Randomized, Double-Masked, Sham-Controlled Trial of
`Ranibizumab for Neovascular Age-related Macular
`Degeneration: PIER Study Year 1
`
`CARL D. REGILLO, DAVID M. BROWN, PREMA ABRAHAM, HUIBIN YUE, TSONTCHO IANCHULEV,
`SUSAN SCHNEIDER, AND NAVEED SHAMS, ON BEHALF OF THE PIER STUDY GROUP
`
`● PURPOSE: To evaluate the efficacy and safety of ranibi-
`zumab administered monthly for three months and then
`quarterly in patients with subfoveal choroidal neovascu-
`larization (CNV) secondary to age-related macular de-
`generation (AMD).
`● DESIGN: Phase IIIb, multicenter, randomized, double-
`masked, sham injection-controlled trial in patients with
`predominantly or minimally classic or occult with no
`classic CNV lesions.
`● METHODS: Patients were randomized 1:1:1 to 0.3 mg
`ranibizumab (n ⴝ 60), 0.5 mg ranibizumab (n ⴝ 61), or
`sham (n ⴝ 63) treatment groups. The primary efficacy
`endpoint was mean change from baseline visual acuity
`(VA) at month 12.
`● RESULTS: Mean changes from baseline VA at 12
`months were ⴚ16.3, ⴚ1.6, and ⴚ0.2 letters for the
`sham, 0.3 mg, and 0.5 mg groups, respectively (P <
`.0001, each ranibizumab dose vs sham). Ranibizumab
`arrested CNV growth and reduced leakage from CNV.
`However, the treatment effect declined in the ranibi-
`zumab groups during quarterly dosing (e.g., at three
`months the mean changes from baseline VA had been
`gains of 2.9 and 4.3 letters for the 0.3 mg and 0.5 mg
`doses, respectively). Results of subgroups analyses of
`mean change from baseline VA at 12 months by baseline
`age, VA, and lesion characteristics were consistent with
`the overall results. Few serious ocular or nonocular
`adverse events occurred in any group.
`● CONCLUSIONS: Ranibizumab administered monthly
`for three months and then quarterly provided significant
`VA benefit to patients with AMD-related subfoveal
`CNV and was well tolerated. The incidence of serious
`ocular or nonocular adverse events was low.
`(Am J
`Ophthalmol 2008;145:239 –248. © 2008 by Elsevier
`Inc. All rights reserved.)
`
`Supplemental Material available at AJO.com.
`Accepted for publication Oct 5, 2007.
`From the Retina Service, Wills Eye Institute, Philadelphia, Pennsyl-
`vania (C.D.R.); Vitreoretinal Consultants, The Methodist Hospital,
`Houston, Texas (D.M.B.); BH Regional Eye Institute, Rapid City, South
`Dakota (P.A.); and Genentech, Inc, South San Francisco, California
`(H.Y., T.I., S.S., N.S.).
`Inquiries to Carl D. Regillo, Wills Eye Institute, 840 Walnut Street,
`Suite 1020, Philadelphia, PA 19107; e-mail: cregillo@aol.com
`
`R ANIBIZUMAB (LUCENTIS; GENENTECH, INC, SOUTH
`
`San Francisco, California, USA) is an intravitreally
`administered recombinant, humanized, monoclo-
`nal antibody antigen-binding fragment (Fab) that neutral-
`izes all known active forms of vascular endothelial growth
`factor-A (VEGF-A). It is the first treatment shown to not
`only prevent loss of visual acuity (VA) but also improve
`VA on average in patients with subfoveal choroidal
`neovascularization (CNV) secondary to age-related macu-
`lar degeneration (AMD). In the two pivotal phase III
`trials—the MARINA Study in patients with minimally
`classic or occult with no classic CNV1 and the ANCHOR
`Study in patients with predominantly classic CNV2—
`ranibizumab was injected monthly.
`The phase IIIb PIER Study was designed to determine
`whether a less
`frequent
`ranibizumab dosing schedule
`(monthly for three months and then once every three
`months) would also prevent loss of VA in patients with
`AMD-related subfoveal CNV with or without a classic
`component, and to provide additional safety information.
`This alternative dosing regimen was selected for testing
`based on evidence from phase I and II studies indicating
`that the pharmacodynamic activity of ranibizumab (0.3
`and 0.5 mg) administered intravitreally monthly for three
`doses may last 90 days.3,4
`
`METHODS
`
`PIER IS A TWO-YEAR, PHASE IIIB, MULTICENTER, RANDOM-
`ized, double-masked, sham injection– controlled study of
`the efficacy and safety of ranibizumab in patients with
`AMD-related subfoveal CNV, with or without classic
`CNV. After providing written informed consent, patients
`entered a screening period (ⱕ28 days), with eligibility
`determined by the investigator. A central reading center
`(University of Wisconsin Fundus Photograph Reading
`Center, Madison, Wisconsin) later re-assessed the CNV
`types based on fluorescein angiograms, but this did not
`affect patients’ eligibility. See Supplemental Table A
`(available at AJO.com) for full eligibility criteria.
