Harry S Geggel1. 1. Virginia Mason Medical Center, Section of Ophthalmology, Seattle, Washington 98101, USA. ophhsg@vmmc.org
Abstract
PURPOSE: To validate the Geggel ratio (GR) and consensus group in myopic excimer laser patients undergoing cataract surgery. METHODS: Two separate consecutively collected case series (group 1, 32 eyes; group 2, 34 eyes) were retrospectively analyzed. Data from group 1 was used to generate a new linear regression formula for the GR based on the Haigis formula. Predicted implant powers based on the Shammas, Latkany, Savini, Seitz, Haigis-L, and GR based on either SRK/T or Haigis formulas and combined consensus methods were computed. Mean error, range, percent within ± 0.5 D, ± 1.0 D and -1.0/+0.5 D were calculated. RESULTS: In group 1, the original GR (SRK/T) formula produced higher mean values (-0.44 D) and lower percentage within ± 0.5 D (41%) compared with the other 5 formulas (-0.03 to -0.33 D; 65%-82%). A linear regression formula (-0.142*ΔSE +0.2) for the GR was derived as a correction factor added to the Haigis formula. Group 2 validated the 5 formulas again plus the GR (Haigis). All group 2 formulas gave excellent results with means ranging from -0.18 to -0.37 D, percentages with ± 0.5 D ranging from 55% to 78%, and percentages within -1.0/+0.5 D ranging from 85% to 100%. Consensus formulas produced mean error -0.30 D, 70% ± 0.5 D, 93% -1.0/+0.5 D with only 2 outliers at 0.6 D and -1.1 D. CONCLUSIONS: These six methods achieve high levels of IOL accuracy. The GR method works better based on the Haigis formula. Consensus calculations minimize both under- and overcorrections.
PURPOSE: To validate the Geggel ratio (GR) and consensus group in myopic excimer laser patients undergoing cataract surgery. METHODS: Two separate consecutively collected case series (group 1, 32 eyes; group 2, 34 eyes) were retrospectively analyzed. Data from group 1 was used to generate a new linear regression formula for the GR based on the Haigis formula. Predicted implant powers based on the Shammas, Latkany, Savini, Seitz, Haigis-L, and GR based on either SRK/T or Haigis formulas and combined consensus methods were computed. Mean error, range, percent within ± 0.5 D, ± 1.0 D and -1.0/+0.5 D were calculated. RESULTS: In group 1, the original GR (SRK/T) formula produced higher mean values (-0.44 D) and lower percentage within ± 0.5 D (41%) compared with the other 5 formulas (-0.03 to -0.33 D; 65%-82%). A linear regression formula (-0.142*ΔSE +0.2) for the GR was derived as a correction factor added to the Haigis formula. Group 2 validated the 5 formulas again plus the GR (Haigis). All group 2 formulas gave excellent results with means ranging from -0.18 to -0.37 D, percentages with ± 0.5 D ranging from 55% to 78%, and percentages within -1.0/+0.5 D ranging from 85% to 100%. Consensus formulas produced mean error -0.30 D, 70% ± 0.5 D, 93% -1.0/+0.5 D with only 2 outliers at 0.6 D and -1.1 D. CONCLUSIONS: These six methods achieve high levels of IOL accuracy. The GR method works better based on the Haigis formula. Consensus calculations minimize both under- and overcorrections.