Comparison of intraocular lens power calculation formulas with LENSTAR LS 900® Tsunekazu Hamada Hamada Eye Clinic, Osaka, Japan Disclosure statement of Financial Interest: Nothing to disclose
Purpose & Methods Purpose: To evaluate the accuracy of Olsen formula and commonly used three other intraocular lens power calculation formulas (SRK/T, Haigis, Hollady 2) using LenSTAR LS 900 Methods: The parameters were measured using LenSTAR LS 900 Abbott Medical Optics aspheric: Tecnis ZCB00V was studied At the 2-months post operative visit, apparent refractive results were obtained The prediction error and absolute prediction error were evaluated Off-set error was canceled out such that a mean postoperative refractive error of zero resulted
Clinical data of 204 eyes 130 patients 204 eyes Axial length(mm) 24.72± K Reading(D)42.84 ± ACD(mm) 3.22± Lens Thickness(mm) 4.44± Olsen: C-constant = 4.6(PhacoOptics) Haigis:A0 = , A1 = 0.311, A2 = 0.208(ULIB) SRK/T:A-constant = 119.9(personalized with HIC SOAP) Holladay 2:ACD constant = 5.97(personalized with HIC SOAP) IOL constants
Prediction Error ParameterHolladay 2SRK/THaigisOlsen N = 204 ME ±SD(D)0 ± ± ± ± 0.37 MAE(D) MedAE(D) <0.5D(%) <1.0D(%) MAE = mean absolute error; ME = mean arithmetic error; MedAE = median absolute error; SD = standard deviation of the error Wilcoxon signed-ranks test revealed significant difference in MAEs between Olsen and Holladay 2 (p=0.0001), SRK/T (p=0.0003), Haigis (p=0.0027)
Formula bias with the axial length was observed with Haigis, SRK/T and Holladay 2. No formula bias with the axial length was observed with Olsen. Prediction Error
Conclusions Olsen formula is more accurate than Haigis, SRK/T and Holladay 2. Formula bias with the axial length was observed with Haigis, SRK/T and Holladay 2. No formula bias with the axial length was observed with Olsen.