`Only patients ⱖ50 years old were eligible. One eye per
`subject (the “study eye”) received study treatment. If both
`eyes were eligible, the one with better VA was selected
`
`0002-9394/08/$34.00
`doi:10.1016/j.ajo.2007.10.004
`
`© 2008 BY ELSEVIER INC. ALL RIGHTS RESERVED.
`
`239
`
`Samsung Bioepis Exhibit 1021
`Page 1
`
`

`

`unless, for medical reasons, the other was more appropriate.
`Key inclusion criteria for the study eye were primary or
`recurrent subfoveal CNV secondary to AMD, with the total
`CNV area (classic plus occult CNV) composing ⱖ50% of the
`total AMD lesion area; total AMD lesion size ⱕ12 disk areas
`(DA); and best-corrected VA of 20/40 to 20/320 (Snellen
`equivalent) measured per a standard testing protocol using
`Early Treatment Diabetic Retinopathy Study (ETDRS)
`charts at a distance of 4 meters. Eyes with minimally classic or
`occult with no classic CNV were eligible only if they met any
`of three criteria for presumed disease progression: ⱖ10%
`increase in lesion size based on a fluorescein angiogram
`obtained ⱕone month before day zero, inclusive, vs one
`obtained ⱕsix months before day zero, inclusive; or ⬎one
`Snellen line (or equivalent) VA loss within the prior six
`months; or CNV-associated subretinal hemorrhage ⱕone
`month before day zero. Eyes with predominantly (⬎50% of
`the lesion) classic CNV were not required to meet the criteria
`for presumed disease progression. Key exclusion criteria were
`any prior treatment with verteporfin photodynamic therapy
`(PDT), external-beam radiation therapy, transpupillary ther-
`motherapy, or subfoveal laser photocoagulation (or juxtafo-
`laser photocoagulation ⱕone month
`veal or extrafoveal
`before day zero); permanent structural damage to the central
`fovea; or subretinal hemorrhage involving the fovea if ⱖ1
`DA or ⱖ50% of the total lesion area. Patients were excluded
`if either eye had been treated in a prior antiangiogenic drug
`trial, or if the nonstudy eye received PDT ⱕseven days before
`day zero.
`Using a dynamic randomization algorithm, subjects were
`randomly assigned 1:1:1 to receive 0.3 mg ranibizumab, 0.5
`mg ranibizumab, or sham injections. Randomization was
`stratified by VA score at day zero (ⱕ54 letters [approximately
`worse than 20/80] vs ⱖ55 letters [approximately 20/80 or
`better], CNV type (minimally classic vs occult with no classic
`vs predominantly classic CNV), and study center.
`To achieve double-masking of treatment assignment, at
`least two investigators participated at each study site: an
`“injecting” ophthalmologist unmasked to treatment as-
`signment (ranibizumab vs sham) but masked to ranibi-
`zumab dose, and a masked “evaluating” ophthalmologist
`for efficacy and safety assessments. All other study site
`personnel (other than those assisting with study treatment
`administration), central reading center personnel, and the
`subjects were masked to treatment assignment.
`The ranibizumab groups received their assigned dose by
`intravitreal injection every month for three doses (day
`zero, months one and two), followed by doses every three
`months (months five, eight, 11, 14, 17, 20, and 23).
`Ranibizumab injection procedures have been described
`previously.1,2 For the sham-injected control group, an
`empty syringe without a needle was used, with pressure
`applied to the anesthetized and antiseptically prepared eye
`at the site of a typical intravitreal injection. Pre- and
`postinjection procedures (described previously1,2) were
`identical for all groups.
`
`The original study protocol specified that each treat-
`ment group would follow the same injection schedule.
`Thus, during the 24-month study, a total of 10 ranibizumab
`or sham injections were to be given, with six of the 10
`during the first 12 months. After careful review of recent
`clinical data,
`including 12-month data from the two
`studies,1,2
`pivotal phase III
`the study protocol was
`amended on February 27, 2006 to allow control subjects
`who had completed the month-12 visit (the assessment
`timepoint for the primary efficacy analysis) to cross over to
`0.5 mg ranibizumab for the remainder of the treatment
`period (subjects in the ranibizumab groups continued their
`originally assigned dose of 0.3 or 0.5 mg). On August 21,
`2006, the protocol was again amended to increase assess-
`ments from quarterly to monthly after month 12, and to
`switch subjects randomized to the 0.3 mg dose to the 0.5
`mg dose for the remainder of their study treatment. Also,
`because ranibizumab was by this time approved by the U.S.
