SILICONE OIL MIGRATION DURING INTRAVITREAL ANTI–VASCULAR ENDOTHELIAL GROWTH FACTOR INJECTIONS

Silicone oil microdroplets can enter the vitreous after anti–vascular endothelial growth factor injections, mainly because of siliconized syringes and improper handling. Reported effects were mostly floaters and imaging artifacts, with no consistent evidence of sterile inflammation or lasting intraocular pressure elevation.

Abstract

Purpose:

To evaluate the incidence, mechanisms, and clinical implications of silicone oil droplet transmission during intravitreal anti–vascular endothelial growth factor injections and to review strategies to mitigate contamination risks.

Methods:

A systematic review of observational and review studies from PubMed and MEDLINE was conducted, focusing on incidence, contributing factors, and clinical outcomes of silicone oil contamination during IVIs. Specific attention was given to syringe design, storage conditions, variations among anti–vascular endothelial growth factor agents, and injection techniques.

Results:

The review found that prevalence is high in chronically injected eyes, although per-injection rates are low, often clustering because of syringe batches or techniques. Contributing factors include freeze–thaw cycles, mechanical agitation, spray-siliconized low dead space syringes, and improper plunger handling, which can lead to asymptomatic droplets or symptomatic floaters. No severe complications such as endophthalmitis were reported. Hypotheses link silicone oil to ocular hypertension or inflammation via trabecular obstruction or immunogenic responses, but the reviewed studies show no direct evidence. Mitigation involves the use of silicone-free/prefilled syringes, strict storage, gentle handling, and refined techniques such as priming and smooth plunger depression.

Conclusion:

Silicone oil contamination during IVIs can be minimized through careful optimization of syringe design, handling practices, and injection techniques. Embracing silicone-free or low-silicone syringe systems, alongside meticulous procedural protocols, offers the potential to enhance the safety and clinical outcomes of intravitreal anti–vascular endothelial growth factor therapy.

With more than 10 million injections administered annually, intravitreal injections (IVIs) have become the primary route of pharmaceutical delivery for managing a range of retinal vascular diseases.¹ However, alongside their widespread use, an iatrogenic phenomenon has emerged: the appearance of silicone oil (SO) droplets in the vitreous humor after intravitreal anti–vascular endothelial growth factor (VEGF) injections.

In 2006, Freund et al² were the first to document the presence of SO droplets within the vitreous cavity after repeated intravitreal injections. In 2008, Bakri and Ekdawi³ reinforced Freud's findings documenting the transmission of SO in patients receiving multiple intravitreal injections. In their series, SO contamination was agent-independent, occurring with various anti-VEGF therapies (bevacizumab, pegaptanib, ranibizumab), implicating the silicone-coated syringes and needles as the source rather than the therapeutic agents.³ Although patients in both studies were asymptomatic, the authors raised concerns about the potential long-term effects of silicone droplets incorporated into ocular tissues. This review aims to evaluate strategies for mitigating SO contamination during IVIs, while also providing a comprehensive narrative of human studies related to the topic.

Methods

For this review, a comprehensive literature search was performed to identify all relevant studies addressing SO droplet transmission from intravitreal injections. Multiple databases were queried, including PubMed/MEDLINE, Embase, and Google Scholar. The search strategy incorporated both vocabulary and keyword terms related to intravitreal injections and SO. Following best practices for search reporting (PRISMA-S guidelines), we documented the search process in detail, including the screening practices (Figure 1).

Fig. 1.

Identification, screening, and inclusion of records for this systematic literature review.

Eligible studies were human reports of intravitreal injections (primarily anti-VEGF) that measured SO outcomes. Included designs were randomized controlled trial, cohort and case–control studies, relevant observational studies, and case series. Data extracted from each study included identifiers, design/sample characteristics, injection agent, syringe and needle details (type, material, siliconization, dead space, gauge), compounding/storage conditions, injection technique (handling, priming, speed, angle), and outcomes such as SO presence, visual symptoms, imaging artifacts, other complications, and their management.

Results

The systematic search yielded 15 relevant studies, including cross-sectional prevalence studies, retrospective case series, and a randomized trial. These studies varied in design and scope, but collectively, they provide data on the incidence of SO droplets, contributing risk factors, clinical symptoms, and preventive measures. Key characteristics and findings of each included study are summarized in Table 1. Given the heterogeneity of methodologies, a quantitative meta-analysis was not conducted; instead, we present a narrative synthesis of the results, organized by significant themes: incidence, syringe type, compounding and storage practices, and injection technique.

