Comparison of videokeratoscope and autokeratometer measurements on ellipsoid surfaces and human corneas.

BACKGROUND Various studies have compared the accuracy and repeatability of autokeratometers and videokeratoscopes using calibrated convex surfaces. We investigate the agreement between the Topcon KR-3500 autokeratometer and the EyeSys videokeratoscope on human corneas and calibrated convex surfaces. METHODS Measurements were obtained from 30 convex ellipsoidal buttons and 20 right eyes of 20 young normal human subjects. Vertex radius and p-values were compared for the two instruments. RESULTS The two instruments showed excellent agreement on convex buttons. The human data showed no such relationship. For vertex radius, a good level of agreement was obtained only for surfaces whose p-values were near unity. Repeatability was also calculated and was shown to be better with the Topcon autokeratometer than with the EyeSys videokeratoscope. Editing the EyeSys data to encompass the same corneal area as that of the Topcon improved its repeatability, although it did not reach the level of the Topcon autokeratometer. CONCLUSIONS The Topcon autokeratometer and the EyeSys videokeratoscope showed reasonable agreement for surface topography on convex conicoidal plastic test buttons but not for human corneas. Alterations in the data-capture mechanisms of videokeratoscopes could improve their ability to accurately image paraboloidal surfaces.