Abstract
Purpose:
This study investigates the effect of UVA irradiation time on corneal collagen cross-linking (CXL) in an oxygen diffusion-limited approach. Recent reports suggest a dependence of the reaction on environmental oxygen while current models of the reaction do not account for this diffusion limit. We hypothesize that an oxygen dependent model of cross-linking can predict material property changes due to CXL. Here, we aim to develop such a model and validate it with experimental data.
Methods:
An empirical oxygen diffusion-limited reaction model of CXL was developed. The model was compared to tensile tests on porcine corneal strips before and after CXL to determine the elastic modulus increase at various strains. Fluorescence recovery after photobleaching (FRAP) on corneal stroma in whole intact porcine eyes was also performed to determine the resulting increase in lateral-direction diffusivity, a marker for the CXL effect. The duration of the UVA light exposure was varied to realize the effect of time-dependent oxygen diffusion on the resulting changes in mechanical and transport properties.
Results:
The model predicts a nonlinear relationship between the total increase in oxidized substrate and UVA irradiation time. The percent increase in elastic modulus at 8 percent strain calculated from tensile tests fit well with the predictions from the model, and this increase after 30 minutes was used to determine the assumed linear relationship between the increase in oxidized substrate and increase in modulus. The model describes cross-linking occurring in a layer by layer fashion, as oxygen diffuses from the surface and reacts with the substrate. Results of FRAP agree with this description and reveal that not only the overall intensity, but the depth of the CXL effect is greatly dependent on UVA irradiation time.
Conclusions:
CXL depends on oxygen diffusion from the corneal surface. We present a CXL model dependent on oxygen that predicts the regional changes in material properties and validates these predictions with experimental data. The diffusion limit suggests that accelerated, high fluence CXL protocols, performed with decreased UVA irradiation time but equivalent total energy, do not yield equivalent results. In addition, this study supports exploring high-oxygen techniques in pursuing accelerated or enhanced protocols.