SHORT-TERM SAFETY AND EFFICACY
`OF INTRAVITREAL BEVACIZUMAB
`(AVASTIN) FOR NEOVASCULAR AGE-
`RELATED MACULAR DEGENERATION
`
`RYAN M. RICH, MD, PHILIP J. ROSENFELD, MD, PHD,
`CARMEN A. PULIAFITO, MD, MBA, SANDER R. DUBOVY, MD,
`JANET L. DAVIS, MD, HARRY W. FLYNN JR., MD,
`SERAFIN GONZALEZ, PHARMD, WILLIAM J. FEUER, MS,
`RICHARD C. LIN, MD, PHD, GEETA A. LALWANI, MD,
`JACKIE K. NGUYEN, MD, GAURAV KUMAR, BA
`
`Purpose: To evaluate the safety and efficacy of intravitreal bevacizumab (Avastin, Genen-
`tech Inc.) for the treatment of neovascular age-related macular degeneration (ARMD).
`Methods: A retrospective review was performed on consented patients with neovas-
`cular ARMD receiving intravitreal bevacizumab therapy. All patients received intravitreal
`bevacizumab at baseline with additional monthly injections given at the discretion of the
`treating physician. At each visit, a routine Snellen visual acuity assessment was performed
`followed by an ophthalmic examination and optical coherence tomography (OCT) imaging.
`Results: Fifty–three eyes of 50 patients received an intravitreal bevacizumab injection
`between May and August 2005. Including the month 3 visit, the average number of
`injections was 2.3 out of a maximum of 4 injections. No serious drug-related ocular or
`systemic adverse events were identified. Improvements in visual acuity and central retinal
`thickness measurements were evident by week 1 and continued through month 3. At
`month 3, the mean visual acuity improved from 20/160 to 20/125 (P⬍0.001) and the mean
`central retinal thickness decreased by 99.6 ␮m (P⬍0.001).
`Conclusion: Off-label intravitreal bevacizumab therapy for neovascular ARMD was well
`tolerated over 3 months with improvements in visual acuity and OCT central retinal
`thickness measurements. While the long-term safety and efficacy of intravitreal bevaci-
`zumab remain unknown, these short-term results suggest that intravitreal bevacizumab
`may be the most cost effective therapy for the treatment of neovascular ARMD.
`RETINA 26:495–511, 2006
`
`From Bascom Palmer Eye Institute, Department of Ophthalmol-
`ogy, University of Miami School of Medicine, Florida.
`Presented in part at the Retina Subspecialty Day during the
`American Academy of Ophthalmology meeting; Chicago, Illinois;
`October 14, 2005.
`No financial support was received from Genentech, Inc. to perform
`this retrospective review. Carmen A. Puliafito, MD, MBA, is listed on a
`
`patent for optical coherence tomography and receives royalties.
`Supported by the Department of Ophthalmology at the Bascom
`Palmer Eye Institute, Miami, Florida, and by an unrestricted grant
`from Research to Prevent Blindness, Inc., New York, New York.
`Reprint requests: Philip J. Rosenfeld, MD, PhD, Bascom Palmer
`Eye Institute, University of Miami School of Medicine, 900 N.W.
`17th Street, Miami, FL 33136; e-mail: prosenfeld@med.miami.edu
`
`495
`
`Novartis Exhibit 2284.001
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`496
`
`RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ● 2006 ● VOLUME 26 ● NUMBER 5
`
`Vascular endothelial growth factor-A (VEGF) has
`
`been implicated as the major angiogenic stimulus
`responsible for neovascularization in age-related mac-
`ular degeneration (ARMD).1– 6 The first anti-VEGF
`drug approved by the Food and Drug Administration
`(FDA) for the treatment of neovascular ARMD is
`pegaptanib sodium (MACUGEN, Eyetech/OSI Phar-
`maceuticals).7 Pegaptanib binds and inhibits the ex-
`tracellular isoforms of VEGF that are at least 165
`amino acids in length.8 In the phase III clinical trial
`using pegaptanib for the treatment of neovascular
`ARMD, repeated intravitreal injections of pegaptanib
`every 6 weeks slowed the rate of vision loss compared
`with a sham-injected control group after 1 year.7 Among
`the patients receiving pegaptanib therapy (0.3 mg) in this
`study, only 6% had significant vision improvement
`compared with 2% in the sham-treated group.
