AMERICAN ACADEMY
`OF OPHTHALMOLOGY®
`
`Prospective Study of Silicone Oil
`Microdroplets in Eyes Receiving lntravitreal
`Anti-Vascular Endothelial Growth Factor
`Therapy in 3 Different Syringes
`
`John T. Thompson, MD
`
`Purpose: To compare the prevalence of intravitreal silicone oil microdroplets detected by slit- lamp
`biomicroscopy in eyes with 6 or more injections of the same anti vascular endothelial growth factor (VEGF) drug.
`Design: Prospective, cross-sectional case series.
`Participants: A total of 260 consecutive eyes receiving 1 of 3 intravitreal anti-VEGF drugs for choroidal
`neovascularization, diabetic macular edema, or venous occlusive disease. The control group included 147 fellow
`eyes with no prior intravitreal injections.
`M ethods: The anterior and mid-vitreous were carefully examined using 12x to 16x magnification through
`dilated pupils with ocular saccades before an injection. Silicone oil microdroplets were graded on a scale from
`0 to 4+ based on the number and size of droplets.
`Main Outcome M easures: Presence and severity of silicone oil microdroplets in the vitreous.
`Results: Silicone oil microdroplets were observed in 78.3% of eyes receiving bevacizumab in Becton
`Dickinson (BD, Franklin Lakes, NJ) 0.3- ml polypropylene syringes, 14.4% of eyes receiving ranibizumab in 1.0-
`ml BD polypropylene syringes or more recently glass prefilled syringes, 48.5% of eyes receiving aflibercept in
`1.0- ml BD polycarbonate syringes, and 0% of eyes in controls. The differences among the 4 groups were
`statistically significant at P < 0.001 . The severity of silicone oil microdroplets was significantly greater in eyes
`using BD 0.3-ml polypropylene syringes than BD 1.0- ml polypropylene syringes, BD 1.0- ml polycarbonate
`syringes, or controls (P < 0.001). The severity of silicone oil microdroplets in eyes using BD 1.0-ml polycarbonate
`syringes was significantly greater than BD 1.0-ml polypropylene syringes (P = 0.012) and controls (P < 0.001).
`There was no significant difference between silicone oil microdroplet severity between BD 1.0- ml polypropylene
`syringes and controls (P = 1.0).
`Conclusions: The BD 0.3-ml polypropylene syringes with repackaged bevacizumab and the BD 1.0- ml
`polycarbonate syringes with afl ibercept cause a higher li kelihood of silicone oil microdroplets. lntravitreal in(cid:173)
`jections
`in eyes receiving multiple regular anti-VEGF injections should be supplied
`in silicone-free
`syringes. Ophthalmology Retina 2021;5:234-240 © 2020 by the American Academy of Ophthalmology
`
`The presence of intravitreal silicone oil microdroplets has
`been recogniz.ed most frequently in eyes receiving multiple
`intravitreal injections of anti vascular endothelial growth
`factor (VEGF). These anti-VEGF injections have become
`common in the treatment of choroidal neovascularization
`associated with age-related macular degeneration and other
`causes, such as myopia or histoplasmosis. Other common
`indications for treatment with anti-VEGF drugs include
`macular edema associated with venous occlusive disease
`and diabetic retinopathy. Many patients require these
`intravitreal injections every 4 to 12 weeks, and some pa(cid:173)
`tients have been receiving these injections since 2005.
`The 3 anti-VEGF drugs commonly used are bevacizumab
`(Avastin, produced by Genentech/Roche, South San Fran(cid:173)
`cisco, CA), ranibizumab (Lucentis, produced by Genentech/
`Roche), and aflibercept (Eylea, produced by Regeneron
`
`Pharmaceuticals, Tarrytown, NY). Each of the 3 drugs was
`typically used in the United States in a different syringe
`type. Minute amounts of silicone oil are typically placed in
`the barrel of most syringes to lubricate the plunger. 1 The
`presence of silicone oil microdroplets was recof nized
`more than 30 years ago with insulin syringes. The
`silicone is important to prevent the plunger from sticking
`during manufacturing and shipping. It also helps
`to
`facilitate a smoother injection by minimizing the break(cid:173)
`loose force
`to start the injection and glide force to
`complete the injections.2 The syringe type and needle used
`are important because they can have an impact on the
`amount of silicone oil that is injected into the eye with an
`intravitreal injection. Bevacizumab is repackaged from
`vials
`into syringes by a compounding pharmacy or
`outsourcing facility. The prefilled syringe has typically
`
`234
`
`© 2020 by lhe American Academy of Ophlhalmology
`Published by Elsevier Inc.
