Changes of corneal topography indices after CXL in progressive keratoconus assessed by Scheimpflug camera.

PURPOSE: To evaluate the accuracy of Scheimpflug camera topography indices in detecting the therapeutic effect of corneal collagen cross-linking (CXL) on progressive keratoconus in the long term. METHODS: Fifty eyes of 25 patients with keratoconus were enrolled. CXL was performed in 25 eyes with progressive keratoconus (CXL group) and 25 fellow eyes with nonprogressive keratoconus served as controls. Thinnest corneal thickness, anterior keratometry (flat, steep), and keratoconus indices were measured with Scheimpflug camera before and 12 to 25 months after CXL. Regression analysis was used to evaluate the influence of corneal thickness and follow-up time on flattening effect of CXL. RESULTS: At baseline, steep keratometric values were significantly higher and thinnest corneal thickness values were lower in the CXL group (P = .027, .034), parallel with increased values of keratoconus indices: index of surface variance (P = .013), index of vertical asymmetry (P = .038), keratoconus index (P = .019), center keratoconus index (P = .039), index of height asymmetry (P = .037), index of height decentration (P = .0016), and radius minimum (P = .008). After adjustment for thinnest corneal thickness and follow-up time, CXL showed significant flattening effect expressed by changes in radius minimum (P < .001), index of surface variance (P = .03), keratoconus index (P = .006), center keratoconus index (P = .03), and index of height asymmetry (P = .026). Thinnest corneal thickness had significant influence on changes of index of surface variance (P = .049), index of vertical asymmetry (P = .01), and center keratoconus index (P = .03). Follow-up time showed no significant influence in any models (P > .05). CONCLUSIONS: Topographic indices indicate corneal flattening after CXL in the long term. Monitoring keratoconus index and index of height asymmetry should be the preferred choice in daily clinical practice because changes in values of these indices are independent from initial corneal thickness.