`In contrast to pegaptanib therapy, injections with
`a different anti-VEGF drug known as ranibizumab
`(LUCENTIS, Genentech Inc.) resulted in vision im-
`provement in phase III clinical trials for the treatment
`of neovascular ARMD. 9 –11 Ranibizumab was shown
`to improve average visual acuity with 95% of patients
`having stable or improved visual acuity after receiving
`monthly injections for at least 1 year. In addition, 40%
`of patients achieved a level of visual acuity of at least
`20/40. The superior results using ranibizumab com-
`pared with pegaptanib can most likely be explained by
`the differences in how the two drugs bind and inhibit
`VEGF. Unlike pegaptanib, ranibizumab binds all the
`biologically active forms of VEGF, including the iso-
`forms and proteolytic breakdown products that con-
`tain fewer than 165 amino acids.12,13
`Bevacizumab (AVASTIN, Genentech, Inc.), a full-
`length, humanized monoclonal antibody against
`VEGF, also binds and inhibits all the biologically
`active forms of VEGF much like ranibizumab.14 –16
`The similar VEGF binding properties of bevacizumab
`and ranibizumab can be explained by their common
`molecular lineage. Both drugs are proteins that were
`genetically modified from the same murine monoclo-
`nal antibody against VEGF. The two proteins differ in
`their size and affinity for VEGF. While bevacizumab
`is a humanized, murine full-length antibody with two
`binding sites for VEGF, ranibizumab is a humanized,
`murine antigen binding fragment (Fab) with only a
`single affinity-matured binding site for VEGF.16 The
`single binding site of ranibizumab has a dissociation
`constant (Kd) for VEGF of approximately 0.140 nM
`compared with an overall Kd for bevacizumab of
`approximately 0.5 nM to 1.0 nM.14 –16
`Ranibizumab is not yet approved by the FDA, but
`bevacizumab is approved for the intravenous treat-
`ment of metastatic colorectal cancer. When bevaci-
`
`zumab was approved by the FDA in February 2004,
`we initiated the Systemic Avastin for Neovascular
`ARMD Study or SANA Study to investigate the use of
`systemic bevacizumab for the treatment of neovascu-
`lar ARMD. Our first cohort of 9 patients were ob-
`served for 12 weeks and had improved visual acuity
`with resolution of leakage from their neovascular le-
`sions after just two or three doses of bevacizumab (5
`mg/kg).17 Except for a mild elevation in blood pres-
`sure that was easily controlled with antihypertensive
`medications, no other adverse events were identified.
`This study was expanded to include a second cohort of
`9 patients and all 18 patients have been followed
`through 24 weeks with similar results.18 While no
`additional adverse events were identified after 6
`months, there was always a concern that high dose
`systemic therapy with bevacizumab (5 mg/kg) could
`result in an increased risk of thromboembolic events,
`the most frequent life-threatening drug-related adverse
`event associated with bevacizumab therapy in cancer
`patients and of particular concern in the older ARMD
`population.19
`One way to decrease the potential risk of drug-
`related adverse events would be to decrease the dose
`of bevacizumab, and one way to decrease the dose
`would be to inject a small amount of drug directly into
`the eye. Bevacizumab was thought to be too large to
`penetrate the retina, a presumed requirement for any
`drug intended to treat neovascularization under the
`retina.20 However, upon review of the literature, we
`found that bevacizumab was never tested to determine
`if it could penetrate the retina. Instead, a different
`antibody against an antigen known as HER-2 was
`used in those penetration experiments, and HER-2 is
`expressed in the inner retina and may have inhibited
`penetration from the vitreal cavity through the reti-
`na.20 More importantly, the basic premise that retinal
`penetration was a requirement for the treatment of
`choroidal neovascularization was never tested. Intra-
`vitreal bevacizumab was never injected in any animal
`model of choroidal neovascularization. Even if retinal
`penetration was important for efficacy, it was unclear
`how much retinal penetration was necessary to achieve
`efficacy and whether penetration through a diseased hu-
`man retina would be the same as the penetration through
`a normal animal retina. We subsequently learned that
`Han et al had shown that a full-length antibody was
`capable of penetrating a normal rabbit retina.21
`The full-length murine precursor of bevacizumab
`was shown to be effective for the treatment of iris
`neovascularization in an animal model of neovascular
`glaucoma.22 When cynomolgus monkey eyes received
`multiple injections of a murine anti-VEGF antibody,
`not only was iris neovascularization prevented but
`
`Copyright© by Ophthalmic Communications Society, Inc. Unauthorized reproduction of this article is prohibited.
