Abstract
Purpose :
While it is generally assumed that different devices will produce the same output for a given formula and that the third generation formulas results may differ mainly for extreme axial lengths, it is legitimate to question these assumptions. The goal of this study is to compare IOL power results for emmetropia for SRK/T, Hoffer Q and Holladay I formulas using the IOLMaster, DGH 6000 A-scan biometer, and UniversIOL calculators.
Methods :
Eyes with axial lengths ranging from 20 to 30mm with increments of 0.5 mm and mean corneal powers ranging from 38 to 50 diopters (D) with increments of 0.5 D generated a set of 525 cases that were entered in each calculator for each device with a target refraction of plano. SRK/T A constant was 119, HofferQ pACD constant 5.64 and Holladay I surgeon factor, 1.84. IOL power calculation according to these formulas were obtained using the IOL Master (IOL Master 500), A scan (DGH6000) and UniversIOL universal calculator.
Results :
The mean absolute difference between all three formulas was less than 1/50 D on all three calculators. For Holladay 1 and Hoffer Q the maximum absolute difference between calculators results did not exceed 0.1 D for any combination of axial length and mean corneal power, and with the exception of a zone in parameter space where the corneal height square root approaches zero, SRK/T yielded similar results. Results for comparison between formulas confirmed the known dependence and differences depending on axial length but revealed a persistent non physiological "cusp" in all three implementations with UniversIOL calculation being closest to the physiological result.
Conclusions :
Implementation of third generation formulas in multiple calculators on different platforms are consistent, with some rare but clinically relevant exceptions that are detailed. Differences between formulas are consistent with generally accepted axial length dependence recommendations but also differ in areas less often recognized. These differences can be manually/visually inspected by the surgeon on all three calculators considered and automatically flagged in UniversIOL.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.