Limbal and conjunctival epithelium after corneal cross-linking using riboflavin and UVA.

PURPOSE: To assess possible cellular damage in the corneal and conjunctival epithelium after corneal cross-linking treatment.

METHODS

Riboflavin-dextran solution was applied every 5 minutes to the right eyes of 3 rabbits 10 minutes before and 30 minutes during the irradiation with UVA light of 370 nm and an irradiance of 3 mW/cm at the 8- to 10-o'clock position, including conjunctival, limbal, and central corneal epithelium. In addition, 3 rabbits were treated with UVA light only. The rabbits were killed 24 hours later. The treated eyes were examined histologically using hematoxylin-eosin and periodic acid-Schiff staining, immunohistochemistry, and a terminal deoxynucleotidyl transferase (TdT)-mediated deoxyuridine triphosphate-biotin nick-end labeling (TUNEL) apoptosis assay and compared with the left fellow eyes, which served as controls.

RESULTS

There was no epithelial defect on fluorescein staining. No apoptosis was found in the corneal limbal epithelium, keratocytes, or endothelial cells in the irradiated area. In the adjacent conjunctival epithelium, rare superficial conjunctival epithelial cells were positive in TUNEL staining in all animals. Anti-multicytokeratin-positive staining was demonstrated in both the limbal corneal and the conjunctival epithelium. The proliferation marker Ki-67 was identified in the basal cell layer of the limbal epithelium. The frequency and distribution pattern of goblet cells, multicytokeratin, and the proliferation marker Ki-67 were the same in all eyes compared with the untreated fellow control eyes.

CONCLUSIONS

Standard corneal cross-linking does not induce significant cellular epithelial damage as assessed by histological methods. Further studies on possible genotoxic and long-term changes would be helpful to complete the risk assessment.