Ranibizumab versus Verteporfin
`Photodynamic Therapy for Neovascular
`Age-Related Macular Degeneration:
`Two-Year Results of the ANCHOR Study
`
`David M. Brown, MD,1 Mark Michels, MD,2 Peter K. Kaiser, MD,3 Jeffrey S. Heier, MD,4 Judy P. Sy, PhD,5
`Tsontcho Ianchulev, MD, MPH5
`
`Objective: The 2-year, phase III trial designated Anti-vascular endothelial growth factor (VEGF) Antibody for
`the Treatment of Predominantly Classic Choroidal Neovascularization (CNV) in Age-related Macular Degenera-
`tion (ANCHOR) compared ranibizumab with verteporfin photodynamic therapy (PDT) in treating predominantly
`classic CNV.
`Design: Multicenter, international, randomized, double-masked, active-treatment-controlled clinical trial.
`Participants: Patients with predominantly classic, subfoveal CNV not previously treated with PDT or
`antiangiogenic drugs.
`Intervention: Patients were randomized 1:1:1 to verteporfin PDT plus monthly sham intraocular injection or
`to sham verteporfin PDT plus monthly intravitreal ranibizumab (0.3 mg or 0.5 mg) injection. The need for PDT
`(active or sham) retreatment was evaluated every 3 months using fluorescein angiography (FA).
`Main Outcome Measures: The primary, intent-to-treat efficacy analysis was at 12 months, with continued
`measurements to month 24. Key measures included the percentage losing ⬍15 letters from baseline visual acuity
`(VA) score (month 12 primary efficacy outcome measure), percentage gaining ⱖ15 letters from baseline, and
`mean change over time in VA score and FA-assessed lesion characteristics. Adverse events were monitored.
`Results: Of 423 patients (143 PDT, 140 each in the 2 ranibizumab groups), the majority (ⱖ77% in each
`group) completed the 2-year study. Consistent with results at month 12, at month 24 the VA benefit from
`ranibizumab was statistically significant (P⬍0.0001 vs. PDT) and clinically meaningful: 89.9% to 90.0% of
`ranibizumab-treated patients had lost ⬍15 letters from baseline (vs. 65.7% of PDT patients); 34% to 41.0% had
`gained ⱖ15 letters (vs. 6.3% of PDT group); and, on average, VA was improved from baseline by 8.1 to 10.7
`letters (vs. a mean decline of 9.8 letters in PDT group). Changes in lesion anatomic characteristics on FA also
`favored ranibizumab (all comparisons P⬍0.0001 vs. PDT). Overall, there was no imbalance among groups in
`rates of serious ocular and nonocular adverse events. In the pooled ranibizumab groups, 3 of 277 (1.1%) patients
`developed presumed endophthalmitis in the study eye (rate per injection ⫽ 3/5921 [0.05%]).
`Conclusions:
`In this 2-year study, ranibizumab provided greater clinical benefit than verteporfin PDT in
`patients with age-related macular degeneration with new-onset, predominantly classic CNV. Rates of serious
`adverse events were low.
`Financial Disclosure(s): Proprietary or commercial disclosure may be found after
`Ophthalmology 2009;116:57– 65 © 2009 by the American Academy of Ophthalmology.
`
`the references.
`
`Neovascular age-related macular degeneration (AMD) is
`the process of anomalous pathologic blood vessels arising
`from the choroid and disrupting the anatomy and function of
`the neurosensory retina. Choroidal neovascularization
`(CNV) can be classified on the basis of its appearance on
`fluorescein angiography (FA) as “occult” or “classic.” The
`clinical course of vision loss associated with occult CNV,
`which is usually confined to the space beneath the retinal
`pigment epithelium (RPE), is typically indolent compared
`with “classic” CNV lesions, which often penetrate the RPE
`and grow in the subretinal space.1–3 “Predominantly clas-
`
`sic” CNV are lesions composed of at least 50% classic
`CNV. Before the approval of verteporfin (Visudyne; No-
`vartis Pharmaceuticals Corp., East Hanover, NJ) in 2001,
`predominantly classic CNV typically led to permanent loss
`of the majority of central vision within 3 to 9 months after
`diagnosis.4 The Treatment of Age-Related Macular Degen-
`eration with Photodynamic Therapy (TAP) study demon-
`strated the efficacy and favorable adverse events profile of
`verteporfin photodynamic therapy (PDT) in patients with
`predominantly classic CNV over the natural history of the
`disease, with 59% of patients treated with PDT losing fewer
`
`© 2009 by the American Academy of Ophthalmology
`Published by Elsevier Inc.
