A Prediction Model for the Intraocular Pharmacokinetics of Intravitreally Injected Drugs Based on Molecular Physicochemical Properties

Purpose: This study was aimed at determining the intraocular pharmacokinetics based on molecular physicochemical properties in a rabbit model. Methods: The entire dataset was obtained from previous literature, and research articles regarding 70 molecular compounds were investigated. The intravitreal half-lives in rabbit eyeballs of 22 macromolecules and 48 micromolecules were analyzed. Multiple linear regression analysis was carried out with non-collinear independent variables (molecular weight [MW] and lipophilicity) influencing intravitreal half-lives. The best-fit equations were selected based on the correlation coefficients and goodness-of-fit statistics. Results: The best-fit models obtained from the entire dataset, macromolecules, and micromolecules suggest the correlation between molecular physicochemical properties (MW and lipophilicity) and intravitreal half-life. Exclusion of outlier molecules (amphotericin B and foscarnet) leads to a better-fit correlation. MW is the definite single factor affecting intravitreal half-lives of macromolecules (Log t1/2 = 0.148 + 0.370 Log MW, R2 = 0.769), while both MW and lipophilicity influence the intraocular pharmacokinetics of micromolecules (Log t1/2 = –1.213 + 0.762 Log MW – 0.115 Log p, R2 = 0.554). Conclusion: The present study indicates that intravitreal half-life could be predicted based on molecular physicochemical properties (MW and lipophilicity). Also, increasing MW while reducing lipophilicity would be a reliable method for prolonging the intravitreal half-life of small chemical drugs, while MW is the single major determinant for large biologic drugs.