Finite element analysis demonstrates the importance of stromal oxygenation in corneal crosslinking

Abstract

Purpose : Corneal crosslinking (CXL) slows progression of keratoconus and decreases the need for corneal transplantation by reinforcing the stroma with covalent bonds. These bonds are formed by a photochemical reaction, in which molecular oxygen (O₂) is converted to reactive oxygen species (ROS) by UV light and riboflavin (Rf). Numerous protocols for performing CXL are in clinical use. These vary greatly in the amount of O₂, Rf, and UV they deliver to the cornea, and whether viscous Rf is applied to the cornea during CXL, generating a pre-corneal film. However, how these differences change the efficacy of CXL protocols is currently unknown.

Methods : To address this question, we created a 1D finite element model of epithelium-off CXL in COMSOL (v6.1) that measures a key indicator of CXL efficacy (ROS). The rate of ROS formation was derived from the underlying kinetics of the CXL reaction and the amounts of the three reactants (O₂, UV, and Rf). Table 1 lists the CXL parameters we evaluated. We used literature values for corneal dimensions and other constants, and we validated our model using measurements of corneal O₂ during CXL.

Results : We found that performing CXL with a pre-corneal Rf film generated significantly fewer ROS than CXL without a film (Figure 2A). Depletion of O₂ from corneal stroma was primarily responsible for this result (Figure 2B). 'Accelerated’ CXL protocols that used higher UV irradiance depleted O₂ more than protocols using less UV, leading to lower levels of ROS generation. Supplementation of O₂ during the CXL process offset the O₂-depleting effects of both the pre-corneal film and accelerated CXL protocols. Protocols that maximized stromal O₂ through lower UV irradiance, O₂ supplementation, and absence of a pre-corneal film generated the most ROS.

Conclusions : Our results indicate that availability of O₂ within the corneal stroma during CXL is a major determinant of ROS generation. Furthermore, they suggest that applying viscous riboflavin during UV exposure could decrease CXL efficacy, and that ‘accelerated’ CXL protocols could require O₂ supplementation to generate ROS effectively.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.

 

Table 1: CXL parameters evaluated in the FEM

 

Figure 1: (A) Heatmap of ROS generated over the entire CXL period and corneal thickness (i.e. integrated over time and corneal depth). (B) Heatmap of steady-state O₂ concentration during CXL integrated over the entire corneal depth.