Abstract
PURPOSE To investigate whether the riboflavin dosing frequency affects corneal cross-linking efficacy or safety, given that isotonic riboflavin solution is viscous and each installation coats the corneal surface with a film that absorbs some of the incident ultraviolet A light. DESIGN Prospective, randomized, single-center equivalence trial. PARTICIPANTS Patients with progressive keratoconus or ectasia after refractive surgery (n = 510). METHODS One eye per patient was prospectively randomized to 2-minute or 5-minute riboflavin dosing intervals with standard corneal cross-linking (epithelial removal and 30-minute irradiation with 3 mW/cm2 ultraviolet A light). Block randomization resulted in comparable representation of keratoconus and ectasia after refractive surgery in the 2 treatment arms. Treatment equivalence was assessed using the 2 one-sided test. Fellow eyes (n = 207) were treated with 5-minute dosing and considered in the safety analysis. MAIN OUTCOME MEASURES The primary hypothesis was equivalent change in the topography-derived maximum keratometry value from baseline to 6 months with 2-minute vs. 5-minute dosing. A ±0.75-diopter margin of equivalence for the treatment difference between dosing regimens was considered clinically relevant. Adverse events and changes from baseline to 6 months in corrected distance visual acuity (CDVA), uncorrected distance visual acuity, and minimum corneal thickness were assessed. RESULTS The mean reduction in maximum keratometry from baseline was equivalent with 2-minute and 5-minute riboflavin dosing intervals at 6 months (0.97 and 0.76 diopters, respectively; 90% confidence interval for treatment difference, -0.23 to 0.66; per-protocol population). With both dosing intervals, the mean improvement in CDVA was 0.07 logarithm of the minimum angle of resolution or 3.5 letters at 6 months. Of the 635 study and fellow eyes examined at 6 months, 134 (21%) gained and 32 (5%) lost 2 or more lines of CDVA. Three eyes (0.4%) developed sterile infiltrates, 1 (0.1%) had delayed epithelial healing with dendrites, and 3 (0.4%) had recurrent epithelial defects. Three eyes (0.4%) were re-treated. CONCLUSIONS The 2 riboflavin dosing regimens produced equivalent reduction in the maximum keratometry value, with a favorable safety profile.