`Food and Drug Administration (FDA), subjects were
`allowed to receive ranibizumab in the fellow eye as well as
`the study eye. No subjects were unmasked to their original
`treatment assignment as a result of
`these protocol
`amendments.
`Assessments were performed at scheduled clinic visits.
`The first ranibizumab (0.3 or 0.5 mg) or sham treatment
`was administered on day zero. At subsequent injection
`visits, subjects underwent a preinjection safety evaluation.
`In addition to injection visits (day zero and months one,
`two, five, eight, 11, 14, 17, 20, and 23), clinic visits were
`scheduled at months three, 12, and 24. At each scheduled
`visit, subjects received a full ophthalmologic assessment,
`including VA testing using ETDRS charts at a test
`distance of 4 meters, slit-lamp biomicroscopy, fundoscopy,
`and intraocular pressure (IOP) measurement. Fundus pho-
`tography and fluorescein angiography (FA) were done at
`day zero and months three, five, eight, 12, and 24. Optical
`coherence tomography (OCT) was done at selected study
`sites at day zero and months one, two, three, five, eight, 12,
`and 24. The primary efficacy endpoint was mean change
`from baseline to 12 months in VA score. The following
`key secondary VA endpoints were also assessed at 12
`months: proportion of subjects losing ⱕ15 letters (⬇3
`lines) from baseline; proportion gaining ⱖ15 letters from
`baseline; proportion with a Snellen equivalent of 20/200 or
`worse (legal blindness ⫽ 20/200 or worse in both eyes);
`mean change from baseline in the near activities, distance
`activities, and vision-specific dependency NEI VFQ-25
`subscales; and mean change from baseline in total area of
`CNV and total area of leakage from CNV (based on
`central reading center assessment). Prespecified explor-
`atory endpoints included the proportion of subjects who
`had lost ⱕ30 letters (⬇6 lines) from baseline VA at 12
`months, the mean change from baseline at three months,
`and mean change from three months to 12 months.
`Key safety assessments were the incidence and severity
`of ocular and nonocular adverse events, changes in vital
`
`240
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`FEBRUARY 2008
`
`Samsung Bioepis Exhibit 1021
`Page 2
`
`

`

`TABLE 1. Ranibizumab for Neovascular Age-Related Macular Degeneration: Subject Demographics and Baseline Study
`Eye Characteristics
`
`Characteristic
`
`Gender—no. (%)
`Male
`Female
`Race—no. (%)
`White
`Other
`Age—years
`Mean (SD)
`Range
`Age group—no. (%)
`50–64 years
`65–74 years
`75–84 years
`ⱖ85 years
`Prior therapy for AMD—no. (%)
`Any
`Laser photocoagulation
`Medication*
`Supplements
`Years since first diagnosis of neovascular AMD†
`Mean (SD)
`Range
`Visual acuity (letters with approximate Snellen equivalent)‡
`Mean ( SD)
`ⱕ54, 20/80—no. (%)
`ⱖ55, 20/80—no. (%)
`Visual acuity (approximate Snellen equivalent)‡—no. (%)
`20/200 or worse
`Better than 20/200 but worse than 20/40
`20/40 or better
`CNV lesion subtype—no. (%)
`Occult with no classic
`Minimally classic
`Predominantly classic
`Cannot classify
`Total area of lesion§
`Mean (SD) (DA)
`Range (DA)
`ⱕ4 DA—no. (%)
`⬎4 DA—no. (%)
`Total area of CNV (DA)§
`Mean (SD)
`Range
`Leakage from CNV, plus RPE staining (DA)§
`Mean (SD)
`Range
`
`Sham
`(n ⫽ 63)
`
`20 (31.7)
`43 (68.3)
`
`59 (93.7)
`4 (6.3)
`
`77.8 (7.1)
`59–92
`
`4 (6.3)
`12 (19.0)
`36 (57.1)
`11 (17.5)
`
`35 (55.6)
`3 (4.8)
`1 (1.6)
`34 (54.0)
`
`0.3 (0.5)
`0.0–3.0
`
`55.1 (13.9)
`25 (39.7)
`38 (60.3)
`
`10 (15.9)
`42 (66.7)
`11 (17.5)
`
`20 (31.7)
`29 (46.0)
`14 (22.2)
`0
`
