(T1030-05-28) Enhanced Intravitreal Pharmacokinetics in the Simcyp™ Ocular Model

Purpose: Accurate modelling of the pharmacokinetics (PK) of the vitreal cavity and its surrounding tissues are crucial for the development of drugs and formulations for diseases such as macular degeneration that affect hundreds of millions worldwide. These diseases are a leading cause of low visions and blindness [1], but are treatable, with one of the most popular treatments involving the use of intravitreal (IVT) injections. Since the drug is injected into a large cavity filled with a viscous medium, the inclusion of diffusive processes in the modelling are critical to make reasonable predictions of the time course of the compounds, and in turn to accurately determine the PK of the target tissues that surround the cavity. The current work aimed to develop and validate the algorithms necessary to capture these phenomena within the Simcyp Ocular module.

Methods: We present here a model of the vitreal cavity using the finite element method (FEM) that replaces the single compartment that was previously available, within the greater context of the complete Simcyp Ocular model. The vitreal cavity geometry is represented in all three dimensions with accurate spatial connections to the retina, iris and ciliary-body, aqueous humour and lens. With the introduction of a physically realistic topology, we have also added parameters to calculate the injection geometry. This allows for variability in the injection location, both intentional and otherwise. Multiple validation studies have been performed, beginning with IVT injections of candesartan into pre-clinical (rabbit) species for which full time scale results exist in the literature [2]. A rabbit physiology was created in the ocular module, and along with the use of first-order absorption and minimal PBPK, five trials of 10 rabbit subjects were performed. All parameters used were either the default or predicted values provided by Simcyp, with the exception of the vitreous humor diffusivity that was predicted using the Stokes-Einstein relation and a viscosity of 6.7cP for the liquid humor. The variability of the dose volume was also reduced to a CV of 10%. In addition to candesartan, we ran population simulations to compare against human data for IVT injections of vancomycin (200 μg) and gentamicin (50 μg) [3]. Both simulations were run over 5 trials of 17 ‘Healthy-Volunteers’ to match the clinical data, and also used the default and predicted parameters provided by Simcyp. All simulations were conducted in Simcyp Simulator V24 unless otherwise noted

Results: We were able to recover the observed vitreal exposure of candesartan within the population’s confidence bands (see Figure 1), as compared with the single compartment model previously available in Simcyp V23 that clears at a substantially higher rate than the observed data. Further optimization of either the diffusivity or the vitreal clearance lead to highly accurate reproductions of the observed data (see poster). The human population runs for vancomycin and gentamicin matched both the mean and variance of the data from the published work (see Figures 2 and 3) without any further optimization of the parameters. The gentamicin shows 92% of the population within the confidence bands (total width of 90%), while the vancomycin shows 88% of the subjects within the bands.

Conclusion: The vitreal cavity is the largest volume within the eye, filled with viscous fluid and collagen fibrils, and for which diffusion plays an outsize role in the modelling of transport kinetics within it. The validation results show that the 3D FEM model of the vitreal cavity is able to accurately capture the inter-subject and dosing geometry variability without any optimization for the three studies simulated. More complex systems may require appropriate handling of metabolic clearance or melanin binding, both of which are also provided for in the ocular model. Many other IVT dosed therapeutics, such as anti-VEGF treatments, come in the form of monoclonal antibodies [4]. Although the Simcyp ocular model is currently focused on small molecules, the extension of the vitreal FEM compartment for use with biologics and large molecules – for which diffusive processes become even more important with the increased size of the drugs – will be included in future releases.

References: 1. Zhang, S., et al., Global burden of low vision and blindness due to age-related macular degeneration from 1990 to 2021 and projections for 2050. BMC Public Health, 2024. 24(1): p. 3510.
2. Lee, J.E., et al., Intraocular Toxicity and Pharmacokinetics of Candesartan in a Rabbit Model. Investigative Ophthalmology & Visual Science, 2011. 52(6): p. 2924-2929.
3. Gan, I.M., et al., Intravitreal vancomycin and gentamicin concentrations in patients with postoperative endophthalmitis. Br J Ophthalmol, 2001. 85(11): p. 1289-93.
4. Tah, V., et al., Anti-VEGF Therapy and the Retina: An Update. J Ophthalmol, 2015. 2015: p. 627674.


Figure 1: Simulated (solid line) versus observed (blue points) candesartan mean concentration of the vitreous humor in New Zealand white rabbits. The dashed line shows the simulated exposure using a single compartment model previously available in Simcyp V23. The diffusivity was predicted using the Stokes-Einstein relation for the liquid humour, giving a value of 4.0e-5 cm2/s; no other phenomena (such as metabolic clearance) was considered.


Figure 2: Concentration profile of vancomycin after an IVT injection of 200 μg into a population of ‘healthy volunteers’. The diffusivity was predicted at 2.36e-5 cm2/s, the dose volume CV was reduced to 10%, while all other ocular parameters used the default or predicted values provided by Simcyp v24.


Figure 3: Concentration profile of gentamicin after an IVT injection of 50 μg into a population of ‘healthy volunteers’. The diffusivity was predicted at 3.66e-5 cm2/s, the dose volume CV was reduced to 10%, while all other ocular parameters used the default or predicted values provided by Simcyp v24.