We have read the letter submitted by Cione et al. and would like to clarify the issues raised by them. We will comment on each topic brought up
Testing distance: This is important for studies where constant optimization is required because comparable constant values can be obtained. In our study, where constant optimization was not performed, it is less relevant. A difference of 0.08 diopter (D) in this context is negligible and does not affect our findings.
Sample size: It is true that in our 2021 guidelines (that were actually written in 2019, when we received the invitation from the editor of Ophthalmology to prepare them), we stated that “it would be preferable … a minimum of 50 eyes that have undergone laser in situ keratomileusis or photorefractive keratectomy.”1 This was a suggestion, but the minimum sample of 32 eyes recommended by Wang et al. in their study is a valid recommendation.2 That study was published after we had already submitted our guidelines, and therefore, we were not able to include their calculated minimum sample size in the manuscript. However, when we prepared the new manuscript, we relied on the calculation performed by a respected and trusted group such as the one including Wang and Koch. Our publication had been delayed because of COVID.
Different intraocular lens (IOL) models: Since we investigated only the spherical equivalent refraction, there was no reason to exclude toric IOLs.
Constant optimization: We agree that constant optimization may be preferable in postrefractive surgery eyes. On the other hand, there is an ongoing debate whether constant optimization should be performed in specific subgroups, such as long eyes, and a consensus among experts has not yet been reached. Most authors believe that constants should be optimized for large samples including short, medium, and long eyes and that these constants should be applied to subgroups as we did in this study. Of course, this can lead to myopic or hyperopic prediction errors.
We thank Cione et al. for highlighting our major error. It is true that we accidently enrolled 9 bilateral cases and that the sample consists of 39 eyes of 30 patients. We therefore repeated the analysis on the subset of 30 unilateral eyes after excluding the second eyes. The results did not change because the method with the lowest mean PE was again by ray tracing with the Holladay 2 axial length (AL) adjustment (+0.025 ± 0.468 D), followed by ray tracing with the Holladay 1 AL adjustment (−0.145 ± 0.495 D). Using ray tracing with an unadjusted AL produced a mean hyperopic PE (+0.561 ± 0.540 D), whereas the original Wang-Koch AL adjustment produced a mean myopic PE (−0.458 ± 0.559 D), which were slightly higher than the Barrett True-K with predicted posterior corneal power (−0.378 ± 0.617 D) and that with measured posterior corneal power (−0.415 ± 0.642 D).
We also thank them for pointing out the lack of anonymity in the attached Excel data file. Again this was an inadvertent mistake, and it has already been corrected by the journal.
We believe our past studies have been proven completely trustworthy, and human errors do occur on occasion. We appreciate the opportunity to correct this one.
1. Hoffer KJ, Savini G. Update on intraocular lens power calculation study protocols: the better way to design and report clinical trials. Ophthalmology 2021;128:e115–e120
2. Wang L, Spektor T, De Souza R, Koch D. Evaluation of total keratometry and its accuracy for IOL power calculation in eyes following corneal refractive surgery. J Cataract Refract Surg 2019;45:1416–1421