We read with great interest the article by Kenny et al. concerning intraocular lens (IOL) power calculation in short eyes.1 In this article, 15 IOL power calculation formulas were compared using traditional axial length (AL) measurements and the segmented AL calculated with the Cooke-modified AL nomogram (CMAL). The authors found that the use of segmented AL for calculating IOL power in short eyes did not improve the formula performances. We congratulate the authors for their interesting study, but we would like to make a few comments.

The authors correctly reported that IOL power calculation is challenging in short eyes, and they declared that although the Wang-Koch and Holladay adjustments have reduced prediction errors (PEs) for long eyes, no such solution exists in short eyes. This is not completely true. In fact, De Bernardo et al. recently reported that cataract can modify the lens refractive index, and this could alter AL obtained by biometers using the group refractive index.2 On these bases, they proposed a correction factor to improve the reliability of AL measurement, applicable to the entire range of AL. This corrected AL (ALc) can eliminate any systematic error resulting from the biometer.2 Because the authors used LENSTAR LS900, based on a group refractive index combined with the Grieshaber correction algorithm to measure AL, they should have taken in consideration ALc. From this point of view, ALc could be considered an alternative to CMAL or an additional correction in case of formulas that already correct AL to CMAL, like the Castrop formula.

The authors did not perform lens constant optimization through zeroing out the mean error (ME), but they used IOL Con website lens factors. These constants are recommended when the sample size is small or more IOL models are included, but this is not the case.3 Lens constant optimization appears imperative: in fact, Table 2 of their study shows that 5 formulas have high ME (>±0.15 diopters), meaning that a systematic error is present. These formulas cannot be compared with each other without eliminating any systematic error.3

We have also some concerns regarding the statistical analysis. We applaud the adoption of the generalized estimating equation to account correlations between both eyes of the same patient, which is mandatory in this scenario.4 On the other hand, we wonder why the authors chose the mean absolute error (MAE) instead of the median absolute error (MedAE). The MedAE is less affected by outliers than the MAE, and the distribution of the absolute PE is not Gaussian, as reported by the same authors, and it is described better by the MedAE than by the MAE.5

In conclusion, the authors reported as a point of strength of the study that it analyzed firstly ZEISS AI IOL Calculator in short eyes, but they did not specify when and how they analyzed this method. Did they access to an online website, or did they use a specific device including ZEISS IOL Calculator? Access modality to IOL power calculation formulas must be descripted to guarantee the reproducibility of the study.

1. Kenny PI, Kozhaya K, Truong P, Weikert MP, Wang L, Hill WE, Koch DD. Efficacy of segmented axial length and artificial intelligence approaches to intraocular lens power calculation in short eyes. J Cataract Refract Surg 2023;49:697–703 2. De Bernardo M, Cione F, Capasso L, Coppola A, Rosa N. A formula to improve the reliability of optical axial length measurement in IOL power calculation. Sci Rep 2022;12:18845 3. Cione F, Gioia M, Pagliarulo S. Bias that should be avoided to obtain a reliable study of IOL power calculation after myopic refractive surgery. J Refract Surg 2023;39:68 4. De Bernardo M, Cione F, Rosa N. Re: Turnbull et al.: methods for intraocular lens power calculation in cataract surgery after radial keratotomy (Ophthalmology 2020;127:45–51). Ophthalmology 2020;127:e87 5. 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