To determine the reliability and repeatability of keratometry (K) measurements obtained with the Pentacam high resolution (HR), automated keratometry, and corneal topography systems.
The right eyes of 46 healthy subjects were examined prospectively. Keratometry measurements in the flat (Kf) and steep (Ks) meridians were taken by 2 independent investigators with the Pentacam HR (Oculus, Wetzlar, Germany) followed by automated keratorefractometry (KR-8100; Topcon, Tokyo, Japan), and corneal topography (TMS-4; Tomey, Erlangen, Germany).
The mean K readings of the Pentacam HR, automated keratometry, and corneal topography were 43.40/43.34 diopter (D), 43.99/43.98 D, and 43.80/43.83 D, respectively. The difference between the values was statistically significant (P < 0.0001, repeated measures analysis of variance). Strong significant correlation was observed between the Pentacam HR and keratometry (Kf: r = 0.952/0.954; Ks: r = 0.845, Spearman rank test), and Pentacam HR and corneal topography (Kf: r = 0.933/0.930; Ks: r = 0.838/0.829) (P < 0.0001). No significant difference was presented between the 2 investigators for any of the instruments (P = 0.215–0.983). Moreover, high correlation was found between the K readings of the observers (interoperator intraclass correlation coefficients ranged from 0.95 to 0.99).
The Pentacam HR provided reliable K measurements in clinical practice in comparison with an automated keratometer and a corneal topographer. Based on the results, for patient follow-up, one keratometry device is recommended.
From the Medical and Health Science Center, Department of Ophthalmology, University of Debrecen, Debrecen, Hungary.
Received for publication December 11, 2009; revision received September 17, 2010; accepted November 2, 2010.
The authors have no proprietary interests or conflict of interest in this study.
Reprints: László Módis, Jr, Medical and Health Science Center, Department of Ophthalmology, University of Debrecen, H-4012 Debrecen, Nagyerdei krt. 98, Hungary (e-mail: firstname.lastname@example.org).