`
`ISSN 0161-6420/09/$–see front matter
`doi:10.1016/j.ophtha.2008.10.018
`
`57
`
`Exhibit 2085
`Page 01 of 14
`
`

`

`Ophthalmology Volume 116, Number 1, January 2009
`
`than 15 letters at 2 years (compared with 31% of patients
`treated with placebo).5 On the basis of these findings, PDT
`became the standard of care for patients with this angio-
`graphic subtype of CNV.
`The 2-year, phase III trial designated Anti-vascular
`endothelial growth factor (VEGF) Antibody for the Treat-
`ment of Predominantly Classic Choroidal Neovasculariza-
`tion in Age-related Macular Degeneration (ANCHOR)
`compared the recombinant, humanized anti-VEGF mono-
`clonal antibody antigen-binding fragment (Fab) ranibi-
`zumab (Lucentis; Genentech, Inc., South San Francisco,
`CA) with PDT in patients with predominantly classic, sub-
`foveal CNV secondary to AMD. At 12 months (the pre-
`specified primary efficacy analysis), ranibizumab had supe-
`rior efficacy to PDT as indicated by both visual acuity (VA)
`measures and changes in CNV lesion characteristics.6 The
`percentage of patients who had lost fewer than 15 letters
`from baseline VA (primary efficacy end point) was 94.3%
`and 96.4% in the 0.3-mg and 0.5-mg ranibizumab groups,
`respectively, compared with 64.3% of patients in the PDT
`group. Also, ranibizumab-treated patients, on average, had
`improved VA compared with baseline at month 12, whereas
`VA declined in the PDT group. This was the first demon-
`stration that a therapy could not only prevent further VA
`loss but also provide clinically meaningful improvement of
`VA in a substantial proportion of patients with predomi-
`nantly classic CNV. Serious ocular events associated with
`treatment were uncommon. These first-year results, together
`with positive 2-year results in a similarly designed, sham-
`injection-controlled phase III trial in patients with minimally
`classic or occult with no classic CNV lesions (the Minimally
`Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab
`In the Treatment of Neovascular Age-Related Macular Degen-
`eration (MARINA) study7), led to United States Food and
`Drug Administration approval of ranibizumab for treatment of
`all angiographic subtypes of CNV secondary to AMD in June
`2006. The ANCHOR study is completed, and the 2-year re-
`sults are reported here.
`
`Materials and Methods
`
`The methods for the ANCHOR study have been reported6 and are
`summarized briefly below.
`ANCHOR was a multicenter (83 sites), international, random-
`ized, double-masked, active-treatment-controlled phase III trial
`evaluating the efficacy and adverse events profile of ranibizumab
`in treating predominantly classic subfoveal CNV secondary to
`AMD that, on the basis of FA and fundus photography, was
`confirmed by an independent central reading center (the University
`of Wisconsin Fundus Photograph Reading Center) to be predom-
`inantly classic in composition and suitable for treatment with PDT.
`Predominantly classic lesions were defined as those where the
`classic component made up 50% or more of the total lesion area,
`which could include, in addition to CNV, components such as
`contiguous subretinal hemorrhage, blocked fluorescence not from
`hemorrhage, serous detachment of the RPE, and fibrosis.
`Patients provided written, informed consent for study partici-
`pation. Institutional Review Board, National Competent Authority,
`or Ethics Committee approval was obtained at each participating
`clinical center before the start of the study. All US study sites
`complied with the Health Insurance Portability and Accountability
`
`58
`
`Act of 1996. Patients were excluded if they had permanent struc-
`tural damage to the central fovea or a history of treatment for
`subfoveal neovascular AMD (including any prior PDT) that by its
`nature or timing might compromise valid assessment of the effects
`of the study treatment. There were no exclusion criteria regarding
`preexisting cardiovascular, cerebrovascular, or peripheral vascular
`conditions.