`4.24 (3.25)
`0.10–17.00
`33 (52.4)
`30 (47.6)
`
`3.56 (3.25)
`0.02–17.00
`
`4.25 (3.55)
`0.20–19.00
`
`Ranibizumab 0.3 mg
`(n ⫽ 60)
`
`Ranibizumab 0.5 mg
`(n ⫽ 61)
`
`26 (43.3)
`34 (56.7)
`
`57 (95.0)
`3 (5.0)
`
`78.7 (6.3)
`60–93
`
`1 (1.7)
`12 (20.0)
`37 (61.7)
`10 (16.7)
`
`35 (58.3)
`5 (8.3)
`1 (1.7)
`33 (55.0)
`
`0.7 (1.6)
`0.0–9.1
`
`55.8 (12.2)
`29 (48.3)
`31 (51.7)
`
`3 (5.0)
`49 (81.7)
`8 (13.3)
`
`29 (48.3)
`22 (36.7)
`8 (13.3)
`1 (1.7)
`
`4.38 (3.30)
`0.09–20.30
`32 (54.2)
`27 (45.8)
`
`3.80 (3.43)
`0.00–20.30
`
`4.49 (3.58)
`0.00–22.50
`
`28 (45.9)
`33 (54.1)
`
`56 (91.8)
`5 (8.2)
`
`78.8 (7.9)
`54–94
`
`4 (6.6)
`12 (19.7)
`31 (50.8)
`14 (23.0)
`
`33 (54.1)
`7 (11.5)
`3 (3.3)
`28 (45.9)
`
`0.7 (1.2)
`0.0–5.0
`
`53.7 (15.5)
`27 (44.3)
`34 (55.7)
`
`10 (16.4)
`36 (59.0)
`15 (24.6)
`
`30 (49.2)
`18 (29.5)
`13 (21.3)
`0
`
`4.01 (2.64)
`0.03–10.00
`31 (50.8)
`30 (49.2)
`
`3.29 (2.27)
`0.03–9.65
`
`3.99 (2.61)
`0.50–9.70
`
`AMD ⫽ age-related macular degeneration; CNV ⫽ choroidal neovascularization; DA ⫽ disk areas; RPE ⫽ retinal pigment epithelium;
`SD ⫽ standard deviation.
`*Triamcinolone acetonide in the sham and 0.3 mg ranibizumab groups; alteplase and a multiple vitamin / mineral formulation in the 0.5 mg
`ranibizumab group.
`†For this parameter, the numbers of subjects are as follows: sham, n ⫽ 62; 0.3 mg ranibizumab, n ⫽ 59; 0.5 mg ranibizumab, n ⫽ 61.
`‡Measured using Early Treatment Diabetic Retinopathy Study (ETDRS) charts at a starting distance of 4 meters.
`§For this parameter, the numbers of subjects are as follows: sham, n ⫽ 63; 0.3 mg ranibizumab, n ⫽ 59; 0.5 mg ranibizumab, n ⫽ 61.
`
`VOL. 145, NO. 2
`
`RANIBIZUMAB FOR AMD: PIER STUDY YEAR 1
`
`241
`
`Samsung Bioepis Exhibit 1021
`Page 3
`
`

`

`signs, and the incidence of positive serum antibodies to
`ranibizumab. Slit-lamp examination and indirect ophthal-
`moscopy were performed before each study injection.
`Grading scales for flare/cells and vitreous hemorrhage
`density (see Supplemental Tables B1 to B3 for grading
`criteria) were used to grade intraocular inflammation or
`vitreous hemorrhage, assessed by slit-lamp examination.
`IOP was measured using applanation tonometry before and
`60 ⫾ 10 minutes after each study treatment.
`Safety analyses, performed using descriptive statistics
`and including all treated subjects, were based on the
`treatment actually received. Efficacy analyses used the
`intent-to-treat approach and included all subjects as random-
`ized. Missing values were imputed using the last-observation-
`carried-forward method. All pairwise comparisons between
`the ranibizumab groups and the sham group used a statistical
`model including only two treatment groups (active vs con-
`trol) at a time. For the primary efficacy endpoint, a Hochberg-
`Bonferroni adjustment5 was made for multiple treatment
`comparisons of each ranibizumab dose group with the sham
`group. For secondary efficacy endpoints, a Type I error
`management plan was used to adjust for multiplicity of
`treatment comparisons and secondary endpoints. Unless
`otherwise noted, efficacy analyses were stratified by CNV
`classification at baseline (minimally classic vs occult with
`no classic vs predominantly classic CNV), as determined
`by the central reading center, and by baseline VA (ⱕ54 vs
`ⱖ55 letters). For binary endpoints, stratified Cochran ␹2
`tests were used for between-groups comparisons of propor-
`tions of subjects meeting the endpoint. Analysis of vari-
`ance or analysis of covariance models were used to analyze
`continuous endpoints.
