ESCRS - PP03.06 - Intraocular Lens Power Calculation In Anatomically Narrow Angle Eyes: A Large Multicenter Real-World Comparative Study Of Twelve Formulas

Intraocular Lens Power Calculation In Anatomically Narrow Angle Eyes: A Large Multicenter Real-World Comparative Study Of Twelve Formulas

Published 2026 - 44th Congress of the ESCRS

Reference: PP03.06 | Type: Free paper | DOI: 10.82333/kaj5-m893

Authors: Matthew D. Cooke 1 , Mr Timothy L. Cooke 2 , Dr David L. Cooke 3 , Dr Colya N. Englisch* 4

1Department of Ophthalmology,Mayo Clinic,Eau Claire,United States, 2Great Lakes Eye Care,Saint Joseph,United States, 3Great Lakes Eye Care ,Saint Joseph,United States;Department of Neurology and Ophthalmology, College of Osteopathic Medicine, Michigan State University,East Lansing,United States, 4Department of Experimental Ophthalmology,Saarland University,Homburg,Germany

Purpose

Eyes with primary angle closure suspect (PACS), primary angle closure (PAC), and primary angle closure glaucoma (PACG) frequently undergo phacoemulsification. However, intraocular lens (IOL) power calculation in these eyes is uncertain because eyes with anatomically narrow angles (ANAs) may have unexpected biometric relationships. The aim of this study was to understand how IOL power formulas treat this group of unusual eyes. Should modifications be undertaken to improve results? We evaluated the refractive predictive accuracy of twelve modern and legacy IOL power calculation formulas in mostly non-Asian eyes with ANAs.

Setting

This retrospective cohort study included 1,220 eyes from 1,220 patients who underwent cataract surgery at multiple centers across the USA between 01/01/2023 and 12/27/2024. Data were extracted from a de-identified database (Veracity, Carl Zeiss). ANA is an accepted indication for cataract surgery in the USA. If this code was submitted with surgery, we considered it an ANA eye. Gonioscopy and intraocular pressure findings were not recorded. We presume “ANA” included PACS, PAC, or PACG eyes.

Methods

SRK/T, Holladay 1 and 2, Haigis, Barrett Universal II (BUII)*, Hoffer Q, Hoffer QST*, RBF 3.0*, EVO 2.0*, Pearl DGS*, Kane*, and K6* formulas were evaluated with IOLcon lens constants (* = available at ESCRS.org). Subanalyses included short eyes (axial length [AL] < 22 mm) and deep eyes (anterior chamber depth [ACD] > 2.5 mm). Metrics included mean absolute error (MAE), mean refractive error (MRE), and proportions of eyes within ±0.50 D. MAE was used to compare formulas with a two-sided Wilcoxon–Holladay–Wang–Koch. Constant optimization was not possible because 44 IOL models were used, 23 models had less than 10 eyes. If MRE was ≤ 1/8th D, it was considered „on target.“

Results

Overall, 9 of 12 formulas were on target. The worst MRE was only 0.15 D. Optimization would thus have changed results minimally. Formulas had from 69% to 76% of eyes within 0.5 D of predicted refraction (3 formulas > 75%). Melles et al. formulas had from 72% to 84% within 0.5 D (7 of 9 > 75%) in normal eyes (PMID: 30980854). In 180 short eyes, Kane was best (MAE 0.46 D). Holladay 1 was 2nd (MAE 0.46) and BUII was 11th (MAE 0.52). 6 formulas were on target and all but the worst formula had between 60% and 68% of eyes within 0.5 D; only about 6% or 7% lower than Kenny et al. where 67% and 73% of short eyes without ANAs were within 0.5 D (PMID: 36940190). ESCRS formulas ranked 1 to 7 in 948 deep eyes, with 75% to 78% within 0.5 D.

Conclusions

In eyes with ANAs, optimizing lens constants would offer minimal improvement over using IOLcon.org constants. Formulas performed fairly well in all eyes, about 5% fewer eyes within 0.5 D than normal eyes. Short ANA eyes had about 6% fewer eyes within 0.5 D than other short eyes. Results were almost normal for deep eyes. Generally the 7 formulas available on the ESCRS website performed well, except for BUII in short eyes. In such situation, either Kane or Holladay 1 was the best choice.