Randomized, Double-Masked, Sham-Controlled Trial of
`Ranibizumab for Neovascular Age-Related Macular
`Degeneration: PIER Study Year 2
`
`PREMA ABRAHAM, HUIBIN YUE, AND LAURA WILSON
`
`● PURPOSE: To evaluate efficacy and safety of quarterly
`(and then monthly) ranibizumab during the 2-year Phase
`IIIb, multicenter, randomized, double-masked, sham in-
`jection– controlled study of the efficacy and safety of
`ranibizumab in subjects with subfoveal CNV with or
`without classic CNV secondary to AMD (PIER) study.
`● DESIGN: Phase IIIb, multicenter, randomized, double-
`masked, sham injection– controlled trial in patients with
`choroidal neovascularization (CNV) secondary to age-
`related macular degeneration (AMD).
`● METHODS: Patients were randomized 1:1:1 to sham
`injection (n ⴝ 63) or 0.3 mg (n ⴝ 60) or 0.5 mg (n ⴝ
`61) intravitreal ranibizumab monthly for 3 months and
`then quarterly. During study year 2, eligible sham-group
`patients crossed over to 0.5 mg ranibizumab quarterly.
`Later in year 2, all eligible randomized patients rolled
`over to 0.5 mg ranibizumab monthly. Key efficacy and
`safety outcomes of the 2-year trial are reported.
`● RESULTS: At month 24, visual acuity (VA) had de-
`creased an average of 21.4, 2.2, and 2.3 letters from
`baseline in the sham, 0.3 mg, and 0.5 mg groups (P <
`.0001 for each ranibizumab group vs sham). VA of sham
`patients who crossed over (and subsequently rolled over)
`to ranibizumab decreased across time, with an average
`loss of 3.5 letters 10 months after crossover. VA of 0.3
`mg and 0.5 mg group patients who rolled over to monthly
`ranibizumab increased for an average gain of 2.2 and 4.1
`letters, respectively, 4 months after rollover. The ocular
`safety profile of ranibizumab was favorable and consistent
`with previous reports.
`● CONCLUSIONS: Ranibizumab provided significant VA
`benefit in patients with AMD-related CNV compared
`with sham injection. Ranibizumab appeared to provide
`additional VA benefit to treated patients who rolled over
`to monthly dosing, but not to patients who began receiv-
`ing ranibizumab after >14 months of sham injections.
`(Am J Ophthalmol 2010;150:315–324. © 2010 by
`Elsevier Inc. All rights reserved.)
`
`Accepted for publication Apr 18, 2010.
`From Black Hills Regional Eye Institute (P.A.), Rapid City, South
`Dakota; and Genentech, Inc (H.Y., L.W.), South San Francisco, Cali-
`fornia.
`Inquiries to Prema Abraham, Black Hills Regional Eye Institute, 2800
`Third Street, Rapid City, SD 57701; e-mail: retina@bhrei.com
`
`N EOVASCULAR AGE-RELATED MACULAR DEGENER-
`
`ation (AMD) is characterized by new vessel
`growth and leakage in the choroidal vascular
`network beneath the macula, with extension and leakage
`into the subretinal space. Although the pathologic events
`that precede choroidal neovascularization (CNV) are not
`clearly understood, disrupting the activity of vascular
`endothelial growth factor A (VEGF-A), a diffusible cyto-
`kine that promotes angiogenesis and vascular permeability,
`effectively treats CNV secondary to AMD.
`Ranibizumab (Lucentis; Genentech, South San Fran-
`cisco, California, USA) is an intravitreally administered
`recombinant, humanized, monoclonal antibody antigen-
`binding fragment that neutralizes all known active forms of
`VEGF-A. In 2 Phase III pivotal studies––the MARINA1
`study in patients with minimally classic or occult with no
`classic CNV and the ANCHOR2,3 study in patients with
`predominantly classic CNV––monthly intravitreal injec-
`tions of 0.3 mg or 0.5 mg ranibizumab not only prevented
`vision loss but also improved visual acuity (VA) compared
`with sham injections or photodynamic therapy (PDT)
`with verteporfin.
`Subsequently, a Phase IIIb, multicenter, randomized,
`double-masked, sham injection– controlled study of the
`efficacy and safety of ranibizumab in subjects with subfo-
`veal CNV with or without classic CNV secondary to AMD
`(PIER) evaluated adverse events and VA benefit of
`quarterly dosing in patients with neovascular AMD.
`The PIER dosing schedule—monthly for 3 months and
`then quarterly—was selected based on Phase I and II
`studies, which indicated that the VA benefits of 0.3 mg
`and 0.5 mg ranibizumab administered intravitreally
`monthly for 3 months may last up to 90 days.4
`While ranibizumab administered on the PIER dosing
`schedule provided significant VA benefit compared to
`sham injections in patients with neovascular AMD, quar-
`terly dosing with ranibizumab did not provide the VA
`benefit demonstrated by monthly dosing in the MARINA
`and ANCHOR studies.5 In fact, during study year 2, after
`careful review of available clinical data, including the
`12-month data from MARINA and ANCHOR, the PIER
`protocol was amended to provide all PIER patients the
`opportunity to receive ranibizumab.
