 years). The
current study, the Seven-Year Observational Update of
Macular Degeneration Patients Post-MARINA/ANCHOR
and HORIZON Trials (SEVEN-UP), is a multicenter,
cross-sectional follow-up of the initial ranibizumab-treated
patients from the ANCHOR or MARINA trials who also
continued ranibizumab treatment in the HORIZON trial.
The aim was to assess long-term outcomes within this
cohort of exudative AMD patients in terms of visual acuity
and disease status 7 to 8 years after initiation of intensive
ranibizumab therapy and to evaluate the risk of late visual
decline over years 4 to 8 since exit from the HORIZON
study.
Methods
Study Design
The SEVEN-UP study (ClinicalTrials.gov identifier, NCT01256827)
is a multicenter, cross-sectional study of the long-term outcomes of
a cohort of patients originally treated with ranibizumab within the
ANCHOR and MARINA trials and subsequently enrolled and treated
with ranibizumab in the HORIZON study. The study was
conducted in accordance with the Health Insurance Portability
and Accountability Act and the tenets of the Declaration of
Helsinki and with approval of respective institutional review
boards of the participating centers. All participating patients
provided informed consent. Each of the 14 clinical sites that
participated in the SEVEN-UP study previously had served as
investigators in the ANCHOR or MARINA trials, or both, as
well as in the HORIZON trial. All study subjects from these
participating centers were eligible and subjects were enrolled by
invitation; each center was provided with a list of its eligible
subjects compiled from the ANCHOR, MARINA, and
HORIZON databases. Patients were not excluded from partici-
pating in the study on the basis of interim treatment and could
have been managed with any regimen or treatment deemed
appropriate by their interim physician, including intravitreal
ranibizumab, bevacizumab or steroid, photodynamic therapy, or
investigational treatments.
Study Cohort
The ANCHOR and MARINA trials were pivotal phase 3, multi-
center, randomized, controlled trials to evaluate the safety and
efficacy of intravitreal ranibizumab for the treatment of exudative
AMD. In the ANCHOR or MARINA protocols, patients were
randomized to receive monthly intravitreal injections of either 0.3
mg or 0.5 mg ranibizumab or sham injection. Protocol designs for
the MARINA and ANCHOR trials have been described previ-
ously.7,8 A total of 233 patients in the ANCHOR trial and 415
patients in the MARINA trial who were randomized to ranibizu-
mab treatment completed their respective studies. Subjects quali-
fied for entry into the HORIZON open-label extension study upon
completion of participation in the ANCHOR, MARINA, or the
phase 1/2 FOCUS studies14; for the current study, FOCUS
participants were not included because all patients received
photodynamic therapy and there was no ranibizumab-only arm in
the FOCUS study. Upon entering the HORIZON trial,13 526
patients who had been in the initial ranibizumab arms of the
ANCHOR (n ¼ 173) or MARINA (n ¼ 353) trials were
assigned to additional treatment with ranibizumab (0.5 mg) as
needed; 357 of these patients completed 2 years of participation
in the HORIZON study. In the HORIZON study, visits were
required at least quarterly, with retreatment at physician
discretion based on visual acuity alterations, biomicroscopic
signs of active exudation, and findings on ocular coherence
tomography (OCT). The 14 SEVEN-UP clinical trial sites, repre-
senting 155 of the total 357 subjects meeting the above criteria,
recruited their original subjects to return for a single reassessment
visit, and assessments were conducted from November 2010
through November 2011.
Outcome Measures
The primary outcome of the study was the percentage of patients
with Snellen equivalent best-corrected visual acuity (BCVA) of 20/
70 or better in the study eye, as measured with standardized Early
Treatment Diabetic Retinopathy Study (ETDRS) vision testing.
Secondary outcome measures included the percentage of eyes with
good vision (20/40 or better) and poor vision (20/200 or worse),
the percentage of eyes with disease quiescence, and the percentage
of eyes with central macular atrophy. Change in ETDRS letter
score was calculated compared with previous time points for each
subject. The ETDRS vision at baseline, upon exiting the MARINA
or ANCHOR trials, and at month 24 in the HORIZON study were
provided from original databases. All data were collected for both
original study eyes and fellow eyes.
2293
 Ophthalmology Volume 120, Number 11, November 2013
Examination Procedures, Ocular Imaging, and
Laboratory Studies
Patients were evaluated during a single study visit to obtain
ETDRS BCVA with a certified examiner using back-illuminated
charts and standard protocols15 starting at 4 m. Imaging studies
were performed with spectral-domain OCT, fundus photography
(FP), fluorescein angiography (FA), and fundus autofluorescence
(FAF). Full ophthalmologic examination including slit-lamp
examination, fundus biomicroscopy, and indirect ophthalmoscopy
was conducted. No intervention or treatments were performed in
the context of the SEVEN-UP study.
