Abstract Excised rabbit, porcine, and bovine eyes were injected in the suprachoroidal space with a cassette of 27 drugs, including 10 beta-blockers, 8 NSAIDs, and 9 corticosteroids, spanning a lipophilicity range (LogD 7.4 : −4.18 to 4.70). Drug concentrations in retina and vitreous humor were quantified by LC-MS/MS on both injected and non-injected sides. Most drugs were detected in both tissues, with higher concentrations on the injected side (0.04–3.69 fraction dose or FD/g) compared to the non-injected side (0.00–0.74 FD/g), after excluding eight outliers (drug–prodrug pairs and unstable drugs). Total retinal concentrations (0.07–1.88 FD/g) exceeded vitreous concentrations (0.02–0.40 FD/g). Across species, delivery decreased with eye diameter (rabbit > porcine > bovine). For individual species, multiple linear regression explained 70–83% of variance in retinal concentrations and 45–62% in vitreous concentrations using drug physicochemical properties. A unified equation incorporating 1/d² (d is eye diameter in cm), lipophilicity (LogD 7.4 ), molar solubility (LogS), and total polar surface area (TPSA) predicted concentrations in all species for a given tissue, explaining 74% of retinal and 93% of vitreous variance. Retinal delivery scaled for all species: Log FD/g = -1.121 + 0.255*LogD 7.4 + 0.106*LogS + 0.004*TPSA + 0.395*1/d 2 . Vitreous humor delivery scaled for all species: Log FD/g = -1.898–0.047*LogD 7.4 − 0.051*LogS + 0.001*TPSA + 0.833*1/d 2 . While this study evaluated static barriers, in vivo delivery is expected to be lower due to dynamic clearance mechanisms.
Patil et al. (Fri,) studied this question.