`The study sample size was based on the primary efficacy
`endpoint. Calculations were based on a 1:1:1 randomiza-
`tion ratio (0.3 mg vs 0.5 mg ranibizumab vs sham), the
`Student t test for comparing mean changes from baseline
`to 12 months in VA (for each ranibizumab group vs sham),
`and the Hochberg–Bonferroni multiple comparison proce-
`dure at an overall ␣ level of
`.05. The power of the
`Hochberg–Bonferroni multiple comparison procedure was
`evaluated using Monte Carlo simulations. The target
`sample size of 180 subjects provided 90% power in the
`intent-to-treat analysis to detect a nine-letter difference
`between one or both ranibizumab dose groups and the
`sham group in mean change in VA at month 12, according
`to the Hochberg–Bonferroni criterion (assumptions based
`on results of the TAP6 and VIP7 trials and anticipated
`proportions of each CNV type).
`Prior PDT in the study eye was an exclusion criterion,
`but subjects with predominantly classic CNV at study
`entry or whose CNV was confirmed by the central reading
`center to have converted during the study from minimally
`classic or occult with no classic to predominantly classic
`CNV could receive verteporfin PDT treatment in the
`study eye given according to the Visudyne prescribing
`information8 (i.e., the physician should reevaluate the
`
`patient every three months and if CNV leakage is detected
`on FA, therapy should be repeated) and at the discretion of
`the investigator per standard of care. Treatment of mini-
`mally classic or occult with no classic CNV with PDT is
`not approved by the U.S. FDA, but was permitted in this
`study if the investigator deemed PDT to be indicated and
`the lesion met all of the following criteria: ⱖ 20-letter loss
`from baseline VA recorded at all study visits over a
`three-month period that included at least two study visits,
`total CNV lesion area ⱕ4 DA, and active CNV as defined
`in the inclusion criteria (Supplemental Table A). Subjects
`receiving PDT in the study eye could continue study
`treatment, but PDT could not be given less than 28 days
`before or less than 21 days after a study injection. Also,
`PDT in the nonstudy eye could not be given less than five
`days before or less than 21 days after a study injection. No
`independent check was done to determine if investigators
`followed the instructions regarding PDT administration
`that were provided in the study protocol, nor was the
`clinical judgment of the investigator regarding suitability
`of the subject for PDT questioned or independently
`verified.
`Treatment of either eye with other anti-VEGF drugs was
`prohibited. When pegaptanib sodium (Macugen) was ap-
`proved by the U.S. FDA in January 2005, subjects were
`allowed to opt for treatment with this agent but were to be
`discontinued from their randomized study treatment and
`followed for the remainder of the study period.
`
`RESULTS
`
`BETWEEN SEPTEMBER 7, 2004 AND MARCH 16, 2005, 184 SUB-
`jects were enrolled at 43 investigative sites in the U.S. and
`were randomly assigned to study treatment: 60 to 0.3 mg
`ranibizumab, 61 to 0.5 mg ranibizumab, and 63 to sham
`injection. Subject disposition is summarized in Supple-
`mental Table C (available at AJO.com). Treatment com-
`pliance was good in the ranibizumab groups, with 85% or
`more of subjects receiving each scheduled injection. In the
`sham group, 27% of subjects permanently discontinued
`treatment before month 12, most often because the sub-
`ject’s condition mandated another therapeutic interven-
`tion. A month-12 VA score was obtained from 97% of
`each ranibizumab group and 86% of the sham group.
`The treatment groups were well balanced overall for
`demographic and baseline ocular characteristics (Table 1).
`Each group was predominantly White and nearly two-
`thirds female, with a mean age of ⬇78 years. The baseline
`mean VA score was 53 to 56 letters (approximate Snellen
`equivalent, 20/63 to 20/80) across groups. The first diag-
`nosis of neovascular AMD was within the prior year in
`87% of subjects. Overall, 80% of subjects had either occult
`with no classic or minimally classic CNV lesions, but
`occult with no classic CNV was more common in the
`ranibizumab groups than in the sham group (nearly half vs
`
`242
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`FEBRUARY 2008
`
`Samsung Bioepis Exhibit 1021
`Page 4
`
`

`

`FIGURE 1. Ranibizumab for neovascular age-related macular
`degeneration (AMD). Mean change from baseline visual acuity,
`measured as letters read on the Early Treatment of Diabetic
`Retinopathy Study (ETDRS) chart, at monthly intervals. At
`month 12, the 0.3 mg ranibizumab group and the 0.5 mg
`ranibizumab group differed from the sham group by 14.7 and
`16.1 letters, respectively (P < .0001). The arrows indicate that
`ranibizumab or sham injections occurred at day zero, month
`one, month two, month five, month eight, and month 11.