`
`Novartis Exhibit 2284.002
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`INTRAVITREAL BEVACIZUMAB FOR NEOVASCULAR ARMD ● RICH ET AL
`
`497
`
`Table 1. Changes in Central Retinal Thickness through 3 Months
`
`Baseline
`Central
`Retinal
`Thickness
`(␮m), n ⫽ 53
`Eyes
`
`Median
`P value*
`Mean
`P value†
`
`313
`
`351
`
`Week 1
`Central
`Retinal
`Thickness
`(␮m), n ⫽ 32
`Eyes
`
`249
`P ⬍ 0.001
`260.7
`P ⬍ 0.001
`
`Month 1
`Central
`Retinal
`Thickness
`(␮m), n ⫽ 51
`Eyes
`
`242
`P ⬍ 0.001
`253.5
`P ⬍ 0.001
`
`Month 2
`Central
`Retinal
`Thickness
`(␮m), n ⫽ 42
`Eyes
`
`232
`P ⬍ 0.001
`240.2
`P ⬍ 0.001
`
`Month 3
`Central
`Retinal
`Thickness
`(␮m), n ⫽ 53
`Eyes
`
`228
`P ⬍ 0.001
`251.4
`P ⬍ 0.001
`
`Decrease in
`Central Retinal
`Thickness (␮m)
`from Baseline to
`Month 3
`
`⫺85
`
`⫺99.6
`
`*Paired Wilcoxon signed rank test.
`†Paired Student t test.
`
`there was no inflammation in this cross-species exper-
`iment. This study provided support for the idea that an
`anti-VEGF antibody could be injected into the eye
`without causing ocular complications even across dif-
`ferent species.
`Even though there were limited data available, we
`offered intravitreal bevacizumab, a humanized murine
`monoclonal antibody, to a patient with neovascular
`ARMD who had failed verteporfin photodynamic
`therapy (PDT) and pegaptanib therapy and was con-
`tinuing to lose vision.23 One week after the bevaci-
`zumab injection, OCT revealed dramatic improve-
`ment in the central retinal thickness of the macula, and
`1 month after the injection, fluorescein angiography
`showed no evidence of leakage from the neovascular
`lesion. The response to intravitreal bevacizumab was
`very similar to the responses previously seen with
`intravenous bevacizumab and with intravitreal ranibi-
`zumab in the early phase I/II studies.24,25 Based on our
`preliminary experience with intravitreal bevacizumab,
`we began to offer off-label intravitreal injections of
`bevacizumab primarily to patients with neovascular
`ARMD who were losing vision despite receiving
`FDA-approved therapies such as PDT and pegaptanib
`therapy. This retrospective report describes our initial
`3-month experience using intravitreal injections of
`bevacizumab for the treatment of neovascular ARMD
`at the Bascom Palmer Eye Institute.
`
`Patients and Methods
`
`Approval for this retrospective review was obtained
`from the Institutional Review Board (IRB)/Ethics
`Committee at the University of Miami School of Med-
`icine. All patients signed an informed consent to par-
`ticipate in this retrospective review. To be eligible for
`this retrospective review, patients had received an
`intravitreal injection of bevacizumab as part of their
`routine clinical care for the treatment of neovascular
`ARMD at the Bascom Palmer Eye Institute. Intravit-
`
`real bevacizumab was primarily offered to patients
`who were losing vision while undergoing treatment
`with FDA-approved therapies for neovascular ARMD
`or as primary therapy only after a thorough discussion
`of all their therapeutic options. All patients had evi-
`dence of increased 1 mm central retinal thickness as
`determined by optical coherence tomography (Stratus
`OCT, Version 4.0.2, Carl Zeiss Meditec, Dublin, CA).
`This increased central retinal thickness consisted of
`subretinal fluid and/or cystic changes within the ret-
`ina. Before each injection of bevacizumab, patients
`signed a standard institutional consent describing the
`potential risks and benefits of treatment. Although
`there were no formal exclusion criteria, patients with
`a history of uncontrolled hypertension and recent
`thromboembolic events were not usually injected with
`bevacizumab, but this decision was at the discretion of
`the treating physician.
`At each visit, patients underwent Snellen visual
`acuity measurements according to the procedures fol-
`lowed by individual physicians within their practice.
