Accurate biometry in silicone oil-filled eyes

Abstract

Sir, Kunavisarut et al1 should be congratulated on their study ‘Accuracy and reliability of IOL master and A-scan immersion biometry in silicone oil-filled eyes'. This is an important work, given the difficulties that can be encountered in IOL power prediction in these cases. As stated by the authors, it is not always possible to obtain accurate axial length measurements in these eyes preoperatively and the use of data from the fellow eye can be inaccurate. This paper compared the preoperative axial length measurements for IOL master and immersion ultrasound in oil-filled eyes with the measurement from the IOL master taken after removal of the oil. Accurate axial length measurement is a critical part of biometry in cataract surgery, as an error in this measurement is multiplied by a factor of around 2.5 in the process of intraocular lens power calculation.2 In this study, the authors used correlation coefficients to compare the axial length measurements obtained pre and post oil removal. Although the IOL master axial length measurements for silicone oil-filled eyes showed a high correlation with the postoperative axial length measurement (0.966), the postoperative refractive outcomes varied between −2.74 and +2.33 for this relatively small sample of eyes. As stated by Bland and Altman,3 the use of correlation coefficients to compare two different measurements of the same thing is problematic since even a very high correlation does not necessarily translate into useful clinical agreement. The method described by Bland and Altman allows 95% limits of agreement to be calculated. To achieve this, the differences between the measured values are plotted against their means. This method may have better illustrated why there were clinically significant errors in the refractive outcomes for the IOL master, despite a high correlation between the axial length measurements and no statistically significant difference between the mean measurements taken pre and post oil removal.