Table 1.

Summary of 15 Studies (Case Series, Cross-Sectional, Retrospective Cohorts, Outbreak Reports, One RCT) on Silicone-Oil Contamination After Intravitreal Injections, Listing Design, Devices/Handling, Detection Method, Symptoms/Imaging, and IOP

Study (Author, Year)	Design	Population	Anti-VEGF Agent(s)	Syringe and Needle Type	Dead Space	Compounding/Storage	Injection Technique	Incidence of SO Droplets (Detection)	Patient Symptoms	Imaging Findings	IOP	Ocular Inflammation	Adverse Events
Freund et al, 2006	Case series	∼3 eyes	Pegaptanib or bevacizumab (early anti-VEGF era)	Standard IVI syringes (likely siliconized; needle type not specified)	Normal (standard 1 mL)*	On-label (pegaptanib kit) or repackaged bevacizumab	Standard injection (no special handling noted)	Small silicone oil droplets noted in vitreous immediately postinjection (identified on fundus examination)	No new visual symptoms	Slit-lamp and fundus examination—visible refractile droplets	No IOP rise noted	No inflammation noted	No adverse events noted
Bakri and ekdawi, 2008	Retrospective case series	15 eyes	Pegaptanib, bevacizumab, ranibizumab, or triamcinolone	Pegaptanib—sterile, 1-mL, USP type 1, graduated, prefilled glass syringe fitted with 27-gauge needle Triamcinolone—1.0-mL disposable syringes + detachable needles (siliconized)* Ranibizumab—1.0-mL disposable syringes + detachable needles (siliconized)* Bevacizumab—fractioned from a large vial into 0.15-mL doses supplied in a tuberculin syringe (siliconized*) + detachable needle	High (∼0.07-mL hub)	Bevacizumab drawn from vial (clinic use); others from vials	Standard technique (no specific priming/flicking noted)	Detected in 15 eyes after 1–16 injections (noted on dilated examination)	None—0% symptomatic; patients did not report floaters	Slit-lamp examination—clear, mobile droplets in vitreous	No IOP rise noted	No inflammation noted	No adverse events noted
Scott et al, 2009 (SCORE report 7)	Prospective, randomized, phase III clinical trial	464 RVO patients (1,205 injections)	Triamcinolone (IVT for RVO)**	Staked-on 27-gauge needle with glass syringes vs. Luer-cone 30-gauge or 27-gauge glass syringes (both siliconized)	Low (staked) vs. high (Luer)	No compounding (drug drawn fresh into syringes)	Standard (initially staked syringes; mid-study switched to Luer)	Staked-needle group: 141/319 eyes (44%) had SO droplets; Luer-only: 0/58 eyes had droplets Slit-lamp examination	None reported (incidental finding during study examinations)	Biomicroscopy—vitreous microdroplets noted on examination	Not reported	Not reported	Not reported
Khurana et al, 2017​	Retrospective series	6,632 injections	Bevacizumab (1.25 mg)	0.3-mL insulin syringe, 31-G staked needle (siliconized)	Very low (∼0.005 mL)	Preloaded by compounding pharmacy (stored up to weeks)	Compared nonprimed vs. primed syringe injections	Overall incidence: 60/6,632 (0.9%). Early 2015–2016: 0.03%; late 2016: 1.7%. Nonprimed syringes: 6.4% vs. primed: 0.5% risk (detected by dilated slit-lamp examination)	68% of affected patients had symptomatic floaters ± sparkles of light​ (others asymptomatic)	Slit-lamp examination—small, refractile spheres in vitreous	No significant IOP rises reported	No inflammation noted	No serious sequelae—floaters tended to improve over months despite droplets persisting
Melo et al, 2019	Cross-sectional study	67 eyes (37 injected vs. 30 controls)	Mixed anti-VEGF (bevacizumab, ranibizumab, aflibercept)	0.5-mL or 1.0-mL syringes (polypropylene) + 30-G needles (SR brand in use)	Low (staked insulin-style) used for injections	No compounding (drug drawn fresh into syringes)	Standard clinic technique (≥7 days after last injection)	Treated eyes: 25/37 (67.6%) had visible silicone droplets on slit lamp; controls: 0/30. B-scan ultrasound detected droplets in 28/37 (75.7%) treated vs. 1/30 controls. More previous injections → higher ultrasound droplet index (r=0.61)	Not reported (mostly incidental findings during examination)	Ultrasound B-scan: hyperechoic spots (quantified via ImageJ). Good agreement with biomicroscopy	No IOP rise noted	No inflammation noted	No adverse events noted (background: some patients with heavy droplet load elected vitrectomy in the literature)
Thompson et al, 2021​	Prospective cross-sectional	260 treated eyes; 147 controls	Mixed anti-VEGF (bevacizumab, ranibizumab, aflibercept) each in separate groups	Bevacizumab: 0.3-mL BD insulin syringe (polypropylene, 31 G staked); ranibizumab: 1-mL BD tuberculin (polypropylene) or manufacturer glass prefilled syringe; aflibercept: 1-mL BD polycarbonate syringe (provided in kit)	Bev: very low; ran: high (tuberculin) or none (glass PFS); Afl: moderate	Bevacizumab prefilled by pharmacy (Avella); ranibizumab partly in PFS (since 2017); aflibercept drawn from vial into provided syringe	All injections given by single practice; standard prep (likely syringes primed, minimal agitation)	Bevacizumab (insulin syr): 78.3% of eyes had intravitreal SO droplets; ranibizumab (1 mL or PFS): 14.4%; aflibercept (polycarbonate): 48.5%; controls: 0%. Severity was highest with insulin syringes, intermediate with polycarbonate, and lowest with tuberculin/glass (≈controls)	Mostly none—findings were largely subclinical (routine examination before injection). Patients receiving bevacizumab occasionally noted floaters, but not quantified here	Slit-lamp biomicroscopy of vitreous (12–16×)—graded 0–4+ droplets. Larger droplets visible as refractile globules; smaller ones required careful examination	No IOP rise noted	No inflammation noted	No adverse events noted