`An attempt was made to obtain best-corrected visual
`acuity at each visit; however, visual acuity measure-
`ments were not standardized and were performed as
`part of routine clinical care. At each visit, an ophthal-
`mic examination was performed consisting of a slit-
`lamp evaluation and a biomicroscopic fundus exami-
`nation. Ocular
`imaging consisted of fluorescein
`angiography and/or OCT at the time of the first bev-
`acizumab injection and at each follow-up visit. Most
`patients did undergo OCT imaging at each visit (Table
`1). OCT imaging consisted of 6-diagonal fast, low-
`density (low resolution, 128 a-scans per diagonal) 6
`mm scans and 6-diagonal slow, high-density (high
`resolution, 512 a-scans per diagonal) 6 mm scans
`performed at 30 degree intervals. The 1 mm central
`retinal thickness measurements were determined from
`the fast macular thickness maps calculated from the 6
`low resolution diagonal scans. The 6 high-density,
`
`Novartis Exhibit 2284.003
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`498
`
`RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ● 2006 ● VOLUME 26 ● NUMBER 5
`
`high resolution radial diagonal scans were used to
`qualitatively evaluate the macula and to determine if
`retreatment was needed.
`In May 2005, when the off-label use of intravitreal
`bevacizumab was initiated, a patient’s blood pressure
`was not routinely measured before an intravitreal in-
`jection at the Bascom Palmer Eye Institute. This pol-
`icy was subsequently changed so that all patients
`receiving an intravitreal injection of any drug under-
`went blood pressure monitoring. However, no stan-
`dard protocol for measuring blood pressure was im-
`plemented. As a result, blood pressure was measured
`manually or by using an automated blood pressure
`monitor. Only one measurement was routinely taken
`just before injection. Patients with systolic blood pres-
`sures above 150 mmHg and diastolic blood pressures
`above 90 mmHg were routinely referred to their in-
`ternists for further evaluation and management.
`All intravitreal injections were performed using a
`standard protocol at the Bascom Palmer Eye Institute.
`Preinjection antibiotic drops were not routinely used.
`The eye was topically anesthetized and a povidone-
`iodine (10%) scrub was performed on the lids and
`lashes. A sterile speculum was placed between the
`lids. Povidone-iodine (5%) drops were then applied
`over the ocular surface three times several minutes
`apart. Additional topical anesthesia was achieved by
`applying a sterile cotton swab soaked in sterile 4%
`lidocaine to the area designated for injection in the
`inferotemporal quadrant. Bevacizumab (0.05 mL;
`1.25 mg) in a tuberculin syringe with a 30-gauge
`needle was then injected through the pars plana into
`the vitreous cavity inserted through the sclera 3 to 4
`mm posterior to the limbus. Postinjection light per-
`ception was assessed and the intraocular pressure
`(IOP) was monitored until the IOP was below 30
`mmHg. After the injection, the patient was instructed
`to apply topical antibiotics to the injected eye 4 times
`a day for 3 days. All patients received a call within 24
`hours to assess their status and remind them to take
`their antibiotic drops.
`The timing of the postinjection follow-up visits and
`the need for retreatment with intravitreal bevacizumab
`were at the discretion of the treating physician; however,
`most patients were seen monthly and injected if there
`was evidence of cystic maculopathy and/or subretinal
`fluid as determined by OCT imaging. Some patients also
`underwent a safety visit 1 week after the first injection.
`
`Preparation of Bevacizumab for Intravitreal
`Injection
`
`Bevacizumab is commercially available as a solu-
`tion (100 mg; 25 mg/mL) and was not diluted, recon-
`
`stituted, or altered in any way. All manipulations of
`bevacizumab were performed using proper aseptic
`technique under a laminar flow hood (IS0 Class 5) by
`a licensed and registered pharmacy in compliance
`with United States Pharmacopeia (USP) guidelines in
`Chapter 797. The vial of bevacizumab was punctured
`with a device called a Mini-Spike Dispensing Pin with
`Security Clips (B-Braun, catalog # DP-1000SC). The
`use of this or a similar device is recommended be-
`cause bevacizumab is available in a single-use vial so
`entering the vial multiple times is not recommended.
`While this device contains a bacterial retentive air-
`venting filter, it does not actually filter the drug itself,
`and filtering of the drug was not performed due to the
`possibility that the protein could stick to the filter.