`
`htq,s://doi.org/10. 1016/j.oreL2020.07 .021
`ISSN 246$~530/20
`
`Novartis Exhibit 2321.001
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`Thompson  Silicone Oil Microdroplet in Vitreous
`
`been shipped in the past to the ophthalmologist in a 0.3-mL
`insulin (Fig 1A) or 1.0-mL polypropylene tuberculin syringe
`(Fig 1B). It is important to recall that the bevacizumab is
`stored in the syringe for up to 3 months before use, based
`on the expiration date supplied by the repackager. These 2
`types of
`syringes are typically produced by Becton
`Dickinson (BD, Franklin Lakes, NJ), a large supplier of
`syringes to the medical field. Ranibizumab was supplied
`in glass vials with a rubber stopper, and no BD syringe
`was in the kit supplied with the drug until early 2017
`when a prefilled glass syringe became available for the
`0.5-mg dose. The 0.3-mg ranibizumab glass prefilled
`syringes became available in 2018. Most ophthalmologists
`used the BD 1-mL tuberculin syringe (Fig 1B) to draw up
`and deliver ranibizumab before the introduction of the
`ranibizumab prefilled syringes. Aflibercept is supplied in a
`glass vial with a different BD 1.0-mL polycarbonate
`syringe supplied in the kit with the aflibercept (Fig 1C)
`before the introduction of glass prefilled aflibercept
`syringes in 2019. The 0.3-mL BD insulin syringe comes
`with a 31-gauge needle attached (“staked syringe”). Most
`ophthalmologists have attached the 30-gauge BD needle or
`32-gauge needle (TSK Laboratory, Tochigi-Ken, Japan)
`because BD does not sell a disposable 32-gauge needle.
`The presence of microdroplets of silicone oil
`in the
`vitreous of eyes receiving these injections was first reported
`in 2006 and was noted to be an infrequent complication after
`intravitreal injection of pegaptanib, an anti-VEGF drug in a
`prefilled syringe but rarely used today, and after triamcin-
`olone acetonide injections.3 A report 2 years later found
`silicone oil in 15 eyes of 1529 injections, but the number
`of patients is not reported, so a per patient prevalence
`cannot be calculated.4 Subsequent publications have noted
`
`treated with
`silicone oil droplets primarily in eyes
`repackaged bevacizumab.5-13 Silicone oil microdroplets
`have been reported less frequently after injection of ranibi-
`zumab5,9 and aflibercept.12 The purpose of the current study
`was to evaluate the prevalence and severity of silicone oil
`microdroplets in the vitreous of eyes that have received
`numerous intravitreal anti-VEGF injections of the same
`drug using the same syringe type for that particular drug.
`
`Methods
`
`The presence and severity of silicone oil microdroplets were
`evaluated prospectively in a consecutive series of all eyes in the
`author’s practice on the day of an intravitreal anti VEGF injection
`before the injection for choroidal neovascularization or macular
`edema due to venous occlusive disease or diabetic retinopathy. The
`data were collected on eyes examined between November 5, 2018,
`and January 31, 2019. The evaluation was performed only when
`the eye was dilated because the eye is not routinely dilated on
`every visit. Eyes had to have a minimum of 6 injections of the same
`intravitreal drug (bevacizumab, ranibizumab, or aflibercept) in the
`same syringe type at the time of grading. Any eye that had a
`different anti VEGF drug/syringe injected at some point during
`their prior treatment was excluded. A total of 227 eyes were
`excluded primarily because they had more than 1 different anti
`VEGF drug injected or had fewer than 6 injections. This would
`allow comparison of the prevalence of silicone oil microdroplets
`among the 3 drugs that were each delivered using 3 different BD
`syringes, respectively, in addition to the prefilled syringe intro
`duced for ranibizumab. Bevacizumab was supplied by Avella
`(Phoenix, AZ) in the BD 0.3 mL insulin syringes. Ranibizumab
`was delivered in the BD 1.0 mL tuberculin syringe until the pre
`filled syringes became available for the 0.5 mg and 0.3 mg dose in
`2017 and 2018, respectively. The ranibizumab prefilled syringes
`were used once they became available. Aflibercept was delivered in
`
`Figure 1. A, Becton Dickinson (BD, Franklin Lakes, NJ) 0.3-mL polypropylene insulin syringe. B, BD 1.0-mL polypropylene syringe. C, BD 1.0-mL
`polycarbonate syringe.