Sivertsen et al, 2021	1-year observational (“outbreak”)	50 eyes (46 patients) with symptomatic floaters; +50 eyes asymptomatic controls	Mixed anti-VEGF (majority on bevacizumab injections)	Initially: 0.5-mL insulin-type syringes with silicone (in use until November 2018); after November 2018: switched to silicone-free syringes	Very low (initially) → low dead space silicone-free	Bevacizumab compounded by hospital pharmacy (before switch); stored refrigerated preinjection	Standard technique (no specific details; change: new syringe type introduced without silicone oil)	Symptomatic cases: 50 eyes out of ∼13,429 injections (∼0.37%) in 2018. Control group (no symptoms). Postswitch: 0 new cases in 12 months after replacing syringes. Detection by slit-lamp fundus examination (often clustered near vitreous cortex)	All 50 cases eyes had acute floaters (described as clouds of small dots or bubbles) appearing after the injection. ∼79% had persistent floaters at follow-up, causing annoyance in 42%	Slit lamp: numerous transparent droplets in vitreous; often near post. hyaloid or behind lens (ultra-widefield fundus photos documented droplets). OCT was sometimes obscured by hyperreflective artifacts from droplets (not detailed in text)	No IOP rise noted	No inflammation noted	Purely vitreous floaters. Once clinic switched to silicone-free syringes, no further adverse events occurred
Avery et al, 2019	Retrospective case series	Patients from a single private practice receiving intravitreal bevacizumab injections, with noted large or symptomatic silicone oil droplets (23 cases noted in the past 5 months)	Bevacizumab (Avastin)	Single-piece insulin syringes with an attached 8-mm, 31-gauge needle (BD Ultrafine insulin syringe; Becton, Dickinson and Company)	Minimal dead space (because of attached needle, insulin syringe design)	Bevacizumab compounded at pharmacies and shipped prefilled in insulin syringes; syringes were not drawn immediately before injection but were provided by compounding pharmacies prefilled	Standard intravitreal injection (no special agitation or flicking technique described in methodology)	Marked increase in large, symptomatic droplets; 23 recent cases vs. only 1 over the previous decade. Detection was by slit lamp and B-scan ultrasound (revealing characteristic reverberation patterns)	Patients frequently reported a distinctive circular floater—a dark ring surrounding a bright central area immediately after injection. No visual acuity loss reported, but the symptoms were bothersome enough for some patients to consider vitrectomy	Prominent reverberations on B-scan ultrasonography characteristic of large silicone oil droplets, confirming their presence and size	No IOP rise noted	No inflammation noted	No visual acuity loss; adverse events were primarily limited to bothersome floaters
García et al, 2022	Cross-sectional study	152 eyes from 140 patients receiving anti-VEGF intravitreal injections	Aflibercept, ranibizumab, bevacizumab	Aflibercept: 0.3-mL BD Microfine + Demi syringes (recently also using prefilled glass syringes) Ranibizumab: prefilled glass syringe (0.165 mL, sterile solution, Luer lock adapter) Bevacizumab: 1.0-mL sterile disposable Luer lock syringes (DH Material Médico)	Syringes designed to minimize dead space (particularly the insulin and microfine types)	Aflibercept: vial split into syringes by pharmacy and stored at 5–8°C Ranibizumab: commercial prefilled glass syringes stored at 5–8°C Bevacizumab: repackaged from vials into syringes by hospital pharmacy, stored at 5–8°C	Standard intravitreal injection technique, no special agitation or flicking described	71.7% (109/152 eyes), assessed via slit-lamp examination and classified into “Scarce,” “moderate,” and “Abundant” categories based on quantity	Floaters reported by 39.5% of patients; symptom severity rated mild (∼4/10); no significant symptoms reported in central scotoma or metamorphopsia	Silicone droplets detected by slit-lamp biomicroscopy (anterior and posterior vitreous visualization)	No significant differences in intraocular pressure (IOP) related to silicone oil droplets, although the group with abundant droplets had slightly higher mean IOP.	No inflammation noted	No adverse events noted
Sanabria et al, 2021​	Cross-sectional study	142 eyes from 98 patients with wet age-related macular degeneration (wAMD), treated with anti-VEGF injections for at least 1 year	Ranibizumab, aflibercept, bevacizumab​	Disposable plastic syringes with staked-in needles (BD Microfine+/BD Ultrafine)	Not explicitly stated; however, BD Microfine+/Ultrafine syringes typically have minimal dead space because of the staked-in needle design	Repackaged at the hospital pharmacy; aliquots prepared from original drug vials into disposable BD Microfine+ syringes with staked-in needles; the pharmacy is located separately (4 km away) from the ophthalmology department	Standard intravitreal injection (no specific mention of flicking or agitation)	High incidence (89.4%, 127 of 142 eyes). Detection by mydriatic slit-lamp biomicroscopy. Quantified as scarce (1–10 droplets), moderate (10–30), or numerous (>30)​	Permanent and significant floaters reported by 36.7% of patients. However, there was no significant correlation between floaters and presence of microbubbles	Silicone microbubbles induced artifacts in OCT scans in 7.7% of eyes, obscuring retinal layer details and interfering with clinical decision making	Not reported	Not reported	Not reported
