Observation of Silicone Oil Within the Vitreous and Sclera Following Intravitreal Administration of Biotherapeutics Using Insulin Syringes in Cynomolgus Monkeys

Silicone oil droplets have been reported in the eyes of human patients following intravitreous (IVT) injections with several marketed biotherapeutic products. Intravitreous administration of a novel biotherapeutic in a 14-week cynomolgus monkey study using insulin syringes was associated with 2, non-test-article-related phenomena: “vitreous floater/clear sphere” on indirect ophthalmoscopy and intrascleral “foreign material near injection track” on histopathology. Retrospective analysis of 81 other preclinical studies of IVT administration of novel biotherapeutics found a greater frequency of clear spheres in monkey IVT studies using insulin syringes and formulations containing polysorbate. We were able to correlate microscopic findings of clear circular to oval areas in the sclera near the injection track with an energy-dispersive X-ray spectroscopy (EDS) signal for silicon at the same location in the sclera. These observations provide further evidence that silicone lubricant in insulin syringes/needles is the source of clear spheres noted in the vitreous and foreign material noted near the injection track in the sclera. Although considered inert and toxicologically insignificant, silicone deposition within the eye should form part of the risk–benefit equation in a clinical setting.

Silicone Oil Droplets are More Common in Fluid From BD Insulin Syringes as Compared to Other Syringes

Purpose: During 2016 and 2017, numerous retina specialists reported silicone oil droplets after intravitreal injection, primarily associated with insulin syringes with a fixed needle. By contrast, adverse events were relatively rare for tuberculin syringes. We hypothesized that fixed-needle syringe injections have more silicone droplets due to the lack of dead space at the syringe–needle junction. Methods: Fluid from Becton Dickenson (BD) insulin syringes, BD tuberculin syringes, Henke Sass Wolf (HSW) insulin syringes, and HSW silicone-free syringes was examined microscopically for silicone oil droplets. Fluid was sampled at the beginning, middle, and end of injection; and droplets were photographed and quantified. Results: Silicone oil droplets were identified in 30% of the BD insulin syringes, measuring 0.9 ± 0.2 mm3; no oil droplets were found in BD tuberculin syringes or HSW silicone-free syringes (analysis of variance, P < .05). All 6 samples with silicone oil droplets were from the end of injection (full depression of the plunger), while samples from the beginning and middle of injections did not have droplets. Challenging syringes with extra silicone oil resulted in increased droplets for fixed-needle syringes (but not for syringes with detachable needles). Conclusion: Significant silicone oil droplets were present in BD insulin syringes and were expressed at the end of injection (with plunger depressed maximally), supporting the hypothesis that insulin syringe injections have more droplets due to the lack of dead space where the plunger meets the hub of the needle. However, there are additional factors such as the amount or type of silicone lubricant in a syringe.

High Shear Stress Behavior of Some Representative Lubricants

Shear stress independent behavior was observed for representative, synthetic, nonblended lubricants to about 4.8 × 106 N/m2 (700 psi) shear stress in high pressure viscometric measurements. This shear stress is of the same magnitude as the shear stress in sliding elastohydrodynamic contacts. It is shown that dissipation heating is the only mechanism of importance in the generation of the deviations from constant viscosity as measured with capillary tube viscometric methods. The Newtonian end corrections for the capillary tubes were found to be constant for the nonblended, liquid lubricants. Newtonian behavior will be expected of the fluids in a high shear lubrication situation. Shear induced, nonliquid behavior was found for the silicone lubricant at about 106 N/m2 and for the polymer-blended mineral oil at about 104 N/m2 at a relatively low pressure level. The observations might provide a key to an understanding of the generation of the anomalous low elastohydrodynamic film thickness as found with these lubricants. The polymer-blended oil showed shear thinning effects. The apparent viscosity was found to increase (∼30 percent) with increasing shear stress in the range of the second Newtonian viscosity level.