`
`235
`
`Novartis Exhibit 2321.002
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`Ophthalmology Retina Volume 5, Number 3, March 2021
`
`Table 1. Grading Scale for Silicone Oil Microdroplets in the Vitreous
`
`Silicone Oil Severity Grade
`
`0
`0.5
`1.0
`2.0
`3.0
`4.0
`
`the BD 1.0 mL polycarbonate syringe supplied in the afli
`bercept kit, and none of the eyes in this study had aflibercept
`from the prefilled syringe that was not available at completion
`of data collection. The BD 30 gauge needle was used on all
`ranibizumab and aflibercept doses until switched to the TSK
`Laboratory 32 gauge needles
`in mid 2018. Other exclusion
`criteria were prior vitrectomy because the oil would float
`to
`the top of
`the eye and not be visible by slit lamp bio
`microscopy, poor pupillary dilation, active uveitis with vitreous
`cells, and asteroid hyalosis because the author found it diffi
`cult
`to accurately identify silicone oil microdroplets under
`these circumstances. Eyes were excluded if
`they had not
`received anti VEGF injections within the past year. All eyes
`with prior injection of other drugs such as triamcinolone or
`dexamethasone implant were excluded.
`Institutional Review
`Board approval was obtained by The Greater Baltimore
`Medical Center Institutional Review Board. The study adhered
`to the tenets of
`the Declaration of Helsinki. The data was
`collected during routine care of the patients, and the IRB did
`not require consent.
`The presence or absence of silicone oil microdroplets was
`assessed by an unmasked observer using an oblique slit beam at
`12 to 16 magnification focused into the mid vitreous. The sil
`icone oil microdroplets could be distinguished from other floaters
`or cells in the vitreous because they reflected light, causing them to
`appear like shiny bubbles. Patients were asked to saccade up and
`down, side to side several times because this improved detection of
`the silicone oil microdroplets. This examination of the vitreous
`typically took approximately 30 seconds, depending on patient
`cooperation with the saccades, and all grading was performed by
`the author. It was not possible to mask the author to the intravitreal
`anti VEGF drug received by the patient, but the goal was to find
`silicone oil microdroplets in all eyes, including controls. Silicone
`oil severity was graded using the following approximate scale
`because it was somewhat difficult to precisely count the number of
`silicone oil microdroplets
`(Table 1). The
`control group
`included fellow eyes of patients who were receiving intravitreal
`anti VEGF injections, but not necessarily those included in the
`study. They could be any fellow eye from a patient receiving
`intravitreal anti VEGF injection that was examined during the
`same time interval.
`
`Statistical Analyses
`
`The prevalence of any silicone oil was determined for each group
`and compared with one another using the chi square test with
`correction for multiple comparisons. The severity of silicone oil
`was compared among the 4 groups, including the controls using
`univariate analysis of variance with Bonferroni correction for
`multiple comparisons. Linear regression analysis was used to
`compare the severity of silicone oil and the number of injections
`for each of the 3 anti VEGF drugs. All analyses were performed
`using Statistical Package for the Social Sciences Subscription
`(IBM, Chicago, IL).