Retrospective Analysis
Chart review of the previous 6 months was performed for evidence
of CNV activity as determined by the interval physician. Retro-
spective review of the interval chart record from exit of the
HORIZON study was performed to determine the number of
interval treatments with anti-VEGF agents (ranibizumab, bev-
acizumab) or other AMD treatments (photodynamic therapy,
corticosteroid, thermal laser, or other investigational treatment).
Because interval care often was continued within the original study
site, most patients had interval charts available for review. If not,
investigators were instructed to request the interval treatment
history. We were able to collect interval history for 63 of 65
patients (97%).
Image Analysis
Images (OCT, FP, FA, FAF) were analyzed by the Doheny
Imaging Reading Center, University of Southern California School
of Medicine, Los Angeles, California, using standardized and
validated metrics. Main anatomic outcomes for the purpose of this
study included presence of intraretinal or subretinal fluid on OCT,
leakage and lesion size on FA, presence and area of decreased
autofluorescence on FAF, and presence of subretinal fibrosis or
hemorrhage on FP.
Statistical Methods
Data were collected on case report forms and then were entered
manually into an encrypted online database (REDCap Software,
Vanderbilt University). Deidentified data then were exported for
statistical analysis with Stata10 (StataCorp, College Station, TX).
Two sample tests of proportions and means were used to compare
ocular and demographic characteristics between enrolled SEVEN-
UP patients and original cohorts. Paired-samples t tests were used
to compare mean ETDRS visual acuities from the SEVEN-UP
study compared with previous time points. The chi-square and
Student t tests were used to explore differences in proportions and
means among categorical and continuous variables, respectively.
One-way analysis of variance testing was used to compare means
in categorical variables with more than 2 groups, using Wilcoxon
scores (Kruskal-Wallis) for nonparametric data when applicable.
Results
Study Population and Baseline Characteristics
The study cohort was defined as the total 357 subjects who
initially were treated with ranibizumab for 24 months in the
MARINA or ANCHOR trial and who then also completed 24
months in the ranibizumab arm of the HORIZON study. Of this
cohort, 155 subjects were eligible from the 14 sites participating
in the SEVEN-UP study, 65 of whom were enrolled in the
2294
Table 2. Distribution of Visual Status (Snellen Equivalent) of
Study Eyes in SEVEN-UP Subjects
Visual Acuity Threshold SEVEN-UP
(Snellen Equivalent) (n [ 65), n (%)

20/40 15 (23.1)

20/70* 24 (36.9)

20/100 30 (46.2)
<20/100 35 (54.8)
20/200 24 (36.9)
SEVEN-UP ¼ Seven-Year Observational Update of Macular Degeneration
Patients Post-MARINA/ANCHOR and HORIZON Trials.
*Primary outcome.
current study (42% enrollment). Study design, eligibility, and
enrollment are shown in Figure 1 (available at http://
aaojournal.org). Reasons for nonenrollment included death,
illness, loss to follow-up, and patient decision not to partici-
pate. Baseline demographic and ocular characteristics at entry to
the MARINA and ANCHOR studies, comparing enrolled
patients in the SEVEN-UP study with the overall study cohort,
are shown in Table 1 (available at http://aaojournal.org). No
significant difference was found between the SEVEN-UP
sample and the total cohort from which it was drawn in terms
of age, sex, initial ranibizumab dose assignment, and eye later-
ality. Mean baseline BCVA was approximately 20/80 for the
ANCHOR/MARINA treated pool, the HORIZON month 24
cohort, and the SEVEN-UP study sample (mean baseline
ETDRS letter score, 52.6, 53.2, and 54.3, respectively). Distri-
bution of CNV lesion subtypes for the SEVEN-UP sample at
baseline was 31% predominantly classic, 25% minimally classic,
and 45% occult with no classic, which closely reflected the
overall cohort at baseline that was correspondingly 29%, 27%,
and 44%. Total area of CNV leakage at baseline in the SEVEN-
UP sample was comparable with that of the overall cohort (3.17
disc areas [DAs] compared with 3.15 DAs), as was total area of
lesion (2.75 DAs compared with 3.00 DAs). The SEVEN-UP
study did have a higher proportion of nonwhite patients
included in the sample compared with the overall cohort.