Yu et al, 2018	Retrospective case series	7 patients experiencing silicone oil droplets after intravitreal bevacizumab injections, selected from a larger group estimated at 90,413 injections performed over 10 years	Bevacizumab	Plastic insulin syringes preloaded by pharmacy (specific type/manufacturer not explicitly named, but described as insulin syringes)	Not explicitly stated; typical insulin syringes generally have minimal dead space	Preloaded syringes from pharmacy (sterile conditions), then delivered to ophthalmology offices	Standard intravitreal injection, no mention of flicking or agitation	Extremely rare, symptomatic droplets observed in only 7 cases over a period of 4 months after 90,413 injections; droplets detected via clinical examination and fundus imaging	Transient floaters developed approximately 6.7 days after injection; symptoms typically resolved within 9 months	Visible silicone oil droplets documented using Optos® fundus photography; no significant retinal pathology reported	No IOP rise noted	No inflammation noted	No adverse events related to silicone oil droplets noted
Dessouki et al, 2023​	Retrospective case series	55 patients treated with 62 intravitreal injections of pegcetacoplan	Pegcetacoplan (complement C3 inhibitor)	1-mL McKesson Luer lock syringes; needles from the Apellis pharmaceuticals injection kit	Not explicitly stated, but typical Luer lock syringes generally have some minimal dead space	Medication drawn from glass vial using filter needle (provided by Apellis); syringe was inverted with plunger repeatedly moved to manage air bubbles and prime because of pegcetacoplan's high viscosity	Standard intravitreal injection; notable agitation (flicking, pulling, and pushing plunger) to remove air bubbles because of high viscosity of pegcetacoplan	16 of 55 patients (29%) had presumed intravitreal silicone droplets	14 of 16 affected patients (88%) reported new persistent floaters, some described as numerous or “hundreds”; 2 patients (13%) asymptomatic	Presumed silicone oil droplets clearly visible on color fundus photographs as clear bubbles with distinct borders, persisting for weeks	No IOP rise noted	No inflammation noted	One patient considering vitrectomy because of severity of floaters
da Cruz et al, 2022	Randomized, double-blind, controlled clinical trial	41 patients (41 eyes) with proliferative diabetic retinopathy undergoing aflibercept injection 48 hours before vitrectomy	Aflibercept	Siliconized syringe: SR syringe (Saldanha-Rodrigues, Brazil, 1 mL) Silicone oil-free syringe: HSW Norm-Ject syringe (Germany, 1 mL) Standard needles used with syringes; no needle details explicitly mentioned	Not explicitly stated, but generally low (staked needles)	Drug drawn into syringe immediately before injection, without extended storage	Intervention group: syringe agitated by standardized flicking technique (five taps, needle down) Control group: no agitation Syringe priming explicitly prohibited	80% of agitated siliconized (SR) syringe group showed anterior chamber cells at 48 hours postinjection, compared with only 10% in the nonagitated siliconized group. Silicone-free syringes showed no significant inflammation regardless of agitation (detailed anterior chamber inflammation used as an indirect indicator of presumed silicone droplets)	No subjective symptoms (no visual change, no pain); inflammation identified solely through slit-lamp examination	Anterior chamber cell count at slit-lamp examination 48 hours postinjection; no specific vitreous imaging findings mentioned	No IOP rise noted	Transient sterile inflammation significantly associated with agitated siliconized syringes	No adverse events noted
Akbulut et al, 2022	Case series (3 cases)	Three patients receiving intravitreal anti-VEGF injections for various retinal conditions	Aflibercept and ranibizumab​	Not explicitly stated, standard intravitreal injection syringes presumed	Not explicitly stated, presumed standard preparation immediately before injection	Not explicitly stated, presumed standard preparation immediately before injection	Standard intravitreal injection technique; no mention of agitation or special preparation	Silicone oil droplets detected during routine clinical follow-up after multiple injections in all three patients	One patient reported subjective visual symptoms (floaters), whereas others had no reported symptoms	Silicone droplets identified on fundus imaging clearly visualized as distinct, reflective droplets	No IOP rise noted	No inflammation noted	No significant adverse events noted in any patient
Miller et al 2022	Prospective observational double-blind study	426 consecutive patients receiving intravitreal bevacizumab injections	Bevacizumab	Comparison of Normject syringes (control) vs. BD syringes (test)	Not explicitly stated	Syringe preparation was retrospective; patients previously received injections with either Normject or BD syringes	Standard injection technique, no specific agitation or priming method mentioned	BD syringes: 67.2% Normject syringes: 7.8% Detected via clinical examination	Increased symptomatic floaters (described as “clear circles with black outline”) occurred more frequently with BD syringes (10.4%) compared with Normject syringes (2.0%), although this difference was borderline statistically significant (P = 0.069)	Clinical examination–based detection; no specific imaging modality mentioned	No IOP rise noted	No inflammation noted	No adverse events noted

