Comparison of Keratoconus Specific to Standard IOL Formulas in Patients With Keratoconus Undergoing Cataract Surgery.

To assess the performance of multiple intraocular lens (IOL) formulas in eyes with keratoconus. Eyes with stable keratoconus scheduled for cataract surgery with biometry measurements on the Lenstar LS900 (Haag-Streit) were included. Prediction errors were calculated using 11 different formulas, including two with keratoconus modifiers. Primary outcomes compared standard deviations, mean and median numerical errors, and percentage of eyes within diopter (D) ranges across all eyes with subgroup analysis according to anterior keratometric values. Sixty-eight eyes from 44 patients were identified. In eyes with keratometric values less than 50.00 D, prediction error standard deviations ranged from 0.680 to 0.857 D. Percentages of eyes within ±0.50 D of target ranged from 57.89% to 73.68% with no statistical differences among formulas. In eyes with a keratometric value of more than 50.00 D, prediction error standard deviations ranged from 1.849 to 2.349 D and were not statistically different with heteroscedastic analysis; percentages of eyes within ±0.50 D of target ranged from 0% to 18.18% with no statistical differences among formulas. Only keratoconus-specific formulas (Barrett-KC and Kane-KC) and the Wang-Koch axial length adjustment version of SRK/T resulted in median numerical errors not significantly different than 0, regardless of keratometric values. In keratoconic eyes, IOL formulas are less accurate than in normal eyes and result in hyperopic refractive outcomes that increase with steeper keratometric values. Using keratoconus-specific formulas and the Wang-Koch axial length adjustment version of SRK/T for axial lengths of 25.2 mm or greater improved IOL power prediction accuracy compared to other formulas.