`Only 1 eye per patient (the study eye) received the study
`treatment. Eligible patients were randomized 1:1:1 to either verte-
`porfin PDT plus monthly sham ocular injection or sham vertepor-
`fin PDT plus monthly intravitreal ranibizumab (0.3 mg or 0.5 mg)
`injection. Ranibizumab was injected into the study eye every 30⫾7
`days for a total of 24 injections beginning on day 0; sham injec-
`tions were administered on the same dosing schedule. Patients’
`CNV lesions were evaluated using FA at screening and then every
`3 months to assess the need for additional PDT (active or sham
`intravenous verteporfin injection). The central reading center as-
`sessed all images, but the decision to retreat with PDT (active or
`sham) was based on the evaluating physician’s assessment of CNV
`leakage on the FA images. Active PDT treatment was administered
`according to the Visudyne prescribing information8 (i.e., the phy-
`sician should reevaluate the patient every 3 months, and if CNV
`leakage is detected on FA then standard fluence PDT should be
`repeated). After careful review of the 12-month data, the study
`protocol was amended to allow all patients to receive active
`ranibizumab injections if they had not yet completed their month
`23 visit (the last possible injection visit). Double-masking was
`maintained. Patients in the active PDT/sham ocular injection arm
`who participated in the amendment received monthly injections of
`0.3 mg ranibizumab for the remainder of the trial, whereas patients
`in the ranibizumab groups who participated continued to receive
`ranibizumab according to their original randomization (0.3 or 0.5
`mg). Active or sham PDT was no longer administered to patients
`who participated in the amendment but was continued (if needed)
`per randomization in patients who did not.
`Best-corrected VA measured per the study protocol (i.e., mea-
`sured with Early Treatment Diabetic Retinopathy Study charts at a
`starting distance of 2 m and using a standardized refraction and
`testing protocol) and CNV lesion characteristics (based on FA and
`fundus photography) were assessed at the regularly scheduled
`study visits. Key FA evaluations were the area of classic CNV,
`total lesion area, total area of CNV, and total area of leakage from
`CNV.
`Intraocular pressure measurement (before and 60⫾10 minutes
`after each study treatment) and indirect ophthalmoscopy and slit-
`lamp examination (before each study treatment) were performed.
`The incidence and severity of ocular and nonocular (systemic)
`adverse events and systemic immunoreactivity (i.e., the presence
`of serum antibodies against ranibizumab) were assessed.
`Efficacy end points were evaluated using an intent-to-treat
`analysis for randomized patients on the basis of their original
`treatment assignment. Missing data were imputed using the last-
`observation-carried-forward method and compared for consistency
`with those obtained using observed data. All available data were
`included in analyses of efficacy end points for year 2, including
`those that occurred after ranibizumab treatment initiation in pa-
`tients randomized to PDT who crossed over to ranibizumab as part
`of the protocol amendment.
`The primary efficacy end point was the proportion of patients
`who at 12 months lost fewer than 15 letters (⬃3 lines) from
`baseline VA in the study eye. The proportion of patients who lost
`fewer than 15 letters from baseline at 24 months was a secondary
`efficacy end point. Other prespecified secondary VA end points
`assessed at 12 months and 24 months included the mean change
`from baseline (letters), proportion of patients who gained 15 or
`more letters from baseline, and proportion of patients with a
`
`Exhibit 2085
`Page 02 of 14
`
`

`

`Brown et al
`
`䡠 Ranibizumab vs PDT for Neovascular AMD
`
`Snellen equivalent of 20/200 or worse. Severe VA loss (30 letters
`[⬃6 lines] or more from baseline) was an exploratory efficacy end
`point. Prespecified secondary end points involving characteristics
`of the CNV lesion at months 12 and 24 were mean changes from
`baseline in the area of the classic CNV component and the total
`area of leakage from CNV (including leakage and intense progres-
`sive RPE staining). Mean changes in the area of CNV and the area
`of the entire lesion were exploratory efficacy end points.
`Visual acuity outcomes were compared between each ranibi-
`zumab dose group and the control group with stratification by
`baseline VA score (⬍45 letters vs. ⱖ45 letters). Binary VA end
`points were analyzed using the Cochran chi-square test,9 and the
`mean change from baseline was analyzed using the t test from an
`analysis of variance model. The mean changes from baseline to
`month 24 in the CNV lesion characteristics were compared be-
`tween each ranibizumab dose group and the control group using
`the t test from a stratified, covariate-adjusted analysis of covari-
`ance model, with baseline VA score as the stratification variable
`and baseline value of the corresponding end point as a covariate.
`The percentage of patients with CNV leakage was compared
`between groups at month 24 using the Pearson chi-square test.
`The main analyses comparing adverse events in the treatment
`groups were performed using all data for the entire study period,
`except for patients randomized to PDT who crossed over to ranibi-
`zumab as part of the protocol amendment. For these patients,
`adverse events data collected after their crossover were excluded
`from the main summaries and summarized separately.
