In the present era of refractive surgery, various methods are used to correct myopia and myopic astigmatism. Transepithelial photorefractive keratectomy (transPRK), femtosecond laser-assisted in situ keratomileusis (FS-LASIK), and small-incision lenticule extraction (SMILE) are the 3 major refractive surgeries performed worldwide, and each has its characteristics.1 TransPRK has the earliest recovery of normal corneal sensitivity but also the longest recovery time, the most amount of discomfort, and the possibility of corneal haze postoperatively.2 FS-LASIK possesses the fastest visual recovery, whereas the corneal nerve recovery is the slowest, and flap displacement is possibly lifelong.3,4 Unlike LASIK, SMILE is a flapless and minimally invasive procedure that involves greater tectonic strength and less dry eye;4,5 however, the lack of both static cyclotorsion compensation and automated centration is regarded as a major drawback.4,5

For myopic astigmatic correction, accurately locating and correcting the vector becomes the key during surgery. As a vector parameter, both the magnitude and axis are crucial for obtaining satisfactory surgical outcomes in patients with myopic astigmatism.6 Few studies have discussed whether SMILE without an automated iris registration and cyclotorsion control could perform as well as FS-LASIK or transPRK for astigmatism correction. Controversial results of astigmatic correction between SMILE and LASIK groups; some obtained similar outcomes, whereas some studies found a better visual outcome in the LASIK group.7–11 A previous randomized contralateral study showed that residual manifest refraction cylinder of less than 0.25 diopters (D) representing the accuracy of cylinder correction was achieved in 81.8% of the eyes (18/22) in the LASIK group and 50% (11/22) of the eyes in the SMILE group (P < .001).11 With the advanced methods in SMILE, Zhao et al. proposed a corneal marking optimized technique that improved the astigmatism correction outcomes in SMILE and achieved no significant difference compared with the FS-LASIK group.12 Besides, a retrospective contralateral study showed that SMILE with a triple centration technique was comparable with corneal wavefront-guided transPRK.13

Our previous study confirmed that cross-assisted alignment for head positioning in the SMILE procedure was safe and effective for astigmatism correction.14 Therefore, the question was posed whether cross-assisted SMILE without a static eye-tracking system could perform as well as FS-LASIK or transPRK. The purpose of the current study is to compare the postoperative outcomes among cross-assisted SMILE, FS-LASIK, and transPRK for the correction of astigmatism.

METHODS Patients

This study is part of a prospective study of consecutive patients who received SMILE, FS-LASIK, and transPRK surgery between January 2019 and August 2022 by the same refractive surgeon (Y.-F.Y.) at the Eye Hospital of Wenzhou Medical University in Hangzhou. The study conformed with the tenets of the Declaration of Helsinki, and written informed consent was obtained from each participant. This study was approved by the Ethics Committee of the Eye Hospital of Wenzhou Medical University (No. KYK2018-29).

Inclusion criteria for this study were myopia <10 D with refractive astigmatism of −1.00 to −2.75 D; age 18–40 years, with stable refraction for at least 1 year; and have corrected distance visual acuity (CDVA) better than 20/25 Snellen refraction. Exclusion criteria for refractive surgery were a corneal thickness of <480 μm, suspicion of keratoconus on corneal topography, cataract, ocular inflammation, or infection. Patients were instructed to stop wearing soft contact lenses at least 2 weeks preoperatively and stop wearing rigid contact lenses at least 4 weeks preoperatively. All patients met the surgical indications and chose one of the corneal refractive surgeries by themselves. The residual corneal stromal bed thickness after SMILE was ≥280 μm, FS-LASIK ≥280 μm, and transPRK ≥360 μm.

Preoperative and Postoperative Assessments

Preoperatively, all patients underwent a detailed ophthalmological examination, including manifest refraction, evaluation of the logMAR for uncorrected distance visual acuity (UDVA), CDVA, pupil size (Atlas, Carl Zeiss Meditec AG), slitlamp examination, intraocular pressure (IOP) (noncontact tonometer, Topcon Corp.), corneal topography (Keratron Scout, Optikon SpA), Scheimpflug-based corneal topography (Pentacam HR Type 70900, Oculus Optikgeräte GmbH), and indirect fundoscopy. Experienced optometrists performed the manifest refractions. The corneal refractive power and aberrations in the central zone of 6 mm were also measured using Scheimpflug camera imaging. The same examinations were also performed at 1 month, 3 months, and 6 months postoperatively.

