Aims To evaluate the safety and efficacy of corneal cross-linking (CXL) followed by photorefractive keratectomy (PRK) for refractive correction in patients with bilateral asymmetric topography.

Methods Forty-four patients (88 eyes) were enrolled in this prospective randomised clinical trial. CXL with subsequent PRK after 6 months was performed in one eye (study group), and PRK alone was performed in contralateral eyes (control group). Patients were followed for 24 months after PRK. Outcome measures investigated included visual acuity (VA), refraction, aberrometry, topography, pachymetry and endothelial cell count. Groups were compared with linear mixed regression and repeated measures logistic regression. Multiple comparison adjustment with the Holm procedure was performed.

Results At baseline, the logMAR VA (best spectacle corrected) in study and control groups was 0.12±0.13 (mean±SD) and 0.08D±0.14, respectively, and axial inferior–superior index (IS) (topographic IS) in study and control groups were 0.59D±0.31D and 0.58D±0.32D, respectively. After 24 months, a mean under correction of −0.50D was observed in both groups. Change from baseline in logMAR VA in study and control groups was 0.00D±0.08D and −0.02D±0.10D, respectively. Frequency of haze at 30 months in study and control group eyes was, respectively, 18.2% and 4.6% (p=0.05). There was no statistical difference between groups in spherical aberration and coma after adjustment for multiple comparisons.

Conclusions Non-simultaneous CXL followed by PRK may be performed safely, and refractive results over a 2-year follow-up are highly similar in virgin and previously cross-linked corneas. Despite using mitomycin C, corneal haze can be significantly higher in the first year after PRK in eyes pretreated with CXL.