Comparison of different IOL calculations for the implantation of toric IOLs in patiens with keratoconus
Free Paper Details
First Author: V.Webers THE NETHERLANDS
Co Author(s): M. Vandevenne N. Visser N. Bauer T. Berendschot R. Nuijts
To evaluate the accuracy of different IOL calculations for the implantation of toric IOLs in patients with keratoconus
University Eye Clinic Maastricht, the Netherlands
A total of thirty-six eyes of 25 patients with stable mild to moderate keratoconus (Krumeich stage 1 and 2) underwent cataract surgery with toric IOL implantation (Acrysof SN6AT3-T9 or Rayner T-flex). Back-calculations were performed for four different IOL calculations: first generation toric calculator (1st), K-values of biometry (IOLmaster, IOlm-Barrett), Scheimpflug (Pentacam, SimK-Barrett), and measured posterior cornea astigmatism (PCA) (Pentacam, PCA-Barrett) in the Barrett Toric Calculator. For residual astigmatism, vector analyses was performed for the mean prediction error (ME), mean absolute prediction error (MAE) and correction index (CI).
The mean preoperative corneal astigmatism was 3.88±1.30D (range 1.42-6.38D). ME was respectively 0.55±0.59D*, 0.67±0.65D*, 0.06±0.72D, -0.19±0.87D for 1st, IOLm-Barrett, SimK-Barrett, and PCA-Barrett (*significant different from zero). MAE was respectively 0.63±0.50D, 0.77±0.53D, 0.55±0.45D, and 0.67±0.57D for 1st, IOLm-Barrett, SimK-Barrett, and PCA-Barrett (p>.05). CI was significantly closer to ideal 1.0 for the SimK-Barrett (1.07±0.24) and PCA-Barrett (1.06±0.45) methods compared to 1st (1.20±0.21) and IOLm-Barrett (1.25±0.24) methods (p<.05). Percentage of patients with correction of astigmatism within 0.5D was respectively 46%, 34%, 59%, 45% for 1st, IOLm-Barrett, SimK-Barrett, and PCA-Barrett method.
Performing toric IOL calculations in keratoconus eyes is complex. The use of a 1st generation toric calculator or the K-values measured with biometry in the Barrett Toric Calculator results in a general overcorrection of pre-existent corneal astigmatism. Using K-values measured with Scheimpflug imaging (with or without measured posterior astigmatism) showed lower mean prediction errors and less overcorrection.