Surgical Procedure

All surgeries were performed by the same surgeon (Y.-F.Y.). In the SMILE group, the head position was realigned under a cross-beam as described previously.14 Before the standard SMILE procedure, the head position was meticulously aligned using a cross-beam assistant (Figure 1 and see Supplementary Figure 1, available at https://links.lww.com/JRS/A980). First, the horizontal and vertical lines of the cross-beam were meticulously aligned with the corresponding green reference lines on the headrest. Second, with the patient laid and eyes closed, the horizontal line was aligned with the outer canthus of both eyes; meanwhile, the vertical line was carefully aligned with the brow and the midline of the nasal bridge. In the final step, the patient was instructed to remain still, and the cross-beam apparatus was deactivated before the patient opened their eyes.14 The procedure was performed using a 500 kHz VisuMax Femtosecond Laser System (Carl Zeiss Meditec AG) with a pulse energy of 140 nJ. After the femtosecond laser application was completed, the lenticule was dissected and extracted completely.15 The lenticule diameter (optical zone) was set between 6.00 and 6.70 mm, and the cap thickness was within 110 to 120 μm. In the FS-LASIK group, corneal flaps were created using the VisuMax Femtosecond Laser System, with flap thicknesses of 110 μm and flap diameters of 8.10 mm; ablation was performed by the AMARIS 1050S excimer laser (Schwind eye-tech-solutions GmbH & Co. KG). In the transPRK group, ablation of the cornea was performed in a single step using the AMARIS 1050S excimer laser platform. In the SMILE group, the rotation of eyes was minimized by the cross-assisted alignment of the head position; in FS-LASIK and transPRK groups, the head position was adjusted by the static cyclotorsion control system in the AMARIS platform with the value of static cyclotorsion within ±2 degrees.14,16

Figure 1.:

Procedures of the cross-axial alignment method. A: The horizontal green line on the headrest represents the ideal alignment of the outer canthus of the patient's eyes. The vertical green line represents the line alignment of the patient's nasal septum. B: The cross-beam is turned on and aligned by aligning the horizontal and vertical lines with the corresponding green lines on the headrest (B, left). After the patient lies down with eyes closed, the horizontal line is aligned with the outer canthus of both eyes, and the vertical line is aligned with the middle line of the brow and nasal bridge.

Postoperatively, topical levofloxacin 0.5% eyedrops (Cravit, Santen, Inc.) were applied 4 times daily for 1 week and preservative-free artificial tears 4 times daily for at least 6 months in all groups. The fluorometholone 0.1% eyedrops (Allergan, Inc.) were used 4 times a day (tapered over 4 weeks) in the SMILE and FS-LASIK groups and 2 to 3 months in the transPRK group. Once epithelial closure was complete, the contact lens was removed.

Vector Analysis of Astigmatism

The astigmatic correction was compared using vector analysis of Alpins using Assort software (Assort Pty Ltd.).17 The manifest refraction at the spectacle plane was converted to the corneal plane, considering a vertex distance of 12 mm. The efficacy index was the ratio of the postoperative UDVA to preoperative CDVA, and the safety index was the ratio of the postoperative to preoperative CDVA in decimals. Efficacy, predictability, stability, and safety were assessed using standard graphs.18 Target-induced astigmatism (TIA) was the change in astigmatic expected to be caused by surgery; surgically induced astigmatism (SIA) was the change in astigmatic induced by the surgery in actuality; the difference vector (DV) was the vector difference of residual astigmatism; the correction index (CI) was the ratio of SIA to TIA; the magnitude of error (MofE) was the arithmetic difference between SIA and TIA; the angle of error (AofE) was the angle between the axis of SIA and TIA; and the index of success was defined as DV divided by TIA.

Statistical Analysis

Statistical analyses were performed using SPSS (v. 26.0). Statistical power calculation was performed using PASS (v. 11.0). Only the right-eye data were analyzed. Independent t tests or Mann-Whitney U tests were used to compare continuous variables between the groups. Categorical variables were evaluated using the χ2 test or Fisher exact test. Differences were considered statistically significant at P < .05.

