Variability in Scheimpflug image-derived posterior elevation measurements in keratoconus and collagen-crosslinked corneas.

PURPOSE: To determine the variability in Scheimpflug image-derived posterior elevation measurements in eyes with keratoconus or corneal collagen crosslinking (CXL).

SETTING

Eye Institute of Thrace, Alexandroupolis, Greece.

DESIGN

Prospective observational study.

METHODS

Two operators in 2 sessions obtained posterior elevation measurements using the Pentacam Scheimpflug system. Variability in measurements was assessed using Bland-Altman analysis and intraclass correlation coefficients (ICCs). Measurements were associated with a series of Scheimpflug-derived keratoconus indices using the Pearson correlation and multivariate linear-regression analysis.

RESULTS

Mean interobserver differences for the central elevation point (e1) and the highest elevation point (HEP) in the keratoconus group were -0.9 μm (95% limits of agreement [LoA], 5.7 to -7.5 μm) and 1.7 μm (95% LoA, 26 to -22.6 μm), respectively. In the CXL group, the mean differences were 0.2 μm (95% LoA, 6.0 to -5.5 μm) and -6.1 μm (95% LoA, 16.8 to -29 μm), respectively. The ICC values in the keratoconus group were satisfactory for most elevation points; the ICC values were unsatisfactory for a series of points in the lower corneal hemisphere. The ICC values for points in both hemispheres were unsatisfactory in the CXL group; HEP had adequate reproducibility. The e1 and HEP parameters correlated significantly with most Scheimpflug-derived indices (P<.05) in the keratoconus group but not in the CXL group. Regression analysis indicated that the central keratoconus index was the primary determinant of the variability in keratoconus eyes.

CONCLUSIONS

Scheimpflug-derived posterior elevation measurements had acceptable reproducibility in keratoconus eyes and post-CXL eyes. Less consistent measurements were obtained in the lower hemisphere.