Autorefraction as an outcome measure of laser in situ keratomileusis

PURPOSE: To determine the limits of agreement between subjective refraction and autorefraction before and after laser in situ keratomileusis (LASIK) to assess whether autorefraction is a valid refractive outcome measure of refractive surgery. SETTING: Ultralase, Leeds, United Kingdom. METHOD: The prospective study involved consecutive preoperative normal patients and post-LASIK patients who had autorefraction using the Nidek ARK 700A autorefractor and careful subjective refraction (masked to autorefraction). Inclusion criteria were age greater than 18 years and healthy eyes with a visual acuity better than 0.1 logMAR (6/7.5) with or without previous LASIK. Refractions were compared by spherical equivalent (SE) using Bland-Altman limits of agreement and astigmatic vector difference using the median and the 95th percentile. The effect of time after treatment and treatment strength were explored. RESULTS: Data were collected from 208 preoperative patients and 237 post-LASIK patients. Preoperatively, the agreement between subjective refraction and autorefraction for the SE was -0.10 diopter (D) +/- 0.35 (SD) and the median difference for the astigmatic vector was 0.28 D with a 95th percentile of 0.72 D. Post-LASIK, the SE agreement was similar, -0.09 +/- 0.39 D, but the astigmatic vector agreement decreased slightly with a median of 0.31 D and a 95th percentile of 1.02 D. This decrease reflected poorer agreement in patients whose pre-LASIK refractive error was greater than +4.00 D. Removing this group brought the median astigmatic difference post-LASIK to 0.27 D with a 95th percentile of 0.87 D, similar to that in the preoperative normals. The percentage within +/-0.50 D and +/-1.00 D of the attempted correction was 56.1% and 78.5%, respectively, with subjective refraction and 51.2% and 78.1%, respectively, with autorefraction. CONCLUSIONS: Autorefraction showed excellent agreement with subjective refraction and was unaffected by refractive surgery except after LASIK for high hyperopia. Most outcomes were correctly classified in the standard categories (+/-0.50 D, +/-1.00 D), illustrating that autorefraction is a valid outcome measure of refractive surgery.