Abstract
Background: Keratoconus affects the anterior segment of the eye, which directly affects the refractive state of the eye. There are three components for the measurement of a corneal curvature or central corneal power (CCP) of the eye, namely, the power along the flat meridian, the power along the steep meridian and the axis of the flat meridian. Traditionally, CCP is analysed using univariate methods that processes each component separately; however, because of the trivariate nature of CCP, the use of multivariate methods and statistics may be beneficial. Aim: The aim of this study was to compare the short-term variation of the anterior and posterior CCP in eyes with and without keratoconus using multivariate methods of analysis. Setting: Data were extracted from a doctoral study by the first author. The group with keratoconus (KC) was obtained from patients attending a university-based contact lens clinic in Johannesburg, South Africa. Methods: A total of 28 eyes with KC and 28 eyes of 28 healthy control eyes without KC were included in this prospective quantitative study. Measurements were taken with the Oculus Pentacam (Wetzlar, Germany) and data related to the anterior and posterior CCP were analysed using multivariate methods and analysis. Results: For both KC and control groups, short-term variation of CCP of the anterior corneal surfaces was significantly greater than that for the posterior corneal surfaces. Whilst shortterm variation was similar for both corneal surfaces in the KC group, variation of the posterior corneal surfaces was significantly different from that of the anterior corneal surfaces for the control group. Conclusion: Multivariate analysis of short-term variation of CCP of both surfaces of the cornea in eyes with or without KC contributed towards a more complete understanding of the disease.