Experimental and mathematical approaches for drug delivery for the treatment of wet age-related macular degeneration

“

TL;DR

This review summarizes polymer-based drug delivery systems (particulate systems, hydrogels, implants, and composites) for sustained intravitreal anti-VEGF release in wet AMD. It also covers mathematical models for drug release (diffusion-, swelling-, and erosion-controlled) and ocular pharmacokinetic models (compartmental and physiologically based) that can be coupled to optimize DDS design. The clinical significance lies in reducing injection frequency, improving patient compliance, and lowering costs.

Full summary

4,867 CHARS

**Background** Age-related macular degeneration (AMD) is a leading cause of visual impairment globally, with wet AMD driven by choroidal neovascularization mediated by vascular endothelial growth factor (VEGF). Current standard therapy involves monthly intravitreal injections of anti-VEGF agents (ranibizumab, bevacizumab, aflibercept), which suffer from poor patient compliance, high costs, and complications. To address these limitations, numerous polymer-based drug delivery systems (DDSs) have been developed to extend drug release. This review covers experimental DDSs and complementary mathematical modeling approaches for predicting drug release and ocular pharmacokinetics.

**Methods** The authors conducted a narrative review of the literature, focusing on four categories of polymer-based DDSs for intravitreal delivery: particulate systems (nano/microspheres and capsules), non-stimuli responsive hydrogels, implants, and composite systems. For each category, they summarized preparation methods, polymer materials, drug release profiles, and in vitro/in vivo bioactivity. They also reviewed mathematical models for drug release classified by mechanism: diffusion-controlled (Fickian, reservoir/matrix systems), swelling-controlled (semi-empirical power law, mechanistic moving-boundary models), and erosion-controlled (surface vs. bulk erosion, stochastic cellular automata, deterministic reaction-diffusion models). Ocular pharmacokinetic models were divided into classical compartmental (one-, two-, three-compartment) and physiologically based (3D geometry, convection-diffusion) models.

**Key Results** - Particulate systems: PLGA microparticles loaded with ranibizumab showed ~70% release over 26 days (Zhang et al.). Bevacizumab-loaded PLGA nanoparticles released ~18% over 7 days at pH 7.4 (Sousa et al.). PLGA/PCADK microspheres achieved 59.6–75.7% cumulative release over 50 days (Liu et al.). Sunitinib-loaded PLGA-PEG microparticles released for ~85 days with minimal burst (Tsujinaka et al.). PolyActive microparticles released anti-VEGF for 6–18 months (Adamson et al.). Polydopamine nanoparticles released bevacizumab completely in 3 months under 10 mM H2O2 (Jiang et al.). Chitosan-PCL microparticles released for 6 months (Jiang et al.). - Hydrogels: Silk hydrogels released 40–62% of bevacizumab over 91 days (Lovett et al.). Hyaluronic acid-dextran hydrogels maintained therapeutic bevacizumab levels for up to 6 months in rabbits (Yu et al.). - Implants: HPMC implants released 1.9% of bevacizumab over 96 h; in vivo, bevacizumab was present at ~4.72 ng/mL after 12 weeks (Burgalassi et al.). PHBV implants released 29.3% over 32 days (Pakdel et al.). Bi-layered chitosan-PCL implants released 25–45% of bevacizumab over 9 months (Jiang et al.). PCL implants released 54–100% of triamcinolone acetonide over 180 days (Annuryanti et al.). PEGDA implants released 25.61–53.51% over 252 days (McAvoy et al.). - Composite systems: PLGA nanoparticles in hyaluronic acid hydrogel released 11.70–14.94% over 56 days (Hsu et al.). Chitosan nanoparticles in HA/ZnSO4 implant released 47% over 60 days (Badiee et al.). Serpin-derived peptide NPs in PLGA MPs maintained anti-angiogenic effect for ≥14 weeks (Shmueli et al.). Bevacizumab-loaded PLA NPs in PLGA MPs released 81% over 4 months (Yandrapu et al.). Chitosan-based NPs in PLGA MPs released ranibizumab over 180 days (Elsaid et al.). - Mathematical models: Diffusion-controlled models include Fick’s law solutions for spheres, cylinders, core-shell, and multi-layer geometries. Swelling-controlled models include the Peppas-Korsmeyer power law and mechanistic moving-boundary models. Erosion-controlled models include Hopfenberg’s surface erosion model, cellular automata (Zygourakis, Göpferich), and deterministic reaction-diffusion models (Zhang, Ford Versypt, Batycky, Faisant). - Ocular PK models: Compartmental models (one-compartment by Sarkhel et al.; two-compartment by Rimpela et al.; three-compartment by Tamhane et al. and Dosmar et al.) and physiologically based models (3D rabbit eye by Park et al.; 3D human eye by Stay et al., Jooybar et al., Li et al., Ferreira et al.).

**Clinical Implications** Polymer-based DDSs can significantly extend the duration of anti-VEGF release, potentially reducing injection frequency from monthly to every 3–6 months or longer. This improvement could enhance patient compliance, lower healthcare costs, and decrease complication risks. Mathematical modeling, when coupled with PK models, can accelerate DDS development by predicting in vivo performance and reducing experimental burden. However, challenges remain: many DDSs show initial burst release, potential toxicity, and need for surgical implantation. Future directions include dual/triple therapies, PROTACs, and gene delivery to address the multifactorial pathogenesis of AMD.