Reply : Comparison of different methods to calculate the axial length measured by optical biometry.

Abstract

De Bernardo et al. compare the results of their study investigating the axial length (AL) in phakic and pseudophakic eyes with those of our study investigating AL calculation with the sum of segments and with the group refractive index. The adoption of the group refractive index to convert the optical path length into a geometrical distance is certainly responsible for the discrepancies observed in both articles. However, some differences exist. In pseudophakic eyes, optical biometers modify AL calculation according to the corrections recommended by Haigis.1 These corrections are applied throughout the whole AL range, and for this reason, De Bernardo et al. found a similar AL shortening in long, medium, and short eyes. Other authors have also reported almost identical outcomes using swept-source optical coherence tomography optical biometry.2 On the other hand, the differences between the Argos and IOLMaster 700 are not due to correction factors but due to the higher percentage of vitreous in long eyes and different refractive indices used for each segment. In fact, the two instruments provide the same mean AL in eyes with medium AL, whereas the AL measured with individual refractive indices leads to higher values in short eyes and lower values in longer eyes.3,4 As regards the choice of the SRK/T formula to assess the influence of different AL values on intraocular lens power calculation, we thank De Bernardo et al. for their suggestion. As they state, the SRK/T formula adjusts the AL of eyes longer than 24.2 mm, and using the Holladay 1 formula might have been more appropriate. Therefore, we repeated our analysis with this formula, using a surgeon factor of 1.84, and found the following mean intraocular lens powers: 10.24 ± 4.77 diopters (D) (Argos), 9.98 ± 4.86 D (IOLMaster 700), 11.10 ± 4.21 D (Wang-Koch adjustment), 10.71 ± 4.30 D (Holladay 1 polynomial equation), 10.68 ± 4.57 D (Holladay 2 polynomial equation), and 10.30 ± 4.71 D (CMAL). The Friedman test showed a statistically significant (P < .0001) difference, equal to the one achieved for the SRK/T formula; the Dunn posttest revealed that no statistical differences existed between Argos measurements and CMAL values as well as between the AL values obtained with the two Holladay polynomial equations. Again, these findings mirror those achieved with the SRK/T formula. Finally, we agree that using only one eye for each patient is correct, and this is the reason why our primary analysis was performed on the longer eye. We subsequently decided to also perform a binocular analysis only after reading the paper by Sabatino et al., who included 218 eyes of 112 patients.5