`
`236
`
`No droplets
`1e2 microdroplets < 0.1 mm diameter
`3e5 microdroplets < 0.1 mm diameter
`5e7 microdroplets < 0.1 mm diameter or 1 large droplet > 0.1 mm diameter
`8e10 microdroplets < 0.1 mm diameter or 2 larger droplets > 0.1 mm
`>10 microdroplets < 0.1 mm or 3 larger droplets > 0.1 mm diameter
`
`Results
`
`There were 69 eyes receiving bevacizumab in BD 0.3 mL
`polypropylene syringes, 125 eyes receiving ranibizumab in BD
`1.0 mL polypropylene syringes, and 66 eyes receiving aflibercept
`in BD 1.0 mL polycarbonate syringes. The control group consisted
`of 147 fellow eyes of patients receiving any anti VEGF injection
`with no prior intravitreal injections in that eye. The eyes receiving
`bevacizumab had a mean of 29 injections, eyes receiving ranibi
`zumab had a mean of 33 injections, and eyes receiving aflibercept
`had 26 injections. There was no significant difference in the
`number of injections among the 3 syringe types by analysis of
`variance using Bonferroni correction for multiple comparisons
`(Table 2). Figure 2 shows the percentage of eyes with any visible
`silicone oil droplets in the vitreous. Silicone oil microplets were
`visible in 78% of bevacizumab delivered with the BD 0.3 mL
`BD polypropylene insulin syringe and 14% of
`ranibizumab
`delivered with the 1.0 mL BD polypropylene syringe or prefilled
`syringe. Silicone oil microdroplets were not seen in any of the
`14 eyes receiving ranibizumab in the prefilled syringe only, but
`these eyes had many fewer injections because this syringe was
`more
`recently introduced. Silicone oil microdroplets were
`detected in 49% of aflibercept eyes delivered with the BD 1.0
`mL polycarbonate syringe and none of the control eyes. The
`prevalence of silicone microdroplets was significantly greater in
`BD 0.3 mL polypropylene (bevacizumab) eyes than all other
`groups as well as BD 1.0 mL polycarbonate (aflibercept)
`compared with ranibizumab and control (P < 0.001 for all
`comparisons by Pearson chi square). The prevalence of any sili
`cone oil microdroplets was significantly greater in BD 1.0 mL
`polypropylene (ranibizumab) eyes compared with control eyes
`(P < 0.001 by Pearson chi square). Figure 3 shows a graph of the
`severity of silicone oil for each of the 4 groups.
`The silicone microdroplet severity score was compared using
`univariate analysis of variance with Bonferroni correction for
`multiple comparisons (Table 3). The severity grade of silicone oil
`microdroplets with bevacizumab in 0.3 mL BD insulin syringes
`was significantly greater (P < 0.001) than aflibercept
`in BD
`polycarbonate
`1.0 mL,
`ranibizumab
`in
`BD
`1.0 mL
`polypropylene,
`and controls. The
`severity of
`silicone oil
`microdroplets with aflibercept
`in 1.0 mL BD polycarbonate
`syringes was significantly greater than ranibizumab in BD 1.0
`mL polypropylene (P ¼ 0.012) and control eyes (P < 0.001).
`The severity of silicone oil microdroplets with ranibizumab in
`BD 1.0 mL polypropylene or prefilled syringes was not
`significantly different from that of controls (P ¼ 1.0). Linear
`regression analysis was performed to see if
`the severity of
`silicone oil microdroplets
`correlated with the number of
`injections. The number of injections was associated with the
`
`Novartis Exhibit 2321.003
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`Thompson  Silicone Oil Microdroplet in Vitreous
`
`Table 2. Number of Eyes and Intravitreal Injections
`
`Anti-VEGF Drug
`
`No. of Eyes
`
`Bevacizumab
`
`Ranibizumab*
`Aflibercept
`Control (fellow eye with no prior
`intravitreal injections)
`
`69
`
`125
`66
`147
`
`Mean No. of
`Injections (Range)
`
`29 (6e121)
`
`33 (6e90)
`26 (6e69)
`0
`
`Median
`
`P Value (with Bonferroni Correction)
`
`19
`
`29
`26
`
`Bevacizumab vs. Ranibizumab
`Bevacizumab vs. Aflibercept
`Ranibizumab vs. Aflibercept
`
`0.0593
`1.00
`0.055
`
`vascular endothelial growth factor.
`VEGF
`*A total of 14 eyes had ranibizumab delivered only in prefilled syringes. Most eyes treated with ranibizumab had a combination of injections with BD
`polypropylene and later with prefilled syringes.
`
`severity of
`silicone oil
`for bevacizumab in BD 0.3 mL
`polypropylene (P ¼ 0.011) and ranibizumab in BD 1.0 mL
`polypropylene (P ¼ 0.005) but was not significant for aflibercept
`(P ¼ 0.085). The
`in BD 1.0 mL polycarbonate syringes
`correlation was weak for all 3 anti VEGF drugs with R2¼0.093
`for bevacizumab, R2¼0.064 for ranibizumab, and R2¼0.046 for
`aflibercept. The relatively low value for R2 means that less than
`10% of the silicone oil severity is explained by the mean number of
`injections. This suggests that the amount of silicone oil is not
`constant in each syringe but varies from lot to lot. Otherwise, eyes
`with a greater number of injections would be expected to have
`more silicone oil microdroplets than eyes with fewer injections.
