Comparison of Hill-RBF method with third and fifth generation formulas for intraocular lens power calculations in short eyes

Session Details

Session Title: Cataract
Session Date/Time: Saturday 10/02/2018 | 08:30-11:00
Paper Time: 10:24
Venue: Blue Hall

First Author: D.Lopes PORTUGAL
Co Author(s): S. Barros  S. Parreira  I. Machado  N. Campos        

Abstract Details

Purpose:

To evaluate the accuracy of 4 methods for intraocular lens (IOL) calculation in predicting refractive outcome after cataract surgery in eyes with a short axial length.

Setting:

Hospital Garcia de Orta

Methods:

Thirty-seven eyes of 31 patients with a short axial length (equal to or less than 22,00 mm) who had uneventful cataract surgery in the last 2 years were selected. Biometric data were measured by IOLMaster 500.  IOL models included Alcon SA60T (4 eyes), AMO Sensar AAB00 (5 eyes), AMO Tecnis PCB00 (12 eyes) and B&L Akreos MI60 (16 eyes). Four formulas for IOL power calculation were evaluated: Hill-RBF, Hoffer Q, Barrett Universal II and Holladay 1. The refractive prediction by each of the 4 formulas was calculated and compared with the actual refractive outcome to give the refractive prediction error.

Results:

The Hill-RBF and Holladay formula presented a lower median absolute error (0,34 and 0,33 respectively) than Hoffer and Barrett (0,44 and 0,41 respectively). No statistically significant differences in the median absolute error were found between the 4 formulas (P=0,068). The Hill-RBF and Holladay 1 formulas were able to achieve the highest percentage of eyes within 0.50 D of the target (72 % of eyes in both). Minimum and maximum prediction error (range of errors) for each formula: -0,46 to 1,39 in Hill- RBF; -0,84 to 1,33 in Holladay 1; -0,99 to 1,31 in Barrett and -1,15 to 1,00 in Hoffer.

Conclusions:

This is the first study comparing the new extended beta version of Hill-RBF calculator with 3rd and 5th generation formulas in short axial length eyes. The Hill-RBF and Holladay 1 formulas had the highest percentage of eyes within ±0.50 D of the target and the lowest median absolute errors. No statistically significant differences in the median absolute error were found between the 4 formulas. Hill-RBF presented the lowest range of refractive prediction errors. Although we have not used optimized lens constants and different IOL types, the Hill-RBF method performed at least at the same level as the other three formulas.

Financial Disclosure:

None