`
`Results
`Patient Disposition
`Patient disposition is summarized in Table 1 (available at http://
`aaojournal.org). Of 423 patients enrolled and randomized, 143
`were assigned to active PDT and 140 each were assigned to the 2
`ranibizumab dose levels. Three patients assigned to 0.3 mg ranibi-
`zumab withdrew before starting study treatment, and 1 patient in
`the 0.5-mg group did not have a baseline VA score. The study was
`completed by 110 patients (76.9%) in the PDT group, 117 patients
`(83.6%) in the 0.3-mg ranibizumab group, and 116 patients
`(82.9%) in the 0.5-mg ranibizumab group. Of those patients who
`discontinued early from the study, only 3 patients (2.1%), 1 patient
`(0.7%), and 3 patients (2.1%) from the PDT, 0.3-mg, and 0.5-mg
`groups, respectively, were reported as having been discontinued
`because of “loss to follow-up.” Other reasons for early discontin-
`uation as reported by investigators on the case report form (i.e.,
`death, adverse event, patient’s decision, physician’s decision, pa-
`tient noncompliance, patient’s condition mandated other therapeu-
`tic intervention) were similarly distributed among the treatment
`groups, with the exception of discontinuation because of “patient’s
`decision,” which was more frequent among patients in the PDT
`group (17/143, 11.9%) than in the 0.3-mg (6/140, 4.3%) and
`0.5-mg (8/140, 5.7%) ranibizumab groups.
`
`Baseline Patient Characteristics
`Demographic and baseline characteristics of the patients, summa-
`rized in Table 2 (available at http://aaojournal.org), were well
`balanced among the treatment groups. Although only patients with
`predominantly classic CNV, based on initially expedited assess-
`ment by the central reading center, were to be enrolled, the central
`reading center subsequently categorized a few patients in each
`treatment arm (2 in the PDT group, 6 in the 0.3-mg group, and 5
`in the 0.5-mg group) as having minimally classic or occult with no
`classic CNV lesions; these patients were included in all analyses.
`
`Study Treatment Exposure
`
`The mean number of ranibizumab injections administered during
`the 2-year treatment period was 21.5 in the 0.3-mg group and 21.3
`in the 0.5-mg group. Patients in the PDT group received a mean of
`19.2 sham ocular injections. Including the required administration
`on day 0, active PDT was administered a mean of 3.8 times in the
`PDT group and sham PDT was administered a mean of 2.2 and 1.9
`times in the 0.3-mg and 0.5-mg ranibizumab groups, respectively,
`during the 24-month study period. This calculation for the active
`PDT group includes patients who crossed over to ranibizumab
`(and thus became ineligible for further PDT) as part of the protocol
`amendment. Ranibizumab exposure and treatment results for the
`patients in the PDT group who crossed over are discussed below.
`Starting as early as month 18, 50 of the 143 patients random-
`ized to the PDT group (35%) crossed over to receive monthly
`injections of 0.3 mg ranibizumab for the remainder of the treat-
`ment period. Patients could receive up to 6 ranibizumab injections
`if crossover occurred at month 18 or 1 injection if crossover
`occurred at month 23. The 50 patients who crossed over received
`a mean of 3.3 ranibizumab injections.
`
`Visual Acuity End Points
`
`Visual acuity outcomes results, which include data from those
`patients who crossed over (but analyzed according to their ran-
`domized treatment assignment) are summarized in Table 3. (An
`additional table, Table 4 available at http://aaojournal.org, pro-
`vides a frequency distribution of changes in VA relative to base-
`line in the study eye at month 24.) As previously reported by
`Brown et al,6 the study met its objectives for the primary VA
`efficacy end point and all secondary VA and FA end points at the
`end of the first treatment year (i.e., each of the ranibizumab groups
`was superior to the PDT group for each end point). All second-year
`efficacy objectives (both VA and FA) concerning secondary end
`points were also met (the primary analysis for VA end points was
`at the end of the first year). A statistically significant and clinically
`meaningful effect of ranibizumab on VA was seen in all VA end
`points at month 24. Statistical analyses performed using observed
`data were consistent with the results using the last-observation-
`carried-forward method described above (i.e., P⬍0.0001 for all
`treatment comparisons vs. PDT using either method).
`At month 24, 90.0% of patients in the 0.3-mg ranibizumab
`group and 89.9% of patients in the 0.5-mg ranibizumab group had
`lost ⬍15 letters from baseline VA, compared with 65.7% of
`patients in the PDT group. A gain of 15 or more letters from
`baseline VA was seen in 34.3% of patients in the 0.3-mg ranibi-
`zumab group and 41.0% of patients in the 0.5-mg ranibizumab
`group, compared with 6.3% of patients in the PDT group. The
`mean change in VA over the 24-month treatment period is shown
`in Figure 1. On average, VA had improved from baseline by 8.1
`letters in the 0.3-mg group and 10.7 letters in the 0.5-mg group at
`month 24, compared with a mean decline of 9.8 letters in the PDT
`group. The superior VA benefit of ranibizumab compared with
`PDT was statistically significant as early as month 1.