Follow-up Time
On average, the SEVEN-UP visit occurred 7.3 years after the
subject’s initial entry into the ANCHOR or MARINA trial (stan-
dard deviation [SD], 0.46; range, 6.33e8.49) and 3.4 years after
the subject completed HORIZON month 24 (SD, 0.45; range,
2.46e4.50).
Vision Outcomes
Twenty-four of 65 study eyes (37%) achieved the BCVA threshold
of 20/70 or better, which was defined a priori as the primary
outcome of the study. Twenty-three percent had good vision,
defined as BCVA of 20/40 or better, whereas 37% were legally
blind in the study eye, defined as BCVA 20/200 or worse
(Table 2). Mean change in ETDRS letter score was calculated
compared with earlier time points for individual study eyes,
shown in Figure 2. Since baseline at entry into the ANCHOR or
MARINA trials approximately 7 years previously, there was
a mean loss of 8.6 letters (standard error of mean, 2.8; P<0.005,
paired t test; 95% confidence interval [CI], 2.9e14.2 letters lost).
From the therapeutic peak upon completion of 24 monthly
injections in the ANCHOR or MARINA trials, mean vision had
declined by 19.8 letters (standard error of mean, 2.64;
 Rofagha et al 
7-Year Outcomes in ANCHOR, MARINA, and HORIZON

Figure 2. Graph showing the mean change in Early Treatment Diabetic Retinopathy Study (ETDRS) letter score. Change in ETDRS vision was calculated
for each patient and means are shown at various timepoints in the Treatment of Predominantly Classic Choroidal Neovascularization in AMD Trial
(ANCHOR), the Minimally Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular AMD (MARINA), and the
Open-Label Extension Trial of Ranibizumab for Choroidal Neovascularization Secondary to Age-Related Macular Degeneration (HORIZON): (1) entry
into ANCHOR or MARINA, (2) exit from ANCHOR or MARINA, (3) month 24 HORIZON, and (4) the Seven-Year Observational Update of Macular
Degeneration Patients Post-MARINA/ANCHOR and HORIZON Trials (SEVEN-UP). The white line presents 388 ranibizumab-treated patients who
completed month 24 HORIZON and the black line represents the SEVEN-UP sample.

P<0.0001; 95% CI, 14.5e25.1 letters lost). Since the HORIZON
month 24, the subset of 50 patients with data available at the
HORIZON month 24 time point lost a mean of 6.9 letters
(standard error of mean, 1.9; P<0.001; 95% CI, 2.9e10.7 letters
lost). Figure 2 displays the mean vision change over time for the
SEVEN-UP subjects compared with published data for all
subjects initially treated with ranibizumab in the HORIZON study
who had been assigned to a total of 4 years of ranibizumab on trial,
again confirming that our sample tracks closely with the overall
population from which it was drawn.
In another analysis, each study eye was compared with its
individual result within the previous clinical trials (Table 3,
available at http://aaojournal.org; Fig 3). Forty-three percent of
study eyes had a stable or improved ETDRS letter score (
0-letter
gain) at the SEVEN-UP visit compared with ANCHOR or
MARINA readings at baseline; 12% improved by 15 letters or more,
whereas 34% declined by 15 letters or more. Compared with visual
acuity readings obtained when exiting the ANCHOR or MARINA
trials after completing 24 monthly ranibizumab injections, 15% of
eyes had stable or improved vision. Thirty-two percent of eyes had
stable or improved vision during the SEVEN-UP study compared
with the month 24 HORIZON measurement, with 48% gaining 5
letters or more; however, only 2% had gained 15 letters or more in
that period, and 22% had lost 15 letters or more.

Figure 3. Stack diagram showing distribution of change in Early Treatment Diabetic Retinopathy Study (ETDRS) letter score. The diagram shows the
distribution of visual acuity (VA) change from previous time points in the Treatment of Predominantly Classic Choroidal Neovascularization in AMD Trial
(ANCHOR), the Minimally Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular AMD (MARINA), and the
Open-Label Extension Trial of Ranibizumab for Choroidal Neovascularization Secondary to Age-Related Macular Degeneration (HORIZON): entry in
ANCHOR or MARINA (left), exit from ANCHOR or MARINA (middle), and month 24 HORIZON (right). SEVEN-UP ¼ the Seven-Year Obser-
vational Update of Macular Degeneration Patients Post-MARINA/ANCHOR and HORIZON Trials.

