Comparison of repeat videokeratography: repeatability and accuracy.

PURPOSE

We compared the repeatability and accuracy of four commercially available videokeratography instruments and a manual keratometer.

METHODS

Ten optometry students and two university employees who had no history of rigid contact lens wear and no soft contact lens wear within 6 months of the study were measured. Two independent measurements were taken on the right eyes only using the Alcon EyeMap EH-290 topography system (both manual and automatic focus), the EyeSys System 2000 corneal topography system, the Humphrey Mastervue topography system, the Humphrey Atlas topography system, and a Marco manual keratometer. A three-factor repeated measures analysis of variance with all factors repeated was performed using SAS and BMDP statistical software. Separate analyses were conducted for paracentral and peripheral data. To test accuracy, measurements were taken with the Alcon EyeMap automatic focus system, the EyeSys System 2000, the Humphrey Atlas system, and the Humphrey Mastervue system using four calibration spheres.

RESULTS

There were no statistically significant differences in the repeatability of the instruments. The EyeSys corneal topography system had the most narrow 95% limits of agreement around the mean difference. The two Humphrey instruments performed similarly to the EyeSys. Although the Alcon EyeMap EH-290 had the widest 95% limits of agreement around the mean difference, it showed better repeatability when the manual focus was used compared with the automatic focus. The instruments varied in their level of accuracy, but the highest percentage of data points within +/- 0.37 D of the known values of the calibration spheres was found with the Humphrey Atlas.

CONCLUSIONS

Although there were no statistically significant differences in repeatability, the EyeSys System 2000 had the highest repeatability, which may have clinical relevance. The accuracy measured in this study varied tremendously; however, the Humphrey Atlas was found to be the most accurate of the instruments tested.