Effectiveness and Safety of Cross-Linking in Keratoconus Patients With Corneal Thickness <400 µm: A Systematic Review and Meta-analysis.

Abstract

To investigate the effectiveness and safety of corneal cross-linking (CXL) in patients with keratoconus and thin corneas (corneal thickness <400 µm). Systematic review and meta-analysis. MEDLINE Ovid, Embase Ovid, and Scopus were searched in July 2024 for studies published in peer-reviewed journals in English. Prospective and retrospective case series, cohort studies, and randomized clinical trials of people with corneal thickness of <400 µm (with or without epithelium) who underwent CXL for keratoconus were eligible, regardless of CXL approach. Key outcomes included maximum and mean corneal curvature (K_max and K_mean), corrected and uncorrected distance visual acuity, refractive error, corneal thickness, and endothelial cell density at 3, 6, 12, and 24 months. Random effects meta-analysis was used to pool average change from preoperative values across CXL protocols. Study quality was assessed using the JBI Critical Appraisal Checklist for Case Series (PROSPERO registration: CRD42023403190). A total of 29 records published between 2011 and 2024 were included with a total of 470 eyes. Most studies were retrospective case series with substantial potential for bias. Substantial heterogeneity was observed between studies in terms of patients, CXL approach, and observed outcomes. At 12 months, flattening of the maximum corneal curvature (mean change in K_max = -1.4 D, 95% CI = - 2.0, -0.9, n = 17/402 studies/eyes, I²= 93%) and improvement in corrected distance visual acuity (mean change = -0.09 logMAR units, 95% CI = - 0.13, -0.06], n = 19/443 studies/eyes, I²= 68%) were observed relative to baseline. No study reported major complications, and there were minimal reductions in corneal thickness and endothelial cell density. Visual and keratometric parameters improved following CXL without major complications, suggesting that CXL may be suitable for patients with stromal thicknesses of <400 µm, providing that adequate modifications have been applied. Prospective comparative studies are required to guide the choice of approach.