Comparison Between Laser Scanning In Vivo Confocal Microscopy and Noncontact Specular Microscopy in Assessing Corneal Endothelial Cell Density and Central Corneal Thickness

Abstract

To compare central corneal thickness (CCT) and endothelial cell density (ECD) with laser scanning confocal microscope Heidelberg Retina Tomograph (HRT) II Rostock Corneal Module and noncontact specular microscope Tomey EM-3000 and to assess intra- and interobserver agreement in normal corneas. This prospective study included 48 normals (69.6 ± 7.2 years, range: 55-80 years) who underwent CCT and ECD with both Tomey and HRT 3 times by 2 independent observers. Measurement differences between instruments, agreement between devices, and test-retest variability (TRV) were determined. Mean CCTs with Tomey and HRT were 529.4 ± 35.4 and 536 ± 37.6 μm (P = 0.06), respectively; average ECDs with Tomey and HRT were 2473.5 ± 242.2 and 2539.7 ± 338.6 cells per square millimeter (P = 0.04), respectively. The mean of the differences (HRT minus Tomey) was 6.5 ± 17 μm for CCT and 65 ± 135.1 cells per square millimeter for ECD. Differences between instruments were not related to CCT (P = 0.35), whereas significantly increased with increasing ECD (P = 0.0001). Intraexaminer TRV for Tomey and HRT were 3.9 ± 3.7 and 22.2 ± 18.4 μm for CCT and 73 ± 63.4 and 152.2 ± 148.4 cells per square millimeter for ECD, respectively; interexaminer TRV was 4.6 ± 4.2 and 23.8 ± 17.3 μm for CCT and 84.9 ± 72.3 and 159.8 ± 149.8 cells per square millimeter for ECD. HRT II Rostock Corneal Module and the Tomey EM-3000 showed an overall good intermethod agreement. HRT showed a tendency to slightly overestimate CCT measurements, significantly underestimate ECD measurements in eyes with a reduced cell density (< 2290 cells per square millimeter), and overestimate ECD in eyes with a high cell density. Both instruments showed low intra- and interobserver TRV for both CCT and ECD measurements, which tended to be less for Tomey.