The aim of this study was to evaluate the repeatability of central corneal thickness (CCT) measurements in normal eyes using 2 new noncontact specular microscopes (NCSMs) EM-3000 (Tomey, Japan) and SP-02 (CSO, Italy) and to compare the results with those obtained from an SP-3000P NCSM (Topcon, Japan) and ultrasound pachymetry (USP).
Seventy subjects were enrolled in a prospective study. A single experienced ophthalmologist performed tests with each of the 4 instruments. Measurements were obtained in the right eye during the same session. The testing sequence of the NCSM was randomly selected. After performing noncontact examinations, the USP was performed to derive the CCT measurements. Intraoperator repeatability was analyzed using within-subject coefficient of variation and intraclass correlation coefficients. The agreement between NCSMs or NCSM and USP was assessed with Bland–Altman plots and 95% limits of agreement (LoA).
The mean CCT values measured by SP-3000P, EM-3000, SP-02, and USP were 513.66 ± 33.14 μm, 529.12 ± 33.22 μm, 549.06 ± 40.27 μm, and 539.01 ± 35.73 μm, respectively. All coefficients of variation were <1.3%, and the intraclass correlation coefficients were >0.95. There were statistically significant differences between any 2 devices as determined by CCT measurements. The mean difference between paired comparisons was >9 μm. The 95% LoA ranges were broad, and the greatest 95% LoA was found to exist between SP-3000P and SP-02.
The new NCSMs and USP all show a high intraoperator repeatability for CCT measurements in normal eyes. However, interdevice agreement was poor and prevented the comparison of CCT measurements taken with different instruments.