Am Orthopt J  49: 123-130  (1999).

Naked Autorefraction in Children: Pitfalls and Perils

Mark Silverberg, M.D.
Eileen Schuler, C. 0.
Jay Rosin, M.D.
Julia Katz, M.D.
David Maberly, M.D.
Norman Medow, M.D.

Purpose: To evaluate the accuracy of autorefraction in a pediatric population.

Methods: Prospectively, 178 eyes (ages 2-16) underwent noncycloplegic (naked) and cycloplegic autorefraction (Nidek 1600), and cycloplegic retinoscopy (1% cyclopentolate). Compared to retinoscopy, autorefraction was considered accurate if within 0.5 diopters of sphere, 0.5 diopters of cylinder and 15 degrees of axis. When possible, a manifest refraction helped confirm retinoscopy.

Results: Naked autorefraction was accurate for sphere, cylinder and axis in 14.0%, 69.6% and 51.1% respectively. After cycloplegia, accuracy improved to 56.2%, 80.3%, and 51.1%. Naked spherical determination was most inaccurate in the youngest age groups (less than 8 years old). Cylinder and axis accuracy were age independent. Accuracy in patients with strabismus or amblyopia did not differ significantly from the general pediatric population.

Conclusions: Noncyloplegic (naked) autorefraction is relatively inaccurate compared to retinoscopy. Cycloplegic autorefraction is more reliable, and may serve as a useful adjunct to retinoscopy, especially when determining the astigmatism.

© 1999 The Board of Regents of the University of Wisconsin System