Effectiveness and safety of cross-linking in keratoconus patients with corneal thickness <400μm: A systematic review and meta-analysis.

Abstract

TOPIC In this systematic review and meta-analysis, we aimed to investigate the effectiveness and safety of corneal cross-linking (CXL) in patients with keratoconus and thin corneas (corneal thickness <400μm). CLINICAL RELEVANCE Guidelines for the conventional (Dresden) CXL protocol to treat keratoconus recommend patients have a minimum stromal thickness of ≥400μm to avoid endothelial damage. Newer CXL methods have been developed for thinner corneas. Evaluation of clinical outcomes is required to inform surgical decisions and maximize the number of people who can safely benefit from CXL. METHODS 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 <400μm (with or without epithelium) who underwent CXL for keratoconus were eligible, regardless of CXL approach. Key outcomes included maximum and mean corneal curvature (Kmax and Kmean) 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. RESULTS Twenty-nine 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 Kmax: -1.4D, 95% CI [-2.0, -0.9], n=17/402 studies/eyes, I2 =93%) and improvement in corrected distance visual acuity (mean change -0.09 log MAR 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. CONCLUSIONS Visual and keratometric parameters improved following CXL without major complications, suggesting CXL may be suitable for patients with stromal thicknesses of <400μm, providing adequate modifications have been applied. Prospective comparative studies are required to guide the choice of approach.