Predictive Factors For Topographic Response Following Corneal Collagen Cross-Linking In Keratoconus: A Prospective Regression Analysis

To identify preoperative predictive factors associated with favorable topographic response following standard corneal collagen cross-linking (CXL) in progressive keratoconus and to develop a predictive model for clinical decision-making.

Department of Ophthalmology, Recep Tayyip Erdogan University Training and Research Hospital, Rize, Turkey. Prospective analytical study.

Sixty eyes of 60 progressive keratoconus patients (aged 14-35 years) who underwent standard epithelium-off CXL were analyzed. Favorable response was defined as Kmax reduction ≥0.5 D at 6 months. Preoperative variables analyzed included age, sex, keratoconus stage  Kmax, Kmin, simK, central corneal thicknes, thinnest corneal thickness, corneal astigmatism, best-corrected visual acuity, endothelial cell density, ISV, IVA, and KI. Univariate analysis identified candidate predictors (p<0.20), which were entered into multivariate binary logistic regression. Receiver operating characteristic (ROC) curves determined optimal cutoff values. Hosmer-Lemeshow test assessed model fit. Statistical analysis was performed using SPSS 23.0 (p<0.05).

Favorable topographic response (Kmax reduction ≥0.5 D) was achieved in 43.3% (26/60) of eyes. In univariate analysis, higher baseline Kmax (p=0.003), younger age (p=0.018), lower CCT (p=0.042), higher ISV (p=0.006), and higher corneal astigmatism (p=0.024) were associated with favorable response. Multivariate logistic regression identified baseline Kmax (OR=1.34, 95% CI 1.12-1.61; p=0.001) and age (OR=0.88, 95% CI 0.79-0.97; p=0.014) as independent predictors. ROC analysis yielded optimal cutoffs of Kmax ≥54.0 D (AUC=0.78, sensitivity 73%, specificity 76%) and age ≤22 years (AUC=0.69, sensitivity 65%, specificity 68%). 

The combined model (Kmax + age) achieved AUC=0.84 (sensitivity 77%, specificity 82%). Keratoconus stage was not an independent predictor after adjusting for Kmax.

Higher baseline Kmax and younger age are independent predictors of favorable topographic response after CXL. A combined predictive model incorporating both parameters achieves good discriminative ability (AUC 0.84). These findings suggest that younger patients with more advanced keratometric values benefit most from CXL, supporting early intervention in progressive keratoconus. This model may assist clinicians in preoperative counseling regarding expected treatment outcomes.