^*Syringe/needle details inferred from context when not explicitly stated (common practice circa 2006–2008). **SCORE trial used triamcinolone (steroid) rather than anti-VEGF, but findings on syringe design are applicable to anti-VEGF injection technique.

Findings: high cumulative SO prevalence, device/handling-linked outbreaks, floaters/artifacts common, no consistent IOP or inflammation signal.

Summary of 15 studies on silicone-oil contamination after intravitreal injections, listing design, devices/handling, detection method, symptoms/imaging, and intraocular pressure is presented in Table 1.

Incidence of Silicone Oil Contamination

Recent cross-sectional studies have revealed a high cumulative prevalence of SO droplets when sensitive detection methods are used. Melo et al (2019) found that 25 of 37 eyes (67.6%) with a history of anti-VEGF injections had detectable SO microdroplets on dilated slit-lamp examination (0% in control eyes). Ultrasonography further increased detection from 25 to 28 of the same set of 37 eyes (75.7%).⁴ Similarly, Banderas García et al (2022) observed vitreous SO droplets in 109 of 152 treated eyes (71.7%), and Sanabria et al (2021) reported microdroplets in 127 of 142 eyes (89.4%) after 1 year of ongoing anti-VEGF therapy.^5,6 Of the studies included in this review, three found a positive correlation between the number of injections received and the burden of intravitreal silicone.^4–7