`Approximately 0.12 mL of bevacizumab was drawn
`up into multiple 1 mL polypropylene tuberculin sy-
`ringes (Becton Dickinson & Co., Franklin, NJ, Ref
`# 309602, NDC # 08290309602). A sterile cap was
`then placed on each syringe. The syringes were then
`labeled, placed in light-resistant brown bags to protect
`bevacizumab from light, and stored in a refrigerator at
`2 to 8 ˚C until used for injection. Approximately 25
`syringes were prepared from each 4 mL vial of bev-
`acizumab. At least two syringes were submitted to the
`Microbiology Department for sterility and endotoxin
`testing. A 14-day expiration was assigned to each
`syringe based on USP Chapter 797 for a low-risk,
`refrigerated preparation. A 14-day expiration date was
`considered to be a conservative estimate of stability
`because the drug was known to be stable in its original
`glass vial for 18 months. If all testing results were
`negative, then the syringe was relabeled to have a
`90-day expiration date; however, this practice was
`abandoned because the long-term stability of bevaci-
`zumab in syringes was unknown and the syringes
`were routinely used within the original 14-day expi-
`ration date. The syringe containing bevacizumab was
`removed from the refrigerator before injection and did
`not remain at room temperature for longer than 30
`minutes. Before injection, a sterile standard 30-gauge
`needle (5/8 inch) was placed on the syringe, and the
`plunger was advanced to 0.05 mL (50 ␮L) so that all
`the dead space was removed. Stability information for
`the drug in the syringes is not currently available.
`
`Statistical Analysis
`
`For purposes of statistical analysis, all Snellen vi-
`sual acuity data were converted to an equivalent letter
`score from a standard 2-m protocol using an Early
`Treatment Diabetic Retinopathy chart.26 Data were
`statistically analyzed using the paired Student t test for
`changes in mean blood pressure measurements, visual
`
`Copyright© by Ophthalmic Communications Society, Inc. Unauthorized reproduction of this article is prohibited.
`
`Novartis Exhibit 2284.004
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`INTRAVITREAL BEVACIZUMAB FOR NEOVASCULAR ARMD ● RICH ET AL
`
`499
`
`Table 2. Change in Visual Acuity through 3 Months
`
`Baseline Visual
`Acuity Letters/
`Snellen
`Equivalent,
`n ⫽ 53 Eyes
`
`Week 1 Visual
`Acuity Letters/
`Snellen
`Equivalent,
`n ⫽ 32 Eyes
`
`Month 1 Visual
`Acuity Letters/
`Snellen
`Equivalent,
`n ⫽ 51 Eyes
`
`Month 2 Visual
`Acuity Letters/
`Snellen
`Equivalent,
`n ⫽ 42 Eyes
`
`Month 3 Visual
`Acuity Letters/
`Snellen
`Equivalent,
`n ⫽ 53 Eyes
`
`Median
`P value*
`
`Mean
`P value†
`
`35
`20/200
`
`38.1
`20/160-1
`
`35
`20/200
`P ⫽ 0.003
`43.8
`20/125-1
`P ⫽ 0.005
`
`35
`20/200
`P ⬍ 0.001
`44.1
`20/125-1
`P ⬍ 0.001
`
`55
`20/80
`P ⫽ 0.001
`46.9
`20/125⫹2
`P ⫽ 0.001
`
`50
`20/100
`P ⬍ 0.001
`46
`20/125⫹1
`P ⬍ 0.001
`
`Change in Visual
`Acuity from
`Baseline to
`Month 3 in
`Letters
`⫹15
`
`⫹7.9
`
`*Paired Wilcoxon signed rank test.
`†Paired Student t test.
`
`acuity letter scores, and central retinal thickness mea-
`surements at week 1 through month 3 compared with
`mean baseline values. Median measurements at week
`1 through month 3 were compared with median base-
`line values using the paired Wilcoxon signed rank test.
`Systolic blood pressure values were analyzed sepa-
`rately from the diastolic values. Statistical significance
`was defined as P ⬍ 0.05.