`
`less than one-third of study eye lesions, respectively).
`Nearly half of each group had lesion sizes ⱖ4 DA. The
`mean total areas of the AMD lesion, the CNV component,
`and leakage from CNV plus retinal pigment epithelium
`(RPE) staining were similar among the groups.
`Of the 184 randomized subjects, 19 (10.3%) received
`one or more treatments with PDT in the study eye during
`the first treatment year: 17 subjects in the sham-injection
`group (27.0%), one subject in the 0.3 mg group (1.7%),
`and one subject in the 0.5 mg group (1.6%). Of the 14
`subjects (22.2%) in the sham group who had predomi-
`nantly classic CNV at study entry, four received at least
`one PDT treatment in the first year (total ⫽ five PDT
`administrations). None of the 21 subjects (17.4%) in the
`ranibizumab groups with predominantly classic CNV at
`study entry received PDT.
`Figure 1 shows the mean change from baseline VA by
`study month for the first treatment year. At 12 months
`(primary endpoint), sham-treated subjects had lost a mean
`of 16.3 letters, whereas ranibizumab-treated subjects had
`lost a mean of 1.6 letters (0.3 mg dose; P ⫽ .0001 vs sham)
`or 0.2 letters (0.5 mg dose; P ⬍ .0001 vs sham). Thus, the
`difference from the sham group after one year of treatment
`was 14.7 letters in the 0.3 mg ranibizumab group and 16.1
`letters in the 0.5 mg ranibizumab group. Moreover, each of
`the ranibizumab groups was statistically significantly dif-
`ferent from the sham group at month one, following a
`single injection of ranibizumab (P ⫽ .02 for 0.3 mg dose,
`P ⬍ .0001 for 0.5 mg dose), and at each monthly assessment
`(all P ⬍ .02). After the initial three monthly doses, both
`
`FIGURE 2. Ranibizumab for neovascular AMD. Percentages
`of the three treatment groups who (Top) at 12 months had lost
`fewer than 15 letters from baseline visual acuity score, (Mid-
`dle) at 12 months had gained 15 or more letters from baseline
`visual acuity score, and (Bottom) had a Snellen equivalent
`visual acuity of 20/200 or worse at baseline (left) and at month
`12 (right). P values are vs the sham treatment group.
`
`ranibizumab groups showed a more than 10-letter benefit in
`mean VA compared with the sham group.
`Results for key vision-related secondary endpoints at 12
`months are summarized in Figure 2. Significantly greater
`proportions of the ranibizumab groups than the sham group
`
`VOL. 145, NO. 2
`
`RANIBIZUMAB FOR AMD: PIER STUDY YEAR 1
`
`243
`
`Samsung Bioepis Exhibit 1021
`Page 5
`
`

`

`TABLE 2. Ranibizumab for Neovascular Age-Related Macular Degeneration: Mean Change from Baseline in the Total Area of
`Choroidal Neovascularization and the Total Area of Leakage from Choroidal Neovascularization at Month 12
`
`Change from Baseline at Month 12
`
`Sham Injection (n ⫽ 63)
`
`Ranibizumab 0.3 mg (n ⫽ 59*)
`
`Ranibizumab 0.5 mg (n ⫽ 61)
`
`Change in total area of CNV (DA)
`Mean (SD)
`95% CI for mean
`P value (vs Sham)†
`Change in total area of leakage from
`CNV ⫹ intense RPE staining (DA)
`Mean (SD)
`95% CI for mean
`P value (vs Sham)
`
`2.08 (2.66)
`(1.41 to 2.75)
`
`1.40 (3.77)
`(0.45 to 2.35)
`
`0.18 (2.13)
`(⫺0.37 to 0.74)
`.0001
`
`⫺1.41 (2.69)
`(⫺2.12 to ⫺0.71)
`⬍.0001
`
`0.43 (1.86)
`(⫺0.04 to 0.91)
`.0002
`
`⫺1.29 (2.48)
`(⫺1.93 to ⫺0.66)
`⬍.0001
`
`CI ⫽ confidence interval; CNV ⫽ choroidal neovascularization; DA ⫽ disk areas; RPE ⫽ retinal pigment epithelium; SD ⫽ standard
`deviation.
`Note: The last-observation-carried-forward method was used to impute missing data. Strata were defined using two factors: baseline CNV
`classification (minimally classic vs occult with no classic vs predominantly classic) and baseline visual acuity score (ⱕ54 vs ⱖ55 letters).
`*One subject randomized to the 0.3 mg ranibizumab group withdrew before receiving any study treatment.
`†P values are based on pairwise analysis of covariance (ANCOVA) models adjusted for the two stratification factors and baseline value of
`the endpoint.