RESULTS Demographics of Patients

We evaluated a total of 154 right eyes of 154 patients, aged 18 to 40 years, who had myopic astigmatism of −1.00 to −2.75 D and underwent SMILE, FS-LASIK, and transPRK with fully manifest refraction correction, with 64 eyes in the SMILE group, 42 in the FS-LASIK group, and 48 in the transPRK group. The characteristics of the 3 groups are presented in Table 1. Preoperatively, there were no significant differences in baseline characteristics among the 3 groups, including age (P = .09), sex ratio (P = .63), manifest sphere error (P = .43), cylinder error (P = .2), manifest spherical equivalent (SE; P = .38), IOP (P = .84), logMAR CDVA (P = .3), and central corneal thickness (CCT; P = .06).

Table 1. - Characteristics of eyes that underwent SMILE, FS-LASIK, and transPRK Characteristics SMILE (n = 64) FS-LASIK (n = 42) transPRK (n = 48) Overall
P value P value (pair comparison)a 1 vs 2 1 vs 3 2 vs 3 Gender (M) (%) 50.0
M:F = 32:32 42.9
M:F = 18:24 41.7
M:F = 20:28 .63 .47 .38 .91 Age (y) 24.44 ± 5.77 26.55 ± 5.77 24.21 ± 5.01 .09 .17 .99 .14 Preop  Sphere error (D) −4.82 ± 1.53 −5.23 ± 2.06 −4.81 ± 1.65 .43 .72 .99 .78   Range −7.75, −1.75 −9.50, −1.25 −8.00, −1.00  Cylinder error (D) −1.42 ± 0.46 −1.55 ± 0.56 −1.59 ± 0.57 .2 .64 .27 .99   Range −2.75, −1.0 −2.75, −1.0 −2.75, −1.0  Spherical equivalent (D) −5.53 ± 1.55 −6.00 ± 2.11 −5.61 ± 1.71 .38 .54 .99 .88   Range −8.75, −2.25 −10.0, −2.0 −8.75, −1.75  CDVA (logMAR) −0.01 ± 0.03 −0.02 ± 0.03 −0.01 ± 0.02 .3 .99 .99 .37  CCT (μm) 539.05 ± 25.86 533.21 ± 26.97 527.44 ± 23.41 .06 .75 .054 .85  IOP (mm Hg) 14.76 ± 2.34 15.02 ± 2.43 14.84 ± 2.05 .84 .99 .99 .99 Postop  Sphere error (D) 0.30 ± 0.35 0.38 ± 0.35 0.44 ± 0.36 .12 .76 .13 .99   Range −0.50, 1.00 −0.25, 1.50 −0.25, 1.50  Cylinder error (D) −0.23 ± 0.25 −0.40 ± 0.28 −0.42 ± 0.32 <.001* .009* .001* .99   Range −0.75, 0 −1.00, 0 −1.00, 0  Spherical equivalent (D) 0.19 ± 0.33 0.18 ± 0.33 0.23 ± 0.33 .76 .99 .99 .99   Range −0.75, 0.75 −0.375, 1.50 −0.50, 1.00  UDVA (logMAR) −0.06 ± 0.04 −0.05 ± 0.04 −0.05 ± 0.04 .41 .99 .59 .99  CDVA (logMAR) −0.03 ± 0.04 −0.03 ± 0.04 −0.02 ± 0.04 .35 .99 .44 .99  Efficacy index 1.12 ± 0.13 1.10 ± 0.14 1.11 ± 0.13 .57 .88 .99 .99  Safety index 1.06 ± 0.13 1.04 ± 0.12 1.04 ± 0.11 .59 .99 .99 .99

CCT = central corneal thickness

Results are expressed as mean ± SD

*Statistically significant

aMultiple correction was performed by Bonferroni correction

Visual Acuity, Efficacy, and Safety

At 6 months postoperatively, all eyes (100%) achieved a UDVA of 20/25 or better. Sixty-four (100%), 41 (98%), and 48 (100%) eyes had a UDVA of 20/20 or better in the SMILE, FS-LASIK, and transPRK groups, respectively (Figure 2, A). The postoperative UDVA was similar among the 3 groups (P = .41; Table 1), and relative to the preoperative CDVA, the UDVA improved in 39 (61%), 21 (50%), and 28 (58.3%) of the treated eyes in the SMILE, FS-LASIK, and transPRK groups, respectively, displaying a gain of at least 1 line in the 6-month postoperative UDVA (Figure 2, B). Five (7.8%), 5 (11.9%), and 4 (8.3%) eyes in the SMILE, FS-LASIK, and transPRK groups, respectively, lost 1 line of CDVA postoperatively (Figure 2, C). At 6 months postoperatively, there was no significant difference in the mean efficacy index among the SMILE (1.12 ± 0.13), FS-LASIK (1.10 ± 0.14), and transPRK (1.11 ± 0.13) groups (P = .57; Table 1). In addition, there was no difference in the mean safety index among the 3 groups (P = .59; Table 1).

