Development of computational fluid dynamics methodology for characterization of exhalation delivery system performance in a nasal airway with Draf-III surgery

Abstract

Computational fluid dynamics (CFD) has been used to investigate the intranasal deposition of corticosteroid sprays. CFD Lagrangian tracking models terminate particle motion once particles come into contact with a surface (“trap” boundary condition) and disregard the liquid film buildup on the mucosal surfaces of the sinonasal cavity walls. However, the interaction between spray droplets and the mucosal surface is critical for corticosteroid delivery using the exhalation delivery system (EDS) nasal spray device. Droplets coalesce on sinonasal walls, forming a thin “wall-film” on which droplets spread along the surface, splash, and/or break into secondary particles, depending on impact energy. To advance nasal drug delivery modeling, three CFD models of fluticasone propionate deposition using the EDS device in a Draf III post-surgical geometry were developed: (1) using traditional “trap” boundary conditions; (2) using “wall-film” boundary conditions and one-way coupling; and (3) using “wall-film” boundary conditions with two-way coupling (to account for the high-mass loading in the near-nozzle spray field). Contour plots and data sets for each CFD model were qualitatively compared with physical models in the same geometry (visualized in a 3D-printed sinonasal cast). The CFD simulations showed that EDS delivers fluticasone to all sinonasal regions, including areas superiorly and posteriorly in the sinonasal cavity. In addition, deposition improved and more closely correlated with physical experiments as the CFD model complexity evolved. “Wall-film” modeling of EDS delivery demonstrated increased surface coverage compared to the “trap” model. The “wall-film/two-way coupling” model demonstrated substantially higher deposition in the remote frontal and maxillary sinuses, agreeing with physical cast experiments. In conclusion, CFD simulation of EDS delivery using the “wall-film” boundary/two-way coupling model appears to most accurately represent observations from physical model experiments in the same geometry.