2295
 Ophthalmology Volume 120, Number 11, November 2013

Treatment Status Table 5. Anatomic Outcomes of SEVEN-UP Study Eyes by

Fundus Photography, Fluorescein Angiography, Ocular Coherence
Total number of anti-VEGF treatments was reviewed for each Tomography, and Fundus Autofluorescence
study eye since exit from the HORIZON study; charts were
available for review in 62 of the 65 patients. A total of 641 anti- SEVEN-UP
VEGF injections (147 bevacizumab and 494 ranibizumab) were
administered in the interval between HORIZON exit and the Fundus photography (n ¼ 57), n (%)
SEVEN-UP evaluation, representing a mean of 6.8 injections per Macular subretinal or intraretinal heme 14 (24.1)
eye (range, 0e46 injections) during the mean 3.4-year period. Involving the foveal center 3 (5.1)
Macular subretinal fibrosis 35 (61.4)
Forty-one percent of study eyes received 0 injections, 18%
Involving the foveal center 22 (38.6)
received 1 to 5 injections, 18% received 6 to 10 injections, and
Fluorescein angiography (n ¼ 56)
23% received 11 or more injections. The mean ETDRS letter Presence of CNV leakage, n (%)
change at the SEVEN-UP evaluation study compared with the Definite 11 (19.6)
HORIZON month 24 measurement by quartiles of injections is Questionable 16 (28.6)
shown in Table 4 (available at http://aaojournal.org). In Total area of leaking CNV* (SD), mm2 3.2 (3.0)
a univariate analysis, patients who received 11 injections or more Total area of lesion and nonlesion 19.5 (11.6)
were more likely to gain vision, averaging a gain of 3.9 letters componentsy (SD), mm2
since the HORIZON month 24 measurement (P<0.05, Kruskal- Ocular coherence tomography
Wallis test). (n ¼ 56), n (%)z
To determine the rate of continuing anti-VEGF therapy, IRF or SRF 38 (67.8)
investigators were queried about current and recent treatments for Questionable IRF without SRF 2 (3.6)
each patient. Fifteen of 65 study eyes (23%) were indicated for Mean foveal thicknessx (SD), mm 264.6 (93.0)
anti-VEGF treatment at the SEVEN-UP visit. Based on clinical
250 24 (42.9)
assessment, 18 patients (28%) had active exudation at the SEVEN-
300 14 (21.5)
UP visit, 8 of whom were deemed as being indicated for anti-
350 11 (19.6)
VEGF treatment at the visit. The mean BCVA of these 8 Retinal thickness (neurosensory only)k
(SD), mm
patients who were indicated for treatment was 55.9 letters (SD,
Central subfield 170.0 (73.9)
23.5 letters) compared with 21.1 letters (SD, 25.9 letters) for the 10 Foveal central point 107.0 (89.4)
patients who had active exudation but no indicated treatment at the Total retinal thickness{ (SD), mm 234.3 (126.4)
time of the SEVEN-UP visit (P<0.005). This marked difference Fundus autofluorescence (n ¼ 58)
suggests that physicians may have deemed treatment futile for Macular atrophy, n (%) 57 (98.2)
patients with active exudation and poor vision at long-term follow- Involving the foveal center 52 (89.7)
up. Within the 6 months before and including the SEVEN-UP visit, Definite decreased autofluorescence 9.4 (7.4)
30 of 65 study eyes (46%) had undergone anti-VEGF treatment, of (SD), mm2
which 80% had been deemed to have exudation at the SEVEN-UP Median (range) 8.0 (0.17e27.43)
visit or within the same period. That a portion of patients receiving
treatment did not have active exudation in this period suggests that
CNV ¼ choroidal neovascularization; IRF ¼ intraretinal fluid; SD ¼
some patients may have been maintained on a treat-and-extend standard deviation; SEVEN-UP ¼ Seven-Year Observational Update of
protocol in which treatment was administered independent of Macular Degeneration Patients Post-MARINA/ANCHOR and
active exudation.16 HORIZON Trials; SRF ¼ subretinal fluid.
*n ¼ 26 (Excludes 1 patient with missing data).
y
Sum of lesion components (leaking CNV components, staining scar,
Anatomic Outcomes subretinal hemorrhage, blocked fluorescence, serous pigment epithelial
Sixty of the 65 SEVEN-UP patients had imaging performed (any detachment) plus retinal pigment epithelium (RPE) disturbance and
atrophy; n ¼ 56.
combination of FA, FP, spectral-domain OCT, or FAF). Fifty- z
Measured by Heidelberg Spectralis (n ¼ 26; 46.4%) or Cirrus Zeiss
seven underwent FA; CNV characteristics and lesion size as (n ¼ 30; 53.6%).
measured by FA are shown in Table 5. Twenty-seven study eyes x
Mean foveal thickness defined as sum of central subfields of retinal
(48%) had had either definite leakage from occult or classic CNV thickness, subretinal fluid, and subretinal tissue.
