Comparison of non-contact methods for the measurement of central corneal thickness

• Published in: Ophthalmic surgery, lasers & imaging Vol. 42; no. 5; p. 400
• Main Authors: Gorgun, Ebru, Yenerel, Nursal Melda, Dinc, Umut Asli, Oncel, Banu, Kucumen, Raciha Beril, Oral, Deniz, Ciftci, Ferda
• Format: Journal Article
• Language: English
• Published: United States 01-09-2011
• Subjects: Adult; Analysis of Variance; Cornea - anatomy & histology; Cornea - diagnostic imaging; Cornea - pathology; Corneal Topography - methods; Female; Humans; Male; Photography - methods; Reproducibility of Results; Tomography, Optical Coherence - methods; Ultrasonography; Young Adult
• ISSN: 1938-2375
• DOI: 10.3928/15428877-20110812-02

This study examined the repeatability of and agreements between central corneal thickness measurements obtained by four different non-contact pachymetry devices.
Seventy-eight eyes of 39 subjects were included. Central corneal thickness of each eye was measured by Visante optical coherence tomography (OCT) (Carl Zeiss Meditec Inc., Dublin, CA), Pentacam (Oculus Optikgeräte GmbH, Wetzlar, Germany), Orbscan IIz topography (Bausch & Lomb Surgical Inc., San Dimas, CA), and slit-lamp OCT (SL-OCT) (Heidelberg Engineering GmbH, Heidelberg, Germany). Inter-device agreements and correlations and repeatability of each device were examined.
All measurement methods correlated well with each other with a correlation coefficient greater than 0.90 and P value of less than .001 for all comparisons. However, Pentacam overestimated central corneal thickness: 546.7 ± 38.2, 535.5 ± 42.7, 531.7 ± 37.6, and 531.2 ± 36.0 μm for Pentacam, Orbscan IIz, Visante OCT, and SL-OCT, respectively (P < .001 for all comparisons versus Pentacam). Despite good correlation, magnitude of differences was high and this bias was proportional (ie, not constant across a range of corneal thickness values) for the following pairs: Orbscan versus Visante OCT, Orbscan versus SL-OCT, and Orbscan versus Pentacam (P < .001 for all comparisons).
Although measurements obtained by various non-contact methods correlate well, numerical agreement of the results may not be sufficient for their interchangeable use in clinical practice.