The accuracy of the double-K adjustment for third-generation intraocular lens calculation formulas in previous keratorefractive surgery eyes

Purpose: To evaluate the effect of the double-K (DK) modification on third-generation formulas. Methods: Thirty-eight previously myopic and 24 previously hyperopic eyes that underwent phacoemulsification with intraocular lens (IOL) insertion after Laser in situ keratomileusis (LASIK) were evaluated. Pre-LASIK refraction and keratometry, post-LASIK topography, axial length (AL), IOL type and power, and 1-month postphacoemulsification refraction were recorded spherical equivalent after phacoemulsification (SEpostphaco). Measured corneal power was adjusted using published and validated methods for postmyopic and posthyperopic LASIK. For each eye, and using SEpostphaco, different DK-IOL formulas were used to calculate the corresponding IOL power, the outcome measure, which was compared with the implanted IOL. Results: DK-Holladay 1 yielded the highest Pearson correlation coefficient (PCC), 0.955 for myopes and 0.943 for high myopes (AL>26 mm). Mean error (ME) and mean absolute error (MAE) for myopes for DK Sanders-Retzlaff-Kraff theoretical formula [DK-SRK/T] were 0.44±0.84 D and 0.75±0.61 D for DK-SRK/T compared with-0.04±0.67 D and 0.52±0.40 D for DK-Holladay 1 (P<0.001 and P=0.016, respectively), and 0.03±0.88 and 0.64±0.58 for DK-Hoffer Q. For high myopes, ME and MAE were 0.75±0.81 D and 0.84±0.69 D for DK-SRK/T, and-0.05±0.74 D (P<0.0001) and 0.57±0.45 D (P=0.019) for DK-Holladay 1. About 29% of DK-SRK/T eyes with large AL had MAE>1.5 D, compared with 0% for DK-Holladay 1 and 14% for DK-Hoffer-Q. Eyes with previous hyperopic LASIK faired similarly for all formulas, with similar PCCs, and only 8% in each category with MAE>1.5 D. Conclusions: DK-SRK/T overestimates IOL power in eyes with large AL, especially with concomitant steep pre-lasik keratometry. Among third-generation formulas, DK-Holladay 1 seems more accurate to use in postmyopic LASIK eyes.