(n ¼ 11) or questionable leakage (n ¼ 16). Of these, 26 had k
Excludes subretinal or sub-RPE fluid and tissue.
{
measurable areas of CNV leakage on FA, with a mean area of CNV Sum of retinal thickness, subretinal fluid, subretinal tissue, and sub-RPE
leakage of 3.2 mm2 (SD, 3.0 mm2). The total lesion (sum of areas space at foveal center point.
of leaking CNV components, staining scar, subretinal hemorrhage,
blocked fluorescence, serous pigment epithelial detachment) plus
nonlesion components (sum of areas of retinal pigment epithelium letter score of 45 (95% CI, 37.0e53.3 letters) compared with 50
[RPE] disturbance and atrophy) measured 19.5 mm2 (SD, 11.6 letters (95% CI, 37.3e62.7 letters) for eyes without exudation,
mm2). Best-corrected visual acuity at SEVEN-UP was not asso- representing no statistically significant difference in visual outcome
ciated with evidence of CNV (P<0.8), presence of leakage on a univariate analysis (P<0.30). Three study eyes had OCT
(P<0.4), or total area of CNV leakage (P<0.70). The BCVA at the images that were not adequate for quantitative measurements. The
SEVEN-UP visit was statistically significantly associated with the mean foveal thickness (sum of retinal thickness, subretinal fluid,
total area of lesion and nonlesion components (P<0.001). and subretinal tissue in the mean central subfield) was 264.6 mm
Retinal anatomic results as measured by OCT and FAF are (SD, 93.0 mm). Analyzing the neurosensory retina only (excluding
shown in Table 5. Spectral-domain OCT was performed in 56 subretinal or sub-RPE fluid and tissue), the mean retinal thickness
study eyes (Spectralis Heidelberg Engineering, n ¼ 26; Cirrus, in the central subfield was 170.0 mm (SD, 73.9 mm), and the mean
Carl Zeiss Meditec, n ¼ 30), and the results showed evidence of retinal thickness at the foveal center point (mean central point
exudation (intraretinal or subretinal fluid) in 68%. Study eyes with thickness) was 107.0 mm (SD, 89.4 mm). Total retinal thickness at
evidence of exudation on spectral-domain OCT had a mean BCVA the foveal center point (sum of retinal thickness, subretinal fluid,

2296
 Rofagha et al 
7-Year Outcomes in ANCHOR, MARINA, and HORIZON
subretinal tissue, and sub-RPE space at the fovea) was 234.3 mm
(SD, 126.4 mm). The BCVA at the SEVEN-UP visit was not
statistically significantly associated with mean foveal thickness
(P<0.15), mean retinal thickness in the central subfield (P<0.45),
mean central point thickness (P<0.95), or total retinal thickness at
the foveal center point (P<0.80).
Fifty-seven of 58 study eyes (98%) undergoing FAF were
detected to have macular atrophy as indicated by definite decreased
autofluorescence, with the remaining eye noted as questionable for
evidence of decreased autofluorescence. In 90% of study eyes
(52 of 58 eyes), the decreased autofluorescence involved the fovea.
The average area of definite decreased autofluorescence measured
9.4 mm2 (SD, 7.4 mm2; median, 8.0 mm2; range, 0.17e27.43
mm2). Decreased BCVA at the SEVEN-UP visit was associated
with both increased area of macular atrophy (P<0.0001) and the
presence of subfoveal macular atrophy (P<0.01). For every 1 mm2
increase in macular atrophy, there was a 2.3-letter decrease in
BCVA (95% CI, 1.6e2.9; t statistic, e6.82; P<0.0001).
Fellow Eyes
Thirty-three patients (51%) were diagnosed with bilateral exuda-
tive AMD at the SEVEN-UP visit compared with 30 patients
(46%) who had bilateral disease at the start of the ANCHOR or
MARINA trials. Fifteen patients (23%) were determined to have
active exudation in the fellow eye at the SEVEN-UP visit, and an
additional 8 patients (12%) had active exudation in the fellow eye
at a previous point within the last 6 months. Eight patients (12%)
had active exudation in both eyes at the time of the SEVEN-UP
visit, with an additional 9 patients (13.8%) having active exuda-
tion in both eyes within the previous 6 months. Therefore, more
than one quarter of the patients (26%) had ongoing bilateral CNV
activity. Four patients (6%) were legally blind, with bilateral 20/
200 vision or worse. For fellow eyes, anti-VEGF treatment was
deemed to indicate at the time of the SEVEN-UP evaluation in 15
of 65 (20%), and 16 of 65 fellow eyes (25%) had been treated with
an anti-VEGF injection within the last 6 months. Four patients
(6%) were receiving ongoing anti-VEGF therapy in both eyes.
