Keratometry with five different techniques: a study of device repeatability and inter-device agreement

The purpose of this study was to determine and compare the repeatability of keratometry measurements with the Javal manual keratometer, Topcon automated kerato-refractometer, IOLMaster, EyeSys Corneal Analysis System, and Pentacam Scheimpflug Topography System, and determine the agreement of measurements of the latter four devices with the Javal keratometer as the gold standard. In this crosssectional study, 21 people with no history of ocular disease or surgery were examined twice with each device. Minimum, maximum, and average keratometry readings were recorded in diopters (D) for each eye and used in the analyses. For statistical analysis, we determined correlation coefficients and used the Bland-Altman method and calculated the 95 % limits of agreement (LoA). All repeatability coefficients were satisfactorily high. Best repeatability for minimum keratometry and maximum keratometry readings was seen with IOLMaster (95 % LoA -0.23 to 0.19 D and-0.31 to 0.32 D, respectively). Agreement with Javal manual keratometry was best with Topcon and IOLMaster for minimum keratometry readings (95 % LoA -0.67 to 0.28 D and -0.57 to 0.38 D, respectively), and with Topcon for maximum keratometry readings (95 % LoA -0.85 to 0.63 D). In our series of normal eyes, keratometry readings with Topcon, IOLMaster, and Pentacam showed very good agreement with Javal measurements, and inter-device agreements were better than the repeatability of the manual gold standard. However, differences between Javal and EyeSys may not be clinically acceptable, and these two devices should not be used interchangeably. Further studies are needed to investigate these issues in non-virgin eyes.