Reproducibility of normal corneal power measurements with a keratometer, photokeratoscope, and video imaging system

To determine the reproducibility of currently available instruments for measuring corneal power, two investigators measured comparable locations on 18 normal human corneas using a keratometer (Bausch & Lomb, Rochester, NY), the Corneascope (Kera Corp, Santa Clara, Calif), and the Corneal Modeling System (Computed Anatomy Inc, New York, NY). (For the two keratoscopes, average powers around rings at comparable locations were used.) Comparisons made between instruments at the significance level of .05 indicated that the keratometer was more reproducible than the Corneal Modeling System and the Corneascope, and the Corneal Modeling System was more reproducible than the Corneascope. Of the 31 rings projected on the surface of each cornea by the Corneal Modeling System, rings 2 through 13 were read reasonably reproducibly in that 76% of the measurements on these rings differed by no more than 0.5 diopter. The Corneal Modeling System (software version 1.16) is 83% as reproducible as a keratometer reading at approximately the edge of the 3-mm central zone on normal human corneas and provides information about corneal topography in a more reproducible and visually useful manner than the other two instruments.