`
`Discussion
`
`Intravitreal anti-VEGF injections have become common
`because of their potent effect on decreasing leakage from
`neovascular and damaged blood vessels in the retina and
`choroid. Frequent injections are required to maintain the
`
`leading to numerous
`beneficial response in most eyes,
`intravitreal injections over a period of many years. The BD
`syringes often used for these injections were not designed
`for delivering intravitreal drugs and contain silicone oil
`droplets to lubricate the barrel of the syringe to make the
`injection easier. The needles may also contain small
`amounts of silicone oil, but
`this appears to be a less
`important source of silicone oil microdroplets based on the
`results of studies of the BD syringes and syringes from other
`manufacturers. The same needles were used for the ranibi-
`zumab and aflibercept, so this would not explain the dif-
`ference in frequency of silicone oil between the different
`syringe types. One limitation of the current study is that the
`anti-VEGF drug and syringe type in each patient was not
`masked to the grader. It was the goal of the grader to find
`silicone oil microdroplets in all eyes, even the control eyes.
`There was also no precise quantitative way to measure the
`number and size of the silicone oil microdroplets, but all
`eyes were examined by 1 grader using the same criteria.
`
`100.0%
`
`90.0%
`
`80.0%
`
`70.0%
`
`60.0%
`
`50.0%
`
`40.0%
`
`30.0%
`
`20.0%
`
`10.0%
`
`0.0%
`
`Any Visible Silicone Oil Microdroplets
`
`Bevacizumab in BO 0.3-
`ml insulin syringe,
`78.3%
`
`Aflibercept in BO 1.0·
`ml proypropylene
`syringe, 48.5%
`
`Ranibizumab in BO 1.0
`ml TB syringe or
`prefilled syringe,
`14.4%
`
`Ill
`
`Control, 0.0%
`
`Bevacizumab
`
`Ranibizumab *
`
`Afli bercept
`
`Control
`
`* 14 eyes received only ranibizumab prefilled syringes
`and none had any visible silicone oil microdroplets
`
`Figure 2. Percentage of eyes with any silicone oil microdroplets visible in the vitreous by slit-lamp biomicroscopy for bevacizumab in BD 0.3-mL poly-
`propylene syringes, ranibizumab in BD 1.0-mL polypropylene syringes, aflibercept in BD 1.0-mL polycarbonate syringes, and controls. BD
`Becton
`Dickinson.
`
`237
`
`Novartis Exhibit 2321.004
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`Ophthalmology Retina Volume 5, Number 3, March 2021
`
`Severity of Silicone Oil Microdroplets
`
`100.0%
`90.0%
`80.0%
`70.0%
`60.0%
`50.0%
`40.0%
`30.0%
`20.0%
`10.0%
`0.0%
`
`Bevacizumab in BD 0.3- Ranibizumab in BD 1.0- Aflibercept in BD 1.0-ml
`ml polypropylene
`ml polypropylene
`polycarbonate syringe
`syringe
`syringe
`
`Control
`
`■ 0 ■ 0.5 ,, 1 JC 2
`
`Ill 3 11 4
`
`Figure 3. Percentage of eyes with silicone oil visible in the vitreous for bevacizumab in BD 0.3-mL polypropylene syringes, ranibizumab in BD 1.0-mL
`polypropylene syringes, aflibercept in BD 1.0-mL polycarbonate syringes, and controls using a severity scale from 0 to 4þ. BD Becton Dickinson.
