Comparison of astigmatic prediction errors associated with new calculation methods for toric intraocular lenses

Purpose: To compare the prediction errors in residual astigmatism associated with new calculation methods for toric intraocular lenses (IOLs). Setting: Hospital da Luz, Lisbon, Portugal. Design: Retrospective case series. Methods: In eyes having cataract surgery with toric IOL implantation (Acrysof IC)), the predicted residual astigmatism by each calculation method was compared with the manifest refractive astigmatism. The prediction error in residual astigmatism was calculated by vector analysis. Results: The study evaluated 86 eyes (86 patients). All calculation methods resulted in overcorrection of with-the-rule astigmatism and undercorrection of against-the-rule astigmatism. For the original Alcon calculator, the centroid prediction error was 0.43 @ 170, which was reduced by the application of the Baylor nomogram (0.35 @ 169) or the Abulafia-Koch formula (0.34 @ 170). For the Holladay toric calculator, the centroid prediction error was 0.40 @ 168, which was reduced by the Baylor nomogram (0.35 @ 169), the Abulafia-Koch formula (0.25 @ 158), and the Goggin coefficient of adjustment (0.38 @ 170). The Barrett calculator and the newly introduced Alcon calculator yielded the lowest centroid prediction errors (0.17 @ 165 and 0.19 @ 164, respectively). The centroid prediction error of ray-tracing calculations (PhacoOptics) using real posterior corneal surface measurements was 0.32 @ 171. Conclusions: The Barrett toric calculator and the new Alcon calculator yielded the lowest astigmatic prediction errors. Of the nomogram methods, application of the Abulafia-Koch formula achieved the best results. The outcomes of toric IOL implantation might be improved by using 1 of these calculation methods. (C) 2017 ASCRS and ESCRS