`
`had lost fewer than 15 letters from baseline VA: 83.3% and
`90.2% of the 0.3 mg and 0.5 mg groups, respectively,
`compared with 49.2% of the sham group (P ⬍ .0001 for
`each dose level vs sham). However, the three treatment
`groups did not differ significantly in the proportions
`gaining at least 15 letters: 9.5% in the sham group, 11.7%
`in the 0.3 mg ranibizumab group, and 13.1% in the 0.5
`ranibizumab group. Significantly smaller proportions of the
`ranibizumab groups than the sham group had VA of 20/200
`or worse Snellen equivalent at month 12: 23.3% and
`24.6% of the 0.3 mg and 0.5 mg groups, respectively,
`compared with 52.4% of the sham group (P ⫽ .0002 and
`P ⬍ .001, respectively vs sham). However, the proportion
`with VA this poor at baseline was smaller in the 0.3 mg
`dose group (5%) than in the 0.5 mg ranibizumab group
`(16.4%) and sham group (16.4%), which may partially
`account for the smaller proportion in this dose group at
`month 12. There was no statistically significant difference
`between either ranibizumab dose group and the sham
`control for any of the three NEI VFQ-25 subscales that
`were prespecified as secondary endpoints. However, post
`hoc analysis indicated that significantly more subjects in
`the ranibizumab groups reported clinically meaningful
`(ⱖ10-point increases) in the near activities subscale scores
`compared with sham (32% for 0.3 mg ranibizumab group,
`31% for 0.5 mg ranibizumab group and 14% for sham
`group, P ⬍ .05 vs sham for both ranibizumab groups).
`Prespecified subgroup analyses of the month 12 VA
`results were performed for several baseline characteristics:
`age (less than 75 and 75 or over), gender, and, for the study
`eye, VA score (ⱕ54 letters vs ⱖ55 letters), total lesion size
`(ⱕ4 vs⬎4), occult CNV present at baseline (yes vs no),
`and CNV angiographic type at baseline. In these subgroup
`analyses the treatment effect of ranibizumab at both doses
`
`FIGURE 3. Ranibizumab for neovascular AMD. Mean change
`over time in the area of leakage from choroidal neovasculariza-
`tion (CNV) plus intense progressive retinal pigment epithelium
`staining. At month 12, each of the ranibizumab-treated groups
`showed significantly less leakage than the sham-treated group
`(P < .0001). The arrows indicate that ranibizumab or sham
`injections occurred at day zero, month one, month two, month
`five, month eight, and month 11.
`
`compared with sham injection was consistent with the
`overall results. For the subgroups of age, gender, baseline
`VA greater than or equal to 55 letters, baseline lesion size
`less than or equal to 4 DA, occult CNV present at
`baseline, and occult with no classic CNV, both ranibi-
`zumab dose groups were significantly different from the
`sham injection group (P ⬍ .05). For the subgroups of
`occult CNV absent at baseline and predominantly classic
`CNV, the 0.3 mg dose group was significantly different
`from the sham injection group (P ⬍ .05). For the baseline
`
`244
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`FEBRUARY 2008
`
`Samsung Bioepis Exhibit 1021
`Page 6
`
`

`

`zumab and sham groups for change from baseline in total
`area of CNV (both ranibizumab doses P ⬍ .001 vs sham).
`Ranibizumab also reduced the total area of leakage from
`CNV plus intense progressive RPE staining on average,
`whereas the sham group exhibited an increase trend (both
`ranibizumab doses P ⬍ .0001 vs sham); this pattern was
`evident at each of the four postbaseline assessments, and
`was most pronounced at month 12 (Figure 3).
`Comparisons between month three and month 12 for
`the VA endpoints were considered indicative of the
`efficacy of the quarterly dosing schedule as a maintenance
`therapy, and therefore several prespecified exploratory
`analyses were conducted. On average, there was 4.5-letter
`decline in VA between month three and month 12 for
`both ranibizumab dose groups. Because neither of the 95%
`confidence intervals for these values contained zero ([⫺8.6
`to ⫺0.3] and [⫺7.2 to ⫺1.7] for the 0.3 mg and 0.5 mg
`groups, respectively), these declines were considered sta-
`tistically significant. Over the same timeline, VA in the
`sham group declined by a mean of 7.6 letters. The differences
`between the ranibizumab dose groups and the sham group in
`mean change in VA from month three to month 12 were not
`statistically significant.
