Comparison of Refractive Error Measures by the IRX3 Aberrometer and Autorefraction

Purpose To compare the aberrometry-derived refractive error measurements from the IRX3 aberrometer (Imagine Eyes, Orsay, France) with a standardized measure of refractive error from the Shin-Nippon SRW-5000 (Japan) autorefractor in a large sample of school-aged children. Methods Participants were a subgroup of children from the Northern Ireland Childhood Errors of Refraction Study phase 2 (n = 161 9 to 10 years; n = 147 15 to 16 years). Refractive error was measured under cycloplegia (1.0% cyclopentolate HCl) with the IRX3 aberrometer followed by measurement with the Shin-Nippon autorefractor. Mean differences and 95% confidence intervals and limits of agreement were calculated for refractive vector components (M, J(0), and J(45)). Results Participants had a wide range of refractive error ranging from -6.00 to +8.00 diopters (D) spherical equivalent refraction. Fixed measurement biases (95% confidence interval) between instruments were small for both groups (9- to 10-year-olds: M, -0.20 +/- 0.65 D; J(0), -0.005 D; J(45), 0.05 D; 15- to 16-year-olds: M, 0.03 +/- 0.61 D; J(0), -0.04 D; J(45), -0.02 D). Statistically significant differences were found between instruments for M and J(45) for the 9- to 10-year-old group (p < 0.0001) and for J(0) for the 15- to 16-year-old group (p = 0.003). A statistically significant proportional measurement bias was found for the cylindrical components J(0) and J(45) for both groups (p < 0.0001), but no statistically significant proportional bias was found for M for either group. Conclusions This is the first study to explore refractive error measurements from the IRX3 aberrometer in children. The differences between instruments for all refractive components (M, J(0), and J(45)) were small for both groups (<0.25 D) and may not be considered clinically meaningful. Levels of agreement were also comparable to other studies investigating the validity of instruments measuring automated refraction in both adults and children. The results would suggest that these objective techniques produce similar results for assessment of refractive error in children.