Discussion
In this cross-sectional cohort study, 65 patients with exuda-
tive AMD were recalled for evaluation at 7 to 8 years after
they had been enrolled in the ANCHOR or MARINA trials.
This group had received 2 years of monthly ranibizumab
treatment, followed by an additional 2 years of as-needed
ranibizumab treatment in the HORIZON protocol. This
cohort provides an opportunity to study the longest period of
ranibizumab therapy with the greatest duration of follow-up
for this chronic disease. The findings indicate that even this
late in the therapeutic course the risk of visual decline and
disease activity persists, and the need for anti-VEGF treat-
ment continues in a substantial portion of patients.
At a mean 7.3 years after initiation of intensive ranibi-
zumab treatment, approximately one third of patients had
visual acuity of 20/70 or better Snellen equivalent, and
almost one quarter had good vision of 20/40 or better.
However, a third had poor vision of 20/200 or worse. In half
of patients, fellow nonstudy eyes also had wet AMD, and
6% of patients were legally blind, with 20/200 vision or
worse in both eyes. Since exit from the HORIZON study at
a mean of 3.4 years previously, approximately one third of
eyes had stable or improved vision (
0-letter gain), but half
had declined by 5 letters or more. Compared with baseline
vision 7 to 8 years earlier at entry into the original
ANCHOR or MARINA trials, approximately half of the
patients had stable or improved vision (
0-letter gain), but
overall there was a mean decline of 8.2 letters over the
period. The differential outcomes, with a group of patients
achieving good visual acuities over the long term and
another continuing to decline over time, confirm the clinical
impression that, for unknown reasons, neovascular AMD
patients with similar baseline characteristics can exhibit
divergent response patterns to therapy.
On OCT imaging, the macula was found to have intra-
retinal or subretinal fluid in two thirds of subjects, and
nearly half of study eyes were recommended for anti-VEGF
therapy at the time of the study visit, had received injections
within the prior 6 months, or both. Because most patients do
not attain durable, treatment-free cessation of exudative
AMD 7 to 8 years after initiation of intensive ranibizumab
therapy for a full 2 years, close clinical monitoring for
ongoing treatment may need to be extended through many
years of the disease course, perhaps for the rest of the
patients’ lives. From the disease characteristics observed in
this long-duration ranibizumab-treated cohort, in particular
the absence of fibrotic lesions in almost 40% of study eyes,
it is plausible that anti-VEGF therapy alters the natural
course of exudative AMD by circumventing the quiescent or
burnt-out disease end point brought on by retinal destruction
and disciform scar formation. However, doing so may
prolong the active phase of the disease because ongoing
VEGF overproduction is enabled by hypoxic or inflamma-
tory stimuli from therapeutically preserved outer retina and
RPE.
The overall decline in mean ETDRS letter score over the
3.4-year interval since exit from the HORIZON study,
continuing that seen in the course of the HORIZON study
itself, may reflect the inexorable nature of this disease even
in the face of treatment, but other exigent factors also may
be relevant. The retrospective analysis of treatment history
during the interval between exit from the HORIZON study
and the SEVEN-UP evaluation revealed a mean 1.6 anti-
VEGF injections per study eye per year. Such low treat-
ment frequencies may reflect the contemporaneous
management during those years and may have contributed
to the decline in mean visual acuity. More recent evidence
for using anti-VEGF therapy on an as-needed basis has
demonstrated favorable outcomes with a much higher
frequency of use, in the range of 6.9 (ranibizumab) to 7.7
(bevacizumab) injections per year in the initial years of
exudation.9 In the SEVEN-UP study, there were better
visual outcomes among the patients in the highest quartile of
anti-VEGF treatments (
11 injections since HORIZON
exit) who gained a mean 3.9 letters of visual acuity since
exiting the HORIZON study.
