Quantitative pharmacokinetic analyses of anterior and posterior elimination routes of intravitreal anti-VEGF macromolecules using published human and rabbit data.

The purpose of this study was to evaluate the contribution of the anterior elimination route for four anti-vascular endothelial growth factor (anti-VEGF) macromolecules (aflibercept, bevacizumab, pegaptanib and ranibizumab) after intravitreal injection using published human and rabbit data and three previously described pharmacokinetic (PK) modeling methods. A PubMed search was used to identify published studies with concentration-time data. The data were utilized only if the intravitreally injected drugs were used as plain solutions and several criteria for a well-performed PK study were fulfilled. The three methods to analyze rabbit data were (1) the equation for vitreal elimination half-life based molecular size assuming anterior elimination, (2) Maurice equation and plot for the ratio of aqueous humor (AH) to vitreal concentration assuming anterior elimination, and (3) the equation for amount of macromolecule eliminated anteriorly based on the area under the curve in AH. The first and third methods were used for human data. In the second and third methods, AH flow rate is a key model parameter, and it was varied between 2 and 3 µl/min. The methods were applied to data from 9 rabbit studies (1 for aflibercept, 5 for bevacizumab, and 3 for ranibizumab) and 5 human studies (1 for aflibercept, 3 for bevacizumab, and 1 for ranibizumab). Experimental half-lives of anti-VEGF macromolecules in both vitreous and aqueous humor were close to those calculated with the equations for vitreal elimination half-life in humans and rabbits. Rabbit data analyzed with Maurice plot indicated that the contribution of anterior elimination was usually at least 75%. In most human and rabbit studies, the calculated percentage of anterior elimination was at least 51%. Variability between studies was extensive for bevacizumab and ranibizumab. The results suggest that the anterior elimination route dominates after intravitreal injection of anti-VEGF macromolecules. However, the clinical data are sparse and variability is extensive, the latter emphasizing the need of proper experimental design.

Extended Pharmacokinetic Model of the Intravitreal Injections of Macromolecules in Rabbits. Part 2: Parameter Estimation Based on Concentration Dynamics in the Vitreous, Retina, and Aqueous Humor.

PURPOSE: To estimate the diffusion coefficients of an IgG antibody (150 kDa) and its antigen-binding fragment (Fab; 50 kDa) in the neural retina (Dret) and the combined retinal pigment epithelium-choroid (DRPE-cho) with a 3-dimensional (3D) ocular pharmacokinetic (PK) model of the rabbit eye. METHODS: Vitreous, retina, and aqueous humor concentrations of IgG and Fab after intravitreal injection in rabbits were taken from Gadkar et al. (2015). A least-squares method was used to estimate Dret and DRPE-cho with the 3D finite element model where mass transport was defined with diffusion and convection. Different intraocular pressures (IOP), initial distribution volumes (Vinit), and neural retina/vitreous partition coefficients (Kret/vit) were tested. Sensitivity analysis was performed for the final model. RESULTS: With the final IgG model (IOP 10.1 Torr, Vinit 400 µl, Kret/vit 0.5), the estimated Dret and DRPE-cho were 36.8 × 10-9 cm2s-1 and 4.11 × 10-9 cm2s-1, respectively, and 76% of the dose was eliminated via the anterior chamber. Modeling of Fab revealed that a physiological model parameter "aqueous humor formation rate" sets constraints that need to be considered in the parameter estimation. CONCLUSIONS: This study extends the use of 3D ocular PK models for parameter estimation using simultaneously macromolecule concentrations in three ocular tissues.

Extended Pharmacokinetic Model of the Rabbit Eye for Intravitreal and Intracameral Injections of Macromolecules: Quantitative Analysis of Anterior and Posterior Elimination Pathways.

PURPOSE: To extend the physiological features of the anatomically accurate model of the rabbit eye for intravitreal (IVT) and intracameral (IC) injections of macromolecules. METHODS: The computational fluid dynamic model of the rabbit eye by Missel (2012) was extended by enhancing the mixing in the anterior chamber with thermal gradient, heat transfer and gravity, and studying its effect on IC injections of hyaluronic acids. In IVT injections of FITC-dextrans (MW 10-157 kDa) the diffusion though retina was defined based on published in vitro data. Systematic changes in retinal permeability and convective transport were made, and the percentages of anterior and posterior elimination pathways were quantified. Simulations were compared with published in vivo data. RESULTS: With the enhanced mixing the elimination half-lives of hyaluronic acids after IC injection were 62-100 min that are similar to in vivo data and close to the theoretical value for the well-stirred anterior chamber (57 min). In IVT injections of FITC-dextrans a good match between simulations and in vivo data was obtained when the percentage of anterior elimination pathway was over 80%. CONCLUSIONS: The simulations with the extended model closely resemble in vivo pharmacokinetics, and the model is a valuable tool for data interpretation and predictions.