Evaluating the forces involved in bubble management in DMEK surgery: mathematical and computational model with clinical implications.

ABSTRACT

PURPOSE: To model postoperative forces involved in Descemet membrane endothelial keratoplasty (DMEK) tissue adherence and bubble management, including the impact of surface tension on graft support, with a view towards clinical applications. SETTING: Tennent Institute of Ophthalmology, Glasgow, and James Watt School of Engineering, University of Glasgow, Glasgow, United Kingdom. DESIGN: Mathematical modelling and computer simulation. METHODS: Theoretical modelling of biphasic flow and interaction of gas, liquid and tissue within the anterior chamber for static horizontal scenario A (adherent DMEK with mobile bubble) and dynamic vertical scenario B (release of bubble due to pupil block following DMEK). RESULTS: The model assumed incompressibility for both fluids within realistically achievable pressure ranges. Cahn-Hilliard Navier-Stokes equations were discretised through the application of the Finite Element Method. Mathematical modelling and computer simulation showed bubble size, corneal curvature and force intensity influences surface tension support for DMEK tissue in scenario A. Scenario B demonstrated complex, uneven distribution of surface pressure on the DMEK graft during uncontrolled bubble release. Uneven pressure concentration can cause local tissue warping, with air/fluid displacement via capillary waves generated on the fluid-air interface adversely impacting DMEK support. CONCLUSIONS: We have quantitatively and qualitatively modelled the forces involved in DMEK adherence in normal circumstances. We have shown releasing air/gas can abruptly reduce DMEK tissue support via generation of large pressure gradients at the liquid/bubble/graft interfaces, creating negative local forces. Surgeons should consider these principles to reduce DMEK graft dislocation rates via optimised bubble size to graft size, longer acting bubble support and avoiding rapid decompression where possible.