`At month 24, the percentage of patients with a VA Snellen
`equivalent of 20/200 or worse was significantly higher in the PDT
`group (60.8%) than in the ranibizumab groups (22.9% in the
`0.3-mg group and 20.0% in the 0.5-mg group; P⬍0.0001 vs.
`PDT). Only 1.4% of patients in the 0.3-mg group and none of the
`patients in the 0.5-mg group experienced severe vision loss (loss
`ⱖ30 letters; an exploratory end point) compared with baseline,
`whereas 16.1% of patients in the PDT group had severe vision loss
`at month 24.
`
`59
`
`Exhibit 2085
`Page 03 of 14
`
`

`

`Ophthalmology Volume 116, Number 1, January 2009
`
`Table 3. Key Visual Acuity Outcomes Relative to Baseline in
`the Study Eye at Month 12 and Month 24
`
`Efficacy Outcome
`Lost ⬍15 letters—n (%)*
`Month 12†
`Month 24
`Lost ⱖ30 letters—n (%)*
`Month 12
`Month 24
`Snellen VA 20/200 or
`worse—n (%)
`Baseline*
`Month 12
`Month 24
`Gained ⱖ0 letters—n (%)*
`Month 12
`Month 24
`Gained ⱖ15 letters—n (%)*
`Month 12
`Month 24
`Gained ⱖ30 letters—n (%)*
`Month 12
`Month 24
`Change from baseline
`(letters)*
`Month 12
`Mean (SD)
`Month 24
`Mean (SD)
`
`Verteporfin
`PDT
`(n ⴝ 143)
`
`Ranibizumab
`0.3 mg
`(n ⴝ 140)
`
`Ranibizumab
`0.5 mg
`(n ⴝ 140)
`
`92 (64.3)
`94 (65.7)
`
`132 (94.3)
`126 (90.0)
`
`134 (96.4)
`125 (89.9)
`
`19 (13.3)
`23 (16.1)
`
`0
`2 (1.4)
`
`0
`0
`
`46 (32.2)
`86 (60.1)
`87 (60.8)
`
`35 (25.0)
`31 (22.1)
`32 (22.9)
`
`32 (23.0)
`23 (16.4)
`28 (20.0)
`
`43 (30.1)
`41 (28.7)
`
`104 (74.3)
`109 (77.9)
`
`108 (77.7)
`108 (77.7)
`
`8 (5.6)
`9 (6.3)
`
`0
`3 (2.1)
`
`50 (35.7)
`48 (34.3)
`
`9 (6.4)
`12 (8.6)
`
`56 (40.3)
`57 (41.0)
`
`17 (12.2)
`20 (14.4)
`
`⫺9.5 (16.4)
`
`8.5 (14.6)
`
`11.3 (14.6)
`
`⫺9.8 (17.6)
`
`8.1 (16.2)
`
`10.7 (16.5)
`
`PDT ⫽ photodynamic therapy; SD ⫽ standard deviation; VA ⫽ visual
`acuity.
`NOTE: P⬍0.0001 for all comparisons of each ranibizumab dose group with
`the verteporfin PDT group with the exception of Gained ⱖ30 letters,
`where at month 12 P ⫽ 0.0018 for the 0.3-mg ranibizumab group, and at
`month 24 P ⫽ 0.0132 and P⫽0.0001 for the 0.3-mg and 0.5-mg ranibi-
`zumab groups, respectively.
`*For ranibizumab 0.5-mg group, the number of patients with observations
`is 139.
`†Primary efficacy endpoint.
`
`Angiographic End Points
`Persistent, statistically significant beneficial effects of ranibizumab
`on FA-assessed lesion characteristics were also demonstrated at
`the end of the second study year (Table 5). At month 24, the total
`area of lesion, on average, remained essentially stable in the
`ranibizumab groups, increasing from baseline by 0.52 disc areas
`(DA) and 0.39 DA in the 0.3-mg and 0.5-mg groups, respectively.
