Transepithelial corneal collagen cross‐linking using iontophoresis versus the Dresden protocol in progressive keratoconus: A meta‐analysis

Abstract Background Standard epithelium‐off collagen cross‐linking using Dresden protocol (S‐CXL) is the standard of care for progressive keratoconus. Despite its efficacy, epithelial debridement is associated with pain, delayed visual rehabilitation, stromal oedema and haze. Minimising these complications while achieving a comparable efficacy remains an unmet need. Methods Comparative studies between transepithelial iontophoresis‐assisted CXL (I‐CXL) and S‐CXL reporting the outcomes of visual, refractive, topographic, aberrometry, demarcation line, endothelial cell density, confocal microscopy or complications were identified from databases. Assessments of publication bias, meta‐analyses, sensitivity analysis, subgroup analysis, and meta‐regressions were performed. Results In this meta‐analysis, 586 eyes from three randomised controlled trials and seven comparative studies were analysed. No differences were found in the change in uncorrected/corrected distance visual acuities, mean/maximum keratometry, central corneal thickness, higher order aberration, spherical aberration, coma, subbasal nerve/anterior stromal keratocyte density and demarcation line depth in both CXL protocols ( P ≥ .052). However, I‐CXL resulted in less thinning at the minimum pachymetry (standardised mean difference 0.25; 95% confidence interval [CI] 0.06‐0.44). More importantly, there was a significant reduction in complications following I‐CXL (odds ratio 0.30; 95% CI 0.12‐0.75). Meta‐regression analyses on demarcation line depth and complication suggested that I‐CXL was more effective than S‐CXL when baseline maximum keratometry was >55.2 D and the risk of complication was independent of other baseline covariates. Conclusion I‐CXL has a more favourable safety profile, as evidenced by the available literature, with less thinning at the minimum pachymetry and reduced risk of complications while achieving comparable effects on visual, refractive, topographic, aberrometry, and morphological outcomes as S‐CXL.