`Here, we present VA and safety outcomes over 2 years in
`the PIER study, showing that the VA benefit of quarterly
`
`0002-9394/$36.00
`doi:10.1016/j.ajo.2010.04.011
`
`© 2010 BY ELSEVIER INC. ALL RIGHTS RESERVED.
`
`315
`
`Exhibit 2087
`Page 01 of 11
`
`

`

`increase in lesion size based on fluorescein angiography
`[FA] obtained 1 month prior to study initiation [ie, day 0]
`compared to FA obtained 6 months prior to day 0; ⱖ5
`ETDRS letter [1 Snellen line] VA loss within 6 months
`prior to day 0; or CNV-associated subretinal hemorrhage 1
`month prior to day 0). Patients who had fibrosis or atrophy
`involving the center of the fovea or subretinal hemorrhage
`ⱖ1 DA or ⱖ50% of total lesion area with foveal involve-
`ment were excluded. One eye per patient was studied.
`Eligible patients were randomized 1:1:1 to sham injections,
`0.3 mg intravitreal ranibizumab, or 0.5 mg intravitreal ranibi-
`zumab (Figure 1). Patients were masked to treatment. Ran-
`domization was stratified by best-corrected VA (ⱕ54 ETDRS
`letters, ⬃20/80 or worse Snellen equivalent vs ⱖ55 ETDRS
`letters, ⬃20/80 or better Snellen equivalent) at day 0, CNV
`type (minimally classic vs occult with no classic vs predom-
`inantly classic), and study center. The protocol mandated
`that patients receive sham injections or intravitreal injections
`of their assigned ranibizumab dose once a month for 3 months
`(day 0, month 1, month 2) and every 3 months thereafter
`(months 5, 8, 11, 14, 17, 20, and 23), for the duration of the
`2-year study. Fluorescein angiography and fundus photogra-
`phy were performed at months 3, 5, 8, 12, and 24 and were
`evaluated by a central reading center (Fundus Photograph
`Reading Center, University of Wisconsin, Madison, Wiscon-
`sin, USA). Patients underwent complete ocular examination,
`including VA assessment at each study visit (ie, the first 3
`months and then quarterly). The Vision Functioning Ques-
`tionnaire-25 (VFQ-25) was administered at baseline and at
`months 3, 8, 12, and 24, prior to patients completing any
`other study-related procedures. In addition to injection visits,
`clinic visits were scheduled at months 3, 12, and 24. Subse-
`quent protocol amendments (crossover and rollover amend-
`ments described below) increased subject assessments from
`quarterly to monthly. The monthly assessments were identi-
`cal to the previous quarterly assessments.
`The incidence and severity of ocular and nonocular ad-
`verse events (AEs) and changes in vital signs were assessed at
`all study visits. In accordance with the criteria established by
`the worldwide Antiplatelet Trialists’ Collaboration,6 arterial
`thromboembolic events (ATEs), such as vascular death,
`nonfatal myocardial infarction, nonfatal ischemic stroke, and
`nonfatal hemorrhagic stroke, were documented.
`After careful review of 12-month data from the pivotal
`MARINA1 and ANCHOR2 trials, the study sponsor be-
`lieved it to be in the best interest of sham group patients
`to be treated with ranibizumab. Thus, the protocol was
`amended on February 27, 2006 to provide sham-injection
`patients the opportunity to cross over to receive 0.5 mg
`ranibizumab quarterly after completing the month-12 visit
`(ie, the assessment time point for the primary analysis).
`Subsequently, after careful review of the 12-month PIER
`data, the protocol was amended again, on August 21, 2006,
`to provide all patients remaining in the study the oppor-
`tunity to roll over to receive 0.5 mg ranibizumab monthly
`for the remainder of the 2-year study. No patients were
`
`FIGURE 1. Ranibizumab for neovascular age-related macular
`degeneration trial: PIER randomization, crossover, and rollover
`scheme. The PIER study was initiated in September 2004 and
`completed in March 2007. Patients were randomized 1:1:1 to
`sham injection, 0.3 mg intravitreal ranibizumab, or 0.5 mg
`intravitreal ranibizumab. A February 2006 protocol amend-
`ment allowed sham patients to cross over to receive 0.5 mg
`intravitreal ranibizumab after completing the month-12 visit.
`An August 2006 amendment allowed all patients to roll over to
`receive 0.5 mg intravitreal ranibizumab monthly.
`
`ranibizumab treatment was maintained well into the second
`year of the study. Furthermore, switching to monthly ranibi-
`zumab treatment late in year 2 appeared to provide increased
`VA benefit to patients who had previously been treated
`quarterly, while ranibizumab treatment appeared not to pro-
`vide a VA benefit to control patients who began receiving
`ranibizumab after a year without treatment.
