Accuracy of intraocular lens power formulas in eyes with keratoconus: Multi-center study in Japan

Purpose

To assess the accuracy of intraocular lens (IOL) power formulas, namely, SRK/T, Haigis, Barrett Universal II, Barrett True-K for keratoconus, Kane formula, and Kane formula for keratoconus, for cataract with keratoconus in Japanese eyes.

Setting

Five surgical sites in Japan.

Design

A retrospective case series.

Methods

Eyes with keratoconus undergoing cataract surgery were included. Postoperative refraction was compared with the prediction by the formulas. Visual acuity, manifest spherical equivalent, prediction error (PE), and mean absolute errors (MAEs) were determined 1 month postoperatively. The PE within 0.50 diopter (D), 1.00 D, and 2.00 D were compared between IOL formulas. Subgroup analysis based on the steepest keratometry (stage 1, ≤ 48 D; stage 2, > 48 D and ≤ 53 D; and stage 3, > 53 D) was performed. The relationship between PE and preoperative biometric data were assessed.

Results

Fifty eyes were included. The MAE of the Barrett True-K for keratoconus, Kane keratoconus, and Kane formulas were significantly lower than that of Haigis. A statistically significant difference in the prediction accuracy within ± 0.50 D was found between Kane keratoconus and Haigis. The prediction accuracy of the Barrett True-K for keratoconus, SRK/T, and Kane within ± 1.00 D was statistically significant compared with that of Haigis. In stage 3, the Barrett True-K for keratoconus had a significantly lower MAE than SRK/T and Haigis.

Conclusion

Keratoconus-specific formulas were more accurate than existing formulas in Japanese eyes. The Barrett True-K formula for keratoconus had higher prediction accuracy in severe keratoconus.