Subjects

Twenty-one subjects were recruited from patients attending the refractive surgery service at Mayo Clinic. All patients had myopia or myopic astigmatism, were >21 years old, and were determined to be suitable candidates for LASIK after a rigorous screening examination. Subjects were excluded if they had any corneal abnormalities; a history of ocular disease, trauma, or surgery; or diabetes mellitus or other systemic disease known to affect the eye; or if they used ocular medications. Systemic medications were permitted unless they were known to affect the cornea or anterior segment. Subject age at surgery was 38 ± 10 years (mean ± standard deviation; range, 22-54 years). Patients who developed new ocular conditions during the follow-up period were excluded from subsequent analysis if it was determined that the new condition would interfere with the outcomes. Fellow eyes of unoperated normal myopic controls were examined concurrently; age of controls was 43 ± 7 years (range, 29-55 years). This study complied with the Health Insurance Portability and Accountability Act and was approved by the Mayo Clinic Institutional Review Board. Informed consent was obtained from all subjects after explanation of the nature and possible consequences of the study.

Randomization

Patients were stratified by ocular dominance and then 1 eye of each patient was randomized to LASIK with the flap created by a femtosecond laser, and the other eye to LASIK with the flap created by a mechanical microkeratome. Ocular dominance was tested by asking patients to use both hands to frame a distant target while an observer determined with which eye the target was aligned.

Laser In Situ Keratomileusis Procedure

Bladeless flaps were created with a 15-kHz femtosecond laser (IntraLase FS; IntraLase Corp, Irvine, California, USA). All flaps had a superior hinge and intended thickness of 120 μm. Raster line and spot separation were 9 and 11 μm, respectively; raster energy was 2.3 μJ, and side-cut energy was 2.5 μJ. Flaps created by the mechanical microkeratome (Hansatome; Bausch & Lomb, Rochester, New York, USA) had a superior hinge with intended thickness of 180 μm. Non-wavefront-guided ablation of the stromal bed was performed with a VISX Star S4 excimer laser (VISX, Santa Ana, California, USA) with radiant exposure of 160 mJ/cm ² . Emmetropia was attempted in all cases by using an ablation zone that ranged from 6.5 × 6.5 mm for spherical corrections to 6.5 × 5.0 mm for astigmatic corrections. Both eyes of each patient were treated on the same day. All procedures followed a standard protocol: the bladeless flap was created first on the eye randomized to the femtosecond laser; the fellow eye then received a full LASIK procedure with the flap created by the mechanical microkeratome; finally, LASIK was completed on the first eye by separating and lifting the flap created by the femtosecond laser and ablating the stroma. It was not possible to mask patients as to which treatment was received in each eye. Postoperative topical medication regimens were identical for each eye and consisted of ciprofloxacin ophthalmic solution 4 times per day for 5 days, and fluorometholone 0.1% 4 to 8 times daily with a taper over 3 weeks.

Outcome Measures

Patients were examined before LASIK and at 1, 3, 6, 12, and 36 months after LASIK. At each examination, corneal topography, high-contrast visual acuity, low-contrast visual acuity, and manifest refraction were recorded. Whole-eye aberrometry was introduced to the study protocol during the enrollment phase and thus data were not available for all eyes at every examination; low-contrast visual acuity data were only available at 12 and 36 months after LASIK. Measurements were made by observers masked as to which treatment was received in each eye.

Corneal topography was recorded using a Humphrey Atlas corneal topography system (Humphrey Systems, Pleasanton, California, USA). Two to 4 topographic maps were recorded for each eye, centered over the line of sight. The videokeratography maps of each cornea were examined by 1 masked observer, and the map with the most complete image and the smallest nondigitized areas was selected for assessment of wavefront errors from the anterior corneal surface. Aberrations for the whole eye were measured with a Hartman-Shack aberrometer (VISX Wavescan, Santa Ana, California, USA). Two to 4 whole-eye wavefront analyses were acquired for each eye, and the examination with the highest quality and largest pupil diameter, as indicated by the aberrometry software, was selected by a masked observer.

High-contrast visual acuity was measured using the electronic Early Treatment of Diabetic Retinopathy Study testing protocol. Uncorrected visual acuity (UCVA) and best spectacle-corrected visual acuity (BCVA) were recorded as letter scores, which were converted to logarithm of the minimal angle of resolution (logMAR).

Best spectacle-corrected low-contrast visual acuity (LCVA) was measured by using a backlit 10% Sloan Translucent Low Contrast Chart (Precision Vision, La Salle, Illinois, USA) with a testing distance of 4 meters. LCVA was measured in a darkened room and recorded as letter scores under photopic (screen brightness, 139 cd/m ² ) and mesopic (screen brightness, 1.1 cd/m ² ) conditions; mesopic conditions were achieved by placing a neutral density (2 ND) filter in front of the low-contrast chart. ETDRS letter scores were converted to logMAR and Snellen equivalent.

Wavefront Errors

From the corneal topography (anterior corneal surface) and whole-eye aberrometry data, the wavefront errors over 4-mm and 6-mm pupils were calculated using VOLCT (Sarver and Associates, Inc, Carbondale, Illinois, USA) and decomposed into Zernike polynomials to the 6th order. All high-order aberrations were summarized for Zernike orders 3-6 as:

Statistical Analysis

The study was powered a priori to detect a difference of 0. 15 logMAR in UCVA or BCVA at 3 years after LASIK by assuming that the standard deviation of the difference in visual acuity would be 0.15 logMAR. This required a minimum sample size of 16 subjects (α = 0.05/6, β = 0.20, paired test). Differences between treatments at each examination, and differences between preoperative and postoperative examinations for each treatment, were assessed by using 2-tailed paired t tests if the data were distributed normally and Wilcoxon signed rank tests if they were not. P values were adjusted for multiple comparisons by using the Bonferroni method, and P < .05 was considered statistically significant. All statistical analyses were performed with Statistical Analysis System Version 9.1.3 (SAS Institute Inc, Cary, North Carolina, USA). Minimum detectable differences were calculated post hoc for nonsignificant differences assuming α = 0.05/5 or 0.05/6 (depending on the comparison) and β = 0.20.

Incomplete whole-eye wavefront error data and no LCVA data were available in patients before LASIK. To determine if differences in whole-eye wavefront errors and LCVA exist between fellow unoperated eyes, we compared these between fellow eyes of the normal myopic controls.