`
`METHODS
`
`PIER METHODOLOGY, INCLUDING STUDY DESIGN, ELIGIBIL-
`ity, masking, treatment, assessments, and analyses, has
`been published in detail.5 All patients provided informed
`written consent prior to participation. Briefly, eligible
`patients were at least 50 years of age with a diagnosis of
`primary or recurrent subfoveal CNV (predominantly clas-
`sic, minimally classic, or occult with no classic) secondary
`to AMD and baseline best-corrected VA of 20/40 to
`20/320 Snellen equivalent, measured using the Early
`Treatment Diabetic Retinopathy Study (ETDRS) chart at
`a distance of 4 meters.
`Classic and/or occult CNV comprised ⱖ50% of the total
`AMD lesion area, and the total lesion was ⱕ12 disc areas
`(DA). If a CNV lesion was minimally classic or occult with
`no classic component, the treated eye was required to meet
`protocol-defined criteria for disease progression (ie, a 10%
`
`316
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`SEPTEMBER 2010
`
`Exhibit 2087
`Page 02 of 11
`
`

`

`TABLE 1. Ranibizumab for Neovascular Age-Related
`Macular Degeneration Trial: Patient Demographics and
`Baseline Ocular Characteristicsa
`
`Gender
`Male
`Female
`Race/ethnicity
`White
`Other
`Age, years
`Mean (SD)
`Range
`Age group, years
`50–⬍65
`65–⬍75
`75–⬍85
`ⱖ85
`Prior therapy for AMD
`Any
`Laser photocoagulation
`Medication
`Supplements
`Years since first
`diagnosis of
`neovascular AMD
`
`n
`Mean (SD)
`Range
`Visual acuity (ETDRS
`letters)
`
`n
`Mean (SD)
`Range
`ⱕ 54, 20/80
`ⱖ 55, 20/80
`Visual acuity
`(approximate Snellen
`equivalent)
`Median
`20/200 or worse
`Better than 20/200 but
`worse than 20/40
`20/40 or better
`CNV lesion subtype
`n
`Predominantly classic
`Minimally classic
`Occult without classic
`Not classified
`Total area of lesion (DA)
`n
`Mean (SD)
`Range
`ⱕ4 DA
`⬎4 DA
`
`Sham
`(n ⫽ 63)
`
`0.3 mg
`(n ⫽ 60)
`
`0.5 mg
`(n ⫽ 61)
`
`20 (31.7)
`43 (68.3)
`
`26 (43.3)
`34 (56.7)
`
`28 (45.9)
`33 (54.1)
`
`59 (93.7)
`4 (6.3)
`
`57 (95.0)
`3 (5.0)
`
`56 (91.8)
`5 (8.2)
`
`77.8 (7.1)
`59–92
`
`78.7 (6.3)
`60–93
`
`78.8 (7.9)
`54–94
`
`4 (6.3)
`12 (19.0)
`36 (57.1)
`11 (17.5)
`
`35 (55.6)
`3 (4.8)
`1 (1.6)
`34 (54.0)
`
`1 (1.7)
`12 (20.0)
`37 (61.7)
`10 (16.7)
`
`35 (58.3)
`5 (8.3)
`1 (1.7)
`33 (55.0)
`
`4 (6.6)
`12 (19.7)
`31 (50.8)
`14 (23.0)
`
`33 (54.1)
`7 (11.5)
`2 (3.3)
`28 (45.9)
`
`62
`0.3 (0.5)
`0.0–3.0
`
`59
`0.7 (1.6)
`0.0–9.1
`
`61
`0.7 (1.2)
`0.0–5.0
`
`61
`60
`63
`55.1 (13.9) 55.8 (12.2) 53.7 (15.5)
`25–76
`18–79
`13–79
`25 (39.7)
`29 (48.3)
`27 (44.3)
`38 (60.3)
`31 (51.7)
`34 (55.7)
`
`20/63
`10 (15.9)
`
`20/63
`3 (5.0)
`
`20/80
`10 (16.4)
`
`42 (66.6)
`11 (17.5)
`
`49 (81.6)
`8 (13.3)
`
`36 (58.9)
`15 (24.6)
`
`63
`13 (20.6)
`30 (47.6)
`20 (31.7)
`0
`
`60
`8 (13.3)
`22 (36.7)
`29 (48.3)
`1 (1.7)
`
`61
`12 (19.7)
`19 (31.1)
`30 (49.2)
`0
`
`61
`59
`63
`4.34 (3.23) 4.36 (3.27) 4.04 (2.61)
`0.1–17.0
`0.1–20.3
`0.05–10.0
`32 (50.8)
`32 (54.2)
`31 (50.8)
`31(49.2)
`27 (45.8)
`30 (49.2)
`
`TABLE 1. (Continued )
`
`Total area of CNV (DA)
`n
`Mean (SD)
`Range
`Leakage from CNV plus
`RPE staining (DA)
`Mean (SD)
`Range
`
`Sham
`(n ⫽ 63)
`
`0.3 mg
`(n ⫽ 60)
`
`0.5 mg
`(n ⫽ 61)
`
`61
`59
`63
`3.61 (3.23) 3.77 (3.40) 3.29 (2.27)
`0.02–17.0
`0.0–20.3
`0.03–9.6
`
`4.25 (3.55) 4.47 (3.56) 3.99 (2.61)
`0.20–19.0
`0.0–22.5
`0.50–9.70
`
`AMD ⫽ age-related macular degeneration; CNV ⫽ choroidal
`neovascularization; DA⫽disc area; ETDRS ⫽ Early Treatment of
`Diabetic Retinopathy Study; RPE ⫽ retinal pigment epithelium;
`SD ⫽ standard deviation.