Figure 2.:

Visual outcomes of the alignment and control groups after SMILE. A: Cumulative 6-month postoperative UDVA and preoperative CDVA. B: Changes in the Snellen lines of the postoperative UDVA compared with the preoperative CDVA. C: Changes in the Snellen lines of the postoperative CDVA compared with the preoperative CDVA. D: Attempted vs achieved changes in SER at 6 months postoperatively. E: The accuracy of the SER compared with the intended target. F: The stability of the SER over 6 months. G: The distribution of preoperative and 6-month postoperative cylinders. H: TIA vs SIA vectors at 6 months postoperatively are shown. I: The refractive astigmatism angle of error distribution is shown at 6 months postoperatively. SER = spherical equivalent refraction; SIA = surgically induced astigmatism; TIA = target-induced astigmatism

Manifest Refraction

At 6 months postoperatively, no significant difference was observed in the SE (P = .76) or manifest sphere error (P = .12) among the 3 groups (Table 1 and Figure 2, F). The R2 of the linear regression model of the attempted SE vs achieved SE was 95% in the SMILE group, 98% in the FS-LASIK group, and 96% in the transPRK group, without any significant difference (Figure 2, D). SE predictability was within ±0.5 D in 92.8% of eyes in the FS-LASIK group compared with 89.2% in the SMILE group and 89.7% in the transPRK group (Figure 2, E). The predictability of astigmatism correction was significantly better with SMILE, with 92.2% of eyes within ±0.5 D, compared with the FS-LASIK (78.6%) and transPRK (79.2%) groups (Figure 2, G). Moreover, there was a significant difference in the postoperative cylinder among the SMILE, FS-LASIK, and transPRK groups (−0.23 ± 0.25 D vs −0.40 ± 0.28 D vs −0.42 ± 0.32 D, P < .001; Table 1).

Vector Analysis

Vector analysis of astigmatism was performed using the Alpins method. At 6 months postoperatively, the TIA was not different among the 3 groups (P = .12), but the mean SIA was significantly smaller in the SMILE group than in the other 2 groups (P = .003; Table 2 and Supplementary Figure 2, available at https://links.lww.com/JRS/A980). The R2 of the linear regression model between TIA and SIA was 65% in the SMILE group, 78% in the FS-LASIK group, and 64% in the transPRK group (Figure 2, H). However, statistically significant outcomes were observed for DV (P = .001; Table 2). In addition, the index of success was also significantly better in the SMILE group (0.20 ± 0.22) than that in the FS-LASIK (0.31 ± 0.26) and transPRK groups (0.32 ± 0.27; P = .02, Table 2). Respectively, 53 (82.8%), 36 (85.7%), and 37 (77.1%) eyes obtained an AofE within ±5 degrees (P = .55; Figure 2, I).