`
`Results
`
`Baseline Characteristics
`
`A total of 53 eyes from 50 consecutively consented
`patients received an initial intravitreal injection of
`bevacizumab during the period from May 2005 to
`August 2005. Three patients had bevacizumab in-
`jected into both eyes. These 50 patients had a mean
`age of 78 years and a median age of 80 years (range:
`62 to 91 years). There were 28 women (56%). Of the
`53 eyes, 40 eyes (75%) had received some prior ther-
`apy before receiving an intravitreal injection of bev-
`acizumab. Prior therapy consisted of PDT in 23 eyes,
`pegaptanib therapy in 33 eyes, and PDT progressing
`to pegaptanib therapy in 16 eyes. Thirteen eyes (25%)
`received intravitreal bevacizumab as primary therapy.
`The baseline median and mean visual acuity and OCT
`central retinal thickness measurements are shown in
`Tables 1 and 2. At baseline, median and mean visual
`
`acuity measurements were 20/200 and 20/160, respec-
`tively. Baseline median and mean 1 mm central retinal
`thickness measurements were 313 ␮m and 351 ␮m,
`respectively. Blood pressure measurements were per-
`formed on only 20 patients at baseline. The median
`and mean systolic/diastolic blood pressure values
`were the same at 130/80.
`
`Safety
`
`Table 3 summarizes the number of injections per-
`formed during the first 3 months. A total of 123
`injections of bevacizumab were performed. On aver-
`age, an eye received 2.3 injections out of the maxi-
`mum of 4 injections if a patient had received an
`injection at baseline, month 1, month 2, and month 3.
`There were no episodes of inflammation or severe
`vision decrease immediately after an injection. There
`were no cases of endophthalmitis, retinal detachment,
`or lens damage. During the 3 months, there were no
`thromboembolic events which included cerebrovascu-
`lar accidents, transient ischemic attacks, myocardial
`infarctions, or peripheral vascular disease.
`Based on the 1 year data from the Phase III pe-
`gaptanib trial, an annual thromboembolic rate of 6%
`and an annual death rate of 2% would be expected.7 In
`this retrospective study with a sample size of 50
`patients over 3 months, the probability of detecting at
`least one thromboembolic event given a true 3 month
`
`Table 3. Distribution of Injections through Month 3
`
`Follow-up Visit
`Month 1 (n ⫽ 51)
`Month 2 (n ⫽ 42)
`Month 3 (n ⫽ 53)
`
`Number of eyes
`injected at this
`visit (%)
`
`Total of one
`injection,
`n (%)
`
`Total of two
`injections,
`n (%)
`
`Total of three
`injections,
`n (%)
`
`Total of four
`injections,
`n (%)
`
`28 (55)
`21 (50)
`21 (40)
`
`23 (45)
`12 (29)
`11 (21)
`
`28 (55)
`19 (45)
`19 (36)
`
`0
`11 (26)
`18 (34)
`
`0
`0
`5 (9)
`
`Average
`number of
`injections per
`eye ⫾ SD
`1.5 ⫾ 0.5
`2.0 ⫾ 0.7
`2.3 ⫾ 0.9
`
`Novartis Exhibit 2284.005
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`500
`
`RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ● 2006 ● VOLUME 26 ● NUMBER 5
`
`Table 4. Frequency Distribution of Changes in Visual Acuity from Baseline through Month 3
`
`Change in Visual Acuity from
`Baseline through 3 Months
`
`Month 1, 51 eyes,
`n (%)
`
`Month 2, 42 eyes,
`n (%)
`
`Month 3, 53 eyes,
`n (%)
`
`ⱖ6-line increase
`ⱖ3-line to ⬍6-line increase
`ⱖ1-line to ⬍3-line increase
`No change
`ⱖ1-line to ⬍3-line decrease
`ⱖ3-line to ⬍6-line decrease
`ⱖ6-line decrease
`
`2 (4)
`11 (22)
`12 (24)
`20 (39)
`4 (8)
`2 (4)
`0
`
`2 (5)
`12 (29)
`9 (21)
`13 (31)
`3 (7)
`3 (7)
`0
`
`3 (6)
`20 (38)
`8 (15)
`14 (26)
`4 (8)
`3 (6)
`1 (2)
`
`rate of 1.5% would be 53% and the probability of
`observing at least one death given a true 3 month
`death rate of 0.5% would be 22%. Of course, this
`assumes that the events would be evenly distributed
`over 1 year. Since no thromboembolic events or
`deaths were observed among these 50 patients, the
`95% confidence interval around this incidence rate
`extends from 0% to 7%.