Four large retrospective series included in this review documented that a small fraction of injections result in clinically evident droplets. Khurana et al⁸ (2017) observed 60 cases of intravitreal SO droplets out of 6,632 bevacizumab injections, corresponding to an incidence of 0.9%. Over 12 months, the incidence in their practice initially was as low as 0.03%, but it spiked to 1.7% in the latter half of the year after a change in syringe manufacturing.⁸ In another analysis, Yu et al⁹ (2018) identified only 7 symptomatic cases among 90,413 bevacizumab injections over a decade (an incidence of 0.0077%). Notably, the 7 cases clustered within a 4-month interval​. These reports highlight that “outbreaks” of SO droplet cases tend to cluster in time or by injector, likely triggered by specific syringe batches or lapses in injection technique. Likewise, Avery et al¹⁰ (2019) reported a sudden surge of 23 cases of large, symptomatic silicone floaters over 5 months in one practice, compared with only 1 case in the previous decade, after switching to a particular brand of insulin syringe​. Similarly, Sivertsen et al (2021) documented an incidence of 50 eyes with silicone floaters over 13,429 injections (0.37%) during 6 months. All cases occurred when using a silicone-coated insulin syringe, with no new cases occurring after the clinic switched to silicone-free syringes.¹¹

Among the studies reviewed, symptomatology associated with intravitreal SO droplets varied widely. Of the 13 studies that addressed visual symptomatology, 10 reported that some proportion of patients experienced floaters in their vision. In controlled trials and large series where silicone microdroplets were detected incidentally (e.g., the SCORE study and Melo et al), patients generally did not notice them.^4,12 Even in the aforementioned high-prevalence studies (Melo, García, Sanabria), only about one third of patients reported perceiving floaters or visual symptoms from the droplets.^4–6 However, when SO droplets are larger or more numerous, patients are more likely to become symptomatic. Khurana et al⁸ reported that 68% of their cases with visible droplets noted new floaters. Similarly, Sivertsen et al¹¹ described 50 cases experiencing an acute shower of tiny dots or bubbles in vision after injection, with 79% experiencing persistent floaters at 1-year follow-up. The limitation with these large retrospective studies is that the authors did not report whether the vitreous was carefully examined for silicone oil droplets or whether only the most severe cases were documented in the medical records, leading to substantial ascertainment bias.

Factors Contributing to Silicone Oil Transmission

Multiple studies have demonstrated that improper storage conditions and handling can lead to increased formation of SO droplets in injection syringes. Liu et al¹³ found that a single freeze–thaw cycle of bevacizumab in plastic insulin syringes resulted in particle counts on the order of 1.2 million per mL, representing a roughly fourfold increase compared with nonfrozen controls. Repeated freeze–thawing compounded the contaminant burden.¹³ By contrast, bevacizumab stored at the recommended refrigerator temperature (4°C) without freezing did not exhibit such large particle increases. Consistent with these laboratory findings, Melo et al¹⁴ warn that even brief exposure to subzero temperatures (e.g., cold spots in a refrigerator or ice packs in transit) can cause repackaged syringes to freeze and later thaw, potentially increasing the release of the drug with SO droplets.

Mechanical agitation is another key factor that increases SO release. Simulated shipping “shock” experiments demonstrated that rough handling can cause a increased number of droplets to be shed from the syringe's siliconized interior. In Liu et al.’s study, syringes tossed vigorously to mimic transport-induced mechanical stress contained over a million particles per mL, far exceeding the particle count of undisturbed syringes.¹³ Funke et al¹⁵ reported that shaking siliconized containers (with air headspace) led to the highest levels of shed silicone particles, whereas gently expelling the fluid produced minimal additional particles​. The duration of storage also has an impact on the formation of particulate matter in syringes used for intravitreal injections. Specifically, oleamide-lubricated plastic transfer syringes demonstrate marked time-dependent particle formation, with particle loads notably increasing after storage periods beyond 30 days. By 60 and 90 days, substantial increases (up to 4- to 6-fold higher than baseline) are evident, occasionally resulting in visible particle formation within the syringe solution.¹⁶

The physical properties of the injection syringe and needle, including the siliconization method, design (staked vs. detachable needle), dead space volume, and construction material, strongly influence the amount of SO released. Most commercial syringes are lubricated with polydimethylsiloxane to ensure a smooth plunger glide.¹⁷ Traditional disposable syringes (e.g., insulin syringes) use a spray-on siliconization, depositing approximately 0.2 to 1.0 mg of liquid silicone into each syringe barrel.¹⁵ In contrast, newer prefilled syringes (Becton, Dickinson and Company; Franklin, New Jersey, USA) often use a baked-on silicone coating, in which a diluted silicone oil emulsion is evenly spread on the interior and then heat-cured onto the glass surface.¹⁸ Funke et al showed that particle shedding is efficiently minimized with low baked-on silicone load. In their experiments using baked-on siliconized cartridges, aggressive agitation produced particle counts predominantly composed of SO droplets in the range of 10⁵ particles/mL. By comparison, studies of conventional spray-siliconized syringes report particle levels an order of magnitude higher (∼10⁶ particles/mL).¹⁵