`aValues are n (%) except where otherwise noted.
`
`unmasked to their original treatment assignment as a result
`of the crossover and rollover amendments.
`The primary endpoint of PIER was mean change in
`best-corrected VA at month 12. Key visual outcomes at
`month 24 were mean change from baseline VA, propor-
`tion of patients who lost ⬍15 VA letters from baseline,
`proportion of patients who gained ⱖ15 VA letters from
`baseline, proportion of patients with Snellen equivalent
`VA of 20/200 or worse, mean change from baseline
`VFQ-25 near and distance activities and vision-specific
`dependency subscale scores, mean change from baseline
`total area of CNV, and total area of CNV leakage plus
`retinal pigment epithelium (RPE) staining. Safety end-
`points were incidence and severity of ocular and nonocular
`AEs, incidence of positive serum antibodies to ranibi-
`zumab, and changes in vital signs.
`The intent-to-treat approach was used for visual and
`anatomic analyses and included all patients as randomized.
`Missing values were imputed using the last-observation-
`carried-forward method. All pairwise comparisons between
`the ranibizumab groups and the sham group were based on
`statistical models with 2 groups (ranibizumab vs sham) at
`a time. A type I error management plan was used to adjust
`for multiplicity of treatment comparisons and visual and
`anatomic endpoints. Unless otherwise noted, analyses were
`stratified by CNV type at baseline (minimally classic vs
`occult with no classic vs predominantly classic), as deter-
`mined by the central reading center, and by baseline VA
`(ⱕ54 vs ⱖ55 letters). For binary endpoints, stratified
`Cochran ␹2 tests were used for between-group comparisons
`of the proportion of patients meeting the endpoint. Anal-
`ysis-of-variance and analysis-of-covariance models were
`used to analyze continuous endpoints.
`The study sample size was based on the primary endpoint
`(ie, change from baseline best-corrected VA at month 12).
`The target sample size of 180 subjects (determined by clinical
`trial simulation) provided 90% power in the intent-to-treat
`
`VOL. 150, NO. 3
`
`RANIBIZUMAB FOR NEOVASCULAR AGE-RELATED MACULAR DEGENERATION: PIER STUDY YEAR 2
`
`317
`
`Exhibit 2087
`Page 03 of 11
`
`

`

`TABLE 2. Ranibizumab for Neovascular Age-Related Macular Degeneration Trial: Patient
`Disposition and Discontinuation During 2 Years in the PIER Studya
`
`Received assigned treatment
`Completed study
`Discontinued from study
`Patient’s decision
`Patient noncompliance
`Patient’s condition mandated other therapeutic intervention
`Discontinued treatment
`Adverse event
`Patient’s decision
`Physician’s decision
`Patient’s condition mandated other therapeutic intervention
`Eligible to participate in crossover
`Crossed over and received 0.5 mg ranibizumab
`Visit at which patient crossed over to quarterly 0.5 mg
`ranibizumab
`Month 14
`Month 17
`Month 20
`Mean (SD) duration of crossover treatment, days
`Eligible to participate in rollover
`Participated in rollover amendment
`Visit at which patient rolled over to monthly 0.5 mg
`ranibizumab
`Month 19
`Month 20
`Month 21
`Month 22
`Month 23
`Mean (SD) number of rollover injections
`Randomized patients (intent-to-treat efficacy analysis)
`
`SD ⫽ standard deviation.
`aValues are n (%) except where otherwise noted.
`
`Sham
`(n ⫽ 63)
`
`62 (98.4)
`46 (73.0)
`17 (27.0)
`8 (12.7)
`1 (1.6)
`3 (4.8)
`27 (42.9)
`6 (9.5)
`7 (11.1)
`2 (3.2)
`12 (19.0)
`40 (63.5)
`39 (61.9)
`
`15 (38.5)
`17 (43.6)
`7 (17.9)
`188.3 (75.5)
`35 (55.6)
`34 (54.0)
`
`Ranibizumab
`
`0.3 mg
`(n ⫽ 60)
`
`0.5 mg
`(n ⫽ 61)
`
`59 (98.3)
`53 (88.3)
`7 (11.7)
`1 (1.7)
`2 (3.3)
`0
`11 (18.3)
`4 (6.7)
`4 (6.7)
`1 (1.7)
`2 (3.3)
`—
`—
`
`—
`—
`—
`
`61 (100.0)
`54 (88.5)
`7 (11.5)
`4 (6.6)
`1 (1.6)
`0
`10 (16.4)
`4 (6.6)
`4 (6.6)
`1 (1.6)
`1 (1.6)
`—
`—
`
`—
`—
`—
`—
`
`43 (71.7)
`43 (71.7)
`
`44 (72.1)
`44 (72.1)
`
`3 (6.8)
`3 (7.0)
`3 (8.8)
`16 (36.4)
`14 (32.6)
`14 (41.2)
`4 (9.1)
`6 (14.0)
`2 (5.9)
`5 (11.4)
`7 (16.3)
`3 (8.8)
`16 (36.4)
`13 (30.2)
`12 (35.3)
`2.5 (1.3)
`2.6 (1.3)
`2.6 (1.5)
`63 (100.0) 60 (100.0) 61 (100.0)
`
`analysis to detect a 9-letter difference between 1 or both
`ranibizumab dose groups and the sham-injection group in
`mean change from baseline VA at month 12, according to
`the Hochberg-Bonferroni criterion (assumptions based on
`results of the TAP7,8 and VIP9 trials and anticipated propor-
`tions of each CNV type).10 Safety analyses were performed
`using descriptive statistics and included all treated patients.