`
`Early reports of silicone oil microdroplets after intra-
`vitreal injections suggested this was an infrequent finding
`and related to particular batches of syringes with greater
`amounts of silicone oil.3,4 Several studies have noted
`increased detection of silicone oil microdroplets in the
`vitreous during one time period.10 A larger series of 6632
`bevacizumab injections found 60 of 6632 eyes (0.904%)
`with silicone oil droplets with a greater prevalence from
`injections during one
`time period in 2016 (1.7%)
`compared with a different
`time period in 2015 2016
`(0.03%).7 The variability in silicone oil detection during
`different
`time intervals was confirmed in a more recent
`study.14 The storage of the prefilled bevacizumab syringes
`also influences the likelihood of
`injecting silicone oil
`microdroplets
`into
`the
`eye. For
`example,
`if
`the
`bevacizumab is
`frozen and then thawed because of
`temperature variability during refrigeration, the amount of
`siliconle oil increases by approximately 4-fold in 0.3-mL
`insulin syringes, but not 1.0-mL tuberculin syringes.3
`Several
`studies have
`recognized that
`syringes
`from
`different manufacturers have variable amounts of silicone
`oil. The BD 0.3-mL syringes were noted to have more
`silicone oil when fluid was expressed through them
`compared with BD tuberculin syringes and HSW silicone-
`free syringes
`(Henke Sass Wolf
`[HSW], Tuttlingen,
`
`Germany).8 More oil
`the end of
`is also expressed at
`injection if the plunger is depressed all of the way to the
`end with the BD 0.3-mL insulin syringes compared with
`the beginning of the injection.6,8 Agitation of the syringes
`by flicking them to remove bubbles also appears to
`increase delivery of silicone oil.15 These 2 factors could
`also explain why the eyes treated with aflibercept had
`more silicone oil microdroplets. The volume of aflibercept
`in the glass vial supplied with the drug is 0.1-mL, so this
`means there is less drug compared with ranibizumab,
`which contained 0.23-mL in their glass vial. This means
`that it was more common to have to push the plunger of
`the syringe containing aflibercept all
`the way down to
`deliver the full 0.05-mL dose compared with ranibizumab,
`for which there was sufficient drug to start with the
`syringe at 0.1-mL and depress the plunger until it read
`0.05-mL to deliver the same dose. Aflibercept
`is also
`more viscous than ranibizumab, so there is a tendency for
`the aflibercept
`to contain bubbles when it
`is drawn up
`from the glass vial into the syringe. It is common practice
`to flick the syringe and withdraw additional air into the
`syringe to expel the air bubbles during priming, before the
`aflibercept
`is injected into the eye because patients are
`bothered by the small air bubbles for 1 or 2 days after
`their intravitreal injection.
`
`Table 3. Comparison of Silicone Oil Microdroplets Severity in Different Syringe Types by Univariate Analysis of Variance
`
`Anti-VEGF Drug/Syringe
`
`Compared to Anti-VEGF Drug/Syringe
`
`Bevacizumab in 0.3-mL BD insulin syringe
`
`Aflibercept in 1.0-mL BD polycarbonate syringe
`
`Ranibizumab in 1.0-mL BD polypropylene/prefilled
`syringe
`
`Aflibercept in 1.0-mL BD polycarbonate syringe
`Ranibizumab in 1.0-mL BD polypropylene/prefilled syringe
`Control (no injections)
`Ranibizumab in 1.0-ml BD polypropylene/prefilled syringe
`Control (no injections)
`Control
`
`P Value (with Bonferroni Correction)
`P <0.001
`P <0.001
`P <0.001
`P
`0.012
`P <0.001
`P
`1.0
`
`BD Becton Dickinson.
`
`238
`
`Novartis Exhibit 2321.005
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`Thompson  Silicone Oil Microdroplet in Vitreous
`
`A comparison of repackaged bevacizumab in Braun sy-
`ringes (B. Braun Medical Inc, Bethlehem, PA) showed a 4-
`to 9-fold increased particle count using light obscuration and
`micro-flow imaging compared with ranibizumab in prefilled
`glass syringes.9 The use of silicone-free syringes results in
`no detectable silicone oil microdroplets in another study
`comparing syringes with silicone lubrication with silicone-
`free syringes.11 It is believed that most of the intravitreal
`silicone oil from anti-VEGF injections arises from the
`syringes, but
`the needles used to inject
`the drug are a
`potential source of silicone contamination because tiny
`amounts of silicone oil were also found in 26- and 30-gauge
`BD PrecisionGlide needles when tested without syringes.16
`Silicone oil microdroplets in the vitreous are usually
`asymptomatic with milder grades such as trace or 1þ, but
`most patients with 2þ or greater did report floaters or
`admitted seeing them when asked. Patients with very poor
`visual acuities (20/400 or less) did not seem to recognize the
`silicone oil microdroplets as often. A small percentage of
`patients with the
`silicone oil microdroplets become
`sufficiently bothered by these floaters that vitrectomy was
`performed to remove them. This may require more frequent
`anti-VEGF injections because it appears that the half-life of
`the ranibizumab and aflibercept is decreased in an animal
`eyes with vitrectomy.17 The
`model of
`silicone oil
`microdroplets in the vitreous usually does not cause any
`substantial inflammation, but there is one report of silicone
`oil droplets after aflibercept injection that was linked to
`intraocular inflammatory reactions.12 The US Food and
`Drug Administration released new guidelines affecting the
`preparation of the repackaged bevacizumab. Compounding
`pharmacies and outsourcing facilities were previously
`bound by USP 797 guidelines. In 2017, the Food and Drug
`Administration released more extensive guidelines for
`repackaged biologics, which became finalized in January
`2018 as USP 789.18 These more stringent guidelines have
`
`Footnotes and Disclosures
`
`Originally received: April 20, 2020.