`However, in the PDT group, the area increased by 2.89 DA
`(P⬍0.0001 for each dose group vs. PDT). At month 24, the total
`area of CNV, on average, also remained essentially stable in the
`ranibizumab groups, increasing from baseline by 0.33 DA and 0.27
`DA in the 0.3-mg and 0.5-mg groups, respectively. In contrast, in
`the PDT group, the area increased by 1.60 DA (P⬍0.0001 for each
`dose group vs. PDT). At month 24, the mean area of classic CNV
`had decreased from baseline by 0.57 DA and 0.72 DA in the
`0.3-mg and 0.5-mg groups, respectively; in the PDT group, it had
`increased by 0.41 DA (P⬍0.0001, vs. PDT). On average, although
`the area of classic CNV decreased from baseline in the ranibizumab
`groups, the area of occult CNV with no classic component increased
`(by 0.91 DA in the 0.3-mg group and by 0.99 DA in the 0.5-mg
`group), resulting in small mean increases in the total area of CNV.
`
`60
`
`The total area occupied by other lesion components showed small
`mean increases in the ranibizumab groups, reflecting mean increases
`in the area of subretinal fibrous tissue (or fibrin) or disciform scar and
`area of atrophic scar, mean changes in the area of blood that was part
`of the lesion, and negligible mean changes in the area of serous
`pigment epithelial detachment (data not shown). These changes in the
`total area of CNV and the area of other lesion components with no
`CNV account for the small overall mean increase from baseline in the
`total area of the entire neovascular lesion.
`At month 24, the total area of leakage from CNV plus intense
`progressive RPE staining, on average, had decreased from baseline
`by 2.23 DA and 2.37 DA in the 0.3-mg and 0.5-mg groups,
`respectively, and had decreased by 0.78 DA in the verteporfin PDT
`group (P⬍0.0001, vs. PDT). The percentage of patients with
`leakage from CNV plus intense progressive RPE staining declined
`in all 3 treatment groups from month 12 to month 24, but the
`percentage of patients whose lesions were still leaking at month 24
`was significantly smaller
`in the ranibizumab-treated groups
`(P⬍0.0001, vs. PDT). Both the mean and standard deviation for
`the total area of leakage from CNV were identical (to 2 decimal
`places) with those for the total area of leakage from CNV plus
`intense progressive RPE staining, indicating that the mean area of
`intense progressive RPE staining was small.
`
`Patients Who Crossed Over to Ranibizumab
`Demographic and baseline characteristics of patients in the PDT
`group who switched to ranibizumab treatment were comparable to
`those of patients in the PDT group who did not switch to ranibi-
`zumab treatment (Table 2, available at http://aaojournal.org). Ef-
`ficacy outcomes for patients who did and did not cross over are
`summarized in Table 6 (available at http://aaojournal.org). After
`18 months or longer in the PDT group, patients who switched to
`ranibizumab treatment, on average, maintained the VA measured
`just before crossover. The overall mean change in VA was ⫹0.2
`letters (median change, 0 letters; range, ⫺20 to ⫹24 letters) at
`month 24. The 33 patients who received 3 or more ranibizumab
`injections after crossover had a mean change in VA of ⫺0.6 letters
`(median change, 0 letters; range, ⫺20 to ⫹24 letters). At month
`24, patients who crossed over to ranibizumab had a mean decrease
`of 5.7 letters compared with a mean decrease of 12.1 letters for
`patients who did not cross over. There were no notable differences
`between these PDT groups in the mean changes in total area of
`lesion, total area of CNV, and area of classic CNV at month 24.
`However, patients who crossed over had better control of leakage
`from CNV at month 24 (mean decrease of 1.9 DA in the total area of
`leakage from CNV plus intense progressive RPE staining and 40% of
`patients with leakage from CNV plus intense progressive RPE stain-
`ing) compared with patients who did not cross over (mean decrease of
`0.2 DA and 79% of patients with leakage). Patients randomized to
`PDT who crossed over to ranibizumab as part of the amendment
`were, on average, doing better on their original treatment regimen
`both in VA measures and in control of leakage from CNV than were
`patients who did not cross over (see month 12 and month 18 outcomes
`in Table 6, available at http://aaojournal.org).
`
`Adverse Events
`The cumulative rates of key ocular and nonocular adverse events
`during the 2-year study period are summarized in Table 7. Overall,
`there was no imbalance among the 3 treatment groups in the rates
`of serious and nonserious ocular adverse events in the study eye.
`The percentages of patients with any serious ocular adverse event
`in the study eye were similar among the PDT (7.7%), 0.3-mg
`ranibizumab (7.3%), and 0.5-mg ranibizumab (9.3%) groups.