`All analyses were performed with SAS software (SAS Insti-
`tute Inc, Cary, North Carolina, USA).
`
`RESULTS
`
`BETWEEN SEPTEMBER 7, 2004 AND MARCH 16, 2005, 184
`patients were randomized 1:1:1 to receive sham injec-
`tion (n ⫽ 63), 0.3 mg ranibizumab (n ⫽ 60), or 0.5 mg
`ranibizumab (n ⫽ 61) at 43 US investigative sites.
`Baseline demographic and ocular characteristics were
`
`similar across treatment groups (Table 1). Groups were
`predominantly white and nearly two-thirds female, with a
`mean age of 78 years. Mean baseline VA was 53 to 56 letters
`(Snellen equivalent ⬃20/63 to 20/80) across groups.
`The first diagnosis of neovascular AMD was within the
`previous year for 87% of patients. Overall, 80% of patients
`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 injection
`group (nearly 50% vs ⬍33% of study eye lesions, respec-
`tively). Nearly 50% of the study eyes in each group had
`lesions ⱖ4 DA. The mean total area of CNV lesion and
`CNV leakage plus RPE staining at baseline was similar
`across groups.
`Forty-six of 63 (73%), 53 of 60 (88.3%), and 54 of 61
`(88.5%) patients randomized to the sham-injection, 0.3 mg,
`and 0.5 mg groups, respectively, completed the study through
`
`318
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`SEPTEMBER 2010
`
`Exhibit 2087
`Page 04 of 11
`
`

`

`TABLE 3. Ranibizumab for Neovascular Age-Related
`Macular Degeneration Trial: Mean Change From Study
`Eye Baseline Visual Acuity at Months 12 and 24 of the
`PIER Study
`
`ETDRS Letters
`
`Month 12
`Mean (SD)
`95% CIa
`P value (vs sham)b
`Month 24
`Mean (SD)
`95% CIa
`P value (vs sham)b
`
`Ranibizumab
`
`Sham
`(n ⫽ 63)
`
`0.3 mg
`(n ⫽ 60)
`
`0.5 mg
`(n ⫽ 61)
`
`⫺16.3 (22.3) ⫺1.6 (15.1) ⫺0.2 (13.1)
`⫺21.9 to ⫺10.7 ⫺5.4 to 2.3 ⫺3.5 to 3.2
`⬍.0001
`.0001
`
`⫺21.4 (21.8) ⫺2.2 (15.6) ⫺2.3 (14.4)
`⫺26.8 to ⫺15.9 ⫺6.3 to 1.8 ⫺6.0 to 1.4
`⬍.0001
`⬍.0001
`
`CI ⫽ confidence interval; ETDRS⫽ Early Treatment Diabetic
`Retinopathy Study; SD ⫽ standard deviation.
`aDerived from t distribution.
`bBased on pairwise analyses of variance adjusted for stratifi-
`cation of baseline choroidal neovascularization classification
`(minimally classic vs occult without classic vs predominantly
`classic) and baseline visual acuity (ⱕ54 vs ⱖ55 letters).
`
`month 24 (Table 2). By month 24, 48 of 184 (26.1%)
`patients had discontinued treatment (25 of 184 [13.6%] at
`month 12), usually because the patient’s condition mandated
`other therapeutic intervention.
`At the time of the February 2006 crossover amendment,
`40 of 63 (63.5%) patients in the sham-injection group who
`had not discontinued study treatment were eligible to cross
`over to receive 0.5 mg ranibizumab quarterly, and 39
`(61.9%) of those received at least 1 intravitreal injection,
`beginning at month 14. At the time of the August 2006
`rollover amendment 34 of 63 (54.0%), 43 of 60 (71.7%),
`and 44 of 61 (72.1%) patients in the sham-injection, 0.3
`mg, and 0.5 mg groups, respectively, who had not discon-
`tinued study treatment or completed the month-24 visit,
`rolled over to receive 0.5 mg ranibizumab monthly, begin-
`ning month 19. Results are presented according to group
`assignment at randomization and include post-crossover
`(sham) and post-rollover (sham, 0.3 mg, 0.5 mg) data.