Retinal imaging demonstrated central fibrotic scars in
approximately one third of study eyes and suspicion of CNV
leakage in half, but virtually all eyes showed the presence of
macular atrophy. In univariate analyses, anatomic outcomes
associated with poorer visual outcomes included a larger
combined area of total lesion and nonlesion components,
a greater area of decreased autofluorescence, and a subfoveal
2297
 Ophthalmology Volume 120, Number 11, November 2013
location of decreased autofluorescence. The factor corre-
lating most strongly with poor visual outcome was increased
area of macular atrophy as detected by FAF. Anatomic
outcomes with no significant effect on visual outcome
included presence of subretinal fibrosis, subfoveal fibrosis,
CNV area and leakage, and OCT findings of active exuda-
tion and measures of retinal thickness. Because OCT is
invaluable in clinical practice in the management of
exudative AMD, it is interesting that current OCT findings at
the SEVEN-UP visit had no association with the vision
outcome among these long-term patients. The relationship
and interaction of these various anatomic outcomes in
determining visual acuity is complex and beyond the scope
of this descriptive report.
The study has several limitations. Although the main
outcome measures were predetermined and the data to
evaluate current clinical status were gathered prospectively,
there was also a retrospective component for interval treat-
ment history and disease activity based on record review and
physician report, with the attendant liabilities of retrospec-
tive analyses. There is no control group because not only
were the original ANCHOR and MARINA sham control
subjects allowed to cross over to a ranibizumab treatment
arm upon entry into the HORIZON study, but maintaining
an untreated control group through the current era of
effective AMD treatments is no longer viable. Instead, this
study relied on comparisons of individual patient’s current
visual results with their respective clinical trial data at earlier
time points to understand disease progression and long-term
visual changes. Another limitation was the recruitment of
the study cohort: the 14 clinical investigators were able to
return only a portion of their eligible subjects for the current
study, so a selection bias may have resulted. Although we
found the enrolled subjects to be comparable with the cohort
population from which they were drawn, our sampling
method was not random. The lack of a uniform treatment
protocol for the cohort in the 3 to 4 years since completion
of the HORIZON study is another drawback; patients were
managed at the discretion of their physicians in their time
off trial, so the SEVEN-UP subjects represent heteroge-
neous interval treatment histories, of which we did not
attempt a complete enumeration or analysis.
The dramatic benefits of ranibizumab therapy were
established in the ANCHOR and MARINA trials, in part with
the same individuals evaluated here. The results of the
SEVEN-UP study point to divergent long-term outcomes
among these AMD patients, with a significant portion at risk
for ongoing vision loss many years after treatment initiation.
Exudative AMD is a chronic disease, and long-term outcomes
cannot be extrapolated from the 2-year results available from
the ANCHOR and MARINA trials. Continuation of such
large-scale clinical trials for many subsequent years is likely
prohibitive for study sponsors, but revisiting patient cohorts
from pivotal clinical trials at later intervals offers a feasible
way to update the clinical course of treated AMD. This
approach has a relatively low cost, leverages high-level data
collected in previous randomized trials, and may be appli-
cable to other chronic diseases where data over an extended
duration can help to characterize long-term clinical trajecto-
ries and guide ongoing management strategies.
2298
Acknowledgments. The authors thank Travis Porco, PhD,
UCSF Proctor Foundation, for statistical consulting, and Suzanne
Staveley, Envision Scientific Solutions, for graphics assistance.
References
1. Macular Photocoagulation Study Group. Laser photocoagula-
tion of subfoveal neovascular lesions of age-related macular
degeneration. Updated findings from two clinical trials. Arch
Ophthalmol 1993;111:1200–9.
2. Wong TY, Chakravarthy U, Klein R, et al. The natural history
and prognosis of neovascular age-related macular degenera-
tion: a systematic review of the literature and meta-analysis.
Ophthalmology 2008;115:116–26.
3. Pauleikhoff D. Neovascular age-related macular degeneration:
natural history and treatment outcomes. Retina 2005;25:
1065–84.
4. Shah AR, Del Priore LV. Natural history of predominantly
classic, minimally classic, and occult subgroups in exudative
age-related macular degeneration. Ophthalmology 2009;116:
1901–7.
5. Rosenfeld PJ, Moshfeghi AA, Puliafito CA. Optical coherence
tomography findings after an intravitreal injection of bev-
acizumab (Avastin) for neovascular age-related macular
degeneration. Ophthalmic Surg Lasers Imaging 2005;36:
331–5.
6. Brown DM, Michels M, Kaiser PK, et al. Ranibizumab versus
verteporfin photodynamic therapy for neovascular age-related
macular degeneration: two-year results of the ANCHOR
Study. Ophthalmology 2009;116:57–65.