`
`Exhibit 2085
`Page 04 of 14
`
`

`

`Brown et al
`
`䡠 Ranibizumab vs PDT for Neovascular AMD
`
`Figure 1. Mean change from baseline visual acuity (VA) score (letters) over time. Vertical bars represent ⫾1 standard error of the mean. The mean
`change at some visits in the first year differed slightly from those previously reported6 because the present analysis is based on the final data. P⬍0.001 for
`all comparisons versus verteporfin photodynamic therapy (PDT) at each month. Pairwise analysis of variance models adjusting for VA score at day 0 (⬍45
`letters vs. ⱖ45 letters) were used to analyze mean VA change from baseline at each monthly assessment. The last-observation-carried-forward method was
`used to impute missing data. All tests were 2-sided.
`
`Serious ocular adverse events considered to be potentially
`related to intravitreal ranibizumab treatment
`include endoph-
`thalmitis, uveitis, vitreous hemorrhage, rhegmatogenous retinal
`detachment, retinal tear, and lens damage. In the combined ranibi-
`zumab groups, “presumed” endophthalmitis (i.e., including the
`patient in Table 7 whose adverse event was reported as “serious
`uveitis,” but was treated with systemic antibiotics) in the study eye
`occurred in 3 of 277 patients (1.1%) in the pooled ranibizumab
`groups and in no patients in the PDT group. The rate of presumed
`endophthalmitis in the study eye per injection was 3 of 5921
`injections (0.05%) in the pooled ranibizumab groups; all 3 of these
`patients had gains in VA at month 24 compared with baseline
`
`(⫹13, ⫹26, and ⫹32 letters, respectively). No patient other than
`the one mentioned above experienced uveitis classified as serious.
`Vitreous hemorrhage was reported in 2 of 277 patients (0.7%) in
`the pooled ranibizumab groups versus 0 of 143 patients in the PDT
`group. Rhegmatogenous retinal detachment occurred in 2 patients
`(0.7%) in the pooled ranibizumab groups and 1 patient (0.7%) in
`the PDT group; the rates per ocular injection were 2 of 2571
`(0.07%) in the PDT group (sham injection) and 2 of 5921
`(0.03%) in the pooled ranibizumab groups.
`The percentage of patients who experienced any serious or
`nonserious adverse event of intraocular inflammation (i.e., iritis,
`iridocyclitis, vitritis, uveitis, anterior-chamber inflammation, or
`
`Table 5. Anatomical Characteristics in the Study Eye at Month 24
`
`Month 24 Outcome Measure
`Change in total area of lesion (DA)
`Mean (SD)
`95% CI of the mean
`Change in total area of CNV (DA)
`Mean (SD)
`95% CI of the mean
`Change in area of classic CNV (DA)
`Mean (SD)
`95% CI of the mean
`Change in total area of leakage from CNV ⫹ intense
`progressive RPE staining (DA)
`Mean (SD)
`95% CI of the mean
`Patients with leakage from CNV ⫹ intense
`progressive RPE staining
`
`Verteporfin PDT
`(n ⴝ 143)
`
`2.89 (3.33)
`(2.34, 3.44)
`
`1.60 (2.42)
`(1.20, 2.00)
`
`0.41 (2.30)
`(0.03, 0.79)
`
`⫺0.78 (3.44)
`(⫺1.35, ⫺0.21)
`65.0%
`
`Ranibizumab
`
`0.3 mg (n ⫽ 140)
`
`0.5 mg (n ⫽ 140)
`
`0.52 (1.34)
`(0.30, 0.75)
`
`0.33 (1.21)
`(0.13, 0.54)
`
`⫺0.57 (1.12)
`(⫺0.76, ⫺0.39)
`
`⫺2.23 (2.09)
`(⫺2.58, ⫺1.88)
`37.9%
`
`0.39 (1.34)
`(0.16, 0.61)
`
`0.27 (1.28)
`(0.05, 0.48)
`
`⫺0.72 (1.12)
`(⫺0.91, ⫺0.54)
`
`⫺2.37 (2.14)
`(⫺2.72, ⫺2.01)
`39.3%
`
`CI ⫽ confidence interval; CNV ⫽ choroidal neovascularization; DA ⫽ disc areas; PDT ⫽ photodynamic therapy; SD ⫽ standard deviation; RPE ⫽ retinal
`pigment epithelium.
`NOTE: P⬍0.0001 for all comparisons of each ranibizumab dose group with the verteporfin PDT group.