`At month 24, VA had decreased from baseline an
`average of 21.4 letters in the sham group, 2.2 letters in
`the 0.3 mg group, and 2.3 letters in the 0.5 mg group (P
`⬍ .0001 each ranibizumab dose vs sham), with about a
`19-letter difference between sham-group and treated
`patients. The group differences at month 24 were similar
`to those at month 12 (Table 3). At month 24, 47 of 60
`(78.2%) patients in the 0.3 mg group and 50 of 61
`(82.0%) of patients in the 0.5 mg group had lost ⬍15
`letters
`from baseline VA compared with 26 of 63
`(41.3%) sham-injection patients (P ⬍ .0001 each
`ranibizumab dose vs sham) (Figure 2); and 21 of 63
`(33.3%) patients in the sham group had lost ⱖ30 VA
`
`letters from baseline. Such severe vision loss was un-
`common (⬃3.0%) in patients who were originally
`randomized to ranibizumab treatment groups. Ranibi-
`zumab groups did not differ significantly from the sham
`group in the proportion of patients who gained ⱖ15 VA
`letters: 3 of 63 (4.8%) in the sham-injection group, 9 of
`60 (15.0%) in the 0.3 mg group, and 5 of 61 (8.2%) in
`the 0.5 mg group.
`A Snellen equivalent VA of 20/200 or worse was more
`common in the sham-injection group (55.6%) than in the
`0.3 mg (25.0%) and 0.5 mg (27.9%) ranibizumab groups
`(P ⬍ .0001 for 0.3 mg vs sham; P ⫽ .0013 for 0.5 mg vs
`sham). The 0.3 mg and 0.5 mg ranibizumab groups did not
`differ significantly from the sham group on the near
`activities, distance activities, and vision-specific depen-
`dency VFQ-25 subscales.
`Subgroup analyses of the mean change from baseline
`VA at month 24 were performed for several baseline
`characteristics, including age (⬍75 years vs ⱖ75 years),
`gender, race (white vs other), VA (⬍54 vs ⱖ54), lesion
`size (ⱕ4 DA vs ⬎4 DA), presence of occult CNV (yes vs
`no), and prior laser photocoagulation (yes vs no). The
`treatment effects of the ranibizumab groups compared with
`the sham-injection group were consistent with the overall
`results for all subgroups except race and prior photocoag-
`ulation, for which the sample sizes were too small to draw
`conclusions (data not shown).
`At month 24 total area of CNV had increased from
`baseline an average of 1.90 DA in the sham group, 0.29 DA
`in the 0.3 mg group, and 0.64 DA in the 0.5 mg group (P ⫽
`.0015 0.3 mg vs sham, P ⫽ .0021 0.5 mg vs sham) (Table 4).
`The total area of CNV leakage plus RPE staining decreased
`from baseline an average of 0.78, 1.52, and 1.22 DA in the
`sham-injection, 0.3 mg, and 0.5 mg groups, respectively (P ⱖ
`.20 for each ranibizumab group vs sham).
`
`● CROSSOVER: Thirty-nine of 40 eligible sham-injection
`group patients crossed over to 0.5 mg quarterly ranibi-
`zumab, beginning month 14 (38.5%), 17 (43.6%), or 20
`(17.9%) (Table 2), and received a mean of 4.1 ⫾ 1.7
`injections from the time of crossover to study discontinu-
`ation or completion. On average, VA of sham-injection
`patients who crossed over (and subsequently rolled over)
`to ranibizumab treatment during study year 2 continued to
`decrease until study completion or discontinuation, with
`an average loss of 3.5 letters 10 months after crossover
`(Figure 3). Small sample sizes and variations in treatment
`time and dose prevented formal statistical analyses of the
`post-crossover data.
`
`● ROLLOVER: Thirty-four, 43, and 44 eligible patients in
`the sham, 0.3 mg, and 0.5 mg groups, respectively, rolled
`over to receive monthly 0.5 mg ranibizumab, beginning
`month 19 (Table 2). Patients in the sham, 0.3 mg, and 0.5
`mg groups received an average of 2.6, 2.6, and 2.5
`intravitreal injections, respectively, from the time of roll-
`
`VOL. 150, NO. 3
`
`RANIBIZUMAB FOR NEOVASCULAR AGE-RELATED MACULAR DEGENERATION: PIER STUDY YEAR 2
`
`319
`
`Exhibit 2087
`Page 05 of 11
`
`

`

`FIGURE 2. Ranibizumab for neovascular age-related macular degeneration trial: visual acuity outcomes at months 12 and 24 of the PIER
`study. Percentage of patients in each of the 3 treatment groups who (Top) lost <15 ETDRS letters, (Middle) gained >15 ETDRS letters,
`or (Bottom) had 20/200 or worse Snellen equivalent VA at months 12 and 24. Post-crossover and post-rollover data are included. Error
`bars are ⴞ1 standard error of the mean. Numbers below the bars in each graph are the number of patients for whom visual acuity data
`were available in the corresponding group. ETDRS ⴝ Early Treatment Diabetic Retinopathy Study; *P < .0002 vs sham.