7. Rosenfeld PJ, Brown DM, Heier JS, et al, MARINA Study
Group. Ranibizumab for neovascular age-related macular
degeneration. N Engl J Med 2006;355:1419–31.
8. Brown DM, Kaiser PK, Michels M, et al, ANCHOR Study
Group. Ranibizumab versus verteporfin for neovascular age-
related macular degeneration. N Engl J Med 2006;355:
1432–44.
9. CATT Research Group. Ranibizumab and bevacizumab for
neovascular age-related macular degeneration. N Engl J Med
2011;364:1897–908.
10. Comparison of Age-related Macular Degeneration Treatments
Trials (CATT) Research Group, Martin DF, Maguire MG,
Fine SL, et al. Ranibizumab and bevacizumab for treatment of
neovascular age-related macular degeneration: two-year
results. Ophthalmology 2012;119:1388–98.
11. Heier JS, Brown DM, Chong V, et al, VIEW 1 and VIEW 2
Study Groups. Intravitreal aflibercept (VEGF Trap-Eye) in wet
age-related macular degeneration. Ophthalmology 2012;119:
2537–48.
12. IVAN Study Investigators, Chakravarthy U, Harding SP,
Rogers CA, et al. Ranibizumab versus bevacizumab to treat
neovascular age-related macular degeneration: one-year find-
ings from the IVAN randomized trial. Ophthalmology
2012;119:1399–411.
13. Singer MA, Awh CC, Sadda S, et al. HORIZON: an open-
label extension trial of ranibizumab for choroidal neo-
vascularization secondary to age-related macular degeneration.
Ophthalmology 2012;119:1175–83.
14. Antoszyk AN, Tuomi L, Chung CY, Singh A; FOCUS Study
Group. Ranibizumab combined with verteporfin photodynamic
therapy in neovascular age-related macular degeneration
(FOCUS): year 2 results. Am J Ophthalmol 2008;145:
862–74.
 Rofagha et al 
7-Year Outcomes in ANCHOR, MARINA, and HORIZON

15. Ferris FL III, Bailey I. Standardizing the measurement of 16. Gupta OP, Shienbaum G, Patel AH, et al. A treat and extend
visual acuity for clinical research studies: guidelines from the regimen using ranibizumab for neovascular age-related
Eye Care Technology Forum. Ophthalmology 1996;103: macular degeneration: clinical and economic impact.
181–2. Ophthalmology 2010;117:2134–40.

Footnotes and Financial Disclosures

Originally received: December 22, 2012. Robert B. Bhisitkul - Consultant - Genentech, Santen, Pfizer; Financial
Final revision: March 28, 2013. support - Genentech, GlaxoSmithKline
Accepted: March 29, 2013. David S. Boyer - Consultant - Alcon, Allergan, Genentech, Regeneron,
Available online: May 6, 2013. Manuscript no. 2012-1924. Neurotech, Novartis/QLT, Pfizer, and iCo; Financial support - Alcon,
1
Department of Ophthalmology, School of Medicine, University of Cal- Allergan, Genentech, and Regeneron SriniVas R. Sadda - Consultant -
ifornia, San Francisco, San Francisco, California. Genentech, Allergan, Heidelberg Engineering; Financial support - Gen-
2 entech, Carl Zeiss Meditec, Optovue; Patents - Topcon Medical Systems
Retina-Vitreous Associates, Los Angeles, California.
3 Kang Zhang - Consultant - Genentech, Acucela; Financial support e
Doheny Eye Institute, University of Southern California School of Genentech.
Medicine, Los Angeles, California.
4 Supported by Genentech, Inc., South San Francisco, California; Research to
Shiley Eye Center, University of California, San Diego, La Jolla, Prevent Blindness, Inc., New York, New York; the National Eye Institute,
California. National Institutes of Health, Bethesda, Maryland (core grant no.:
A full listing of the SEVEN-UP Study Group is available in Appendix 1 at EY002162); and the That Man May See Foundation, San Francisco, CA.
http//aaojournal.org.
Correspondence:
Presented in part at: American Academy of Ophthalmology Annual
Robert B. Bhisitkul, MD, PhD, UCSF Beckman Vision Center, 10 Koret
Meeting, Chicago, Illinois; November 2012; and Association for Research
Way, K301, San Francisco, CA 94143. E-mail: BhisitkulR@vision.ucsf.
in Vision and Ophthalmology Annual Meeting, Fort Lauderdale, Florida,
edu.
May 2012.
Financial Disclosure(s):
The author(s) have made the following disclosure(s):

2299