Corneal thickness measurements with the Orbscan Topography System and ultrasonic pachymetry.

PURPOSE

To compare corneal thickness measurements obtained with a new instrument, the Orbscan Topography System, with those obtained with the DGH ultrasonic pachymeter and to assess the agreement and repeatability of the two devices.

SETTING

LSU Eye Center, New Orleans, Louisiana, USA.

METHODS

Measurement agreement was assessed in 51 eyes of 26 normal volunteers using both Orbscan and ultrasonic pachymetry. Repeatability for the instruments was measured in 10 eyes of 5 additional volunteers. Corneal thicknesses were compared using the analysis of variance (ANOVA). The relationship between the devices was assessed by analysis of regression (ANOR).

RESULTS

In the measurement agreement experiment, the mean corneal thickness was 571.3 microm +/- 6.21 SEM with the Orbscan system and 543.3 +/- 7.49 microm with ultrasonic pachymetry; these values were significantly different (F test, ANOVA, P = .0048). In the repeatability experiment, the mean thickness was 561.1 +/- 8.42 microm with the Orbscan system and 537.4 +/- 5.84 microm with ultrasound pachymetry; these values were also significantly different (F test, ANOVA, P = .0003). Analysis of regression showed a significant linear regression between the values obtained with the devices (P = .0001, F test, ANOR).

CONCLUSIONS

In both studies, the Orbscan system obtained statistically significantly different and higher values for corneal thickness. Regression analysis suggests that over the range of values in this study, the two devices differ by a constant amount (intercept and slope). The nonzero intercept of this regression shows that the values from the devices differ and cannot be directly substituted for each other. We therefore conclude that in this study, Orbscan system measurements of corneal thickness were 23 to 28 microm greater than ultrasonic pachymeter measurements. Linear regression equations may be developed for the results of measurements from the two devices and used as a precise transformation factor for the values obtained with the two devices.