`
`over to study discontinuation or completion. On average,
`VA of patients in ranibizumab groups who rolled over to
`monthly treatment with 0.5 mg ranibizumab increased
`across the first 4 rollover injections, with an average gain
`of 2.2 and 4.1 letters in the 0.3 mg and 0.5 mg groups 4
`months after rollover, respectively (Figure 4). Small sam-
`ple sizes and group differences (ie, sham-group VA at the
`time of rollover differed from that of ranibizumab-treated
`patients) prevented formal statistical comparisons of the
`post-rollover data.
`
`● SAFETY: Key safety results through month 24 are sum-
`marized in Table 5. During the 2-year study period, safety
`was observed for an average of 626.4 (⫾ 197.1), 712.6 (⫾
`43.1), and 689.8 (⫾ 108.0) days for the sham, 0.3 mg, and
`0.5 mg groups, respectively. Ocular AEs that occurred at a
`
`rate ⱖ10% in ranibizumab-treated patients compared with
`pre-crossover sham-injection patients were conjunctival
`hemorrhage (29% [18/62] of
`sham-injection patients,
`50.8% [30/59] of 0.3 mg patients, 52.5% [32/61] of 0.5 mg
`patients) and increased intraocular pressure (4.8% [3/62],
`23.7% [14/59], and 31.1% [19/61], respectively). No inci-
`dents of endophthalmitis were reported. The rate of
`intraocular inflammation was low, with no notable differ-
`ence across groups: 2 of 62 (3.2%) in the sham-injection
`group, 3/59 (5.1%) in the 0.3 mg group, and 3 of 61 (4.9%)
`in the 0.5 mg group. No serious intraocular inflammation
`was reported during the study. Five of 62 (8.1%) patients
`in the pre-crossover sham-injection group, 5 of 59 (8.5%)
`patients in the 0.3 mg group, and 11 of 61 (18.0%) patients
`in the 0.5 mg group had cataract, nuclear cataract, or
`cortical cataract in the study eye.
`
`320
`
`AMERICAN JOURNAL OF OPHTHALMOLOGY
`
`SEPTEMBER 2010
`
`Exhibit 2087
`Page 06 of 11
`
`

`

`TABLE 4. Ranibizumab for Neovascular Age-Related Macular Degeneration Trial: Choroidal
`Neovascularization–Related Changes From Baseline at Months 12 and 24 of the PIER Study
`
`Change in total area of CNV (DA)
`Month 12
`Mean (SD)
`95% CIa
`P value (vs sham)b
`Month 24
`Mean (SD)
`95% CIa
`P value (vs sham)b
`Change in total area of CNV leakage ⫹ RPE
`staining (DA)
`Month 12
`Mean (SD)
`95% CIa
`P value (vs sham)b
`Month 24
`Mean (SD)
`95% CIa
`P value (vs sham)b
`
`Sham
`(n ⫽ 63)
`
`Ranibizumab
`
`0.3 mg (n ⫽ 59)
`
`0.5 mg (n ⫽ 61)
`
`0.18 (2.13)
`2.08 (2.66)
`1.41 to 2.75 ⫺0.37 to 0.74
`.0001
`
`0.43 (1.86)
`⫺0.04 to 0.91
`.0002
`
`0.29 (2.73)
`1.90 (2.46)
`1.28 to 2.52 ⫺0.42 to 1.00
`.0015
`
`0.64 (2.16)
`0.08 to 1.19
`.0021
`
`⫺1.29 (2.48)
`⫺1.41 (2.69)
`1.40 (3.77)
`0.45 to 2.35 ⫺2.12 to ⫺0.71 ⫺1.93 to ⫺0.66
`.0001
`.0001
`
`⫺1.22 (2.74)
`⫺1.52 (2.99)
`⫺0.78 (4.13)
`⫺1.82 to 0.26 ⫺2.30 to ⫺0.75 ⫺1.92 to ⫺0.52
`.273
`.199
`
`CI ⫽ confidence interval; CNV ⫽ choroidal neovascularization; DA ⫽ disc areas; RPE ⫽ retinal
`pigment epithelium; SD ⫽ standard deviation.
`The last-observation-carried-forward method was used to impute missing data. Strata were
`defined using baseline CNV classification (minimally classic vs occult without classic) and baseline
`visual acuity score (ⱕ54 vs ⱖ55 letters).
`aDerived from t distributions.
`bBased on pairwise ANCOVA models adjusted for the 2 stratification factors and baseline value of
`the endpoint.
`
`Nonocular AEs were experienced by 48 of 62 (77.4%) of
`patients in the pre-crossover sham-injection group, 50 of
`59 (84.7%) patients in the 0.3 mg group, and 53 of 61
`(86.9) patients in the 0.5 mg group. Nonocular AEs
`included urinary tract infection, nasopharyngitis, constipa-
`tion, and upper respiratory tract infection, all of which
`might be expected in an elderly population.
