A Multifactorial Treatment Analysis and Algorithm for Corneal Collagen Crosslinking

Abstract

Purpose: : To determine preoperative predictors of best spectacle corrected visual acuity (BSCVA) and maximum keratomtery (K_max) outcomes after corneal collagen crosslinking (CXL), and to determine a treatment algorithm to select those patients who will benefit most from CXL.

Methods: : CXL was performed on 104 eyes with either keratoconus (kc) or post LASIK ectasia (ec). A multiple regression and odds ratio analysis were performed to determine predictors of BSCVA and K_max outcomes. In the regression analysis, preoperative predictors included uncorrected and best corrected visual acuity, K_max, thinnest pachymetry, haze, disease, cone location, gender, and age. Improvement of BSCVA was defined as an improvement of ≥2 Snellen lines of BSCVA at 1 year postoperatively. Improvement of K_max was defined as an improvement of ≥2D of K_max at 1 year postoperatively. Worsening of BSCVA and K_max was defined as a loss of ≥1 Snellen lines of BSCVA, and ≥1D of K_max steepening at 1 year postoperatively.

Results: : In a multifactorial regression analysis, all preoperative measurements were found to be significantly predictive of 1 year postoperative BSCVA (r²=0.5, p<0.001), and 1 year postoperative K_max (r²=0.9, p<0.001). When analyzed individually, within the above multivariable regression analysis, preoperative BSCVA (coefficient=0.4, p<0.001) was a significant predictor of 1 year postoperative BSCVA. Patients with a preoperative BSCVA of 20/40 or worse were 5.9x (CI = 2.2x-16.3x) more likely to improve by 2 or more Snellen lines than patients with a BSCVA better than 20/40. Patients with a K_max ≥ 55D were 5.2x (CI = 2.0x-13.0x) more likely to experience an improvement of ≥2D of K_max, compared to those patients with a K_max< 55D at 1 year.

Conclusions: : Our multifactorial analysis suggests that all patients with progressive kc or ec will benefit from CXL treatment with regard to stabilization of the disease. In general, patients with worse BSCVA and higher keratometry, particularly when their preoperative BSCVA is 20/40 or worse and/or their preoperative K_max is 55D or worse, are most likely to experience improvement after CXL. With regard to patients with stable keratoconus, a treatment algorithm based on these results may suggest those patients who may obtain a significant clinical improvement.

Clinical Trial: : http://www.clinicaltrials.gov NCT00647699

Keywords: keratoconus • cornea: clinical science • refractive surgery