In Vitro to ex Vivo/In Vivo Correlation (IVIVC) of dissolution kinetics from inhaled particulate solutes using air/blood barrier models: Relation between in vitro design, lung physiology and kinetic output of models

Abstract

Correlating data from in vitro dissolution testing of inhaled particulate pharmaceuticals or air pollutants with the dissolution and distribution to the blood circulation of the same aerosols following inhalation in vivo is a major challenge. Better methods could benefit both inhaled drug development and the risk assessment of air pollutants. Five properties known to influence the results produced by in vitro methods are the particle deposition pattern, barrier thickness, surfactant distribution, perfusion strategy and distribution volume of dissolution. The purpose with this manuscript was to compare the rates of dissolution- and absorption of two particulate solutes, nicotine and Fluticasone Propionate (FP), in the Isolated, ventilated, and Perfused Lung of the rat (IPL) with that of non-biological test barriers simulating the same process in vitro. Three published test model barriers were compared; the DissolvIt with a mucous gel simulant and data for two other models, utilizing particle deposition on glass fiber filters; one Paddle over Disk model, and one Modified Transwell model. The more water-soluble nicotine, only tested in the ex vivo IPL and the in vitro DissolvIt model, had a very rapid first absorption half-time in both the IPL and the DissolvIt of respectively 1.9 and 0.55 min. For more lipophilic FP (Flixotide, MDI) the correlation of dissolution- and absorption data was compared for the inhaler formulation between the ex vivo rat IPL and all three in vitro test models. The results varied substantially between the in vitro models regarding simulated appearance in the blood circulation and, thus, regarding correlation with the ex vivo IPL model. The first half-time of in vitro dissolution- and absorption of FP to perfusate when normalized to the 8.6 h half-time measured for FP in the ex vivo IPL ranked as; 2.9, 0.021, and 0.17 for respectively the DissolvIt, the Paddle over Disk model, and the Modified Transwell model. Despite similar physicochemical properties of the solvents involved in all three in vitro models, it seems like resemblance of the models with the in vivo physiology has a profound influence on the dissolution- and absorption data produced.