Table 2. - Comparison of the vector analysis parameter of eyes among SMILE, FS-LASIK, and transPRK groups Parameter SMILE (n = 64) FS-LASIK (n = 42) transPRK (n = 48) Overall
P value P value (pair comparison)a 1 vs 2 1 vs 3 2 vs 3 TIA 1.23 ± 0.40 (0.82, 2.30) 1.37 ± 0.48 (0.81, 2.59) 1.40 ± 0.48 (0.85, 2.57) .12 .46 .16 .99 SIA 1.19 ± 0.37 (0.47, 2.32) 1.52 ± 0.71 (0.41, 3.17) 1.53 ± 0.73 (0.43, 3.27) .003* .02* .01* .99 DV 0.24 ± 0.24 (0.01, 0.77) 0.40 ± 0.29 (0.00, 1.01) 0.43 ± 0.31 (0.01, 1.01) .001* .01* .001* .99 CI 0.98 ± 0.20 (0.54, 1.77) 1.10 ± 0.29 (0.48, 1.86) 1.09 ± 0.38 (0.46, 2.17) .07 .15 .18 .99 IOS 0.20 ± 0.22 (0.01, 0.88) 0.31 ± 0.26 (0.00, 1.24) 0.32 ± 0.27 (0.01, 1.19) .02* .09 .03* .99 AofE −1.85 ± 6.04 (−29.38, 16.82) −2.79 ± 7.04 (−39.30, 4.89) 1.43 ± 4.98 (−13.82, 15.68) .002* .99 .01* .003* |AofE| 2.80 ± 5.66 (0.0, 29.38) 3.49 ± 6.7 (0.0, 39.30) 3.10 ± 4.14 (0.0, 15.68) .82 .99 .99 .99 MofE −0.04 ± 0.24 (−0.66, 0.67) 0.17 ± 0.37 (−0.50, 0.93) 0.13 ± 0.45 (−0.70, 1.0) .004* .01* .03* .99 |MofE| 0.16 ± 0.18 (0.0, 0.67) 0.30 ± 0.26 (0.0, 0.93) 0.35 ± 0.30 (0.01, 1.0) <.001* .008* <.001* .99

AofE = angle of error; CI = correction index; DV = difference vector; IOS = index of success; MofE = magnitude of error; SIA = surgically induced astigmatism; TIA = target-induced astigmatism

Results are expressed as mean ± SD (range)

*Statistically significant

aMultiple correction was performed by Bonferroni correction

Higher-Order Aberrations

The changes in corneal aberrations in the 3 groups at 6 months postoperatively were shown in Table 3. The root mean square (RMS) of higher-order aberrations (HOAs) and coma was significantly increased postoperatively in each group (Table 3). The corneal spherical aberration exhibited a significant increase compared with the preoperative level in the FS-LASIK and transPRK groups (both P < .001), whereas it did not change significantly in the SMILE group (P = .32; Table 3). The trefoil increased significantly in the SMILE and transPRK groups (P = .007 and P < .001, respectively), but not in the FS-LASIK group (P = .66). The change in total RMS HOAs, spherical aberration, and coma was significantly increased in the FS-LASIK group, followed by transPRK and SMILE groups at 6 months (total RMS HOAs, 0.93 ± 0.43 vs 0.75 ± 0.22 vs 0.67 ± 0.21, P < .001; spherical aberration, 0.40 ± 0.30 vs 0.34 ± 0.22 vs 0.21 ± 0.14, P < .001; coma, 0.67 ± 0.43 vs 0.45 ± 0.21 vs 0.47 ± 0.24, P = .001; Table 3).

Table 3. - Comparison of corneal aberrations of eyes that underwent SMILE, FS-LASIK, and transPRK Parameter RMS HOAs Spherical aberration Coma Trefoil SMILE FS-LASIK transPRK P value SMILE FS-LASIK transPRK P value SMILE FS-LASIK transPRK P value SMILE FS-LASIK transPRK P value Preop 0.38 ± 0.08 0.41 ± 0.20 0.40 ± 0.11 .49 0.19 ± 0.08 0.23 ± 0.16 0.20 ± 0.09 .24 0.21 ± 0.11 0.21 ± 0.16 0.22 ± 0.10 .82 0.12 ± 0.06 0.15 ± 0.10 0.15 ± 0.08 .10 Postop (6 mo) 0.67 ± 0.21 0.93 ± 0.43 0.75 ± 0.22 <.001* 0.21 ± 0.14 0.40 ± 0.30 0.34 ± 0.22 <.001* 0.47 ± 0.24 0.67 ± 0.43 0.45 ± 0.21 .001* 0.16 ± 0.10 0.16 ± 0.10 0.23 ± 0.13 <.001* P value (vs preop) <.001* <.001* <.001* .32 <.001* <.001* <.001* <.001* .002* .007* .66 <.001* ∆ (preop vs postop) 0.28 ± 0.20 0.52 ± 0.40 0.34 ± 0.22 <.001* 0.02 ± 0.14 0.17 ± 0.27 0.15 ± 0.20 <.001* 0.26 ± 0.25 0.46 ± 0.44 0.23 ± 0.22 .001* 0.04 ± 0.11 0.006 ± 0.12 0.08 ± 0.13 .011*

