Repeatability of central corneal thickness measurements measured with the Topcon SP2000P specular microscope

The non-contact specular microscope has become the method of choice for a quick, accurate and non-invasive assessment of central corneal thickness (CCT), which is an important variable to monitor before and after refractive surgery. The consistency of the results produced by such widely used methods/equipment must be assessed to determine their reliability. The purpose of this study was to assess within- and between-observer repeatability of, and to determine if a systematic bias exists in the measurements made by, the Topcon SP2000P specular microscope.

The CCT of the right eyes of 70 adult subjects, divided equally between men and women, was assessed on two separate occasions (4-7 days apart) by each of two examiners using the low-intensity auto mode of the SP2000P specular microscope.

The average CCT values for men and women, measured by one observer, were 0.52+/-0.03 mm (mean +/- SD) and 0.52+/-0.04 mm, respectively. The average for the entire sample was 0.52+/-0.04 mm. Within- and between-observer repeatability were assessed by plotting the mean difference (for each subject) between two readings made by the same observer or one each by both observers against the combined average CCT reading of both sessions; the mean difference between two sets of observations was not significantly different from zero (P<0.05). For the first observer, the 95% limits of repeatability were between -0.015 and 0.017 mm. For the second observer, the 95% limits of repeatability were between -0.018 and 0.018 mm. For the between-observer repeatability, the 95% limits of agreement were between -0.016 and 0.016 mm. For both within- and between-observer repeatability, we found no systematic bias of the mean difference with the average CCT reading.

The within- and between-observer limits of agreement we found were similar to those previously reported for the Topcon SP2000P specular microscope; the range of the 95% limits of repeatability were within +/-1 SD of the average CCT reading for both sessions. We suggest that a technique be considered reliable if: (1) the mean difference between two measurements does not vary significantly from zero, (2) there is no systematic bias of the mean difference with the magnitude of the measured quantity and (3) the error inherent in a measurement technique is within +/-1 SD of the average measurement of the two sessions.