`Data for comparisons of OCT-assessed anatomical out-
`comes over time were available from 40, 37, and 41
`subjects from the sham, 0.3, and 0.5 mg groups, respec-
`tively (Figure 4). On average, both ranibizumab groups
`showed decreases in retinal thickness over the 12-month
`period. In ranibizumab-treated subjects, a statistically sig-
`nificant within-group reduction from baseline was seen as
`early as month one for both foveal center point retinal
`thickness (P ⬍ .001, each dose group) and central subfield
`retinal thickness (P ⬍ .007, each dose group). At month
`12, compared with the sham group, ranibizumab-treated
`subjects showed significantly greater mean decreases from
`baseline in foveal center point thickness (P ⫽ 0.01 for 0.3
`mg and P ⫽ .0006 for 0.5 mg) and, in the 0.5 mg group,
`central subfield retinal thickness (P ⫽ .04 for 0.5 mg).
`There was a maximal decrease in foveal center point
`thickness at months two and three for both ranibizumab
`groups, compared with a continued increase in thickness in
`the sham group during the same period. During the
`ensuing quarterly dosing interval, at assessments made
`three months after the previous dose (months five and
`eight),
`foveal center point thickness was on average
`greater than that at months two and three, and also greater
`than that at month 12, which followed a ranibizumab dose
`at month 11.
`Fluorescein angiographic data for comparisons of change
`over time were available from 63, 59, and 61 subjects from
`the sham, 0.3, and 0.5 mg groups, respectively. The leakage
`from CNV plus RPE staining increased by a mean of
`approximately 1.4 DA at 12 months in sham-treated
`subjects while decreasing by a mean of 1.4 DA and 1.3 DA
`for the 0.3 mg group and the 0.5 mg group at 12 months.
`During the quarterly dosing interval (months three to 12),
`
`FIGURE 4. Ranibizumab for neovascular AMD. Mean change
`over time in (Top ) foveal center point retinal thickness (P ⴝ .01
`for 0.3 mg ranibizumab and P < .001 for 0.5 ranibizumab, vs
`sham at 12 months) and (Bottom) central subfield retinal thick-
`ness (P < .05 for 0.5 ranibizumab, vs sham at 12 months). The
`number of subjects contributing data for central subfield retinal
`thickness is less than that for foveal center point retinal thickness
`because calculation of the former requires nine data points on the
`optical coherence tomography (OCT) scan whereas calculation of
`the latter requires only six. Therefore, there was a greater
`likelihood of missing data points (and consequently an inability to
`calculate thickness) for central subfield retinal thickness than for
`foveal center point retinal thickness. The arrows indicate that
`ranibizumab or sham injections occurred at day zero, month one,
`month two, month five, month eight, and month 11.
`
`lesion size less than 4 DA subgroup, the 0.5 mg ranibi-
`zumab group was significantly different from the sham
`injection group (P ⬍ .05). The other subgroup compari-
`sons did not achieve statistical significance.
`The results for key secondary endpoints related to lesion
`morphologic characteristics at 12 months are summarized
`in Table 2. Ranibizumab arrested CNV growth, on aver-
`age, as indicated by the difference between the ranibi-
`
`VOL. 145, NO. 2
`
`RANIBIZUMAB FOR AMD: PIER STUDY YEAR 1
`
`245
`
`Samsung Bioepis Exhibit 1021
`Page 7
`
`

`

`TABLE 3. Ranibizumab for Neovascular Age-Related Macular Degeneration: Summary of Safety During First 12 Months
`
`Sham Injection (n ⫽ 63)
`
`Ranibizumab 0.3 mg (n ⫽ 59*)
`
`Ranibizumab 0.5 mg (n ⫽ 61)
`
`Key serious ocular adverse events†—no. (%)
`Endophthalmitis, uveitis, retinal detachment,
`or lens damage
`Ocular hemorrhage‡
`Macular edema
`Most severe ocular inflammation, regardless of
`cause (slit-lamp examination)†#—no. (%)
`None
`Trace
`1⫹
`2⫹
`3⫹ or 4⫹
`Key nonocular adverse events—no. (%)
`Nonocular hemorrhage
`Hypertension
`Arterial thromboemobolic events
`Myocardial infarction
`Cerebrovascular accident
`Ischemic cardiomyopathy
`Death
`
`0
`2 (3.2)
`2 (3.2)
`
`59 (95.2)
`2 (3.2)
`0
`1 (1.6)
`0
`
`3 (4.8)
`5 (8.1)
`
`0
`0
`1 (1.6)
`0
`
`0
`2 (3.4)
`1 (1.7)
`
`55 (93.2)
`1 (1.7)
`2 (3.4)
`1 (1.7)
`0
`
`2 (3.4)
`4 (6.8)
`
`0
`0
`0
`0
`
`0
`0
`0
`
`59 (96.7)
`1 (1.6)
`0
`1 (1.6)
`0
`
`4 (6.6)§
`6 (9.8)
`
`0
`0
`0
`0
`
`Note: Multiple occurrences of the same event for a subj