Intraocular lens power calculation in virgin eyes: Accuracy of the Barrett Universal II formula and a Ray tracing software

Purpose To compare the accuracy of Sirius ray tracing software with the Barrett Universal II formula for intraocular lens power prediction in virgin eyes. Methods Retrospective case series including 86 eyes that have undergone uneventful cataract surgery with SN60WF implantation. The median absolute error, mean prediction error, variance, and the percentage of eyes within ± 0.25 D, ± 0.50 D, ± 0.75 D, and ± 1.00 D of the prediction error in refraction were calculated. The correlation of prediction error with different baseline parameters was investigated. Results No differences were found between the median absolute error of the Barrett Universal II formula (0.226 D) and the ray tracing software with different intraocular lens centerings; apex (0.331 D), limbus (0.345 D), and pupil (0.342 D) ( p = 0.084). The variance, from lowest to highest, was the Barrett Universal II (0.144 D2), ray tracing-limbus (0.285 D2), ray tracing-pupil (0.285 D2), and ray tracing-apex (0.287 D2) ( p = 0.027). The Barrett Universal II formula showed a higher percentage of eyes within ± 0.25 D (56.98%), ± 0.50 D (82.56%), and ± 0.75 D (93.02%) compared to ray tracing software ( p < 0.01). A significant correlation between the prediction error of the Barrett Universal II formula and corneal diameter (r = 0.322, p = 0.002) and pupil diameter (r = 0.230, p = 0.033) was found. Also, a positive correlation between the prediction error of Sirius ray tracing and axial length ( p < 0.001) and pupil diameter ( p = 0.01) was found. Conclusion There is a trend of the Barrett Universal II formula to be more accurate than Sirius ray tracing software for intraocular lens power calculation in virgin eyes. This should be confirmed in future prospective comparative studies.