Purchase this article with an account.
Carlos Rocha de lossada, Matthias Brunner, Dr. Davide Borroni, Esmaeil Arbabi, Stephen B Kaye, Vito Romano, Gabriela Czanner; Improving precision for determining change in patients with keratoconus. Invest. Ophthalmol. Vis. Sci. 2019;60(9):340.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To investigate a method for precision analysis to discriminate corneal change from measurement imprecision in patients with keratoconus (KC).
Patients with KC and healthy control were included in a prospective study. Posterior corneal tomographic parameters using a Pentacam HR were analysed. Patients had 3 measurements undertaken on 2 occasions. Coefficients of repeatability and limits of agreement (LOA) were compared using single and multiple measurements for each observer. Correlation of repeated measurements was evaluated using a linear mixed effect model. A formula was derived for the theoretical expected change in precision and compared with measured change
Sixty eyes of sixty patients were included: 30 healthy subjects (mean age of 31.6 ± 10 years) and 30 patients with KC (mean age of 26.7 ± 8 years). High quality scans were achieved in all healthy eyes (180 scans) and in 91.7% of the keratoconic eyes (165 scans). Posterior corneal curvature values were significantly higher in the KC group (p<0.01).Intra- and inter-observer 95% and 99.73% limits of agreement (LOA) for K1 flat [D], K2 steep [D], K mean [D], and K2-K1 [D], in healthy and keratoconic eyes are reported in the table. The 99.73% LOA using the mean of 3 measurements were narrower than the 95% LOA using one measurement
The 99.73% LOA (3 SD) using the mean of three measurements, provides better precision than 95% LOA (2 SD) using a single measure. A change in the back surface of the cornea of more than 0.21D K1 or 0.17D K2 using the mean of 3 measurements can be used as a cut-off for change in the back surface of the cornea in patients with KC
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.
This PDF is available to Subscribers Only