`Blood pressure was monitored during the 3 months
`on a subset of patients. Blood pressure measurements
`were usually obtained at the time of injection. Mea-
`surements were obtained on 20 patients at baseline, 15
`patients at month 1, 10 patients at month 2, and 27
`patients at month 3. Both baseline blood pressure and
`follow-up blood pressure measurements were ob-
`tained on only 8 patients at month 1, 4 patients at
`month 2, and 12 patients at month 3. While there was
`no apparent change in systolic blood pressure and
`diastolic blood pressure at month 1 and month 2 (P ⫽
`0.68 and P ⫽ 0.21, respectively), the mean systolic
`blood pressure increased from 131 mmHg at baseline
`to 148 mmHg at month 3 (P ⫽ 0.009), while the mean
`diastolic blood pressure remained unchanged at 80
`mmHg (P ⫽ 0.83). Among the 12 patients with mea-
`surements at baseline and month 3, one patient forgot
`to take his usual antihypertensive medication and the
`other patients with elevated measurements were re-
`ferred to their internists where their blood pressures
`were found to be unchanged. No new antihypertensive
`medications were initiated. In addition, there was no
`association between the number of injections and the
`3 month change in mean systolic (P ⫽ 0.99) or dia-
`stolic (P ⫽ 0.80) blood pressure.
`
`Visual Acuity and Central Retinal Thickness
`Outcomes
`
`Within 1 week after the initial bevacizumab injec-
`tion, improvements in visual acuity and central retinal
`thickness measurements were observed and these sig-
`nificant changes continued through 3 months (Tables
`1, 2, and 4). By 1 week, the mean visual acuity
`
`improved from 20/160 to 20/125 (P ⫽ 0.005) and this
`was maintained through 3 months (P⬍0.001). At the
`month 1 follow-up visit, 13 eyes (26%) had at least a
`3-line improvement, and at month 3, 23 eyes (44%)
`had at least a 3-line improvement in visual acuity
`(Table 4). This overall improvement in visual acuity at
`1 month was associated with an overall decrease in the
`1 mm central retinal thickness as measured by OCT
`(Table 1). At 1 week, the mean 1 mm central retinal
`thickness measurements decreased from 351 to 261
`␮m (–90 ␮m, P⬍0.001) and this overall improvement
`continued to 3 months.
`Overall, 4 eyes (8%) experienced at least a 3-line
`decrease in visual acuity at month 3 compared with
`baseline, and upon review of the OCT and angio-
`graphic images from these eyes, it was apparent that
`these eyes had chronic macular lesions with fibrosis.
`Even though the macular fluid resolved in these eyes
`following bevacizumab treatment, the visual acuity
`still decreased.
`The use of prior treatment appeared to have no
`effect on the visual acuity outcomes at 3 months. In
`particular, there was no difference in outcomes when
`the 23 eyes with prior PDT exposure were compared
`with the 30 eyes without prior PDT. When comparing
`the proportion of patients at 3 months with a 3-line
`improvement in visual acuity we found no difference
`(P ⫽ 0.85), and when comparing average change in
`visual acuity, we found no difference (P ⫽ 0.45).
`Figure 1 shows the distribution of patients with
`respect to their visual acuity at baseline and their
`visual acuity at 3 months. The solid line with a slope
`of one depicts no change in visual acuity with the
`symbols above the line representing patients with
`some gain in visual acuity and the symbols below the
`line representing patients with some loss in visual
`acuity. The dotted line is a fitted linear regression line
`with a slope of 1.02 (P ⬍ 0.001) showing that pa-
`tients, on average, improved approximately 1.5 lines
`as shown by the intercept on the Y-axis. The amount
`of improvement was not dependent on baseline visual
`
`Novartis Exhibit 2284.006
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`INTRAVITREAL BEVACIZUMAB FOR NEOVASCULAR ARMD ● RICH ET AL
`
`501
`
`ments at baseline and at 3 months. The dotted line
`with a slope of one depicts no change in central retinal
`thickness measurements with the symbols below the
`line representing patients with diminished macular
`fluid and the symbols above the line representing
`patients with increased macular fluid. As with the
`visual acuity data in Figure 1, these OCT data repre-
`sent only the 3 month time point in an otherwise
`dynamic process of fluid resorption and fluid reaccu-
`mulation usually dependent on whether an injection
`was given at the month 2 follow-up visit. Not surpris-
`ingly, the distribution of responses suggests that the
`greatest decrease in central
`retinal
`thickness
`is
`achieved when a larger amount of macular fluid is
`present at baseline; however, resolution of fluid is not
`always correlated with an improvement in vision at
`month 3 as shown in Case 4.