`Nonocular AEs known to be associated with systemic
`VEGF inhibition were of particular interest. Hypertension
`AEs occurred in 7 of 62 (11.3%) pre-crossover sham-injec-
`tion patients, 6 of 59 (10.2%) 0.3 mg patients, and 13 of 61
`(21.3%) 0.5 mg patients for 0.5 mg patients (Fisher exact ␹2,
`2-sided P ⫽ .13 for 0.3 mg vs 0.5 mg; P ⫽ .15 for 0.5 mg vs
`pre-crossover sham). One of 61 (1.6%) patients in the 0.5 mg
`group had a serious hypertension AE.
`The incidence of nonocular hemorrhage was higher in
`the 0.3 mg and 0.5 mg groups (4 of 59 [6.8%] and 6 of 61
`[9.8%], respectively) compared with the pre- and post-
`crossover sham-injection group (3 of 62 [4.8%] and 1 of 39
`[2.6%], respectively).
`The rate of Antiplatelet Trialists’ Collaboration ATEs
`(vascular deaths, nonfatal myocardial infarctions, nonfatal
`ischemic strokes, and nonfatal hemorrhagic strokes) during
`
`the 2-year treatment period was 1.6% (1 of 62 patients) in the
`pre-crossover sham group, in the post-crossover sham group,
`1.7% (1 of 59 patients) in the 0.3 mg group, and 0% in the
`0.5 mg group.
`
`DISCUSSION
`
`RANIBIZUMAB, ADMINISTERED MONTHLY FOR 3 MONTHS AND
`then quarterly, provided a VA benefit to patients with all
`angiographic subtypes of CNV compared with sham injec-
`tions. Because ranibizumab-group patients switched from
`quarterly to monthly ranibizumab treatment beginning at
`month 19, it was not possible to assess the VA benefits of
`quarterly dosing for the entire 24-month study. However, the
`VA differences between sham and ranibizumab group pa-
`tients that were observed at the primary month-12 analysis
`were maintained prior to the first rollover treatment, indicat-
`ing that the VA benefit of quarterly ranibizumab persisted
`well into the second year of the study.
`The finding that VA scores tended to increase after
`patients in the ranibizumab treatment groups rolled over to
`receive monthly ranibizumab suggests that, in order to
`
`VOL. 150, NO. 3
`
`RANIBIZUMAB FOR NEOVASCULAR AGE-RELATED MACULAR DEGENERATION: PIER STUDY YEAR 2
`
`321
`
`Exhibit 2087
`Page 07 of 11
`
`

`

`FIGURE 3. Ranibizumab for neovascular age-related macular
`degeneration trial: mean change in visual acuity for sham-group
`patients who crossed over to receive quarterly injections of 0.5 mg
`ranibizumab (and subsequently rolled over to monthly 0.5 mg
`ranibizumab). Numbers below the x-axis are the number of sham
`patients for whom visual acuity data were available at the corre-
`sponding month post crossover. Error bars are ⴞ1 standard error of
`the mean. ETDRS ⴝ Early Treatment Diabetic Retinopathy Study.
`
`obtain the greatest benefit from ranibizumab treatment,
`some patients may require more frequent dosing. This is
`supported by the finding that the degree of VA benefit
`obtained with quarterly dosing in the PIER study was not
`as robust as that obtained with monthly dosing in the
`ANCHOR and MARINA studies. In those studies, pa-
`tients who received monthly injections of ranibizumab
`experienced a gain of 5 to 11 letters from baseline at
`month 24 compared to a loss of approximately 2 letters
`with the PIER dosing regimen; and between 25% and 41%
`of ANCHOR and MARINA patients gained ⱖ15 letters
`from baseline at month 24 compared to ⱕ15% in the PIER
`study. A current challenge is to determine a dosing
`regimen that will prove optimal for physicians and patients
`while realizing the full benefit of ranibizumab (eg, Brown
`and Regillo, 2007; Fung and associates, 2007).11,12
`At month 24 (ie, 5 months after patients began rollover
`to receive 0.5 mg ranibizumab monthly), VA scores tended
`to increase in the sham group and decrease in the 0.5 mg
`group, while remaining stable in the 0.3 mg group. The
`small number of evaluable patients at that time point (ie,
`n ⫽ 3 for each group) prevents interpretation of the
`observation.
`Sham-group patients who began receiving ranibizumab
`during study year 2 continued to experience a loss of VA
`relative to baseline. While the extent of any benefit of
`ranibizumab treatment to sham-group patients was not
`determinable (since it is not known how sham-injection
`VA would have changed without crossover/rollover treat-
`ment), the implication is that ranibizumab had limited
`
`FIGURE 4. Ranibizumab for neovascular age-related macular
`degeneration trial: mean change in visual acuity for patients
`who rolled over to receive monthly injections of 0.5 mg ranibi-
`zumab. Numbers below the x-axis are the number of patients for
`whom visual acuity data were available at the corresponding
`month post rollover. Note the small number of patients in each
`group who were assessed at month 5 post rollover. Error bars
`are ⴞ1 standard error of the mean. ETDRS ⴝ Early Treatment
`Diabetic Retinopathy Study.
`
`benefit in neovascular AMD patients after ⱖ12 months
`without treatment. Given that baseline disease character-
`istics were similar across all randomized patients, it is likely
`that sham-group patients exp