A Comparison of the Accuracy of 6 Modern Toric Intraocular Lens Formulas

To compare the accuracy of the Abulafia-Koch, the Barrett, the EVO 2.0, the new Holladay 2 with total surgical-induced astigmatism, the Kane, and the Næser-Savini toric intraocular lens (IOL) power formulas using a large database of toric IOL refractive outcomes.

Retrospective consecutive case series.

Eight hundred twenty-three eyes of 823 patients who had a toric IOL inserted during surgery.

One eligible eye from patients having uncomplicated cataract surgery with insertion of an Alcon SN6AT(2-9) IOL (Alcon Laboratories, Inc, Fort Worth, TX) from 1 surgeon were included in the study. Both preoperative and postoperative biometry were measured using either the IOLMaster 500 or 700 (Carl Zeiss Meditec AG, Jena, Germany). Using vector calculation, the predicted postoperative refractive astigmatism was calculated for each formula. This was compared with the actual postoperative refractive astigmatism to give the prediction error.

Mean absolute prediction error, standard deviation of the prediction error, and percentage of eyes with a prediction error within ±0.50 diopter (D).

The Kane formula showed the highest proportion of eyes with a prediction error within ±0.50 D with 65.6%, followed by the Barrett formula (59.9%), Abulafia-Koch formula (59.5%), EVO 2.0 formula (58.9%), Næser-Savini formula (56.7%), and Holladay 2 formula (53.9%). The Kane formula showed a statistically significantly lower mean absolute prediction error (P < 0.001) and a significantly lower variance of the prediction error (P < 0.01) compared with all other formulas. No statistically significant difference existed among the mean absolute prediction errors for the Abulafia-Koch, Barrett, and EVO 2.0 toric formulas.

Use of the Kane toric formula significantly improved the prediction of postoperative astigmatic outcome compared with the other formulas studied.