Related Experiment Video
Updated: Jan 9, 2026
Comparison of Agreement and Accuracy using Binocular Wavefront Optometer with Autorefractor and Phoropter
Published on: September 16, 2025
Refining Intraocular Lens Calculation in Phacoemulsification and DMEK: Comparison of Three Methods
David Mingo-Botín1,2, Pablo de-Arriba-Palomero1, Oscar Gris2
1From Cornea and Ocular Surface Unit, Ophthalmology Department, Ramón y Cajal University Hospital, Madrid, Spain.
Purpose:
To describe tomographic alterations leading to refractive changes after Descemet membrane endothelial keratoplasty (DMEK) and to evaluate three keratometric adjustment methods compared to conventional unadjusted formulas.
Methods:
Biometric, topographic, and refractive data from a retrospective case series of patients who had undergone phacoemulsification and DMEK were collected. Eyes with corneal scarring or preoperative epithelial edema preventing accurate keratometric readings were excluded. Mean prediction error, mean (MAE), and median absolute error (MedAE), standard deviation, and percentage of eyes within ±0.50, ±1.00, and ±1.50 diopters (D) were calculated using four methods: conventional (SRK/T, Haigis, Barrett Universal II); with a keratometric adjustment using a fictious keratometric index based on the change of preoperative to postoperative back to front (BF) ratio; applying keratometric adjustment only when residual hyperopia was expected based on preoperative parameters (HaigisCombi); and a linear regression keratometric adjustment (Haigis-D).
Results:
A total of 112 eyes (93 patients) from two centers were included. Significant changes were found in all tomographic parameters. Mean posterior keratometry and asphericity, BF ratio, and relative pachymetry correlated with postoperative hyperopic error. The three keratometric adjustments significantly reduced the hyperopic prediction error. The Haigis-D formula obtained the lowest MAE and MedAE (0.62 and 0.52 D, respectively), whereas the highest percentage of eyes within ±1.00 D was achieved by the HaigisCombi (83.8%).
Conclusion:
The three methods demonstrated a reduction in prediction error and the proportion of hyperopic results when compared to the unadjusted formulas. The most balanced method was based on the BF ratio as a predictor of residual hyperopia, whereas the linear keratometry correction seemed more extrapolable. Results were mixed when adjustments based on data from one center were applied to data from the other center. The proposed keratometric adjustments need to be validated by a prospective study.
