Posterior corneal shape: Comparison of height data from 3 corneal topographers.

Purpose: To compare the ability of 3 clinical corneal topographers to describe the posterior corneal shape.

Setting: University Medical Center Groningen, the Netherlands.

Design: Prospective observational study.

Methods: Corneas of healthy participants were measured twice with a dual Scheimpflug instrument (Galilei G2), a scanning-slit system (Orbscan IIz), and a single Scheimpflug instrument (Pentacam HR). Height data describing the posterior corneal shape were fit with Zernike polynomials. Mean values with standard deviations (SD), test-retest variability (coefficient of repeatability [CoR]), and interdevice variability were determined for the defocus Z(2,0), astigmatism Z(2,-2) and Z(2,2), and higher-order terms coma Z(3,-1) and Z(3,1), trefoil Z(3,-3) and Z(3,3), and spherical aberration Z(4,0) coefficients for 5.5 mm and 8.0 mm diameters.

Results: For the 5.5 mm diameter, CoRs ranged from 0.3 to 4.3 μm with the dual Scheimpflug instrument, 1.6 to 5.2 μm with the scanning-slit system, and 0.3 to 2.0 μm with the single Scheimpflug instrument. The CoR was similar for the Scheimpflug instruments (P = .43) but poorer for the scanning-slit system (P < .001). The CoRs of the Scheimpflug instruments were smaller than the corresponding population SD for defocus, cardinal astigmatism, coma, and spherical aberration. The scanning-slit system failed to provide 8.0 mm diameter data. There was a significant bias (interdevice variability) between the Scheimpflug instruments in the higher-order coefficients at both diameters.

Conclusions: Repeatability in assessing the posterior corneal shape was generally good for the Scheimpflug instruments but poor for the scanning-slit system. Interdevice variability between the Scheimpflug instruments compromised the interchangeability of higher-order coefficients. For astigmatism, CoR and 95% limits of agreement of the Scheimpflug instruments typically corresponded to 0.1 diopter per astigmatism term.