`
`Retreatment
`
`All eyes received an intravitreal injection at the
`initial visit; however, retreatment was at the discretion
`of the treating physicians. Table 3 summarizes the
`number of eyes that received 1, 2, 3, or 4 injections.
`Twenty-three eyes (43%) required 3 to 4 injections
`over 3 months to achieve complete resolution of mac-
`ular fluid (Case 1). Nineteen eyes (36%) received 2
`injections during the 3 months and these injections
`could have been given on sequential months or given
`at baseline followed by a reinjection at month 2 or 3.
`Usually, the reinjections were performed because of
`persistent or recurrent fluid observed using OCT im-
`aging. Visual acuity was not usually affected by the
`early recurrence of macular fluid as detected using
`OCT (Cases 2, 3). Finally, 11 eyes (21%) received just
`1 injection with resolution of macular fluid through
`the month 3 visit (21%) (Case 4). Case 4 is of interest
`because the vision initially improved with resolution
`of the fluid, but then subsequently declined to the
`baseline level of vision over 3 months without evi-
`dence of recurrent macular fluid.
`
`Case Reports
`
`Case 1
`
`An 81-year-old man had received two prior treatments to his
`right eye with combination PDT and intravitreal triamcinolone
`acetonide for a subfoveal predominantly classic neovascular lesion.
`During these treatments, his vision decreased from 20/100 to
`20/400, the neovascular lesion increased in size, and he developed
`steroid-induced ocular hypertension. At the time of the second
`treatment with PDT, both the right and left eyes received PDT. The
`left macula was treated because of a new small, extrafoveal,
`predominantly classic lesion. Three months after this treatment to
`his left eye, his vision deteriorated from 20/25 to 20/200. Fluores-
`
`..
`
`..
`
`80
`
`,0
`Visual Acuity (Equivalent Letter Score) at Baseline
`Fig. 1. Visual acuity at 3 months compared with baseline visual acuity
`(n ⫽ 53 eyes). The solid line represents the same visual acuity at
`baseline and at 3 months. The dotted line represents the linear regres-
`sion line showing the relationship between visual acuity at baseline and
`at 3 months following treatment. These visual acuities were highly
`correlated (P ⬍ 0.001, r2 ⫽ 0.71) with a slope ⫽ 1.02 and an intercept
`indicating that bevacizumab therapy was associated with an increase in
`visual acuity of 1.5 lines regardless of baseline visual acuity.
`
`acuity, as reflected by the regression slope of almost
`exactly 1.0. The square of the correlation coefficient
`(r2) for this relationship is 0.71, indicating that almost
`75% of the variability in the 3 month visual acuity is
`explained by the acuity at baseline. Thus, patients with
`poor initial acuity are likely to experience an improve-
`ment, but will still have relatively poor acuity at 3
`months after initiation of injections.
`Figure 2 depicts the distribution of patients with
`respect to their OCT central retinal thickness measure-
`
`1000
`
`800
`
`i 0
`:I: ..,
`~ e
`.. .. .
`a. 600
`! :c ... , oo
`C ·:, . Q:
`C . u
`
`;;
`
`;;
`t,
`
`200
`
`~
`~
`~
`~
`Central Retinal Thickness (um) at Baseline
`Fig. 2. Central retinal thickness at 3 months compared with central
`retinal thickness at baseline as measured using optical coherence to-
`mography. The dotted line represents the same central retinal thickness
`at baseline and at 3 months.
`
`1~
`
`Novartis Exhibit 2284.007
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`502
`
`RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ● 2006 ● VOLUME 26 ● NUMBER 5
`
`Fig. 3. Case 1. Left eye. Color fundus photographs with early and late-phase fluorescein angiographic images of the left eye. A, Baseline;
`bevacizumab injection given at this visit. B, Month 1 follow-up visit; bevacizumab injection No. 2 given at this visit. C, Month 3 follow-up visit, 2
`months after the second bevacizumab injection; bevacizumab injection given at this visit.
`
`cein angiography of the left eye showed an enlarged, juxtafoveal,
`predominantly classic neovascular
`lesion (Figure 3A). OCT
`showed macular cysts with subretinal fluid and a central retinal
`thickness of 663 ␮m (Figure 4A). He received an intravitreal
`injection of bevacizumab after discussing his treatment options
`which