The syringe type and dead space volume are also critical factors in the transmission of SO droplets. A key distinction can be made between syringes with staked-on (integral) needles and those with detachable needles (Luer-slip or Luer-lock). In staked-needle designs, such as by 0.3-mL insulin syringes, the needle is fixed and the “dead space” (fluid remaining in the hub after the plunger is fully depressed) is very small. This allows nearly the entire drawn 0.05 mL to be injected, including any SO microdroplets suspended in solution.¹⁴ Although having negligible dead space may be desirable to minimize waste of medication, a less-recognized drawback of zero dead space syringes (Becton, Dickinson and Company; Franklin, New Jersey, USA; SJJ Solutions B.V.; Netherlands) is that they cannot retain stray droplets. Laboratory tests have confirmed that syringes with no hub dead space are more likely to expel SO if it is present than syringes with a small dead space.¹⁴ In contrast, syringes with detachable needles (e.g., many 1-mL tuberculin syringes) have a hub that holds residual volume after injection.¹⁹ This dead space phenomenon was linked in one report of an outbreak: an American Society of Retina Specialists alert in 2016 noted a cluster of patients with intravitreal SO droplets directly linked to a type of low dead space insulin syringe.¹⁴

Plastic syringes, typically composed of polypropylene, generally require a silicone coating to achieve an acceptable plunger glide. In contrast, glass syringes can be manufactured with very low friction, even with minimal silicone. The ranibizumab prefilled glass syringe (Genentech; Canada), for example, is only lightly siliconized and has a low void volume, which reduces both air bubbles and SO droplet injection.^14,17 Needles may also be siliconized by manufacturers to ease insertion through tissue. If a needle is prelubricated with silicone, it could introduce trace amounts during the injection process.^20,21 Furthermore, larger-bore needles have greater surface area and can carry or dislodge more material. In an experimental model using fluorescent microspheres to simulate external contamination, Nakashizuka et al²² found that a 27-gauge needle introduced significantly more fluorescent microspheres into the eye compared with finer 30-G or 32-G needles under identical conditions.

Beyond storage and syringe selection, the injection technique itself plays a pivotal role in minimizing SO transmission. It is common to tap or flick the syringe barrel to aggregate and remove air bubbles before injection. However, this action can be a deliberate agitation of the fluid, which may dislodge SO from the walls. A laboratory study by Souza Dias Júnior et al quantified this effect, showing that when syringes were agitated by flicking, 93% of them released detectable SO droplets. A single vigorous flick increased the odds of SO droplet release by 265-fold relative to no agitation.²³ A randomized control trial by Santos da Cruz et al²⁴ of 41 patients showed that in eyes receiving an intravitreal aflibercept injection, if the syringe was agitated with five firm taps before injection, 80% of those eyes showed an inflammatory cell reaction in the anterior chamber 48 hours later (possibly provoked by SO microdroplets) compared with only 10% of eyes injected without preinjection agitation (using the same type of syringe).

Priming the needle by expelling a small amount of solution before injection can help flush out any silicone-rich fluid that has collected in the hub. Because SO has a higher viscosity than the aqueous solution, a greater injection force on a nonprimed syringe may increase the likelihood of SO passing through the needle.⁸ In a comparative study, Khurana et al⁸ reported a markedly lower incidence of visible silicone droplets with primed versus nonprimed injections using identical insulin syringes, 0.5% versus 6.4%, respectively. Overall, minimizing plunger manipulation and using steady, moderate force is recommended to reduce the risk of introducing silicone oil into the eye.^4,14,23

Discussion

Collectively, these technique modifications can meaningfully reduce SO transmission during intravitreal injections. These precautions are supported by clinical observations: eyes that have undergone many injections often contain appreciable amounts of SO, and the cumulative burden correlates with the number of injections. The literature reviewed here strongly supports that proper storage, careful syringe selection, and refined injection technique can significantly reduce the inadvertent transfer of SO during intravitreal anti-VEGF therapy.