`Final revision: July 3, 2020.
`Accepted: July 23, 2020.
`Available online: July 28, 2020.
`
`Manuscript no. ORET D 20 00291.
`
`Retina Specialists, Baltimore, Maryland.
`
`The results were presented in part at: the American Society of Retina
`Specialists Meeting 2019, Chicago, Illinois, July 26e30, and The Retina
`Society Meeting, September 11e15, 2019, London, United Kingdom.
`Disclosure(s):
`All authors have completed and submitted the ICMJE disclosures form.
`The author(s) have made the following disclosure(s): JTT: Grant support:
`Genentech (Roche), Regeneron Pharmaceuticals, Opthea Consultant e
`Genentech (Roche).
`Dr. John T. Thompson serves on the editorial board of Ophthalmology
`Retina.
`HUMAN SUBJECTS: Human subjects were included in this study. The
`human ethics committees at the Greater Baltimore Medical Center approved
`the study. All research adhered to the tenets of the Declaration of Helsinki.
`
`caused most compounding pharmacies and outsourcing
`facilities to stop using the BD 0.3-mL insulin syringe to
`repackage bevacizumab. Many have shifted to the use of
`silicone-free 1-mL tuberculin syringes to allow them to
`adhere to USP 789 requirements. It is unknown if the afli-
`bercept delivered in the BD 1-mL polycarbonate syringes
`would pass the USP 789 guidelines, but Regeneron Phar-
`maceuticals introduced a prefilled glass syringe for delivery
`of aflibercept 2019, although supply issues initially slowed
`transition to all glass prefilled syringes.
`
`Conclusions
`
`This study found large differences in the presence of
`silicone oil microdroplets in the vitreous using 3 different
`syringes
`commonly
`used
`for
`intravitreal
`injections.
`Repackaged bevacizumab in BD 0.3-mL insulin syringes
`and aflibercept delivered in the supplied BD 1.0-mL
`polycarbonate syringes results in detectable silicone oil
`microdroplets in 49% of eyes treated long term with
`aflibercept delivered in BD 1.0-mL polycarbonate sy-
`ringes and 78% of eyes with repackaged bevacizumab
`delivered in BD 0.3-mL insulin syringes. In late 2019, a
`new intravitreal anti-VEGF drug became available, bro-
`lucizumab (Beovu produced by Novartis Pharmaceuticals
`Corporation, Basel, Switzerland). This drug is currently
`only available in glass vials, so silicone-free syringes
`should be preferred for intravitreal
`injection. It
`is rec-
`ommended that pharmaceutical companies load intravitreal
`anti-VEGF drugs in silicone-free syringes. It is also rec-
`ommended that compounding pharmacies and outsourcing
`facilities repackage intravitreal drugs such as bevacizumab
`in silicone-free syringes to minimize the risk of silicone
`microdroplets
`in the vitreous
`in eyes with repeated
`injections.19
`
`The data was collected during routine care of the patients, and the IRB
`exempted this study from collecting patient consent.
`No animal subjects were used in this study.
`Author Contributions:
`Conception and design: Thompson
`Data collection: Thompson
`Analysis and interpretation: Thompson
`Obtained funding: N/A
`Overall responsibility: Thompson
`Abbreviations and Acronyms:
`BD Becton Dickinson; VEGF
`Keywords:
`intra
`ranibizumab, bevacizumab, aflibercept, silicone oil microdroplet,
`vitreal
`injections, Becton Dickinson syringe, anti vascular endothelial
`growth factor.
`Correspondence:
`John T. Thompson, MD, 6569 North Charles St., Suite 605, Baltimore, MD
`21204. E mail: JThompson@retinaspec.com.
`
`vascular endothelial growth factor.
