Three-year clinical observation of the outcomes of transepithelial and epithelial-off accelerated corneal collagen crosslinking treatment for different types of progressive keratoconus.

In the present study, the clinical and long-term effects of accelerated transepithelial corneal collagen crosslinking (ATE-CXL) and accelerated epithelial-off corneal collagen crosslinking (A-CXL) for the treatment of different types of progressive keratoconus were compared. A total of 70 patients, including 96 eyes with advanced keratoconus, were enrolled in the study. ATE-CXL or A-CXL was performed on one or two eyes of each subject according to corneal thickness, keratoconus type and surgical approach. Patients were divided into the following four groups: Group A, ATE-CXL for central keratoconus; group B, A-CXL for central keratoconus; group C, ATE-CXL for peripheral keratoconus; and group D, A-CXL for peripheral keratoconus. Uncorrected distant visual acuity (UDVA), best-corrected distant (BD)VA and corneal astigmatism (CA) were evaluated in all patients by routine ophthalmology pre-operatively and 3 years post-operatively. Topographical features, including maximum corneal curvature (K), thinnest corneal thickness (TCT), anterior corneal elevation (ACE) and corneal endothelial cell density (ECD) were also compared across groups. The results suggested that pre- and post-operative UDVA, BDVA, K, CA and ACE values differed in all four groups (P<0.05), whereas no differences were observed between pre- and post-operative TCT and ECD (P>0.05). Concordant results were obtained between groups A and C and groups B and D. ATE-CXL achieved better control of central keratoconus UDVA, K and CA as compared with A-CXL. The difference between pre- and post-operative UDVA, K and CA as compared with A-CXL was highly correlated with the change in intraocular pressure and treatment effectiveness. There was a statistically significant improvement in BDVA with ATE-CXL for treatment of central keratoconus compared with that after A-CXL treatment (P=0.032). There were statistically significant improvements in BDVA (P=0.047), CA (P=0.045) and ACE (P=0.012) with A-CXL treatment of peripheral keratoconus when compared with ATE-CXL treatment. Central, and to a lesser extent, peripheral, keratoconus may be effectively controlled by either approach, with disease stabilization 3 years later. ATE-CXL is suggested to be the most suitable treatment for keratoconus of <400 µm with a corneal thickness of >400 µm; however, A-CXL yields superior long-term outcomes.