Repeated intravitreal anti-VEGF injections have been associated with sustained increases in intraocular pressure, which is a concerning complication potentially exacerbated by SO contamination.^25–28 Hypothetically, SO droplets may physically obstruct the trabecular meshwork, thereby mechanically reducing outflow. Indirectly, SO droplets can provide a surface for protein adsorption and aggregation within the aqueous humor, further clogging the trabecular meshwork, contributing to persistent ocular hypertension.^13,15,29 However, in 13 of the 15 studies included in this review that examined intraocular pressure, there were no significant elevations across the reports.

Sterile inflammation after intravitreal injections has been increasingly recognized as a clinically significant complication.^30,31 Hahn et al reported a comprehensive analysis of aflibercept-related sterile intraocular inflammation, emphasizing the role of immune activation and protein adsorption as potential mechanisms. SO droplets introduced via silicone-lubricated syringes and needles may adsorb proteins, altering their conformation and rendering them immunogenic.³⁰ In the studies evaluated in this review, none explicitly reported sterile inflammation attributed to SO droplets, as multiple sources indicate that such droplets are frequently present in the vitreous postinjection but remain asymptomatic and do not elicit noticeable immune responses or clinical complications.

One common complication found in this review is that SO droplets may cause visual disturbances, manifesting clinically as mobile visual opacities or “floaters” that patients perceive.^8,10 Once introduced, these droplets may persistently remain within the vitreous, making them challenging to remove without vitrectomy. The prevalence and symptomatic severity of SO droplets can vary based on factors such as syringe design, injection technique, and whether syringes are primed before injection, highlighting critical considerations for clinical practice and patient counseling.^8,12

One practical implication of these findings is the effect on diagnostic imaging and monitoring. SO droplets can confound ocular imaging modalities, most notably optical coherence tomography (OCT). Sanabria et al⁵ reported that approximately 7.7% of eyes with intravitreal silicone exhibited noticeable artifacts on OCT scans, including hyperreflective spots that obscured retinal details. These artifacts could be mistaken for pathology (e.g., vitreous opacities or cells) if the clinician is unaware of the silicone. Knowing that a patient has silicone floaters might explain unexpected OCT reflections and prevent misinterpretation.

Perhaps the most impactful intervention is to use syringes that are less prone to shedding SO. Silicone-free, sterile syringes (made of plastic, such as polypropylene or cyclo-olefin polymer, but manufactured without silicone) are now available and have been tested for ophthalmic use (Gerresheimer AG; Germany, Terumo Medical Corporation; Somerset, New Jersey, USA, Merit Medical Systems; South Jordan, Utah, USA, SJJ Solutions BV; Netherlands, Ocuject LLC; Salt Lake City, Utah, USA, etc.).^4,14,16 On the research front, alternative drug delivery systems that eliminate the need for traditional syringes, such as microinjectors, novel vial adapters, or even needle-free jet injectors, are being developed for ocular use.^14,17,32 Although such technologies are in early stages, they represent potential future directions for enhancing patient safety. In addition, continued pharmacovigilance is crucial. Tracking any new cases of silicone droplet–related issues will help identify if further improvements are necessary.

Although this narrative review provides a comprehensive evaluation of SO contamination during intravitreal anti-VEGF injections, several limitations should be acknowledged. First, the heterogeneity of study designs, outcome measures, and detection methods across the included literature precluded formal meta-analysis. Variability in how SO droplets were identified (e.g., slit-lamp biomicroscopy, ultrasonography, OCT artifacts) introduces differences in sensitivity and diagnostic thresholds, limiting direct comparability across studies. Second, many studies relied on retrospective data, which are inherently subject to selection bias and incomplete documentation. In addition, the duration of follow-up varied widely, with few studies providing long-term data on the natural history of SO droplets or their potential delayed effects. Third, a substantial proportion of the literature stems from single-center experiences or case series, which may limit generalizability. Geographic and practice-specific factors, such as access to silicone-free syringes, compounding protocols, and injection techniques, vary significantly between institutions and countries.

Conclusion

In conclusion, this review highlights that SO microdroplet contamination is an underrecognized but preventable phenomenon from intravitreal therapy. Although the most common effects of these microdroplets include transient floaters and imaging artifacts, their presence reflects a modifiable risk with uncertain long-term risks. Ongoing improvements in syringe design and injection protocols, guided by the evidence summarized here, offer promising opportunities to reduce the incidence of SO droplets and minimize symptomatic floaters observed by some patients.