Accuracy of a universal theoretical formula for power calculation in pediatric intraocular lens implantation.

Purpose: To compare the accuracy of Barrett Universal II formula with other formulas (Holladay 2, Hoffer Q, and SRK/T formulas) in the prediction of postoperative refraction for pediatric intraocular lens implantation.

Setting: Academic medical center/children's hospital, San Francisco, California.

Design: Retrospective case series.

Methods: Children aged 16 years or younger who underwent cataract extraction and IOL implantation (2012 to 2019) and had refraction assessed at 3 to 16 weeks postoperatively were included. Prediction error (PE) was calculated as postoperative mean spherical equivalent minus the target refraction. Mean, median, and standard deviation was calculated for PE and absolute PE. Performance across covariables (axial length, age, biometry type, keratometry, etc.) was studied, and a multivariate regression analysis was performed using a single prediction model for each formula.

Results: Sixty-four eyes of 64 patients, aged 1.5 to 15.5 years, were included. Barrett Universal II formula had the lowest mean PE (-0.22 diopters [D]), SD (1.18 D), median PE (-0.26 D), and median absolute PE (0.71) compared with those of the other formulas. Holladay 2 formula performed similarly to Barrett Universal II formula, and SRK/T formula had the greatest mean PE (-0.50 D) and SD (1.22 D). Barrett Universal II formula predictions were stable across all variables.

Conclusions: Barrett Universal II formula demonstrated similar or superior performance when compared with other formulas in this pediatric study. Holladay 2 formula performed similarly to Barrett Universal II formula, and SRK/T formula had the least reliable performance, across several key biometric characteristics. Although PEs can be highly variable in pediatric populations, this study supports Barrett Universal II formula as a reasonable and reliable option for lens power calculation in children, including those with extreme biometric measurements.