∆ = change; RMS = root mean square

Results are expressed as mean ± SD

*Significance was set as P value less than 0.05

aMultiple correction was performed by Bonferroni correction

DISCUSSION

Less is known about whether SMILE without an automated iris registration and cyclotorsion control could perform as well as FS-LASIK or transPRK for astigmatism correction. In this study, we evaluated the surgical outcomes of SMILE, FS-LASIK, and transPRK for myopic astigmatism. All 3 types of refractive surgery performed well to correct myopic astigmatism during the 6-month follow-up visit, with comparable postoperative UDVA, CDVA, SE, efficacy, and safety index. Notably, the postoperative cylindrical error was smaller in the SMILE group than that in the FS-LASIK and transPRK groups, indicating that the SMILE was more accurate for astigmatic magnitude correction than the other 2 groups. Regarding axial misalignment, there was no significant difference in the absolute value of AofE among the SMILE, FS-LASIK, and transPRK groups (P = .82), indicating that the SMILE was comparable with FS-LASIK and transPRK in axial alignment.

Comparing the effectiveness of LASIK and SMILE for the correction of astigmatism has been a controversial topic in recent years.7–11 Zhang et al. did not find significant differences in the postoperative cylinder between eyes having LASIK and SMILE, with preoperative astigmatism ranging from −1.00 to −4.50 D.8 By contrast, Chan et al. compared FS-LASIK and SMILE with a mean preoperative cylinder of 1.37 ± 0.98 D and 1.08 ± 0.71 D, respectively.9 After 3 months, they found FS-LASIK favorable. Of interest, a previous study presented that the postoperative cylinder was not significant between corneal marking optimized SMILE and FS-LASIK groups (P = .20), with −0.19 ± 0.23 D (−1.0 to 0) and −0.12 ± 0.26 D (−0.75 to 0.50), respectively.12 In addition, a retrospective contralateral study showed that SMILE with a triple centration technique was comparable with corneal wavefront-guided transPRK (P = .299).13 In the current study, we specifically compared astigmatic correction among SMILE, FS-LASIK, and transPRK groups. The residual cylinder was significantly smaller in the SMILE group than that of FS-LASIK and transPRK groups (−0.23 ± 0.25 D vs −0.40 ± 0.28 D vs −0.42 ± 0.32 D, P < .001; respectively). The better outcome of astigmatism magnitude correction for SMILE could be explained by the following reasons. First, the head position was aligned with a cross-axial method in SMILE, which has been proven to effectively reduce residual astigmatism.14 Second, the fixed cutting pattern of transPRK is a central corneal epithelial thickness of 55 μm and a peripheral corneal epithelial thickness of 65 μm. If the distribution of individual corneal epithelial thickness does not conform to this fixed cutting pattern, 0.25 D will be introduced in the correction process, on average, up to 0.63 D of astigmatism.19 Third, flap creation can induce astigmatism. Previous studies reported that LASIK had a mean of 0.24 D of induced astigmatism.20,21 Fourth, patients with high astigmatism tended to have undercorrect results in both SMILE and LASIK.9,10 The higher proportion in FS-LASIK and transPRK groups may increase residual astigmatism postoperatively.

As a vector parameter, both the magnitude and axis are crucial for obtaining satisfactory surgical outcomes in patients with myopic astigmatism.6 For FS-LASIK and transPRK, the Schwind system allows for both static and dynamic cyclotorsion control during surgery, compensating for preoperative errors by incorporating topography-guided data into the laser system. Meanwhile, in SMILE, the lack of a static cyclotorsion detection and compensation system for distortion for the docking of eyes might cause a shift in the treatment axis.22 In the current study, the head position was aligned with a cross-axial method to minimize the axial misalignment in SMILE, and the astigmatism axial correction was comparable among SMILE, FS-LASIK, and transPRK groups.14 Respectively, 53 (82.8%), 36 (85.7%), and 37 (77.1%) eyes achieved an AofE within ±5 degrees (P = .55). No significant difference was observed in the absolute AofE value (|AofE|) among groups (2.80 ± 5.66 vs 3.49 ± 6.7 vs 3.10 ± 4.14, P = .82). However, the absolute value of MofE (|MofE|) was significantly smaller in the SMILE group (0.16 ± 0.18) than that in the FS-LASIK (0.30 ± 0.26) and transPRK (0.35 ± 0.30) groups (P < .001), which was consistent with the result of residual cylinders in this study. Therefore, our