INTRODUCTION: Information regarding the predictability of modern intraocular lens (IOL) calculation formulas in combined cataract surgery and trabeculectomy is lacking. This study aimed to evaluate the prediction error of the IOL calculation formulas incorporated in the European Society of Cataract Carl Zeiss). Prediction errors (PE), defined as the difference between postoperative and formula-predicted SE based on the IOL power implanted, were calculated at 3-6 months follow-up. Mean absolute error (MeanAE), median absolute error (MedAE) and the percentage of eyes within ±0.25 diopters (D), ±0.50D and ±1.00D of PE were determined. RESULTS: A total of 41 eyes from 41 patients with a mean axial length (AL) of 23.39±0.82mm were included. The mean postoperative SE was -0.55±0.71D. The formula presenting the lowest MedAE was EVO (0.29D), followed by SRK/T (0.30D), Cooke K6 (0.32D), Kane (0.34D), Hoffer QST (0.35D), Hill-RBF (0.35D), PEARL GDS (0.38D), and finally Barrett Universal II (0.43D). The percentage of eyes within ±0.50D of PE was superior for Hill-RBF (65.9%), followed by Cooke K6 (64.1). For eyes within ±1.00D of PE, the formula showing a higher proportion was Hoffer QST (92.7%), followed by Kane (90.2%). Barrett Universal II was the formula with less percentage of eyes in all groups (29.3% for ±0.25D PE, 53.7% for ±0.50D PE, and 78.0% for ±1.00D PE). CONCLUSION: Overall, all formulas evaluated showed a robust performance. The finding that none of the modern formulas exhibited a significant advantage when compared to SRK/T suggests that eyes undergoing combined cataract surgery and trabeculectomy should be evaluated as a singular subgroup in terms of IOL power calculation.
Geada et al. (Tue,) studied this question.