Accuracy of thick-lens intraocular lens power calculation based on cutting-card or calculated data for lens architecture

Purpose: To compare the accuracy of a thick-lens intraocular lens (IOL) power formula using the manufacturer's cutting-card information (the Naeser 1 formula) and calculated data from open sources (the Naeser 2 formula) for IOL architecture; and to compare these results with the achievements of the Barrett Universal II, Haigis, Hoffer Q, Holladay 1, and SRK/T formulas. Setting: IRCCS GB Bietti Foundation, Rome, Italy. Design: Retrospective case series. Methods: For each IOL power formula, the prediction error in refraction was retrospectively calculated in eyes after phacoemulsification with implantation of the SN60WF posterior chamber IOL. The predictions made using the different formulas were optimized in retrospect by adjusting the respective constants. The mean arithmetic error (ME), the variance, and the median absolute error (Med-AE) were calculated, as well as the dependency of the arithmetic error on the axial length. Results: The study comprised 151 eyes. The Naeser 1 and Naeser 2 formulas were identical clinically and statistically. The ME (P = .278 with a one-way analysis of variance) and the variances for arithmetic errors (P = .248 with Levene's homogeneity test) were similar for all 7 formulas. A combined metric of ME, variance, MedAE, and arithmetic error on the axial length suggested the Naeser 2 formula as the most accurate, followed by the Naeser 1, Barrett Universal II, Haigis, SRK/T, Hoffer Q, and Holladay 1 formulas. Conclusions: The Naeser thick-lens IOL power equations based on calculated and manufacturer's IOL data provided similar accuracies and performed approximately as the best thin-lens formulas. Cutting-card information is not necessary in current IOL power calculation. (C) 2019 ASCRS and ESCRS