`
`61
`
`Exhibit 2085
`Page 05 of 14
`
`

`

`Ophthalmology Volume 116, Number 1, January 2009
`
`Table 7. Key Adverse Event Findings: Cumulative for 2 Years of ANCHOR Study
`
`Verteporfin PDT
`(n ⴝ 143)
`
`Ranibizumab 0.3 mg
`(n ⴝ 137)
`
`Ranibizumab 0.5 mg
`(n ⴝ 140)
`
`Key Serious Ocular Adverse Events—no. (%)
`Presumed endophthalmitis*
`Uveitis
`Rhegmatogenous retinal detachment
`Retinal tear
`Vitreous hemorrhage
`Lens damage
`Most Severe Ocular Inflammation, Regardless of Cause
`(Slit-Lamp Examination)—no. (%)
`1⫹
`2⫹
`3⫹
`4⫹
`Key Nonocular Adverse Events—no. (%)
`Treatment-emergent hypertension
`Key arterial thromboembolic events (nonfatal)
`Myocardial infarction
`Cerebrovascular accident
`Death
`Vascular (APTC criteria)储
`Nonvascular#
`Nonocular hemorrhage
`Total (serious or nonserious)**
`Reported as a serious adverse event
`
`0
`0
`1 (0.7)†
`0
`0
`0
`
`1 (0.7)
`0
`0
`0
`
`23 (16.1)
`
`2 (1.4)
`2 (1.4)
`5 (3.5)§
`3 (2.1)
`2 (1.4)
`
`7 (4.9)
`1 (0.7)
`
`0
`0
`2 (1.5)
`0
`2 (1.5)
`0
`
`3 (2.2)
`2 (1.5)
`2 (1.5)
`1 (0.7)
`
`13 (9.5)
`
`1 (0.7)
`3 (2.2)‡
`5 (3.6)
`2 (1.5)
`3 (2.2)
`
`12 (8.8)
`4 (2.9)
`
`3 (2.1)
`1 (0.7)*
`0
`1 (0.7)
`0
`0
`
`9 (6.4)
`0
`4 (2.9)
`1 (0.7)
`
`17 (12.1)
`
`5 (3.6)
`0
`3 (2.1)
`2 (1.4)
`1 (0.7)
`
`13 (9.3)
`3 (2.1)
`
`ANCHOR ⫽ Anti-VEGF Antibody for the Treatment of Predominantly Classic Choroidal Neovascularization in Age-related Macular Degeneration;
`APTC ⫽ Antiplatelet Trialists’ Collaboration; PDT ⫽ photodynamic therapy.
`*Presumed endophthalmitis was defined as cases in which intravitreal or systemic antibiotics were administered. One patient had 2 episodes of intraocular
`inflammation that were reported as uveitis, but one of the episodes was classified as presumed endophthalmitis because it was treated with systemic
`antibiotics. In neither of these 2 episodes was a vitreous culture obtained, and neither was treated with intravitreal antibiotics. Vitreous culture was positive
`for Staphylococcus epidermidis for 1 patient, and culture was negative for another patient.
`†Patient had 2 episodes of rhegmatogenous retinal detachment.
`‡Includes 1 nonserious adverse event of cerebral ischemia.
`§Two patients died after withdrawing from the study because of an adverse event.
`储Verteporfin PDT group: deaths due to cardiac arrest, congestive heart failure, and coronary artery disease (1 patient each). Ranibizumab 0.3-mg group:
`deaths due to cardiac arrest and exacerbation of bradycardia (1 patient each). Ranibizumab 0.5-mg group: deaths due to cardiac failure and worsening of
`chronic heart failure (1 patient each).
`#Verteporfin PDT group: deaths due to chronic obstructive pulmonary disease and complications secondary to renal cell carcinoma (1 patient each).
`Ranibizumab 0.3-mg group: deaths due to respiratory arrest, viral syndrome, and pneumonia (1 patient each). Ranibizumab 0.5-mg group: death due to
`adult failure to thrive.
`**Detailed nonocular hemorrhage adverse events are summarized in Table 8 (available online at http://aaojournal.org).
`
`hypopyon) in the study eye was higher in the ranibizumab groups
`(11.7% in the 0.3-mg group and 17.1% in the 0.5-mg group) than
`in the PDT group (3.5%). Findings from objective slit-lamp ex-
`aminations were consistent with the reports of intraocular inflam-
`mation adverse events. As in all previous trials of ranibizumab,
`transient increases in intraocular pressure in the study eye were
`common in the hour after intravitreal injection of ranibizumab
`(data not shown).
`No traumatic lens d