A Systematic Review and Meta-Analysis Comparing Epithelium-On and Epithelium-Off Corneal Collagen Cross-Linking in Keratoconus Management: Integrating New Evidence

Purpose: Keratoconus is a progressive corneal ectasia leading to visual impairment. Corneal collagen cross-linking (CXL) is a widely used intervention to halt progression, with 2 main approaches: epithelium-on (Epi-On) and epithelium-off (Epi-Off). This study presents a meta-analysis integrating findings from existing literature and newly analyzed data. Methods: A systematic search was performed to compare Epi-On and Epi-Off CXL outcomes from 2014 to 2024. Meta-analysis techniques were employed to evaluate the changes in visual and imaging outcomes based on the differences between baseline to available 6-month, 1-year, 2-year, and 3-year follow-up. In addition, subgroup analyses were conducted based on study design (randomized controlled trials vs. non-RCT and observational studies) and risk of bias (high vs. low), using standardized risk assessment tools. Results: This meta-analysis included 28 studies, covering 1764 eyes including 875 (49.6%) Epi-Off and 889 (50.4%) Epi-On treated eyes. Maximum keratometry flattening was observed at 6 months (ES = 0.183, P = 0.039), became more pronounced at 1 year (ES = 0.639, P = 0.004), and remained significant at 2 years (ES = 0.377, P <0.001). Central corneal thickness significantly reduced after Epi-Off treatment at 6 months (ES = 0.327, P = 0.012) and 1 year (ES = 0.247, P = 0.040) more than Epi-on group. In addition, spherical equivalent improved significantly at 2 years (ES = −0.395, P = 0.048). In contrast, no significant differences were observed in steepest keratometry, average keratometry, best-corrected distance visual acuity, and uncorrected distance visual acuity between the 2 groups across all follow-up durations. Subgroup analysis supported the robustness of structural outcomes in Epi-Off, particularly in lower risk, observational studies and nonrandomized clinical trials, whereas functional outcomes remained not significant. Conclusions: Epi-Off CXL demonstrates superior topographic and tomographic improvements compared with Epi-On, although visual acuity outcomes remain equivalent. These findings support a personalized approach in selecting CXL technique based on patient profile and treatment goals.