`
`239
`
`Novartis Exhibit 2321.006
`Regeneron v. Novartis, IPR2021-00816
`
`

`

`Ophthalmology Retina Volume 5, Number 3, March 2021
`
`References
`
`1. Changelau E, Berger M, Bohlken B. Silicone oil released from
`disposable insulin syringes. Diabetes Care. 1986:9672e9673.
`2. Yoshino K, Nakamura K, Yamashita A, et al. Functional eval
`uation and characterization of a newly developed silicone oil free
`prefillable syringe system. J Pharm Sci. 2014;103:1520e1528.
`3. Freund KB, Laud K, Eandi CM, Spaide RF. Silicone oil droplets
`following intravitreal injection. Retina. 2006;28:701e703.
`4. Bakri SJ, Ekdawi NS. Intravitreal silicon oil droplets after
`intravitreal drug injections. Retina. 2008;28:996e1001.
`5. Liu L, Ammar DA, Ross LA, et al. Silicone oil microdroplets and
`protein aggregates in repackage bevacizumab and ranibizumab:
`Effects of long term storage and product mishandling. Invest
`Ophthalmolol Vis Sci. 2011;52:1023e1034.
`6. Pitcher JD, Roybal CN. Silicone oil microdroplets in repack
`aged bevacizumab syringes: Identification, quantification, and
`strategies for reducing clinical impact. ARVO Annual Meeting
`Abstract. Invest Ophthalmol Vis Sci. 2017;58:434.
`7. Khurana RN, Chang LK, Porco TC. Incidence of presumed
`silicone oil droplets in the vitreous cavity after intravitreal
`bevacizumab
`injection with
`insulin
`syringes.
`JAMA
`Ophthalmol. 2017;135:800e803.
`8. Emerson GG. Silicone oil droplets are more common in fluid
`from BD insulin syringes as compared to other syringes.
`J Vitreoretin Dis. 2017;1:401e406.
`9. Shargus M, Werner BP, Geerling G, Winter G. Contamination
`of anti VEGF drugs
`for
`intravitreal
`injection: how do
`repackaging and newly developed syringes affect the amount
`of silicone oil droplets and protein aggregates? Retina.
`2018;38:2088e2095.
`10. Yu JH, Gallemore E, Kim JK, et al. Silicone oil droplets
`following intravitreal bevacizumab. Am J Ophthalmol Case
`Rep. 2017;10:142e144.
`
`11. Olea JL, Gomez Resa M, Cervera Peris MM, Aragon JA.
`Silicone oil droplets in repackaged anti vascular endothelial
`growth factors for intravitreal injection: in search of the main
`source of contamination. Eur J Ophthalmol. 2020;30:
`774e779.
`12. Melo GB, Figueira ACM, Batista FAH, et al. Inflammatory
`reaction after aflibercept intravitreal injection associated with
`silicone oil droplets released from syringes: a case control
`study. Ophthalmic Surg Laser Imaging Retina. 2019;50:
`288e294.
`13. Melo GB, Emerson GG, Dias Jr CS, et al. Release of silicone
`oil and the off label use of syringes in ophthalmology. Br J
`Ophthalmol. 2020;104:291e296.
`14. Avery RL, Castellarin AA, Dhoot DS, et al. Large silicone
`droplets after intravitreal bevacizumab (Avastin). Retin Cases
`Brief Rep. 2019;13:130e134.
`15. Melo GB, Dias Jr CS, Carvalho MR, et al. Release of
`silicone oil droplets from syringes. Int J Retina Vitreous.
`2019;5:1.
`16. Melo GB, Emerson GG, Lima Filho AAS, et al. Needles as a
`source of silicone oil during intravitreal injection. Eye (Lond).
`2019;33:1025e1027.
`17. Niwa Y, Kakinoki, Sawada T, et al. Ranibizumab and
`aflibercept: intraocular pharmacokinetics and their effects on
`aqueous VEGF level
`in vitrectomy and nonvitrectomized
`macaque
`eyes.
`Sci.
`2015;56:
`Invest Ophthalmol Vis
`6501e6505.
`18. US Pharmacopeia: USP 789. http://ftp.uspbpep.com/v29240/
`usp29nf24s0 c789.html. Accessed January 1, 2020.
`19. Sharma A, Kumar N, Bandello F, et al. Understanding
`intravitreal silicone oil droplets due to intravitreal injections.
`Retina. 2019;39:1233e1235.
`
`240
`
`Novartis Exhibit 2321.007
`Regeneron v. Novartis, IPR2021-00816
`
`