Corneal Cross-Linking for Paediatric Keratoconus: A Systematic Review and Meta-Analysis

Abstract

All corneal cross-linking techniques attenuated disease progression in patients with pediatric keratoconus for at least one year based on a meta-analysis. A standard and accelerated technique led to marked improvement in visual acuity. We determined the efficacy and safety of corneal cross-linking (CXL) in pediatric keratoconus by conducting a systematic review and meta-analysis. The PubMed and Cochrane databases were searched for relevant studies on the effects of standard, transepithelial, and/or accelerated CXL protocols in patients aged 18 years or younger. Standardized mean differences with 95% confidence intervals were calculated to compare the data collected at baseline and 12 months. The primary outcomes were maximum keratometry (Kmax) and uncorrected visual acuity (UCVA), and the secondary outcomes were the thinnest corneal thickness (TCT), best-corrected visual acuity (BCVA), and manifest refraction spherical equivalent or cylindrical refraction. Our search yielded 7913 publications, of which 26 were included in our systematic review and 21 were included in the meta-analysis. Standard CXL significantly improved the Kmax, UCVA, and BCVA, and significantly decreased the TCT. Accelerated CXL significantly improved UCVA and BCVA. In the transepithelial and accelerated-transepithelial CXL methods, each measurable parameter did not change after treatments. All CXL techniques attenuated disease progression in patients with pediatric keratoconus for at least one year. Standard and accelerated CXL led to marked improvement in visual acuity.

Keywords: corneal cross-linking; keratoconus; pediatric.

Figure 1

Preferred reporting items for systematic reviews and meta-analyses flow chart outlining search process to identify relevant articles from abstract identification to full paper review and inclusion of relevant publications.

Figure 2

Forest plot of maximum keratometry (Kmax) 1-year standardized mean differences in dioptres in studies included in meta-analysis. (A), standard cross-linking (SCXL); (B), accelerated CXL (ACXL); (C), transepithelial CXL (TCXL); (D), accelerated and transepithelial CXL (ATCXL). IV = inverse variance, CI = confidence interval, Tau² = tau-square statistic, Chi² = chi-square statistic, df = degrees of freedom, I² = I-square heterogeneity statistic, z = Z-statistic.

Figure 3

Forest plot of thinnest corneal thickness 1-year standardized mean differences in micrometres in studies included in meta-analysis. (A), standard cross-linking (SCXL); (B), accelerated CXL (ACXL), (C), accelerated and transepithelial CXL (ATCXL). IV = inverse variance, CI = confidence interval, Tau² = tau-square statistic, Chi² = chi-square statistic, df = degrees of freedom, I² = I-square heterogeneity statistic, z = Z-statistic.

Figure 4

Forest plot of uncorrected visual acuity 1-year standardized mean differences in logMAR in studies included in meta-analysis. (A), standard cross-linking (SCXL); (B), accelerated CXL (ACXL); (C), transepithelial CXL (TCXL), (D), accelerated and transepithelial CXL (ATCXL). IV = inverse variance, CI = confidence interval, Tau² = tau-square statistic, Chi² = chi-square statistic, df = degrees of freedom, I² = I-square heterogeneity statistic, z = Z-statistic.

Figure 5

Forest plot of best-corrected visual acuity 1-year standardized mean differences in logMAR in studies included in meta-analysis. (A), standard cross-linking (SCXL); (B), accelerated CXL (ACXL); (C), transepithelial CXL (TCXL); (D), accelerated and transepithelial CXL (ATCXL). IV = inverse variance, CI = confidence interval, Tau² = tau-square statistic, Chi² = chi-square statistic, df = degrees of freedom, I² = I-square heterogeneity statistic, z = Z-statistic.

Figure 6

Forest plot of manifest refraction spherical equivalent 1-year standardized mean differences in dioptres in studies included in meta-analysis. (A), standard cross-linking (SCXL); (B), accelerated CXL (ACXL); (C), transepithelial CXL (TCXL), (D), accelerated and transepithelial CXL (ATCXL). IV = inverse variance, CI = confidence interval, Tau² = tau-square statistic, Chi² = chi-square statistic, df = degrees of freedom, I² = I-square heterogeneity statistic, z = Z-statistic.

Figure 7

Forest plot of cylindrical refraction 1-year standardized mean differences in dioptres in studies included in meta-analysis. (A), standard cross-linking (SCXL); (B), accelerated CXL (ACXL); (C), transepithelial CXL (TCXL), (D), accelerated and transepithelial CXL (ATCXL). IV = inverse variance, CI = confidence interval, Tau² = tau-square statistic, Chi² = chi-square statistic, df = degrees of freedom, I² = I-square heterogeneity statistic, z = Z-statistic.

Figure 8

Forest plot of corneal endothelial cell density 1-year standardized mean differences in cells/mm² in studies included in meta-analysis. (A), standard cross-linking (SCXL); (B), transepithelial CXL (TCXL), (C), accelerated and transepithelial CXL (ATCXL). IV = inverse variance, CI = confidence interval, Tau² = tau-square statistic, Chi² = chi-square statistic, df = degrees of freedom, I² = I-square heterogeneity statistic, Z = Z-statistic.Table 2. Summary of complications of CXL after a 1-year followup.