Outcome indicators for cross linking in pediatric keratoconus

Abstract

Purpose: To evaluate the predictive factors for successful corneal collagen cross-linking (CXL) in pediatric patients with Keratoconus (KC). Methods: This retrospective study was conducted using a prospectively built database. Patients (18 years old or younger) underwent CXL for KC between 2007 and 2017, with a 1-year follow-up period or longer. The outcomes included changes in Kmax (delta [Δ] Kmax = Kmaxlast − Kmaxpre) and LogMAR visual acuity (ΔLogMAR = LogMARlast − LogMARpre). The effects of CXL type (accelerated or non-accelerated), demographics (age, sex, background of ocular allergy, ethnicity), preoperative LogMAR visual acuity, maximal corneal power (Kmax), pachymetry (CCTpre), refractive cylinder, and follow-up (FU) time on the outcomes were analyzed. Results: One hundred thirty-one eyes of 110 children were included (mean age, 16 ± 2 years; range, 10–18 years). Kmax and LogMAR improved from baseline to last visit: from 53.81 D ± 6.39 D to 52.31 D ± 6.06 D (p < 0.001) and from 0.27 ± 0.23 LogMAR units to 0.23 ± 0.19 LogMAR units (p = 0.005), respectively. A negative ΔKmax (meaning corneal flattening) was associated with a long FU, low CCTpre, high Kmaxpre, high LogMARpre, and non-accelerated CXL on univariate analysis. High Kmaxpre and non-accelerated CXL were associated with negative ΔKmax in the multivariate analysis. A negative ΔLogMAR (meaning vision improvement) was associated with a high LogMARpre in univariate analysis. Conclusion: CXL is an effective treatment option in pediatric patients with KC. Our results showed that the non-accelerated treatment was more effective than the accelerated treatment. Corneas with advanced disease had a greater effect on CXL.